11/22/2024 - Quantum Rap, AI Education & Creative Exploration: A Mind-Blowing AI Journey

Description

Chapters

Transcript

0:00 [Music]
0:10 [Music]
0:17 [Music]
0:33 [Music]
0:39 good
0:40 day good day champy
0:44 [Music]
0:59 [Applause]
1:01 [Music]
1:08 there been something baby I've been
1:10 trying to
1:13 say age and it seems I don't know
1:20 how the past and the future now
1:22 surrounding
1:25 me surrender to whatever
1:30 [Music]
1:31 there's been a little
1:34 trouble since you came to my
1:37 [Music]
1:40 rescue and if you like all of the rest I
1:43 would have quit you long quit you long
1:46 ago but I couldn't do
1:51 that oh tell me now women and why never
1:56 went too
1:58 well make a man C Mak him cold as
2:04 hell I know woman you wish me
2:08 well but
2:10 SP still going to have to find my way
2:15 through
2:16 [Music]
2:25 [Music]
2:36 [Music]
2:41 is this place I can rest my
2:48 forhead my thoughts in sweet
2:54 silence and buest place where the
2:57 feelings AR dead
2:59 [Music]
3:01 from an over exposure to
3:05 violence buest place I can slowly face
3:09 the only one I truly can
3:13 know these are tears from a long time
3:17 ago got these tears from a long time ago
3:21 I need to cry 30 years so these are
3:26 tears from a long time
3:30 [Music]
3:37 go well oh darling Oh Darling you say
3:41 unto
3:42 [Music]
3:45 me where have you been all my
3:51 life well I have been swimming seven s
3:58 Seas all women have toss me their
4:02 lifelines I pull them into the waters
4:05 darker in out have one them but I didn't
4:10 know these are tears from a long time
4:13 ago got these tears from a long time ago
4:18 I need to cry the years or
4:22 so these are tears from a long time
4:30 a
4:34 [Music]
4:47 go good evening good people hello people
4:51 of the Tik Tock people of the YouTube
4:53 people of the X it it would sound so
4:55 much better if it were people of the Tik
4:57 Tock people of the YouTube people of the
4:59 Twitter
5:00 but no Elon had to go and [ __ ] it
5:05 up anyway
5:09 hi my name Kyle Shannon hey KY hey ky ky
5:14 Cy
5:16 Cy this is the AI learning lab if you're
5:19 new here welcome we'll talk about AI
5:21 stuff
5:22 tonight this week we've been talking a
5:25 lot about a little rap song I put
5:28 together it's been getting been getting
5:30 a lot of Attraction which is cool
5:36 [Music]
6:04 [Music]
6:07 oh Freedom came out way that
6:13 night just like a jet plane in and out
6:17 of
6:18 side I was hauling ass at a million
6:21 miles an hour wonder how died
6:29 here when they came into the
6:32 [Music]
6:35 station they said I was mad beond
6:40 repairs but I got no quals with my
6:46 situation say here I
6:49 [Music]
6:53 am
6:55 so say Shere Shere Shere won't you d
7:02 say sh sh sh w't you
7:06 D say sh sh
7:11 sh
7:13 uhhuh yeah leave a message and your
7:17 number
7:20 please take a time to want to satisfy
7:26 me take all these old fantasies
7:30 send them care of
7:40 me I know champ good song solid song
7:43 solid song Marty Martin Saxton you are
7:46 brilliant thank you very much Steve I
7:48 appreciate that brilliant uh we will
7:51 leave for the brilliant people but I
7:53 appreciate the kind
7:55 words
7:58 um so what we're going to do tonight
8:00 there's there's a bunch of stuff Silver
8:02 Fox I've been inviting Alyssa Milano to
8:05 your times for about a week maybe
8:07 tonight's theight awesome hopefully she
8:10 comes bring her up here and give her a
8:13 little platform to chat
8:16 [Music]
8:21 away Brandon has figured out how to get
8:24 someone if we want to invite someone to
8:26 join the YouTube live stream which
8:28 you'll watch on that little window when
8:30 I flip it over here he's figured out how
8:32 to do that so if she does show up we can
8:36 make that
8:39 happen lots of lots of Hearts going on
8:41 on Tik Tok thank you if you're on uh X
8:44 or YouTube so here here's how this works
8:46 if you're on Tik Tok if you go to my
8:49 name and click on Youtube it'll jump you
8:52 over to the channel uh if you don't feel
8:54 like searching for it just go to YouTube
8:56 and search for at AI or no at
8:59 learninglab
9:01 Das
9:02 ai ai learning lab was taken I was too
9:07 slow
9:11 Che oh lid was
9:19 closed um but what we're gonna do
9:23 tonight is we're
9:26 gonna we're we're going to uh we're
9:29 going to continue playing around with
9:33 um the quantum Cipher and it you know
9:36 it's funny
9:39 like I always like that word Cipher
9:41 right you know for for for rap and and
9:45 just just in general it's just a cool
9:47 word and so we made this thing the
9:49 quantum Cipher or you know that that was
9:52 what I said I wanted an Eminem style
9:55 Cipher about quantum mechanics and then
9:58 I just didn't have a better name so just
10:00 called it the quantum Cipher but it
10:01 really is a cipher there's a lot in it
10:05 there's a lot in it um the other thing
10:08 that's kind of fascinating
10:10 is
10:12 um so the the maybe I'll maybe I'll just
10:16 play the video for you and then we can
10:18 we can talk about it um yeah why don't I
10:21 do that enough of me singing slightly
10:24 off key and slightly off beat let's
10:27 let's witness something that's actually
10:29 on the beat
10:31 a lot of people are like that's really
10:32 remarkable but it's soulless well here's
10:35 Soulful and you
10:38 know Divine Cipher for the people in the
10:41 back you know what's funny I asked I
10:43 think it was in this
10:45 chat um I said I said is cipher the
10:48 right word for the name of this song is
10:53 cipher
10:57 um the right
11:00 word for this
11:03 song and how is it
11:07 different
11:08 than Cipher with a Y which I don't even
11:11 know the answer to that maybe Cipher is
11:13 not even a
11:17 word oh I'm in 01 preview it's thinking
11:20 about
11:21 it um Cipher traditionally refers to a
11:24 secret disguise way of
11:27 writing um Cipher hipop culture oh so it
11:30 should be
11:32 why in hip-hop culture a cipher with a Y
11:36 is a gathering of rappers beatboxers
11:38 dancers in a
11:40 circle
11:41 Cipher is an alternative spelling of
11:44 Cipher but in modern usage Cipher with a
11:46 Y has become associated with the hip-hop
11:50 context well I can't go change all my
11:52 tweets I could change it on
11:55 Spotify this is what happens when we
11:57 move fast people
12:00 and when you trust chat GPT it was funny
12:02 in in uh in 40 I asked it and it said it
12:07 was right um given quantum mechanics
12:10 theme Cipher could oh oh
12:13 wait given the quantum mechanics theme
12:16 Cipher could cleverly underscore the
12:19 complexity okay so this is good all
12:22 right so here's what I learned I learned
12:24 that I used the wrong word my intent was
12:26 to use the one that was used in hipop um
12:29 and only had chat GPT tell me so it told
12:32 me wrong [ __ ] um and and so now that
12:35 I'm learning here the Y is the one for
12:38 hip hop the I is the one that means like
12:41 like a a
12:43 puzzle but the I that that's a good play
12:46 on words it it's it's it's very much in
12:49 theme with Eminem So anyway um that's
12:52 pretty
12:54 cool
12:55 so Quantum Cipher or Quantum Cipher
12:58 welcome to my Cipher
13:00 where every Line's a code to crack and
13:03 that's so that's that's what hit me so I
13:04 want to I want to go through a thing
13:07 that I posted on Twitter today about the
13:10 layers of complexity in the lyrics of
13:13 this song but before we do that why
13:15 don't we go play the song um
13:22 okay I gotta do this hang on a sec we
13:25 gotta do that we gotta do that and then
13:30 what we're going to do you know Kyle hey
13:32 Kyle KY Kyle hey Kyle all right there's
13:38 that there's two types of
13:41 Cipher there's an encryption Cipher and
13:43 Cipher in Hip Hop and yeah the cipher
13:45 and Hip Hop uses the why and that's what
13:48 I [ __ ] up on but I will just claim all
13:51 of you know the secret that I'm I was
13:54 just a dumb dumb but now that I know
13:56 that there's two
13:58 different versions and I chose the wrong
14:01 one I'm just going to say I chose the
14:04 right one it's a play on words like
14:06 Eminem I'm layered I'm
14:08 deep I'm hipop Rick riic one of the best
14:12 comments so so so for those of you who
14:15 don't know uh this this song that I'm
14:17 about to play you um it's sort of taken
14:20 off so there's a there's a Tik Tok video
14:22 in my channel where it says holy crap
14:24 and then I'm sitting in the car I'm just
14:26 like a fat old guy sitting in the car
14:29 talking about this hip-hop song that I
14:31 made and and it's gotten now like
14:34 330,000 views like it is it is eclipsed
14:37 anything I've ever done um one of the
14:41 one of the comments one of the comments
14:43 I got today was oh so what I guess like
14:46 fat old men sitting in cars are are
14:48 hip-hop stars now and I was like yep
14:53 apparently
14:55 so someone else was like do you even
14:58 know anything about hip-hop like not
15:02 really but I know words I know words and
15:06 I appreciate
15:08 words ah and that's that's one of the
15:11 lessons I think is that um one of the
15:14 lessons with
15:16 AI is if you're new to
15:20 AI the trick with it is use your words
15:24 right there's actually you know it's
15:26 funny there's there's a lot of parallels
15:29 between hip hop and and AI because it's
15:34 all about the words right it's all about
15:36 you know crafting and navigating using
15:39 words to navigate the world right and
15:42 that's kind of what hip hop is it's like
15:44 how do we how do we take this shitty
15:46 situation or this shitty life or this
15:48 shitty whatever it is and and you know
15:51 recontextualize it with
15:52 words that's all AI is it's like use
15:55 your words figure out what is the
15:57 combination of words that you can put in
15:59 into this bizarre machine to get it to
16:01 spit something back out that is a
16:04 version of what you wanted and it's you
16:07 know it's funny a lot of the comments
16:09 that are negative there's they're mostly
16:11 positive which is cool but a lot of the
16:13 comments that are negative are just like
16:16 it's it's that assumption that well say
16:17 I you just push a button there's no work
16:19 there's no craft and it's just not true
16:22 there's there's linguistic craft to be
16:25 able to get to the things to do what you
16:28 want them to do it's like I'm getting a
16:29 lot of comments like well how did you
16:30 get it to talk fast like that I can't
16:33 get it to do that well like I don't know
16:34 like maybe I just got lucky maybe it was
16:37 some combination of how I asked it I
16:39 don't know so anyway kind of fascinating
16:42 all right so how's 40 for screenwriting
16:46 I haven't explicitly
16:50 um
16:54 well okay so e so
16:57 okay 40 for screen writing it's a good
17:00 question 40 for screenwriting here let
17:02 me let me turn this thing
17:08 off chat gp4 has always been better at
17:11 writing than than the other llms until
17:16 within the past four or five months
17:19 Claude got really good at creative
17:20 writing so claude's good at creative
17:22 writing gp4 is good at creative writing
17:25 I think chat gp40
17:29 is really good at um story
17:34 structure but it's historically been bad
17:36 at dialogue right so if you want to get
17:39 the the scenes it would just do them too
17:41 short it it was just
17:43 shitty um and then when 0an preview came
17:47 out um as a test I uploaded a fragment
17:50 of a screenplay outline and a fragment
17:52 of a screen play that I had you know
17:54 started years back and I basically just
17:56 said finish the outline and it thought
17:59 about it for I don't know 59 minutes or
18:02 or 59 seconds or something like that
18:03 thought about it a lot
18:06 um
18:08 okay okay um and
18:12 um it it finished the outline
18:15 brilliantly and then I said I just
18:17 picked a random scene and I said write
18:19 this scene and it it was shockingly good
18:23 so 01 preview is good at writing I
18:27 assume you're going to need to give it
18:28 some sample writing that it can follow
18:30 but once once I gave it that it was
18:32 quite good I haven't tried doing
18:35 dialogue writing with the new creative
18:37 writing version of chat GPT that came
18:39 out this week so but I would imagine
18:41 it's probably pretty good all right what
18:45 are we going to do we are going
18:49 to stop that we're gonna come over here
18:55 [Music]
19:11 what the world needs now is love sweet
19:16 love all
19:19 right not the only
19:22 thing that there's just to Little
19:28 Love all right
19:33 [Music]
19:39 kids okay
19:41 so so this is the video in question
19:44 here's a fascinating thing about
19:46 [ __ ] Tik Tok yeah I said it uh
19:49 one streamyard started coming okay
19:52 um
19:54 so I put up the video of old man in the
19:59 car talking about holy crap this AI
20:02 shit's
20:04 good followed by the song that thing's
20:07 gotten 330,000 views it just keeps you
20:10 know it just keeps getting watched and
20:13 shared then I put up the song itself
20:15 with the lyrics and then I put up this
20:17 video on Tik Tok as Standalone
20:21 files they're like Tik tok's blocking
20:23 him and to the point that Kelly Bosch
20:26 reached out to me and she goes is Tik
20:27 Tok blocking the video cuz like normally
20:29 I get three or four thousand views like
20:31 that and I've only got a thousand I was
20:33 like I've got like
20:34 800 so the the shitty video of the fat
20:38 guy in the car that's getting all the
20:40 views nice and then but this this is
20:44 uh so the history if you haven't been
20:47 here the past couple of
20:48 nights five nights ago
20:52 now um I was just living my my life we
20:56 were doing these lives and then
20:59 the next day um chat GPT came out with
21:03 this new creative writing upgrade and so
21:05 I tried it and for the first time in two
21:07 years it wrote lyrics for an Eminem
21:11 style
21:12 rap really close to how Eminem does word
21:15 play and sophisticated word play Within
21:18 the lyrics I was like [ __ ] and so I
21:21 updated you know I refined it a bit and
21:24 then I took those lyrics put them into
21:27 suo V4 tuno version
21:30 4 and had chat GPT write me the prompt
21:34 for that right give me something that is
21:36 describes his music without using his
21:38 name and so I ended up with this song
21:41 and then the next day I asked Kelly
21:43 Bosch if she would do a video she showed
21:46 up last night with this video so in less
21:49 than like a 36h hour period this thing
21:53 went from not existing in the world to
21:55 this video so here you go so here's
21:58 here's what everyone's
21:59 been sharing and excited
22:01 [Music]
22:11 about I'm tangled with the beat spooky
22:13 action in the boo the Adam with a
22:15 pattern every bar burst of Truth super
22:17 position in the mission Fusion fusing my
22:19 intentions paralleling every line I'm a
22:21 prism of Dimensions mulic Antics I'm
22:24 manic with the semantics my rhs Collide
22:26 like particles Dynamic and gigantic
22:28 oscillating I'm weaving Quantum States
22:30 spit a verse and watch the Multiverse
22:32 reciprocate I'm discreet like planks
22:34 beat energy and tight packets when I
22:36 rapid M motion fing flows in the E
22:39 equals every I my P like antim when it
22:43 cles with quick as a flick of a I'm a
22:46 catalyst atomically spit in the script
22:47 I'm with the m wh fing the tempo theal
22:51 Crescendo sending my Sy slicing through
22:53 time like a sharpen new window qu
22:55 Precision I'm ripping the Rhythm I'm
22:56 spitting in prising Visions I'm shifting
22:58 Dimensions with lcal
23:01 in turul Sp I'm the velocity ripping the
23:04 your dation Cadence I stretch the second
23:07 I'm pacing Heisenberg with the mic
23:09 uncertainties the foundation observer
23:11 effect in effect the second I start to
23:12 speak collapse reality potential with
23:15 every word that I tweak yo my flows a
23:17 firm spinning in lyrical collisions
23:19 every bars a quk spark in a fusion of my
23:21 Visions I'm a wave in a particle
23:23 unpredictable art entangled in the
23:25 cipher splitting atoms with my heart
23:27 from wormholes to portals I'm immortal
23:29 the sches through the fabric of this rap
23:30 i s the threads of my dreams dark matter
23:32 in my patterns unseen but colossal black
23:35 whole soul pulling rhythms through the
23:37 gospel Double the Trouble I'm stacking
23:38 the rubble mys are com speeding and
23:40 weaving believe it I'm leaving your
23:42 speakers erupting I'm the equation of
23:43 pacing erasing the limits of sound
23:45 taking the crown breaking the bounds I'm
23:47 a force you can't slow down rapidly
23:49 tapping the map of my mind I Define
23:51 every pattern emotion bars of a bullet I
23:53 pull it and shoot it with ruthless
23:54 devotion I'm the ignition of friction
23:56 depicting a world that is vivid and
23:58 bright fire Fury got the track in the
24:00 blaze at the speed of
24:08 light shifting perspectives midline like
24:10 Quantum flips in my grind infinite
24:12 chances to shine I'm the master of space
24:14 every rhes of Singularity collapsing the
24:17 infinite I'm the big bang slang
24:18 expanding my imprint Limitless so step
24:20 into the quantum where the chaos aligns
24:22 my bars of entropic logic redefining the
24:24 Divine from the smallest of particles to
24:26 the cosmos that we know this is w limit
24:29 Quantum flow let
24:31 itow so so that's the video and it
24:34 sounded like some of you didn't have um
24:37 audio there so sorry about that don't
24:40 know what happened there let
24:41 me Let Me Entertain y I'm GNA
24:47 change change how we presting because
24:50 that's what we got to
24:52 [Music]
24:55 do there we go all right
24:59 beautiful buffering and blurring YouTube
25:02 can't keep up with it YouTube's trying
25:04 to understand the lyrics it's like I
25:06 can't get I can't do it I can't do it
25:09 people all right Emmy winner B and Emmy
25:12 category that would be cool wouldn't it
25:14 I would I would like that beautiful
25:16 thank you very much so um so even before
25:20 we had the video I mean the the visuals
25:22 in the video Kelly Bosch is just she's
25:25 just stunning the the work that that she
25:29 does
25:33 like she's so
25:35 hip like you know science science flying
25:38 all
25:42 around so even before we had the visuals
25:46 there's something about
25:48 this song that people so there were a
25:51 couple of things one is people literally
25:53 couldn't believe that AI could do that
25:56 lyrically or musically and again the
26:00 capabilities both lyrically and
26:03 musically went up a level this week
26:06 right so a week ago this project wasn't
26:10 possible at this level like you could do
26:11 a version of it but it wouldn't be at
26:13 this
26:14 level and then and then you know the the
26:17 other thing is just they're like you
26:18 know like the lyrics and the and the and
26:21 just it's like a beat I would listen to
26:22 and a lot of people like put it on
26:24 Spotify so it's on Spotify now so if you
26:26 search for the quantum Cipher C i p h e
26:30 r which we now know that's I I chose
26:33 that spelling correctly on
26:36 purpose it was not a mistake for
26:39 definitely for sure that's the correct
26:42 spelling as I intended because I am a
26:45 master of the
26:53 words oh good lord okay but
27:00 something struck me about the
27:03 lyrics and that was they're more
27:07 sophisticated well if I really took the
27:10 time to parse them I could parse them
27:15 but there's a lot going on so let's go
27:17 find the lyrics and we'll look at them
27:30 I can't even do it I can't even fake
27:31 word
27:37 it I probably have this somewhere else
27:39 here they are is this it
27:43 yeah all
27:46 [Music]
27:50 right you can see that right you can see
27:53 that good oh [ __ ] Tik Tok sorry about
27:55 that oh I am totally screwing up Tik
27:57 tok's life over here I don't have the
27:59 black thing up there you go let's see
28:02 let's get the words on
28:04 screen my big fat heads in the way but
28:07 you'll get the basic idea all
28:11 right what did she use to make the
28:14 visuals so Kelly used Dolly 3 so chat
28:17 GPT to make the visuals and then she
28:21 used Runway ml to animate
28:25 them Cipher is an ancient mimetic symbol
28:28 that rep represents life with soul at
28:30 its base oh nice oh Cipher with the
28:33 Y maybe take chunks of the L
28:37 lyrics and have chat
28:40 GPT prompt mid Journey that's not a bad
28:43 idea well so anyway
28:45 so so yo I I'm tangled with the beat
28:49 spooky action in the booth split the
28:52 atom with the pattern every bar is a
28:54 burst of Truth split the atom with a
28:56 pattern
28:59 um so amazing I'm tangled with the beat
29:03 spooky action in the booth so spooky
29:04 action that's what Einstein called
29:07 Quantum quantum
29:08 mechanics uh this is about quantum
29:11 entanglement um split the atom I assume
29:14 with a pattern means like you know it's
29:16 both a wave and
29:18 a um and a w wave and a
29:24 um physical object whatever the [ __ ]
29:27 [ __ ] that is
29:29 um superposition in the mission F Fu
29:33 Fusion fusing my intentions like that's
29:36 Fusion fusing my intentions that's such
29:39 a cool line parallel in every line I'm a
29:42 prism of
29:43 Dimensions so I had um chat GPT analyzed
29:49 this I thought rather than use my stupid
29:54 monkey
29:55 brain why don't I
29:58 have chat GPT use its crazy ass PhD
30:03 level
30:10 [Music]
30:24 brain let's see
30:37 so I gave it an objective understand the
30:39 key concepts of quantum mechanics
30:41 presented in the lyrics analyze the use
30:43 of language literary devices and poetic
30:45 structures explore the social and
30:48 cultural context of blending science and
30:50 hipop and then engage creatively with
30:53 the material through discussion and
30:56 activities so what's what struck me
30:59 about this project
31:01 is a lot of the comments that I'm
31:04 getting on both Twitter and Tik Tock are
31:09 like I want to use this for education or
31:11 this would be great to use for
31:13 education and so that it opened up kind
31:17 of a whole world to me that
31:21 that if you think critically and if you
31:24 think
31:27 creatively a about anything you produce
31:29 with
31:30 AI it gives you the opportunity to kind
31:33 of open the door to other kinds of
31:37 output right so if I was just if I was
31:39 just a music
31:41 producer and I produced that song or
31:45 even just that beat or the lyrics or
31:46 whatever it was then I'd go I'm gonna
31:49 make the song and then I might you know
31:50 make the video or that might come
31:54 later and I'd just be focused on on the
31:56 song and the track and maybe the
31:58 marketing for the
31:59 song but because AI can basically go
32:03 anywhere you want it to
32:07 go if you get an idea like someone says
32:09 oh you should use this for Education
32:12 well [ __ ] what could you do well the
32:14 lyrics are pretty dense what if we went
32:16 and looked at
32:17 them
32:20 so here are the lenses that chat
32:24 jpt used the language lens vocabulary
32:29 and diction literary devices rhythmic
32:32 structure social and cultural lens
32:35 Fusion of genres representation cultural
32:38 impact scientific lens Quantum Concepts
32:41 entanglement superposition planks
32:43 constant metaphorical interpretation how
32:46 are quantum Concepts used metaphorically
32:50 to describe the artist's skill and
32:54 experiences like
33:00 that it did that is just you know crazy
33:03 and that it can now analyze it and tell
33:05 us what's in there that's that amazing
33:08 mathematical lens philosophical lens
33:10 artist artistic and creative lens
33:14 right so then it designs activities for
33:17 us Activity one vocabulary building
33:19 enhance understanding of complex terms
33:21 in create instructions create flash
33:24 cards for each scientific term used in
33:26 the lyrics on one side write the term on
33:28 the other it's
33:30 definition literary device hunt um
33:33 identify and analyze literary devices in
33:35 small groups find examples of metaphors
33:38 similes alliteration internal Rhymes
33:41 discuss how they contribute to the
33:42 song's
33:44 meaning um so so you're getting you know
33:47 science you're getting
33:49 language um deep in the knowledge of
33:52 quantum mechanics choose one scientific
33:54 concept from the lyrics and prepare a
33:56 short presentation explaining in simple
33:58 terms creative expression write your own
34:01 verse or poem incorporating scientific
34:04 concepts with at least two literary
34:05 devices so this is this is one of those
34:08 things
34:10 where critics of AI sitting on the
34:12 outside will say oh AI produced it so
34:15 it's all
34:18 garbage but it's not and how you make it
34:22 obvious that it's not garbage is you use
34:24 it as a teaching device for students to
34:28 to do other things like you go write
34:30 your own verses now and if you think
34:33 these suck because AI wrote them go
34:35 write some better ones right I think
34:38 that's brilliant um interdisciplinary
34:41 connections metaphorical meaning um
34:43 innovative ways to engage students so
34:45 it's giving us a whole lesson plan
34:48 here
34:52 um all right let me scroll to the Bottom
34:55 now because I did some stuff this
34:57 afternoon that I that was pretty
35:01 good this is the teachers
35:07 guide and I put this if if you're here
35:10 on X one of my recent posts I I posted
35:15 some of this stuff on X today which I
35:17 thought was was pretty
35:22 [Music]
35:25 cool this one of one of the Challen es
35:28 of um of large language models is they
35:32 produce so much content that you're like
35:34 ah [ __ ] I got to go through the
35:38 [Music]
35:50 content huh what did I where did I
35:56 lose did I do
36:00 and hold please hold
36:14 please I did a statistical analysis
36:17 actually let me just go to my let me go
36:19 to my
36:21 Twitter and I'll I'll show you on my
36:23 Twitter
36:33 one of the one of the
36:36 uh um whatever you call it criticisms
36:41 that I got from people were that the
36:43 lyrics were word
36:45 salad now I might not know the
36:47 definition of word salad but I do know
36:50 that it's a disparaging remark that
36:53 means your words are meaningless sir
36:56 your words suck and my words are much
37:01 better um
37:05 so so I so I did this post on Twitter
37:08 word salad my ass you know some some
37:10 haters basically say it's AI Dil
37:14 um so what we created what I created
37:18 using chat GPT was a linguistic
37:20 scavenger
37:21 hunt and um so this is the analysis that
37:27 um that chat GPT did it basically said
37:31 there's 18 metaphors used five similes
37:33 12 instances of
37:35 alliteration 15 instances of internal
37:37 Rhymes 20 instances of multi- cabic
37:40 Rhymes um six double on tandas one of
37:43 the things that um the know--it alls
37:46 said in the comments was um yeah M&M's
37:50 use double on tandas there's none in
37:52 here you big fat loser they didn't say
37:54 that last part I I add that in my head
37:56 just to keep the shame stick rolling um
38:00 uh unique vocabulary words 120 literary
38:02 devices used over 50 in instances right
38:05 of of all these different things then
38:07 scientific references 25 terms um
38:11 quantum mechanics uh Concepts 12 of them
38:14 are in there um mathematical Concepts
38:18 philosophical Concepts right so so
38:20 numerically here's all the [ __ ] that's
38:21 in there and then so here's this the
38:24 scavenger hunt challenge find all 18
38:26 metaphors and identify the compar being
38:28 made like this is just brilliant so the
38:31 other thing that hit me is I'm looking
38:32 through this little scavenger hunt which
38:34 it it just sort of whipped this up and I
38:36 you know copied and pasted some of the
38:37 best parts of it and then I said here's
38:39 your lyrics go go have fun with it
38:44 um there's so many so many ways that we
38:48 could dissect this project we could do
38:51 um keep the songs in a similar style but
38:55 you know this is the song about quantum
38:57 mechanics and then we could do a song
38:59 about I don't know Agriculture and we
39:02 could do a song about biology and we
39:04 could do a song about geology or
39:07 geography and do a song about political
39:09 science right you could do all these
39:11 different songs and then every one of
39:13 them becomes a teaching opportunity for
39:15 that discipline but all these other
39:17 layers of linguistics and culture and
39:21 Society like that that's just stunning
39:24 right um or you could do everything's on
39:27 quantum mechanics but you do all sorts
39:29 of different musical styles or you do
39:31 all sorts of different hip-hop artists
39:32 like like there's just an infinite
39:34 number of ways you can cut
39:36 this the other thing that I thought
39:38 would be fun tonight is um is
39:43 to go over to Claude and see if Claude
39:47 can actually do some analysis and come
39:49 up with a an interactive dashboard for
39:52 us we'll we'll basically watch Claude
39:54 build an interactive
39:56 dashboard of these lyrics and so so I
39:59 don't know how we're going to do that
40:00 like one of the things that this channel
40:03 is about is just playing right just
40:05 explore very often that playing and
40:08 exploring leads to my red-faced
40:12 humiliation at how awful the evening
40:15 was but I'm willing to do that for you
40:19 all right let's see Kyle makes me laugh
40:22 it loud at work thank you very much um
40:25 let me see there's some starred comments
40:26 here let's let's let me just take a
40:28 break before we go off and and play
40:30 What's This say kakee on Tik Tok did
40:32 anyone hear about Claude and
40:34 mistl talking smexy to each other anyone
40:38 um there was a there was a
40:41 uh I forget the dude's name Brandon the
40:44 the bearded dude on Tik Tok isn't that a
40:47 music theory class like you could do
40:50 this with human generated songs yeah you
40:52 could you could certainly do it with
40:54 human generated songs um it would just
40:57 take you longer to do it Nate
40:59 Nate Nate
41:01 B on on Tik Tok he's this bearded dude
41:04 and he talks about AI a lot he talked
41:06 about um Claude talking sexy to to um
41:12 mistel so yeah I I heard about that I
41:14 didn't I didn't see the transcript ocean
41:17 on Tik
41:19 Tock what will AI not be able to do um
41:30 I just going to say have sex I'm like no
41:32 we're gonna have robots soon so so so
41:35 that's out what will it not be able to
41:37 do
41:41 um I don't know I don't
41:53 know it's a really good
41:55 question it's a really good question
41:59 I mean it sounds like in 2025 we're
42:02 going to get AGI which is artificial
42:04 general intelligence the point at which
42:06 these machines are
42:09 um you know smart smarter than the
42:12 average bear smarter than the average
42:13 person and not only smarter capable of
42:16 doing the economically viable work of
42:19 most people right initially that starts
42:22 out with knowledge work and computers
42:24 but
42:25 as these systems get into physical
42:28 robots it's also going to happen with
42:30 physical
42:31 work um Elon Musk says you know in the
42:35 next 10 years there are going to be
42:37 billions of you know um these humanoid
42:41 robots out there doing all sorts of
42:45 stuff so I don't know and as these
42:47 things get smarter and smarter and
42:49 smarter I think they start to do things
42:51 like solve all of our problems Ethan
42:53 mullik
42:54 today this was fascinating let me go
42:56 find this quote from him
42:59 if you don't follow Ethan mullik you
43:02 should he's a
43:05 um Wharton business school
43:09 Professor he's written a book called
43:11 co-intelligence not to be confused with
43:14 my book collective intelligence in the
43:16 age of AI two different
43:20 concepts um but
43:24 related but he said
43:29 what's
43:40 that for $24 billion doar we could have
43:43 prototype vaccines ready for each of the
43:46 26 known viral families that cause human
43:50 disease and be deployed in a 100 days
43:53 Jesus so that kind of [ __ ] could happen
43:58 okay here's the
44:00 post organizations including University
44:03 Labs should stockpile problems they
44:06 could use hon from the new set of
44:09 agentic and 01 like systems that are
44:11 coming soon I suspect that there will be
44:14 less useful they they they will be less
44:17 useful to non-experts at first but open
44:20 up with some new opportunities for what
44:21 AI can do I just I think this concept of
44:25 stockpiling problems
44:27 is genius one of my one of my challenges
44:30 with 01 the reasoning engine in cat
44:34 GPT it's way beyond my capabilities I
44:37 don't know what to ask it like my my
44:39 joke has been it's like I've got a best
44:42 friend that has 30 phds and can answer
44:44 any questions and I'm like uh yeah could
44:47 you tell me the direction to the gas
44:49 station I want to get some
44:51 twinkies and like he's like well it's
44:54 down there and you turn left at the Rock
44:56 is does that it is that all you're going
44:58 to ask me uh what's in a
45:05 Twinkie you know you know what I'm
45:08 saying here
45:09 people stockpiling problems do you know
45:12 how many unsolved problems there
45:14 are yeah but but I think like what it
45:18 implies to me Pate is like start
45:22 now you know identifying the problems
45:25 prioritizing the problems categorize the
45:27 problem s right put them in order of
45:30 order of operations if we solve this
45:32 problem then we could solve that problem
45:34 then we could solve that problem like
45:35 there's there's a lot of work to do to
45:37 be able to put these things forward
45:40 so I don't really like mik takes
45:43 interesting I find I find him fairly
45:46 straightforward and I don't find any of
45:48 his stuff particularly
45:50 controversial uh uh yeah have the
45:53 problems queued up and then just let let
45:55 the AI things knock them off
45:57 off knock him off knock him off all
46:00 right let's see what else we got
46:03 here all right what's
46:08 this you should reference We Didn't
46:12 Start the Fire from Billy Joel
46:17 nice early sunno days I made a song
46:20 about way faren for a pharmacist friend
46:23 of mine Kevin McAllister moment for him
46:26 that's super cool
46:27 yeah
46:29 like we were talking about this today
46:31 where was I talking about this oh I
46:34 think on the AI office hours we did
46:36 today um on on LinkedIn
46:41 um the the um the idea was basically so
46:47 so there's a guy from from Silicon
46:49 Valley that moved to the black forest in
46:51 Germany German guy I forget his name
46:53 Mike Mike something or other um or Mark
46:56 I think I'm might him in mark
46:58 um and and mark saw one of our Lives you
47:02 know where we made this hip-hop song and
47:05 he was like oh I've got this conference
47:07 I'm putting on let me make a song for
47:08 the conference and then like in the in
47:10 the
47:11 conversation the idea came up well he
47:13 could actually make a song for every
47:15 person coming to the conference right
47:18 they could each get their own
47:19 personalized song like I'm like yeah go
47:21 to their LinkedIn profile copy their
47:23 LinkedIn profile throw that into chat
47:25 GPT have it write the lyrics throw it
47:27 over into sunno have it write the song
47:29 and send it off to them it's like why
47:32 [Laughter]
47:35 not um Kelly was talking today about
47:37 digital Twins and we discussed what if
47:41 something happens to the person who the
47:43 business was built around would people
47:46 trust your digital twin to run your
47:49 business certainly not right now I think
47:51 you know the digital we're calling them
47:53 digital advisors now at at uh content EV
47:57 ution but I think
48:02 that so here's one of the things that's
48:04 probably going to happen in 2025 Google
48:06 said they've basically already cracked
48:08 this pate I'm curious if you know
48:09 anything about this is
48:13 um they've basically said that that the
48:17 that our these AI systems are going to
48:19 have infinite
48:21 memory of what of what we interact with
48:24 them so they'll be able to remember
48:26 anything you've ever talked about with
48:28 them and so as these systems get
48:31 infinite memory and they get agentic and
48:34 they get logic that's way beyond our
48:36 problem solving
48:38 ability um you're G to have this
48:41 personalized thing that knows exactly
48:43 what you want and how you write and how
48:45 you joke and how you like to do things
48:47 and where you like to go and the things
48:52 that you rated low on Yelp right it's
48:54 just they're going to get more and more
48:55 intelligent about who you are and what
48:57 you like and that that's um I think when
49:00 we start to have systems like that then
49:02 we'll be able to build digital twins
49:04 that are you know in the hood of being
49:09 able to run a business without us but
49:12 but
49:19 I it might not be advantageous to have a
49:23 digital twin of me run the business
49:27 right because I'll probably make
49:31 shittier
49:32 decisions than the AI system that's got
49:35 all of the knowledge of all of humanity
49:36 in it right so it might just be you know
49:41 maybe you just trust the AI system to
49:43 run your business and that's so
49:48 um open AI has created five levels of
49:52 AGI the first level of AGI is chat Bots
49:55 we've got that the second level is
49:57 reasoning which 01 preview is a
49:59 reasoning engine we've got you know the
50:01 the shitty version of that it's actually
50:02 pretty good but it's it's the shittiest
50:04 it will ever be it's a preview right now
50:06 it's kind of low on the totem pole um
50:09 the next level level three is Agents so
50:12 the ability for the computer to take
50:15 action on your behalf is level three
50:18 Claude has one of those out right now
50:20 and January we're going to get open AI
50:23 version of that called
50:24 operators um the fourth level up
50:27 innovators or inventors so if you
50:29 combine chatbots reasoning and agentic
50:33 behavior you can get them to do new and
50:36 novel things and then the fifth level of
50:39 AGI is what open AI calls organizations
50:43 so you take all of
50:45 this and then you dynamically spin up
50:48 employees for every person on on Earth
50:52 that's got an open AI account right so I
50:55 could say I want to go start a business
50:57 business doing
50:59 X and then the system just spins up
51:02 workers worldclass workers to go do all
51:06 that [ __ ] and I just sit back and go
51:08 yeah no yeah I can just direct
51:11 it we're quickly moving into a world
51:15 where we get to become directors like
51:19 creative directors and director
51:21 directors and you know storytelling
51:24 directors business directors things like
51:26 that so what am I supposed to chime in
51:29 on whether we have invented infinite
51:32 memory yeah I'm just curious if you've
51:33 heard anything like not not so much that
51:36 you've invented it I assume that you're
51:38 probably pretty close to it internally
51:40 just have you heard any Rumblings about
51:42 you know is it going to be released
51:45 anytime
51:51 soon oops
51:57 was the science accurate in your song
52:00 most of it was there was one
52:03 line there was one line in the
52:15 [Music]
52:17 song these are the orchestras of Our
52:20 Lives that's pretty
52:21 good Silver Fox Silver Fox coming in
52:25 with the with the silver tongue always
52:27 she's good with the
52:28 words there's all sorts of stuff in
52:31 Gemini coming soon no no idea what order
52:34 or when they will be made public yeah I
52:36 bet because it's been a while since I
52:39 mean Gemini's been rolling out a decent
52:41 amount of stuff
52:42 but mostly to the developer
52:51 Community I'll show you the line in the
52:53 lyrics that a I I've had a fair amount
52:56 of quantum physicist come into the come
52:59 come into my comments and like well you
53:01 know well you know technically the the
53:04 concepts in there are are relatively
53:05 sound however I I don't think any any of
53:08 my uh colleagues would would use the
53:10 term uh
53:13 metaverse so so there's been there's
53:15 been a fair amount of
53:20 that um this line right here got called
53:24 out by a Quantum physicist
53:28 my pen ignites antimatter when it
53:31 clashes with the
53:34 light so what they said is antimatter
53:36 when it clashes with light does not
53:38 ignite so they said this is inaccurate I
53:42 would argue that there might be other
53:44 layers of metaphor going on here so you
53:47 could you could probably parse this in
53:49 such a way this couplet you could
53:52 probably parse it in such a way that it
53:53 could make it work logically but that
53:55 line in particular was called called out
53:57 by a Quantum physicist oh that's not
53:58 quite accurate there but I think most of
54:01 the other [ __ ]
54:03 is I've been getting a fair amount of
54:05 quantum physicists commenting on the
54:08 lyrics which is hey it's [ __ ] awesome
54:10 I've also been as getting a lot of
54:11 teachers going hey can I use this song
54:13 for my physics class you know and I'm
54:15 like yeah go go grab it by the way this
54:18 is available on Spotify so if you search
54:20 the quantum Cipher c p e
54:23 r i have the same problem Kyle quantum
54:26 physicist coming into my comments I know
54:29 I've I've never you know it's funny an
54:31 ex- actor that talks about generative AI
54:34 with no um no skill or real interest in
54:38 mathematics or physics uh it's it's
54:41 surprising I haven't had more Quantum
54:43 physicist in my feed but but now I
54:49 do oh man um Can the system spin up
54:53 world class students so I don't lose any
54:57 more hair grading essay exams uh the
55:01 answer is
55:03 yes but I don't I don't know that's a
55:05 world we want to live
55:07 in
55:08 h just education in general
55:14 is it's Bonkers
55:16 man it's like I wonder I wonder if
55:21 teachers I don't know I don't know how
55:23 it's going to go I like I feel like
55:25 teachers are like right now teachers if
55:27 they Embrace AI they've got this really
55:29 cool collaborative assistant where
55:32 they're doing the teaching and they've
55:33 got this collaborative assistant as
55:35 these tools get better and better and
55:37 better and as they
55:38 become you know sort of brought to life
55:41 with voice and with with you know
55:43 physical robots and [ __ ] like that
55:45 you're gonna have a robot that's gonna
55:48 that's
55:49 gonna have way more capacity to teach in
55:53 an incredibly personalized way than the
55:55 teacher and so does the role of teacher
55:57 shift to more of you know counselor
56:01 sociologist
56:02 psychologist you know I I don't know I
56:05 like it it just feels Bonkers to
56:08 me I I don't know I don't have a good
56:10 answer for that
56:13 one
56:14 um all right what's this one say daisy
56:17 May Since the models were trained on
56:19 human knowledge at some point do you
56:21 think they could exceed human knowledge
56:23 enough to solve problems like incurable
56:25 diseases yes in fact that thing I just
56:27 read Ethan mik posted they said with $24
56:31 billion they could design vaccines for
56:34 all 26 classes of viruses that make
56:37 humans sick and deploy them in a 100
56:40 days for $24 billion doar which seems
56:42 like a lot but for the trillions of
56:45 dollars that sickness costs um that's a
56:48 pretty remarkable statement so yeah I
56:50 think that we're going
56:51 to so there's two things going on Daisy
56:55 this is this a really good question
56:59 the models are getting more capable and
57:02 and as much as as much as people say the
57:07 models can't create new novel ideas
57:10 because they've only been trained on
57:11 known
57:13 knowledge therefore by definition they
57:16 can't create new novel ideas and to that
57:19 I say how do human beings create new
57:22 novel
57:23 ideas well they learn [ __ ]
57:27 right so they jam into their head a
57:30 known concept and then they jam into
57:32 their head another known concept and
57:34 then two more and then two more and then
57:35 10 and 20 and they fill their head full
57:37 of all these
57:40 Concepts and at some point a human says
57:44 I'm G to combine these three
57:46 concepts that's
57:48 gonna you know spawn into some new thing
57:52 that no one's ever thought of
57:54 before but it's it's all fragments of
57:57 other Concepts
57:59 so large language models are good at
58:02 that so anyway so I think when prompted
58:04 right you can get these things to do
58:05 pretty novel stuff I mean I can at least
58:07 do it with with you know creative
58:11 development um and so I think that's
58:14 going to get better and better and
58:16 better there's an argument out there
58:18 that we're scraping all the data we're
58:21 running out of data but one of the
58:22 things that um uh open AI is apparently
58:25 working on P might know if they're doing
58:28 this on on Google is they're taking the
58:33 gp5 class models and they're using those
58:36 to generate synthetic data or wait yeah
58:40 uh they're taking some high powerful
58:42 model and they're using it to create
58:45 synthetic data so basically vetted
58:48 organized well um well
58:52 tagged um data about anything in the
58:57 world and use that synthetic data as the
58:59 training data um so so I don't think
59:03 we're going to run out of data which is
59:05 one of the one of the fears right now of
59:08 of these models as we're going to run
59:09 out of data but they're they're
59:10 inventing their own but all that you
59:13 know says to say that these things keep
59:15 growing and keep getting better and
59:16 better um and so yeah I think we're
59:18 going to probably within five years
59:21 start to see remarkable advances in
59:25 science and technology and remarkable
59:27 advances in Creative output I'm I'm I'm
59:31 calling the era that we're entering the
59:34 great
59:35 Renaissance and I don't know if you know
59:37 this but within the middle of the word
59:38 renaissance is AI
59:41 huh so I always spell it with capital AI
59:44 Ren Aon
59:47 AI all
59:49 right so I hope that was
59:52 useful have it on Tik Tok Khan Academy
59:55 started an AI based training app KH
59:57 Academy is doing really interesting
59:59 stuff with education so they always have
1:00:02 um but yeah they're they're they had
1:00:04 access to GPT
1:00:06 4 four or five months before everyone
1:00:08 else did so they've been doing really
1:00:10 good stuff with it Tom most of us
1:00:13 realize our job will no longer be the
1:00:15 dissemination of
1:00:18 information yeah I think I think that
1:00:20 our jobs shift more into that director
1:00:23 role like people that are going to be
1:00:24 really good at using AI systems are
1:00:27 people with good judgment people with
1:00:29 good critical thinking skills people
1:00:32 that can initiate ideas you know it's
1:00:36 funny one of the one of the the shitty
1:00:39 comments that I get on the on the uh the
1:00:43 hipop song is um oh [ __ ] I just lost my
1:00:48 train of thought doesn't matter all
1:00:49 right
1:00:52 um ocean so with AI
1:00:56 how the learning process will go how can
1:00:59 we convince our kids to attend
1:01:02 school they they will just say that they
1:01:05 should use
1:01:08 AI well human systems are going to be
1:01:12 the bottleneck here so I think I think
1:01:14 AI technology is going to advance
1:01:17 quickly um education education will get
1:01:21 personalized quickly but the
1:01:22 institutions of Education are not going
1:01:25 to move quickly so I think you'll start
1:01:27 to see fragmenting where some schools
1:01:30 will get really Innovative with AI and
1:01:32 they'll start to accelerate away from
1:01:34 the big
1:01:36 institutions
1:01:40 um and yeah I just I
1:01:43 think every human being on Earth is
1:01:46 going to be able to learn anything they
1:01:48 want instantly at
1:01:51 exactly the way they
1:01:53 learn exactly the pace they need it
1:01:55 exactly when they need it so I think
1:01:57 from kindergarten you know up and
1:02:00 probably even younger than that um we're
1:02:03 going to pretty quickly have a world
1:02:05 where we've just got instant access to
1:02:08 knowledge instant instant I tell you
1:02:15 instant I mentioned this before if you
1:02:18 write a song about lessons you can start
1:02:20 your own
1:02:22 University oh like if I made an album
1:02:24 with well so I thought about that like
1:02:27 my my problem is ADD now that I've kind
1:02:30 of done this thing with this song and
1:02:32 I've sort of figured out oh you could
1:02:33 run and you could do all this education
1:02:35 stuff off of it that's where my brain
1:02:38 starts to shut down and someone who's
1:02:40 got a more um operations kind of brain
1:02:43 should take over because it would be
1:02:46 really easy to design this this
1:02:49 educational system based on music in
1:02:50 fact it would be relatively
1:02:52 straightforward to build an app that you
1:02:55 just put in a topic in a musical style
1:02:57 and it generates all of this it
1:02:59 generates the lyrics it generates the
1:03:01 songs and you just go I like those
1:03:03 lyrics here let's let's tweak those it
1:03:05 does them in order right here's the
1:03:06 lyrics here's the song Here's visuals
1:03:09 for the video here's the lesson plans
1:03:11 right just it just automatically spits
1:03:13 all that [ __ ]
1:03:15 out like like you probably knock that
1:03:18 app out in a in a
1:03:21 weekend but but that's when the work
1:03:23 starts right then you got to start
1:03:24 selling it that hip-hop song is what
1:03:26 brought me in here welcome welcome
1:03:28 welcome so if you're new here so a
1:03:30 couple of things if you're new
1:03:33 here
1:03:35 um I want to share something with you
1:03:38 that I think will serve you
1:03:40 well and it is this the AI
1:03:44 salon so go to the salon. so so there's
1:03:48 kind of two components to what I do with
1:03:50 community building and AI education one
1:03:53 is this five nights a week I'm here by
1:03:55 the way happy Friday night date night
1:03:57 everybody forgot to say that earlier I
1:03:59 just went right into it we didn't have
1:04:01 any pleasantries about nachos or Hot
1:04:04 Pockets um so I do these five nights a
1:04:08 week and then I've also got the AI salon
1:04:10 so the AI Salon is a community that I
1:04:13 started in um December 2022 December 7th
1:04:18 2022 so this coming December 3rd is our
1:04:21 second year
1:04:22 anniversary um it's a group that is a
1:04:25 community of about 2,000 AI optimists so
1:04:29 if you go to the salon. scroll down and
1:04:31 say join our
1:04:33 community and then what you will
1:04:36 see in there is this site this is our
1:04:41 community site and it'll jump you to the
1:04:43 welcome page and you can watch a video
1:04:45 with myself and the co-founder Leah
1:04:47 fasten talking about the salon and why
1:04:49 it's
1:04:50 important um you can learn about this
1:04:55 which is the the journey to AI literacy
1:04:58 that we like to talk about in the salon
1:05:00 and that's basically just start with
1:05:02 playing don't take this AI [ __ ] so
1:05:05 seriously everyone's talking about it
1:05:07 like it's gonna [ __ ] ruin the world
1:05:09 it might but in the meantime just play
1:05:12 right like like you there's a lot of
1:05:14 things we can't control if the robots
1:05:16 are going to kill us like okay that
1:05:19 that's somewhere in the future in the
1:05:21 meantime we've got these amazing Tools
1:05:24 in our hands
1:05:26 and I think people are taking them
1:05:27 serious you know seriously on a lot of
1:05:30 levels that's things that they don't
1:05:33 need to worry about a lot of the fear of
1:05:36 AI is well I I I'm not good at math and
1:05:39 I you know I'm not an engineer and I'm
1:05:40 not good with computers you don't need
1:05:42 to be good with computers you just need
1:05:44 to be good with words you don't even
1:05:45 need to be good with
1:05:47 words there are tools out there let me
1:05:49 go show you one
1:06:05 [Music]
1:06:15 this is an app called Deco
1:06:18 here and Deco
1:06:21 here is an image generation tool and a
1:06:24 video generation tool
1:06:26 and what you're looking here looking at
1:06:28 here is Deco here right now is
1:06:31 generating original
1:06:34 images one of Deco here's claim to fame
1:06:36 it uses uh stable diffusion sdxl their
1:06:40 really fast image generation
1:06:43 thing and all of these images we're
1:06:46 looking at have never existed before and
1:06:49 what's amazing about this is I can click
1:06:51 on any image that I
1:06:54 like so I like this kitchen cool it's
1:06:56 kind of a 50s retro orange Punky kind of
1:06:59 kitchen there's this little button here
1:07:01 called create station oops sorry about
1:07:04 that kids there you go create station
1:07:07 and if I click on create
1:07:13 station it takes that image as the
1:07:16 source and now I've got my own station
1:07:19 think radio station right where instead
1:07:22 of
1:07:24 um in instead of songs just playing in a
1:07:28 linear fashion this is images being
1:07:31 generated in a linear
1:07:34 fashion these are all original images
1:07:37 every one of these images if I like it I
1:07:38 can click on it and save it or upscale
1:07:40 it or turn it into a
1:07:42 video I didn't use any words right you
1:07:45 know there was a year or so ago it was
1:07:47 like you got to be a prompt engineer
1:07:49 which I thought I've always hated that
1:07:51 term because it implies that you have to
1:07:54 be an engineer to be able to do AI
1:07:57 go to a thing like Deco here and just go
1:08:00 uh I I like that one uh let me save that
1:08:04 let's see how do I save this upscale
1:08:06 share add to library done saved uh I
1:08:10 like him he's
1:08:12 funny save the library save oh that's a
1:08:16 cool car save the library done right or
1:08:21 if you like I I don't I haven't uh paid
1:08:23 for a subscription here so I can't make
1:08:25 movies and I can't upscale but any one
1:08:27 of these you can you could use as a
1:08:29 starting point for some other thing I
1:08:32 need no words for
1:08:34 this most people don't know that this
1:08:37 exists that that that even something
1:08:39 like this
1:08:41 exists all right it's 2024 what in the
1:08:44 hell is a radio station ah sorry good
1:08:46 point P yeah a radio station in the
1:08:48 olden timey days we used to have a
1:08:51 box that's a really good point actually
1:08:58 oh good lord all right what do we got
1:09:06 here all right I got all the all the
1:09:08 starred starred comments I'm good with
1:09:11 they're done um oh that looks like a
1:09:13 trash
1:09:18 80 that's essentially when I walked into
1:09:20 my my math class the first day of sth
1:09:22 grade in
1:09:23 1977 there was a TRS 80 on every desk in
1:09:28 the classroom that's essentially what
1:09:29 they look
1:09:30 [Laughter]
1:09:34 like oh
1:09:37 man yeah the the deoh
1:09:40 here. station is is how you get here how
1:09:44 you get there in the interface is you
1:09:46 click on the let's see you go home and
1:09:50 you click on the explore button this is
1:09:52 stations right here
1:09:56 but
1:09:58 anyway let me go to one of my one of my
1:10:01 stations and I I'll actually share
1:10:04 something
1:10:05 here let's see maybe this one this is a
1:10:08 good
1:10:14 one all right I can't type and I don't
1:10:17 like voice to text and I still have my
1:10:20 socks knocked off every day yeah yeah
1:10:24 exactly
1:10:26 exactly it's and this gets back oh we
1:10:29 were talking about the salon this gets
1:10:31 back to the salon this idea of play
1:10:33 first mindfully create generously
1:10:36 lead the only way you can really
1:10:38 discover all of these goodies that are
1:10:41 out there for you to play with is
1:10:44 to play to navigate to be in a community
1:10:49 like the AI salon so so there's kind of
1:10:51 this seven-step welcome process here and
1:10:55 you know first we ask you to do is
1:10:56 introduce yourself so use your real name
1:10:59 put your picture in there get known in
1:11:01 this
1:11:02 community um you can you can learn our
1:11:06 values the values at the salon are
1:11:08 curiosity
1:11:10 playfulness
1:11:12 empathy like a lot of people I'm seeing
1:11:15 it because I'm getting all these
1:11:17 comments a lot of people are scared some
1:11:20 people are pissed off one woman today it
1:11:22 broke my heart she goes I'm I'm crying
1:11:24 I'm literally crying and not in a good
1:11:27 way right she heard that song and that
1:11:30 that song represented for
1:11:33 her you know an existential
1:11:38 crisis and my only response to her was I
1:11:41 completely get it and get your ass into
1:11:45 the AI learning lab and get your ass
1:11:47 over to the AI
1:11:48 Salon the the the thing that I've said
1:11:51 more than anything I'm trying to go
1:11:52 through and respond to all the comments
1:11:54 it's like it's mathematically imp
1:11:55 possible but I'm
1:11:57 trying and the thing I'm saying more
1:11:59 than anything is those people that think
1:12:02 of AI as their collaborator rather than
1:12:05 their enemy are the ones that are going
1:12:08 to
1:12:09 win
1:12:11 because every single comment that's like
1:12:14 ai's evil ai's [ __ ] up AI is going to
1:12:16 kill us are not using
1:12:19 AI but that fear is real and it's
1:12:22 visceral and it's existential and so
1:12:25 this community empathy is at the core of
1:12:30 it um bravery right you got to be a
1:12:35 little [ __ ] nuts and willing to go on
1:12:38 the adventure to choose to be on the
1:12:42 adventure with these tools that you're
1:12:44 like I might not be able to do this and
1:12:45 I don't know how to do this and what's a
1:12:47 GPT and what does GPT mean and they're
1:12:49 talking about embeddings in latent space
1:12:51 and it sounds like math and it makes my
1:12:52 head hurt who are you people
1:12:56 why would you step into
1:13:00 that because AI is not going
1:13:06 away it's not going
1:13:09 away it's not
1:13:13 hype it's profoundly powerful technology
1:13:17 that countries do you know what China
1:13:20 and the US are talking about right now
1:13:23 as a way to maintain their competitive
1:13:25 advantage who can build the most power
1:13:27 stations to be able to process the most
1:13:30 AI to be able to do like this is this
1:13:33 [ __ ] ain't going away it's being talked
1:13:35 about at a a fundamental infrastructure
1:13:38 level as how we keep our advantage in
1:13:41 the in the
1:13:43 global
1:13:45 Enterprise so it's not going
1:13:48 away so you can be pissed off about it
1:13:51 and you can be pissed off that it was
1:13:52 trained on data that you know they
1:13:54 didn't compensate the people for you can
1:13:56 be pissed off about that it works too
1:13:59 fast you can be pissed off that it does
1:14:00 work that you know I took 20 years of my
1:14:03 career to get good
1:14:05 at and it's not going away so like
1:14:09 independent of your feelings about it
1:14:12 it's here and it's
1:14:15 accelerating and My Philosophy is and
1:14:18 has been from day one just me
1:14:22 personally I've got two choices with
1:14:27 AI everyone has two
1:14:29 choices
1:14:32 two you can sit on the sidelines and be
1:14:37 right or you can pick up a surfboard and
1:14:40 start paddling out and catch a wave I
1:14:43 don't know how to surf well those are
1:14:45 your two
1:14:46 choices do you want to sit in your cozy
1:14:48 beach hut knowing that the tsunami is
1:14:51 coming you'll be like Oh I'm fine the
1:14:53 water's the what this Tide's really low
1:14:56 yeah that's cuz all the water's getting
1:14:57 sucked up into that
1:14:59 wave or do you want to be out there
1:15:01 paddling with a bunch of other people
1:15:03 trying to figure this [ __ ]
1:15:05 out and then the last um value of the
1:15:08 salon is
1:15:19 generosity I save so much time not
1:15:22 having to deal with dumb coding problems
1:15:24 and can focus on big things that's from
1:15:26 Pate modder so Pate Pate is an irregular
1:15:29 guy that shows up to these things um you
1:15:31 know with with us night after night
1:15:33 which love that you come here Pate um
1:15:36 Pate works for Google he's got a PhD in
1:15:40 mathematics
1:15:43 AI is allowing Pate to take the big part
1:15:47 of his brain which
1:15:50 is which is which is you know
1:15:53 disturbingly bigger than mine
1:15:56 and let him focus on the big stuff right
1:15:58 in in any given eight- hour
1:16:00 workday what is most of your time spent
1:16:04 on being on hold with text support
1:16:06 trying to write an email trying to just
1:16:09 do it's all this noly little [ __ ] in in
1:16:12 his case small little coding problems I
1:16:15 gotta write the
1:16:16 code but what he really needs to do is
1:16:18 be thinking about what is the problem
1:16:20 he's got to be thinking about that
1:16:22 caused him to write the code in the
1:16:23 first place right well now ai can do all
1:16:26 that so he can operate at a much higher
1:16:29 level and that's been my experience with
1:16:30 it as well like it AI does the shitty
1:16:35 part and you just get to live up here
1:16:39 right with the with the the hipop
1:16:43 song I've been wanting AI to create
1:16:46 lyrics that I thought were interesting
1:16:48 for two years and it hasn't been able to
1:16:50 do it that I
1:16:52 found and then it could oh great that's
1:16:55 cool let me take those lyrics and put
1:16:57 them in sunno and see if it sounds like
1:16:59 a cool song oh that's a cool song oh let
1:17:01 me reach out to Kelly Bosch who makes
1:17:03 cool videos and see if she'd make a
1:17:05 video oh that's a cool video here's the
1:17:08 thing I never
1:17:11 left the higher level idea
1:17:16 plane but I still created that thing
1:17:19 co-created it with
1:17:23 Kelly and people like well cuz you used
1:17:25 AI it's not real
1:17:27 work did it make you do
1:17:33 this did you share it with your
1:17:36 friends then it's real work if you like
1:17:39 the song and then you found out it was
1:17:42 Ai and then you hated the
1:17:44 song values are all [ __ ]
1:17:47 up anyway all
1:17:51 right the tsunami is coming is not
1:17:53 helping with those with anxiety that's a
1:17:54 good point I I got to find another
1:17:58 metaphor but listen I I mean let's not
1:18:02 let's not kid ourselves like if you've
1:18:04 seen the
1:18:06 um the world economic Forum I think they
1:18:09 just put out a a study and it's like
1:18:12 like this AI is GNA be it's going to
1:18:15 decimate the job market it really is
1:18:17 it's going to be it's going to be
1:18:19 crazy um Kyle whoever is still in denial
1:18:23 about AI do you blame them
1:18:26 completely agree that we have to try do
1:18:27 I blame them I don't blame them if if
1:18:31 you're still on the sidelines about
1:18:40 AI I just
1:18:43 listen I spend two to three hours a
1:18:46 night every night here at least five
1:18:49 nights a week I've got the AI Salon
1:18:52 that's probably another 10 hours of my
1:18:54 week organized doing meetings things
1:18:56 like that I have a Friday afternoon
1:18:59 office hours on
1:19:00 LinkedIn all of them have the same
1:19:04 Mission I think it is absolutely [ __ ]
1:19:07 criminal for those of us those of us who
1:19:12 understand what's happening with AI not
1:19:14 to try to help people up into the
1:19:17 boat and so like for the last two years
1:19:21 my mission has been to demystify AI make
1:19:25 make it safe and fun and playful you
1:19:28 know why are our why is the first thing
1:19:31 we talk about
1:19:32 play because when you play with AI what
1:19:35 you really quickly understand is
1:19:38 that it is inspiring and it lights up a
1:19:43 creative spark within you that in many
1:19:46 cases you thought had died 30 or 40
1:19:49 years ago or if you're only 30 you know
1:19:52 15 years ago but but
1:19:56 it's it's this remarkable thing and it's
1:19:58 one of those thing the thing the thing
1:20:00 about AI right now I think it will be
1:20:03 for a
1:20:04 while is that it's kind of like when
1:20:08 when you do a self-help program or you
1:20:10 go to one of those weekend seminars like
1:20:12 a Tony Robbins thing or the Forum or one
1:20:14 of those stupid [ __ ] things I mean
1:20:15 I've done I've done them all so I speak
1:20:18 from experience and listen they're
1:20:19 incredibly
1:20:22 powerful but you know how those
1:20:24 self-help we
1:20:26 weekends you can't describe
1:20:29 them like you you try to say to someone
1:20:32 hey was just amazing thing and
1:20:34 everything's opened up for me and it's
1:20:35 just everything's possibility and people
1:20:37 like what are you talking about and then
1:20:40 especially if you're in something like
1:20:41 the Forum where they teach you a whole
1:20:42 new vocabulary you're using your new
1:20:45 vocabulary and people are looking at you
1:20:46 like what the [ __ ] are you talking
1:20:49 about but inside you're like alive and
1:20:52 you're like oh my God you gotta you
1:20:53 gotta feel this thing and see this you
1:20:55 got to see what's
1:20:56 possible that's what AI is like right
1:20:59 now that the minute you actually use it
1:21:02 for one thing the fear just disappears
1:21:05 because you realize oh [ __ ] that could
1:21:08 help me do my job better oh wait I've
1:21:10 always wanted to start a job and maybe I
1:21:11 could learn how to do that thing I never
1:21:13 knew how to do or maybe it could just do
1:21:14 it for
1:21:16 me but you can't explain that to someone
1:21:19 so the the only way I know how to do it
1:21:21 is like join a community and start
1:21:23 hanging out with people that are already
1:21:24 Cur ious about this and are not shaming
1:21:27 you for it that's what this group is
1:21:30 that's what the AI salon's all
1:21:32 about all
1:21:36 right people who don't understand AI are
1:21:39 really not going to understand why Ubi
1:21:43 will be necessary well they will they
1:21:46 will understand why Ubi is necessary
1:21:49 unfortunately
1:21:50 [Laughter]
1:21:55 I've been in a Kevin McAllister moment
1:21:57 for the past past several hours writing
1:22:00 the last few novel chapters I can't even
1:22:03 explain how far into the singularity
1:22:05 I've been
1:22:07 diving yep there you
1:22:09 go
1:22:11 that is what it's
1:22:17 about it really is amazing it's amazing
1:22:21 all right so anyway so go join the AI
1:22:22 salon and again here's the URL
1:22:25 if you want to you can see it in the The
1:22:28 Mighty Network's URL is in the in the
1:22:30 screen there but that's our website
1:22:33 that'll take you
1:22:35 there I've got to get uh this site back
1:22:39 into hang on I should probably do that
1:22:41 now shouldn't I can I do it
1:22:45 [Music]
1:22:48 here no [ __ ] all right whatever I got to
1:22:52 get it back into dark mode because that
1:22:53 that white screen is just
1:22:56 obnoxious oh this is a cool thing that I
1:22:59 wanted to do okay
1:23:03 um so wait how do I do this I go
1:23:07 here I go
1:23:09 [Music]
1:23:14 here okay for those folks on
1:23:20 YouTube I am going
1:23:23 to share the URL for that
1:23:27 station so all the images what what's
1:23:30 cool about de this is decco here again
1:23:32 Deco here d c h re.
1:23:38 a and if you if you copy that link or
1:23:42 click on that link I don't know if it's
1:23:44 clickable if you click on that link it's
1:23:46 going to take you to my station what's
1:23:48 cool about this
1:23:49 feature is that you're going to see the
1:23:52 exact same images that I'm seeing any
1:23:54 one of the images that you see that you
1:23:56 like you can click on it and turn it
1:23:57 into your its own
1:24:00 station but like this could like one of
1:24:03 the ideas I had wouldn't it be cool
1:24:05 being in a doctor's office uh well maybe
1:24:08 not with these images maybe these are a
1:24:09 little too radical for a doctor's office
1:24:11 maybe an art gallery for this one but um
1:24:15 you know to to be in some public space
1:24:17 and it's just constantly it's just
1:24:19 rainning this beautiful art and then any
1:24:22 any one of that art maybe you scan a QR
1:24:24 code and you can just go put that on
1:24:26 your phone
1:24:29 and and someone could be curating
1:24:38 these I mean what if you
1:24:41 took an image like this
1:24:45 one and you took an image like this
1:24:48 one and you put those into chat GPT
1:24:51 because it can see images in fact let's
1:24:53 do this right now I'm Gonna Save
1:24:58 that and where's that other image was it
1:25:01 that one no we do this one I'm Gonna
1:25:05 Save
1:25:10 that Silver Fox this group is so
1:25:13 embedded in my life that I don't know
1:25:16 what the outside Salon life is a silver
1:25:20 fox that's cool y'all sweet um the the
1:25:23 people that come here night after night
1:25:25 we call them the Irregulars because
1:25:26 they're just not regular like why would
1:25:28 you come here and listen to you know me
1:25:33 [ __ ]
1:25:34 Babble over and over again because
1:25:37 here's
1:25:39 why what happens
1:25:42 here to a great degree has nothing to do
1:25:45 with me it's a bunch of people that show
1:25:47 up and what what I try to create here
1:25:50 and what this space really
1:25:52 is is a space that allows us all
1:25:55 including myself to be in the
1:25:57 conversation to be in the conversation
1:25:59 about
1:26:02 AI what's that mean be in a conversation
1:26:04 I it just means be in the conversation
1:26:06 most of the Irregulars most of the time
1:26:08 are like doing the dishes or falling
1:26:10 asleep or you know not NE necessarily
1:26:13 paying attention to every single thing
1:26:15 they're just in the conversation and at
1:26:17 some point maybe I show something
1:26:18 they're like well let me go try that oh
1:26:20 yeah I forgot about Deco here let me go
1:26:22 try that let me grab this this image
1:26:25 that one's
1:26:28 crazy and so being in the
1:26:31 conversation about AI is the only way
1:26:35 that I know like right now this all
1:26:37 these tools are so new no one can teach
1:26:41 you this stuff by the time someone gets
1:26:43 together a course to teach you something
1:26:46 about generative AI the tech will have
1:26:47 moved along like that's happening
1:26:50 regularly so you can take those courses
1:26:52 and they might help give give you you
1:26:55 know give yourself permission to go okay
1:26:57 now I'm willing to try it because I took
1:26:59 a course but you don't need to take a
1:27:00 course just hang out and start playing
1:27:03 and be in the conversation that's why
1:27:06 they're irregular right it's not normal
1:27:09 to think about this stuff all the time
1:27:13 except
1:27:14 that AI is going to change everything
1:27:17 and I'd much rather be in the
1:27:19 conversation than on the outside looking
1:27:20 at okay so let me show you what we're
1:27:24 going to do here I'm gonna I'm going to
1:27:25 go to chat
1:27:27 GPT um so I got to go to my images first
1:27:30 let's
1:27:32 see where oh
1:27:36 Library okay so here's my that image
1:27:40 here's my that
1:27:42 image and here's my that
1:27:46 image so
1:27:48 without prompting anything I opened a
1:27:52 station that I liked it made some images
1:27:54 I picked some that I that I I
1:27:59 dug and now I'm going to go to chat
1:28:02 GPT you can make money with I'm g go to
1:28:08 chat and I'm gonna create a new chat
1:28:11 we're going to do a 40
1:28:15 chat and I'm GNA click on this little
1:28:18 paperclip and we're going to upload
1:28:20 those pictures
1:28:29 okay and I'm just going to
1:28:32 go describe
1:28:38 these so again prompt engineering is you
1:28:43 got to learn all this promp thing you
1:28:44 got to learn the way of the mysterious
1:28:46 prompt
1:28:47 Masters what I'm doing is there are
1:28:50 tools out there where if you want to get
1:28:52 into AI
1:28:54 you don't need many
1:28:56 words what are these describe
1:29:00 these um YouTube Spotlight I see no
1:29:06 Spotlight um okay Drew I did my first AI
1:29:10 seminar last week have to thank Kyle for
1:29:12 helping me on this journey Drew
1:29:13 congratulations that's awesome I hope
1:29:16 they gave you lots of lots of cash um
1:29:19 president AI on Tik Tok any one of those
1:29:22 images could be worth $100 million yeah
1:29:24 did you see the bosot um there was a
1:29:27 bosot that went for I think it was was
1:29:30 it $22 million today it was like a
1:29:32 little dirty
1:29:35 sketch let's see basat is it was it with
1:29:38 the Q wait BOS
1:29:41 bosot isn't it
1:29:46 that yeah this one this this this went
1:29:49 for 20 22 oh no yeah yeah $22 million
1:29:53 today
1:29:57 it's like dirty like like you know it's
1:29:59 it was like laying on his Studio
1:30:03 floor he's
1:30:05 like $22 million but look at it it's
1:30:09 [ __ ]
1:30:10 beautiful all right first image is a
1:30:14 dystopian cityscape second image is a
1:30:16 vibrant abstract portrait of a human
1:30:19 face third image um
1:30:24 so here's what I'm going to
1:30:26 do so we've put in those three
1:30:29 pictures we've had chat GPT 40 which is
1:30:33 like the B it's like your Toyota Corolla
1:30:35 the the car you drive to the grocery
1:30:37 store we had it described the images a
1:30:40 vibrant abstract portrait of a human
1:30:42 face the artwork integrates geometric
1:30:44 shapes Vivid colors patterns blah blah
1:30:47 blah
1:30:48 right oh I I don't think I can switch to
1:30:51 preview now can I all right so we'll
1:30:54 just do it within within within this so
1:30:57 let's do this I'm going to switch to GPT
1:31:01 40 with canvas so if you don't know what
1:31:03 canvas is canvas is going to create an
1:31:06 editable document beside the the chat
1:31:09 that we're in so if you're just in
1:31:11 normal gp4 you just have this long
1:31:14 scrolling cat or chat if you go to
1:31:17 Canvas you can create these editable
1:31:19 documents within a chat it's pretty cool
1:31:22 and I'll show you I'll show you what we
1:31:23 do here
1:31:25 can you combine the images in Deco here
1:31:28 I think so I think there's a way that
1:31:31 you can smash images together in decco
1:31:34 here decco here is worth playing with
1:31:36 there's all sorts of Cool Tools there
1:31:37 like I just showed one tool there's like
1:31:40 20 and that's again that idea of play
1:31:42 first like just go find a tool like Deco
1:31:45 here and take the weekend with it right
1:31:48 you know take it to bed with you on the
1:31:51 laptop if it's got an app go get the app
1:31:53 just start
1:31:55 playing um but here's what we're going
1:31:57 to
1:31:59 do I guess what I'm trying
1:32:05 to what I'm trying to share here
1:32:09 is there's no right way to do this stuff
1:32:12 and you don't even need to have an
1:32:15 idea right I went to this thing that was
1:32:17 raining images I picked three images I
1:32:20 like I took them into chat GPT I said
1:32:23 describe them and so now here's what I'm
1:32:25 going to do I'm going to flip over to
1:32:26 Canvas and I'm going to
1:32:29 say in a
1:32:40 canvas
1:32:41 um yeah maybe not in a canvas let's
1:32:44 let's dig a little bit deeper before we
1:32:45 go to a canvas I want to do I want to do
1:32:47 something specific with the canvas so
1:32:49 we're just going to do this in regular
1:32:50 GPT 4 all right um
1:33:00 I want you to analyze the three images
1:33:10 together and the creative elements and
1:33:17 Aesthetics and I want you
1:33:21 to come up with 10 different possible
1:33:26 story
1:33:28 Concepts that could
1:33:30 incorporate those three
1:33:34 images and what they
1:33:40 represent now that's sort of a like
1:33:43 that's sort of a prompty engineering
1:33:44 thing but it really is just you know
1:33:48 like it's just KN like I know I can ask
1:33:50 it to just give me 10 ideas that's
1:33:53 basically I've uploaded three images
1:33:55 you've described them now come up with
1:33:57 10 story ideas
1:34:00 right and then it does its
1:34:04 thing so so let's see urban urban
1:34:07 isolation and Chaos image one human
1:34:10 identity and fragmentation image two so
1:34:13 let's go look at these images and see if
1:34:15 we
1:34:16 agree let's see here they are
1:34:22 um so that's image
1:34:26 one and so it described that as uh hang
1:34:30 on Urban isolation and Chaos yeah that's
1:34:34 good no people very chaotic second one
1:34:38 human identity and
1:34:40 fragmentation so
1:34:43 that's I think that
1:34:49 one vibrant layered portrait of humanity
1:34:52 as as a mosaic of culture sure yep I can
1:34:56 see that and then posthuman Evolution
1:34:59 nice so that's this one postum Evolution
1:35:03 all right here's here's where we were
1:35:06 optimistic here's where we've been
1:35:08 through it a bit
1:35:10 okay these Aesthetics suggest themes of
1:35:14 dystopia identity human machine
1:35:15 integration and the fragility of social
1:35:18 constructs 10 story Concepts the
1:35:21 patchwork profit in aopi
1:35:24 Mega City a mosaic Oracle image 2
1:35:28 emerges a human machine hybrid capable
1:35:31 of decoding the fractured psyche of
1:35:33 society she she's hunted by the
1:35:36 corporate overlords that's pretty good
1:35:38 Rus and Ry fragments of a forgotten City
1:35:42 the architect's Abyss so let's just take
1:35:46 one of these and
1:35:51 say the broken circuit so so the broken
1:35:54 circuit is number six I'm not really
1:35:56 reading these and let's do number one so
1:35:59 let's
1:36:02 say I like number one and number six
1:36:10 period I'd like you to write a creative
1:36:19 brief that combines these two concepts
1:36:24 and completely fleshes
1:36:29 out the story The
1:36:32 aesthetic the characters the intended
1:36:37 audience and anything else that you
1:36:39 think should be in a creative brief
1:36:42 period I want to give this
1:36:46 to lots of different kinds of creators
1:36:49 from musicians to
1:36:51 filmmakers to poets
1:36:55 to business
1:36:56 people to
1:36:59 Educators period the
1:37:02 document should be easily
1:37:06 readable
1:37:12 and make it clear what each of them
1:37:15 could do to bring it to life through
1:37:18 their view of the
1:37:22 world okay
1:37:24 now that was a lot
1:37:34 right I don't know if you got this from
1:37:36 what I just did but the way I'm talking
1:37:38 to this thing is just like a
1:37:41 collaborator like this is how I if I
1:37:43 were sitting down with a brilliant you
1:37:45 know creative director if I were sitting
1:37:47 down with a brilliant writer or or
1:37:49 whatever it might be and I'd say okay
1:37:51 I've got these three pictures I want to
1:37:52 turn them into something
1:37:54 okay and you know they would say well
1:37:56 here's what they here's what I see in
1:37:57 them okay that's great and then I would
1:37:59 just sort of say I'd like to come up
1:38:01 with some Concepts okay here's 10
1:38:03 Concepts so like the interaction here is
1:38:06 not prompt engineering the interaction
1:38:08 here is a [ __ ]
1:38:10 conversation a conversation with
1:38:13 who a conversation with the reflection
1:38:16 of the collective intelligence of of
1:38:20 everyone that's put something on the
1:38:21 Internet in the past 70 years
1:38:25 right including all of the history that
1:38:28 got put on the Internet by those people
1:38:31 so I'm actually interacting with
1:38:32 Humanity so I I'm I'm literally having a
1:38:35 conversation here so now this is
1:38:38 something that I want I ask for a very
1:38:40 specific document I want a creative
1:38:42 brief so now I'm going to switch it to
1:38:43 GPT canvas I'm going to say um put the
1:38:47 brief in a canvas
1:38:57 there it is W there it is W there it
1:39:03 [Laughter]
1:39:07 is I'm a shitty rap T tune from your
1:39:10 [Laughter]
1:39:10 [Music]
1:39:19 childhood all right
1:39:24 oh y'all it's
1:39:27 funny Kyle you should doing you should
1:39:30 consider doing supplementary Tik toks in
1:39:32 which he does prompting tutorials to his
1:39:35 Tik
1:39:36 toks that's good that's
1:39:42 good Kyle Kyle Kyle settle
1:39:46 down okay so so what is canvas doing
1:39:49 okay so canvas is is wild so here's
1:39:51 what's going on with canvas
1:39:54 so on the left hand side here this is my
1:39:56 prompt history right there's our
1:39:58 pictures we put in there's the
1:39:59 description of the pictures this is all
1:40:01 the stuff we've been talking about so
1:40:03 it's maintaining that
1:40:04 chat you notice here that it's got this
1:40:07 little block with a document on it this
1:40:10 link will always link us back to this
1:40:13 document what I can now do is I could
1:40:15 say okay so that created a creative
1:40:17 brief I'm going to say
1:40:20 okay now um
1:40:25 based on that creative
1:40:29 brief I want you to
1:40:31 create a document of four main
1:40:37 characters and six ancillary
1:40:49 characters that includes
1:40:55 their backgrounds what they currently
1:40:59 do and a detailed description of
1:41:03 exactly what they look
1:41:08 like now this should
1:41:12 oops yeah this should give us a new oh
1:41:15 no wait hang on stopped editing
1:41:20 um let me go back here I want to go back
1:41:22 to this thing
1:41:24 uh let's
1:41:26 see where why
1:41:34 what stopped
1:41:36 editing um no make this document in a
1:41:43 new canvas I know I haven't answered
1:41:45 what's a canvas
1:41:47 yet but I
1:41:52 will Patrick characters
1:41:56 okay so there's that now I should be
1:41:59 able to go let's let's let this
1:42:01 build so there's the fourth character
1:42:03 there's the ancillary
1:42:12 characters
1:42:14 okay are we done yeah we're done now I
1:42:17 should be able to go back to this
1:42:18 document
1:42:24 creative brief
1:42:26 um yeah okay so here's the creative
1:42:28 brief
1:42:29 okay so now I have two canvas documents
1:42:33 in this chat I've got the character
1:42:35 document and then I've got the
1:42:38 main um creative
1:42:41 brief what's cool about canvas is I can
1:42:43 edit all this so this says creative
1:42:45 brief the patchwork of rust and revery
1:42:47 so that's the title of it I'm going to I
1:42:50 can I can go up here I can highlight
1:42:51 that text and I can actually talk to
1:42:52 chat GB
1:42:54 in context and
1:42:56 say um this
1:42:59 title doesn't
1:43:03 resonate uh
1:43:05 reso res resonate resonate with
1:43:09 me make
1:43:11 it
1:43:14 suckless right you can you can talk to
1:43:17 these GPT
1:43:18 things any way you want now they might
1:43:22 go no need to use use nasty words uh
1:43:25 let's see on
1:43:28 S okay on s y I don't know what s y is U
1:43:34 fragments of rust and Revelation um
1:43:37 fragments of rust and Revelation so
1:43:41 let's just
1:43:43 go let's just go with r and
1:43:51 r oops oh
1:43:56 wait so fragments of rust and Revelation
1:44:00 I can I don't need to use chat GPT for
1:44:02 that so that's one thing you can do um
1:44:05 the other thing you can do is you can
1:44:07 you can grab whole sections so that
1:44:09 let's see Mosaic City
1:44:11 dystopia here's three characters the the
1:44:15 Oracle rust is a biomechanical assassin
1:44:19 and the city an unnamed sprawling
1:44:22 Metropolis let's say
1:44:31 um um where
1:44:35 why oh something's not working here I
1:44:37 gotta reload oh no it was
1:44:46 [Music]
1:44:50 there 10 story Concepts where's my did I
1:44:53 lose
1:44:57 my so there's that one okay here it is
1:45:00 so here's Patchwork character so if you
1:45:02 have this little arrow here that means
1:45:03 it's a canvas so I can split that back
1:45:05 out I could come in here and I could go
1:45:08 so here's the here's the characters so I
1:45:14 could wait oh these are the characters
1:45:16 that's wrong document we want this
1:45:19 document um so here's these two
1:45:22 characters I can highlight that and now
1:45:24 I can say s chat gbt um add a
1:45:31 character for comic
1:45:37 relief and then what it's going to
1:45:41 do so it's editing it and then it's
1:45:44 going to at some point rewrite this
1:45:47 little
1:45:49 section oh failed to edit it
1:45:57 uh it it kind of lost track of where it
1:46:00 was for some
1:46:05 reason
1:46:07 um add the care the new character below
1:46:15 these
1:46:20 two uh uh uh uh
1:46:30 oh eight okay so that's
1:46:37 bad so if you need to go back a version
1:46:40 you can go back a version so I can
1:46:42 restore this version the other thing let
1:46:44 me show you what else is going on here
1:46:45 because like at this point I'm just
1:46:47 confusing the [ __ ] out of
1:46:48 you um we've got these two documents
1:46:51 this little pencil on the lower lower
1:46:53 right hand corner here is really
1:46:55 powerful I can adjust the length of all
1:46:57 of this um so I can click on this and I
1:47:00 can say you know make this document
1:47:03 longer or shorter or keep it the same um
1:47:07 I can click on this and change the
1:47:09 reading level I could say I want this to
1:47:11 be graduate school level writing or I
1:47:14 want this to be kindergarten writing or
1:47:16 anything in between
1:47:19 um you can add a final polish to it you
1:47:22 can even add emojis to it um I think one
1:47:24 of these is an editing I thought there
1:47:26 was an editing
1:47:28 one um I don't see it now maybe they
1:47:31 changed
1:47:33 that I thought there was one that would
1:47:35 oh this one suggest edits so if I click
1:47:38 on this and hit the little up Arrow it's
1:47:40 now going to go through the
1:47:43 document and it's going to find things
1:47:45 in the document that it doesn't like and
1:47:47 suggest edits for us failed to comment
1:47:50 on it some something's going on with
1:47:52 with canvas tonight
1:47:54 it's a little
1:47:57 weird I tried adding them but it didn't
1:48:00 work we'll try it one more time suggest
1:48:10 edits no it's not it's not doing it all
1:48:13 right
1:48:14 anyway that's what you can do but you
1:48:16 can see we could go on and on and on but
1:48:18 the basic idea was you know I started
1:48:21 with three pictures with with no prompts
1:48:24 I just said describe them and then I
1:48:25 said give me some story ideas and now
1:48:27 give me a creative like you you can
1:48:29 literally just run down Rabbit Hole
1:48:31 after rabbit hole if you want so all
1:48:34 right it's been acting up for two weeks
1:48:35 for me okay yeah they're they're clearly
1:48:38 they've I what I could what I've heard
1:48:40 about the 40 update the the update that
1:48:42 I got so excited about for creative
1:48:44 writing apparently is totally [ __ ] up
1:48:46 coders it's like not doing good coding
1:48:48 anymore um and it yeah it seems to be
1:48:51 not stable right now so anyway all right
1:48:54 that's Chad [ __ ] let me let me get my
1:48:58 head out of the details for a minute and
1:49:01 hop over here and see what comments we
1:49:05 got going joer the mighty on Tik Tok are
1:49:07 you aware of any collaborative tools
1:49:09 either a teacher enters a prompt and the
1:49:12 class gets to interact or ask follow-up
1:49:14 questions in real time no that's a
1:49:16 really good idea
1:49:18 um that's a really good idea all right
1:49:22 oh stream yard was syy got
1:49:25 it um I don't know of any but that would
1:49:27 be that would be a good app to build
1:49:31 quite
1:49:32 frankly um and with coding tools you can
1:49:36 get them built pretty
1:49:39 easily so let's see here what why is
1:49:42 this not
1:49:44 working the hell is going on
1:49:48 here um
1:49:56 I'm trying to fix the Tik Tock
1:50:01 view there you go ticktockers hopefully
1:50:03 that's a little better all right Daisy
1:50:05 may my daughter's class is allowed to
1:50:08 use AI yay for that school with heavy
1:50:11 guard rails for things like asking about
1:50:13 life in the 1920s where AI is heavily
1:50:16 restricted to relevant
1:50:17 information great like awesome I mean
1:50:21 even if even if a school is limiting it
1:50:23 to very limited kind of kinds of things
1:50:25 like that it's better than we're not
1:50:29 letting students use the most profoundly
1:50:32 powerful knowledge tool in the history
1:50:33 of humanity I like the fact that
1:50:36 educational institutions cannot get
1:50:39 their [ __ ] heads around the fact that
1:50:42 everything is about to change they're
1:50:43 the educational
1:50:46 institutions
1:50:49 right here's what AI is like like
1:50:54 Hollywood has kind of [ __ ] this all up
1:50:56 for everyone right because what
1:50:58 Hollywood has done is they've said AI or
1:51:01 robots and the robots are going to kill
1:51:04 us and chat GPT quite
1:51:07 frankly [ __ ] up as well they created
1:51:11 this thing called chat GPT which they
1:51:13 called a
1:51:14 chatbot right so we call these things
1:51:16 chat
1:51:17 Bots and then the first person to use
1:51:20 the Microsoft chatbot Bing chat is Kevin
1:51:24 Roose from The New York Times and the
1:51:26 chatbot tells him that she loves him and
1:51:30 he should leave his wife and that gets
1:51:32 on the cover of the New York Times and
1:51:34 so so it just reinforces that the robots
1:51:37 are out to kill us they're you know
1:51:39 they're all [ __ ]
1:51:41 evil what large language models actually
1:51:44 are is they're technically this really
1:51:47 cold calculator it's just a probability
1:51:51 engine but the way it works is you take
1:51:54 as much information as you can get and
1:51:57 you jam it into this thing called an
1:51:59 embeddings engine that takes all of
1:52:02 those documents and and breaks them
1:52:05 apart into these things called tokens
1:52:08 and it puts all those tokens in Thousand
1:52:10 dimensional mathematical space
1:52:13 semantically clustered whatever the [ __ ]
1:52:16 that means but that's what it's doing
1:52:18 none of the original
1:52:20 documents exist anymore more it like it
1:52:24 it's not like the the uh this PDF of you
1:52:28 know this paper written in a college is
1:52:30 sitting up in the cloud somewhere no
1:52:32 it's taken all of the words broken them
1:52:33 into these little things called tokens
1:52:35 and said where do these how do these
1:52:37 things all relate together with all the
1:52:39 other [ __ ] that's been put in here and
1:52:41 so what it is is they've basically taken
1:52:44 as much of the output of humanity is
1:52:47 they can get their hands on and they
1:52:49 essentially squash it into this little
1:52:51 you know softball side thing that we get
1:52:54 to play with so we get to go hey you
1:52:57 know what do you know about X Y and Z
1:52:59 and it's just in this magical [ __ ]
1:53:01 space I know it's not magical paid it's
1:53:03 mathematic it's mathemagical there's
1:53:05 this mathemagical space that we ask and
1:53:08 it's sort of figures out what we're what
1:53:11 what the answer is and
1:53:15 so
1:53:18 Educators they like
1:53:20 libraries they like books they like
1:53:23 knowledge they like
1:53:25 teaching and here's this grapefruit size
1:53:29 thing that's got all of Humanity's
1:53:31 knowledge in it well as much as we could
1:53:32 get our hands on up to 2024 and it's in
1:53:36 this
1:53:37 ball and and any student could ask it
1:53:40 and get stuff and they're like no no no
1:53:43 we're going to we we people could cheat
1:53:46 with that so so we're not going to let
1:53:48 them use that I just I find it
1:53:51 staggering I find it
1:53:54 staggering anyway enough Soap
1:54:03 [Laughter]
1:54:07 Box all right
1:54:09 listen um it's Friday night if you
1:54:12 didn't know what night it was it's
1:54:13 Friday night um I don't know if this was
1:54:16 valuable tonight I I kind of just jumped
1:54:19 right into tactical stuff before I
1:54:21 didn't do my normal
1:54:23 happy go-lucky thing
1:54:25 so I hope this was valuable
1:54:32 um I'm trying to think so over the
1:54:35 weekend what what I like to do so I take
1:54:37 Saturday and Sunday off what I like to
1:54:40 do is if you want a homework assignment
1:54:42 if you want to play this weekend trying
1:54:44 to think what I want you to do I think
1:54:46 here's two things that you could do one
1:54:50 is um on my Tik Tok Channel you'll see a
1:54:53 a a purple Tik Tock
1:54:57 video of
1:54:59 um the hip hop song This &em Style song
1:55:03 on quantum mechanics and in the comments
1:55:05 of that video you you can listen to the
1:55:07 video you can see the lyrics on screen
1:55:10 in the comments of that video is the
1:55:13 process for how I made that song so one
1:55:16 thing you could do this weekend is play
1:55:19 with song generation generate lyrics and
1:55:22 chat GPT take them over toso or Udo or
1:55:26 some other music generation tool and go
1:55:30 make some songs and just experiment with
1:55:32 that use that process and and use this
1:55:35 sort of playful conversational way of
1:55:38 you like don't take any of this [ __ ]
1:55:40 seriously just go make some songs so
1:55:42 that could be one thing another another
1:55:44 thing you could do this weekend is just
1:55:47 find
1:55:48 any any AI tool go on Twitter and just
1:55:52 type in you know AI tool or just AI or
1:55:57 just go to Google and say what's a good
1:55:59 AI tool I can play with and just go
1:56:03 play all right like like even in this
1:56:06 let's um let me show you one quick thing
1:56:09 before we
1:56:12 go so I had this this uh
1:56:22 the conversation I had about you know
1:56:25 the the stories right and then it kind
1:56:27 of screwed
1:56:28 up but let's
1:56:32 say um let's go find a character
1:56:34 Patchwork characters the Mosaic Oracle
1:56:38 so I'm going to
1:56:39 say based on what you
1:56:43 know show me the
1:56:48 Mosaic
1:56:50 Oracle if you didn't know that chachy
1:56:52 gbt can make images it
1:56:55 can and it uses Dolly 3 and if the the
1:56:59 video that I showed you before about uh
1:57:02 the quantum Cipher um Kelly Bosch that
1:57:06 did the video for that she used do she
1:57:09 used chat GPT just like this to make
1:57:11 those
1:57:12 images um I'm gonna say make this a 16
1:57:16 by9 wide image and make the oracle
1:57:23 older
1:57:26 more rough and
1:57:31 scarier so again just ask it for what
1:57:34 you want you don't like what it gave you
1:57:35 tell it it was
1:57:40 shitty nice there you go look at
1:57:44 that
1:57:46 gorgeous gorgeous I tell you gorgeous
1:57:58 really
1:57:59 cool you want to see something cool yeah
1:58:02 Kyle that's why we're
1:58:03 here go over to a video generation tool
1:58:06 like say Luma
1:58:08 Labs Luma laabs dream machine click on
1:58:11 the little picture
1:58:14 button go to
1:58:17 your go to
1:58:19 your Oracle and say
1:58:24 um
1:58:27 scary
1:58:29 woman
1:58:31 woman's
1:58:33 eyes glow with
1:58:36 green fire as
1:58:40 camera and lumabs has added camera moves
1:58:44 as let's see pushes out as camera pulls
1:58:49 out boom and now we're going to turn
1:58:52 this into a video so this process is how
1:58:55 Kelly Bosch um she used Runway ml to
1:58:59 animate this thing but we're going to
1:59:02 use Luma Labs there's there's a bunch of
1:59:03 them out there that you can
1:59:07 use these are some these are some images
1:59:10 I made this is a project I'm working on
1:59:12 at work where we've got a doctor and a
1:59:14 patient we're trying to bring them to
1:59:18 life that this was a failed thing for
1:59:20 her but let's see this is one we ended
1:59:23 up using this is one we didn't use
1:59:25 because he floats off
1:59:35 screen all right
1:59:37 but there you go her eyes aren't burning
1:59:41 fire
1:59:44 but we went from nothing to
1:59:49 something based on a silly little prompt
1:59:52 right all right let me get out of here
1:59:56 okay um next week is Thanksgiving week
2:00:02 so first of all thank you all for all of
2:00:05 what you do for hanging out for you know
2:00:08 taking the time to be here thank you
2:00:10 mods thank you
2:00:12 Brandon
2:00:15 um so have a good
2:00:20 Thanksgiving trying to think if I'm
2:00:21 going to be here Thanksgiving night I
2:00:23 might be here Thanksgiving night um mods
2:00:26 you don't have to come but
2:00:28 maybe maybe what we do Thanksgiving
2:00:31 night I'll I'll pour myself a nice
2:00:33 blandin I'll just drink some whiskey
2:00:35 we'll just hang out um Thanksgiving
2:00:37 night maybe I can be your excuse to get
2:00:40 away from the family just tell them you
2:00:42 have to take a business meeting you've
2:00:43 been working too hard you just got to
2:00:45 take this one meeting and we'll just
2:00:47 hang out on
2:00:48 Thanksgiving okay but over the weekend I
2:00:50 want you to try a tool that haven't
2:00:52 played with before or even chat GPT if
2:00:55 you haven't dug into what chat GPT can
2:00:57 do do that this weekend
2:01:00 okay
2:01:03 um my deep deep request is save the date
2:01:07 December 3D in the evening 5:00 pm
2:01:10 mountain time so it's 7 PM Eastern 4 pm
2:01:15 Pacific is the two-year anniversary of
2:01:18 the AI Salon okay so if you don't know
2:01:21 about the AI Salon if you came in late
2:01:23 and don't know about it there's the URL
2:01:26 the salon. say join our community and
2:01:29 then I think the third or fourth button
2:01:32 down says RSVP to events click on that
2:01:35 and RSVP for December 3rd we want to get
2:01:37 a lot of people there it's a second year
2:01:39 anniversary would love to have new
2:01:41 people if if there's anyone that's new
2:01:42 here tonight because you saw my hipop
2:01:45 video and you're like who's this fat old
2:01:47 man is he a
2:01:50 producer oh I
2:01:53 produce I produce a lot of [ __ ] uh
2:01:56 but I did produce a pretty sick song
2:01:58 with some pretty sick
2:02:00 beats
2:02:02 um but just keep coming coming back I I
2:02:05 will be here you know all week next week
2:02:07 and we got the anniversary after that so
2:02:10 all right with that thank you all very
2:02:12 much for hanging out this was
2:02:15 fun
2:02:16 and what else what else people what do
2:02:20 we want to say kak
2:02:23 thank you so
2:02:24 much thank you Brandon Emilio's wife
2:02:26 great to see you papa JJ president AI
2:02:29 Call Me Al great to see you call me Al
2:02:32 coming in from Down Under
2:02:34 Apocrypha Apocrypha is that how you say
2:02:37 your name I'm probably mispronouncing
2:02:39 that because I'm not so good with the
2:02:42 words Dr Jay in the
2:02:45 house um deal of the day in the house
2:02:48 Very Cool Jock are the mighty or play
2:02:51 Dungeons and Dragons with chat GPT is
2:02:53 the dungeon master that's really fun
2:02:54 it's really good at that oh and you can
2:02:56 do that with put it into um advanced
2:02:59 voice mode and just have it be a dungeon
2:03:01 master for you I think the other thing
2:03:02 I'm gonna do is just leave this
2:03:05 creepy This creepy woman on screen while
2:03:07 we sign off here because that's just
2:03:16 creepy oh
2:03:18 yeah all right Bob tell him what he's
2:03:21 won he hasn't won anything but that's a
2:03:23 creepy image
2:03:25 dude on Tik Tock Danielle good to see
2:03:28 you Bob donigan Jeff Flanigan oceans Tom
2:03:32 ndla Forever hooked in the
2:03:36 H Lori Blair Silver Fox thank you thank
2:03:39 you Daisy May great to see
2:03:43 [Music]
2:03:48 you all right what do we got going here
2:03:50 all right that's that cool
2:03:53 all right I am out of here peace out
2:03:55 everybody have a fantastic weekend and I
2:03:58 will see you on Monday um 8 o00 mountain
2:04:04 time all right YouTuber