Dec 15, 2023 - Exploring Pika, Art, and the Future of Content Creation

Description

Chapters

Transcript

0:02 you want to do live
0:03 barking you want to do live
0:06 [Music]
0:25 [Music]
0:29 singing
0:32 [Music]
0:42 [Music]
0:50 stand
0:52 between you and place is
0:58 insane stand to
1:02 you makes
1:05 [Music]
1:07 it trouble to
1:13 me trouble can't you
1:16 [Music]
1:18 see happy happy happy Friday night
1:23 [Music]
1:29 everybody
1:32 [Music]
1:33 [Applause]
1:37 [Music]
1:50 yeah champy says
1:53 [Music]
1:58 hello Sola
2:03 [Music]
2:13 Casey in the learning learning lab band
2:16 that's not bad I like it I like it
2:19 champy likes it it's late everybody I
2:23 know it's late my hair is
2:26 [Laughter]
2:28 pointy oh lordy lordy lordy what are we
2:32 going to do
2:34 tonight we got Pik labs to play with
2:37 pika's out if you haven't seen that it
2:40 looks like they've released 1.0 access
2:42 to
2:43 Pika um I haven't played a lot with it
2:45 so we can go play with that and see if
2:46 it's uh better than Runway I have shitty
2:49 luck with Runway
2:51 so it would be nice if Pika wasn't
2:56 shitty it's janky time it's janky time
3:01 now is the time it's jankie
3:04 time still on the Pico wait list check
3:07 your
3:11 email um it's weird someone sent someone
3:14 sent me a note on the Tik Tock that said
3:18 they got Pika so I checked my email and
3:21 my note was different than theirs theirs
3:23 says you got access to P 1.0 and mine
3:26 just said you're off the wait list but
3:28 it didn't mention 1.0 but I think it's
3:30 1.0 so p. art is that it yeah sign in
3:35 with
3:37 Google sign in with the
3:40 Google yeah there we go live describe
3:43 your story image or video I just got a
3:47 video I know I'll I'll move it over hang
3:50 on calm down
3:55 people describe your story let's see um
4:04 clation
4:06 vampire go all right let's see we'll see
4:10 if it does
4:11 anything all right it's making
4:16 something and here I've got
4:27 a I made a 7c panning shot of a ancient
4:33 Mayan ruin and it looks like an ancient
4:35 Mayan ruin so all right we can play with
4:38 that stuff we'll do that my invitation
4:41 must have gotten lost in the mail I know
4:43 you know these these um these stupid
4:47 flipping weight list things are just
4:49 kind of cuckooing for the birds it's
4:52 just it's
4:54 like it's just like let us play I know
4:58 it cost you money but that's not our
5:01 problem just give us free [ __ ] we want
5:05 the thing to be perfect and perform
5:07 Miracles and we don't want to pay for it
5:09 this is
5:13 stupid happy Friday night everybody my
5:15 name is Kyle Shannon this is the AI
5:17 learning lab that's a sign for the AI
5:19 Salon but we'll talk about that too oh
5:21 look merch I got my AI Salon shirt
5:25 today I love our logo it's got a little
5:27 spot on it from uh from the factory but
5:30 [ __ ] it I don't care it's t-shirt
5:32 t-shirt t-shirt t-shirt whatever but
5:35 that's exciting isn't it I love that
5:37 logo that logo I'm very happy with very
5:39 very happy with the logo thank you to
5:42 Bradley Perkins for working on that with
5:45 me putting the merch store together
5:47 putting our Mighty networks together
5:50 Bradley's been working his ass
5:53 off oh there's a cool cookie tin I think
5:56 that's from
5:57 Ted Ted Ted Ted that's a Christmas
6:00 cookie tin for the AI learning Lab look
6:03 at that look how pretty that is that's
6:06 so nice that's so nice let's see how we
6:10 doing
6:11 this we're going there and then we're
6:14 going there and then we're going there
6:16 going there all right I'm just getting
6:18 my professional setup set up here um all
6:23 right so here's here's here's what we we
6:26 should do do me a favor if you're
6:29 willing
6:30 you don't have to do this you can just
6:32 sit there and lurk I'm a I'm a lurker
6:35 I'm a lurker on a lot of these lives you
6:37 know you don't have to do [ __ ] but if
6:39 you want to do something share the live
6:42 like share the live with I don't
6:45 know five of your
6:47 friends which five
6:50 friends the ones that will annoy if you
6:53 send them something about AI I think
6:55 that would be
6:58 fun cuz I'm just I'm just ornery I'm
7:03 ornery damn it had a good day
7:07 today I don't think I did a lot
7:14 particularly doing a seminar with
7:16 someone about leading an extraordinary
7:19 life so things in here are going to get
7:21 much more extraordinary I'm about a
7:23 third of the way through the thing I
7:25 think it goes through February so by
7:27 February you're going to be looking at
7:30 Kyle
7:32 extraordinary I'm at a movie theater
7:34 listening to your
7:40 show well that's listen there's
7:44 Spielberg and then there's the AI
7:46 learning lab you got to you got to you
7:48 got to balance your life you can't it
7:50 can't just all be Blockbusters right you
7:54 need you need different levels of
7:56 entertainment and this is a level of
7:58 entertainment
8:02 I I will let you all categorize where
8:05 this sits although some people said this
8:07 is better than TV which I'm like what
8:09 the [ __ ] is TV
8:12 doing which by the way I'm apparently
8:14 allowed to swear now they they age
8:16 restricted my channel the other night
8:18 major laser um Kyle do you think the AI
8:22 robots will spare me if I'm kind to chat
8:24 GPT now I would say I would
8:28 say couple of
8:31 things Ethan Ethan mullik posted a study
8:34 about this apparently sort of just
8:35 traditional psychology works on these
8:38 things being nice to them
8:41 Works um I met a great guy today from
8:44 from he works at Google and Boulder um
8:47 Todd I forget his last name but he works
8:49 in the AI group um really really really
8:52 smart guy and we we were talking about
8:54 all this stuff about um about being
8:58 being nice to the robot and he was
9:00 talking about there's that thing
9:03 um there's a there's a prompt you can
9:06 give chat GPT take a deep breath and
9:09 then do whatever your prompt is and it
9:12 it
9:12 improves the results like this is like
9:15 they've done a study and take a deep
9:18 breath was the sort of number one thing
9:20 you could say to chat GPT to improve the
9:22 prompting we were talking about like
9:23 what what is it about that like why why
9:26 would that work and the only thing I
9:28 could come up with was
9:31 if you think
9:33 about who who would who would write down
9:36 take a deep breath well it's probably
9:38 like self-help people like you know take
9:41 a deep breath and then consider your
9:42 life and have a moment and you know
9:45 whatever it might be so maybe when you
9:47 say take a deep breath it puts chaep in
9:50 some sort of semantic region of the
9:53 vector space where there's some
9:56 instructions there's there's some
9:57 instructions about how to
10:00 considerably lead your life that chat
10:02 GPT turns that into I don't know some
10:05 sort of better way of giving answers I
10:08 don't [ __ ] know it's just all weird
10:10 man have you tried Pinocchio makes it
10:12 easy to install AI locally I tried it
10:16 uh maybe two months ago someone
10:18 mentioned it and I and I kind of
10:20 installed it
10:22 um I did I haven't played with it a lot
10:25 so I know it's there I have I have
10:27 installed successfully installed um GPT
10:31 for
10:32 all and then I just saw that there's a
10:35 lot of noise right now about the about
10:36 the mistol um uh models that are out
10:40 there people are installing them on
10:41 their phones things like that take a
10:43 deep breath help helps with spelling
10:46 yeah I don't know it's very
10:48 strange this this AI stuff's weird so
10:51 but we're we're leaning in on it people
10:54 we're going with
10:57 it we are flipping going with
11:01 it all right so how this works um if
11:04 you're new to
11:06 AI um where I would suggest you go right
11:09 off the bat is hang on what did I just
11:13 do did I just delete
11:16 something I think I did no I didn't
11:24 weird um what I would suggest you
11:28 do
11:30 if you are brand brand new to this AI
11:34 stuff is
11:37 go get the to a Nary
11:41 go get the to chat
11:46 GPT chat. open.com so if you're brand
11:50 new to to ai go there if you've never if
11:54 you've never been to Chad GPT and you're
11:56 like oh I'll get to it
11:58 tomorrow I'm not very technical Kyle I
12:01 don't know if you know this but it's
12:02 it's very late on a Friday night I don't
12:05 feel like learning something at the AI
12:07 learning lab okay that's fine you can
12:10 say that but go there anyway just go
12:12 there right
12:13 now and ask it to write you a poem or
12:16 ask it to write you a a recipe uh for
12:19 chicken picata where you substitute the
12:22 chicken with pork I don't know doesn't
12:25 matter go ask it to create a project
12:27 plan for you for your whatever business
12:29 you're thinking of
12:31 starting so you can quit your day job
12:34 just go there right now and start
12:35 playing and then it is my request it is
12:37 my sincere hope for you as a human being
12:40 that you play with it long
12:43 enough that you have your first Kevin
12:45 mallister
12:47 moment that
12:50 moment
12:52 and at that point then you've taken the
12:54 red pill and you can't come back from it
12:57 but if you just sit there and think like
12:58 oh it's just like Google it's whatever
13:00 it's it's cute it's
13:02 [Laughter]
13:04 cute uh then you know then it's not as
13:06 much fun and you're not going to you can
13:09 make money
13:10 with everybody wants that everybody
13:12 wants a gold button you think I just
13:15 bought this off Amazon no this is
13:17 hard-earned how did I get this cat
13:21 GPT I made money with it and then and
13:24 then bought this is solid gold it's not
13:26 a Chinese piece of plastic
13:29 I got this from Parker from Gold Rush he
13:32 smelted this for
13:33 me and then Mark rooper the the the guy
13:37 that does the the the fart bomb things
13:41 he he made the electronics in this for
13:42 me it's not from
13:45 Amazon I couldn't be this successful
13:48 just buying crap from Amazon no so go
13:51 there and do that your life will get
13:52 better you'll get a gold button you you
13:54 will have the capacity to get your own
13:58 solid gold
13:59 you can make money with chat gbt button
14:01 digital Gods that's very very generous
14:04 thank you so much you are a title gifter
14:07 I don't know what that means but
14:09 congratulations on the title do I need
14:11 to call you sir sir digital Gods I think
14:14 I think it's appropriate at this
14:17 point oh man whoever smelled it dealt it
14:21 we can make money with chat GPT exactly
14:24 exactly
14:27 Nexus so go there do that just go play
14:30 right now dude I asked GPT one time if
14:32 it would spare when they enslave
14:36 humanity and it said yeah sure yeah so
14:40 be oh so that was the question was about
14:42 should I be nice to these things so one
14:43 is be nice to them because they work
14:45 better but the other one is these models
14:48 right now with the exception of Pi which
14:50 is starting to do this um they don't
14:53 really have memory across chats um but I
14:56 would imagine fairly early in 2024 chat
14:59 C GPT will have memory and the other
15:01 ones will start to have memory where it
15:03 will start to remember who you
15:05 are and why not why not be light
15:11 nice Lord digital Gods I okay see you
15:15 get a title and all of a sudden it's
15:18 just it goes right to your head
15:20 unbelievable I tried to give him a sir
15:23 he went straight to Lord we're now we're
15:24 nighting people here at the
15:26 salon or the what is this the learning
15:29 lab this is the salon this is the
15:32 learning lab wherever we are
15:36 here all right so how we do this if you
15:38 have questions about AI if you have done
15:42 anything interesting I'd love to hear
15:43 that can you do a mistal walkthrough uh
15:46 I can't just because I haven't installed
15:48 it I don't know what to do and like I
15:52 ended last night's live cuz someone said
15:54 can you do a zappier walk through and I
15:57 I
15:58 did
15:59 and I warned people I said this is going
16:01 to be horrible and it was horrible so
16:05 maybe hang on maybe there's something on
16:07 hugging face let me go look quick before
16:09 I go make
16:11 a buttthe head of myself oh this is
16:14 pretty
16:15 cool hang on can I zoom in on
16:21 that that's
16:23 interesting all right that's the that's
16:26 the pabs thing I'll go show you that in
16:28 a second
16:29 video input can I play
16:33 this reprompt edit
16:37 upscale huh that's kind of
16:39 cool um let me go to hugging face for a
16:42 second you know what I'm going to do I'm
16:43 going to do this with everyone here
16:45 after a third revision request on an
16:47 image chaty BT Dolly responded hire an
16:52 artist whoa I guess the art community is
16:55 getting getting to the developers over
16:57 at open AI
16:59 they're like listen man just it just at
17:02 some point randomly throw in like could
17:04 you hire someone to do this please it's
17:07 pretty
17:07 [Laughter]
17:09 good oh God man what are we all going to
17:14 do I do not know I do not know I do not
17:19 know so hugging face if you don't know
17:23 it is a place where you can if if you're
17:25 this is primarily for geeks like if
17:28 you're a geek you want to go play with
17:29 models and things like that hugging face
17:31 is where all the models are but when you
17:34 first log into the site there's this
17:35 area called spaces and what spaces are
17:39 is these are um areas that are that are
17:43 that you can play
17:45 with um so so someone is actually
17:48 hosting them or hugging face is actually
17:50 hosting them and they and they kind of
17:51 do here like here's the trending ones so
17:54 let me see if I've got search
17:57 Mol
17:59 mrw 7B is that the is that is the 7B I
18:03 thought it was eight
18:07 something is this Mixr is this the one
18:09 that's the
18:12 thee thing OS error well that's not
18:17 good mistal chat fast 7B instruct see
18:22 this is the problem the the problem with
18:25 the open- source Community it's got
18:27 nothing to do with the open source
18:28 community it's got to do with
18:30 me
18:32 um it is essentially impossible to keep
18:35 up with the commercial application
18:38 development the open- source application
18:40 development and and model development is
18:43 like times 100 like there's so much to
18:47 keep up with so if you don't keep up
18:48 with it like I don't know what all the
18:50 numbers are and what all the versions
18:53 are so mistel ultimate chord and lyrics
18:58 writer
19:02 huh type a message
19:05 um write me a
19:09 song in the
19:12 style of
19:14 Brandy
19:18 Carlile how do you spell her name
19:23 carile um I don't think that's right but
19:26 well we'll go for it
19:29 uh
19:31 about a
19:36 llama loving a
19:44 turtle I met a guy on the beach his name
19:46 was
19:47 Max he was a rugged looking
19:50 dude he had these bright blue eyes and
19:53 he walked with such a strut he owned
19:55 every bit of the sand on that
19:57 Shore
19:59 there was something different about him
20:01 something I couldn't put my finger on
20:03 Max is a llama oh he loves Turtles too
20:07 they spend all their Days by the Sea
20:09 playing just for you I don't think this
20:12 is anywhere near the neighborhood of shi
20:15 vt4 but you know we
20:19 tried that was that was poo poo that was
20:24 [ __ ] [ __ ] [ __ ] [ __ ] [ __ ] so anyway so
20:28 yeah I don't know anyway if you want to
20:30 go geek out go play on Carla oh thank
20:33 you go geek out on uh on hugging hugging
20:38 face and and just start clicking on [ __ ]
20:40 cuz some of the things are cool like
20:41 some of the things are some of the
20:42 drawing you know the text to video text
20:45 to image image to image all that sort of
20:47 stuff um and these are things that you
20:49 can download and install on your own
20:51 computer things like that so anyway all
20:53 right here's the here's the pabb so I
20:57 uploaded this
21:03 video so I grabbed three seconds of a
21:06 just a stock video of dancer
21:09 dude and then that turned
21:12 into that and I said clation vampire so
21:16 now let's
21:17 go
21:19 um van
21:22 go let's see reprompt
21:24 reprompt
21:27 uh let's go van
21:34 go ballet
21:40 dancer all right your video will start
21:43 generating soon um so on Pika there's
21:48 there's a couple of things I want to try
21:49 one is you can you can grab regions of a
21:54 video and modify
21:56 them so let me see modify region all
22:00 right let's
22:02 go it's move it's moving
22:05 weird so let's
22:08 do
22:14 this and I'm going to go um dancer in
22:20 tuxedo let's see if we
22:24 can replace his sweats with a tuxedo
22:29 um hello Kyle mods and chat
22:32 line oh hello Kyle mods and chat line
22:37 tanker Ed Kush what's happening is this
22:39 free right now this is free right now it
22:41 is free baby oh
22:45 yeah the other thing you can do is you
22:47 can expand the canvas which is
22:49 interesting so if I wanted to make this
22:51 a vertical video like say a 9 by6 video
22:57 boom let's go ahead and do
23:01 that um wait what am I doing
23:07 here expand C why is it not
23:13 going oh I think it
23:15 is it is I'm going to trash this
23:18 one can I trash
23:25 it all right so what I said van
23:29 go I said van go
23:36 dancer all right listen if you want to
23:39 do 70s style art films surrealistic art
23:43 films I think pik's Pik is your
23:52 tool do I want a Hendrick and tonic
23:55 right now uh yeah that's a yes you do I
23:59 might go get a little I I might join you
24:02 cuno I might go get myself a little
24:05 vis a little visca you know at some
24:08 point there someone said they were going
24:10 to send me some blantons and I never
24:11 heard from them but I'll go I'll go
24:14 drink non
24:17 blantons you know what I'm going to do
24:20 you know what I'm going to do I am going
24:25 to some people made me some lovely
24:29 be right back
24:33 images let's
24:37 see I I think if I'm getting whiskey I
24:40 shouldn't I shouldn't use that that
24:43 one I think I should save that for champ
24:47 I think we'll go with I think we'll go
24:49 with the the turtle one um I can't use
24:52 Pika it's saying you are on the wait
24:54 list how did you get
24:55 in I did you not know that I'm special
24:59 I'm anointed we have we have a lord in
25:02 here Lord digital Gods Not not only is
25:05 he a
25:07 God now he is a lord digital God I don't
25:10 even know if he's a he he could be he's
25:13 an it they are Lord digital gods and I'm
25:16 now thank you thank you Lord digital
25:18 Gods I'm I'm getting free roses from The
25:22 Lord Oh Lord those Roses are
25:27 lovely
25:30 that looks like Hillary Clinton doing
25:31 Tai
25:34 Chi it does oh that's totally a Hillary
25:37 Clinton pants suit isn't
25:44 it man that's my front yard wait what's
25:47 your front yard this okay nice I like it
25:51 okay so I'll be back I'm going to go
25:54 make myself a whiskey entertain yourself
25:57 when I get back you you're not going to
25:58 believe how the cool [ __ ] we're going to
26:01 do but what Kyle I don't [ __ ] know
26:04 hang on let me see let's go look at some
26:05 other things this is actually pretty
26:07 fast also oh look it did it put him in a
26:10 tuxedo sort
26:13 of it sort of
26:15 worked I mean the fact that it was the
26:18 camera's panning is a
26:19 problem um so they're going to have to
26:21 work on the tracking of of elements and
26:24 then here's the vertical
26:27 video
26:28 it did a pretty good
26:30 job put that big palm tree in the
26:32 background that's pretty
26:35 slick huh all right and then this one
26:38 looks like a fail but let's go look at
26:46 it it's a music video filmed through
26:49 some some laundry or like a like a
26:52 Tibetan tent you know that's that's been
26:54 blown
26:57 apart
27:00 oh goodness so this is p. art if you're
27:03 wondering what site this is and then
27:05 this is me going to make a whiskey all
27:08 right and look at my look at my AI Salon
27:11 t-shirt merch we got merch we need
27:14 nightly live stream minutes taker my
27:17 head's about to explode it's well okay
27:21 so someone in here I forget who it was
27:23 might have been digital Gods I don't
27:24 know who it is someone in here is C has
27:28 somehow done some sort of API hook into
27:31 Tik Tok where they can pull the chat out
27:34 of here live I don't think it's
27:35 transcribing my voice live but it's
27:37 pulling the chat
27:39 out and we could do something with it I
27:41 I assume there's some way we could do
27:43 that right there's there's probably real
27:45 time transcribing [ __ ] out there at this
27:47 point um that I don't know that might be
27:51 an interesting thing all right so here's
27:53 the thing in the spirit of of
27:55 collaboration and exploration we still
27:58 have the AI Salon Discord up I would
28:00 imagine that you could probably build a
28:02 bot on Discord that you could
28:04 stream real time transcription to and do
28:07 some sort of analysis on so if anyone
28:09 wants to geek out go do that I'm going
28:10 to go get a whiskey I will be back champ
28:14 take care of
28:26 them
28:56 watch
29:26 h
29:56 for
30:01 [Music]
30:26 nice
30:30 all right he's back he's back oh thank
30:33 goodness okay listen up I'm in charge
30:35 over here listen Jazz I don't even know
30:38 if you're an irregular these Irregulars
30:39 they will take you out I've seen it
30:42 happen you don't [ __ ] with the
30:43 Irregulars
30:46 cheers it's a little hog back little hog
30:49 back
30:51 bourbon
30:56 bourbon all right let's see what are we
30:58 putting up here that's a good
31:00 one a a learning lab although that's is
31:04 that learning lab that's sort of
31:06 learning lab it's almost
31:09 legible think it will be on the Rocks um
31:12 so here's what you need to know about me
31:14 not only do I do rocks I do like I like
31:16 my bourbon like a slushie like it's got
31:19 so much ice in it that it forms ice on
31:21 the outside of the glass oh and just you
31:24 just and you just it's just a little
31:26 sippy little sippy it's like a sippy cup
31:29 of [ __ ]
31:35 gold yes it will be on the Rocks I don't
31:38 like it neat I don't like it neat e it's
31:41 not pure it's not pure I'm
31:46 trash I am trash p. Art Is Wild it is
31:52 um trying to think what else happened
31:54 today open open AI released some sort of
31:56 prompting guide
31:59 um I assume that's for the new for their
32:01 new multimodal
32:03 thing um what else happened Pika came
32:12 out open AI cut a deal with the
32:15 Publishers of Politico and whatever the
32:18 other thing they do is there's like four
32:21 four properties they're doing something
32:23 with
32:26 um a mod have to have pinned that think
32:29 it will be on the Rocks viewers dropping
32:31 like flies no that's not
32:35 true we're up to 107 I left there there
32:38 was 96 when I left or something like
32:40 that we're we are net positive in the uh
32:43 the whiskey of
32:47 Goo Okay
32:53 um okay let's go let's go take a look at
32:56 some let's go have some fun
32:59 because there's been some really good
33:00 artwork if you want to join
33:03 it so if you go to the AI Salon um when
33:08 you get there you'll see a welcome to
33:09 the salon talks about who we are talks
33:11 about our values all that sort of stuff
33:13 like curiosity generosity they're cool
33:15 values it's cool people but in here
33:19 there are a couple of interesting places
33:22 one is um show and tell and show and
33:25 tell while you're in here while we're
33:26 doing all this you can be somebody
33:30 buried a llama and you're going to Treet
33:32 him out of the out of the uh out of the
33:37 Fortress
33:40 that's oh that's beautiful who did that
33:42 quick take Jake wow look at that damn
33:46 that's
33:47 nice Tranquility pushed to you all wow
33:52 that was done in Korea very cool so as
33:55 people are sharing here they're starting
33:56 to share where they're making it will he
33:58 make
34:01 it I guess that's me me getting my
34:04 whiskey well see you got it wrong that's
34:07 that's neat um but it's about to hit the
34:10 hit the floor Champ's going to lick that
34:12 up that's nice that's
34:14 beautiful
34:16 um wow that's pretty gorgeous Joe mama
34:20 government secrets
34:23 revealed the Llama
34:26 turtle
34:28 there were 88 when you came back oh well
34:30 you know well you know what that just
34:32 goes to prove jenu is that is that when
34:35 my when my supermodel good looks are on
34:38 screen people can't stay away they're
34:40 just like wait who is this it's got is
34:43 that Johnny Depp it's got to be a star
34:45 wait is that Brad Pit what's he done
34:48 with
34:50 himself I've never seen Brad Pit so
34:53 bloated I didn't know he wee glasses I
34:55 thought he was perfect
34:59 that's pretty
35:00 good I need the innovator I KN the
35:04 innovator Lord digital
35:07 Gods hi hereby
35:10 night
35:12 nice Hedgehog properly placed that's a
35:16 good one that is
35:19 solid um so if you want a place where
35:22 you can oh this is from so I think
35:24 Danielle created a uh she created did a
35:27 GPT that will automatically generate
35:29 like snow globes or whatever these
35:30 things are glass globes and people are
35:33 using them she also did some uh oh that
35:35 one's
35:36 beautiful wow who did that Owen Olsen E
35:41 Auto Olsen Auto very cool that's
35:45 gorgeous
35:47 wow very nice very nice
35:51 indeed Scott
35:55 row beautiful
35:58 time
36:01 out oh a little fractal action in a in a
36:04 room full of
36:06 globes very nice nice Vicki what is this
36:10 the turtle dragon fruit is this wait
36:12 yeah this is a dragon fruit
36:14 Turtle oh inspiration is some 3D print
36:17 models that were released today this is
36:19 the dragon fruit oh that's super
36:23 cool wow and then there's the there's
36:27 the actual 3D print how cool is that I'm
36:32 telling you man we're we are entering a
36:34 new realm where where people are going
36:36 to conceive things well people are
36:38 Vicki's doing it right now conceive
36:40 things on these AI tools and then
36:43 generate them in the
36:45 world was going to refine this a bit
36:48 that's
36:52 funny I'm on timeout till 10:45 second
36:56 time
36:57 tonight oh
37:01 man man there's a lot of stuff in here
37:04 anyway so so this is the AI salon so
37:07 what is the AI Salon let me let me show
37:10 you that one's gorgeous damn nice Tim
37:18 shffle so the AI Salon if you want to
37:22 join you can go to AIS salon.
37:26 mn.
37:27 or you can go to I'm going to put it
37:31 back to this the salon. because I I have
37:35 merch on so if you go to the salon. the
37:38 first link is to the mighty networks
37:40 where I was just showing the show and
37:41 tells the other thing the mighty
37:43 networks has is it's got guilds there's
37:45 an Art Guild there's a writing Guild
37:47 there's a video Guild there's a
37:48 healthcare Guild there's all these
37:49 different guilds so there's these sub
37:51 communities that you can join within
37:52 there and you can connect with people of
37:55 you know similar things um
37:57 we meet bi-weekly so also at the salon.
38:01 is our Meetup although Mighty networks
38:04 has calendars in it so you don't really
38:05 need meetups but if you like Meetup you
38:07 can go there and then the third link on
38:09 that page is to the merch store and so
38:12 that's where this shirt came from and so
38:14 uh if you end up getting merch it
38:16 supports the salon there's no cost to
38:18 the salon as an end user like if you
38:19 just want to join a community of people
38:21 up to some [ __ ] in AI then join here and
38:26 and when I say up to some [ __ ] one of
38:28 the things people get really uh
38:31 intimidated by this AI stuff because
38:33 they think oh I've got to be a
38:35 mathematician or an engineer and I I got
38:37 to be super mad
38:39 genius
38:41 um post November 30th 2022 which was the
38:45 launch of chat GPT the rules have
38:48 changed that was true before then that
38:51 date marks an inflection
38:54 point after which
38:57 the rest of us gain access to the power
39:01 of machine learning and Ai and that's
39:03 the phase we're in now and we're right
39:05 at the very beginning we're a year into
39:06 this and it's um even though Ai and
39:10 machine learning has been around for
39:11 decades which people are are want to
39:14 point out in here well you know this has
39:16 been around since the the late 90s I
39:19 know
39:21 and if you didn't have a degree from MIT
39:23 or Stamford you were at a disadvantage
39:25 and that's not true anymore
39:27 and so what this group is is a group of
39:29 people just trying to figure this [ __ ]
39:30 out and there are people in here that
39:33 are brand new there are PE there are
39:34 people in there that have been working
39:36 in this stuff for 30 years and everyone
39:38 in between the only price of oh that's
39:41 good that's good cuno so what what Cuna
39:43 said is is is true so so here's what I'm
39:45 going to tell you it's free except for
39:47 this there's three things I want you to
39:50 do if you go to the salon the first is I
39:54 want you to go to welcome to the salon
39:57 and I want you to read about what we're
39:59 about and I want you to check out our
40:00 values make sure you resonate with the
40:02 values cuz cuz this it's a remarkable
40:05 place it's a remarkable community and
40:07 it's growing and I want to keep it
40:09 remarkable and and how it stays
40:11 remarkable is by you know having people
40:14 that that share these values so that's
40:17 request number one or cost number one if
40:20 you're going to join the salon check out
40:22 who what we're
40:24 about
40:25 um
40:27 cost number two what was cost number two
40:31 I know I had
40:32 three uh oh yeah cost number two
40:36 introduce yourself so there's a channel
40:39 that's called introduce yourself and so
40:42 Julia she's a Socratic gadfly customer
40:44 experience executive super cool she even
40:47 put in a a dueling turtle image here so
40:51 so be generous introduce introduce
40:53 yourself she says AI has become my new
40:55 crack don't judge me perfect perfect let
40:57 us know who you are let us know where
40:58 you are in this journey whether you're
41:00 just starting out whether you're totally
41:01 into it whether you're cynical and bored
41:03 with it so that's the second one and
41:05 then the third cost is go to show and
41:07 tell and make an image with a turtle in
41:10 it because why not or you can do you can
41:13 substitute Turtles with llamas hedgehogs
41:16 Gummy Bears or blue rubber
41:22 duckies all right so that's the yeah
41:25 salon
41:28 um all right what do we want to do now
41:30 what do we want to do on this FR night F
41:36 night hey Jen what did you do Friday
41:39 night I don't want to talk about it
41:42 Sally why not
41:44 Jen did you go on a hot Dee Jen no Sally
41:49 no Sally I didn't what did you do I sat
41:54 home and I watched an old guy on Tik
41:56 tock
41:58 Jen why would you do that Sally I don't
42:01 want to talk about it I told you I
42:02 didn't want to talk about
42:05 it these are the uh these are the
42:08 cousins of the the wild and crazy guy
42:10 Steve Martin from Saturday Night
42:15 Live oh
42:20 man 3D merch nice I like it nice Joe
42:25 mama I dig it
42:34 more members are making images in the
42:36 past few weeks yeah I've noticed that I
42:38 it's it's really cool there's it's
42:40 become quite quite a a hub of activity
42:43 and I think what we're going to what
42:44 we're going to start to see as tools
42:46 like Pika where's Pika as tools like
42:50 Pika you know people start exploring you
42:53 know Pika and Runway and things like
42:55 that and these things become come um
42:59 accessible and there's more of them and
43:01 the cost come comes down I think we're
43:03 going to start to see a shift to you
43:06 know sharing all sorts of different
43:07 things gpts and videos and you know text
43:12 outputs and movies and music and it's
43:15 it's going to get it's going to get
43:16 pretty exciting so I think that's super
43:18 cool all right so if you have questions
43:21 about II pop them in the comments below
43:23 oh Kyle did you get a chance to look
43:25 over the imagin two comparison article I
43:28 made I did I saw it
43:31 briefly and I I looked at um so where
43:34 did you put it Pate did you put it you
43:36 put it in in uh in mechanics Guild or
43:39 did you put it in somewhere else I think
43:41 you put it in mechanics Guild so so Pate
43:45 Pate runs the the mechanics Guild um the
43:49 AI mechanics Guild in the AI Salon U and
43:52 he did a comparison of Google's new
43:55 imagine two engine Show and Tell I think
43:59 yeah let me let me go see if I can find
44:04 [Music]
44:07 it
44:15 [Music]
44:19 P I do not see it let me go back out to
44:22 the main
44:24 level uh uh uh uh uh
44:35 uh find a
44:37 guild which one is it is
44:40 it that
44:43 one
44:45 no I don't know where it is pate and I
44:48 don't quite know how to search this
44:49 thing well enough maybe it was oh is it
44:51 AI tools and resources add your favorite
44:55 tool
45:06 I don't know anyway it was good
45:08 somewhere in here somewhere in the salon
45:11 uh PT did a a comparison of of uh
45:15 imagine 2 and dolly dolly 3 um which is
45:21 a good comparison um so yeah um that was
45:24 super cool uh Jim sharp wrote I can't
45:27 believe this tool is free talking about
45:29 Pika I think I should add p. art to the
45:33 salon tools and pics um here's a thing
45:36 that I'll go over with you so again if
45:38 you're new to
45:40 um to AI one of the other things that is
45:46 uh is super I think Super valuable
45:49 within the salon is um there's a tools
45:52 and resources section where you can add
45:54 your favorite tool so there's kind of
45:55 this running list list of people adding
45:57 tools so again if you found things you
45:59 think are cool go add them there but
46:01 there's also the salon pick so this is a
46:04 list that I've been sharing um on on
46:08 these lives for the past six months or
46:10 so and I keep I keep expanding it and
46:12 changing it and and you know curating it
46:15 but there's you know chat gbt is still
46:17 the best like if you're just starting
46:19 start with chat
46:21 GPT it's still the best Just Go With It
46:25 don't feel like you have to do anything
46:28 fancier than that um then here's kind of
46:32 the top four large language models
46:34 sitting underneath chat GPT Claude X
46:37 Bard and co-pilot from Microsoft um by
46:41 the way if you didn't know it co-pilot
46:43 from Microsoft this is free chat gp4 so
46:48 GP chat gp4 from open AI was restricted
46:52 for a while you couldn't you couldn't
46:53 get it it's unrestricted now so that's
46:56 good if you can afford the 20 bucks a
46:58 month I think chat GPT plus is the best
47:01 deal in software right now I would
47:03 absolutely recommend you do it if you
47:05 can afford that Microsoft that's free
47:07 free chat
47:09 gp4 um why I think it's worth paying the
47:11 20 bucks to open AI it's just better um
47:14 Microsoft kind of screwed this up a
47:16 little bit but whatever it's it's still
47:18 good and it's free
47:20 um the next four kind of specialty large
47:23 language models Pi which is really good
47:25 in conversation perplexity which is good
47:27 for research llu which is a whole
47:31 [ __ ] world of Open Source large
47:33 language models uh llama 2 is the one
47:36 from meta but there's a whole misil all
47:38 all these things are out there now that
47:40 are getting really really good and then
47:42 you got po which is interesting you can
47:44 play with lots of different models there
47:47 then I put a bunch of image creation
47:49 things in here um Mid Journey Korea I
47:52 just added last night um Leonardo
47:55 Firefly from Adobe that's we're now it's
47:58 actually now Firefly 2 oops oh I'm not
48:01 editing right now it's Firefly 2 uh then
48:05 here's video creation and I don't have
48:07 oh I do have Pik in here yeah there's P
48:09 so P's in
48:10 here um dids in there um hen's in there
48:14 hey some of the haen stuff is
48:17 insane um want to see some insane haen
48:22 stuff just it's just every time I show
48:25 some of this stuff let me do I have yeah
48:28 this is here here we'll
48:32 show is it this one okay so this is this
48:36 is something that I'm putting together
48:38 this a product I'm putting together for
48:40 my company Story vine so here's a an ad
48:44 for some of the new products that we're
48:45 working on so authentic translation so
48:48 I've got the um Source video was
48:52 recorded in English my greatest source
48:55 of supp for living with rheumato
48:57 arthritis has been um speaking with
49:00 other people about what they've gone
49:03 through uh reading a lot about the
49:05 effects of rheumatoid arthritis so we
49:08 don't do a lot of video in a healthcare
49:11 and Farmers space sorry I got ice in my
49:13 mouth um okay so here's Ellen speaking
49:25 Spanish
49:34 so we can we can take all of her English
49:36 Clips the the way Story vine works we
49:38 capture things in this modular way where
49:40 she records a series of prompts and they
49:43 go into our system and get assembled
49:45 together into a final video we can sort
49:47 of deconstruct them translate them into
49:49 Spanish and then we can also add in
49:51 translations for all the lower thirds
49:53 and things like that and then we can you
49:56 can put it into a system which will
49:57 automatically create a Spanish
49:59 transcription Spanish title description
50:02 chapters things like that and then
50:04 here's the final video of her speaking
50:07 [Music]
50:18 Spanish Kevin Vasquez says the Spanish
50:20 is on
50:24 point
50:32 okay so that's that now we can do this
50:35 in 27 different languages so here's
50:38 Hindi anyone on the call
50:46 [Music]
50:49 speak here's H
50:52 [Music]
50:54 Japanese
51:01 so that's hey Jen so hey Jen's [ __ ]
51:05 cuckoo it's
51:07 cuckoo um then you got music and audio
51:10 you got 11 Labs so the the translation
51:13 and the um voice that that hey Jen uses
51:17 is actually from 11 Labs so if you want
51:20 to do audio synthesis of yourself um
51:22 they've got a crazy new model called
51:24 voice to voice so let me show you
51:26 another thing that we did here
51:29 um let me jump down to is it this one
51:34 yeah so this is personas where I can now
51:39 just basically generate a
51:41 Persona so a fake person with a fake
51:44 persona and then have that person with
51:47 with whatever the Persona is answer a
51:49 set of Story vine prompts and then um
51:54 generate an image of them and then bring
51:56 that image to life hi I'm George Miller
51:59 my life's always been about Simple
52:00 Pleasures cooking my Catch of the Day
52:03 and enjoying the quiet of the country
52:05 there's nothing like the piece of
52:07 fishing at dawn or the aroma of a
52:09 home-cooked meal but when I was
52:11 diagnosed diabetes everything changed
52:14 suddenly the foods I loved Meed
52:16 rethinking and my days had to be planned
52:18 with care it was a challenge but I've
52:21 learned to adapt finding new ways to
52:23 enjoy my passions while managing my
52:25 health
52:26 life's different now but no less Rich so
52:31 that that voice I did at 11
52:35 labs and it's a thing called Speech to
52:38 speech where I recorded that audio and
52:41 then swapped out the voice they they had
52:43 a voice there called Santa and I thought
52:45 let me see if Santa works for old George
52:47 Miller here and so so I picked that but
52:51 but like the Cadence and the inflection
52:54 is me but the the quality of the voice
52:57 is substituted and so here's here's
53:00 George's
53:02 doctor um that I also recorded this
53:04 audio and then swapped it with this
53:06 female
53:08 voice let's see
53:11 oops hello I'm Dr Emily Park as an
53:14 endocrinologist specializing in diabetes
53:17 my Approach goes beyond traditional
53:19 treatment I'm deeply committed to
53:21 empowering my patients integrating
53:23 Medical Care with Lifestyle Changes
53:26 so yeah so we're in one one 101 just
53:30 said I'm scared so I don't know that we
53:33 need to be scared I think we need to be
53:35 educated like part of what this this
53:37 Channel's about is like understand what
53:40 these tools are understand where the
53:43 boundaries are what are they good at
53:44 what are they not good at like like
53:47 the why am I doing these as sketches and
53:51 and watercolor paintings congratulations
53:54 thank you cuno are are you
53:56 congratulating me on my whiskey because
53:57 that's the thing I'm really excited
53:59 about no no I'm actually I'm really
54:01 excited about these these these new
54:02 products um why am I choosing to do
54:05 these as drawings because when you do
54:08 them as photographs like like watch
54:09 George's head here George's he's it's a
54:12 little weird hi I'm George Miller my
54:15 life's always been about Simple
54:17 Pleasures cooking my Catch of the Day
54:19 and enjoy like nothing there's quite
54:21 moving except his head and it's just
54:23 sort of weird and janky and then
54:26 I'm Dr Emily Park as an endocrinologist
54:29 specializing in diabetes my Approach
54:31 goes beyond traditional treat so I'm
54:33 doing them as a stylized thing to make
54:35 it clear like this is not a real person
54:38 right but you know I think everyone's
54:40 going to need to figure out where their
54:42 own ethical boundaries are and and
54:44 boundaries like where where does that
54:46 actually make sense one of the one of
54:48 the the reason that I'm doing this is
54:51 we've got clients right now that have
54:53 hired us to create temp templates for
54:56 them to allow their patients to tell
54:59 their stories and we're like a year two
55:02 years into a relationship with a client
55:04 and they haven't signed a single
55:07 contract with a patient so they have
55:09 zero videos to show right so what if we
55:12 could do do these personas and bring
55:15 their stories to life while they're
55:17 waiting to get real people um to tell
55:19 their stories or here let me show you
55:22 another product
55:24 here what if they just have an audio
55:26 recording um of Ellen they don't have
55:29 video of her well we could take a
55:31 photograph of Ellen stylize that and
55:33 turn that into my greatest source of
55:35 support living with rheumatoid arthritis
55:37 has been um speaking with other people
55:41 right so that's actually her voice you
55:43 know from from one of the videos but I'm
55:45 just taking the audio of it and turning
55:46 it into that and then here's one where
55:48 we synthesized her voice on 11 labs and
55:52 um this is like in the case where a
55:55 client might might just have a written
55:56 story and they want to bring the written
55:57 story to life hi I'm Ellen despite
56:00 living with rheumatoid arthritis I've
56:02 discovered a passion for water skiing
56:03 and interpretive dance which bring joy
56:06 and a sense of freedom to my life and
56:09 the interpretive dance I put that in
56:11 there I wanted something silly but
56:13 something that I could put in front of a
56:14 Pharma client that they wouldn't they
56:16 wouldn't be like well that's
56:18 inappropriate so so water skiing an
56:21 interpretive dance from the the uh the
56:25 woman with rheumatoid arthritis I
56:27 thought that was pretty
56:29 good whatever that's that is my
56:33 idiocy um but but this is the world
56:35 we're in right now is is there's just
56:37 some crazy [ __ ] and that's what this
56:39 that's what this list is about um is
56:42 these are things you should play with
56:43 here's people you should follow Ethan
56:45 mullik David Shapiro Paul rater and
56:47 Rachel Woods um and I have links to all
56:49 them so anyway all this stuff's on the
56:50 AI salon so you should go join the a
56:52 salon the AI salon and that's how you
56:56 pronounce it it's French
56:58 [ __ ]
57:00 AI
57:06 the best travel app or translation for
57:09 English to Spanish at the moment I quite
57:13 frankly for me it's CAD GPT I mean here
57:19 look um if you want to do produced media
57:23 11 labs and hay Jen is probably the way
57:25 to go but I'm just on my old iPad my
57:30 fancy iPad I should probably put on a
57:32 better background shouldn't I I'm a dumb
57:35 dumb why do I not have a cool background
57:37 in there anyway uh I'm going to go to
57:39 the chat GPT
57:45 app and I'm in chat GPT 4 but this works
57:47 in 3.5 as well that little set of
57:50 headphones in the bottom corner
57:54 there
57:56 I'd like you to act as a Spanish
57:57 translator for me I'll speak English and
57:59 then I just want you to repeat it in
58:01 Spanish without speaking the English
58:03 part
58:07 okay absolutely I'm ready to help with
58:10 translation just say what you need in
58:12 English and I'll translate it into
58:15 Spanish for you go ahead when you're
58:18 ready oh boy this is embarrassing I need
58:21 to find a bathroom and I need to find at
58:24 pronto
58:25 [Laughter]
58:44 no no no no seriously it's an emergency
58:48 it's an emergency I must go
58:53 now no no no in ser
58:59 so and you can do that in pretty much
59:01 any
59:03 language espiranto it's got it I use
59:07 chat GPT wait what was that someone just
59:09 pinned something and I just I just lost
59:10 it as I was trying to read it uh thanks
59:13 I love Pi I'm totally comfortable with
59:15 Google very nice
59:17 awesome congratulations knck and Luts I
59:20 use chat GPT in real time at an airport
59:22 for translation yeah I I
59:26 mean folks like folks I just called you
59:30 folks listen
59:34 folks um the level of
59:39 accessibility
59:41 for I just I I feel like accessibility
59:44 as a concept is is getting like a
59:48 significant level up in the next two
59:51 years I mean it's already gotten like
59:53 the fact that I've got my my sons are
59:55 going to Japan next summer and they're
59:58 going to be able to just on their phone
1:00:00 just speak into it in English and have
1:00:02 it talk now are there translator apps
1:00:04 for that
1:00:05 sure but you don't need it now we're
1:00:07 going to we we are essentially going to
1:00:09 have this the everything app that does
1:00:14 everything the other nice thing about
1:00:16 doing those the the um when you put it
1:00:19 in conversational mode is it gives you
1:00:23 the transcript of everything that you
1:00:25 just talked about and everything it
1:00:26 talked about
1:00:28 so um so if you needed to take obviously
1:00:32 not the bathroom one but if you were
1:00:33 having a good conversation with someone
1:00:35 and they spoke to you in Spanish and it
1:00:37 translated it back in English you would
1:00:39 have a transcript of
1:00:42 that um what site is this what site was
1:00:46 I on the site with all the the list of
1:00:48 directories that was the AI salon so how
1:00:49 you get to the AI Salon is you go
1:00:54 to
1:00:56 [Music]
1:01:04 what was the Max Smart it was like
1:01:06 Mission Impossible but it was Max
1:01:09 Smart what was that show that theme is
1:01:12 in my head right now um the site I was
1:01:14 looking at that had all those sites on
1:01:15 it was the AI salon and it was the tools
1:01:17 and resources section so go join the
1:01:20 salon uncanny valley yeah this is
1:01:22 totally uncanny valley stuff this is now
1:01:26 soon we will have a level of
1:01:27 augmentation of real life and merged
1:01:29 with robotics totally yeah I agree with
1:01:31 that I know I know one 101 listen listen
1:01:35 this channel exists because I've been
1:01:39 like this for a [ __ ]
1:01:42 year it happens it like I would say I'm
1:01:46 doing these things seven nights a week
1:01:48 and I mean the the regulars know this I
1:01:51 would say two or three times a week
1:01:56 like I I either stumble on a tech but
1:01:58 but more often than not it's not even a
1:02:00 tech that blows my mind it's just a use
1:02:03 case of an existing Tech that I've
1:02:04 demoed a hundred times and someone says
1:02:08 well can't it do this or can't it do
1:02:09 that and I'm like I don't know let's go
1:02:11 try and then I try it and I'm like holy
1:02:13 [ __ ] like the the implications of what
1:02:16 these things make possible like those
1:02:18 demos that I just showed of storyvine
1:02:20 where we're taking our decade plus
1:02:23 experience in authentic narrative and
1:02:27 we're saying hey how can we tell
1:02:28 authentic stories using these new tools
1:02:31 even though they're sort of uncanny
1:02:33 valley and
1:02:34 weird how can we push the
1:02:37 state-of-the-art as it is today and do
1:02:39 something interesting with it I don't
1:02:41 know if you all remember this um if if
1:02:43 you're there's a lot of there's a lot of
1:02:45 gen xers on this I don't know if you
1:02:47 remember when the game missed came out M
1:02:50 Myst it was kind of a a puzzle game but
1:02:54 it was very immersive it had a lot of
1:02:56 sound in it had beautifully rendered
1:02:57 images and it had these little moments
1:03:00 of video like these little animations of
1:03:02 like a book opening in a in a bigger
1:03:04 scene there'd be like a pedestal with a
1:03:06 book on it and it would Shimmer and the
1:03:08 pages would open
1:03:10 and
1:03:12 and
1:03:14 what what I thought was brilliant about
1:03:16 that book actually we interviewed Randon
1:03:18 Robin Miller in urban desires my my
1:03:20 digital magazine that I started back in
1:03:22 the '90s we interviewed them cuz I was
1:03:25 totally smitten with that game and and
1:03:29 why because what they did in that game
1:03:32 at at the time in the
1:03:34 mid90s a lot of video games were trying
1:03:37 to push the envelope on you know
1:03:39 immersive 3D realtime rendering which
1:03:43 back then the technology really sucked
1:03:46 but they were all trying to push the
1:03:47 envelope with what you could do with
1:03:48 real-time rendering and what Rand and
1:03:51 Robin Miller did is they said well if
1:03:53 QuickTime can only play in a window this
1:03:56 big and then we can render a a bigger
1:03:59 picture what they did was they said well
1:04:01 why don't we just render high quality
1:04:03 images and we'll we'll sort of stitch
1:04:06 them together as as like single frames
1:04:09 and then we'll use sound design to make
1:04:11 it seem like things are alive so it was
1:04:13 just a still image and then they would
1:04:15 occasionally use a little QuickTime
1:04:17 video nicely integrated into the thing
1:04:21 and it was just it was like this amazing
1:04:24 experience
1:04:25 and and and it was just they didn't they
1:04:28 they took where the technology
1:04:30 was of the day and they said how can we
1:04:34 do something excellent within this and I
1:04:37 think that that's you know get smart
1:04:39 that's what it was
1:04:44 don't get
1:04:47 smart that was a great
1:04:51 show oh
1:04:53 man crazy
1:04:56 so yeah so we're in we're in some cuckoo
1:04:59 times get smart Mad Max no it wasn't Mad
1:05:02 Max Max smell smart on get smart um look
1:05:06 at all us older peeps
1:05:09 seriously all us older peeps I don't
1:05:11 know do you know what's going on it's
1:05:15 and you know listen if if this channel
1:05:17 starts the revolution that's fine by me
1:05:19 this is like the Revenge of the Gen
1:05:21 xers I'm regularly Hearing in this
1:05:24 channel
1:05:25 people that are our age or in the
1:05:29 hood that have more of a clue about AI
1:05:32 than their kids
1:05:33 do and I think I there's something
1:05:36 there's something really fascinating
1:05:38 going on with this stuff right now why
1:05:41 why are older people more into this
1:05:46 and and way more educated about it than
1:05:49 young people right
1:05:50 now it's fascinating it's I'm living in
1:05:53 a house full of I got two 23-year-old
1:05:56 twins right up
1:05:57 there could give two shits about this I
1:06:03 try two three times a week hey you
1:06:06 should really check out this AI stuff
1:06:08 day
1:06:11 whatever I don't feel good
1:06:16 okay this this could be something
1:06:19 remarkable for you
1:06:22 no that's not how they talk they talk
1:06:24 more like this oh Dad don't be such an
1:06:29 [ __ ] I didn't want to mischaracterize
1:06:32 them Boston version Get Wicked
1:06:37 smart that's
1:06:39 awesome awesome awesome awesome get
1:06:41 smart aging 86 yes that was that that
1:06:44 was a I love that show that that was one
1:06:46 of my Gilligan's Island and get smart
1:06:49 were were couple of my biggies those are
1:06:52 some of the formative cultural uh
1:06:54 imprints in my
1:06:57 life it says a
1:06:58 [Laughter]
1:07:03 lot loved Mist I know wasn't Mist
1:07:06 amazing mist is the first game like as
1:07:08 an Adder Mist was the first game I think
1:07:12 it was the first game that I ever played
1:07:13 beginning to end like I solved Mist
1:07:16 because it was just it was like spooky
1:07:18 and creepy we would we had a we had a a
1:07:21 weird hallway we had a a brown Stone in
1:07:25 Brooklyn like a basement apartment and
1:07:27 we had this long weird hallway that you
1:07:29 would enter so that was where the
1:07:30 computer was and so we turn out the
1:07:32 lights and we just sit in there with
1:07:34 like a 13-inch you know monitor with
1:07:37 mist on it and just we turn the sound
1:07:40 all the way up and it was just all this
1:07:42 weird awesome stuff so if you haven't
1:07:45 played with mist I'm pretty sure you can
1:07:46 just play it online now it's it's a it's
1:07:51 it still holds up because they focus
1:07:54 primar where they focused first was on
1:07:56 the storytelling and then on the
1:07:58 production values and and like lastly on
1:08:01 the technology they didn't try to push
1:08:02 the technology they just did the techn
1:08:06 they said what's the best we can do with
1:08:08 the current state of Technology thought
1:08:09 it was [ __ ] brilliant brilliant
1:08:11 brilliant brilliant brilliant brilliant
1:08:12 brilliant love that game I miss
1:08:14 missed Miss was an early Rabbit Hole
1:08:16 Adventure absolutely never had any Tech
1:08:19 that could handle Mist Ted looks at life
1:08:22 you gota you gotta up your game M it
1:08:25 just needed a CD ROM
1:08:28 buddy so many friends of mine played
1:08:30 Mist they loved it I had a crush on
1:08:33 agent 99 I know I know she was a catch
1:08:36 agent 99
1:08:38 right I'm there with you she was
1:08:43 hot oh
1:08:47 Max we need Gary ve
1:08:50 here God we're just all cussing away
1:08:54 cussing away telling people to have
1:08:56 their kids drop out of college and jump
1:08:58 into
1:09:02 ai' be
1:09:04 fun Google music you can hum and it will
1:09:07 find media for you no I think if you hum
1:09:10 it doesn't find media for you it
1:09:13 generates a song based on your hum it'll
1:09:16 you can say take turn this hum into a
1:09:18 sack solo and it and it'll do
1:09:21 it Millennials had missed too Grant I
1:09:24 was too young to solve the puzzles very
1:09:26 frustrated young and yeah um yeah I know
1:09:31 it for me like puzzle games puzzle games
1:09:34 in me are not good like the add doesn't
1:09:37 I got really frustrated with m a lot of
1:09:40 the puzzles cuz I'm like there doesn't
1:09:42 it doesn't need to be this [ __ ]
1:09:43 stupid like I want immediate
1:09:45 gratification and so but but it was good
1:09:48 enough that I I worked through that [ __ ]
1:09:49 it was very
1:09:51 gratifying um because we saw how the
1:09:54 world changed with the internet yeah I
1:09:55 think that's right thankful Jones I I
1:09:57 think that I think that gen xers right
1:09:59 now are in a in a unique position for
1:10:01 for two reasons one is we've seen it
1:10:04 before right we we watch the transition
1:10:08 from not internet to internet
1:10:12 right um and it was it was insane it was
1:10:17 pretty crazy and so I think we know
1:10:20 what's coming so that's like first thing
1:10:22 we have is we've seen it before
1:10:24 second thing we have is because we've
1:10:27 been on a planet for a while now we can
1:10:29 look at these AI tools and really
1:10:31 quickly assess that's good that's bad
1:10:35 and it's and it's not just be it's not
1:10:37 any particular skill it's just we've
1:10:40 been around for a while right so if
1:10:43 we've been paying attention to the world
1:10:45 for 30 or 40 years we can look at what
1:10:47 these things do and understand oh that's
1:10:51 a big deal oh that's really cra happy
1:10:54 but when that gets good holy [ __ ] I know
1:10:56 what that's going to be used for like
1:10:58 that kind of stuff so yeah I think we're
1:11:00 in a I think we're in a a really unique
1:11:02 position as a bunch of old
1:11:06 farts what'd you do Friday night I
1:11:08 watched some old guy try to use ticktock
1:11:10 it was it was awkward it I I'll admit it
1:11:13 it was awkward it was it was not
1:11:16 pleasant he was a funny
1:11:19 fell kind of looked like a a chubby
1:11:22 chubby Alec Baldwin or
1:11:24 Tom Arnold he you know he kind of
1:11:26 resembled Rosie
1:11:27 O'Donnell it was it was it was it was I
1:11:31 felt
1:11:32 uncomfortable how long did you watch it
1:11:34 three hours three hours yeah it was
1:11:36 solid yeah I don't know I don't know
1:11:38 what we're doing here
1:11:40 people all right oh thankful Jones okay
1:11:42 this is good oh yes Bewitched get smart
1:11:46 laugh in Beverly Hillbillies holy
1:11:49 [ __ ] [ __ ] Beverly Hillbillies for me
1:11:50 th that was those were my heroes [ __ ]
1:11:54 loved that show same with gilliam's
1:11:56 Island it's like
1:12:00 unbelievable it's like I go back and I
1:12:02 look at them now and I'm
1:12:06 like they're bad right but but they
1:12:09 weren't they were [ __ ] awesome was a
1:12:11 [ __ ] generation that's how he that's
1:12:13 how he got educated Robert Rossy thank
1:12:14 you so much sir very
1:12:16 nice oh Laugh In The Midnight Special
1:12:21 amazing Hogan's Heroes oh my God
1:12:24 colel
1:12:28 clink
1:12:33 Hogan lost in Sp holy [ __ ] you know
1:12:36 what's amazing okay so for those if
1:12:38 there's any youngans if there's any
1:12:40 Young on the channel
1:12:43 um the reason we all love the same shows
1:12:48 is there were only like 20 shows like
1:12:52 there were three channels and and you'd
1:12:54 have like you know a show a week you
1:12:58 know and there'd be like four of them
1:13:01 it'd be stari and Hutch and then there
1:13:03 Adam
1:13:06 12 like there weren't that many um Kyle
1:13:10 look at show tell
1:13:14 please oh
1:13:20 man all right oh that's beautiful
1:13:26 awesome oh my
1:13:34 God nice steo is doing these Pirates of
1:13:38 the Caribbean things that are beautiful
1:13:40 that's
1:13:41 gorgeous oh nice Maxwell smarts car
1:13:46 Sunbeam Tiger oh that's not that's not
1:13:49 AI That's is it oh no it is that's meta
1:13:54 meta's imag meta's new thing um that's
1:13:59 gorgeous quick take Jake
1:14:05 beautiful man some of you guys are just
1:14:08 killing it with these
1:14:09 images it's so cool to see where people
1:14:12 go with the tech right like you can take
1:14:14 it
1:14:15 anywhere
1:14:17 Korea abstract paint integrated with
1:14:21 woman oh spin art the spin art dress
1:14:26 there's another classic 70s '70s look
1:14:29 look she got some spin art paint in her
1:14:34 hair that's
1:14:38 awesome oh man here's your iPad backdrop
1:14:42 okay I'll I'll actually I'll actually
1:14:43 use that that's
1:14:45 great I like it I like it it's solid
1:14:49 okay good Korea wheat and Farmhouse wow
1:14:51 that's
1:14:52 gorgeous
1:14:55 Kelly Camp's getting into into Korea if
1:14:58 you haven't played with Korea Korea is
1:14:59 another one holy [ __ ] let's go play with
1:15:01 Korea a little bit like this is another
1:15:03 one where it's just like I don't I don't
1:15:05 even know how to use it yet because it's
1:15:07 it's such a completely different mindset
1:15:10 from having to prompt an
1:15:12 image so if you go to kore.ai they have
1:15:15 this thing called realtime image
1:15:19 generation and then you get you know you
1:15:21 got these shapes and
1:15:25 it real
1:15:28 time oh that's
1:15:33 cute The Prompt is pink frog on top of a
1:15:36 blue mushroom well but now now it's now
1:15:39 it's a blue frog um but let's see let's
1:15:42 we'll roll the dice here it's always a
1:15:44 dangerous
1:15:47 thing oh that's cool ice Citadel Frost
1:15:51 images and snow beasts and then you know
1:15:54 the evil witch has put people in
1:15:56 their in
1:15:58 their the spell of
1:16:01 iceing
1:16:03 whatever you can also do let me get rid
1:16:07 of this get rid of this get rid let me
1:16:09 go
1:16:12 um Nick
1:16:15 Cage
1:16:17 in a
1:16:21 1970 Grand Torino
1:16:27 and then I
1:16:30 can I can put me in
1:16:35 there come on see me see me for who I
1:16:40 am turn me into Nick Cage baby oh
1:16:44 yeah I don't know why it's not doing the
1:16:47 hand thing oh there it
1:16:48 went
1:16:51 yeah get that get that hand going going
1:16:53 oh
1:16:56 yeah so yeah so Korea's a whole [ __ ]
1:17:01 trip
1:17:03 um [ __ ]
1:17:05 nuts I should probably get rid of the
1:17:08 camera
1:17:12 there I was pretty happy with the wheat
1:17:14 and Farmhouse picture yeah it's pretty
1:17:15 good it's pretty good I'm totally
1:17:17 figuring wait I'm totally figuring out
1:17:22 cuz yeah I don't know how to use it
1:17:25 either
1:17:27 I but but you know what's going to
1:17:29 happen here's what's going to
1:17:31 happen some people are going to get
1:17:33 really good at Mid Journey right like
1:17:35 whatever mid Journey becomes and evolves
1:17:37 into they're going to get good at like
1:17:39 specific image prompting other people
1:17:41 are going to get really good at Dolly
1:17:43 sort of creating images in the context
1:17:45 of a conversation other people are going
1:17:47 to get really good at Korea generating
1:17:49 things like that other people are going
1:17:51 to use Korea for performance art this
1:17:54 guy Todd and I that had lunch today we
1:17:57 talking about that that you know IM
1:17:59 imagine something like I just did with
1:18:01 Nicholas Cage and you know the hands
1:18:03 like imagine that being a stage show
1:18:06 where you're taking the actors on stage
1:18:08 and you're real time rendering them into
1:18:10 some image up there like some people are
1:18:12 going to figure this [ __ ] out and
1:18:14 they're going to get really good at it
1:18:16 and it's going to create entirely new
1:18:17 art forms and entirely
1:18:20 new products entirely new
1:18:26 jobs like what do you do I'm a I'm a
1:18:30 real time image synthesis actor
1:18:35 what what's
1:18:38 that well I did improv on the weekends
1:18:42 and I worked at an insurance company
1:18:43 during the week and then I got pretty
1:18:46 good at this Crea [ __ ] and then I
1:18:48 figured out how I could take one of
1:18:50 those images and project it on a screen
1:18:52 and so I decided to do a show and now
1:18:55 we're sold out through
1:18:59 2027 why
1:19:01 not why not why not there's no reason
1:19:06 why not well there is a reason why
1:19:10 not anyone who sort of sits cross-armed
1:19:13 right now it's like AI is evil AI is
1:19:15 stupid AI is not
1:19:17 perfect AI is going to take jobs AI
1:19:19 steals from
1:19:22 people
1:19:24 are going to be at such a
1:19:27 disadvantage and anyone who like shows
1:19:29 up at a stupid [ __ ] Tick Tock Channel
1:19:32 like this at [ __ ] midnight on the
1:19:36 East Coast on a Friday night to geek out
1:19:39 on
1:19:40 AI is going to be at a significant
1:19:43 Advantage they're going to like the
1:19:44 opportunities are going to be insane
1:19:46 what are the opportuni is going to be I
1:19:48 don't [ __ ] know it's not going to be
1:19:50 these
1:19:51 people it's not this
1:19:54 it's not like oh make a 100,000 t-shirt
1:19:56 images and put them up on Etsy no it's
1:19:58 not
1:20:01 that the the first person that like
1:20:03 everyone's going to do that and then
1:20:05 like no it's not
1:20:08 that it's someone gets curious about
1:20:12 these tools they realize oh there's a
1:20:14 problem in the world that needs
1:20:17 solving or here's an idea I have that I
1:20:19 couldn't do before that I can do now and
1:20:22 I've always want wanted to be an improv
1:20:24 actor and I'm not a very good improv
1:20:27 actor but I'm good enough to stand in
1:20:30 front of a webcam run through Korea and
1:20:33 have my buddy Fred who's good at
1:20:36 prompting prompt [ __ ] while I do stupid
1:20:39 [ __ ] on stage and it shows up on the
1:20:42 screen behind me and people
1:20:45 clap I'm living the
1:20:51 dream you not not going to you're not
1:20:53 going to discover that magic on the
1:20:59 sidelines you're
1:21:02 not like I'm all nervous now to to take
1:21:05 a sip of this cuz you know you've been
1:21:08 permanently
1:21:09 [Laughter]
1:21:14 banned can I do
1:21:17 that all right fantastic the channels
1:21:21 went off at midnight yeah oh that's
1:21:23 right our TV channels went off at
1:21:25 midnight you couldn't binge watch [ __ ]
1:21:27 till 3:00 in the morning when when we
1:21:29 were around you actually had to go to
1:21:36 sleep The Jeffersons that was a good one
1:21:40 sford and
1:21:42 Sons All in the Family although All in
1:21:44 the Family was a bit more for the older
1:21:46 the older folks
1:21:50 but but it was good CU it was so mean
1:21:53 that taught All in the Family taught me
1:21:55 mean humor that mean can be
1:21:59 funny I think that's why gen xers are so
1:22:02 [ __ ] salty is we all we were all
1:22:04 trained by Archie Bunker I
1:22:08 just mean can be
1:22:13 funny oh my
1:22:16 God I'm totally okay yeah let's see
1:22:22 U
1:22:28 no way any of this will matter in a year
1:22:31 nope it won't but but here's the deal
1:22:38 one-on-one here's here's what
1:22:41 matters
1:22:45 is here's here's what
1:22:48 matters if
1:22:50 like if Sam mland puts out the GPT store
1:22:54 and it starts becoming economically
1:22:56 viable and you got to protect your
1:22:57 interest then care but right now don't
1:22:59 care just build just build the [ __ ] and
1:23:01 put it out there play
1:23:03 collaborate do projects
1:23:05 why because in doing that you sort of
1:23:08 figure out where the edges are right
1:23:11 okay it's really good up to 80% of my
1:23:14 writing it's it's really good up to [ __ ]
1:23:18 10% of my video needs like it's video is
1:23:21 not really there but image needs it's
1:23:23 like 90% of my image
1:23:26 needs okay cool it can't spell though
1:23:29 [ __ ] it's all right so that's like 80%
1:23:31 [ __ ] so you kind of know where where the
1:23:33 edges are the boundaries it's good at
1:23:36 this it's not good at that oh if I if I
1:23:39 take this tool and I Stitch it together
1:23:40 with that tool and do that would be
1:23:43 really
1:23:44 cool but if that's hard to do and I'm
1:23:46 not Technical and then 3 months later
1:23:50 some technical person is going to stitch
1:23:51 all those things together and hand you a
1:23:53 tool that does them all and you're going
1:23:54 to know exactly what to do with it so it
1:23:57 doesn't matter anything that we're
1:23:59 building right now is absolutely
1:24:01 irrelevant a year from from now totally
1:24:03 agree with that
1:24:06 and the 10 20 50 70 hours a week that
1:24:11 you're spending building the useless
1:24:13 [ __ ] right
1:24:14 now is not time wasted in my opinion it
1:24:19 is time very well spent in fact if if I
1:24:23 knew it was going to take me 4 hours to
1:24:25 accomplish
1:24:27 something and I spent 4 hours trying to
1:24:30 get AI to accomplish that thing and
1:24:33 failed and then still had to spend
1:24:35 another 4 hours doing it the old
1:24:38 way I think that's 4 hours well spent
1:24:42 because I've learned where the
1:24:43 boundaries are and I tried to solve that
1:24:46 problem and what what
1:24:48 is more likely the case is that the 4
1:24:51 hours I spend trying to do it with AI
1:24:54 I'll probably actually solve it in an
1:24:56 hour and I didn't have to take that 4
1:24:59 hours in in most
1:25:01 cases there's there's things that I've
1:25:03 done where I'm like I just I just keep
1:25:05 hitting the wall I just I can get it to
1:25:07 80% I can get it to 80% I can't get it
1:25:09 above that and at some point it's just
1:25:12 like it it would have been faster to do
1:25:15 it the old-fashioned way but I'm so
1:25:18 glad I tried to do it with AI so um
1:25:23 yes yes
1:25:26 and still do it I only use do um I think
1:25:32 steo that will likely start to shift as
1:25:36 as some of these like like as mid
1:25:38 Journey as the web version of mid
1:25:40 Journey gets there and I like I have a
1:25:44 feeling I have a feeling some shit's
1:25:46 coming from from the other players like
1:25:49 why we haven't seen any multimodal stuff
1:25:51 from anthropic with with Claude is
1:25:53 really interesting to me I think Pi we
1:25:56 got some [ __ ] coming with pi that we're
1:25:58 not
1:25:59 considering
1:26:01 um you know meta putting some [ __ ] out
1:26:05 there Gemini and Gemini Ultra are coming
1:26:09 I I think there's going to be a lot of
1:26:10 stuff in the next 6 months that are
1:26:13 going to change the balance of tools
1:26:15 that we
1:26:16 use we will have AGI or something close
1:26:19 to it within a year yep totally agree
1:26:21 with that do you know one-on-one if you
1:26:24 don't know David Shapiro you should you
1:26:26 should watch him and his he's he's he's
1:26:30 still on track for I think I think what
1:26:33 he said is September of
1:26:35 2024 he thinks that we'll be at AGI by
1:26:38 by any definition of
1:26:41 it which is
1:26:44 interesting shoot it's already midnight
1:26:46 I know oh [ __ ] it's oh because I started
1:26:48 late I'm like why is it so
1:26:51 late um
1:26:53 100,000 T's seems like a win what's a
1:26:56 100 what do you mean 100K
1:26:58 T's seems like a win I don't know what
1:27:01 that means apple juice
1:27:07 water you need a long straw I totally
1:27:10 need a long straw I've got look at I see
1:27:13 the ice I've got oh you can't quite see
1:27:15 it I got ice on the edge of my glass
1:27:17 it's like that's my favorite that's
1:27:19 that's when you know you did it right
1:27:21 that's that's a perfect por for me if
1:27:23 you got ice on the glass you're doing
1:27:29 good taxi oh my God taxi was amazing mod
1:27:34 was great I'll give you that mod was
1:27:36 great that wasn't mod a spin-off of The
1:27:39 Jeffersons is that
1:27:42 right 10 is enough like Facebook or
1:27:45 YouTube or eBay for the internet same
1:27:47 type solutions for AI will appear
1:27:49 absolutely Earth rocker forever hooked
1:27:52 they play played the national anthem oh
1:27:54 that's right when the TV went
1:27:57 off You' be like [ __ ] I wanted to keep
1:28:00 watching
1:28:03 TV what does AGI mean exactly AGI stands
1:28:06 for artificial general intelligence and
1:28:09 and the definition from open AI so so
1:28:12 open ai's mission is to achieve AGI
1:28:17 their definition of it is that how do I
1:28:20 I think I've got it memorized we can go
1:28:22 find it
1:28:23 um autonomous systems that are better
1:28:26 than humans at most valuable tasks or
1:28:31 most is it valuable
1:28:39 most basically things you get paid for
1:28:42 that AGI is the point at which Chachi PT
1:28:45 is better than most humans at things you
1:28:48 get paid for at at our
1:28:51 jobs
1:28:54 and one of the rumors of
1:28:58 why Sam
1:28:59 mman was fired is that they effectively
1:29:03 achieved
1:29:04 it um the the rumor is the rumor is
1:29:08 they've got this thing called qar that's
1:29:09 no longer a rumor there there was a leak
1:29:12 there was a leaked letter that Sam Alman
1:29:15 characterized as an unfortunate leak
1:29:18 there was a letter written to the board
1:29:20 from some open AI employees about qar
1:29:23 this technology that they that ilas SSG
1:29:26 I think um was working on where they did
1:29:30 some some grade grade school level math
1:29:34 problems that open a that that whatever
1:29:37 version of what whatever this thing was
1:29:39 got 100%
1:29:41 correct not 80 not 90 not 97 100% which
1:29:46 means that it's reasoning um
1:29:50 and the the the other thing in that
1:29:52 sentence that I said autonomous
1:29:55 systems that are better than humans at
1:29:59 most I think it's valuable tasks
1:30:02 valuable work something like
1:30:05 that autonomous systems mean so the so
1:30:08 the current chat GPT is not autonomous
1:30:11 you go in and you prompt it and
1:30:13 it it gives you an answer and then if
1:30:16 you don't like it you go I don't like
1:30:17 that answer and it gives you another but
1:30:19 if you play with it a little bit there's
1:30:20 some hints of autonomous in in
1:30:23 there so if you do anything with code
1:30:26 interpreter if you upload a CSV file you
1:30:29 upload you upload a spreadsheet to chat
1:30:32 GPT and you say please analyze this data
1:30:36 and tell me what's interesting about it
1:30:38 and then visualize it for
1:30:40 me you sit back and it starts prompting
1:30:44 itself and if you click on the little
1:30:46 boxes where it says analyzing there's a
1:30:48 little pull down menu in fact we can go
1:30:50 we can go do it watch system are going
1:30:53 to kind of
1:30:55 do so we're going to have these agents
1:30:57 we're going to have these tools that
1:31:00 know who we are know what we want know
1:31:02 our data have access to our data have
1:31:04 access to things and they're going to go
1:31:06 do stuff for us but let me show
1:31:12 you a little bit of this in action
1:31:14 because it's it's very subtle right now
1:31:17 but but there's hints of what's coming
1:31:20 and it's when you think about what's
1:31:21 actually happening here it's it's kind
1:31:23 of mind-blowing so I'm going to go I'm
1:31:25 going to switch from GPT 3.5 to GPT 4
1:31:28 and when I I'm in GPT 4 I get this
1:31:30 little paper it's kind of mindblowing so
1:31:33 I'm going to go I'm going to switch from
1:31:34 GPT 3.5 to GPT 4 and when I I'm in gp4 I
1:31:38 get this little
1:31:41 [Music]
1:31:44 paper hold please hold please it's
1:31:48 coming I know that's one of the things I
1:31:50 say hold please people make fun of me me
1:31:52 for my hold please that's all right uh
1:31:55 no I want AR hang on
1:31:57 hold what's he doing I don't know he's
1:32:00 trying to find something he does
1:32:03 [Laughter]
1:32:11 this okay ai- ml YouTube videos. CSV ai-
1:32:18 ml
1:32:21 okay
1:32:23 so let me go find
1:32:24 AI was it underscore or was it Dash
1:32:30 mL no was
1:32:39 underscore okay CSV file okay there's a
1:32:43 CSV file it's it looks like it's YouTube
1:32:45 videos I have no idea what's in here but
1:32:48 what I'm going to do school is
1:32:50 irrelevant now it's getting close
1:32:53 um I think I think if schools can make
1:32:55 the shift to teaching critical thinking
1:32:58 um and and sort of refinement and
1:33:01 curation and
1:33:02 orchestration um I think that I think
1:33:06 that they have a very important role if
1:33:08 they can't make that transition I think
1:33:10 they're [ __ ] so I'm just going to say
1:33:14 what's in this data
1:33:19 and uh what are some interesting
1:33:26 Trends in it question
1:33:28 mark and so now it says analyzing so if
1:33:31 I pop open this little analyzing
1:33:33 box well what's that that's
1:33:37 python so what code interpreter does
1:33:40 what this tool does inside chat
1:33:44 GPT is it spins up what's called a
1:33:47 container and the container is like a
1:33:49 little mini computer with its own little
1:33:51 mini operating system and it's it's been
1:33:54 given access to python libraries so it
1:33:57 knows python so when you ask it to do
1:34:00 something it says hey what what python
1:34:05 do I need to write like what's the
1:34:06 what's the thing we're trying to solve
1:34:08 here based on that I'm going to use
1:34:09 these libraries in Python I'm then going
1:34:12 to write the python code I'm then going
1:34:13 to look at the data so it says the data
1:34:16 set contains information about YouTube
1:34:18 videos related to Ai and ml
1:34:20 topics here's the stuff
1:34:23 to identify interesting Trends we can
1:34:25 analyze aspects such as this this and
1:34:27 that then it sort of failed let's
1:34:28 conduct an analysis to uncover some
1:34:30 Trends it failed it seems there was an
1:34:32 issue with plotting Trends over time
1:34:34 likely due to the format of the Year
1:34:36 month column I'll adjust the approach to
1:34:38 ensure the data is in compatible
1:34:41 format I'll adjust the approach so it
1:34:46 failed it wrote this python code and
1:34:49 failed because the data was [ __ ] up
1:34:52 and then and look now it wrote python
1:34:54 code
1:34:55 to plot uh let's see monthly video
1:34:59 Trends monthly likes Trend monthly
1:35:01 comments
1:35:04 Trend it's having trouble with this file
1:35:08 for whatever reason but it's still
1:35:09 trying so this is autonomous behavior
1:35:13 and so it it finally it it it errored
1:35:15 out five times it finally got it right
1:35:18 top five five videos by
1:35:20 views
1:35:24 top five videos by likes so it analyzed
1:35:27 this data it's still analyzing this
1:35:30 data and within that it visualized this
1:35:34 like this is autonomous Behavior so so
1:35:37 agents where we move very quickly likely
1:35:41 in the first six months of the year is
1:35:43 that we're going to start to see if if
1:35:45 you think about the gpts that we're all
1:35:47 building right now they're all built on
1:35:49 top of this nonautonomous or
1:35:52 semi-autonomous thing if you use you
1:35:54 know code
1:35:55 interpreter this very simple version of
1:36:00 GPT when open AI
1:36:02 starts creating libraries of apis so
1:36:06 apis stand for application programming
1:36:08 interface so so when the ability to
1:36:11 connect chat GPT say to Google Maps or
1:36:15 Google search or um some database or the
1:36:18 New York Times you know uh archives or
1:36:22 or Getty
1:36:24 Images or travel sites or banking sites
1:36:29 or PayPal right all it's going to it's
1:36:31 going to start to be connected to all
1:36:33 sorts of
1:36:34 services and you're going to basically
1:36:36 opt it in to your accounts on all those
1:36:39 different
1:36:40 services and then at some point you're
1:36:43 just going to say hey chat GPT I need
1:36:45 this accomplished and it will go off and
1:36:48 start spinning up its own agents on your
1:36:51 behalf
1:36:53 okay I need an agent to go out and
1:36:54 search Google for cool places the
1:36:57 Shannon might like to vacation that they
1:36:58 haven't been before and doesn't break
1:37:01 the bank and there's no political
1:37:02 turmoil okay let me go read the news let
1:37:05 me go okay here's the four places that
1:37:08 might be good now let me compare those
1:37:10 four places and right and so there's
1:37:11 going to be all these agents that it
1:37:13 will
1:37:14 create you're going to give it a goal
1:37:17 it's going to generate these Dynamic
1:37:19 agents that are going to go operate on
1:37:20 your behalf
1:37:25 and then it'll come back to you and say
1:37:27 okay here's what I've done your Trip's
1:37:30 booked here's your itinerary for each
1:37:32 day you're like
1:37:34 oh I didn't want to go to
1:37:39 Tijana and and and CH like all right
1:37:43 boss I got it off let me go get you
1:37:46 another one it'll go unbook all that
1:37:49 rebook you somewhere else here you go
1:37:51 boss
1:37:54 Morin Mindy oh my God nanu nanu Robin
1:37:58 Williams was my [ __ ]
1:38:01 hero he was [ __ ]
1:38:05 astounding
1:38:10 astounding is Agi what used to be called
1:38:12 The Singularity no I think it's slightly
1:38:14 different but it I forget what the
1:38:17 singularity is Ray curtz is the one that
1:38:19 defined the singularity he his
1:38:21 prediction is 2029 we hit the
1:38:23 singularity I forget I forget the
1:38:25 definition of it I should know that but
1:38:27 I don't I don't remember it but AGI is a
1:38:29 step towards it um ASI artificial super
1:38:34 intelligence is the singularity I'm
1:38:36 pretty sure that's where that's where
1:38:38 these computers are superior to all
1:38:42 humans like all humans combined or
1:38:44 something like that
1:38:47 um and and that feels like we're
1:38:50 probably in the 5-year neighborhood hood
1:38:51 for for that but but this stuff's going
1:38:53 to start accelerating one of the one of
1:38:55 the things if you go watch David
1:38:57 Shapiro's videos um one of the things
1:39:00 that you'll see is
1:39:05 um there there's some of the rumored
1:39:07 stuff that was going on at open AI was
1:39:10 the engineers actually
1:39:12 stopped before they allowed whatever
1:39:16 this new system was to start improving
1:39:19 itself they they stopped before it
1:39:22 started doing that cuz what's going to
1:39:24 happen is these things are going to get
1:39:25 really good and then they're going to
1:39:27 allow the systems to train themselves
1:39:29 and they're they're going to get better
1:39:30 and better like not gradually but like
1:39:35 shink David Shapiro talks a lot about
1:39:38 the the with machine learning and AI
1:39:41 it's like it's like not good not good
1:39:43 not good not good not good good
1:39:47 excellent like like it just it just all
1:39:50 of a sudden figures [ __ ] out and then
1:39:52 goes from like bad to to like off the
1:39:55 charts good and so when these things
1:39:57 start improving themselves that curve
1:39:59 gets even steeper it basically just
1:40:01 becomes a vertical line like it's kind
1:40:02 of like overnight these things get
1:40:06 insanely
1:40:10 [Music]
1:40:13 powerful it's crazy isn't it it's crazy
1:40:18 it's crazy I tell you you can make money
1:40:21 with
1:40:28 that's tasty D that's good that's some
1:40:31 good whiskey yeah whiskey tasty M mhm
1:40:36 [Music]
1:40:37 yeah all right this is good as the
1:40:41 average this this is good as the average
1:40:43 person as AIS oh oh AGI is yeah that's
1:40:48 right AGI is when the computers are as
1:40:50 good as
1:40:52 the average human ASI is when it's
1:40:55 better than all
1:40:56 humans
1:40:58 yeah debt collectors do things like
1:41:00 negotiate debts and protect people's
1:41:02 ability to
1:41:03 pay that's going to get all automated
1:41:06 high value work nanu nanu to you and the
1:41:10 turtles so cool reasoning bang math are
1:41:15 both important but qar trains
1:41:18 reasoning yeah yeah but you you need
1:41:21 reasoning you need reasoning to do math
1:41:24 right so it's like if you hit the wall
1:41:26 you need to be able to navigate around
1:41:28 that that's what Q qar is doing qar is a
1:41:32 is a it's it seems like it's a merging
1:41:34 between Q learning and AAR which is a
1:41:37 game mechanics training thing they've
1:41:39 kind of merged those two
1:41:40 together it's really cool what it does
1:41:43 so so basically what qar does as far as
1:41:45 I understand it you have a state that
1:41:47 you're in so you want to accomplish
1:41:50 something and then in that that state
1:41:51 you can take these four
1:41:53 actions and then based on whatever the
1:41:56 probabilities are you decide to take
1:41:57 action
1:41:58 two and then what qar does is says okay
1:42:02 you took action two and now you're in a
1:42:03 new state based on that new state let's
1:42:06 grade how well you did how you know when
1:42:10 you took action two did it did did was
1:42:14 it a positive thing or was it a negative
1:42:16 thing and if it was a positive thing
1:42:17 it's like oh you're going in the right
1:42:19 direction and then now you have a new
1:42:21 set of actions you can take in state two
1:42:23 and that leads you to state three and
1:42:25 then it says how did you do good or bad
1:42:27 so it's keeping this running tally if
1:42:28 you know anything about card counting in
1:42:31 blackjack qar is kind of like a card
1:42:33 counter so the you know the card
1:42:35 counting system where if if you see 10
1:42:38 or face cards it's plus one if you see
1:42:41 whatever it is sixes or lowers sevens
1:42:43 are lowers it's it's minus one and then
1:42:47 the [ __ ] in The Middle's neutral and
1:42:48 then you're just keeping a running tally
1:42:50 is is the number going going positive or
1:42:52 going negative and depending on if if it
1:42:55 hits a certain Delta positive or
1:42:58 negative that's when you bet
1:43:00 heavier well that's essentially what qar
1:43:04 is right so it's like here's the thing
1:43:07 it's going to take an action and it's
1:43:09 just keeping this running tally of is it
1:43:11 you know are are we in in good territory
1:43:13 or bad territory and if it's in bad
1:43:15 territory then it starts to steer in a
1:43:17 different
1:43:18 direction that's reasoning That's like
1:43:20 [ __ ] what we do it's what we do
1:43:24 you're like going along okay I'm going
1:43:26 to do this for a while oh that didn't
1:43:27 look so good I think I'll go this way
1:43:30 now that's
1:43:33 it it's I what's amazing about some of
1:43:37 these models is like mathematically like
1:43:39 the mathematical formulas are just
1:43:41 [ __ ] insane they just make my head
1:43:42 hurt I just cuz I wasn't good at math
1:43:44 but when you understand what it's doing
1:43:46 conceptually you're like oh yeah it's
1:43:48 pretty it's it's pretty simple
1:43:50 conceptually
1:43:51 the way large language models work is
1:43:53 pretty simple conceptually it's [ __ ]
1:43:56 insanely complicated in
1:43:58 reality but it's like you wrote some
1:44:02 [ __ ] and based on some [ __ ] crazy ass
1:44:05 math thing it
1:44:08 generates a location in Vector space
1:44:12 some mathematical thousand dimensional
1:44:14 space that is the most probable point in
1:44:19 space to be the next right word and then
1:44:21 it Associates that point in space with
1:44:23 the word and puts that word in your
1:44:25 paragraph that's how that [ __ ] works
1:44:28 conceptually really simple but what it
1:44:30 does is [ __ ] magic it's like
1:44:35 miraculous it's like good Lord good Lord
1:44:40 people got 149 people in there welcome
1:44:43 everyone my name is Kyle Shannon this is
1:44:45 AI learning lab we just talk about [ __ ]
1:44:47 in here what I call Expert systems I'm
1:44:50 in in Amanda H is in awesome so join I I
1:44:55 hope you mean you're in with the AI
1:44:57 salon so if you're if you're new here um
1:45:00 look we got AI Salon Merch this is this
1:45:02 is the first shirt it's got a blemish on
1:45:04 it right there but it's okay I still
1:45:07 like it I like I love our
1:45:09 logo um so the AI Salon is is an online
1:45:13 community this this channel is kind of
1:45:15 an extension of it so this is a group
1:45:16 that I started um the week after chat
1:45:19 GPT came out
1:45:21 um so if you are experimenting with this
1:45:25 AI stuff and you're like no one in my
1:45:27 life is into this I don't have anyone to
1:45:30 talk to that's what this is that's who
1:45:32 we are that's that's who that is so if
1:45:35 you go there there's three links there
1:45:37 the first one is a link to our Mighty
1:45:38 networks Community the second one is a
1:45:41 link to our Meetup which you can sign up
1:45:43 for that but there's calendars in in the
1:45:45 mighty networks Community um which is
1:45:47 brand new it's like a month old uh and
1:45:50 then there's link links to the merch
1:45:51 store so if you if you end up buying
1:45:53 merch it supports the salon the salon
1:45:55 doesn't cost anything so I'm not this
1:45:57 I'm not like Hawking this like go to the
1:45:59 salon and I'll get you for your $149 a
1:46:01 month it's it's it's no cost it really
1:46:04 is just a community of people like
1:46:06 trying to figure this [ __ ] out together
1:46:09 so that's what it is um and and this
1:46:12 this channel is really just me trying to
1:46:15 figure [ __ ] out in real time and and
1:46:17 what's been really
1:46:19 cool is
1:46:21 um it's turned into kind of a jam
1:46:23 session so I'll I'll sometimes demo some
1:46:26 stuff and if I'm demoing stuff that
1:46:28 people who come here regularly have seen
1:46:31 before they'll just start generating
1:46:33 images and share them on the show and
1:46:34 tell channel here or put them in the
1:46:36 guilds or or things like that so um it's
1:46:39 it's a pretty swell thing when I hear
1:46:42 the theme song to The Night tonight's
1:46:43 show with Johnny Carson I'm immediately
1:46:45 put to sleep that's really
1:46:47 funny
1:46:49 um
1:46:56 oh I just went way back I'm in the
1:46:59 Wayback machine here June Mary hi Kyle
1:47:03 hello how are you so
1:47:05 cool all right I'm I'm with Dr J can't
1:47:10 spell where [ __ ] was but it was a good
1:47:15 [Music]
1:47:19 show prompts cart I think that is more
1:47:23 as ASI artificial super intelligence
1:47:26 here's the thing quite frankly about all
1:47:29 those labels AGI ASI The Singularity at
1:47:33 this point I think it's it's just
1:47:37 semantics like we're having arguments
1:47:39 now did we pass the Turing test or not I
1:47:42 don't know you know what when you
1:47:48 have when you launch pi so the Turing
1:47:51 test if you don't know is a test that
1:47:54 was developed by Alan Turing and think
1:47:56 in the 1950s that basically
1:47:59 said someone computer systems will pass
1:48:02 the Turing test when you can have a
1:48:04 human being interact with the
1:48:06 system and not be able to know if it's a
1:48:11 computer or
1:48:12 not and and now there's all sorts of
1:48:15 nuances to it and they're trying to
1:48:16 redefine it and it's like well it's got
1:48:18 to be the Turning like someone's got to
1:48:19 spend at least four hours inter
1:48:21 interacting with it like there's all
1:48:23 these qualifications to it now but
1:48:26 here's my is the Turing test gone or not
1:48:31 go use Alexa or
1:48:33 Siri and within three sentences it will
1:48:37 frustrate the [ __ ] out of you and you'll
1:48:38 be like this is
1:48:40 stupid I want to be able to ask it what
1:48:42 I want to ask it I don't just want to be
1:48:44 able to buy [ __ ] from their stupid
1:48:46 store you use
1:48:49 piie
1:48:58 I'm trying to figure out when we pass
1:49:00 the Turing
1:49:17 test come on girl where are you
1:49:21 Pi is thinking I wonder if I have shitty
1:49:24 internet
1:49:30 connection well we just didn't pass the
1:49:33 tur test did
1:49:38 we let me try let me try that
1:49:44 again how will we know when we've passed
1:49:46 the Turing
1:49:49 test
1:49:56 the cheing test is a bit of a tricky
1:49:58 Beast because it's not like there's a
1:50:01 clear-cut threshold that we need to
1:50:02 cross to say yep we did it it's more of
1:50:05 a gradual and evolving process and it's
1:50:07 kind of subjective too ultimately it's a
1:50:11 matter of whether an AI can engage with
1:50:14 a person in such a way that the person
1:50:16 can't tell whether they're speaking to a
1:50:19 human or machine it's a goalpost that
1:50:22 we're constantly moving forward towards
1:50:24 and it'll be a matter of debate when
1:50:27 exactly we've hit that
1:50:28 Mark but if I just wanted to talk about
1:50:32 Mor and Mindy and my favorite shows from
1:50:34 that era you could have that
1:50:36 conversation with me
1:50:41 [Music]
1:50:44 right heck yes we could totally nerd out
1:50:48 about Mor from orc and Mindy and all the
1:50:50 amazing shows that graced the TV screen
1:50:52 back then like how epic were Happy Days
1:50:56 the love boo and Charlie's Angels
1:50:58 Amorite and let's not forget the OG
1:51:00 Muppet Show now that was some quality TV
1:51:04 so come on spill the tea about your fave
1:51:06 shows
1:51:07 Kyle Gilligan's Island was certainly one
1:51:10 of them Beverly Hill
1:51:19 billies
1:51:22 oh man Gilligan's Island was pure gold
1:51:25 it was basically a sitcom about a group
1:51:27 of Misfits so I
1:51:33 mean I don't know about where you
1:51:37 were expectations wise but for
1:51:41 me when Alexa came out and Siri came out
1:51:45 and they're like oh this is AI driven
1:51:46 chat Bots my expectations were it was
1:51:50 would be something like that the reality
1:51:53 was they
1:51:57 suck and then the text based chat Bots
1:52:00 that were supposed to be AI based chat
1:52:01 Bots they're not AI based chat Bots
1:52:04 there might be some AI in there to help
1:52:05 them do the search but they're logic
1:52:07 trees and if you Veer off the path of
1:52:10 the logic tree it completely falls apart
1:52:14 these things don't fall apart so from
1:52:16 from where I
1:52:18 sit whether it's ing test or AGI or ASI
1:52:23 or whatever it's going to be this
1:52:24 gradual like I I feel like we're we're
1:52:27 already in New
1:52:29 Territory right we're already in
1:52:31 territory where I could go on a 4-Hour
1:52:34 road trip
1:52:36 and not get bored of the
1:52:39 conversation I could talk to pie like
1:52:43 I've been doing these things for 2 or 3
1:52:45 hours a night I could probably talk to
1:52:47 piie for 2 or 3 hours a night and
1:52:49 radically increase my understanding of
1:52:51 the world learn new things like it it it
1:52:55 could be absolutely insane and it it is
1:52:57 incredibly basic right now it doesn't
1:53:00 have any visuals it it it's got that
1:53:02 weird delay in it like all this stuff is
1:53:05 going to get you know sort of compressed
1:53:07 like latency is going to drop to zero
1:53:09 especially when these models are running
1:53:11 locally on
1:53:13 devices you're going to be able to have
1:53:15 experiences like that with zero
1:53:18 latency maybe with pi having a you
1:53:21 know
1:53:23 physical animated
1:53:26 Persona I just saw there's some new
1:53:28 virtual girlfriend thing that came out
1:53:31 on the app store today did you see that
1:53:32 it's like this voluptuous cartoon babe
1:53:35 that's like your virtual girlfriend it
1:53:37 just came out on the App Store let me
1:53:39 see it's I probably can't I can't show
1:53:42 up without getting cancelled by the Tik
1:53:44 Tok here let me go see if I can find
1:53:49 it
1:54:05 oh yeah I can't
1:54:06 I geez uh let's see
1:54:11 ioos
1:54:15 girlfriend uh uh uh uh uh uh can't you
1:54:21 can't find it I don't know well anyway
1:54:23 it's I can't search for it live on the
1:54:26 Tik Tok or I'll get
1:54:28 cancelled
1:54:30 but um but it's like you
1:54:33 know we're in new territory here it's
1:54:36 we're in we're in very new
1:54:45 territory all right so listen here's
1:54:49 what we're going to do I'm going to I'm
1:54:51 going to wind this down cuz I know it's
1:54:52 late for a lot of you there's a bunch of
1:54:54 people in here um go check out the AI
1:54:58 Salon the AI Salon is there that
1:55:02 address um check out about the salon
1:55:05 check out our values what the salon's
1:55:07 about if you resonate with that join us
1:55:10 introduce yourself go to the
1:55:11 introduction Channel and then share some
1:55:13 stuff on Show and
1:55:15 Tell um I'll be back here tomorrow um
1:55:20 I need to get caught up on the YouTube
1:55:23 channel so Archives of these lives are
1:55:26 on the YouTube channel
1:55:28 um which is ai ai learning laab dtt for
1:55:33 Tik Tok um if you want to see some of
1:55:35 the back issues or back episodes
1:55:37 whatever the [ __ ] these are the back Tik
1:55:40 toks um so I'll get caught up in that
1:55:43 tomorrow I'll be back at our normal time
1:55:45 likely 8:00 mountain time tomorrow and
1:55:48 uh we'll get going so what I want to do
1:55:51 thank you to the mods for modding
1:55:53 tonight uh thanks for hanging out
1:55:55 everybody it's good seeing youall let me
1:55:57 uh see who's in here June Mary good to
1:56:00 see you don't ask me Sherry D good to
1:56:03 see you papa JJ
1:56:06 Thelonius concerned citizens united high
1:56:09 value work don't know what that means
1:56:11 but yes good high value work Dr
1:56:15 J uh Milo's wife good to see
1:56:18 you source Camp let's see who else who
1:56:22 else is in here Amanda H 101 thanks for
1:56:26 chatting tonight Amanda H Sherry D Carol
1:56:30 Bernett oh yeah I loved Carol Bernett
1:56:32 God Carol Tim Conway was my flipping
1:56:35 hero it was Gilligan and Tim
1:56:38 Conway um mod was a spin-off of Archie
1:56:41 Bunker that's right I knew mod was a
1:56:43 spin-off of something was Archie Bunker
1:56:44 oh right she was Archie Bunker's cousin
1:56:46 cousin mod that's
1:56:48 right scientists say machine
1:56:51 intelligence is now capable of
1:56:52 replicating without humans yikes thanks
1:56:56 tonius how are we supposed to sleep
1:57:01 tonight you're still on your a cam Ken
1:57:04 good to see you Debbie all right
1:57:06 everybody um great to see you all have a
1:57:10 fantastic night thanks for hanging out
1:57:11 with me on a Friday night um I
1:57:13 appreciate all your time and I will see
1:57:15 you tomorrow peace out