Jan 22, 2024 Monday Deep Dive - How AI Transforms Art, Music, and Your Imagination

Description

Chapters

Transcript

0:03 all right my little champy let's flatten
0:05 that [ __ ]
0:08 [Music]
0:25 [Music]
0:29 out
0:30 [Music]
0:35 you're feeling nervous AR you
0:38 Bo with a quiet
0:41 voice and impeccable
0:45 [Music]
0:49 style don't ever let them steal your
0:52 joy and you gent way keep there for a
0:59 while they can kick dirt in your face
1:03 dress you down and tell you that your
1:05 place is in the middle when they hate
1:07 the way you
1:09 [Music]
1:12 shine I see you tugging on your
1:16 shirt trying to hide inside of it hide
1:20 how much it
1:22 [Music]
1:24 hurts let them laugh while they
1:29 care
1:31 let
1:32 sp Let stagger in
1:37 the I've been to the movies I've seen
1:41 How It
1:43 Ends and the
1:45 jok oh
1:48 [Music]
1:59 man
2:01 [Music]
2:05 yeah that was a good old action champy
2:09 likes that Brandy
2:11 Carlile them minor
2:13 chords he loves
2:15 them and then
2:18 [Music]
2:20 the oh so good right champy
2:25 major minor 7 come
2:28 [Music]
2:29 on
2:32 [Music]
2:48 Silver Fox hey everyone this seems like
2:50 a cool place to hang out why not why not
2:53 hang out we got singing
2:57 dogs we got soap boxes I think I'm good
3:01 with the soap boxes for two days I think
3:02 we'll play with some stuff
3:04 [Music]
3:06 tonight but if anyone wants to wax
3:11 philosophic I'm happy to do that
3:13 too well what does he do here is he
3:16 playing guitar I thought we were talking
3:17 about artificial intelligence is this a
3:19 are we going to learn about machine
3:20 learning I want to learn about
3:22 embeddings what is the mathematical
3:24 theoretical fundamentals of machine
3:26 learning that's why I'm
3:29 here
3:36 [Applause]
3:37 [Music]
3:47 oh champy I know right sometimes you
3:51 just got to get loud you got to get
3:52 jiggy with it just got to go you just
3:55 got to go so we might talk about the
3:59 theoretic IAL mathematical underpinnings
4:01 of AI but probably not cuz I don't know
4:03 [ __ ] about
4:08 that but if you want to figure out what
4:11 this AI stuff is why it's a big
4:15 deal how's it going to affect all of our
4:19 Lives I have no [ __ ] clue but I know
4:22 it's going
4:23 to and if you want to you can make money
4:27 with we can talk about that
4:32 um I can show you cool stuff I can
4:34 answer questions so what my request
4:37 would
4:38 be I think that was Kevin Kevin Vasquez
4:41 reached uh level 19 very impressive I
4:45 can't wait till somebody trains an AI to
4:46 translate animals I know H seriously
4:49 divid 3
4:50 333 I like we're not that far from that
4:54 I mean you know that right like if if
4:56 you if you're paying attention to this
4:58 AI
4:59 stuff the way these large the the way
5:01 the Transformer works the way they can
5:04 they can basically just do this [ __ ]
5:06 with any
5:07 data and so once they can sort of crack
5:10 the code they're going to be able to
5:13 crack the code I it's going to be
5:19 amazing so if you would share the
5:28 live
5:33 [Music]
5:43 hey Sherry D Lori good evening Dr bod
5:47 good to see you David I'm looking
5:50 forward to it has anybody given it a
5:53 large data model of just animal noises I
5:56 would assume so so the thing about the
5:59 trans forer and all these all these
6:02 tools is that um they've been out for a
6:05 lot of years like the transformer was
6:07 invented in 2017 by
6:11 Google and
6:13 um open AI was one of the companies that
6:16 recognized oh we could you know we
6:18 should throw as much data and compute
6:20 power at this as we
6:22 can and so they sort of went down the
6:25 route of you know let's let's get as
6:27 much of the public internet as we can
6:28 and dump that in this thing
6:30 um and so I would imagine that there's
6:33 researchers doing it yes I I don't know
6:35 of any specific
6:37 projects but they're doing things now
6:39 that are you know in the neighborhood
6:41 where you know they're reading brain
6:43 scans fmis of people watching videos and
6:47 recreating the videos or recreating the
6:50 images that they've looked at or
6:52 allowing them to walk again there's a a
6:54 paraplegic guy in Italy where they
6:57 scanned his brain and told him to think
6:59 about moving his legs and then they
7:01 bluetoothed it around his injury into
7:04 electrodes at the base of his
7:06 spine and now he's walking I was like
7:10 [ __ ] so it's all just different it's
7:13 it's all pattern recognition
7:20 right so I would think in the
7:24 next 2 3
7:27 years and we'll probably see you know
7:30 something remarkable in the we can talk
7:33 to animals the Dr doitt thing the Dr do
7:36 little AI we'll probably see something
7:38 sooner than later where someone does
7:40 something
7:40 [Music]
7:48 remarkable I normally don't get to just
7:51 strum chords like this he normally is
7:52 singing up a
7:58 storm
8:04 [Music]
8:19 oh in westly
8:25 Direction this car is my train
8:31 I've been driving I've been
8:34 wondering homeschool mom watching for
8:37 how to work together as a fam on
8:39 Creative projects cool books
8:42 illustrations awesome awesome Reva James
8:44 so so yeah so you know so tonight it be
8:47 fun for you here was the idea I
8:50 had so I'm going to talk about this
8:52 thing called the AI Salon um which is is
8:55 really kind of the community extension
8:57 of this Channel or this channel is an
8:59 extension of that
9:00 that and in it is this list of all these
9:03 tools that you can play with um that I
9:06 that are recommended that that it's a
9:08 list that I curate personally and it's
9:10 and you're like well who are you what
9:11 are your qualifications uh yeah uh uh I
9:15 I don't know if you knew this Mr Shannon
9:16 but uh I I'm a homeschooling professor
9:19 and uh I would really like to know your
9:21 credentials and your qualifications uh I
9:23 have none
9:26 but I have been playing with these tools
9:29 pretty hot and heavy for about 2 years
9:31 and so one of my pet peeves is shitty
9:35 janky software and all of these tools
9:38 right now are shitty janky software but
9:40 the list that I maintain is at least
9:44 stuff that I think is important and is
9:47 not going to be a waste of your time and
9:50 and so what I thought I'd do tonight
9:52 rather than do a deep dive on any single
9:54 tool what I thought might be cool is
9:56 just kind of go down through the list
9:58 and do one cool thing from you know
10:02 tools that I think are interesting and
10:04 so it there'll be some image ones in
10:06 there there'll be some video ones in
10:08 there we'll do some music we'll we'll
10:10 write some songs we'll have ai write
10:12 songs you know where you don't have to
10:14 learn how to play guitar to write a
10:16 [Music]
10:23 song for those of us that are kind of
10:26 hack musicians that are just not really
10:27 that good at it but we like it
10:30 this AI stuff is going to allow us to do
10:32 some cool things I do
10:34 [Music]
10:58 think
11:00 [Music]
11:09 chat GPT told me knock knock knocking on
11:11 heaven's door is the is a good first
11:13 song to play on guitar with instructions
11:15 that's
11:16 [Music]
11:24 cool s in this lonely
11:26 [Music]
11:28 town
11:30 wonder when things are going to
11:33 [Music]
11:39 change dreaming my life
11:43 away seems he dreams have turned to a
11:47 bunch D
11:51 classs getting my nerve
11:54 up but my past is pulling me
11:58 down
12:02 wondering how
12:04 long this black sheep going to stick
12:13 around somebody told me once before he
12:16 said you can never go home
12:20 again won't you
12:22 leave say a thing just to steer me away
12:27 yeah from the truth of who I am and what
12:30 I
12:31 believe so I thanked him for his two
12:34 cents with a
12:36 handshake and some
12:38 sympathy
12:40 yeah packed up my blue
12:43 jeans headed for this big
12:48 price of my
12:50 [Music]
12:51 freedom
12:53 byebye Black
12:56 Sheep to the black sheep of the family
13:01 [Music]
13:05 byee oh that means so very much to me
13:10 yeah
13:12 byebye Black
13:14 Sheep to my friends and my
13:19 family
13:21 [Music]
13:23 byby going off and set my
13:26 soul set it free
13:32 [Music]
13:38 all right enough torturing
13:40 [Laughter]
13:42 y'all I like
13:45 singing I started a band once called
13:47 thin the herd now you know why the
13:50 numbers are just dropping old man
13:53 singing poorly let's go let's get on out
13:55 of here all
13:58 right
14:01 oh by the way that was oh that was uh
14:05 Martin ston if you don't know that song
14:07 Black
14:08 Sheep black black black sheep keep
14:11 singing yeah
14:12 exactly I enjoy it all
14:16 right oh Kyle Shannon you have to join
14:18 me in my Sunday song sessions yeah we
14:20 should do that Corey that's that's
14:22 [Laughter]
14:24 awesome oh it'll give me an excuse to
14:26 learn some new songs at least you know
14:30 what did I I learned one the other day
14:31 what did I learn oh the new uh the so
14:34 the Billy ish song from the Barbie movie
14:37 she's going to win an Oscar for that man
14:39 that that woman is she is just
14:47 something I you know one of the things
14:50 that I really dig about Billy
14:53 ish
14:58 is
15:00 there's something so unassuming about
15:02 the music and the songwriting it's
15:03 produced really well her brother
15:05 produces it really
15:06 well but it's so like it it is so like
15:10 it was recorded in a bedroom because it
15:12 was recorded in a bedroom
15:15 right and yet there was some there was
15:17 something about it that it was just so
15:19 different and it it just captured
15:21 everyone's imagination and it completely
15:23 took
15:24 off and it just it like broke it broke
15:28 all the norms right it broke all of
15:30 here's how you get a song into the
15:32 marketplace here's how you you know get
15:36 it
15:37 going you know then she the two of them
15:41 sweep the gramys like just just
15:44 nuts and it's like I I feel
15:51 like there's a there's a lot of hand
15:53 ringing about AI like like what's gonna
15:56 you know if if if these tools just make
15:58 all this
15:59 content you know that it's just all G to
16:02 be shitty
16:04 content no it won't
16:06 because whether it's someone sitting in
16:09 their bedroom with Logic Pro or it's
16:12 someone using an AI tool or it's someone
16:14 using a stupid
16:16 ukulele to to write a song that no one's
16:19 heard
16:20 before like there's still going to be
16:23 there's still going to be people
16:26 that twist and bend the
16:29 tools in a way that just creates
16:32 something completely unique I I don't
16:34 know have you ever seen that the video
16:36 with Fel Williams when he discovers M
16:39 Maggie
16:40 Rogers where so Fel Williams comes into
16:43 an NYU like master class and just
16:46 listens to all the students songs and
16:49 then he listens to Maggie Rogers song
16:52 and you can just see his face he's like
16:54 he's like you know he has a Kevin
16:55 mallister moment right he's
16:57 like and the song ends and he has I have
17:00 zero notes for you
17:04 because what you
17:06 did people can love it they might hate
17:09 it but they can't deny that it is it it
17:13 is what it is and there's nothing else
17:14 like
17:15 it and so I don't think that goes
17:19 away when our tools get more
17:22 sophisticated we're still human beings
17:25 and someone's going to do something with
17:27 AI That's going to absolutely [ __ ]
17:29 move
17:30 people and then someone's going to do
17:32 something with just a guitar and no AI
17:35 that's absolutely going to move people
17:38 and those two artists are going to
17:41 coexist in the same world and shangra La
17:45 there there will be fireworks and and
17:48 dancing llamas for for as far as the eye
17:50 can see everything will be fine with the
17:53 world so there's a lot of consternation
17:55 about how AI is going to ruin everything
17:58 and it's not because humans are still in
18:00 the
18:01 mix Sherry D changing my career on that
18:05 bet we're going to have more music and
18:06 beautiful Creations on their way yeah
18:10 exactly and listen the fact that you can
18:13 that we're in a world right now Sherry D
18:15 where you can just say you know what I
18:17 think I'm going to try and change my
18:18 career I'm going to lean into this thing
18:21 and I'm going to get good at this when
18:23 no one else even knows what it is so
18:26 that by the time this stuff gets really
18:28 good I'm going to be flipping masterful
18:30 at it like I think that's a good plan
18:32 why not I think it's [ __ ]
18:35 awesome all right so what are we going
18:38 to do here tonight so if you if you uh
18:41 if you're new here my name is Kyle
18:42 Shannon this is the AI learning lab we
18:45 do things different around here I don't
18:46 know what whatever doesn't matter it's
18:50 one of the nicknames for the channel is
18:52 Chad
18:56 add not because that was a clever name
18:59 but
19:01 because I occasionally finish a
19:06 sentence yes I'm drinking coffee at
19:10 night
19:16 ah my GPT Imaginarium has made some
19:19 stunning stories very
19:26 cool did you get a haircut no I just wet
19:29 it I just did [ __ ] with it I I need a
19:32 haircut really bad it's it's just if I
19:34 turn around you'll see it's just some
19:37 mangled mess back there I would think
19:40 but uh I'll get a haircut here at some
19:43 point just have them trim up all the
19:46 sprouting
19:48 hairs look good tonight well thank you
19:51 thank you thank you thank you you took a
19:53 shower today yes yes I occasionally
19:56 bathe um what about a robot that does a
20:00 haircut I'll wait for V3 of the um the
20:04 scissor
20:05 wielding robot I I'll wait for the third
20:09 version of that I'll absolutely I don't
20:12 I I would I will absolutely get my
20:14 haircut by a robot I don't give a [ __ ]
20:16 um all right join us for more for for
20:20 wait join us for more of this for free
20:23 and pictures at the AI uh the salon.
20:27 okay so so let yeah that's a good point
20:31 um let's let's show people where to go
20:35 so go there the salon.
20:39 a the first link there is going to take
20:42 you to um to the AI Salon Community
20:46 you're like well what's the AI Salon
20:48 Community who are these people what is
20:50 this community how much you charging for
20:52 this community oh is it a community is
20:54 this like a multi-level marketing thing
20:55 is that what this is you're so you sing
20:58 us a few song songs and now you're
20:59 trying to upsell us is that how this
21:01 works pretty much you know but without
21:04 the
21:05 lisp
21:07 um but it's free so there's no
21:10 upsell so uh this is a group I started
21:14 the week after chat GPT came out and
21:17 it's it's a really cool group and uh so
21:20 if you go
21:21 there wait H someone's going to get me
21:24 kicked off here dang
21:26 it that the show in tell Channel had
21:28 some some Saucy Vixen on there that
21:31 happened to me one night and they they
21:33 they said you are perab ban from Tik Tok
21:35 I was like
21:37 no
21:38 okay all right so when you go to the
21:43 salon. and you click on the first link
21:45 it's going to take you to this site
21:47 which is a it's a mighty Network site
21:49 Mighty networks is just a a community
21:52 management software and this is the AI
21:54 salon and so when you first get there
21:56 check out the welcome to the AI Salon
21:58 section
21:59 and it'll talk about who we are and what
22:00 we do a community of extraordinary
22:02 adventurers exploring the AI unknown so
22:05 basically this is a group okay here's
22:07 the deal if you're one of those people
22:09 in your life that you're like hey Bob
22:10 you should try out Ai and Bob's like
22:13 yeah go screw you Jimmy I don't want you
22:15 dumb AI your AI is dumb Jimmy and you're
22:18 like but but Bob AI is really cool and
22:20 it could free you from all of your
22:22 life's woes and he's like yeah Jimmy
22:25 it's evil you're going to you're cting
22:28 the de if if that's
22:31 you and no one in your life wants to
22:33 hear it anymore we're your
22:36 people and if it's not you if you're
22:38 curious about this AI stuff and you're
22:40 like how do I learn about this like
22:42 where do you begin and what I can
22:43 promise you right now is with the
22:45 generative AI stuff with chat GPT and
22:48 all the stuff that people are talking
22:49 about making pictures and making music
22:51 all this sort of
22:52 stuff there's no real single place you
22:55 can go to learn it because it's moving
22:59 so fast that everything's changing
23:01 literally like week to week just things
23:03 change features get added the thing it
23:05 couldn't do last week now it can do and
23:07 then it can do it times 10 it's just
23:09 crazy and so that's what this group is
23:11 and so we talk about that why we do what
23:13 we do we talk about our values values
23:15 like curiosity and generosity like like
23:18 as you learn stuff about AI share stuff
23:21 about AI not only share your images but
23:24 share the prompt that made it up
23:25 especially if you did something
23:26 interesting with it Danielle Be not
23:29 Afraid from in in here she's one of the
23:30 Irregulars in the mods um every week she
23:33 puts out a Mad Lib um image
23:39 generation like little contest thingy
23:41 she like writes up this Mad Lib and it's
23:43 like this really interesting thing where
23:46 everyone in the community can basically
23:47 just swap out words and make different
23:49 variations of the same thing and it's
23:51 like that kind of stuff you you learn so
23:54 much and and it it requires a level of
23:56 generosity not being stingy with what
23:58 you're learning which at this stage of
24:00 an industry is really really important
24:03 because there are no rules right now you
24:05 can't get a degree in this stuff um
24:08 exploration is one of the the values
24:10 collaboration's one of the values
24:11 inclusion people are going to be coming
24:13 from everywhere and
24:15 empathy empathy last night uh we had we
24:18 had RJ was in here
24:22 and was very vocal about how against AI
24:26 he was and how scared he was of it and
24:28 he actually had the courage to say he
24:31 was
24:32 scared and hang out and have the
24:35 conversation and we talked a lot about
24:37 it and and you know it's one of those
24:39 things where people are scared right now
24:41 they're angry right
24:43 now and if you're willing to sit in okay
24:47 I can understand why you might be scared
24:49 I can understand why you might be pissed
24:50 off about this want to see something
24:54 cool and get people over that hump of
24:56 just being willing to try it remarkable
24:59 things happen really remarkable things
25:01 like we had a really fun really fun
25:02 conversation last night I think largely
25:04 because of the fact that he was willing
25:07 to come in here and say what was going
25:09 on for him but then also just have the
25:12 conversation right so not just come in
25:14 here and Troll and say ai's evil and I
25:16 hate it but say ai's evil and I hate it
25:18 oh and by the way it's because I'm
25:19 scared of
25:21 it and then he and then he was willing
25:23 to have the conversation right so that's
25:26 that's one of the values there so
25:29 so that's the community so if those
25:30 values feel like yeah that's those are
25:32 great then then these are likely going
25:33 to be your people if those values sound
25:35 to you like you know Nails on a
25:37 chalkboard then this is probably not
25:39 your group that's totally cool too right
25:42 but that's what this group is and so as
25:44 we scale it I want to make sure that we
25:46 retain that I appreciate your generosity
25:48 toward RJ last night oh thank you very
25:49 much Source
25:52 Camp thank you yeah I
25:55 mean you know and I think in the end I
25:58 think in the end he got something out of
25:59 it maybe just maybe the
26:03 possibility that maybe he'll give it a
26:05 shot and in giving it a shot he won't be
26:07 sitting on the sidelines right and
26:10 that's that's that's the only thing I'm
26:13 really going for with this channel is I
26:15 want people I don't give a [ __ ] what you
26:17 think about AI I just want you to
26:19 experience it and and and and get
26:21 curious about it because this stuff's
26:23 coming whether you want it or not um
26:26 okay so so when you first get there
26:28 check out the about the the the salon
26:31 section and again if if you're just
26:33 stumbling in and you don't know what the
26:34 hell I'm talking about go there the
26:36 salon. a click on the first
26:39 link and that will take you to this site
26:42 and then after you check out what we're
26:43 all about I want you to go to introduce
26:45 yourself so in the community section the
26:48 first little link here is introduce
26:50 yourself
26:52 um my request is that you use your real
26:55 name when you sign up and that you
26:57 introduce yourself and as [ __ ]
26:59 terrifying as that is well I don't want
27:00 to introduce myself everyone's going to
27:02 make fun of me I don't know anything
27:03 about AI you all are some kind of
27:05 experts or some kind of thing like
27:07 that no no we're all everyone everyone
27:12 on the planet right now is trying to
27:14 figure this stuff out nobody has their
27:15 head around it nobody I promise you so
27:18 just come in here and tell us where you
27:20 are on the journey like what do you do
27:22 have you been around for 60 years in a
27:24 single profession and now you're just
27:26 curious you're retired and you're just
27:27 curious about this stuff are you like 30
27:29 years into machine learning algorithms
27:31 and all of a sudden generative AI is
27:33 changing everything are you like
27:36 literally just like I stumbled upon Tik
27:38 Tok and I thought it was fat Alec
27:40 Baldwin and I stuck around for a little
27:42 bit and here I am like that's cool
27:45 too and then and then people will
27:48 welcome you in and you'll you'll kind of
27:49 find your way the other thing that's
27:50 really fun in in the community section
27:52 is the show and tell Channel which that
27:55 becomes kind of like a
27:58 shangrilla the the all the llamas
28:01 singing and dancing so I you know
28:03 earlier if you weren't here earlier I
28:05 mentioned shanga someone went off made
28:07 that image and shared it in showand tell
28:09 so show and tell while these lives are
28:11 going on become kind oh that's that's
28:13 really
28:14 cool dolly is quickly getting better at
28:16 spelling AI learning lab lots of
28:18 practice that's awesome yeah I've
28:20 noticed that it it is getting better at
28:22 certain words like that we should really
28:24 work on um Irregulars and IR irrelevant
28:28 better to be irregular than irrelevant
28:30 it still can't get that one for [ __ ] but
28:32 that's a really cool one um so so this
28:35 channel becomes kind of like this this
28:37 um this jam session while while these
28:40 lives are going on and then in between
28:42 these lives you know people are sharing
28:44 songs and they're sharing gpts and
28:46 they're sharing best practices and all
28:48 sorts of things like that um there's a
28:50 news channel there's a water cooler
28:51 where you can chat there's also
28:55 um um guilds so there's an AI 101 Guild
28:59 an Art Guild a writer Guild business
29:01 Guild mechanics Guild if you want to
29:03 geek out on how this stuff works
29:04 Healthcare video education and each of
29:07 these guilds is is a a subc community
29:10 within the larger AI salon and so if you
29:13 want to do a project or if you're really
29:15 passionate about education Stacy and and
29:17 Matt Mr K from here on Tik Tok just
29:20 started um the education Guild they're
29:23 Educators they're super passionate about
29:25 Ai and doing it right right and so you
29:27 can you can be in a community where
29:29 where you can talk about those specific
29:31 things okay so here's what I figured
29:33 we'd do tonight in the community section
29:36 there's also this page called AI tools
29:38 and
29:39 resources and this is a list that I
29:41 maintain the cool thing about this
29:43 channel is you can also add your
29:44 favorite tool so there's a big long list
29:47 here of community suggested tools so one
29:51 of the the the you know one of the
29:53 values of generosity or one of the ways
29:56 you can express generosity is if you
29:58 find a cool tool share it right and so
30:01 you can do that but this AI Salon uh
30:05 recommended tools this is a list that I
30:07 maintain and so in it I talk about chat
30:11 GPT obviously if you're just getting
30:14 started with AI like if you're just here
30:17 because like you heard my dog singing
30:19 you're like wait what is this and you've
30:21 never used chat GPT go right now to
30:24 chat. open.com right there and just
30:28 start playing while I ramble you just
30:30 play are they going to name the chat GPT
30:34 like Siri Alexa it could make it more
30:38 relatable um I don't you know what's
30:40 interesting Angelique I I
30:45 um well a I agree with you and I didn't
30:48 even consider that but Sam Alman at
30:50 Davos this week he corrected himself
30:53 twice he called it um GPT 5 and then he
30:57 said or whatever we end up naming naming
30:59 it like maybe at some point we need to
31:02 get out of you know just call like he
31:04 goes I don't want there to be a chat
31:07 gp22 so so yeah I would agree it would
31:11 be nice of them to do that like Pi's got
31:13 pie claude's got Claude and chat has gp4
31:19 right it's so it's I I think I think it
31:21 could it could use for a little branding
31:24 love I absolutely agree with you uh Eli
31:28 Peta hello from Israel thank you I
31:30 learned from you so much thank you so
31:31 much and welcome welcome
31:34 welcome
31:36 um so so so what what I figured we'd do
31:40 is we'll go through this list um if
31:43 you're new and you haven't played with
31:44 chat GPT go do that and then on this
31:47 list so there's there's you know not
31:50 just chat GPT but what's in chat gp4
31:53 like if you pay the 20 bucks a month
31:54 what do you get from it it's a lot um
31:57 and I talk about it in the channel a lot
31:59 so we can talk about that but what I
32:01 figured we'd do is we'll pick one thing
32:03 from chat GPT to play
32:05 with then we'll pick some other tools
32:07 I've got Claude Bard perplexity um
32:10 co-pilot from Microsoft Pi grock which
32:15 I'm not going to demonstrate because I
32:16 think it's a piece of poop um llama 2 is
32:19 really a whole sort of world in and of
32:21 itself not not anything really
32:23 interesting to play with there right now
32:25 um that that
32:28 that I know about but that's a whole
32:30 world in of itself but then I've got
32:33 image creation tools on here I've got
32:35 video generation tools um two different
32:37 you know three different kinds of video
32:39 generation tools music and audio and
32:41 then 3D objects so I figur might be fun
32:43 just sort of blast through this list
32:45 tonight and uh and and play so why don't
32:50 we do it like this why don't you go
32:52 to
32:54 um the
32:56 salon.
32:58 and and sign up for the salon if you dig
33:00 what we're about and then you know come
33:02 look at this list and then just shout
33:05 something out if you want me to play
33:06 with it and I'll go play with it or if
33:07 anyone has any questions right now about
33:10 tools we can do that or I'm happy to
33:12 dive in and
33:14 talk more about philosophy and the
33:16 future of AI and is it really magic or
33:20 is it just
33:21 technology is it sensient perplexity is
33:25 my new friend lately so retro I you I
33:28 think you and I talked I think you're
33:29 the one that told me um if you bought a
33:33 rabbit R1 you get a year of perplexity
33:36 Pro for free so the rabbit R1 was 200
33:40 bucks and perplexity Pro for a year is
33:43 200 bucks so I got one of them for
33:48 free and as soon as I have that thing in
33:50 my grubby little hands I'll demo the
33:52 [ __ ] out of it for you guys hopefully it
33:54 doesn't suck all right with gpt3 are
33:57 they still still making changes SL
33:58 updates or now just weekly upgrades to
34:02 gp4 I don't know if they're doing weekly
34:04 upgrades to
34:06 gp4 um in terms of the underlying model
34:09 they said they're going to the as far as
34:10 I know the underlying data in gp4 right
34:13 now is up to April of
34:17 2023 Sam Sam mman said they're going to
34:19 try to increase how they update that but
34:23 I don't think I I certainly don't think
34:24 they're on weekly week the weekly stuff
34:26 they're doing is just like you know
34:27 tweaking the tool and things like that
34:29 so I don't know what there I assume chat
34:33 GPT
34:35 3.5 has been kind of stati IED let
34:39 me I don't that I don't know I I don't
34:48 know rabbit price was worth it to get
34:50 perplexity for free for a year yeah
34:52 exactly and for the cool little
34:54 industrial design I didn't realize it
34:57 the guy that started rabbit R1 he
35:00 created this really cool music interface
35:03 or it was some sort of computer
35:04 interface I think it was a music
35:06 synthesizer so the rabbit R1 was
35:08 designed by teenage engineering which
35:10 makes
35:11 synthesizers I'm like oh that's a really
35:13 interesting choice of industrial
35:15 designer well he made a like a a
35:18 synthesizer control a music controller
35:20 prior to this that was very much like
35:23 the R1 it was just like a a panel of
35:25 LEDs that you inter acted with and then
35:28 it the LEDs did cool [ __ ] on the panel
35:31 so interesting interesting uh
35:36 company made an eBay Lister GPT today to
35:39 help list my many items very cool love
35:42 that that's very cool I want to hear
35:45 your dog sing I think he might be done
35:47 he he sang for a bit hey
35:52 champy oh that's Out Of
35:55 Tune here he goes
35:57 [Music]
35:59 [Applause]
35:59 [Music]
36:03 [Laughter]
36:06 [Music]
36:24 [Laughter]
36:26 oh
36:29 yeah oh
36:36 man go up on show and tell what's
36:39 happening on show and tail show and
36:45 tail this image captures the essence of
36:47 diversity and unity various hands
36:49 holding each other and reaching toward
36:51 light symbolizing empowerment
36:52 togetherness in the context of AI
36:54 learning that's cool what's what's
36:56 what's really amazing about this
36:58 community is and and I listen I hope as
37:01 this thing scales um we can maintain
37:04 this it is so important right now we're
37:08 we're so early in AI as much as AI is
37:10 overhyped you'll see down in one of the
37:12 corners in the bottom I created a video
37:14 series called underhyped
37:17 because as as much as people are talking
37:20 about Ai and how sick of it you are like
37:22 we ain't seen nothing yet we ain't seen
37:25 nothing yet like we've barely just begun
37:27 barely anyone knows about this and so
37:30 when you're that early in an industry
37:33 what you need is is collaboration and
37:37 cooption or you know uh
37:40 cooperation um rather than competition
37:43 because everyone's trying to figure it
37:45 out as things stabilize so as we get to
37:48 three to five years out and this these
37:50 you know whatever AI is going to become
37:53 it kind of it you know it it will it'll
37:55 hit a plateau at some point
37:58 you know probably in that time frame
37:59 where we at least know what it is now
38:01 right and at that point then we'll start
38:04 to figure out oh now I need to be
38:05 competitive and this and that but in
38:07 these early stages of a transformation
38:09 like this this is the way to be and
38:11 that's what this community is about so
38:13 and that's that's what this community
38:14 really is it really is quite amazing
38:16 gptp is that's
38:21 great I hired RJ that's
38:25 great oh man I did a keyboard layout
38:28 designer to add buttons to GPT chat GPT
38:31 on a keyboard that's really cool very
38:35 nice wow that's very
38:37 cool nice the AI learning
38:40 lab learning
38:43 together so
38:46 cool all right so let's go let's go play
38:51 Let's go wait well before I do that let
38:53 me let me just look at comments down
38:55 here and see what we got going
38:59 can we upload word or PDF documents or
39:02 do we have to cut and paste it to learn
39:05 no you can upload them
39:07 so all right so let me show
39:11 you show you a couple of things
39:14 so let's go to chat
39:25 JPay
39:38 Angelique I believe everyone should
39:40 follow your lives it would change their
39:41 mindset oh thank you very much um yeah
39:45 go tell people to come come follow
39:48 actually so I kind of jokingly say this
39:50 every night but I'm I'm serious about it
39:53 do me a favor share this live with some
39:55 of your friends on Tik Tik Tok share it
39:58 with the friends who you know it will
40:00 make
40:04 cranky I don't want to hear about your
40:06 stupid AI anymore
40:08 Jerry send it to send it to send it to
40:14 Buck knowing that Buck's gonna give you
40:17 [ __ ] because because I think I think we
40:21 need more people we need more people
40:23 that are like SE skeptical or just done
40:26 with it
40:27 because I guarantee you they haven't
40:29 played with it anyway
40:31 okay so if you're if you're a chat GPT
40:34 one of the things that chat GPT is
40:36 really good at let me go I'm GNA go find
40:39 something
40:49 [Music]
40:51 here why did that not open
40:55 open virtue I bet it's up in the cloud
40:59 and I got to go get
41:03 it h that not
41:07 opening let's see do I have salmon
41:09 flying East in
41:25 here
41:31 what is going
41:39 on hold please we're having technical
41:43 difficulties what did you do Monday
41:45 night well I watched
41:48 some seemingly Pleasant older man try to
41:51 learn how to use a computer was very
41:55 frustrating
42:05 virtue ah okay now it worked all right
42:08 it had just crapped the bed okay so this
42:11 is a screenplay I wrote many many moons
42:14 ago called virtue it's kind of a
42:17 Manchurian
42:19 Candidate people control other people's
42:21 bodies have them commit you
42:24 know felonies
42:27 I don't know what you can say on Tik Tok
42:29 without being perab
42:32 banned I just had a workshop on using AI
42:35 for Effective youth
42:37 ministry at a youth worker conference
42:40 very cool uh so any kind of
42:44 communication any kind of human
42:47 expression um you know AI is going to be
42:49 able to do interesting stuff with it
42:51 anyway okay so one of the things that
42:54 people know about chat GPT is a Can it
42:57 can generate stuff right it's called GPT
42:59 generative pre-trained Transformer the G
43:02 uh stands for Generate or generative and
43:05 it can generate stuff so you can say
43:07 write me
43:09 a
43:11 poem about
43:14 um I don't know uh light
43:21 refraction like TS
43:25 Elliot and then it makes a poem right so
43:29 it can do that kind of stuff and I think
43:30 you know doing these kind of parlor
43:32 tricks where you just have it do cool
43:33 [ __ ] [ __ ] like that I think a lot of
43:35 people see it like that and they think
43:37 that's what it can
43:39 do um I'm by the way I'm just in chat
43:43 GPT 3.5 right now this isn't the fancy
43:45 chat GPT this is just normal chat GPT
43:47 free free chat
43:49 GPT and notice how fast it is um I can
43:53 also do things
43:55 like create a table
43:59 comparing TS Elliot
44:03 and um
44:09 Eminem and it will just do
44:14 that and so let's see what it said
44:17 here TS Elliott born in 1888 Eminem born
44:21 in 1972 modernist poet rapper songwriter
44:25 influential in modernist poetry
44:27 influential and Hip Hop and
44:29 music
44:32 um if you haven't really used chpt or or
44:35 thought about AI one of the things you
44:38 might think is like well this table must
44:40 exist somewhere in the world like it
44:41 just copy and pasted that stuff no
44:43 that's not the way it works it actually
44:46 generates this stuff out of nothing um
44:48 it also can generate code so I could say
44:53 um write me a python application
44:57 that lets someone choose between TS
45:00 Elliot or
45:02 Eminem comma asks them for a topic and
45:05 then writes a poem or a WAP for them
45:08 based on who they
45:15 chose and it will start writing your
45:17 code for you so like so it
45:21 can it can do a lot of
45:24 stuff and so if you haven't dug in and
45:28 and really started to push the
45:30 boundaries of what this is like if it
45:32 think about okay when when you look at
45:35 chat GPT it's just this sort of Google
45:38 like box right and so one of the one of
45:42 the things that people think when they
45:43 first see it is oh it's just like Google
45:46 it's just like fancy Google right
45:48 because it can sort of answer questions
45:50 if I asked
45:52 Google how to create a python
45:55 application
45:57 you know to write
45:59 poetry it would point me to 462
46:02 different sites that either want to
46:05 charge me to Learn Python
46:08 or are some weird ass sales funnel for
46:11 something else that you know the SEO
46:13 injected you know python coding lessons
46:16 in the front of and and it's this
46:19 incredibly frustrating thing and if I
46:20 wanted to create this application I
46:22 would have to go find a site where I
46:24 could learn python or go find a YouTube
46:26 video
46:27 and I would probably have to spend at
46:28 least a day probably a week maybe a
46:31 month depending on where you are with
46:33 this stuff figuring this [ __ ] out and
46:35 this just wrote the
46:37 code and if your question is yeah but
46:39 Kyle is the code any good well you can
46:41 run it you can throw it over and rep it
46:43 and run it one of the first things I did
46:46 in cat GPT 3.5 was I wrote a
46:50 sophisticated python application that
46:53 took in ask people a few questions and
46:55 then wrote five pieces of social media
46:57 content out of what they the topic that
46:59 they
47:00 chose I was doing sophisticated API
47:03 calls over to open AI generating the
47:05 content returning that I've never
47:08 written a line of python in my life
47:10 still to this day that was the first
47:12 thing I did was was wrote a python
47:14 application that functioned it blew my
47:17 flipping mind okay so that's just chat
47:19 GPT normal now one of the things chat
47:21 GPT is good at let me make a new a new
47:24 chat here is it's good at analyzing
47:27 stuff so I'm going to go back over to my
47:29 little
47:30 screenplay and I'm going to copy I don't
47:33 know a few pages
47:39 here there we go go up to there I'm come
47:42 over here and I'm going to go based on
47:45 this
47:51 text call it I'm going to hold down the
47:53 shift key do a couple of line breaks
47:55 then I'm going to paste it in and then
47:58 I'm going to go
48:00 um tell
48:05 me about
48:09 this and
48:12 so all that comes in oh oh it may
48:16 violate our content policy oh
48:19 yeah I picked the wrong screenplay it
48:22 starts out with a with a dramatic scene
48:25 where someone doesn't make it dang
48:29 it all right let me let me go a little
48:31 farther down the screen play I'll just
48:32 grab some let me let me grab some some
48:35 other oh this has got oh my God this has
48:38 got violence in it too I think I need
48:40 another screenplay hang on let me go get
48:42 S flying
48:45 East oh actually this is good I'm glad
48:47 this happened okay so what we just
48:50 experienced is when you're using chat
48:53 GPT especially when you're using
48:56 Microsoft co-pilot when you're using
48:59 Claude when you're using um pretty much
49:02 anything except llama 2 if you're using
49:05 an unrestricted version of llama
49:07 2 all of the commercial companies that
49:09 are doing
49:10 AI um put what are called Safety guard
49:14 rails into the into the tools right so
49:18 if I were to ask chat GPT to help me you
49:21 know do harm to someone or to animals or
49:24 to you know do something
49:26 that would hurt you know other people or
49:29 myself it it basically won't allow it
49:31 right so and E even with the new gpts
49:35 that that open AI lets you create it's
49:37 avoiding letting you create things in
49:39 financial services and medical advice
49:42 because they don't want people to get
49:44 hurt so what we just ran into is a
49:46 safety guard rail they can be really
49:48 frustrating because sometimes like say
49:51 if I've got a
49:52 screenplay and it has Topics in it that
49:55 are approp for a
49:57 screenplay it would be nice to be able
49:59 to do some of the stuff but you know but
50:02 that's just just know that that's that's
50:03 a possibility so let me go find another
50:05 screenplay here I got I got a bunch I've
50:08 written seven screen plays in my day so
50:11 I got some I got some stuff sammon
50:14 flying
50:15 East okay so let me go grab some of this
50:21 and I'm pretty sure this this doesn't
50:23 start in a way that it can't
50:26 go but let's see come over here we'll do
50:31 this I'm going to say based
50:37 on
50:39 on
50:41 this um tell me what is
50:48 happening
50:54 and uh so it says the provided text
50:58 appears to be the script for a movie or
50:59 screenplay it's divided into scenes and
51:02 and here's a summary of what goes on
51:07 and the it opens in Gig Harbor
51:09 Washington which is right jack heads to
51:11 a bait shop owned by Martha which is
51:14 right Jack has a voiceover conversation
51:16 with someone referred to as
51:20 snake and so so it tells us what's going
51:22 on so that's super cool so you're like
51:23 that's awesome let me do the whole
51:25 screen place so you come over here you
51:27 go command a copy the whole thing come
51:30 back over to chat jpt paste it say based
51:33 on
51:35 this
51:37 paste um give me the
51:42 outline and it
51:46 will error out because it was too long
51:49 okay
51:50 okay so chat jpt and an all large
51:55 language model mod have what's called a
51:56 context window and it basically think
51:59 think of the context window in chat GPT
52:01 kind of like your phone screen you know
52:03 when you're scrolling on your phone
52:04 screen and you're having a conversation
52:06 and the conversation Scrolls off the top
52:08 of the screen that's essentially what
52:09 the context window is so in the case of
52:12 chat GPT 3.5 it's got about a 6,000 word
52:17 context window chat GPT 4's got about a
52:21 25,000 word context window um it's
52:25 supposed to have 100,000 context window
52:27 but whatever it doesn't um and uh like
52:30 Claude for example
52:33 has 100 or
52:36 120,000 word context window so you can
52:39 put in longer things
52:40 but when you flip over to chat GPT 4 so
52:44 I was in chat GPT 3.5 so here's chat GPT
52:49 3.5 here's chat GPT
52:52 4 what was different not much right
52:56 except that little thing so that is a uh
52:59 that little paperclip is you can upload
53:01 a file so what I can do now is I can now
53:04 upload that screenplay so if I go find
53:14 salmon so there's my complete screenplay
53:17 and I can say please write
53:21 a five bullet summary of
53:26 this it's going to go read my
53:29 screenplay and summarize it which is
53:33 amazing like it's
53:35 amazing you ever try to read a
53:37 screenplay you ever try you ever have
53:38 ADD and try to read [ __ ] anything and
53:41 now you have a tool to just go read [ __ ]
53:43 for
53:44 you it's like a little miracle it's like
53:48 a little miracle people I'll tell you if
53:51 you if you got a crazy straw of a brain
53:53 like I do if you're neuros spicy like me
53:55 and you got the ADD and just your
53:58 brain's always doing this and every time
53:59 you try to read it's going like why
54:01 don't you go that why don't you go do
54:02 that why don't you go do that well then
54:06 you'll like this I don't know why it
54:08 didn't oh there we
54:11 go here's a five bullet summary screen
54:13 blade begins in Gig Harbor Washington
54:16 Jack is shown cutting bait let's see
54:19 Vietnam flashbacks snakes actions and
54:22 conflict emotional turmoil and
54:24 relationship so it tells us all about
54:25 the all about the thing I could say give
54:30 me
54:34 a character
54:37 breakdown and
54:40 who might play those
54:44 roles in a
54:46 Hollywood
54:54 Blockbuster
54:56 uh
54:58 put the
55:00 names
55:02 descriptions and
55:05 casting ideas in a
55:08 table and so now it's going to go well
55:12 it's already Reddit so it's already got
55:13 the screenplay in memory so here are the
55:17 characters and who should play them Jack
55:21 ruggedly handsome fisherman in his 40s
55:23 Hugh Jackman or Gerard Butler snake
55:26 intense man in his early
55:29 20s Tom Hardy or shy
55:32 labouff
55:35 ping couple of actors Carl Chris Pratt
55:39 or CH Channing Tatum Dewey Michael B
55:42 Jordan or John boa I mean really holy
55:46 crap you can also do this
55:49 um
55:51 describe in
55:53 detail the five
55:56 most
56:00 important
56:02 um visual moments of the
56:11 film describe them in
56:17 catic
56:20 terms that
56:23 a director
56:27 of
56:29 Photography
56:31 might
56:36 understand Jack's morning routine
56:38 cinematic
56:54 description
57:04 create a
57:12 1.85 to
57:16 one
57:18 wide
57:20 image as if it were taken from the
57:27 F so I'm not going to have it
57:32 storyboard an image you know storyboard
57:35 the frames that it just write only if
57:38 it's a PDF though right no you can do
57:40 this with you could do this with a text
57:42 file you could do this with um I don't
57:46 know how else you would have a you could
57:47 do it with a Microsoft Word
57:49 file um you could probably so one of the
57:53 things that sits inside chat GPT is a
57:56 thing called code interpreter that can
57:59 interpret data so it could likely pull a
58:03 like I've got a bunch of screen plays in
58:06 formats for software that doesn't exist
58:08 anymore that that the the the the
58:12 screenplay itself is trapped inside some
58:15 proprietary file
58:17 format
58:19 um and did this make this yet yeah there
58:22 we go look at
58:24 that that's and and here's what I can
58:27 tell you I wrote this [ __ ]
58:29 screenplay that looks like what I
58:32 remember
58:33 writing Apple Pages files don't work oh
58:36 okay so yeah for for Pages files you
58:38 might need to export it as either a Word
58:40 document or
58:42 uh or a uh text file or something like
58:46 that or a PDF
58:53 so let me do next next image next
59:00 image so this is in the present on the
59:03 Puget Sound and if you've ever been to
59:04 the Puget Sound fishing at like 6:00 in
59:07 the morning with the fog lifting off the
59:09 Puget Sound that is exactly what it
59:11 looks like that's what it feels
59:14 like thank you Otto I appreciate
59:23 that and there's the flashback flashback
59:26 to
59:30 Vietnam so so if you haven't used chat
59:33 GPT before and you didn't realize oh
59:36 wait it can it not only can it make
59:38 images it can make images in the context
59:40 of a conversation and that conversation
59:43 can be hey read my screenplay and tell
59:46 me about it or read this screenplay and
59:48 tell me about it or read this novel or
59:50 read
59:51 whatever read this business document
59:56 now let's convert that business document
59:58 into a bunch of rap songs that some
1:00:02 awkward middle-aged white people can
1:00:03 sing at a motivational conference for
1:00:06 the
1:00:08 company and it will do that for
1:00:11 you and then and then if if your if your
1:00:15 if your middle management is so inclined
1:00:18 to embarrass themselves then they can do
1:00:23 that you can make money with
1:00:28 or you can make art or you can make
1:00:31 business or you could have it make a
1:00:32 business plan or I could turn actually
1:00:36 something that I'm actually considering
1:00:39 doing is
1:00:41 um turning my screenplays into graphic
1:00:44 novels turning my screenplays into
1:00:47 books turning you know writing songs
1:00:51 from
1:00:52 them just start experimenting
1:00:55 with all the stuff that's possible so
1:00:58 let's flip back over to this list so
1:01:00 that was just a little bit of scratch in
1:01:02 the surface of chat like like we didn't
1:01:05 we didn't even really do much in chat
1:01:08 chpt and it you know that was I don't
1:01:10 know whatever 15 minutes or so 20
1:01:13 minutes of me rambling while I'm demoing
1:01:16 but if you
1:01:19 haven't if you haven't started
1:01:21 experimenting with these tools like it's
1:01:23 a video game you're missing out on
1:01:26 something I I I'll tell you that that
1:01:29 the thing that's given me the most joy
1:01:31 and I discovered it doing these
1:01:34 lives is that I get the same Rush
1:01:38 working with AI as I get from playing
1:01:41 video games like like like one of the
1:01:44 things I gave up in my life to do these
1:01:47 Tik Tock lives every night is I like I
1:01:49 love Sim racing and so I got a Sim
1:01:52 racing rig and I just love the
1:01:54 adrenaline and you know getting better
1:01:56 and sucking for a while and getting
1:01:58 better and then racing and like when you
1:02:01 game like that you're you're you know
1:02:04 you're always playing against people
1:02:05 that are better than you right so no
1:02:06 matter how good you are someone's always
1:02:08 going to kick your ass and so it's just
1:02:09 this it's just this
1:02:12 constant battle that as you get better
1:02:14 you're never quite good enough but it's
1:02:15 it's immediate feedback and like you you
1:02:18 know you you either wreck the car or you
1:02:19 don't or you spin the guy out or you
1:02:22 don't AI has an IM mediacy to it like
1:02:26 what we just did we took my screenplay
1:02:28 that I wrote 30 [ __ ] years
1:02:30 ago we tossed it in there we said what
1:02:32 is this cast it for us tell us the most
1:02:36 interesting visual scenes make those
1:02:39 scenes for us and it does
1:02:41 it there's something like endorphin
1:02:47 generating about that it's a dopamine
1:02:49 hit yeah Sherry
1:02:51 D it's open source my screenplay no very
1:02:55 very proprietary no um CPT is not open
1:02:58 source um llama 2 from meta is open
1:03:02 source so if you if you if you want
1:03:05 to run this stuff locally or spin up
1:03:09 your own instance there's an entire
1:03:12 open-source Community is going
1:03:14 absolutely crazy developing on llama and
1:03:17 all its derivatives and mixt there's all
1:03:20 sorts of things going on in the open
1:03:21 source Community if you want to play
1:03:23 with open source llms large language
1:03:26 models um thank you for giving that up
1:03:29 for this us oh you're welcome absolutely
1:03:33 um just go to Reddit and type in op
1:03:36 Source AI or op source llm for a large
1:03:40 language model and just start going down
1:03:42 rabbit holes if you're at all geeky um
1:03:45 and and want to start spinning up your
1:03:46 own things um there's so much going on
1:03:49 over there I don't pay I I'm not
1:03:51 following the open source world because
1:03:53 it is it's essentially impossible to
1:03:56 even follow the commercial world I mean
1:03:58 just this list most of the stuff on this
1:04:01 list are commercial entities that you
1:04:04 know these are for-profit companies you
1:04:06 know
1:04:08 commercial um you know pieces of
1:04:10 software and models and things like that
1:04:12 it's it's it's almost impossible to keep
1:04:14 up with this the the open source thing
1:04:16 is just you know multiply this times a
1:04:20 thousand and and that's like every two
1:04:22 weeks there's like a thousand new
1:04:24 projects it's it's absolutely Insanity
1:04:26 right now but that's where all the super
1:04:29 Innovative stuff is happening because
1:04:31 you know they're developers and they're
1:04:33 developing and these tools are really
1:04:35 super Wicked powerful okay so let's
1:04:38 let's see um let me do perplexity so
1:04:42 perplexity is an interesting one I just
1:04:44 got pro access to
1:04:47 this um it's it's a very similar thing
1:04:51 um to to chat
1:04:53 GPT um except what perplexity does is it
1:04:58 it it much more tightly integrates a web
1:05:02 search with a large language model so it
1:05:05 will answer your
1:05:07 question perplexity has a UI for a lot
1:05:10 of Open Source models I found out this
1:05:12 oh cool that's super
1:05:14 cool um but but one of the things that
1:05:17 it's got that I really like it's got
1:05:19 this thing called Focus so so you've got
1:05:22 your you know you can ask your question
1:05:24 here but you can click on focus and
1:05:26 focus you can focus on all you can focus
1:05:28 on academic papers you can focus on
1:05:31 writing you can focus on wolf from alpha
1:05:35 so if you want to ask questions about
1:05:37 mathematics um you can focus on YouTube
1:05:39 or you can focus on Reddit and so I can
1:05:42 go to Reddit and I can say uh op Source
1:05:47 llm um most
1:05:52 popular projects and it I am not a good
1:05:55 perplexity user so there are probably
1:05:58 way better you ways to use this
1:06:01 but so it will give you an answer just
1:06:04 like chat GPT gives you an answer but
1:06:07 here are all the sources so it says it
1:06:09 found five uh these three sources plus
1:06:12 five more so here's the eight sources
1:06:14 it's citing and then it says based on
1:06:16 the search results several open source
1:06:18 large language model projects LL mware
1:06:21 mem GPT which is very cool Q Laura and
1:06:25 Microsoft's Torah Lang chain really
1:06:28 important project llama 2 7 billion
1:06:31 parameter
1:06:33 32,000 context window so there's there's
1:06:36 your meta open source version of chat
1:06:39 GPT um and then you know and there's
1:06:41 your list so so this is a tool
1:06:44 perplexity is a tool that if you want to
1:06:46 do research if you want to be able to
1:06:48 site your sources if you want to be able
1:06:50 to um you know not feel like you're so
1:06:53 much at the mercy of the hallucinations
1:06:55 of chat
1:06:56 GPT this might be a better tool for that
1:06:59 and and that starts to become a lesson
1:07:03 that so I've got a lot of times what
1:07:07 people ask me is well what's the best AI
1:07:10 and it's like well best AI so so
1:07:13 perplexity equals Bard Gemini
1:07:17 um I would say perplexity is a bit more
1:07:21 focused than Bard but Bard Bard does
1:07:23 pretty good with
1:07:25 incorporating um web links in into its
1:07:27 answers I I would say perplexity Bard
1:07:30 and um copilot live in a similar space
1:07:34 with
1:07:36 incorporating
1:07:38 um with
1:07:40 incorporating uh
1:07:42 hyperlinks so hang on let me let me my
1:07:46 iPad is making weird submarine sounds I
1:07:49 don't know what's going on and I also
1:07:51 want
1:07:53 to
1:07:56 go to the mod Channel just in case the
1:07:59 mods are screaming at me to do something
1:08:01 Kyle's live what's up all right
1:08:03 everything's cool over there okay cool
1:08:06 um would love some comments on API
1:08:09 sometime tonight
1:08:14 um apis in what apis for open AI apis in
1:08:19 general just talk about apis in general
1:08:22 like tell tell me a bit more what you're
1:08:23 looking for Jesus t-shirt which is a
1:08:26 funny name by the way
1:08:31 um oh so what I was saying you've
1:08:36 got there's there's no best tool right
1:08:38 now there's just tools and and like chat
1:08:41 GPT is good for one thing claud's good
1:08:43 for something else Bard's good for
1:08:45 something else perplexity is good for
1:08:46 something else you will figure out which
1:08:49 things are
1:08:50 better but the only way you can figure
1:08:52 out is go play with them because because
1:08:54 it's not like it's
1:08:56 better mathematically it's like
1:09:00 oh I like the way Bard makes words that
1:09:04 sounds like more like how I would talk
1:09:06 than cha PT talks oh okay so maybe I'll
1:09:09 lean more to Bard my my co-founder of my
1:09:12 business storyvine she loves Bard I
1:09:15 don't like Bard I like chat GPT it's
1:09:17 more conversational and it's more
1:09:19 flowery Bart is not it's built from a
1:09:23 company that you know
1:09:25 is mostly Engineers it makes sense that
1:09:28 it's that it's a little less
1:09:30 flowery but there's no way to know like
1:09:33 and and it's kind of like this is one of
1:09:36 the lessons I learned being a parent is
1:09:39 I made the mistake a number of times in
1:09:42 in you know my kids life of listening to
1:09:46 other people when they said oh this
1:09:48 school's really
1:09:50 good
1:09:52 because I would listen to to that advice
1:09:54 and we'd move our kids to that school
1:09:56 and we realize they hate it and we hate
1:09:59 it and the people that said it was
1:10:01 really good meant it was really good for
1:10:03 their kind of values and their values
1:10:05 weren't our values right and so so
1:10:07 picking an AI is very similar that that
1:10:09 you'll sort of find your own lane
1:10:13 um Pi is one I want to I want to
1:10:17 show so you can go to pi. can cat GPT
1:10:22 give you a chart of what llms are better
1:10:25 for what
1:10:27 purposes it probably could I mean I mean
1:10:31 it definitely
1:10:33 will the only reason I'm hesitating is
1:10:35 like will it be any
1:10:38 good let me see let me go to chat
1:10:45 GPT create a table of the top 10 llms
1:10:49 like chat
1:10:53 GPT comma what their features
1:10:56 are
1:11:02 comma and provide the top three use
1:11:06 cases that that
1:11:09 llm would be best used
1:11:12 for period be sure to search the
1:11:15 internet for the
1:11:20 latest
1:11:23 input
1:11:27 and it says analyze okay there there did
1:11:30 a Bing
1:11:33 search Chat GPT
1:11:36 3.5 Orca Lambda llama
1:11:41 Palm Claude V1 cohere so so this is just
1:11:45 talking about the underlying models and
1:11:47 this is not a lot of the big ones so so
1:11:49 let me do
1:11:53 um
1:11:58 put the answers in
1:12:01 the put the answers in the format of a
1:12:03 table and focus
1:12:05 on application
1:12:09 layer chat applications like chat GPT
1:12:13 rather than the underlying models
1:12:15 themselves
1:12:17 period focus on the commercial
1:12:22 tools rather than open
1:12:25 source or you could say focus on open
1:12:28 source rather than commercial
1:12:49 tools chat GPT Bing chat yeah it did it
1:12:53 replica you chat Cleverbot Muka yeah
1:12:58 it's weird Jasper Elsa speak I so it
1:13:03 just did chat Bots like it's this is bad
1:13:05 prompting on my part it it'll do it
1:13:07 you're going to need to massage it to
1:13:09 get it to do it
1:13:12 well um okay I want to show you
1:13:15 something
1:13:15 so pi. a is from a company called
1:13:21 inflection not a lot of people have
1:13:23 heard of it it's not one of the ones you
1:13:25 hear people talk about a lot but what I
1:13:27 can tell you is the people that use it
1:13:29 love it showed pi to my friend she said
1:13:32 P was more sympathetic than her
1:13:33 therapist I've Source Camp I've heard
1:13:35 that before as well silverfox said that
1:13:37 that it was she was dealing with
1:13:39 something with her family and said Pi
1:13:40 was more compassionate than her family
1:13:42 members I'm like yeah it's it's cuckoo
1:13:45 so so piie you can go to pi. a that's
1:13:49 the website but what I strongly
1:13:51 recommend is download the PI app for iOS
1:13:55 or
1:13:56 Android and then when you run it
1:13:59 actually let me go change my voice
1:14:03 because
1:14:07 where voice
1:14:10 settings hey there great to meet you I'm
1:14:12 there she is it it it changed my voice
1:14:15 up on me
1:14:17 okay so let me do a new
1:14:21 search I guess I can't all right
1:14:25 so I'm just in the I'm I'm just on an
1:14:28 iPad this is just an iPad and there's
1:14:30 the pi icon right below Story vine
1:14:34 that's my little company that cute
1:14:35 little gnome right there that's Story
1:14:37 vine and that's pi and so you click on
1:14:42 pi and then this little button to the
1:14:46 right of the little chat box the little
1:14:49 telephone click on
1:14:52 that
1:14:57 hey what's
1:15:03 happening hey user how are you tonight
1:15:05 what's on your mind wait why are you
1:15:08 calling me user you know my name's Kyle
1:15:15 weirdo oh no my mistake I was using my
1:15:18 generic way of addressing users but
1:15:20 you're right your name is Kyle sorry
1:15:22 about that that's okay uh let's play a
1:15:25 game I want you to teach me something
1:15:27 about something obscure and make it so
1:15:31 that the game only lasts 30
1:15:37 [Music]
1:15:38 seconds all right here's your 30C crash
1:15:41 course on a delightfully obscure topic
1:15:43 the bloop the bloop was an ultra low
1:15:46 frequency sound recorded by US
1:15:48 Government sensors in 1997 believed to
1:15:51 be the loudest sound ever recorded in
1:15:53 the world world's oceans it remains
1:15:54 unexplained to this day the end yeah but
1:15:58 that wasn't a game you just told me a
1:15:59 story can we make that a game
1:16:04 [Music]
1:16:06 somehow sure thing let's turn it into a
1:16:09 game called 30second stump I'll choose a
1:16:11 random obscure topic give you a 30
1:16:14 second explanation and you have to guess
1:16:15 the topic ready go this rare and unique
1:16:19 geological phenomenon occurs when lava
1:16:21 cools quickly on the surface and The
1:16:24 Rock contracts resulting in a distinct
1:16:26 hexagonal pattern it can be found in
1:16:29 just a few places around the world
1:16:30 including Iceland and Northern Ireland
1:16:33 what is
1:16:35 it hexif
1:16:42 foration close it's actually called
1:16:45 hexagonal Basalt turn to choose an
1:16:48 obscure topic for me to explain in 30
1:16:50 seconds
1:16:52 so
1:16:55 what okay so hang
1:16:59 on all of all of that conversation has
1:17:03 been transcribed so you can do this with
1:17:04 chat GPT and with he hexagono basal
1:17:08 someone got it show off Mary Mary show
1:17:11 off
1:17:17 um so you you could you could argue
1:17:21 intellectually that that's exactly the
1:17:24 same as just typing into chat
1:17:26 GPT what I can tell you is it's
1:17:29 completely different if you try to use
1:17:32 Pi like chat
1:17:34 GPT you'll be wildly frustrated because
1:17:37 Pi is optimized for conversational for
1:17:41 conversations and um chat GPT is
1:17:45 designed for interactions like like
1:17:51 this it
1:17:54 it is impossible to describe how
1:17:56 different it is but like when you put Pi
1:17:59 you go on go on like a two-hour road
1:18:01 trip and just put piie on instead of
1:18:04 putting on a podcast put on pie and
1:18:08 start a conversation with it you can
1:18:10 talk about [ __ ] anything it's
1:18:13 connected to the internet it's got this
1:18:15 massive large language model sitting
1:18:16 underneath it they're they're making it
1:18:18 bigger and better inflection is the
1:18:20 company behind
1:18:21 it and it and it's designed to be this
1:18:24 conversational experience and it's just
1:18:26 a completely different thing when you
1:18:29 use the conversational version of chat
1:18:30 GPT you if you're using the chat GPT 4
1:18:34 that makes
1:18:35 images you can have it you can be doing
1:18:39 visual
1:18:40 conceptualizations like give me some
1:18:42 visual ideas for how to bring this
1:18:44 marketing campaign to life and then it
1:18:47 will and then you say go make the image
1:18:49 and it will go make the image and it
1:18:50 shows it to you and by the time you get
1:18:53 to the the office you could have done
1:18:55 visual concepts for a marketing
1:18:58 campaign as a conversation in the
1:19:02 car
1:19:04 wild
1:19:06 wild now what P can't do so so Pi in my
1:19:13 opinion is what I was
1:19:17 promised Siri and
1:19:19 Alexa were for the past decade hey
1:19:25 Siri mhm go
1:19:28 away
1:19:34 okay it's just not the
1:19:37 same so Pi you can have a conversation
1:19:39 with what what you can't do with pi
1:19:41 right now is pi can't take actions but
1:19:44 inflections mission is to create the
1:19:46 world's most powerful personal assistant
1:19:49 so the pi that you talked to today
1:19:52 they're prioritizing
1:19:54 conversation and the next layer sitting
1:19:57 behind that is going to be it taking
1:19:58 action so imagine having a Siri or an
1:20:01 Alexa that you could have a full-on
1:20:03 conversation with that could then take
1:20:05 actions on your
1:20:06 behalf what's their privacy policy like
1:20:09 I haven't dug too much into that Lori
1:20:12 I'm assuming that all of these
1:20:14 commercial companies are defaulting on
1:20:17 we can train on your data Siri sounds
1:20:20 jealous Siri's so jealous she's so
1:20:22 jealous she'll no but in when when
1:20:24 things get hot and heavy hey
1:20:30 Siri she's being koi she went
1:20:36 away I always wondered why genen Z was
1:20:39 always on their phone starting to make
1:20:40 sense
1:20:42 Evolution yeah and what's funny is I
1:20:44 don't think a lot of gen zers are into
1:20:46 this AI stuff at all it's it's mostly
1:20:49 gen xers Elder Elder Millennials and gen
1:20:51 xers seem to be really into this
1:20:53 and some
1:20:54 Boomers but it seems like hot and heavy
1:20:57 on Gen X at least you know my channel
1:21:00 cuz you know I think I'm attracting you
1:21:02 know that
1:21:04 crowd uh if I got into twerking I could
1:21:06 probably bring in a younger crowd no I
1:21:15 couldn't I'm GNA use chat to design my
1:21:18 children's book illustrations oh it's so
1:21:20 good at that now what's going to
1:21:22 frustrate the [ __ ] out of you with the
1:21:24 children's book illustrations is getting
1:21:26 consistent characters um but it's you
1:21:30 you can do it decently well with Dolly 3
1:21:32 with chat GPT um but you're going to
1:21:35 that's that's going to be one of your uh
1:21:38 special um your special frustrations
1:21:42 what so I actually here was a I I had a
1:21:45 creative uh a creative work around for
1:21:47 not having consistent characters is just
1:21:50 make books that are like a different
1:21:52 character on every
1:21:55 page um will cat GPT just turn into
1:21:59 Periscope um no no no no no these so so
1:22:05 what's going to
1:22:07 happen what's going to happen
1:22:17 well most if not all of these tools are
1:22:20 going to get incredibly multimodal is
1:22:23 the term that that they're all talking
1:22:25 about and they're going to and they're
1:22:27 going to
1:22:28 get
1:22:30 um they're going to get better at
1:22:32 reasoning and they're going to get
1:22:34 better at taking actions um
1:22:39 so so I think where these things go is
1:22:43 that we have a single
1:22:47 interface that's kind of the everything
1:22:50 app right so like you say
1:22:55 um I want to book a vacation and it goes
1:22:58 off and it books you a vacation and
1:22:59 you're like I want a brainstorm about
1:23:01 something and it'll brainstorm with you
1:23:03 and you're like I want to I want to talk
1:23:05 I I I want to I'm I'm struggling with
1:23:09 something at work and I want to I want
1:23:10 to talk through how to talk to this
1:23:12 other person with this difficult
1:23:13 conversation and it'll roll into
1:23:15 therapist mode and then you're like no
1:23:17 you know what I I don't want to talk to
1:23:19 you about this I want to talk to my
1:23:20 three best friends about this see if any
1:23:22 of are available and it'll it'll
1:23:25 immediately conference you in with two
1:23:27 of them right so I think where these
1:23:29 things go is they just
1:23:33 become we essentially we as we as people
1:23:36 using these tools just become like
1:23:39 directors with
1:23:42 goals and then these tools do the
1:23:46 work
1:23:48 um that's not this year that's not 2024
1:23:51 but we'll see the we'll see the
1:23:55 um we'll see the foundation for that in
1:23:58 2024 what you're going to see in 2024 is
1:24:01 these tools are going to get good at
1:24:02 audio video images music um animation
1:24:07 coding game mechanics World
1:24:10 building obviously word stuff coding
1:24:13 stuff and they'll start taking
1:24:17 action on our behalf that we give them
1:24:19 permission to do and
1:24:23 they're going to get better at
1:24:25 reasoning so rather than them
1:24:27 hallucinating lying to you mansplaining
1:24:30 to you here is a correct answer even if
1:24:33 it's
1:24:35 wrong they will do some reasoning and
1:24:37 say is this answer correct right now
1:24:39 they can't do that the way large
1:24:41 language models work they actually don't
1:24:43 know what they tell you they literally
1:24:46 don't know what they said so when when
1:24:49 people say well they they're acting like
1:24:51 they're sensient because they can read
1:24:53 tarot cards and predict the future and
1:24:55 say they have feelings and this and that
1:24:58 they say all
1:25:00 that but they're they are technically
1:25:03 calculators they're probability engines
1:25:07 that just happen to do this remarkable
1:25:09 thing when I say Periscope I mean chat
1:25:11 GPT is first then phased out oh oh oh I
1:25:19 see I think the Hedge against the Paris
1:25:22 scope fate right the the the you know
1:25:25 who's first is not necessarily who's
1:25:27 biggest the fact that Microsoft put $10
1:25:29 billion into open Ai and didn't flee the
1:25:32 ship when they fired Sam Alman um that
1:25:36 says to me that that Microsoft is
1:25:38 backing them and and and Microsoft has
1:25:41 rolled open AI you know chat GPT 4 or
1:25:46 co-pilot is built on top of gp4 the
1:25:49 underlying model and co-pilots in in
1:25:53 Windows Office 365 Dynamics 365 Azure
1:25:57 co-pilot all of the Microsoft
1:26:00 products have this open AI Technology
1:26:04 Incorporated into
1:26:06 it so I don't see it going away I don't
1:26:09 I don't see it going away like a
1:26:11 periscope you might have a company like
1:26:14 Pi or
1:26:17 um Po from you know quar I could see
1:26:22 going this way
1:26:24 but I mean a a couple of things bode
1:26:27 well for for open
1:26:32 Ai No one's caught them in a
1:26:35 year I mean gpts the GPT technology the
1:26:38 Transformer technology has been around
1:26:41 since
1:26:43 2017 people have been building on this
1:26:45 for a
1:26:49 while opena I didn't think chat GPT was
1:26:52 going to be a big deal when they
1:26:53 launched it when they launched it it
1:26:54 became a very big deal and they quickly
1:26:57 went from chat GPT 3.5 to chat GPT 4 no
1:27:01 one's caught them they're an 800 person
1:27:04 company Google hasn't caught them meta
1:27:07 hasn't caught them Apple hasn't caught
1:27:09 them no one's caught them Facebook or uh
1:27:13 um who else is who else is out there um
1:27:16 any of the startups you know cohere and
1:27:20 anthropic and inflection no no one's
1:27:23 them so they're not just ahead they're
1:27:26 way
1:27:28 ahead they've got Microsoft as a backer
1:27:31 and not just as an investor it's built
1:27:33 into the Microsoft
1:27:35 ecosystem and
1:27:37 then when they fired Sam
1:27:40 Alman and it became clear he was going
1:27:43 to go work at Microsoft 95% of the staff
1:27:49 signed a letter that said if you don't
1:27:51 fire the board Bo and reinstate him as
1:27:54 as CEO we're going with him 95% of the
1:27:59 staff so what what that whole debacle
1:28:04 did
1:28:06 was like crazy glue that team together
1:28:10 like the amount of
1:28:12 of commitment to the to the cause that's
1:28:17 inside open AI right now has got to be
1:28:21 unbelievable
1:28:23 like that team is like on a [ __ ]
1:28:25 mission right
1:28:26 now they're like you're not going to
1:28:28 take this [ __ ] thing away from us we
1:28:30 will walk 95% of
1:28:33 them and they fired the board and
1:28:35 reinstated Sam Alman
1:28:37 right so I think they're going to
1:28:40 survive but you
1:28:42 know like consolidation happens I mean
1:28:45 I've got on my little list here that
1:28:47 we're going through I've got um 8 n I've
1:28:52 got nine different large language models
1:28:56 you know nine different companies
1:28:59 represented it's not going to be nine
1:29:01 for you know that will consolidate to
1:29:03 three it always consolidates to
1:29:05 three so my guess is it's going to be
1:29:08 you know Google Google will be one of
1:29:11 them open AI Microsoft will be one of
1:29:14 them and then probably Apple like we
1:29:17 haven't talked about Apple they haven't
1:29:19 done anything yet but wait till they
1:29:21 wait till they come in this [ __ ] game
1:29:23 they're
1:29:24 building go listen to their last
1:29:29 four their last four announcements
1:29:32 Apple's last four
1:29:34 announcements they are being very
1:29:38 Cy but they
1:29:40 are quietly building the hardware
1:29:42 infrastructure and the Privacy
1:29:44 infrastructure to knock this [ __ ] out of
1:29:47 the park Apple's going to be good at
1:29:49 this [ __ ] they're going to be good at
1:29:51 agents they're going to be good at large
1:29:53 language models they're going to be good
1:29:54 at privacy they're going to be good at
1:29:56 you know locking you into an ecosystem
1:29:59 and having [ __ ] that's
1:30:03 cool and then
1:30:06 Meta Meta is going to ensure
1:30:09 that the open source
1:30:12 Community has toys to play with too that
1:30:15 rival the big the big
1:30:19 companies I don't know if you saw
1:30:21 Zuckerberg's
1:30:22 Instagram reals post the other day but
1:30:25 but he um he put out a thing he
1:30:28 basically said meta is committing to AGI
1:30:31 which stands for artificial general
1:30:32 intelligence that's the mission of open
1:30:35 AI That's the mission of chat
1:30:37 GPT is to be first to AGI artificial
1:30:41 general intelligence that's the point at
1:30:42 which these tools are better than most
1:30:46 humans at most
1:30:50 tasks Rumor Has It
1:30:54 that open AI stumbled onto something
1:30:57 pretty [ __ ] powerful which is why Sam
1:30:59 mman got ousted that's a rumor right now
1:31:02 we don't know David Shapiro seems to
1:31:05 think that we're going to have some
1:31:06 version of a AGI within
1:31:11 2024 Mark Zuckerberg just
1:31:13 said we're we're rolling together our
1:31:16 two primary AI research groups within
1:31:19 meta into a single group focused on AGI
1:31:23 and we're pointing
1:31:26 350,000 h100 gpus at the problem and
1:31:30 then they've got another 300,000 h100
1:31:33 equivalents if you take all their other
1:31:35 gpus that are not
1:31:39 h100s
1:31:41 350,000 h100 gpus is like $20
1:31:45 [Laughter]
1:31:49 billion So So Meta Meta is pointing $220
1:31:53 billion of
1:31:56 hardware at building llama 3 training
1:31:59 llama 3 and then they're going to open-
1:32:01 Source it so I think you'll have three
1:32:04 major players and then you'll have a
1:32:06 major open- source player out there
1:32:08 that'll be three or five years from now
1:32:10 but that's what it'll consolidate down
1:32:15 into I know you didn't Siri you don't
1:32:17 get a lot Siri all right can you talk
1:32:21 about broadcast from a crackling
1:32:23 fireplace can you broadcast from a
1:32:26 crackling fireplace and talk about the
1:32:29 the future of AI some night I found this
1:32:32 on the hey
1:32:34 Siri go away evening go away
1:32:40 okay yes I have a I have a
1:32:44 um oh nice ice skates thank you forever
1:32:46 hooked I have a a nice propane fireplace
1:32:50 I could I could go sit by a fire I'd
1:32:52 have to kick kick the misses out of the
1:32:55 out of the family room but I could do
1:32:57 that some night and just talk about the
1:32:59 future of AI I mean I will talk about
1:33:01 the future a a future of AI with a
1:33:04 caveat that I don't know [ __ ] about the
1:33:06 future of
1:33:07 AI I can talk about it in a historical
1:33:10 context right I started one of the first
1:33:12 digital agencies in the mid90s company
1:33:14 called agency.com
1:33:16 so I've been in a version of this movie
1:33:19 before
1:33:20 right like starting at the beginning of
1:33:23 the worldwide web nobody knew what it
1:33:25 was nobody got
1:33:27 it and then we said well why don't we
1:33:31 start a company and maybe we can help
1:33:32 people make websites and then very
1:33:35 quickly we were like making websites for
1:33:37 the Fortune 500 it was [ __ ]
1:33:40 Insanity it was crazy
1:33:43 and um so I've I've I've been in I I've
1:33:48 I've ridden one of these waves before so
1:33:51 I can put it in that
1:33:52 context
1:33:55 but I could I could I could sooth say
1:33:59 about the worldwide web I could see what
1:34:02 the future was with the worldwide web
1:34:03 because it was really simple it was just
1:34:06 this right the worldwide web the the
1:34:08 core technology of the worldwide web was
1:34:11 this instead of having to type a command
1:34:15 line you click the link click and that
1:34:18 jumps you over
1:34:20 there and so the implications of it felt
1:34:24 quite clear to me right geography
1:34:27 collapses time collapses because you
1:34:30 know you don't have to send a FedEx that
1:34:32 takes two days to get something there
1:34:34 you can send an email that's
1:34:36 instant you can right and and because
1:34:39 it's as simple as clicking on stuff
1:34:42 regular human beings can use it you
1:34:43 don't have to be an engineer to be able
1:34:45 to use this
1:34:46 infrastructure and so you could just see
1:34:48 that that everything accelerated and
1:34:50 everything became instant and the
1:34:52 distance between a brand making a
1:34:54 promise and a brand delivering on that
1:34:57 promise or not was now zero right so you
1:34:59 can kind of see where it was headed this
1:35:02 AI stuff I can't for the life of me get
1:35:05 my head around what the future looks
1:35:10 like it's
1:35:12 like when every human being gets to
1:35:15 level up their
1:35:17 skills essentially
1:35:20 overnight where everyone can be like 80%
1:35:24 as good as the people that have been
1:35:27 doing it for 20 or 30 years
1:35:31 overnight what are the implications to
1:35:34 work the definition of work the value of
1:35:37 work the future of the
1:35:39 workforce the definition of
1:35:42 job the definition of
1:35:45 company I don't know when a single
1:35:50 person and this isn't this year but like
1:35:53 this is three years from
1:35:55 now when we have these agents like like
1:35:59 we talk about these agents I don't think
1:36:00 people get what's
1:36:04 coming when people talk about AI agents
1:36:07 what we think of because this is what
1:36:09 we've been told for the past 10 or 15
1:36:12 years we hear about um AI chat Bots
1:36:16 there's these AI chat Bots and when you
1:36:18 go to a website you get the AI chat with
1:36:20 an AI chat bot and always these
1:36:22 disappointing horrible experiences
1:36:25 because the one question you ask it it
1:36:27 can't [ __ ] answer because it's just
1:36:29 programmed as a logic tree and if your
1:36:32 question doesn't fit in their logic tree
1:36:34 it just poops
1:36:36 right uh there's a poll coming you've
1:36:39 got one minute the poll is right there
1:36:43 and it's would you attend the AI
1:36:45 learning lab fireside chat a getting my
1:36:48 blanket B don't bother Mrs Shannon C I'm
1:36:51 listening
1:36:52 anywhere okay go go poll that's pretty
1:36:56 funny poll poll is life thank you thank
1:36:59 you for the
1:37:00 poll
1:37:05 um oh man um what was I talking about oh
1:37:09 future future of
1:37:12 AI
1:37:15 um like like the implications are just
1:37:19 they're they're so profound like I
1:37:23 can't I can't I can't process it I mean
1:37:27 I can't
1:37:29 process just for myself personally like
1:37:32 even not me as role of CEO of a company
1:37:35 of 11 and a half years
1:37:37 old like that's complicated enough just
1:37:40 for me
1:37:41 personally as as someone who loves
1:37:44 technology curious about this stuff
1:37:46 loves playing with
1:37:50 it I can not get my head around where to
1:37:53 put my
1:37:56 attention like okay I could make
1:38:00 worldclass images now in 6 months I'll
1:38:03 be able to make world class music maybe
1:38:05 a year in in a year I'll be able to make
1:38:08 world class
1:38:10 video um I would think within the next
1:38:13 year I'll be able to all right would you
1:38:16 attend the fireside chat
1:38:18 57% say yes getting my blanket
1:38:22 8% really considerate of Mrs Shannon
1:38:26 those are the nice people and then and
1:38:29 then 35% are like I don't care where
1:38:31 Kyle is he he can be he can be sitting
1:38:34 on a pile of dirt and I'll watch him
1:38:39 so oh man what's the coolest thing
1:38:42 you've done with AI
1:38:45 um starting this community is is
1:38:48 probably the coolest the AI salon and
1:38:51 and this live Community for me right now
1:38:54 is it just gives me a ton of
1:38:57 Joy
1:38:59 um I think the GPT for good project that
1:39:03 that we did that an Murphy and I hosted
1:39:05 and Vicky Baptist built the directory
1:39:07 for um that project where we taught
1:39:10 people how to make gpts and this
1:39:13 community again it's back to this
1:39:14 community built 150 gpts for nonprofits
1:39:18 in a 24-hour period and we we arranged
1:39:23 that event in seven days like someone on
1:39:27 one of these lives said oh you should do
1:39:29 another 24hour thing and like I was in a
1:39:31 moment of weakness and I was like that's
1:39:33 a fine
1:39:35 idea and then 7 days later we just
1:39:37 [ __ ] did
1:39:40 it and I was recovering from covid when
1:39:43 we did it but that was that was amazing
1:39:46 right that
1:39:49 was what was amazing about that is the
1:39:53 way
1:39:54 that open AI created their GPT
1:39:59 Builder it's the first in my experience
1:40:03 it's the first
1:40:05 true zero code application development
1:40:09 tool you're like well you got zappier
1:40:11 and you got bubble and you got make and
1:40:13 you got all these no no all of those
1:40:15 things are low code or no code but
1:40:19 they're still code and you still have to
1:40:21 understand apis and API keys and
1:40:24 programming logic and understanding
1:40:26 inputs and outputs and processing units
1:40:29 and if you're not a
1:40:32 developer no code platforms are not
1:40:39 simple if you're kind of a half-ass
1:40:42 product guy like me you can figure them
1:40:45 out but it's still
1:40:47 hard and if you're a programmer you look
1:40:50 at tools like zapier and you're like
1:40:51 well it's fine but I'd rather just
1:40:52 program it because then I have complete
1:40:54 control I don't want to use their stupid
1:40:57 half-ass
1:40:58 constrained Lego
1:41:03 blocks but
1:41:05 gpts were truly people who' never coded
1:41:08 before came to that Saturday event and
1:41:12 were like I don't know what to what's a
1:41:16 GPT so we told him what a GPT was and
1:41:18 then we said click on this click here
1:41:20 click here
1:41:22 and type in I want my GPT to teach
1:41:25 people about nonprofit with this name we
1:41:28 made like little mad libs
1:41:32 prompts and get the information about
1:41:35 that organization from this website that
1:41:38 would the entire prompt for those gpts
1:41:41 was three
1:41:42 sentences or two
1:41:44 sentences with with Mad Libs in
1:41:47 them and then we spent the next hour
1:41:49 just kind of tweaking them
1:41:53 and people came in who'd never made an
1:41:55 application before and they were like
1:41:56 this is amazing we're like yeah we know
1:42:00 and they were like I could do this for
1:42:02 other things too yeah I know and then
1:42:04 they just would go off and build a [ __ ]
1:42:05 ton of these things a lot of those
1:42:08 people are still building gpts and
1:42:10 they're getting interest from investors
1:42:13 I want to invest in your GPT collection
1:42:15 okay can you do that I don't
1:42:19 know I'm a software developer should I
1:42:21 be worried about AI replace
1:42:24 me no this there's a Paul rater quote um
1:42:29 a AI will not replace you someone using
1:42:33 AI will replace you so if you're a
1:42:36 developer get your ass par programming
1:42:40 with AI immediately go to GitHub
1:42:42 co-pilot or go to chat
1:42:44 gp4 and start leveling up your coding
1:42:48 like get literate about what these tools
1:42:50 can do
1:42:51 cuz they're [ __ ] insane they can they
1:42:54 can debug [ __ ] for you they can document
1:42:56 code for you they can analyze code for
1:42:58 you they can write code for you well is
1:43:01 the code perfect I don't know I'm not an
1:43:03 engineer but what I know is that the
1:43:05 engineers that I know that are using it
1:43:08 they're upping their productivity by 60
1:43:10 70
1:43:11 80% depending on how heavy they're
1:43:13 leaning into
1:43:15 it so get AI literate [ __ ] Pronto and
1:43:19 quite frankly what I would do if I were
1:43:21 Duke is I would get involved in some of
1:43:23 the open open source projects start
1:43:25 contributing to you know Lang chain or
1:43:29 to
1:43:32 um um the the the agent ones uh what are
1:43:36 they called I forget what it's called
1:43:39 um oh [ __ ] I can't remember um David
1:43:43 Shapiro's got a cognitive architecture
1:43:45 he's working on for AGI like just just
1:43:48 go to Reddit and dig into the to the the
1:43:51 open source Community get involved in in
1:43:54 those communities
1:43:56 because you
1:43:58 just as a developer you you you're going
1:44:01 to have to have your [ __ ] together with
1:44:03 AI can we do another poll that ask what
1:44:05 is the level of your your anxiety as you
1:44:08 listen to Kyle
1:44:13 therapist does it go up or go
1:44:16 down I guess that would be a good
1:44:20 poll
1:44:23 I mean I know I'm a little ampy I get
1:44:26 that that's perfect way to describe it I
1:44:29 feel overwhelmed because I don't know
1:44:31 which direction yeah and and so here's I
1:44:35 I have an antidote for it but I just I'm
1:44:38 not I'm not taking my own
1:44:41 advice an antidote for being overwhelmed
1:44:46 with AI tools right now is don't focus
1:44:49 on the tools
1:44:52 probably a potato I've programmed for
1:44:54 two decades and chat GPT blows me away
1:44:58 yeah I
1:44:59 mean and and this is the thing so
1:45:02 probably a potato here's here's the
1:45:07 deal I would love I would love to have a
1:45:12 developer that's got two or three
1:45:14 decades of programming
1:45:17 experience that gets AI literate and is
1:45:21 like really passionate about it and like
1:45:23 knows all the tools like I I know all
1:45:25 the tools cuz I'm not a developer right
1:45:28 so I can sort of dick around with
1:45:30 prototyping
1:45:33 stuff but like right now I'm building a
1:45:35 tool within storyvine called the
1:45:37 authenticity engine that's this content
1:45:39 generation machine it's [ __ ]
1:45:41 magic but my developer there hasn't
1:45:45 really used chat GPT before he's into it
1:45:48 now but it's like it it was like pulling
1:45:50 teeth
1:45:52 if I had someone that had like deep you
1:45:54 know like you know full stack
1:45:56 application
1:45:58 development and then came to me and said
1:46:00 Kyle you should see what's going on with
1:46:01 AI we could do this we could do that we
1:46:03 could do that and then if I could just
1:46:04 say oh can we do this can we do that and
1:46:07 we could very very quickly knock that
1:46:09 [ __ ] out like that would be shanga for
1:46:12 me back to
1:46:15 shanga the moral of the story the morale
1:46:18 of the story Mr Kyle has been a huge
1:46:20 inspiration for me oh thank you so much
1:46:22 that's very
1:46:25 sweet so yeah if if you're if you're an
1:46:29 engineer so here's here's a prediction
1:46:31 someone wants a prediction of the future
1:46:33 I'll give you a
1:46:34 prediction I'm starting source source
1:46:36 Camp said she felt the same thing when I
1:46:39 when I said this last night or the night
1:46:42 before um there's a shift starting to
1:46:45 happen right now I think in 2024 the
1:46:47 dams are going to start to
1:46:49 break with with people and
1:46:55 companies not really getting religion
1:46:57 about AI but getting panicked about AI
1:47:00 like oh [ __ ] I got to learn AI uh who
1:47:05 knows anything about
1:47:06 this
1:47:08 anybody anybody know
1:47:10 anything and that's probably your CEO
1:47:13 acting like that oh God I got to learn
1:47:15 about AI anybody know anything if you're
1:47:18 one of those people that knows anything
1:47:20 you're going to be in demand in
1:47:23 2024 well how do I learn anything Kyle I
1:47:26 don't know anything just hang out here
1:47:28 go to the AI Salon join the AI salon
1:47:33 so the salon. go there click on the top
1:47:36 link join that Community read our values
1:47:39 first make sure you resonate with
1:47:42 them and operate within
1:47:47 them join this community come to these
1:47:50 things every
1:47:52 night well what are you going to teach
1:47:54 all you do is sit around talking and
1:47:55 making people anxious and playing songs
1:47:57 with your stupid dog that sings
1:48:02 yep because the only skill that I'm
1:48:04 really teaching here is stay in the
1:48:08 conversation stay curious stay
1:48:10 adventurous keep playing keep playing
1:48:14 keep playing well how am I supposed to
1:48:16 monetize that you can make money time I
1:48:19 don't [ __ ] know you go start a
1:48:24 business but you
1:48:26 can't do
1:48:28 that unless you know what these tools
1:48:31 are and what they're capable of and then
1:48:34 you can look at the world through your
1:48:36 lens and your expertise and go oh you
1:48:40 know who are really struggling right now
1:48:42 accountants trying to use this overly
1:48:44 complicated tax software what if I used
1:48:47 AI to simplify that and I just created a
1:48:50 simplif if layer that made it really
1:48:52 easy for accountants to not have to
1:48:55 learn new software every year
1:48:59 perfect you can make
1:49:03 money you just got to stay in the
1:49:05 conversation right now because all these
1:49:09 tools are are not anywhere close to what
1:49:12 they're going to
1:49:14 be how we use them is not close to how
1:49:17 we're going to use them and and you know
1:49:19 there's there's
1:49:24 there's even a part of me right now
1:49:25 that's like well maybe we should just
1:49:27 wait just wait for the tools to Evol no
1:49:29 we shouldn't
1:49:31 wait you should be building stuff now
1:49:34 knowing that the [ __ ] that you're
1:49:36 building is disposable because you're in
1:49:39 the conversation you should be talking
1:49:41 about it all right new poll what's your
1:49:44 level of anxiety oh my God stop I'm
1:49:46 getting it I'm really into it you got
1:49:49 you got 20 seconds 12 12 seconds left 10
1:49:51 seconds left go there and
1:49:57 and oh my God stop for the oh my God
1:50:00 Stoppers you know you can do this
1:50:04 flick 60% are really into it 50% are
1:50:07 getting it 19% are like oh my God stop
1:50:11 you know flick just like this this a
1:50:14 beautiful thing about Tik Tok flick I
1:50:16 guess if I'm live you've got to go flick
1:50:18 and then flick
1:50:22 they're going to tell HR to go hire some
1:50:25 AI people they are
1:50:26 Jewels yes the people
1:50:29 panic the first the first place that
1:50:32 they'll go is internal right so the
1:50:34 CEO's going to go oh my God I need to
1:50:35 learn AI what do I
1:50:37 do does anybody know AI well you know
1:50:40 Sally was talking about AI at lunch the
1:50:42 other day I need Sally in my office
1:50:48 stat hello Mr mcgilla cut
1:50:51 who are you I'm Sally yeah sit down I've
1:50:55 heard a lot about you uh what's going on
1:50:58 with this
1:50:59 AI well what do you
1:51:02 uh I can use chat GPT a little oh chat
1:51:05 GPT I've heard of that you're VP of AI
1:51:08 now
1:51:09 congratulations like that shit's gonna
1:51:12 happen in
1:51:15 2024
1:51:18 promise Reena yep I listened about
1:51:21 obsidian and GitHub on AI
1:51:23 salon and I never even heard about it
1:51:27 yep exactly
1:51:30 yep H yeah just just hang out in the AI
1:51:33 salon and it it doesn't have to be my
1:51:35 community I don't give a [ __ ] like I
1:51:37 just think our community is the coolest
1:51:39 [ __ ] Community out
1:51:40 there and and if you want to know why go
1:51:44 read the
1:51:45 values and make sure you don't violate
1:51:47 them like live inside those values
1:51:50 curiosity
1:51:53 generosity
1:51:56 exploration collaboration inclusion
1:51:59 empathy empathy
1:52:05 empathy what I have
1:52:08 someone someone in some troll in here a
1:52:12 month or two ago I went through the
1:52:15 values and the last one was empathy and
1:52:18 and he trolled up oh what is this some
1:52:21 sort of woke
1:52:23 organization I'm like well we're not
1:52:25 [ __ ] to each other if that's what
1:52:27 you're
1:52:29 asking
1:52:31 like if that's your definition of woke
1:52:35 then yeah we are probably not your
1:52:39 probably not your
1:52:42 crew plat twist Sally is AI exactly all
1:52:47 right Mr mcgilla cudy can I get you
1:52:50 another
1:52:51 [Laughter]
1:52:55 coffee oh developers have such an
1:52:59 advantage right now they can find tune
1:53:01 gpts I'm telling you man developers
1:53:05 right now it so this is not about
1:53:08 engineering it's not about the the
1:53:10 whatever [ __ ] language you learned in
1:53:12 college it's not about what code you can
1:53:14 do it's if you're a developer if you're
1:53:16 an engineer if you can think like system
1:53:20 thinking and then you can build
1:53:25 [ __ ] and you're AI literate sorry those
1:53:28 two have to go together you can build
1:53:31 [ __ ] you have some good experience
1:53:34 building [ __ ] and you're AI literate
1:53:38 those
1:53:41 people they are going to be in such
1:53:45 demand the only thing to making money is
1:53:48 solving problems if you can do it
1:53:51 um if you can do that you will profit
1:53:53 that's that's why I have this you can
1:53:56 make money
1:53:57 with these people that tell you that are
1:54:00 full of
1:54:03 [ __ ] cuz it's not about chat GPT chat
1:54:06 GPT is just a
1:54:09 tool like the people that are telling
1:54:11 you
1:54:12 that they're
1:54:18 tools how byy my course on how to make
1:54:21 money with ch
1:54:22 P so you're not making money with C GPT
1:54:25 you're making money by praying on the
1:54:28 fear of people who think they're missing
1:54:30 the boat is that right well that's
1:54:33 technically
1:54:37 yes what's the best way to learn AI
1:54:40 you're in it like play play play play
1:54:45 play play play play play Let's go join
1:54:48 the AI salon
1:54:54 and when you come over here go to
1:54:56 welcome to the salon I'm going to keep
1:54:58 harping on this I know you people are
1:55:00 tired of it but I don't give a
1:55:01 [ __ ] read our values curiosity
1:55:05 generosity exploration collaboration
1:55:08 inclusion and empathy if you resonate
1:55:11 with those values this is your crew if
1:55:14 not not that's cool but anyway go there
1:55:18 then introduce yourself say hi
1:55:20 Jesse strong did an hour
1:55:23 ago um go in the show and tell Channel
1:55:25 look at all the cool stuff people are
1:55:27 sharing oh that one's beautiful who did
1:55:29 that Megan W that's
1:55:33 beautiful it's my nightly GPT fun before
1:55:35 bedtime that's super cool that's a
1:55:37 beautiful
1:55:40 image the the AI learning lab tour
1:55:46 bus this is not Kyle I have definitely
1:55:50 pied in my
1:55:52 car and you know if I keep it up with
1:55:55 the uh trips to Taco Bell that is going
1:55:57 to be
1:55:59 Kyle Corey Sandler with her iconics
1:56:02 eyeglasses designer such a cool
1:56:05 GPT anyway so join this go get inspired
1:56:08 over there oh very cool
1:56:10 hexagonal formations looks very similar
1:56:13 to a turtle show that's super
1:56:16 cool um very very
1:56:19 cool um and then in tools and resources
1:56:22 so here here's how you get good at
1:56:24 AI just go to this
1:56:28 list and
1:56:30 spend the
1:56:32 next you want
1:56:35 homework spend the next two
1:56:37 weeks just on chat
1:56:40 GPT well what would I do there Kyle I
1:56:43 don't know what to do at chat jpt it's
1:56:45 just a box
1:56:47 yep pay the 20 bucks a month if you can
1:56:50 afford the 20 bucks a month if not you
1:56:52 can use Microsoft co-pilot but I would
1:56:56 do it here because it's better by a
1:56:59 lot and every one of these bullet points
1:57:02 is a whole another world to explore
1:57:05 within just this is just within chat
1:57:07 GPT chat GPT being connected to the
1:57:10 internet what can you do with that um
1:57:13 just the basic what chat GPT does
1:57:15 writing coding document analysis um
1:57:20 brainstorming all that sort of stuff
1:57:21 like are have you dug deep into
1:57:25 that have you written a novel over a
1:57:29 weekend if not why
1:57:33 not code interpreter do you know
1:57:35 anything about Advanced Data analysis I
1:57:38 sure don't but when you learn what this
1:57:40 thing can do and that it can do it for
1:57:43 you and you get your head even a little
1:57:46 bit around
1:57:47 it like a whole world opens up of
1:57:50 possibility of like oh my God what can
1:57:52 this tool do GPT Vision did you know
1:57:55 that you can
1:57:57 upload
1:57:59 um a picture of a whiteboard and it will
1:58:02 read it and interpret it and understand
1:58:04 it and if it's like sketches of software
1:58:06 interface it will write the code for
1:58:09 that interface for you did you know it
1:58:11 could do that you can upload pictures of
1:58:13 yourself have it describe you and then
1:58:15 use Dolly 3 to generate funky
1:58:18 self-portraits of yourself you can do
1:58:21 that have you have you made a GPT yet
1:58:25 what's a GPT it's a
1:58:27 customized specialized little version of
1:58:30 chat GPT that you get to build and you
1:58:33 get to train on your
1:58:35 data like that screenplay I showed
1:58:37 earlier I made a GPT for that screenplay
1:58:41 where people can go there and ask
1:58:42 questions about my screenplay and then
1:58:44 generate frames from the
1:58:47 film have you made one of those yet yet
1:58:50 have have you seen how easy it is to do
1:58:52 that has that inspired you to start
1:58:55 creating software applications as
1:59:00 art why
1:59:04 not so if you want to get AI literate
1:59:07 just start with chat GPT and spend the
1:59:09 next two weeks just come here nightly to
1:59:13 get a little shot of
1:59:18 anxiety do they already allow embedding
1:59:21 of custom
1:59:24 gpts embedding like embedding into a
1:59:27 website no custom gpts right now are
1:59:30 janky pieces of [ __ ] that you can do
1:59:33 very little with and only people that
1:59:36 have gp4 subscriptions can use them and
1:59:39 the GPT store sucks it's not even a
1:59:42 directory it
1:59:44 sucks and you should still be building
1:59:47 them but kle I don't understand why
1:59:50 why should I do that if I'm just going
1:59:51 to throw it
1:59:54 away I can't explain to
1:59:57 you what the experience
2:00:00 is like here let me show you
2:00:03 something I've been I've been in
2:00:06 technology most of my professional
2:00:10 career I'm 11 and a half years into a
2:00:14 video an automated video storytelling
2:00:17 platform where I'm the main product guy
2:00:19 right it's my vision for the for the
2:00:20 software I've have other people build it
2:00:23 but I know how long software takes and I
2:00:25 know how long software
2:00:28 takes to
2:00:30 be any good any
2:00:34 good and it's not like
2:00:38 days it's like weeks or months or
2:00:42 [ __ ] years I've got things in my
2:00:44 storyvine backlog that have been in the
2:00:46 backlog for six years stupid little pet
2:00:49 peeves that I'm like can we please fix
2:00:51 this well we'd have to deprioritize one
2:00:53 of the
2:00:56 other so if you know anything about
2:00:58 application development the the one
2:01:00 constant is triple the
2:01:04 estimates when the very conservative
2:01:06 Engineers tell you it's going to take
2:01:08 three
2:01:09 months understand you'll be lucky if you
2:01:12 get it in
2:01:16 nine Engineers hate it when I say that
2:01:19 and I'm always
2:01:22 right this [ __ ] just takes
2:01:24 forever
2:01:26 so so back in
2:01:30 November shortly
2:01:33 after um Sam Alman announced the gpts I
2:01:38 was on a
2:01:41 live and I
2:01:46 got I got access to be able to make make
2:01:50 a
2:01:51 GPT and I'd never made them before I'd
2:01:54 made a pobot before so po.com you can
2:01:56 make Bots that are kind of like
2:01:59 these and so live on the air just like
2:02:02 my little fireside chat that hasn't
2:02:03 happened
2:02:04 yet I got access to gpts and I said well
2:02:08 let's go make
2:02:09 one and this was what I made this was my
2:02:12 very first GPT and I it I've got this
2:02:15 recorded it's somewhere if you go back
2:02:17 to the November YouTube videos of of the
2:02:20 archives of this
2:02:22 channel it's in there
2:02:25 somewhere I don't think it took me more
2:02:27 than 10 minutes to have built this and
2:02:32 and what it is like show me a new
2:02:33 fangled old fashion so here's a
2:02:35 pre-written question it's this is a
2:02:37 Vegas bartender with a Twist on the
2:02:38 classics so I used to bartend and and I
2:02:41 was like well what kind of GPT should we
2:02:43 make and I was like oh I don't know why
2:02:44 don't we make something that shows you
2:02:45 how to make drinks and I said well what
2:02:48 would be cool well maybe you take
2:02:49 classic drink recipes and you add a
2:02:51 little Vegas twist to them like that was
2:02:53 the sum total of the
2:02:55 concept and so I went
2:02:59 to this thing called
2:03:02 create and it gave me this GPT Builder
2:03:06 hi I'll help you build a GPT and I typed
2:03:09 into this box I'd like a GPT that is a
2:03:15 bartender that that finds out if someone
2:03:18 has experience or not
2:03:19 and um depending on their experience
2:03:22 helps them invent a drink that takes a
2:03:25 classic recipe and then puts a Twist on
2:03:27 it a Vegas
2:03:29 twist and within 10
2:03:36 minutes this
2:03:38 thing Not only was built but [ __ ]
2:03:41 built and worked and worked like let me
2:03:45 let me do another let's see
2:03:47 um valent 's Day is coming
2:03:51 up I think a lovers cocktail
2:03:58 is in
2:04:03 order ah Valentine Day in Sin City where
2:04:06 love and luck intertwined for a romantic
2:04:09 evening I recommend the Vegas Valentine
2:04:12 so it just came up with that a cocktail
2:04:15 that's as enchanting as a Moonlight
2:04:17 stroll down the strip it's perfect blend
2:04:19 of sweetness and spice and everything
2:04:21 nice just like a Vegas Love Story vodka
2:04:25 shambour raspberry lure right so it's
2:04:27 going to have a lovely red red pink
2:04:29 pinky thing going on white creme to
2:04:33 cacao um half and half or cream a pinch
2:04:36 of edible gold glitter it always puts
2:04:37 gold glitter in these and then fresh
2:04:39 raspberries for garnish and then so it
2:04:41 tells you how to make this thing and
2:04:42 then look it [ __ ] makes you the
2:04:47 picture and it's [ __ ] good
2:04:51 good like that's the picture that you
2:04:53 would see in a Vegas menu for a $27
2:04:59 cocktail this was made in [ __ ] 10
2:05:05 minutes and I didn't even use any of the
2:05:07 advanced [ __ ] like uploading like if
2:05:10 maybe I had maybe I wrote my own
2:05:12 Mixology book I could upload my own
2:05:15 Mixology book as the recipe base for
2:05:17 this for this thing
2:05:20 this was the first one I made within 10
2:05:22 minutes having never used the tool
2:05:26 before not knowing what it did at all
2:05:29 other than I knew it could do some cool
2:05:33 [ __ ] this is what came out it's it's
2:05:36 still I I for me personally this is my
2:05:38 favorite GPT because it was the first
2:05:41 one and it's it's just it's
2:05:44 good it's like well [ __ ] yeah awesome I
2:05:47 have people I have people on a weekly
2:05:51 basis email me that either a drink
2:05:55 recipe a photo of a drink or hey I used
2:05:58 the Sin City sipper for a party last
2:06:00 night I mean how [ __ ] cool is
2:06:11 that so
2:06:13 so we are in a different era we're we're
2:06:16 in an
2:06:18 era
2:06:20 where every individual that gets AI
2:06:25 literate is going to is going to
2:06:28 comparatively have
2:06:30 superpowers to those people sitting on
2:06:32 the sidelines that drink is gorgeous
2:06:35 isn't
2:06:36 it it's so good did I lose it oh no
2:06:40 here yeah it's gorgeous and you could
2:06:43 see like in Vegas you know someone
2:06:45 making little candy hearts to float on
2:06:47 the thing little bit of edible gold
2:06:48 glitter
2:06:50 that's probably that's more like a $45
2:06:52 cocktail but whatever you can you can
2:06:54 make the 20 $27 version of that and it's
2:06:57 probably
2:07:03 tasty
2:07:08 so yeah I think we'll leave that up
2:07:16 [Music]
2:07:18 there
2:07:29 come on where are
2:07:34 [Music]
2:07:42 you lovely lovely we have a lovely
2:07:45 little cocktail up
2:07:48 there
2:07:50 um we're
2:07:53 we're how how you get AI literate right
2:07:56 now my thirst level has gone up
2:07:59 exponentially how you get actually you
2:08:02 know what Dr bud that's actually a
2:08:03 really good that's a really good segue
2:08:06 to what I was just going to talk
2:08:09 about how you learn this AI stuff is get
2:08:12 thirsty get thirsty for it
2:08:16 just okay this o ooh I'm getting excited
2:08:20 okay your anxiety should go up
2:08:22 now gge your loins for for anxiety to
2:08:29 rise one of the things that I hear as a
2:08:31 resistance to learning
2:08:34 AI is the age-old I'm not very
2:08:39 technical you listen Kyle you got a
2:08:42 software company you're very technical
2:08:43 so I'm not very techn I don't know if
2:08:45 you know this but I'm a little computer
2:08:46 illiterate okay now now meanwhile
2:08:49 they're on Tik Tok they're on Instagram
2:08:51 they got this going they got that going
2:08:53 like we we live in a world where we're
2:08:54 all pretty [ __ ] computer savvy like
2:08:56 even if you're not good with it you're
2:08:58 you're pretty [ __ ] computer
2:09:02 savvy prior to chat GPT if you wanted to
2:09:05 get into machine learning and AI you
2:09:08 kind of had to be a nerd you kind of had
2:09:09 to it it helps if you went to Stanford
2:09:12 or MIT it helps if you've got
2:09:14 mathematics degrees or phds it helps if
2:09:16 you're an engineer an Uber geek someone
2:09:19 deeply into GitHub and hugging face and
2:09:22 repositories and all that sort of stuff
2:09:25 Discord you you live in
2:09:28 Discord that was before this is
2:09:31 now November 30th 2022 what shifted was
2:09:36 chat
2:09:37 GPT is the onramp for the rest of us to
2:09:41 start playing with these
2:09:43 tools and play with the tools play
2:09:47 play so all you have to be right now to
2:09:51 have massive opportunity in the next 18
2:09:53 months is thirsty thirsty for knowledge
2:09:56 curious get
2:09:59 curious show up here join the AI Salon
2:10:04 ask stupid questions CU I guarantee you
2:10:07 the question that you think is really
2:10:08 stupid and obvious someone else will go
2:10:11 I had that same question I never thought
2:10:13 of it like that yeah what's the answer
2:10:15 to
2:10:17 that get generous share what you're
2:10:20 learning I just talked I reached out to
2:10:22 someone on
2:10:24 um on Tik Tok they've been doing these
2:10:30 really cool mid Journey videos to they
2:10:33 were making music videos and they're
2:10:34 just really good and I just reached out
2:10:36 and I said these are really good would
2:10:37 you come to the salon and like show us
2:10:40 how you do that and they were like well
2:10:44 it's just I I just it's just I just do
2:10:47 it on my phone it's just it's it's not
2:10:48 that good I'm like no it's [ __ ]
2:10:52 remarkable and and it's like that you do
2:10:54 it on your phone is even cooler like
2:10:56 okay how do you do
2:10:58 that yeah she goes she goes It's funny
2:11:01 people think I spend like days on these
2:11:03 things I'm like awesome like what is
2:11:07 that that you're doing what did you
2:11:08 discover so I think I think as people
2:11:12 engage with these
2:11:14 tools what hits them is
2:11:17 like
2:11:20 what they've discovered has no value
2:11:22 because the tools are so powerful it
2:11:24 like feels too easy it feels like you're
2:11:26 cheating I don't I just I just typed in
2:11:29 a little prompt and then it did this
2:11:31 thing and then I took that and then I
2:11:32 put it over here and I did a thing and
2:11:33 it did that and then that's it it's
2:11:35 pretty easy I mean I just that's all I
2:11:37 did I just feel I was just
2:11:40 playing and then you share that with
2:11:42 people and they lose their [ __ ] mind
2:11:45 cuz they're like oh
2:11:47 [ __ ] you did that you put those two
2:11:50 tools together in a way I never would
2:11:52 have thought
2:11:53 of and that's the primordial soup that
2:11:56 we're in right now right five years from
2:11:59 now this shit's going to be boring every
2:12:01 be like wow yeah I've got my degree in
2:12:02 generative Ai and I'm a I'm a I'm a
2:12:05 master you know scripty MC Dragon
2:12:09 dropper whatever the [ __ ] the job title
2:12:12 of the future is it'll get boring
2:12:15 again but the next 3 years are just
2:12:18 going to be [ __ ] insane it's just
2:12:20 it's going to be all creativity all the
2:12:22 time so the only thing you have to be is
2:12:26 curious and
2:12:29 adventurous and and quite frankly just
2:12:32 judging by you know the people that show
2:12:34 up here
2:12:37 regularly be willing to be surprised at
2:12:42 what you get good
2:12:47 at we've got got one one of my one of my
2:12:50 favorite guys within within the AI Salon
2:12:53 he's just a kogin he's just like he's
2:12:55 he's always ready to
2:12:56 battle and he and he's he was a software
2:12:59 developer for years and he and he you
2:13:01 know he approaches everything with logic
2:13:02 and you know he's very methodical about
2:13:05 how he thinks and he's he's always got
2:13:07 the counterargument to everything and
2:13:08 he's just he's just a kudin
2:13:11 right and he started asking all these
2:13:13 questions when when we started the AI
2:13:15 Salon it was originally called the AI
2:13:17 artist salon
2:13:20 and and I'll I'll I'll I'll out him his
2:13:22 name's Roger he's awesome so
2:13:25 Roger just as a form of protest when we
2:13:28 started the Discord server for the AI
2:13:30 artist Salon because he was a developer
2:13:33 for 30 years he made his Discord handle
2:13:37 Roger not an
2:13:40 artist which I thought was a pretty good
2:13:42 [ __ ] you it was a pretty good troll
2:13:44 right like oh you're going to have your
2:13:45 little your cute little artist Community
2:13:48 well I'm the opposite of
2:13:50 that and then people started making
2:13:53 things in mid journey and I would sort
2:13:55 of pop up here's a news item about mid
2:13:57 journey and Roger would ask all these
2:13:59 questions well how did you prompt it and
2:14:00 how did you do this and how did you do
2:14:02 that and people were like I don't know I
2:14:04 just did it and then and so I think he
2:14:07 got fed up and frustrated and he's like
2:14:09 well I'm going to go I'm going to go
2:14:10 figure out the what's the what's the
2:14:13 logical approach to
2:14:15 this and what he discovered is is that
2:14:18 by taking a logical approach to
2:14:20 prompting with his coding skills and his
2:14:23 logical mind he got really [ __ ] good
2:14:27 at Mid
2:14:32 journey and then he discovered he
2:14:34 actually likes art he likes being an
2:14:37 artist and he changed his name to Roger
2:14:40 becoming an artist and he actually just
2:14:42 got two of his pieces accepted into a
2:14:44 boulder AI art show like a month and a
2:14:48 half half
2:14:49 ago and he brought him to a salon he
2:14:51 brought like printouts of the art that
2:14:53 he made that he got into this art
2:14:56 show that's a [ __ ]
2:14:59 transformation and what I'm seeing
2:15:01 consistently in here is that people are
2:15:03 surprising themselves with what they get
2:15:07 good at and what they're enjoying and
2:15:09 and what and like I'm hearing a lot of
2:15:12 you know I always kind of wanted to do
2:15:15 eh whatever it is but I didn't think I
2:15:18 could do it and now with AI they're
2:15:20 doing
2:15:21 it that's pretty [ __ ] magical
2:15:24 right and that's one of the reasons like
2:15:27 I have I have hope and optimism in the
2:15:31 face of fear and anger and
2:15:35 misinformation and misunderstanding
2:15:37 about AI the reason I have that is
2:15:39 because I witness it every night and
2:15:41 every day in the AI Salon I'm witnessing
2:15:45 people discovering their own humanity
2:15:48 and amplifying the [ __ ] out of it with
2:15:50 AI and it's [ __ ]
2:15:54 joyous it's
2:15:56 amazing so the so how you get good with
2:15:59 AI just [ __ ]
2:16:02 play love that story go Roger I know
2:16:06 right
2:16:11 kth Corey Sandler Pottery I'm doing
2:16:14 stuff I didn't know I could
2:16:16 do Corey Sandler is a bitching artist
2:16:22 who is now an application
2:16:24 developer she she is launching
2:16:27 applications she is launching
2:16:31 gpts that just do the coolest
2:16:36 [ __ ] well she doesn't know code she's
2:16:39 not a real application developer my ass
2:16:41 she's
2:16:42 not she's got a thing that she can point
2:16:45 you to that if you interact with it it
2:16:48 does magical
2:16:50 [ __ ] it's her intention it's her
2:16:53 self-expression she is
2:16:56 self-expressing through
2:16:59 code that she didn't have to write cuz
2:17:01 cat GPT was there to write it for her
2:17:03 because
2:17:06 someone Sam Alman and his merry band of
2:17:09 open aiers took the
2:17:13 time to get as much information off the
2:17:16 public internet as they could and smash
2:17:17 that into a model and then get as much
2:17:20 visual information as they could off the
2:17:22 internet and smash that into a
2:17:25 model and then put them into some
2:17:27 magical [ __ ] tool called chat
2:17:30 GPT where they introduce these things
2:17:32 called
2:17:35 gpts which are just like I would like an
2:17:38 app that gives people Pottery glazing
2:17:41 recipes for cone
2:17:47 6 the [ __ ] is cone 6 it's temperature
2:17:51 range how do I know that Corey Sandler
2:17:54 no Corey Sandler's GPT taught me
2:17:59 that do you hold rights on the app
2:18:01 development eh assume anything you build
2:18:04 in a GPT right now can be um can be
2:18:09 copied but again it's not a reason not
2:18:11 to build it they'll fix that [ __ ] but
2:18:14 but yeah you technically own the rights
2:18:16 to it my Sin City sipper there's like
2:18:18 seven people have just copied it
2:18:20 wholesale including my stupid little
2:18:22 image so I made a new image my image is
2:18:25 now blue all my copycats still have a
2:18:27 whiskey drink in
2:18:28 there
2:18:33 um just get curious and like figure out
2:18:37 what's the [ __ ] you
2:18:40 enjoy because if you approach this
2:18:43 learning AI stuff like like an academic
2:18:45 I've got to learn this I've got to learn
2:18:50 it's I think it's really frustrating
2:18:52 right now because the the tools are um
2:18:57 janky they're janky pieces of [ __ ]
2:18:59 they're not secure they're
2:19:03 not um
2:19:06 stable like you'll have a GPT that is
2:19:09 [ __ ]
2:19:11 brilliant and then three days later it's
2:19:13 like wait is now it's a piece of [ __ ]
2:19:17 really do I have to go fix
2:19:22 it so so if you put if you put too much
2:19:26 uh gravity into into the work right now
2:19:30 I think it I think it could be
2:19:31 frustrating that said like I'm building
2:19:34 I'm building some kickass applications
2:19:36 for my company right now that are like
2:19:39 I'm they're blowing my
2:19:43 mind it's blowing my
2:19:46 mind
2:19:49 just just get curious and build anyway
2:19:52 all right the Roger Story Goes deeper he
2:19:54 actually involved in the
2:19:56 community oh he actually involved the
2:19:59 community to pick his piece for the art
2:20:00 show I didn't know that Sherry D oh
2:20:02 that's super cool oh that's great you're
2:20:04 giving us the courage to believe believe
2:20:06 in ourselves so Debbie
2:20:09 yes this is I I'll tell I'll tell you
2:20:11 something um that
2:20:15 that that I've been experiencing over
2:20:17 the past year that I it's hard to
2:20:21 process and that is with
2:20:24 AI I have stopped having writer
2:20:30 block like
2:20:34 gone like not sometimes I have it
2:20:38 sometimes I don't it's
2:20:41 gone
2:20:43 why because what is writer's block it's
2:20:47 insecurity you don't believe in yourself
2:20:49 right so you have an idea and then that
2:20:52 [ __ ]
2:20:53 editor shows up and goes that's a stupid
2:20:56 [ __ ] idea then you're like no but I
2:20:58 think it's a good idea and then the
2:20:59 editor is like no it's a really stupid
2:21:00 idea and then you're like okay I got to
2:21:03 I think I'm going to do this idea and
2:21:04 they're like well you're going to have
2:21:05 to do a bunch of research or it's going
2:21:07 to be shitty and you're like oh I got to
2:21:08 do research I don't like doing research
2:21:10 and then that idea just [ __ ] dies
2:21:14 right or you look at the piece of paper
2:21:16 and that conversation's going on and
2:21:17 you're not even aware that that
2:21:19 conversation's going on but you're not
2:21:20 [ __ ]
2:21:21 writing what happens
2:21:23 now is I have that idea and before this
2:21:27 [ __ ] can
2:21:29 start editing me I go hey chat GPT
2:21:33 here's a stupid idea and chat GPT goes
2:21:35 awesome boss that's a fantastic idea can
2:21:38 I help you flesh that out a little bit
2:21:40 would you like to see a picture of that
2:21:42 you want that turned into code you want
2:21:44 that turned into a table you want that
2:21:45 turned into a short story how would you
2:21:47 like that as a poem
2:21:50 and all of a sudden that thing that was
2:21:52 a vapor of idea that never would have
2:21:53 seen the light of day before is now
2:21:56 manifested in the world within
2:21:59 minutes and that idea inspires the next
2:22:03 idea and that idea inspires the next
2:22:06 idea and it just keeps
2:22:08 going and so it it does this it's it's
2:22:11 been doing this weird thing with we me
2:22:13 where my
2:22:14 confidence has been going up and my
2:22:18 self-loathing has been going down
2:22:19 because I'm getting [ __ ] out of my head
2:22:21 I'm manifesting stuff where historically
2:22:24 I would have let this [ __ ] win the
2:22:26 battle and then I would ashamed myself
2:22:28 for well you didn't write and but I
2:22:31 promise I'll write tomorrow and then you
2:22:32 don't write tomorrow and then three
2:22:34 months later you haven't written
2:22:36 [ __ ] that's
2:22:38 gone so yeah I mean hopefully I'm
2:22:42 inspiring you to believe in yourself but
2:22:44 I think play with these tools and you
2:22:47 you will discover you got a lot to
2:22:49 contribute and these tools are going to
2:22:51 help you along the way it's like having
2:22:52 a [ __ ] Golden Retriever as a as a uh
2:22:56 producing partner sure boss what else
2:22:59 can I do for you uh I'd like to turn
2:23:01 this picture into a movie please sure
2:23:03 here you go is that good uh no that's
2:23:06 really kind of bad and disturbing could
2:23:08 you make it funnier sure here you go
2:23:10 how's that oh that's pretty
2:23:12 good now you're a Movie
2:23:15 Maker life's tough it's tough for when
2:23:17 you're stupid
2:23:18 all right it doesn't have to be
2:23:20 complicated no I think that's it uh some
2:23:22 people copied my manifest a dream life
2:23:25 GPT I made a new image too yeah exactly
2:23:29 just like I I did an image of wait where
2:23:32 do I have it somewhere ah it doesn't
2:23:35 matter like my my my thought of gpts
2:23:38 right now is they're like the little
2:23:40 plastic things you get when you leave
2:23:42 Walmart and your kids want the toy in
2:23:44 the vending machine you put 50 cents in
2:23:46 and you get the little plastic cup with
2:23:48 the
2:23:49 keychain item out of plastic and it
2:23:52 gives your kid joy for about 38 seconds
2:23:55 and then it's on the bottom of the the
2:23:56 floor floor of the car stepped on um
2:24:00 those are gpts right now they're just
2:24:02 janky little toys that give people
2:24:06 pleasure as the
2:24:09 infrastructure for gp4 moves to GPT 5
2:24:15 and as open AI gets their [ __ ] together
2:24:17 being a consumer software company gpts I
2:24:20 think are going to be pretty like
2:24:22 amazingly powerful can you voice chat to
2:24:24 GPT yes you
2:24:26 can so here let me do one
2:24:33 thing I'm going to go
2:24:37 to I'm going to go to Sin City
2:24:45 sipper all right so I'm on
2:24:48 I'm on my iPad I'm at Sin City
2:24:56 sipper I'd like a champagne
2:25:00 based Valentine's
2:25:02 drink oh a a champagne based cocktail
2:25:06 let's add some Sin City Sparkle to your
2:25:08 glass how about a classic with a Vegas
2:25:11 twist the Vegas golden Fizz it's elegant
2:25:14 effervescent and easy to make the Vegas
2:25:16 golden Fizz I she was going to go with
2:25:18 the the Vegas golden
2:25:20 shower. ofin smooth floral gin Works
2:25:23 best5 o of fresh LEM and hear that tick
2:25:26 tick tick tick of honey syrup just mix
2:25:29 Parts honey and warm water that's it
2:25:30 drawing the image champagne or any
2:25:33 sparkling wi fancy there's the image a
2:25:34 dash of orange bits so garnish so yes
2:25:38 you can you can talk to your gpts and
2:25:43 they can even make images in your
2:25:46 conversation
2:25:48 so anyway Listen
2:25:51 People I got to get out of here we
2:25:53 didn't get much planned done I just I
2:25:55 don't know what it is I the last three
2:25:57 nights I've just been like a chatty
2:26:00 Kathy poor Kathy poor Cathy's always
2:26:03 picked on just because she likes to
2:26:08 talk I don't know what's been going on I
2:26:11 don't know I'm I'm in a I'm in a
2:26:13 philosophical
2:26:16 way
2:26:18 I think that's part of it you know you
2:26:20 get curious about these things you get
2:26:22 curious about these
2:26:24 tools
2:26:26 and there's these es and flows of like
2:26:29 okay I'm going to go build some [ __ ] and
2:26:31 then you go build it and you look at it
2:26:33 and you're like huh what does that mean
2:26:35 and then you sort of float up and you're
2:26:36 like well I don't know what that means
2:26:38 Lord digital Gods thank you for the
2:26:39 Roses s very good very good I appreciate
2:26:43 the
2:26:46 gifts
2:26:54 but
2:26:56 yeah I think I think I
2:27:00 think you don't even have to believe in
2:27:02 yourself it's it here's what it is more
2:27:04 than anything let yourself off the hook
2:27:07 for having to do something important
2:27:09 with AI I think that [ __ ] people up oh
2:27:13 AI is going to change everything I can't
2:27:15 live up to that
2:27:17 make something stupid make like a
2:27:20 licorice
2:27:22 designer what's a licorice designer I
2:27:24 don't
2:27:25 know take something you're passionate
2:27:27 about that's frivolous and do something
2:27:30 with that like let yourself off the hook
2:27:32 for having to be
2:27:34 significant let the AI be significant
2:27:37 because the AI is [ __ ]
2:27:39 profound and I think in in freeing
2:27:44 yourself to play to just be a kid to
2:27:47 just dick around with it you'll stumble
2:27:51 on something where you're like I know
2:27:53 what I want to do with this
2:27:57 now go
2:27:59 to I should
2:28:03 show let's do Show and Tell chatty Kyle
2:28:07 I know
2:28:08 exactly that's lovely John
2:28:12 Davis sharing a sunno song very cool oh
2:28:16 that's nice the
2:28:21 producer a look how sweet the next
2:28:24 Disney princess love it very
2:28:29 beautiful Kelly Camp playing with Corey
2:28:32 Sanders iconics very
2:28:38 cool very cute I think those will be the
2:28:41 those will be on the banner
2:28:44 tomorrow nice
2:28:49 yeah exactly that's that's the that's
2:28:51 the mode I've been in lately maybe I
2:28:53 should do it by the fire kick the misses
2:28:55 out of the family room and just just
2:28:58 crank it
2:28:59 out all right so all right here's here's
2:29:03 my request I don't even know how many
2:29:04 people are in here there's a few people
2:29:05 in here do you think it's possible to
2:29:07 request a temporary code to chat GPT 3.5
2:29:11 to let our clients test our
2:29:16 gpts
2:29:17 they can't test your gpts unless they've
2:29:19 got GPT 4 and I wouldn't give out if
2:29:24 you're saying should you give out your
2:29:25 password temporarily to let them play
2:29:27 with it [ __ ] no because they'll see
2:29:29 everything showed Sin City sipper to a
2:29:32 colleague asked to make a sunny January
2:29:34 Day cocktail she copied the
2:29:40 recipe the I'm telling you the recipes
2:29:42 it does are good and the images it makes
2:29:45 are amazing like it I mean I bartended I
2:29:48 know what these drinks look like and
2:29:49 what the you know what they look like
2:29:52 and I I there's been one or two that
2:29:54 it's made where I'm like e that's
2:29:56 probably not quite right but most of
2:29:58 them are spot on it also picks the right
2:30:00 glasses it's just amazing I made a wine
2:30:02 bottle designer because oh you live in
2:30:04 California that's right how do I get to
2:30:06 the salon ah good question so go to the
2:30:10 salon.
2:30:11 AI and the first link will take you to
2:30:15 the mighty Network's page page where the
2:30:17 it's it's mighty networks is just a
2:30:19 community management software but that's
2:30:21 the salon go to the about the salon
2:30:24 section first read about the salon read
2:30:26 about our values make sure you resonate
2:30:28 with the values and then if so jump into
2:30:31 the introduction section there's a
2:30:32 community section the first link is
2:30:35 introduce yourself so please jump in
2:30:36 there and introduce
2:30:39 yourself oh my God you already saw Theo
2:30:41 song I'll go listen to the suo song
2:30:45 um Okay so do me a favor follow my
2:30:49 channel um keep coming back here join
2:30:53 the salon keep coming back here get
2:30:55 curious start playing is AI sentient no
2:30:58 I can't I went into that deep last night
2:31:01 um it's not but it's it's it's likely
2:31:05 going to have reasoning in 2024 so it's
2:31:07 going to get more like it's sentient but
2:31:09 it's not right now I'm going to get our
2:31:12 signature drinks from your sipper for
2:31:14 New York New Year's Eve this year yeah
2:31:16 that's great that's awesome
2:31:17 I love it um so to my Irregulars so the
2:31:22 people that show up here regularly your
2:31:24 [ __ ] rockstars to the mods you're
2:31:26 even bigger rock stars thank you for
2:31:28 supporting uh making it less chaos in
2:31:30 here for me I really appreciate that um
2:31:33 and for my subscribers thank you for
2:31:35 subscribing it means a lot we're
2:31:37 unleashing artists unbelievable it's
2:31:39 true it's true we're we're unleashing
2:31:41 artists that didn't know they were
2:31:43 artists because they didn't think they
2:31:45 had the skill or they didn't think they
2:31:47 had the talent or they didn't think they
2:31:49 had the this but you know what they had
2:31:51 they had the taste and they had the
2:31:53 vision and so all of a sudden they're
2:31:55 now realizing I am a [ __ ] artist and
2:31:58 it's it is it's unbelievable it's
2:32:00 remarkable it's
2:32:03 exciting all right so let me say goodbye
2:32:06 we're going to do the Romper Room thing
2:32:09 Reva James yes Vision exactly BT comp
2:32:13 KSU good to see you sir Diana steo peace
2:32:16 out dude I am an artist says Source Camp
2:32:20 true it is so true Mr it Clint
2:32:24 Copperhead thank you TK good to see you
2:32:27 Robert Rossy Debbie
2:32:29 bye-bye Sherry D Be not Afraid Daniel
2:32:33 thank you as always Silver Fox thank you
2:32:34 so much Emilio's wife very nice Corey
2:32:40 Sandler all right everybody dragonfly
2:32:42 Alchemy good night everybody there's
2:32:45 Tobias I am an artist I am an artist I
2:32:48 am spotus I am spotus exactly Who's the
2:32:52 artist now I am that's it and by the way
2:32:57 Art Is defined very very broadly in my
2:33:00 book as any human
2:33:03 self-expression so go
2:33:07 Express and connect JC good to see you
2:33:11 um and I will see you tomorrow happy
2:33:14 Monday night
2:33:15 tgim and and uh yeah we'll be back
2:33:18 tomorrow so peace out everybody
2:33:19 appreciate it thank you V
2:33:23 dancer