aiLL 10/7/23 - The Democratization of AI: How Generative Tools Are Changing Everything

Description

Chapters

Transcript

0:28 welcome welcome welcome
0:31 I looking a little dim what's up with
0:34 the lighting could be
0:40 my
0:43 maybe lunar stick what's happening what
0:45 shing lunatick how you doing girl what's
0:49 up girl how you doing how you doing no
0:52 no how you doing harlon Hobs Sunseeker
0:57 was shaking
0:58 niow niow
1:01 print hello world hey you got notified I
1:03 went live I'm just going live print
1:05 world so apparently you asking me if I
1:09 was going live let Tik Tok know you
1:12 wanted to know I was going live I the Mr
1:16 the mysteries of Tik
1:19 Tok happy Saturday night
1:22 everybody finally cool welcome the
1:25 T-shirt tonight
1:27 Fender it looks like I've had it forever
1:30 but it's just one of those shitty
1:33 Target
1:35 distressed
1:40 things I sold all my electric guitars
1:43 cuz I realized that playing electric
1:45 guitar was actually a different
1:46 instrument in playing acoustic
1:48 guitar I was waiting all excitedly like
1:50 I used to for SNL back in Phil Harman
1:53 days Fender a good guitar evaporated ink
1:57 hey buddy hey IR regular steo steo oh
2:01 steo that list steo that movie you and
2:04 Johnny Knoxville did awesome it's
2:06 classic classic steo amazing amazing
2:11 you're
2:15 great that one where you got the Staples
2:17 in your ass that was the best amazing
2:21 just like the the the preparation the
2:25 execution um welcome everyone welcome
2:28 happy Saturday night this is the AI
2:29 learning la my name is Kyle Shannon as
2:31 you can tell I'm a bit of a
2:35 cutup some call it idiot um so what we
2:40 do here is we talk about AI we talk
2:42 about AI we have fun doing it it's like
2:44 you know if if you got nothing to do on
2:46 a Saturday night why not hang out with a
2:49 58-year-old waxing poetic about the
2:51 future of AI That's What We Do
2:55 here uh we also call the channel Chad
2:57 add so if you're expecting uh order an
3:02 orderly progression from this point
3:04 forward oh no it's just chaos um but
3:08 what we will do is I'll pop in if you're
3:10 brand new to this AI stuff I'll start uh
3:13 I'll give give you some some good places
3:15 to
3:17 start with that list well not that list
3:20 we'll start with that list you're like
3:22 but K I can't read it we'll fix that
3:25 we'll fix that um also if you have
3:28 questions or comments about AI or want
3:31 to share something you've done cuz a lot
3:34 of people in here what's been cool about
3:36 what's happening in here is I've been
3:38 doing these lives every night now for
3:40 about what three and a half four months
3:42 people Irregulars so the Irregulars are
3:45 the people that are here regularly and
3:47 they're weird but they're not weird
3:49 they're
3:52 irregular
3:53 um what's been starting to happen is
3:56 that a lot of the people in here that
3:58 just are kind of hanging out and
4:00 learning and getting curious and
4:01 starting with like I don't really know
4:03 what I'm doing
4:04 here are getting more educated but
4:06 they're also sharing it and as they're
4:09 sharing it really good stuff's happening
4:11 to them so I can't finally I can't wait
4:14 to finally learn what this whole AI
4:17 thing is about yes sorry that's not this
4:19 channel
4:23 Pate oh my God I know you and me both
4:27 I'm so ready to learn what this whole AI
4:29 thing's about it's so funny I was doing
4:31 a um I was doing a a LinkedIn live on
4:36 Friday
4:38 afternoon and uh and one of the guys on
4:40 there was I showed the the do stuff the
4:44 image stuff and and he was just like
4:46 it's not perfect he was really [ __ ]
4:49 cranky about how it wasn't perfect I'm
4:52 like you just watched the machine
4:54 perform a [ __ ]
4:56 miracle and because it's not exactly the
4:59 style you want
5:01 it's it's useless so we're in this very
5:05 weird stage right now where you know
5:08 people's expectations of this are either
5:10 like the robots are going to kill us
5:13 which they're not yet you know let's
5:16 give them a couple of years to get their
5:17 [ __ ] together um or or these things are
5:21 perfect right it's be because the tools
5:25 because what they do is so
5:28 remarkable it's really easy to fall into
5:30 the Trap that they're perfect and
5:32 they're not and it's also easy to fall
5:34 into the Trap that it's like oh I'll
5:36 never be able to learn that and you can
5:37 I mean it's just this is just about
5:39 playing best
5:40 tool tool to summarized very long text
5:44 please oh oh best tool to summarize very
5:47 long text all right let's do this Chris
5:50 Martin
5:51 777
5:54 um so number three on this list is what
5:58 you're looking for claw.
6:02 a has a 100,000 token context
6:08 limit which if you're sitting there like
6:11 haven't really used these things before
6:12 you're like what the hell does that mean
6:15 here's what it means um if you go to
6:18 cloud. there's a little plus button you
6:20 can upload a document to it or you can
6:22 just copy and paste text right into the
6:24 thing um the 100,000 tokens translates
6:29 to about 75,000 words so to be safe you
6:32 could put in a 50,000-word document and
6:35 start interacting with it you got 25,000
6:37 spare words to play with with prompts
6:40 and responses and things like that it's
6:43 really good at summarization I do this
6:46 all the time um if you go to my YouTube
6:49 channel AI learning lab that is
6:53 um um archives of these lives or AI
6:57 learning la- TT is the you YouTube
6:59 Channel just regular AI learning lab
7:01 somebody had
7:02 it but
7:04 um AI learning laab
7:07 sltt is all of the video descriptions
7:11 and the chapter isation is done with
7:12 Claude so I take like the the
7:14 transcripts from these lives where I'm
7:16 just rambling incessantly for 3 hours
7:19 pop it into Claude and I've written a
7:20 prompt that summarizes it and makes the
7:23 chapters Allin one go um is it perfect
7:27 no it because it's it's wildly in
7:29 consistent sometimes it's [ __ ]
7:31 brilliant sometimes I have to go back
7:32 and forth with it three or four times um
7:35 but but on balance the fact that it can
7:37 read a three-hour transcript and
7:39 summarize it like that like the
7:41 summaries are always you know decent the
7:43 chapter zations where it gets a little
7:45 confused but um but yeah oh wait what
7:48 did you say that you can't you can't
7:51 access from Quebec Canada oh that sucks
7:53 well so try a uh try a VPN
7:58 Chris can if you can access chat
8:02 GPT you can also use uh Advanced Data
8:06 analysis so here let me show you that
8:09 quick let me pop over
8:11 here are these the right glasses oh they
8:13 are they're just dirty looks like the my
8:16 dog took a dump on the
8:18 lenses how do how do glasses get so
8:21 dirty I know I'm not touching them I
8:23 touched my screen a lot I do not touch
8:25 my glasses in a lot what the [ __ ]
8:27 happens are there like glasses elves
8:30 that just come in
8:32 like [ __ ] your
8:34 lenses
8:38 um so if we're at chat
8:46 GPT um go to so so you need to have chat
8:50 GPT plus and then you're going to want
8:53 to go to Advanced Data
8:56 analysis which is they used to have a
8:58 name it used to be called Cod
8:59 interpreter which was a really bad name
9:01 and now they've named it like what it
9:03 does but it just sounds like a boring
9:06 feature but it's not boring it's pretty
9:09 amazing so see that little plus symbol
9:13 there now now you can upload a document
9:15 so you could upload your big long
9:17 document into this and then start start
9:19 interacting with it all right this
9:23 should be able to do it I I think you
9:24 have access to chat GPT in Canada but I
9:28 do not know I do not no no no no no um
9:33 all right so that answered that for for
9:38 Yan Chris Martin
9:40 777 if you have not tried chat gbt yet
9:44 go to the top one there chat. open.com
9:47 just just you don't have to pay for it
9:49 just use the free version start playing
9:50 with it do it now don't don't be like
9:54 yeah I guess do it tomorrow it's
9:56 Saturday night I don't want to do any
9:57 work tonight I don't want to learn
9:59 anything no no learn it tonight there's
10:01 nothing to learn there's nothing to
10:03 learn except how to use your
10:08 words you know when you when you're
10:10 thinking stuff and people like use your
10:12 words just go there and start using your
10:13 words ask it to write you a poem ask it
10:15 to write you a country song ask it to
10:17 write you a project plan or a computer
10:20 program go ask it to ask write you a c
10:24 program that's a clone of Flappy Birds
10:27 and watch what it
10:28 does
10:30 play with it long enough until you have
10:32 what I call your Kevin mallister
10:35 moment but you
10:38 will and then at that point then you're
10:40 ready to start just get curious about it
10:42 what you'll find out about these tools
10:44 is
10:45 that they're not as intimidating as they
10:48 seem you don't need to have a computer
10:50 science degree you don't need to be a
10:51 mathematician or a researcher or a
10:53 programmer so so they're not as
10:56 difficult as they might seem from the
10:58 outside and yet yet they're
11:01 significantly more profound than you
11:04 think they might be it's it's a it's a
11:07 really cool kind of
11:10 uh it's I don't know it it's it's a
11:13 remarkable thing a a remarkable thing
11:16 that are
11:17 remarkable
11:20 um so what else um b. google.com that's
11:24 that's Google's entry into this it it's
11:27 not bad it doesn't deserve to be in
11:29 second place but they're about to
11:30 announce a thing called Gemini that I
11:33 think is going to put this in second
11:35 place and according
11:37 to them this is it's going to be five
11:40 times better than gp4 I I don't buy it
11:44 Pate I love you brother but I don't buy
11:47 it's five times better than gb4 hey if
11:49 it is awesome it'll go right to the Top
11:52 If it's as good as gp4 it's still
11:55 remarkable so so um Bard has gotten much
11:58 much better over the past three months
12:01 and I think it's going to get
12:02 dramatically better like I said Claude
12:04 can do big things Bing is Microsoft's
12:07 version of chat GPT it's connected to
12:09 the internet it can make images using
12:11 dolly3 which I'll show in a
12:14 second um that's the adults table here's
12:17 the Kids Table you got Pi which Pi is
12:20 like a cute little we call it chatty GPT
12:23 in here chatty GPT is uh you put it into
12:28 conversation mode so go get the IOS app
12:31 the iOS version of Pi company's called
12:34 inflection and then put it into talk
12:37 mode there's a little green phone icon
12:39 and just start talking to it it's quite
12:40 remarkable perplexity doai really good
12:43 tool labs. perplexity let you play with
12:46 Meadow llama 2 models Winston are you
12:48 here Silver Fox thank you for the Cru
12:53 very nice very nice indeed oh Silver Fox
12:57 reached member level 16 M she as
13:01 sends yes the AI learning lab is a
13:04 competition there are winners they're
13:07 losers we don't like the
13:09 [Music]
13:11 losers but Kyle you can't say winners
13:14 and losers it's 2023 we'll get triggered
13:17 no no no sorry I grew up in a time
13:21 where the trophy meant you
13:24 won everyone else went home
13:27 depressed you can't go home from a
13:29 competition where most of the people are
13:32 not depressed it's not a
13:37 competition anyway U po.com is also good
13:40 to check out okay you want to make
13:41 pretty pitches here's the things there
13:44 feel free to screenshot these if you
13:45 want um bing.com create is now dolly3 um
13:50 I've got access to dolly3 inside chat
13:52 GPT which I'll show you in a second it's
13:55 mindboggling I don't well is it
13:58 mindboggling
14:00 I'm I'm I'm tempering my thought on it
14:03 because the dude in my LinkedIn live was
14:05 like it's not perfect none of these
14:07 tools are perfect in fact we kind of
14:09 have a a
14:11 running
14:14 visual gag in here
14:18 that all most of these images were made
14:21 in uh
14:23 ideogram where is it there it is janky
14:27 pieces of [ __ ] all of this AI stuff
14:29 right now lives in the in the level
14:33 of don't expect this software to do all
14:36 the stuff you want it to do cuz cuz
14:39 you'll start playing with it and you'll
14:40 be like oh well it should do this it
14:42 won't do that and you should be like
14:45 well it should be able to spell better
14:46 and it no and then and then you're like
14:49 well it shouldn't make mistakes I yeah I
14:51 get it but it's just the way it is right
14:53 now
14:55 um so so we'll go play with that mid
14:59 Journey if you want really high quality
15:01 art is still the best although Dolly 3
15:03 is I have a feeling they're going to
15:05 give them a run for their money pretty
15:07 soon and and the the convenience of
15:09 making art inside of a chat GPT
15:12 conversation for me for for the bulk of
15:16 what I
15:17 create visually I don't need to go to
15:20 Mid Journey
15:22 anymore and it's weird cuz before I
15:24 thought if I need something good I'll go
15:25 to Mid journey and now it's
15:27 like I just need something good enough
15:30 and I want to put text on it and I want
15:32 to do it within the context of a
15:34 conversation I don't I don't even think
15:36 about it anymore so so this mid Journey
15:39 for me has dropped to like 15 20% of my
15:43 consideration if I want to make an image
15:45 ideogram I was using all the time and
15:47 now I'm not using at all um Adobe but
15:50 but it does text so um if you don't have
15:53 access to Dolly 3 and you don't want to
15:55 play play with it at Bing or you can't
15:58 ideogram is really good uh the Adobe
16:01 products Firefly and Adobe Express and
16:04 Photoshop gener to fill are good and
16:07 getting better they seem to be
16:08 aggressively leaning into this
16:09 generative AI stuff they're doing it in
16:11 a way that's safe for artists which if
16:13 you work for a big company or that's
16:15 important to you pay attention there um
16:18 stable diffusion is not on this list
16:20 just cuz it's it's it requires you to be
16:22 a bit more geeky you can play around
16:25 with a site called uh
16:28 leonardo.da it's a nice web front end on
16:30 top of stable diffusion so that's good
16:33 I've actually canceled my subscription a
16:34 mid Journey yeah interesting steo I
16:36 might do the same I I really might
16:38 because it's
16:40 like if mid Journey hadn't been so like
16:43 the CEO there's just kind of this
16:45 pompous you know nft bro type
16:48 personality dude and he and it like I
16:50 think he's like well mid journey is cool
16:52 it's like like yeah but it's you're not
16:55 going to get to the masses in mid
16:56 Journey like I don't know mid Journey's
16:59 mid journey is not great I mean
17:01 discord's not great the fact that you
17:02 have to do [ __ ] inside Discord is a bit
17:04 of a drag so anyway yeah I might I might
17:06 cancel my subscription there as well um
17:10 these are video things so Runway ml will
17:12 do text to video they actually just cut
17:15 a deal with canva so you can now
17:17 generate text to video inside canva so
17:21 canva by the way canva should probably
17:23 be on one of these lists um because they
17:26 just added magic canva or whatever the
17:28 hell it is where they're rolling ai ai
17:31 out across all of that you you
17:35 take big um sort of proprietary tools
17:39 like canva and you you know pour AI
17:42 gasoline on them
17:44 aggressively I think that they're going
17:46 to I I think canvas's got a really good
17:48 sh shot at at you know capturing a lot
17:51 of AI market share because they're
17:53 they're just going to allow you to play
17:55 with all these different models in an
17:57 interface you already know with the
17:59 ability to add typography and do
18:01 different you know social formats and
18:03 things like
18:04 that and what they will likely do is
18:07 start creating their own models so they
18:09 won't even be using other people's
18:11 models so canvas canvas is one to watch
18:14 um Heen does um video translation into
18:19 multiple languages where it actually
18:20 syns your lips up it's quite remarkable
18:23 blockade Labs does text to 3D worlds
18:26 it's kind of a hack but it's it's
18:28 interesting and then uh D- ID is kind of
18:31 like KY Jen uh in terms of you can do
18:34 text to animation where you write a
18:36 story pick a voice and then it'll make a
18:39 a picture
18:41 talk so those are those and then here's
18:44 some
18:45 resources so there's some some things
18:48 you should look at futur
18:51 pedia.com are directories of AI
18:55 tools Ethan mullock David Shapiro and
18:58 Paul R rer are all kind of badass people
19:02 you should be paying attention
19:04 to um also on Tik Tock there's a woman
19:08 named Rachel Woods who's got the AI
19:10 exchange she's worth paying attention to
19:13 I also like McKenzie Bose on Tik Tok
19:15 although he hasn't been posting as much
19:17 lately but he created a tool called GPT
19:19 boss he's an interesting
19:24 dude all right alrighty all right all
19:28 right all
19:30 right ticktock the ticktock
19:34 turtle private pepper nice hair on that
19:41 turtle look look people are picking on
19:45 me wait let me flip
19:50 [Laughter]
19:53 this I resemble that remark I think the
19:56 other one has glasses yeah
20:00 they're picking up my hair look at that
20:03 that's not right that's not right that's
20:06 IR
20:11 regular oh
20:13 man okay what else
20:17 um if you listen if you're one of those
20:21 people so so Ethan mullik one of the
20:23 guys I told you to follow he created a
20:25 he wrote an article a few months back
20:26 like six months back
20:29 um called detecting the secret cyborgs
20:31 and the secret cyborgs were people that
20:33 were getting secretly AI literate within
20:36 companies and the companies were
20:38 remaining
20:39 clueless and
20:42 um if you're one of those if you're a
20:44 secret cyborg where you're like curious
20:46 about this [ __ ] and you're like I don't
20:47 have anyone to talk to this about
20:49 everyone thinks I'm weird there people
20:51 are angry about AI what do I do every
20:53 time I talk about it they're like AI is
20:55 going to kill us what are you doing what
20:56 are you doing with your life um come
20:58 hang out here I created a community
21:01 called the AI Salon last December
21:04 because I had a feeling there were going
21:05 to be a bunch of us weirdos trying to
21:07 figure this [ __ ] out and the other
21:08 people are like what is that robots are
21:11 going to kill us so we meet every other
21:15 Tuesday um the next meeting is this
21:17 coming Tuesday um we have speakers we
21:20 have guilds now we just added guilds we
21:22 have a Discord um Channel where you can
21:26 hang out and ask questions and things
21:27 like that as much as I hate Discord it's
21:29 it's a decent Community platform so um
21:32 if you go to the AI Salon or the salon.
21:35 there's a it's a link tree with a link
21:37 to our meetups a link to a mission
21:39 statement and values and a link to the
21:43 Discord and there's other [ __ ] in there
21:45 but those are the big three first three
21:47 links all right so that's that let me
21:51 show you one other thing if you're in
21:54 Colorado where is hang on a second l in
21:59 in
22:09 no where's my mouse is it working where
22:13 is
22:14 it come on come on there we
22:22 go so next Saturday October 14th um I'm
22:28 speaking at a thing called AI Peaks
22:31 which is Colorado's inaugural AI
22:35 conference and so I'm one of the keynote
22:38 speakers there and I'm talking about
22:42 um um why why I think um 2023 is the new
22:47 1995 that that you know this uh that
22:51 what's going on with this here AI stuff
22:53 is actually a really big deal so if you
22:55 go to AIP peak.org
22:58 um and you're in Colorado it's free
23:00 tickets are free although I think
23:01 they're limited and it's if if not sold
23:04 out it's getting close so um go sign up
23:08 and uh look stock photography cuz it's
23:12 it's the first inaugural how could we
23:14 how could we have a picture see I think
23:17 that must be that must be mid
23:21 Journey but anyway there's some of the
23:23 people are speaking Laura rich is
23:25 speaking she's great
23:27 so so so that's that and let's see what
23:31 else we want what else we going to do oh
23:33 you know what I said I said I would show
23:35 you you want to see some cool and groovy
23:37 [ __ ] your keynote is the same day as the
23:41 solar eclipse well there you have it
23:45 we're we're entering a new
23:48 realm the sun shall disappear and then
23:51 it shall reemerge and we will be be in a
23:54 new era and together we
23:59 you can make money
24:01 with that's what everybody wants isn't
24:04 it I just want to make money how do I
24:06 make money with this ky ky can you tell
24:08 me how to make money ky ky ky ky ky ky
24:11 ky listen how do you make money with ch
24:13 P KY C can you help c c c c c c
24:17 c um you can make money with it but I'm
24:20 not going to give you the answer you
24:21 want to hear so we'll talk about that
24:23 later but you want to see some cool
24:25 [ __ ] so here we are at ch GPT I am the
24:29 lucky um holder of access to Dolly 3
24:33 within chat GPT which is super
24:37 cool
24:39 um and I'll show you how it works
24:43 so the the the process is very different
24:47 than well you can you can kind of treat
24:50 it like mid
24:52 Journey but what it does is really
24:54 different so let me just I I'll do a
24:56 prompt treating it kind of like like mid
24:58 journey and then and then we'll sort of
25:00 Watch what it does so I'm going to give
25:02 it a really shitty
25:07 prompt oh no Siri are you did you take
25:10 over my microphone Siri she gets jealous
25:13 yep there she is see I knew it
25:18 [ __ ] Siri go
25:21 away
25:27 okay
25:30 come on oh my god do I actually have to
25:32 type
25:34 please there we go
25:40 hello create an image of a 1970s muscle
25:44 car in an abandoned
25:46 Garage in
25:50 landscape orientation
25:53 period the car should be
25:56 pristine and shiny like it's new off the
25:59 showroom floor period the
26:04 factory should
26:06 be grungy and
26:12 dilapidated U let's see
26:20 Factory sunlight should stream
26:23 through the broken
26:25 windows creating dramatic shadow
26:28 and a compelling
26:30 picture so so
26:35 already even though I'm just sort of
26:37 jumping right in and saying make me an
26:39 image
26:40 already the way you can talk to chat GPT
26:44 is you can just describe stuff where I
26:46 wouldn't write this the same way with
26:48 mid Journey right I would write it more
26:50 like like a series of kind of Google
26:53 prompts where where where you've got I
26:55 don't know you've just got
26:57 more you're more prescriptive here I can
26:59 be just more like I'm talking like I'm
27:02 talking to an art director so then you
27:06 hit return and it fires off the plug
27:09 effectively the plug-in the dolly 3
27:11 plug-in and then it writes its own
27:13 prompts so what it's doing right now is
27:15 it's taking what I wrote and it's
27:17 prompting itself so photo of a pristine
27:20 and shiny 1970s muscle car parked inside
27:23 an old grungy and dilapidated Factory
27:26 the contrast between the cars gleaming
27:28 appearance and the factories rundown
27:29 aesthetic is Stark sunlight streams in
27:32 through the broken window casting a
27:34 dramatic Shadow across the scene
27:36 highlighting the J to position cool um
27:40 glossy 70s muscle car looking brand new
27:42 position imin amidst the ruins of a
27:46 decayed Factory so it's changing up the
27:48 prompts for each one of the four images
27:51 and like sometimes it'll just it'll do
27:53 like four of the same subject but it'll
27:55 change out the art
27:57 style oh these are pretty
28:00 good look at that that's pretty nice so
28:05 it made them silver which is a really
28:07 interesting choice silver with a black
28:09 stripe because
28:10 that does not look like a 70s muscle car
28:15 I mean there were
28:17 some that's a
28:19 challenger that's a
28:22 chevel so these are actually pretty
28:24 close to actual
28:26 cars that's sort of a hybrid between
28:29 some sort of Pontiac and some sort of
28:30 mopar thing um so they're not bad um but
28:35 then I can go
28:37 um I can just talk to it like like I'm
28:40 an art D like it's an art director and
28:42 I'm a creative
28:44 director that's
28:46 great but I'd like the cars to have more
28:50 color can you do four more where two of
28:52 the cars are green and two of the cars
28:55 are orange period for the orange cars
28:59 see if you can make them a 1969 GTO
29:03 [Music]
29:06 Judge and and and like it's just going
29:09 to like take this creative Direction and
29:12 go off and [ __ ] do it I mean it's
29:15 just like it's just mindblowing to me
29:19 janky oh nice one lunatick I like that
29:22 that's
29:24 good oh the the DJ the DJ Ninja Turtle
29:28 looking good y all are
29:31 [Laughter]
29:33 nuts all right we got some new folks
29:36 Maho the dragon and zimer joined the AI
29:39 Salon
29:41 welcome welcome welcome
29:44 welcome you can make money with
29:46 Chach what's going on here is it oh it's
29:49 just slow so so this is also one of the
29:52 things I've noticed in the last week GPT
29:55 4 has gotten incredibly slow it's it's
29:57 just gotten really really slow because
30:01 they just launched a bunch of new
30:03 features right it it can read images it
30:07 can make images it can um it's now
30:10 connected to the internet again so a lot
30:13 of the people that might not have been
30:14 using chat GPT they maybe went over to
30:16 Bing or they went to Bard because it
30:18 wasn't connected to the
30:20 internet um are are
30:23 back so there's green that's not the 70s
30:26 green I was look that's 7s green but it
30:28 wasn't the lime green I was thinking of
30:30 but it's cool that's pretty G GTO Judge
30:32 like that's pretty [ __ ] spoton
30:35 doesn't have the
30:37 graphics yeah that's pretty tight that's
30:39 actually that's that's at least in the
30:43 neighborhood at least in the
30:46 neighborhood so all right so so now
30:49 we're going to do something
30:51 fun
30:53 um how are we going to do
30:56 this okay
31:04 okay so one of the other things that
31:07 dolly3 does really well is I don't know
31:10 of a better term for this but uh like
31:13 prompt coherence or prompt
31:15 accuracy so when you're working in mid
31:19 Journey if you say like I want you know
31:22 this kind of person that kind of person
31:24 and that kind of person it'll sometimes
31:25 blend them together or if you want this
31:29 kind of object in the foreground and
31:30 this kind of object in the background it
31:32 doesn't it just doesn't do multiple
31:35 objects well so you have to get clever
31:37 about how you add objects into mid
31:42 Journey
31:44 um this is really good so we're going to
31:46 do this we're going to do
31:56 um
31:58 I want to try an experiment I'd like to
32:05 create a circle of objects surrounding
32:10 each
32:13 car and the objects should be
32:19 components of an engine period there
32:24 should be cylinders spark plugs
32:33 wires
32:38 Wheels rear view
32:43 mirror cylinders oh I said that
32:54 [ __ ] a cam
32:56 shaft
32:59 transmission
33:02 gears and a
33:04 muffler
33:06 period they should be arranged in a
33:09 symmetrical circle around the
33:12 car floating in
33:15 space as if by
33:18 Magic
33:20 period additionally I'd like to have the
33:26 words muscle cars
33:32 Rock incorporated into the
33:36 scene so let me pop that
33:41 into I'm going to put muscle cars Rock
33:44 in quote
33:45 marks and let's see how it does this
33:48 this might fail miserably
33:51 but we should we should be able to see a
33:54 fair amount of the objects that I just
33:56 mentioned
34:00 let's see how it wrote The Prompt I
34:02 always find this
34:04 fascinating photo of a pristine 1970s
34:07 green muscle car in the center with
34:09 engine components like cylinders spark
34:11 plugs wires Wheels rear viiew mirror cam
34:14 shaft transmission gears and Muffler
34:17 simult symmetrically arranged in a
34:19 circle around it these components appear
34:23 to float Defying Gravity the words
34:25 muscles cars Rock wow there they are
34:27 already okay so it went to it it forgot
34:30 to include the abandoned
34:33 Factory um but actually what that came
34:36 up with is kind of a cool happy
34:39 accident that's kind of like one of
34:41 those they call these things noing
34:43 noling photographs k n o l l i n g where
34:47 you got lots of different
34:49 objects did
34:53 it control
34:56 wa
35:01 um copy image Link
35:10 new where why did that not go
35:13 there oh it's downloading it God yeah RI
35:18 them FR them Rock them sock
35:26 them
35:37 I'm trying to see if it
35:39 um if the words are
35:45 right not
35:48 really but there's Wheels there's an
35:51 engine there's gears there's a
35:55 muffler there's a spring so I kind of
35:58 got
36:00 it um that one's really cool look at
36:04 that muscle cars are this is really good
36:06 actually I'm Gonna Save this one I like
36:08 this one this is
36:10 cool see I would actually go I would go
36:13 Photoshop in at this point I go
36:15 Photoshop in some judge graphics on this
36:17 bad boy that's a cool
36:19 [Laughter]
36:23 t-shirt and this this is one of the
36:25 things about the the kind of serendipity
36:28 let's see muscle cars Rock um the
36:33 Serendipity of these tools right now um
36:37 because they're not perfect especially
36:39 when you're doing something like this
36:40 when we're just kind of experimenting
36:43 and playing around um you can you can
36:47 stumble upon you can stumble upon really
36:49 amazing things this is also one of the
36:52 things that I talk about a lot is is
36:54 this idea of what I call lateral play so
36:57 so play around with these tools even if
36:59 it's not what you do um or or are
37:03 interested in because you will stumble
37:06 upon things that can do that you're like
37:08 oh that's pretty something I
37:11 should I should
37:14 go do something like that like this is a
37:17 really good example of
37:20 that did that add that
37:25 no that's going going in the that's
37:28 going in the permanent archive people
37:31 look how cool that
37:33 is muscle cars
37:36 Rock everything's orange and orange and
37:40 silver come on man are you kidding
37:44 me it's a 69 GTO Judge this Line's a
37:48 little low but it's close and that that
37:50 door seems really bad
37:53 but damn that's
37:55 cool um um okay back to this
37:59 thing so I'm going to show you one other
38:02 thing to do here oh let's go look at the
38:04 other two images that one's pretty
38:08 slick that one is too I mean like these
38:12 are these are all really interesting
38:14 kind of uh kind of Graphics it's not
38:17 it's not exactly what I gave it let me
38:19 give it let me give it one other
38:24 adjustment while I like this looking
38:27 feel it seems like we forgot to have the
38:30 car sitting in the abandoned Factory
38:34 period can you combine the first
38:36 Concepts we played with with this new
38:39 floating objects concept into a new
38:42 series of four cars two green and two
38:45 orange period also make the green more
38:48 of a lime green than that 70s muted
38:54 green so like yeah I
38:58 I I because I'm not really necessarily
39:01 seeing your comments in real time right
39:05 now this is mindboggling to me that I
39:08 can that I can essentially have a
39:11 creative director
39:14 conversation and this thing's just doing
39:18 it and it's doing it quickly right it's
39:20 doing it quickly enough that it's
39:22 tolerable to do it on something like
39:25 this um
39:27 I don't know anyone y'all y'all like is
39:32 this this is crazy isn't it is it is it
39:35 me pop in the comments down below like
39:38 what what are your thoughts as as you've
39:39 been watching this I'm just curious
39:43 you're being way too nice uh maybe what
39:47 am I being nice to I'm being nice
39:49 because it gave me a shitty
39:51 scene what was that Ethan guy's last
39:53 name Ethan mullik is is the guy's name
39:57 um his um his blog is called one useful
40:00 thing.
40:02 org Ethan mik so let me let me describe
40:04 these three these three dudes so Ethan
40:06 mik is a Wharton business school
40:09 Professor he's been kind of deep diving
40:12 on this generative AI stuff for a couple
40:15 of years now he does it really
40:17 methodically they're doing a lot of
40:19 studies he's really innovating in
40:22 education he's you know he talks a lot
40:24 about um I just quoted him in a Lin post
40:27 the other day that said um chat GPT
40:30 detectors do not work don't use them um
40:35 um really interesting guy really
40:37 interesting work David
40:39 Shapiro um genius has written five books
40:42 on artificial general intelligence he's
40:45 got a cognitive architecture open-
40:47 source project right now that's
40:49 fascinating and he makes amazing YouTube
40:51 videos and he does like two or three a
40:54 week that are like they they're like
40:57 they're like high enough quality and and
40:59 and like well produced enough with you
41:01 know PowerPoint presentations and [ __ ]
41:03 like that if he did one of them a month
41:06 I'd be impressed and he's doing like two
41:08 or three a week and then Paul rater the
41:10 marketing AI Institute was started like
41:12 5 years ago so he's been looking at this
41:15 [ __ ] long before chat GPT came along and
41:18 then when chat GPT came along he's like
41:20 holy [ __ ] this is why we're here these
41:23 tools are evolving nicely we need to
41:24 invent some new terms for janky we can
41:26 do that motherly advice crappy uh
41:30 incomplete oh that one's that's a cool
41:33 image that's
41:36 cool
41:38 um uh and I don't think janky has an e
41:42 in it does it wait we got to ask chat
41:45 GPT this oh look these are cool we'll go
41:48 look at those in a second but let me ask
41:50 uh let me do uh yeah I can just do it in
41:55 here quick question does the word janky
41:59 have an e in
42:04 it no the word janky does not have an e
42:07 in
42:10 it it's slang often used to describe
42:13 something that is of poor quality
42:14 unreliable or awkward that is the
42:17 perfect description of all these AI
42:19 tools right now um and yet as much as I
42:22 make fun of them and call them janky
42:24 don't get me wrong I wouldn't be
42:26 spending 3 hours a night on [ __ ] Tik
42:28 Tok doing these lives if I didn't think
42:31 this was the most profoundly powerful
42:33 technology in the history of mankind so
42:36 those two things can coexist in the same
42:39 space we're very very early we're very
42:41 early we're very
42:44 early and that's why I point that out
42:49 because couple of things we've been
42:52 hearing about AI for decades in the
42:53 movies ai's been been around for decades
42:57 decades what's different is what
42:59 happened on November 30th last year
43:02 we're coming up on the one-year
43:03 anniversary of chat GPT changed
43:05 everything it made these powerful
43:08 machine learning tools available to the
43:11 rest of us basically and that changes
43:13 everything so there was a similar
43:16 transition that happened with the
43:18 internet and the worldwide web made the
43:20 internet accessible by
43:22 everyone this is the same kind of thing
43:26 so um
43:29 yeah but yeah we can find new words for
43:31 janky it's just janky piece of [ __ ] is
43:33 funny look at that that's cool it's not
43:35 a circle around it but that's pretty
43:38 slick I mean you know think about this
43:40 if you if you were just wanting to do
43:42 like a a Nostalgia piece on 70s muscle
43:45 cars you know for a magazine or for a
43:48 blog post or something like that and you
43:50 wanted to talk about you know car
43:53 culture and tinkering and you know
43:55 blowing the doors off these things I
43:58 mean this is just
44:02 staggering like even to conceive that
44:05 like I I wouldn't look at look at what
44:06 it did it put muscle cars Rock up in the
44:10 um up in the ey beams in the roof of the
44:16 factory it's floating like a god like it
44:19 deserves to be janky piece of [ __ ] kept
44:22 being blocked on Dolly 3 oh did it oh
44:24 that's really
44:25 funny well you know what you could do
44:27 lunar stick is janky poop emoji it'll
44:30 probably do a poop
44:33 emoji that's
44:35 [Laughter]
44:37 hilarious um a that's just impressive
44:41 right
44:43 there I mean that looks pretty accurate
44:46 to a 69
44:50 GTO that's super
44:52 cool I don't know what these weird
44:54 things are they're sort of like super
44:56 wide wheels but they're
44:58 not um anyway that's Dolly 3 inside chat
45:02 GPT it's it's quite remarkable the the
45:04 other thing that I've done with with
45:06 this is instead of just jumping right
45:09 into image generation start talking
45:13 about
45:14 the the the story first so one thing I
45:18 did was I took
45:21 the description of the AI salon and our
45:25 values I think I might still have it
45:29 here because I have prompt history don't
45:31 I
45:34 um let me see if I can find
45:45 [Music]
45:55 it
45:57 did I miss
46:12 [Music]
46:16 [Applause]
46:22 it I do not know let's see the sunset
46:25 lady I don't like that this is so easy
46:27 wait uh pop that back up
46:30 again I don't like that this is so
46:36 easy whoever is modding and pinned that
46:39 could you repin that I just missed
46:42 it I don't like this so easy it takes
46:45 the joy away from humans so Sunset lady
46:48 really interesting point
46:52 and I've got I've got thoughts on it as
46:55 you can imagine in
46:59 um so for me this is a different kind of
47:03 Joy the instant gratification and
47:10 the I'm generating something here I
47:13 would not have generated otherwise right
47:17 so for me this is an invention tool as
47:19 much as a
47:23 um as much as anything else
47:28 and let's see hang on a sec I want to
47:30 copy this one
47:36 too I'll put that there because that's
47:38 cool muscle cars
47:40 Rock
47:44 [Music]
47:47 um
47:48 so so I don't disagree with you that
47:51 there is a there is an
47:54 unsettling nature of this
47:58 that and and that's why I was saying
48:00 like I can't I can't quite wrap my head
48:04 around what's happening right now right
48:08 because this
48:09 is let's use this image generation just
48:12 as a metaphor
48:14 okay the fact that it can do this level
48:17 of illustration or photoshopping or
48:19 whatever you want to call it image
48:21 making this quickly and this
48:24 well is going to happen for literally
48:27 every piece of knowledge work every
48:29 piece of data entry anywhere where
48:32 computers touch human
48:34 beings are going to be impacted in this
48:37 kind of instant execution kind of
48:42 way and so a couple of things happen for
48:45 me
48:48 one I think as individual contributors
48:51 we all need to figure out where the
48:53 boundary is where we're comfortable
48:56 using tools like this or not and where
48:58 we get our joy so for me I get joy out
49:01 of geeking out on this stuff and
49:02 learning about it
49:05 right Paul rer one of the guy the guy
49:07 from the marketing AI Institute he talks
49:09 a lot about writing and he
49:13 says one of the roles of writing in his
49:16 life is that's how he learns how he
49:18 internalizes a concept as he writes
49:20 about it and so he said he will never
49:23 use AI to write his thinky pie pieces
49:27 but he uses aii all the time to write
49:30 like a little blog post or a little you
49:33 know social media thing or to process an
49:35 email or to write an email or to
49:38 transcribe a
49:40 meeting and so he's figured out in his
49:43 life that there's sort of business
49:44 process and then there's creative output
49:47 and for him those are two very different
49:51 modes and one is cool for AI and one's
49:55 not
49:58 I
49:59 think anyone that gets joy out of the
50:04 doing is not going to enjoy these tools
50:07 and there's there's a really good piece
50:09 if you want to read a good piece look up
50:12 um I don't exactly know how to find it
50:14 but it it was written about a lot there
50:16 was a post on Reddit of this 3D artist
50:20 that basically said it's it's a lament
50:22 for his job he's quitting his
50:24 job and he said overnight he he works
50:28 for a 10p person
50:30 company he was a 3D 3D modeling artist
50:33 he would make 3D objects for games I
50:36 think for video games 10p person company
50:39 there were two 3D modelers in the
50:41 company he was the
50:43 best and then there was the other
50:46 guy and then AI came along and his job
50:52 overnight switched from him spending two
50:55 or three days modeling an object to him
50:59 prompting and just choosing the best
51:01 image and then turning that into a 3D
51:04 object and he was like devastated about
51:08 it and the
51:11 coworker that was not as good as him was
51:13 now better than him because he loved
51:15 doing the promp thing because it was
51:17 giving him something so that was a case
51:19 where for him the joy was in the doing
51:22 and that got taken away from him
51:23 overnight so I think that a lot of
51:25 artists are going to have to make that
51:26 choice to say hey where do I actually
51:28 find my joy where's the value for me and
51:31 if it's in the doing if it's in the
51:33 painting then keep
51:36 painting but if you just need to crank
51:38 some [ __ ] out then use
51:40 Ai and if you're like me where I'm doing
51:43 things right now
51:46 that just just this this silly little
51:48 exercise that we did about muscle cars
51:50 in a factory I've done this prompt in
51:53 mid journey and
51:54 Leonardo probably a hundred times on
51:57 this
51:58 channel the adding the objects around it
52:01 just created this new kind of explosion
52:03 of Parts kind of thing like that's a
52:05 concept that was not anywhere in my
52:07 imagin like I I would not have thought
52:09 of this like it invented this out of
52:13 nothing for me that's [ __ ]
52:18 Joy because it's a new creative idea and
52:21 it actually like now I think like oh
52:23 that could be a t-shirt or that could be
52:24 a this or maybe you could turn that into
52:25 into a game
52:27 and like like my creative synapses right
52:32 now are like
52:33 firing because of what it
52:37 did Sherry has been absolutely muted I
52:40 don't have a way to undo it that's
52:43 bad let me see if I can fix
52:45 that Oh Sherry that sucks let me see how
52:49 do I do
52:51 this that's stickers where's
52:54 my
52:56 where uh hold
53:00 please settings
53:03 settings moderator oh wait muted
53:06 accounts no muted accounts so she's not
53:09 been muted
53:12 officially
53:16 moderators oh she's no longer on the
53:18 moderator list all right let me see if I
53:20 can add
53:24 her no she's in
53:28 there
53:30 wait no she's in there
53:35 manage she should be okay just sherar
53:38 just log out log back in see if it's Oh
53:41 Oh Sherry Banks okay wait you're back
53:43 okay you're back um I thought it was
53:45 Sher bear Sherry Banks um let me see if
53:48 you're a moderator I think you are hang
53:49 on a
53:51 sec hold
53:54 please
53:57 I don't maybe you weren't ever a
53:58 moderator I can make you a moderator if
54:00 you want to be
54:04 one no I can't I've got to follow you
54:06 all right I'll I'll figure that out let
54:09 me know if you want to if you want to be
54:10 a moderator Sherry
54:12 Banks um are you seeing my chats yes Ron
54:17 someone just pinned it so I saw it
54:18 because they pinned
54:23 it yes I am I have chat gpc4 can I
54:27 access Dolly 3 from the inside you will
54:30 be able to Chris Martin but I got I
54:34 don't I I don't know how quickly they're
54:36 rolling out the dolly 3 access if you
54:39 can play with Bing right now
54:42 um Bing's image generator is Dolly 3
54:45 it's not as
54:47 slick as the chat GPT implementation but
54:50 it's pretty good so it's the same engine
54:53 right so the same engine that I'm doing
54:54 all this with is in being AI still needs
54:56 human assistance to be good absolutely
54:59 and it and it will I think it will for a
55:00 while and it's not even that so so
55:04 listen will AI need human assistance to
55:08 be good nope I think within three years
55:11 these tools are going to
55:13 be at or
55:15 above almost all
55:18 capabilities I'm super ADHD so a AI has
55:21 been a miracle for me um just to get the
55:25 momentum
55:26 on a project alone is great that's
55:28 that's exactly it for me Tyler it's it's
55:31 this is I can take a vapor of an idea
55:33 and I can manifest it in
55:36 minutes and then I can look at it and go
55:38 is that [ __ ] or not and to your point
55:40 Ron so so I think these tools are going
55:43 to get remarkably good where it really
55:46 won't need human beings
55:49 but these are still tools these are
55:52 tools they're not magic they're not
55:55 doing they're not perform they're not
55:58 using us we're using them right
56:02 Microsoft calls their AI tools co-pilot
56:04 that's the that's the right name in my
56:06 opinion we're the pilot it's the
56:09 co-pilot so it's like this this
56:12 really enthusiastic lap dog that will do
56:15 anything for
56:17 you and then we as humans get to say how
56:20 are we going to use these tools what are
56:22 we going to use them for what are we
56:23 going to create and like what just
56:25 happened happened here where I was just
56:27 sort of demonstrating something I knew
56:30 kind of knowing the result that I'd get
56:33 and then I said oh let me try one other
56:35 thing and it was just kind of a like a
56:37 brain fart like it wasn't even an idea
56:39 really it was just like I know that that
56:42 um that Dolly 3 can do multiple objects
56:45 in a prompt pretty well that like that
56:47 was the only thing I wanted to
56:48 demonstrate and then it came up with
56:50 this new creative concept that you know
56:55 has my world shook so we as humans get
56:57 to say oh that's cool let me take this
57:01 and do something amazing with
57:04 it so
57:07 pogan pogon POG pogon POG pogm meon I
57:14 assume um copyright is gray and
57:18 weird
57:23 so the way these tools
57:25 work the G and GPT stands for generative
57:29 the way these tools work they're
57:30 actually generating original content now
57:34 the models have been trained on in most
57:36 cases copywritten content or copywritten
57:39 content is included in what they've been
57:41 trained
57:42 on what actually ends up going into the
57:45 models into these large language models
57:47 into the
57:48 embeddings is not um is not final work
57:52 one of the misconceptions about um um AI
57:57 is that it's copying and pasting from
57:59 existing work to create a new one it's
58:02 not doing that at
58:03 all it's more Akin like in the way it
58:06 generates these art images it's actually
58:08 more akin to me going to college for
58:11 four years learning about Andy Warhol
58:13 getting really inspired by Andy Warhol
58:15 and then I leave college and I make art
58:17 like Andy Warhol it's still me making
58:19 art but it's heavily influenced by what
58:22 I learned in college thank you for the
58:24 lemonade very
58:26 tasty
58:28 um now it's good enough at doing that
58:31 that sometimes it kind of crosses the
58:32 line right but right now can you use
58:35 that image for a client copyright
58:38 law it is so here's I have a degree in
58:43 acting but I have written seven
58:45 screenplays so I've been paying
58:46 attention to this stuff and the the
58:48 writer Guild strike and the the actors
58:50 and things like that
58:53 um your chances of getting sueded are
58:56 very very low if you use Adobe Firefly
59:00 and generative fill none of their models
59:03 were trained on um copywriting
59:07 copywritten
59:09 images so if you want to be safe use
59:12 Adobe products I I assume that that open
59:16 ai's products were trained on you know
59:19 existing living
59:21 artists um stability AI has now
59:26 um made it possible for artists to opt
59:28 out of the models there are also
59:31 interesting things where Getty right now
59:33 just re released their own generative
59:37 AI um
59:39 tool that is trained on Getty
59:41 copywritten Images where the
59:43 photographers are going to get
59:45 compensated if their images were used I
59:48 don't know how I don't know if it's just
59:50 amortised across all photographs and all
59:52 artists and everyone just gets a cut of
59:55 whatever they make on the AI tool I
59:57 don't know how it's
59:58 working so the only thing that I know
1:00:00 for sure is that for the next 20 years
1:00:03 the intellectual property attorneys are
1:00:04 going to get [ __ ]
1:00:06 Rich cuz everyone's going to sue
1:00:09 everyone so if you you can I use this
1:00:13 image yes might I get sued by someone
1:00:17 possibly like I didn't I didn't invoke
1:00:20 any artist specific name there
1:00:24 um I could take a photograph of a GTO
1:00:27 Judge and put it in a magazine so I
1:00:28 don't think there's an issue there right
1:00:30 so I don't think I'll get sued for this
1:00:32 um but we're in America here I am anyway
1:00:38 we anyone can sue anyone for everything
1:00:40 so everyone's going to sue everyone
1:00:42 right so so is there risk slight now can
1:00:46 I actually copyright this image that's
1:00:50 right now the copyright artist is saying
1:00:53 if it wasn't cra Ed by a human no you
1:00:57 can't could I argue in a
1:01:01 copyright
1:01:03 case
1:01:05 that
1:01:06 um all of the all of the prompting that
1:01:10 I did and all of these
1:01:13 iterations where I said no that's not
1:01:15 right oh that's interesting oh make it
1:01:17 different put it back in a factory that
1:01:19 led to this image and then I
1:01:22 chose this image and then I put that in
1:01:25 Photoshop and I refined it and I stuck
1:01:27 you know judge badging on the car and
1:01:31 did [ __ ] like
1:01:32 that if I make a case to the copyright
1:01:35 office that I as a human contributed to
1:01:39 making that image then it might be
1:01:41 copyrightable right now the the one case
1:01:43 that's there they said no um in the
1:01:46 style of a living artist would that be a
1:01:48 would would that prompt pass a copyright
1:01:52 test I think from an ethical standpoint
1:01:55 one of my friends um um Kenny who's
1:01:59 who's a a really good creative director
1:02:02 and and a um and a uh AI
1:02:08 artist he has he has made it his
1:02:11 personal
1:02:13 um he's drawn a personal Line in the
1:02:16 Sand that he won't
1:02:18 prompt um any image generation tool with
1:02:21 a living artist
1:02:23 name um he just tries to get it the
1:02:27 oldfashioned way and figure it out with
1:02:29 prompting um and so so I think you know
1:02:32 that's that I I think
1:02:34 if if I say in the style of Greg
1:02:38 grotowski is is an artist that got
1:02:41 internet famous because everyone was
1:02:43 putting his name into stable diffusion
1:02:46 prompts and you know he got pissed off
1:02:48 he's like you know this is taking money
1:02:49 out of my pocket in the metadata of
1:02:52 those images his name's in there right
1:02:54 so it's possible that he could go do
1:02:56 cease and desists on anyone who is
1:02:59 sharing an image that's got his name in
1:03:01 it um is it going to be worth his while
1:03:03 to do that in a world where there's sort
1:03:05 of infinite content generation happening
1:03:07 I don't know that's why this is a really
1:03:10 um a really gray area as a prompt
1:03:12 engineer myself I can never use an
1:03:14 artist name and refuse to yeah exactly
1:03:17 and I think that's I think that's a a
1:03:18 very valid place to be and um chat GPT
1:03:23 the the implement ation of this tool in
1:03:26 chat
1:03:27 GPT I asked it to do something in the
1:03:30 style of anel Adams and it wouldn't do
1:03:31 it it it will not invoke an artist's
1:03:33 name but what it did do was it said I
1:03:36 know enough about his work to be able to
1:03:39 describe it and so it chat GPT itself
1:03:41 actually described the look and feel of
1:03:44 of the photograph that I wanted without
1:03:47 using anel Adam's
1:03:48 name
1:03:51 um so and and then and then I think in
1:03:53 the future where this goes is you take
1:03:56 that idea of what Getty Images has done
1:03:59 so they said okay we're going to take
1:04:00 our library we're going to train machine
1:04:02 learning generative AI on our existing
1:04:08 copywritten stuff that we own the
1:04:09 copyrights to and in in conjunction with
1:04:12 the uh with the artists we're going to
1:04:14 do this we're going to copyright them
1:04:16 we're going to figure out some way to do
1:04:17 attribution if you if you scale that
1:04:20 model down to the
1:04:22 individual I can absolutely see a time
1:04:26 in the in the near future probably
1:04:28 within 2 years and I think Adobe is is
1:04:31 perfectly poised to create this platform
1:04:35 so imagine a platform from for from
1:04:37 Adobe that says I as an artist can opt
1:04:40 out of having my stuff trained or you
1:04:43 know my artwork used as training I can
1:04:45 opt into having it used as
1:04:47 training I can get attribution I can get
1:04:50 compensated but I can also create an
1:04:52 official model right so I could so I'm a
1:04:56 photographer I could take my photography
1:04:58 and I could create the official Kyle
1:05:00 Shannon photography
1:05:03 model and if if Adobe made this platform
1:05:07 possible where I could put it in there
1:05:08 and where was someone you know the
1:05:10 prompt engineer that just said that
1:05:11 they're not invoking an artist's name
1:05:14 well what if you could go to Adobe and
1:05:15 you could say let me look for artists
1:05:18 whose work I like and I will invoke
1:05:21 their name because their official model
1:05:23 is in there and if I use Kyle Shannon
1:05:26 photography as my model and make some
1:05:28 images using those cloud pictures Kyle
1:05:31 will get paid for that and he will get
1:05:32 attribution for that I absolutely see
1:05:35 that as a as a major industry moving
1:05:40 forward um put one in comments if this
1:05:43 is your first time two if you've been
1:05:45 here before Turtle if you're an
1:05:46 irregular I like it Zach zero Pippin
1:05:49 that's solid all right put put if if
1:05:53 this is your if you're uh uh at the AI
1:05:55 learning lab this is your first time put
1:05:57 a one if you've been here before put a
1:05:59 two and if you're just here way more
1:06:01 than as [ __ ] reasonable like me put a
1:06:05 turtle I love it and welcome everybody
1:06:10 um my name is Kyle Shannon this is the
1:06:11 AI learning lab I obviously you know wax
1:06:15 philosophic about this stuff I like to
1:06:17 joke around a little bit we haven't been
1:06:18 joking as much tonight I haven't been
1:06:20 making the funny voices and the you can
1:06:22 make money with
1:06:25 but we do that too um we like to have
1:06:27 fun in here happy Saturday night hope
1:06:28 you're all doing well um Let me let me
1:06:32 since the you stumbled onto this channel
1:06:35 like you were flipping through Tik Tok
1:06:36 and you're like wait did did Alec
1:06:38 Baldwin get fat is that Rosie oon why am
1:06:41 I here if you're just like not sure why
1:06:45 you're here or if you've heard about Ai
1:06:47 and you're like I haven't really tried
1:06:48 it yet it scares me I'm not a technical
1:06:50 person any of that sort of stuff you
1:06:53 don't need to be a tech technical person
1:06:55 this is not a technical Channel I not I
1:06:57 have a degree in acting I'm an
1:06:58 entrepreneur I'm the CEO uh and founder
1:07:01 of a company called storyvine it's an 11
1:07:03 and a half year old uh video technology
1:07:06 company a video storytelling
1:07:08 company
1:07:11 I have become pathologically obsessed
1:07:15 with generative AI
1:07:17 because in the mid90s when I stumbled
1:07:20 upon this thing called the worldwide web
1:07:23 I had this Epiphany like oh my God this
1:07:25 is going to change everything and it did
1:07:28 and I started one of the early digital
1:07:29 agencies a company called agency.com and
1:07:32 so I didn't think I'd get to experience
1:07:35 that again in my lifetime and yet here
1:07:37 we are so this generative AI stuff
1:07:40 starting for me the Tipping Point the
1:07:43 inflection point
1:07:46 for the Before Time and the new time is
1:07:50 November 30th 2022 so less than a year
1:07:53 ago chat GPT that site right there chat.
1:07:57 openen AI that launched just about a
1:08:00 year
1:08:01 ago so if you haven't played with chat
1:08:03 GPT just go there right now just go
1:08:05 there right now just [ __ ] go ask it
1:08:08 to write you a country song and describe
1:08:11 your significant other or your pet or
1:08:14 some cliche or non- cliche situation and
1:08:19 it will write you a country song have it
1:08:20 write a program for you have it write a
1:08:23 marketing plan or
1:08:25 project plan or a strategy it doesn't
1:08:28 matter just ask it to do some
1:08:30 stuff and just keep playing with it if
1:08:33 if you want to hang out and listen to me
1:08:34 ramble then we can we can have some good
1:08:37 times together we can have some good
1:08:39 times you be like C but I don't like the
1:08:42 AI the robots are going to kill us c c c
1:08:46 c c
1:08:48 c Kyle why you talk so funny Kyle Kyle
1:08:52 can I make money with Chach you can make
1:08:55 money with ch oh thank you how do you
1:08:57 make money with Chach could you tell me
1:08:59 that k c c c c k
1:09:05 anyway all right let's go let's go get
1:09:08 to some questions shall
1:09:10 we hang
1:09:12 on hang on
1:09:15 people what did I
1:09:19 do yeah yeah see yeah yeah you POA yeah
1:09:23 see
1:09:30 y this is the AI learning lab welcome
1:09:35 this image was made in Dolly last night
1:09:38 it was a cool one it's got a hedgehog
1:09:40 it's got a zombie whatever that is a
1:09:42 zombie dog it's got llamas M llamas is
1:09:46 tasty it's got a Dame it's got a
1:09:51 detective we put all those in the prompt
1:09:54 and and it made a Pulp Fiction mag You
1:09:56 Know cover it's pretty good uh Apple
1:10:00 user thank you for the ruse how do we
1:10:02 convince people in charge to make our
1:10:04 Collective goal to let Ai and automation
1:10:08 do our
1:10:11 work I think people in charge is a
1:10:14 little
1:10:15 bit broad Tyler
1:10:20 um My Philosophy right now why this
1:10:23 channel exists like why am I doing a Tik
1:10:25 Tok Channel
1:10:28 because
1:10:31 um I actually think it's impossible to
1:10:34 convince anyone that AI is anything
1:10:39 other than what's in their head without
1:10:41 them actually using it I don't think
1:10:43 there's an argument that you can make to
1:10:45 anyone whether it's your boss your
1:10:47 boss's boss the CEO of the company the
1:10:50 president of the United
1:10:53 States
1:10:54 until you actually experience what these
1:10:57 things
1:10:59 are and and you recognize two things one
1:11:03 is they're less scary when you use them
1:11:06 than they are when you're outside
1:11:08 looking at what they do they're really
1:11:10 [ __ ] scary when you're outside
1:11:12 looking at what they do when you use
1:11:14 them you realize oh these are just tools
1:11:16 and they're imperfect
1:11:18 tools but then the other thing happens
1:11:20 is they are profoundly powerful tools
1:11:24 and so the only the only solution that I
1:11:27 know on how to convince
1:11:30 people and I'm not even trying to
1:11:32 convince people to use it or love it I
1:11:34 don't give a [ __ ] if you love it or hate
1:11:36 it like I absolutely respect artists and
1:11:39 writers and people that are like [ __ ] it
1:11:41 I'm never going to use it absolutely
1:11:43 respect it I
1:11:45 hope that they take that position having
1:11:48 used it rather than just sitting on the
1:11:51 sidelines because this shit's coming
1:11:54 whether you want it or not you can take
1:11:55 a moral stance that this shit's
1:11:59 evil and if you don't deal with the
1:12:02 reality of what it does you're still
1:12:03 going to be struggling to find work
1:12:06 right taking a moral stance on AI
1:12:09 without being involved in it is kind of
1:12:12 the equivalent of a state like Florida
1:12:14 passing a law that they're no longer
1:12:17 going to accept
1:12:19 hurricanes it's
1:12:22 Noble but it's not necessarily a good
1:12:25 strategy because the hurricanes are
1:12:27 going to keep coming
1:12:30 right this shit's here it's not coming
1:12:33 it's here and the tools as we see them
1:12:36 today are as bad as they're ever going
1:12:39 to be they only get better from
1:12:42 here so that speed of ideation that
1:12:46 speed of
1:12:50 execution just I I mean if you think
1:12:53 about in 2023 the amount of phone calls
1:12:57 that were not
1:12:59 made to artists or
1:13:02 writers let's just take you know image
1:13:05 Image Creators and word makers how many
1:13:09 calls were not
1:13:10 made
1:13:13 because the entrepreneur that used to
1:13:15 hire a copywriter to do social media for
1:13:19 them now has a tool that can do it for
1:13:21 them like that a lot
1:13:27 but you know what ends up happening that
1:13:30 entrepreneur realizes oh I actually do
1:13:34 need a copywriter but I need a
1:13:36 copywriter that's AI literate that
1:13:39 understands how to do these how to use
1:13:41 these tools and do things quickly but I
1:13:43 want them to be more like an editor
1:13:45 right I want them to be to help me pick
1:13:48 the right Concepts and and do all that
1:13:50 sort of stuff so I think copywriters
1:13:52 don't go away but the the the role of
1:13:55 what they do I think changes
1:13:58 dramatically
1:14:02 so all right let's get to some questions
1:14:06 here in other words Irregulars show the
1:14:08 secret sign we've got a secret handshake
1:14:12 good day to you sir good day good day to
1:14:15 you print hello I'm becoming a turtle I
1:14:18 think well at least you got the
1:14:19 invitation
1:14:20 tonight that I think that puts you I
1:14:23 think that put you in the irregular
1:14:25 category and you showed up you actually
1:14:26 clicked on it when you got the
1:14:29 invite I like that um but if you would
1:14:32 need a specific wait but if you need a
1:14:35 specific concept and keep having to
1:14:38 adjust it aren't
1:14:40 you at least creative director yeah
1:14:45 that's what I'm saying I working with
1:14:46 chat GPT is like I think I think in some
1:14:49 sense we all
1:14:51 become it's not even necessarily
1:14:53 creative directors we all become the I
1:14:56 did a I did a video on this that's
1:14:58 that's on the channel
1:15:00 um and the the title of the video is are
1:15:03 you a virtuoso or are you a
1:15:07 conductor
1:15:09 and the the term that I've come up with
1:15:13 since making that that video is
1:15:18 that we're entering an
1:15:21 era where specialization gets
1:15:25 democratized right so if I'm a virtuoso
1:15:28 if I'm a violin
1:15:29 player and actually musical instruments
1:15:31 isn't a good thing because because uh
1:15:34 musical uh capability actually does
1:15:36 require practice right there's some
1:15:38 things like knowledge work where these
1:15:40 tools kind of instantly get you up to
1:15:42 speed there's other things that require
1:15:43 dexterity that don't but just hang with
1:15:46 the metaphor we're moving into a world
1:15:49 where specialization of all the
1:15:51 individual you know instrument in an
1:15:54 orchestra is getting
1:15:56 democratized where I can be as good at
1:15:59 you
1:16:01 know at um I don't
1:16:06 know mathematics over here as someone
1:16:09 who studied mathematics for you know a
1:16:12 bunch of
1:16:13 years um if I focus on it for a week or
1:16:16 two with these AI tools I can be
1:16:18 executing at some level right I might
1:16:20 not be as good like maybe they're going
1:16:21 to always be up here and the AI version
1:16:24 of me and Mathematics is just below them
1:16:26 but it's way higher than I would have
1:16:28 been otherwise so I think each of us
1:16:31 have an opportunity to the the more we
1:16:35 learn about these tools to kind of sit
1:16:37 at the center of this crazy control
1:16:41 panel of like here's words here's music
1:16:43 Here's animations here's Business
1:16:45 Solutions here's strategy here's this
1:16:48 and we get to choose right this gets
1:16:50 back to like what's the role of the
1:16:51 human I think the role of the human is
1:16:53 curator
1:16:55 refiner we get to choose what we put in
1:16:57 the world we get to choose what problems
1:16:59 we solve we're just using different
1:17:01 tools to do it can you be unique with AI
1:17:05 or will it feel like a machine digital
1:17:07 has touched everything I don't think so
1:17:09 man I'm telling you that
1:17:12 the the sophistication of how these
1:17:16 these things are doing images it's it's
1:17:17 a lot of it's in the prompting it's in
1:17:19 the creative
1:17:21 directing if if
1:17:24 if I've got a really interesting concept
1:17:26 all this does is help me bring that
1:17:27 concept to life but it's my
1:17:30 idea so like like the tow dance the the
1:17:34 the kind of subtle implication in what
1:17:37 you're asking and I don't mean
1:17:38 implication in a bad way is that if the
1:17:43 machines are doing the creativity are we
1:17:45 still
1:17:47 creative and and you know the the thing
1:17:50 that I said earlier that that these
1:17:53 images is this
1:17:58 image like I don't know that I've ever
1:18:00 seen anything like that before I mean
1:18:02 I've seen I've seen noling photographs
1:18:05 and I've seen trophies and I've seen you
1:18:07 know sort of tribute things to cars but
1:18:10 like this kind of combines them all in a
1:18:12 really unique and interesting way
1:18:15 and you might look at that and go like
1:18:17 that's boring or I don't like that I
1:18:19 look at that and I'm like that's a new
1:18:20 idea to me so so
1:18:25 I
1:18:26 think I think people that use AI Tools
1:18:29 in a purely automated way like you can
1:18:32 make money with
1:18:33 J make a 100,000 t-shirt designs and
1:18:36 stick them on printify you'll get rich
1:18:39 that work is all going to have a samess
1:18:41 to it it's going to [ __ ] suck and
1:18:43 there's going to be like infinite
1:18:45 amounts of it it's going to be [ __ ]
1:18:49 awful and then what's going to happen is
1:18:52 like my buddy Kenny or me or someone at
1:18:54 the AI learning lab or one of the
1:18:56 Irregulars in here is are going to use
1:18:58 these tools and they're going to use
1:18:59 their [ __ ] brain and they're going to
1:19:02 use critical thinking and they're be
1:19:03 going to be like Oh What If This Were A
1:19:06 game oh I want to make that game okay
1:19:09 let's turn all these things into 3D
1:19:10 objects and I'll 3D print them I know
1:19:12 how I'd make the game and before you
1:19:14 know it someone who just kind of had a
1:19:17 brain fart about something is
1:19:19 manufacturing a game and trying to sell
1:19:21 that to Hasbro like that will be a net
1:19:24 new thing
1:19:25 created so I think you're going to have
1:19:28 both in a world of infinite content
1:19:32 creation there's going to be more
1:19:34 content and it's going to be better
1:19:36 executed and then all the stuff that
1:19:38 doesn't use any imagination is going to
1:19:40 have a sameness to it it's going to
1:19:42 suck and then human beings are going to
1:19:46 choose some [ __ ] to do with these tools
1:19:48 like here's the deal artists are going
1:19:50 to choose to use these tools and and
1:19:53 creative thinkers and innovators in ways
1:19:56 that all of the sort of bubbling masses
1:19:59 will not and they will generate some
1:20:01 [ __ ] that's going to resonate you're
1:20:03 going to be like oh my God that's
1:20:05 amazing how did they design that house
1:20:08 how did they build that house I've never
1:20:10 seen an interior like that hotel before
1:20:12 in my life cuz maybe AI was an
1:20:15 inspiration to do something completely
1:20:17 [ __ ]
1:20:18 different so I think it's both I wonder
1:20:21 if AI will start noticing the user's
1:20:24 patterns and the results get flat like
1:20:28 fyp that's interesting
1:20:33 Barbara I don't think so because of what
1:20:36 I just what I just talked about is what
1:20:39 you don't have with a Tik Tock algorithm
1:20:42 which is your Point's a good one is you
1:20:44 don't have human
1:20:48 intervention humans want
1:20:51 to engage and humans want to contribute
1:20:54 and humans want to do things that are
1:20:57 interesting and new so again these tools
1:21:01 are not making all this [ __ ] on their
1:21:03 own we are choosing to use them to make
1:21:06 all this [ __ ] so it's in what we
1:21:09 choose and what we put into the world is
1:21:12 what determines whether it's interesting
1:21:14 or not or whether it flattens out like
1:21:19 fyp like for
1:21:22 me I I go I go to things like
1:21:27 this why did Billy ish like explode onto
1:21:31 the scene like she
1:21:36 did these were songs that her and her
1:21:38 brother made in her bedroom I mean I
1:21:40 know she was performing she had she had
1:21:42 some stuff there but she just there was
1:21:44 there was something in the
1:21:45 Zeitgeist about who she was how she
1:21:50 sang what the music was how it was
1:21:52 produced it was just completely
1:21:54 different it was completely foreign and
1:21:57 completely familiar at the same time and
1:22:00 people were like when when you heard of
1:22:02 Billy ish song when you heard bad guy
1:22:04 for the first time you leaned in and you
1:22:06 were like what's that holy [ __ ] and then
1:22:09 you heard another one of her songs and
1:22:10 she's
1:22:11 like you're like wait sing out loud and
1:22:14 she's
1:22:16 like like what is what is this what is
1:22:19 this
1:22:21 madness and it became the new
1:22:25 thing so that's not going to stop
1:22:28 happening it's just the tools are going
1:22:30 to be
1:22:32 different what you're going to have is
1:22:34 you're going to have
1:22:37 some 23 year old or 16y old or 75y old
1:22:44 is going to look at these tools and
1:22:46 they're going to make a feature
1:22:49 film on their
1:22:51 phone using gener ative
1:22:54 Ai and they're going to do it in such a
1:22:56 way because of their
1:23:00 worldview they're going to use the exact
1:23:02 same tools that Millions hundreds of
1:23:04 millions of people are using and they're
1:23:06 going to make that feature film in such
1:23:07 a way that the world goes wait what's
1:23:10 that they're going to lean
1:23:12 in and all of a sudden that film or that
1:23:16 director or whatever the hell we call
1:23:18 people that make stories with AI in the
1:23:21 future it's going to explode onto the
1:23:24 scene and it's going to change some it's
1:23:26 going to change it's going to invent a
1:23:30 genre it's somewhere
1:23:33 between you know a Hollywood Blockbuster
1:23:36 and an arcade
1:23:40 game and he's releasing one a
1:23:47 week she created a system that lets you
1:23:51 go in and describe the movie you want to
1:23:53 watch and it just starts playing it for
1:23:55 you so it's it's like Netflix except
1:24:00 every TV show and movie that you watch
1:24:03 are personalized to you imagine a
1:24:07 Spotify where your playlists are not
1:24:09 playlists of other people's songs but
1:24:11 original music written for you based on
1:24:15 your tastes based on your emotional
1:24:17 reaction at the
1:24:20 time who makes that is it spot
1:24:24 or is it
1:24:26 some pimply
1:24:29 faed kid geeking out on the
1:24:34 weekends probably the
1:24:36 latter probably the latter and it's
1:24:39 gonna [ __ ] up
1:24:41 Sony it's gonna [ __ ] up Warner Music and
1:24:44 it's gonna [ __ ] up
1:24:49 Hollywood and it's you know the the
1:24:51 writer Guild they they what they ended
1:24:54 up getting in their contract is good
1:24:56 it's decently future proof but it's not
1:24:58 so restrictive one of my fears was of
1:25:01 the way they were talking about AI in
1:25:04 the in the early part of the strike was
1:25:05 like [ __ ] ai ai is evil we're never
1:25:08 going to use Ai and they didn't do that
1:25:11 which is smart on their part they got
1:25:14 protections against being completely cut
1:25:16 out of the
1:25:18 loop and against um producers
1:25:23 leveraging their talent without
1:25:25 compensating them using
1:25:27 AI um but but they said we can use these
1:25:32 tools but the AI can't get credit
1:25:34 because like the way screenplays work if
1:25:37 someone comes in and they write more
1:25:38 than 50% if they change more than 50% of
1:25:41 a script that new person
1:25:44 gets Buy credit like by Kyle
1:25:47 Shannon and if they do enough work then
1:25:50 the original author drops to Story by
1:25:52 someone else right you know screenplay
1:25:55 by Kyle Shannon Story by Bob right Bob
1:25:57 got cut out they got in there that a an
1:26:01 AI written script can't supersede their
1:26:05 credit
1:26:06 right so that's
1:26:09 good but um like all the shit's going to
1:26:13 change these are always good discussions
1:26:15 I really enjoy your perspective thank
1:26:16 you studmuffin Steve by the way 10 out
1:26:19 of 10 on the name studmuffin Steve's
1:26:21 [ __ ] hilar
1:26:28 ious my God the the Irregulars are going
1:26:31 crazy they're going crazy in the uh in
1:26:35 the Discord hey wait hang on these are
1:26:37 [ __ ]
1:26:40 hilarious wait you got to see some of
1:26:43 these U Can I can I do command plus here
1:26:47 no can I do that no all right here we'll
1:26:51 just we'll just show them this way all
1:26:53 right there's a good
1:26:55 one
1:26:58 janky you can make money with
1:27:03 ch
1:27:06 [ __ ] um that dude's cool little Funko
1:27:09 Pop action AI learning
1:27:12 lab I mean the fact that you can do just
1:27:15 text on stupid [ __ ] images now like
1:27:17 the meme game meme game is going to go
1:27:19 up like like I feel like you know the AI
1:27:22 learned lab is becoming like a meme
1:27:24 Factory at this point we got to start
1:27:26 getting this [ __ ] out in the world maybe
1:27:28 I'll start doing videos of our little
1:27:30 memes that one's
1:27:32 [Laughter]
1:27:36 cute that's really
1:27:42 good and again this one also says you
1:27:45 can make money with chat yout the doorto
1:27:48 door janky
1:27:49 [Laughter]
1:27:51 salesman um if if since since a lot of
1:27:54 you are new here that was pretty good
1:27:56 that's the Billy is you can make money
1:27:57 with Chat gbt thing um since you're new
1:28:00 here one of the reasons we have a turtle
1:28:03 is kind of our mascot is just because
1:28:05 I'm an idiot and one of the videos that
1:28:08 I can not not laugh to every [ __ ]
1:28:11 time I see it now I won't laugh at it
1:28:14 but is this one you know it we all know
1:28:21 it back here live at the waterfront
1:28:24 village with my friend the zombie
1:28:26 Jonathan you're looking good Jonathan
1:28:28 just got an awesome face paint job what
1:28:30 do you think I like turtles all right
1:28:33 you're great zombie and good times here
1:28:36 at the waterfront vill I want to meet
1:28:39 that
1:28:40 reporter oh my
1:28:43 God I don't know how she kept it
1:28:46 together that's a level of
1:28:48 professionalism right there the fact
1:28:49 that she didn't [ __ ] just snap and
1:28:52 break break in
1:28:53 half while while she was live on the
1:28:59 air oh my God that's so
1:29:03 funny so this is what we came up with
1:29:06 last night Turtle pops Shell Shock by
1:29:10 Flavor the ultimate Turtle meat pop M
1:29:13 [Laughter]
1:29:18 tasty H Cliff example ecme widgets want
1:29:22 a new ad campaign you come up with one
1:29:24 aii delivers on your concept
1:29:27 exactly I like turtles oh that's a good
1:29:29 one someone nailed
1:29:32 him who did that steo
1:29:39 steo by the way steo is really into AI
1:29:42 Johnny Knoxville not so much see what
1:29:44 are you gonna
1:29:45 do what are you going to do you're like
1:29:48 is that that steo yeah and I'm Alec
1:29:51 Baldwin
1:29:53 uh uh uh uh uh uh I wanted to do a deep
1:29:56 dive and see what Jonathan is up to
1:30:00 let's invite him to guest the AI which
1:30:03 Jonathan are you talking about
1:30:06 cunu Jonathan
1:30:11 who I got my eyes all dried out or
1:30:14 something it's stinging maybe I rub some
1:30:17 crap in
1:30:21 it
1:30:24 she creates an AI movie production
1:30:26 Studio you enter your story oh the
1:30:29 turtle kid oh that would be fun that
1:30:32 would be fun to bring I
1:30:36 although we we walk a really fine line
1:30:38 between um making fun of Jonathan and
1:30:42 celebrating Jonathan I mean I know it's
1:30:44 been 20 years and he's he's probably
1:30:46 seen all this
1:30:51 but
1:30:53 I love disruptive Tech yeah I do too I I
1:30:55 mean especially disruptive Tech that is
1:30:59 this profoundly powerful like it's it's
1:31:01 just we're just in a we're just in a
1:31:04 we're just in such a different world and
1:31:07 and this was the this was the thing that
1:31:08 I that I experienced when I first um
1:31:12 published Urban desires which was one of
1:31:14 the first digital magazines in
1:31:16 '94 I uploaded I I learned enough HTML
1:31:20 to make a Zen
1:31:22 and then I uploaded these files from my
1:31:24 computer in Brooklyn to a computer in
1:31:26 California with a dialup modem if you're
1:31:29 too young to know what that means you're
1:31:32 an awful
1:31:33 person and 3 weeks later I got a note
1:31:37 from someone on a dialup bulletin board
1:31:39 system that said hey Kyle did you know
1:31:42 that Urban desires has been written up
1:31:45 in France in Liberation own the Parisian
1:31:48 News Daily I was like what no I didn't
1:31:52 know that and I ran down to the
1:31:54 international news stand in Times Square
1:31:57 I was living in New York at the time the
1:31:58 next Sunday and I bought this newspaper
1:32:01 in
1:32:03 French well I bought it in English it
1:32:05 was in French and I flipped through it
1:32:08 and sure enough there's this full page
1:32:10 article of this Zen I made in my
1:32:13 apartment in Brooklyn 3 weeks earlier in
1:32:17 France and I knew in that instant the
1:32:20 world had changed geography had just
1:32:22 collapsed and and time just went from a
1:32:26 week to
1:32:27 instant I knew it in that moment the
1:32:29 world will never be the same and that's
1:32:33 exactly what's happening with this [ __ ]
1:32:35 the world is already different it's just
1:32:37 that most people don't know it
1:32:39 yet and everyone anyone right now that's
1:32:42 sitting on the sidelines I think is
1:32:44 going to be blindsided by how quick this
1:32:47 stuff's coming and how profoundly it's
1:32:49 coming and that's why I do this every
1:32:51 night
1:32:52 have I seen Opus clip can convert Long
1:32:55 YouTube vids into shorter Tik toks um I
1:32:58 haven't I'll check it out there's a lot
1:32:59 of those things they all seem to be
1:33:01 pretty expensive I'd like to find one
1:33:03 that's cheap because I would love to be
1:33:05 able to take these
1:33:07 lives and when I occasionally stumble
1:33:09 into saying something interesting I
1:33:11 would love to be able to turn those into
1:33:13 shorts really easily what's it called
1:33:15 Opus clip I'll check it
1:33:18 out the problem is like all of the
1:33:21 things are like they they give you up to
1:33:23 like 180 minutes for Siri
1:33:27 no bad girl go
1:33:32 away Siri here's what I
1:33:36 found Siri go
1:33:39 away
1:33:45 okay Ser he's my [ __ ]
1:33:49 stalker oh my god oh
1:33:58 oh what was I doing I got distracted oh
1:34:00 Opus
1:34:01 clip welcome to Chad
1:34:04 add Siri doesn't help with the add she
1:34:08 really doesn't AI powered video
1:34:11 repurposing oh what I was saying is like
1:34:13 all these things are like sign up it's
1:34:16 free is it free or is it there's a free
1:34:20 option
1:34:25 get your Clips via email I don't care
1:34:27 how I get them 13 bucks 200 upload
1:34:29 minutes here's the
1:34:31 problem I do a three-hour live every
1:34:35 [ __ ]
1:34:36 night so it's
1:34:40 $13 for 200 minutes it's 180 minutes for
1:34:44 every one of these let's say
1:34:50 150
1:34:54 60 free upload minutes yeah so
1:34:58 anyway the there'll be a free one at
1:35:00 some point or a lowcost one one I can
1:35:02 install on on my own
1:35:04 servers his name's Jonathan we love
1:35:07 Jonathan we do love Jonathan it'll be
1:35:09 fun to have him on just see what he's up
1:35:11 to maybe he's into AI like maybe
1:35:13 Jonathan will become like our new
1:35:16 mascot he was just in a a new commercial
1:35:20 for Teenage Mutant Ninja Tur Turtles
1:35:22 that's very cool I want to know what
1:35:24 Kyle's old online handle was during the
1:35:27 early days of the interwebs I've been
1:35:29 Kyle Shannon
1:35:32 for I think I think I I've been Kyle
1:35:35 Shannon like from the from the
1:35:38 80s I just it's funny when I got into
1:35:41 the I got into blockchain and nft like
1:35:43 three or four years ago and that whole
1:35:46 [ __ ] world is like you know you never
1:35:48 dox yourself man like never put out your
1:35:51 real picture never put out your real
1:35:53 name I'm like uh sorry I got like four
1:35:57 four Decades of that [ __ ] out there too
1:36:00 late for
1:36:02 me can you use AI to create a logo yes
1:36:06 for a brand you can and actually let's
1:36:09 go play with
1:36:11 uh so so here's the here's the the funky
1:36:14 thing about it so you
1:36:16 can mid Journey's good at it
1:36:19 um ideogram a is okay at it it's it's a
1:36:24 little ide's text is a little uh just
1:36:29 janky um I think dolly3 is going to be
1:36:33 really good at it but we'll go check
1:36:35 here um the one thing that ideogram has
1:36:39 that I'm assuming that this doesn't is
1:36:42 ideogram if you put in like the Miami
1:36:44 Dolphins logo ideogram is actually good
1:36:48 at doing existing logos so you could
1:36:51 have it create an existing logo and then
1:36:53 maybe use that as a source to like
1:36:55 iterate off a logo or something like
1:36:57 that but let's go let's go play with
1:36:59 chat GPT for a second what are some of
1:37:01 the websites you all use to create logos
1:37:04 there's listen if you just want like a
1:37:06 really basic logo a lot of the logo
1:37:08 generation sites that are basically
1:37:10 taking like clip art elements and then
1:37:13 quickly mixing those with different type
1:37:16 faces they're fine like you don't need
1:37:18 AI for everything if you want a
1:37:20 sophisticated logo like we did we were
1:37:24 joking around the other night and
1:37:29 did where is it where is it that
1:37:33 one this was Larry the Hedgehog so in
1:37:36 Dolly 3 they were talking about um Larry
1:37:38 the Hedgehog and so we we wanted we came
1:37:41 up with
1:37:42 Larry's wait Larry's Hedgehog's
1:37:45 Landscaping service and so this was the
1:37:47 logo that we did that we did this in mid
1:37:49 Journey but let's go play Let's Go hop
1:37:52 over to uh chat GPT so again if you're
1:37:55 new here um I got access to Dolly 3
1:37:59 within chat gp4 and it's it's it's
1:38:02 really quite something so
1:38:06 so we're going to do this we're going to
1:38:09 do this um I think we're going to do
1:38:11 this in in a good way okay let's
1:38:14 see so I'm not going to go right into
1:38:16 making a logo what I'm going to do is
1:38:18 I'm going to I'm going to start a
1:38:20 convers I'm going to start a strategy
1:38:21 conversation about
1:38:24 branding as if like you know I'm a I'm a
1:38:28 business stakeholder working with a
1:38:30 creative director or if I am the
1:38:32 creative director something like that
1:38:39 okay I want you to act as a branding
1:38:43 expert and we're going to come up with
1:38:46 some branding for a new company period
1:38:49 but before we do that
1:38:55 I
1:38:57 want you to ask me a few questions that
1:39:00 you need to
1:39:03 know to come up with a branding
1:39:07 strategy period from there we'll start
1:39:10 to
1:39:13 explore
1:39:16 Concepts period only when I choose one
1:39:19 of the concepts will we start exploring
1:39:21 what logos might look like period do you
1:39:28 understand and so this approach of
1:39:32 working with chat
1:39:35 gbt as as kind of a a collaborative
1:39:39 partner before jumping into doing images
1:39:41 and things like
1:39:44 that the more you understand about how
1:39:47 these large language models work the
1:39:49 more context that you can provide chat
1:39:51 GPT the better it gets at doing very
1:39:54 specific
1:39:57 tasks okay so company overview can okay
1:40:00 so there's the there's the 10 things
1:40:08 um okay these are great I'm in a bit of
1:40:12 a rush so I might not give you a ton of
1:40:16 detail
1:40:18 period the company is is for a sprinkler
1:40:25 system maintenance
1:40:27 company and the value
1:40:31 proposition is
1:40:34 that for 50 bucks they'll blow out your
1:40:37 sprinkler system in the fall and for a
1:40:41 100
1:40:42 bucks they'll do that plus turn it back
1:40:45 on in the spring
1:40:50 period you never have to remember a
1:40:54 thing they just do it and your pipes
1:40:57 will never freeze again
1:41:01 period I want to create a name for the
1:41:04 company that is
1:41:07 memorable and the target audience is
1:41:10 lazy people who own
1:41:14 houses and have enough money to not
1:41:18 think about this
1:41:20 crap
1:41:21 [Laughter]
1:41:25 company name ideas sprinkle auto spray
1:41:29 care lazy lawu guard that's pretty good
1:41:33 set forget sprinklers seasonal spray
1:41:37 saver branding concept tone and
1:41:39 personality da
1:41:42 taglines set it forget it we've got your
1:41:45 sprinkler sorted that's very English for
1:41:48 the lawn Lo lover with better things to
1:41:50 do okay
1:41:53 so I love the tagline for the lawn lover
1:41:56 with better things to do comma but I'm
1:41:58 not happy with any of the names period
1:42:01 now that we have that as a tagline give
1:42:04 me 10 possible company name
1:42:10 ideas lawn Lux
1:42:14 no where's lawn lover in
1:42:19 here sprinkler sabatical ease
1:42:24 drops um turfscape time these are all
1:42:27 bad Mist
1:42:28 minder Mist minder is not
1:42:36 bad none of those names seem to capture
1:42:39 the idea of lawn lover which is kind of
1:42:42 funny period can you give me some more
1:42:46 creative names that are not quite so
1:42:48 descriptive and capture the essence of
1:42:50 that
1:42:58 tagline and while while I know this is
1:43:01 kind of a ridiculous like you ask can it
1:43:03 make a logo I mean if you're making a
1:43:06 logo this this kind of [ __ ] you want to
1:43:07 answer anyway but this will do a better
1:43:10 job for us because one of the things
1:43:12 that DOI inside of chache BT does is
1:43:15 it's making the image in the context of
1:43:17 this entire conversation or at least
1:43:19 3,000 words of whatever it's you know
1:43:22 it's a token window
1:43:24 is all right what do we got here lawn
1:43:26 lovers leader gra GL grass Glee you
1:43:30 can't say that that's bad Turf treat no
1:43:33 um green Gusto eh Blissful blades that's
1:43:36 that uh Turf tickle oh I want to go with
1:43:39 Turf tickle but it's just too it's too
1:43:41 naughty um we're going to go with
1:43:43 Blissful
1:43:45 blades okay the company's name is
1:43:47 blissful blades And the tagline
1:43:53 is the one we chose before period Please
1:43:57 give me 10
1:44:02 different logo Concepts that
1:44:05 include subject matter typography
1:44:09 description and style
1:44:13 period Please do not generate any images
1:44:16 yet I want to
1:44:19 choose one of the concept
1:44:24 first fantastic choice Blissful blades
1:44:27 here are the 10 logo Concepts based on
1:44:29 the name and
1:44:31 tagline Nature's Embrace blade
1:44:36 ballet Sunrise so wait graceful Blades
1:44:40 of grass Dancing In The Wind typography
1:44:42 elegant ballet inspired no no drop of
1:44:47 delight a water droplet splashing onto a
1:44:50 blade of grass that that's good rounded
1:44:52 joyful font okay drop a delight's pretty
1:44:55 good lawn lounge zenzone no Bliss
1:45:03 bubble heartfelt Heaven okay we're going
1:45:05 to go with whatever the [ __ ] that one
1:45:07 was drop of delight
1:45:12 okay I can't remember the tagline now
1:45:16 can you put the tagline and name of the
1:45:19 company together for me
1:45:23 and
1:45:24 then
1:45:28 select drops of delight as the concept
1:45:32 and just write me up a creative
1:45:36 brief for the
1:45:38 logo that incorporates
1:45:41 everything period do not start making
1:45:44 the image
1:45:48 yet so now I'm having it write a
1:45:50 creative
1:45:51 brief and so now what I'll do now that
1:45:54 we've done all this sort of happy horse
1:45:55 [ __ ] is I'll have it make a logo based
1:45:58 on the creative brief
1:45:59 so company name Blissful blades tagline
1:46:02 for the lawn lover who has better things
1:46:04 to do that's [ __ ] funny um objective
1:46:08 develop a logo for Blissful blades a
1:46:10 sprinkler system maintenance company
1:46:12 that promises hassle-free service for
1:46:14 homeowners the logo by the way I had my
1:46:17 sprinkler system blown out today for 50
1:46:19 bucks in case in case you were wondering
1:46:22 where This brilliant [ __ ] idea came
1:46:25 from there's there's nothing original
1:46:27 going on in this [ __ ]
1:46:32 nogan um concept drops of delight
1:46:35 inspired by joy and convenience the
1:46:37 service brings the logo will feature a
1:46:39 water droplet SP splashing onto a blade
1:46:41 of grass creating ripples of Joy Design
1:46:45 Elements a single pristine water droplet
1:46:48 typography the name Blissful blades will
1:46:50 be written in in a rounded joyful font
1:46:52 color palette Cool Blues for the water
1:46:55 droplet Lush greens for the grass a
1:46:59 natural earthy background style okay
1:47:01 that's good target audience homeowners
1:47:03 who value convenience and are willing to
1:47:05 invest in services that allow more
1:47:07 leisure time they appreciate quality
1:47:09 reliability and a touch of joy in the
1:47:11 everyday things this perfect right so so
1:47:15 so it wait get Henry Winkler on the idea
1:47:18 Winkler sprinkler exactly Ted looks at
1:47:20 life he's got it okay once you're
1:47:23 satisfied with the brief we can Pro
1:47:24 provide we can proceed to the next step
1:47:27 of visualizing the logo I'm just going
1:47:29 to go let's
1:47:33 go and
1:47:35 now there you go it's prompting itself
1:47:38 to make a
1:47:40 logo whether or not it actually makes a
1:47:42 good logo here or if it's you know logo
1:47:45 appropriate we'll see we'll see my
1:47:47 instinct is it's going to be pretty good
1:47:49 at this
1:47:51 but I have not I have not played with
1:47:53 this before I haven't done it people I
1:47:56 haven't done it winklers Winkler
1:47:59 sprinkler this is
1:48:04 awesome defans Winkler sprinkler come
1:48:10 [Laughter]
1:48:15 on I think I think we've got to do
1:48:19 uh
1:48:21 I don't
1:48:22 know it just it feels like these memes
1:48:24 need to go out in the world somehow
1:48:26 these are really funny Blissful
1:48:30 blades not quite spelled right but you
1:48:33 could Photoshop that if you needed to
1:48:36 that's bad that's not
1:48:39 horrible that's getting there but it
1:48:41 looks a little
1:48:43 uh it looks a little squirty and that's
1:48:46 awful all right so I'm going to
1:48:48 [Music]
1:48:49 say I'm going to
1:48:52 say while these are a good start they're
1:48:55 a little too corporate and clean I would
1:48:58 like something more
1:49:00 playful period perhaps you can
1:49:05 anthropomorphize the water
1:49:12 droplet period let's name
1:49:15 him
1:49:19 Winkler
1:49:23 [Laughter]
1:49:27 damn everyone is using Bing lately I've
1:49:29 been using it for months now and it's
1:49:31 really slowed down oh I know it sucks
1:49:33 like gp4 has also just been slower than
1:49:37 crap it
1:49:39 sucks um do you share these prompts um I
1:49:43 can I haven't
1:49:46 um oh see these are better look what we
1:49:50 got here
1:49:52 this is awesome and he's winking get it
1:49:56 winklers sprinklers okay do we so okay
1:50:01 okay
1:50:07 wait we just got a change from corporate
1:50:10 period we're changing the name from
1:50:13 brisel Blades to winklers sprinklers
1:50:18 period and we got Henry win to be our
1:50:23 spokesperson I really like the style of
1:50:25 this last round of logos but can you
1:50:28 make the drop look like The
1:50:33 Fawns and change the name of the
1:50:39 company and and by the way if you've
1:50:41 ever been in Creative that is an
1:50:43 absolutely realistic scenario that you
1:50:47 finally have the thing [ __ ] done and
1:50:48 ready to go and they're like oh yeah we
1:50:50 changed the company name last week you
1:50:51 didn't did we not tell you
1:50:54 that that is absolutely [ __ ]
1:50:57 realistic that is not fiction that's a
1:51:01 [Laughter]
1:51:06 [Laughter]
1:51:09 documentary I wonder if it oh yeah it
1:51:12 said the fonts okay
1:51:14 good what's it say here did styled like
1:51:17 the fonds with a leather jacket and
1:51:19 Iconic Hair style and a confident
1:51:24 wink this going to be
1:51:30 good oh my God I'm telling you man I
1:51:34 don't you know it's funny the the person
1:51:36 earlier that said you know because this
1:51:39 stuff does things so quickly it takes
1:51:41 the joy out of it I I don't know about
1:51:43 you but like I like I find this like
1:51:46 incredibly joyful like this this feels
1:51:49 very very right to
1:51:54 me like I don't know I'm digging it all
1:51:58 right these weren't as good as the other
1:51:59 ones the blades of bliss but anyway you
1:52:02 get the idea so couple of things about
1:52:04 this
1:52:05 um so these the the image generation
1:52:09 tools for the most part are pretty low
1:52:11 um
1:52:13 resolution so you would obviously want
1:52:15 to take these into um into illustrator
1:52:19 and vectorize them they're vector izing
1:52:20 tools out there I know there's one for
1:52:24 um for Mid Journey that you just install
1:52:27 into Discord that'll vectorize things
1:52:30 um so anyway yeah it does
1:52:49 logos
1:52:55 welcome to chat ad D what time is it
1:52:58 what time is it Eddie what time is it
1:53:01 why you washing Dicky in the
1:53:03 sink I got an ice cream and you can't
1:53:07 none's
1:53:12 onare name that
1:53:16 comedian um should I be taking notes
1:53:18 recording this um I do put these up on
1:53:21 YouTube so if you go to you
1:53:25 I I wouldn't take notes of me okay
1:53:28 here's the deal Michael
1:53:33 ATX hopefully what I'm sharing is
1:53:36 valuable and um you're learning stuff
1:53:40 great any tactical thing that I'm doing
1:53:43 today is going to be irrelevant 3 months
1:53:46 from
1:53:48 now these tools are changing ing so F in
1:53:51 fact it's funny that I put that picture
1:53:53 up the Curiosity picture if I can
1:53:56 inspire you if anyone that's in there we
1:53:58 got 113 people in here right
1:54:01 now the thing that I most hope that you
1:54:04 leave this with is curiosity and a sense
1:54:06 of adventure to go start playing with
1:54:08 these tools and just start playing with
1:54:10 them like it it it really doesn't matter
1:54:12 don't worry about how I prompted if you
1:54:14 like how I prompt great if you want to
1:54:16 take notes great just know that anything
1:54:18 that I'm teaching right now is likely
1:54:19 going to be Irrelevant in a few months
1:54:22 this is all right now about play and
1:54:24 exploration what do these tools do
1:54:26 because the the [ __ ] that I just showed
1:54:28 making that logo that was not available
1:54:31 4 days
1:54:33 ago there was a version of it in bing
1:54:36 like a week and a half
1:54:39 ago but that [ __ ] that I just showed
1:54:42 it's four days old and not everyone has
1:54:43 access to it right now and it's rolling
1:54:45 out over the next month and by the time
1:54:47 they roll it out widely it will probably
1:54:49 have different features like right now
1:54:51 if I make an image and I'm like hey I
1:54:53 just want to tweak this image I can't so
1:54:56 I got to just kind of restart everything
1:54:58 or pop it into Photoshop and illustrator
1:55:01 and start working on it you know the
1:55:03 oldfashioned way as if you had to
1:55:06 actually work on [ __ ] but what I what I
1:55:09 would actually do is I'd take it into
1:55:10 Photoshop and use generative fill and
1:55:12 just you know muck with it that
1:55:15 way so anyway um but yeah I I I if this
1:55:21 is a value cool if you want to take
1:55:24 notes have at
1:55:26 it and you're welcome to record this but
1:55:29 I do put these up we've got a YouTube
1:55:30 channel the YouTube channel
1:55:36 is where is my cursor when my screen's
1:55:39 bright enough that I can see my cursor
1:55:41 it's too bright for the Tik
1:55:43 Tock okay the YouTube channel is at the
1:55:46 top right here at AI learning lab
1:55:54 dtt I know that's where dolly is janky
1:55:57 is is the is the iterating on stuff yeah
1:56:00 like being able to refine something yeah
1:56:02 there's no refining right now it's all
1:56:05 Dolly right now is all um you know
1:56:13 ideation Eddie Eddie Murphy Eddie Murphy
1:56:16 Millennials are cool Millennials are
1:56:18 cool millennials don't know the word
1:56:21 winner or loser
1:56:23 loser the dependents are calling I have
1:56:26 to go [ __ ] dependence I know isn't it
1:56:28 rough sorry art thanks for hanging out I
1:56:31 appreciate it I know I'm going to get
1:56:34 the call soon how much longer you going
1:56:36 to be in
1:56:40 there we never get to hang out anymore I
1:56:43 know but I got a thing I got to do
1:56:48 Eddie
1:56:54 [Applause]
1:56:57 who did that Sherry D you're the
1:57:06 best and you can't
1:57:14 have I just installed doia as an
1:57:17 application on Windows right now
1:57:21 e Apple user be very careful about that
1:57:26 so it's probably not the official if if
1:57:29 the manufacturer of that application is
1:57:31 not open AI it's not the dolly
1:57:36 application if it's if it's another
1:57:39 application that's using the dolly 3 API
1:57:42 which I don't think is out
1:57:45 yet I I would be careful with that um
1:57:49 this is what one of the problems with
1:57:51 um with uh with these tools being so
1:57:54 popular right now is that there's a lot
1:57:56 of these [ __ ]
1:57:59 make making stupid little apps that like
1:58:02 people want access to Dolly 3 so they
1:58:04 put put out an app that's like this is
1:58:06 the dolly 3 app it's probably not it's
1:58:08 probably actually calling unstable
1:58:10 diffusion as my guess um or Dolly 2 so
1:58:14 just be careful with it um hey that was
1:58:17 almost a Bill Clinton impression I used
1:58:20 to do Bill
1:58:21 Clinton I haven't done I haven't done
1:58:23 him in ages though he just he went
1:58:25 irrelevant there for a while and I just
1:58:27 kind of forgot him I'll have to dig up
1:58:29 the old the old Bill Clinton things AI
1:58:32 3D modeling good question not really
1:58:34 there yet there's a lot of white papers
1:58:37 that have come out over the summer
1:58:39 taking 2D images and turning those into
1:58:41 3D Point clouds there's a thing right
1:58:43 now going on called gaui and splatting
1:58:46 holy [ __ ] it looks amazing have you guys
1:58:48 seen any of the gaui and splatting stuff
1:58:50 that's what Monica used to
1:58:53 say I just saw an interview with her
1:58:57 yesterday God man
1:58:59 what just it [ __ ] just destroyed her
1:59:03 life it's like she could never Escape
1:59:06 that Legacy of just that
1:59:09 [ __ ] um let's see um the gazi and
1:59:13 splatting stuff is just this crazy
1:59:15 there's these things called
1:59:18 Nerfs neural something Radiance
1:59:23 fields that basically just capture the
1:59:26 the the photons in 3D space and like you
1:59:31 can change lighting it makes 3D and then
1:59:34 now there's a thing called
1:59:37 uh uh ner splatting which which takes
1:59:41 very incomplete information and turns it
1:59:44 into these highly detailed 3D models
1:59:46 like people are making these these
1:59:49 highly detailed 3D models with very very
1:59:51 little information and input just like
1:59:54 scan it with a
1:59:56 phone I mean there's tools like
1:59:58 matterport that have been doing [ __ ]
2:00:00 like that for years but they were were
2:00:02 really
2:00:03 um um they were really like processor
2:00:07 intensive and slow um yeah let me let me
2:00:10 see if I can go find some neural people
2:00:12 ask like how do you find all your [ __ ]
2:00:14 how do you keep up with it mostly just
2:00:15 Twitter
2:00:18 honestly follow good people and follow
2:00:20 if when you find good people like Ethan
2:00:22 mik or Andre karpathy from open AI um
2:00:26 when you find good people
2:00:30 just um follow the people that they
2:00:32 follow and your Twitter feed will get
2:00:34 better and
2:00:36 better all right let's go
2:00:40 neural
2:00:48 splatting m
2:00:50 [Music]
2:00:54 is that in Focus not
2:00:59 really you can now generate 3D Assets in
2:01:03 2 minutes without the use of neural
2:01:04 networks dream gajian is a new model
2:01:07 that uses gazi gajian splatting actually
2:01:10 let me look up gajian splatting gajian
2:01:18 splatting
2:01:22 uh uh uh uh uh
2:01:26 uh new 3D gajian splatting
2:01:30 recording okay so they're recording so
2:01:33 so this is a watch I guess they're
2:01:35 taking pictures of this thing in real
2:01:38 time when looking closer you can see how
2:01:41 the watch hands are
2:01:43 modeled just a couple of elongated
2:01:46 gausian so like oh wow oh so see that
2:01:51 see the the really sparse Point
2:01:56 Cloud it starts with that sparse Point
2:01:59 cloud and it turns it into this detailed
2:02:01 3D model like I'm telling you like the
2:02:06 [ __ ] as as magical as the [ __ ] that
2:02:08 we've seen with images over the past
2:02:11 year and a half two years let's call it
2:02:13 2 years two years ago text to image were
2:02:18 laughably bad and then a year ago they
2:02:22 were surprisingly good and then over the
2:02:24 summer they became like shockingly good
2:02:28 like oh [ __ ] but they didn't do
2:02:32 text and then like a month ago IDR comes
2:02:36 out and it does text and then three
2:02:39 weeks after that Dolly 3 comes out and
2:02:41 it does text and
2:02:44 coherence and
2:02:46 then the video stuff the 3D stuff the
2:02:51 music stuff all of it right now is kind
2:02:53 of where images were 2 years ago so
2:02:56 there's hints of really cool stuff
2:02:58 coming it's all in these white
2:03:02 papers like here's here's someone like
2:03:05 walking through it looks like a like
2:03:08 a a set of like
2:03:12 Boo right and it goes from like these
2:03:14 really loose
2:03:17 dots to semi detailed
2:03:21 stuff the capture came out with a false
2:03:24 result before but now works better than
2:03:26 I
2:03:27 expected this is in akih
2:03:31 Habra yeah I thought that looked
2:03:33 Japanese that's in akih
2:03:37 Habra so there's the gaui in see how see
2:03:39 how sparse it is and then look how it
2:03:41 fills it in like how's it doing that I
2:03:43 don't [ __ ] know I mean this is
2:03:45 similar like the the the the [ __ ]
2:03:48 math wizard that are figuring all this
2:03:50 [ __ ] out right now they're on fire look
2:03:54 at this using gazi and splatting 1.3
2:03:58 seconds to a 3D object and old Nerf
2:04:13 technology 12 seconds right so so like
2:04:17 right now the scientist the Geeks
2:04:19 whatever whatever they're doing they're
2:04:21 going from a 12 15 second thing to a
2:04:24 like a 1.2 second thing like that's
2:04:28 dramatic Improvement and then these are
2:04:30 going to just be built into our
2:04:31 operating systems they're going to be
2:04:33 built into our tools you're going to be
2:04:35 able to ask
2:04:36 for just like I asked for that logo I'm
2:04:39 like oh that's great I want to turn that
2:04:40 into a mascot can I 3D print it sure let
2:04:43 me let me spin that up for you and then
2:04:45 we'll just send it over your 3D
2:04:48 printer
2:04:50 [ __ ]
2:04:52 insane look at this that's not
2:04:57 video that's not video that's someone
2:05:00 moving around a 3D World they captured
2:05:02 with their
2:05:08 phone 3D was either fast or pretty now
2:05:11 it's both interactive SE scenes with
2:05:13 gazi and
2:05:15 splatting [ __ ] nuts what is the link
2:05:19 at Kyle
2:05:22 x.com this is this is Twitter I just
2:05:25 looked up gaui and splatting and I'm
2:05:27 just I'm literally just scrolling down
2:05:29 through the gazan
2:05:35 Splats and this
2:05:37 is this is one of the problems right
2:05:43 now most of what I play with on this
2:05:45 channel is just the word [ __ ] right well
2:05:49 well now the word [ __ ] has images in it
2:05:53 and then pretty soon you're going to be
2:05:54 you're going to be able to it can read
2:05:56 images as well and pretty soon you're
2:05:58 going to be able to talk to it and it'll
2:05:59 talk back and then it's going to be able
2:06:00 to make music then it's going to be able
2:06:03 to make
2:06:04 video all of these
2:06:07 tools are going to get increasingly
2:06:11 sophisticated so the so so if you're
2:06:13 sitting there going but but I haven't
2:06:14 even figured out chat GPT yet exactly
2:06:17 like that's why that's why I say you
2:06:20 know if if you're here go right now and
2:06:25 if if you have not played with
2:06:27 chbt go to chat. open.com right now
2:06:31 today and just start playing it doesn't
2:06:34 matter what you do go start learning how
2:06:36 these things work learn what's possible
2:06:39 today what they're good at what they're
2:06:41 not good at and then just start paying
2:06:43 attention to this [ __ ] and as something
2:06:46 intrigues you like if you're really into
2:06:48 the gaui and splatter idea go run down
2:06:50 that rabbit hole every single discipline
2:06:55 whether it's statistics or data
2:06:57 visualization or videography or special
2:07:00 effects or music
2:07:03 making or images or just [ __ ]
2:07:07 everything there are probably a thousand
2:07:10 projects right now using generative AI
2:07:13 all these tools that are based on the
2:07:15 same [ __ ] that chat GPT is built out
2:07:18 of happening right
2:07:22 now you could spend 24 hours a day 7
2:07:25 days a week and never come close to
2:07:28 catching up with this stuff so if you're
2:07:29 feeling overwhelmed and like you're way
2:07:32 behind everyone is no one's an expert in
2:07:39 this this channel is about me
2:07:42 essentially sharing in real time the
2:07:44 [ __ ] I'm learning in real time like I
2:07:47 might be a few days ahead of you
2:07:49 not much further because it just keeps
2:07:56 changing so I do virtual production for
2:07:59 film where we are using this stuff for
2:08:01 film Productions yeah I know well so
2:08:04 it's it's so amazing um I made it do a
2:08:06 thing that's super cool
2:08:09 um one of the things that blew me away
2:08:11 if if if you haven't if you want to geek
2:08:13 out on some of this stuff Go Geek out
2:08:16 have you ever played with unreal yeah
2:08:17 I've played a little bit with unreal I
2:08:19 think the procedural stuff in unreal 5
2:08:22 where you you just drop in objects and
2:08:25 it creates whole sections of Worlds you
2:08:27 attach that to a large language model
2:08:29 front end um or if uh have you played
2:08:33 with um what's it called um Skybox Labs
2:08:38 at blockade Labs um I made it do a thing
2:08:41 because that's one where it's just text
2:08:43 to 3D world but the Skybox Labs is just
2:08:46 a a 2D image that's mapped onto a sphere
2:08:48 it's still pretty cool but people are
2:08:50 now taking those things dropping them
2:08:52 into Unreal Engine adding objects in
2:08:54 them and making them um you know
2:08:58 interactive but but yeah like we're not
2:09:00 that far from describe a world and like
2:09:05 it will generate it good enough that you
2:09:08 can put it in a
2:09:09 film or you can enter it like the
2:09:12 metaverse or you can turn it into a game
2:09:15 environment we're not that far from that
2:09:18 and if you want to geek out on something
2:09:19 check out how the Mandalorian is filmed
2:09:22 th this is [ __ ] insane they basically
2:09:24 have a cylindrical Sound Stage that's
2:09:27 all
2:09:28 LED so they have real [ __ ] on the floor
2:09:32 the background is all led the sky is all
2:09:37 LED and they're generating those worlds
2:09:39 in Unreal Engine and and essentially
2:09:42 they're saying oh wouldn't it be cool if
2:09:44 it were snowing and the wind were
2:09:47 blowing so they they just whip it up and
2:09:50 unreal and toss it up on the [ __ ] led
2:09:54 psych and that's how they film that
2:09:56 thing so so you have instant locations
2:09:59 well that's that's Hollywood so now
2:10:02 imagine that sort of shrinking down and
2:10:06 each of us individually is going to be
2:10:07 able to create that [ __ ] by asking for
2:10:13 it you think it was a big deal that you
2:10:16 could make a movie on your iPhone cuz
2:10:18 the camera's got good enough wait until
2:10:21 you can do fullon special effects by
2:10:24 asking for them I predicts homes will go
2:10:28 minimalist to support a metaverse
2:10:30 existence that's actually a really
2:10:31 interesting idea I hadn't thought about
2:10:33 that Lord
2:10:34 John that's a really cool
2:10:38 idea
2:10:41 H yeah you you walk into someone's house
2:10:43 and they got a bunch of pictures on the
2:10:45 wall and furniture and you're like oh
2:10:48 you're one of those you still go with
2:10:51 this static art thing
2:10:54 interesting it's not how I roll here
2:10:56 come the holidex Kyle exactly Clint
2:11:04 copperhead look at jrc did this is
2:11:07 [ __ ] awesome causi and
2:11:11 splatting that's [ __ ]
2:11:13 [Laughter]
2:11:17 great
2:11:19 it go go oh that one's
2:11:25 cool I mean that that um the the V the
2:11:29 the uh the AI art at the Metropolitan
2:11:33 Museum of Modern Art right now in New
2:11:35 York City that big 20ft tall thing with
2:11:38 the crazy shapes yeah that's all of that
2:11:42 stuff that's that's sort of bigger than
2:11:44 all of us right now is going to start
2:11:46 shrinking down very very quickly and
2:11:48 we're going to have we're going to have
2:11:49 immediate access Tyler the utopian
2:11:52 unreal is about to start charging
2:11:54 licensing fees for all non-gaming stuff
2:11:56 like virtual production yeah they
2:12:03 should I mean quite frankly Unreal
2:12:12 Engine I mean this is fascinating
2:12:15 this so their current model is you can
2:12:18 use it for free and only if you start
2:12:20 making money on it do you pay them
2:12:25 right but that kind of presumes that
2:12:28 you're going to
2:12:29 take some static set of
2:12:32 assets and you're going to put it in
2:12:34 some static sort of game or some static
2:12:36 sort of environment like a movie
2:12:40 production
2:12:41 thing and you're going to get a bunch of
2:12:44 people to pay for that and then when
2:12:46 enough people pay for that then you give
2:12:47 them a cut
2:12:51 what
2:12:54 if we all just subscribe to Unreal
2:12:59 Engine just like we subscribe to Netflix
2:13:02 today or to Hulu
2:13:05 today and we pay them1 1995 a month to
2:13:10 generate infinite worlds infinite game
2:13:13 environments infinite
2:13:16 right because there model sort of
2:13:20 presumes one Creator to many audience
2:13:25 right I'm going to create a game and
2:13:27 when I get to a million or more users I
2:13:30 think it's 100,000 when I get to 100,000
2:13:32 or more
2:13:33 users then I'll start paying unre
2:13:36 engine well what if the audience is
2:13:40 one I want an environment of a snowy
2:13:43 Tundra with Pac-Man game mechanics in It
2:13:47 Go and Unreal Engine goes here's your
2:13:51 Pac-Man in the Arctic game Kyle I'd pay
2:13:55 1995 for
2:13:58 that and then they become then then
2:14:01 unreal becomes kind of a content
2:14:04 provider it's insane it's insane I'm
2:14:07 telling you every [ __ ] business model
2:14:10 every [ __ ] business model is going to
2:14:13 be
2:14:14 threatened and man is there going to be
2:14:17 opportunity
2:14:19 because if Unreal Engine
2:14:21 doesn't realize that someone else will
2:14:24 someone else will start it they'll be
2:14:25 like wait a minute unreal engin sticking
2:14:27 into that old business model and it
2:14:29 sounds like I like I feel like Unreal
2:14:31 Engine like they're they're they're
2:14:33 really Savvy and it's [ __ ] like that
2:14:36 that technology is staggeringly good so
2:14:39 I can only imagine that they're as good
2:14:42 at the business side as they are on the
2:14:44 technology side how do you think it will
2:14:46 affect crime it's already hard to catch
2:14:49 criminals
2:14:51 without this
2:14:54 Tech well I think I mean surveillance
2:14:58 State's been around for for some
2:15:01 time I think what what this technology
2:15:03 does is it democratizes surveillance
2:15:05 Tech down to the individual like you
2:15:07 don't have to be a big police force to
2:15:09 have you know surveillance Tech in fact
2:15:12 I think there's I've seen some videos on
2:15:14 Twitter or YouTube or maybe even Tik Tok
2:15:17 where people are using object
2:15:18 recognition there was a guy oh you know
2:15:21 there what a guy
2:15:23 did he basically put a webcam in his
2:15:26 window so so parking by his apartment
2:15:29 was like really brutal really bad so so
2:15:33 he put a webcam in his window that runs
2:15:37 24/7 and all it does is analyze the
2:15:40 parking spots and it texts him whenever
2:15:44 whenever there's moving whenever there's
2:15:45 a person walking up to a car that's at
2:15:47 of parking spot it texts him so that he
2:15:51 can run home and get the parking
2:15:54 spot so
2:15:57 so you know that kind of technology is
2:15:59 going to get trivial so it it might it
2:16:02 might Aid in Catching more criminals but
2:16:05 you know the criminals are going to have
2:16:06 access to these tools so so they're
2:16:09 going to get more sophisticated about
2:16:11 what they can do again I go back to this
2:16:12 idea of the democratization of
2:16:14 specialization right
2:16:17 so more and more individuals are going
2:16:19 to have more and more skills because the
2:16:22 AI is augmenting the skills that they
2:16:25 didn't have so I've got a certain set of
2:16:28 skills they're they're like this wide
2:16:30 right and then I have some tangential
2:16:32 skills that maybe go to this
2:16:34 wide but with AI that skill set can be
2:16:38 this
2:16:40 wide and everyone's going to have that
2:16:42 everyone's going to have whatever they
2:16:44 do now and they're going to be be able
2:16:45 to extend out laterally their skill
2:16:49 sets I think despite the AI takeover
2:16:51 rollout Happening Now increased desire
2:16:55 for
2:16:57 entertainment yeah I agree and I think I
2:17:01 think the increased lunatick is in um in
2:17:05 personalized entertainment like I want
2:17:07 my entertainment the way I want it right
2:17:10 now it's that the the reality of that's
2:17:14 not probably for 5 years or 7 years but
2:17:18 the ability to just ask for a movie that
2:17:21 is stars who you want it to Star and
2:17:24 tells the story you want it to tell we
2:17:26 just be there how long would it take
2:17:29 before chat GPT can code a playable
2:17:31 version of pong it can do it right now
2:17:34 let's go let's
2:17:36 go I wonder if I can do it in code
2:17:39 interpreter let's
2:17:43 [Music]
2:17:46 see I go new
2:17:49 chat I'm going to go Advanced Data
2:17:52 analysis and I'm going to
2:17:58 say can you create a playable version of
2:18:04 pong can you create a playable version
2:18:06 of pong that I can play in the web
2:18:09 browser
2:18:15 here might be limited to the inter face
2:18:19 I'm going to use a library like
2:18:20 turtle you'd need to run this on your
2:18:23 local machine to play it interactively
2:18:25 would that work for
2:18:27 you
2:18:30 uh can you make it a downloadable
2:18:33 executable that can run on
2:18:47 OSX
2:18:53 by by the way if you if you've never
2:18:55 seen chat GPT code yeah it
2:19:02 codes so so actually so what we were
2:19:04 just talking about before about
2:19:05 personalized
2:19:07 entertainment so look at what this is
2:19:09 doing here I asked it to create a game
2:19:10 of pong like I don't need to know game
2:19:13 mechanics I don't need to know anything
2:19:15 now whether or not this is any good or
2:19:16 not that's a whole different question
2:19:18 question but it knows enough about code
2:19:21 it knows what pong is right there
2:19:23 there's enough examples in the world of
2:19:25 pong that that you can just do this [ __ ]
2:19:27 so now imagine
2:19:31 this taking this kind of knowledge of
2:19:33 what Unreal Engine can do so so Unreal
2:19:36 Engine has some finite amount of
2:19:39 functions that it does right it can make
2:19:42 a sphere it can make a this it can make
2:19:43 a particle it can change the parameters
2:19:46 of those particles it can link these
2:19:48 things together it can add textures and
2:19:50 3D objects and all that sort of stuff
2:19:52 right it's got all these
2:19:54 functions well you can train large
2:19:56 language models to learn all those
2:19:57 functions and then just like it just
2:20:00 wrote code like that it's going to be
2:20:02 able to just create [ __ ] in Unreal
2:20:05 Engine and then as these things get more
2:20:08 sophisticated it'll be able to take the
2:20:10 game mechanics from this programming
2:20:11 environment the the 3D you know and and
2:20:15 game elements from Unreal Engine the
2:20:17 story telling from you know this cool
2:20:21 storytelling GPT Library thing and smash
2:20:25 them Al together and deliver it to you
2:20:27 essentially in real
2:20:29 time um after verifying that the game
2:20:32 works to run this
2:20:34 code and the turtle Library installed on
2:20:37 your machine um let's
2:20:41 see can I run this on repet and if so
2:20:45 how do I set that up
2:20:49 uh uh uh uh yes you can run this on
2:20:52 repet create an account create a new
2:20:56 repet paste the code all right let's see
2:21:00 if it
2:21:01 works copy
2:21:03 code now I'm going to go to
2:21:12 repet log in log in log
2:21:16 in
2:21:19 welcome
2:21:20 welcome to you all this is fantastic
2:21:24 what did you do Saturday night well I
2:21:26 don't want to talk about
2:21:30 it create a
2:21:32 ripple a rippel a
2:21:37 python be bad
2:21:43 pong all right
2:21:46 paste
2:21:51 why did it not copy that come on come on
2:21:54 you
2:21:58 dang there we go
2:22:04 run
2:22:09 uh it's doing
2:22:16 something yeah I don't know enough about
2:22:18 it to do it but uh the answer is yes you
2:22:21 can do it right now if you took enough
2:22:23 time to sort of figure out I think you
2:22:25 need what it probably didn't do right
2:22:27 there is it probably didn't create an
2:22:28 interface for it it's probably it's
2:22:30 probably running code that you can
2:22:33 technically play it but it's not showing
2:22:34 it so I don't know maybe I need to
2:22:36 install the turtle turtle library but so
2:22:40 the answer is yes but I just don't I
2:22:42 don't want to do it here
2:22:43 cuz I don't think it would be
2:22:45 interesting watching me fail 60 times to
2:22:48 get it
2:22:50 going big
2:22:52 beer I'm here hey steo what's
2:22:59 happening all right let's
2:23:05 see all right I gotta go soon what time
2:23:08 is it it is getting late it's getting
2:23:10 late here people just learning lab we go
2:23:15 late we we rock it out we rock it out
2:23:18 y'all having fun tonight we having a
2:23:20 good time here AR glasses that show you
2:23:23 how to pick a lock oh yeah that'll be
2:23:25 easy yeah I did Cam wait I did Cam 2 and
2:23:29 put music open all lights for thieves at
2:23:32 200 kilometers oh
2:23:34 interesting what's the best way to build
2:23:37 an app hire a
2:23:41 programmer pretty soon you're just going
2:23:43 to be able to like ask it for an app and
2:23:45 it'll just make well I mean you can do
2:23:46 that now the the problem with it right
2:23:48 now so to back to this idea that these
2:23:50 tools are just kind of janky pieces of
2:23:52 [ __ ] right
2:23:53 now
2:23:57 is right now it really behooves you to
2:24:00 be good at engineering if you want to do
2:24:02 programming right it won't be that long
2:24:06 where you won't need to be good at
2:24:07 programming but right now you still need
2:24:09 to be you still need to like like if
2:24:11 you've got basic skills like you know
2:24:13 what a Google collab is and you can you
2:24:15 can you know fire up a Google
2:24:18 collab notebook and you know you know
2:24:20 what GitHub is and how to read the
2:24:22 readme files there and you can go to
2:24:24 hugging face and start playing around
2:24:26 with [ __ ] there if if you're if you're
2:24:28 sort of geek
2:24:30 tinkery if you like Tinker at being a
2:24:32 geek like I do you're fine right now you
2:24:35 can figure this [ __ ] out it just takes
2:24:36 you some time to do it but the good news
2:24:38 is you have chat GPT that can teach you
2:24:41 how to do that [ __ ] even if you're not
2:24:43 good at this stuff but you're curious
2:24:44 and you're like [ __ ] it I want to get
2:24:45 good at programming spend a month with
2:24:48 chat
2:24:50 GPT I bet you'll be
2:24:52 remarkable like you'll be at a
2:24:54 remarkable level of programming you
2:24:56 could have never imagined yourself in if
2:24:59 you spend the next 30 days with
2:25:02 gp4 having it code for you and teach you
2:25:04 how to code and build
2:25:07 applications I I just go spend 30 days
2:25:11 what's the Worst That Could Happen 30
2:25:12 days from now you're like I need a
2:25:13 [ __ ] drink this is horrible Sherry
2:25:15 Banks hey Kyle would you look at the
2:25:18 Discord I'm trying to tell you that
2:25:20 joker and I coded something
2:25:23 oh wait what' you do what did you people
2:25:37 do a spreadsheet
2:25:41 format blood typing what app did you
2:25:45 make oh well damn show and tell okay
2:25:47 fine I'll go to show and
2:25:59 tell what did you make I can't find
2:26:04 it I can't find
2:26:06 [Laughter]
2:26:08 it I'm not fully sure about him
2:26:16 yet
2:26:17 [Music]
2:26:24 where show
2:26:26 tell oh yeah showcase show and
2:26:30 tell what did you make Sherry D need
2:26:35 needing 12 step pretty soon that was
2:26:36 from last night I don't see what you
2:26:39 made but if if it's there Kyle
2:26:42 yes he needed oh he needed a blood
2:26:45 platelet thing he has to tell you but oh
2:26:49 there's a Google collab
2:26:52 Link in showand tell or in
2:26:56 um I saw that in Irregulars hang
2:27:05 on oh I see Google collab oh okay wait
2:27:10 all joking
2:27:13 aside hold
2:27:16 please
2:27:19 um there was banter between myself and
2:27:21 him one
2:27:22 day anyway he needed to create a bot for
2:27:26 health care for healthcare issue and no
2:27:29 surprise there are
2:27:33 several ER
2:27:35 translate so Joker gave me the gave me a
2:27:38 form of what he was working on basically
2:27:40 a form that needed turn into an app chat
2:27:43 GPT produced a working and functioning
2:27:46 program hope to put this on my resume
2:27:49 someday to Learn Python has been so
2:27:52 intimidating until now here are some
2:27:54 screenshots of the code
2:28:00 itself is
2:28:11 it
2:28:13 so so they're so they're building
2:28:17 so Irregulars now people that hang out
2:28:20 in this channel are now starting to
2:28:22 collaborate so I I mean so [ __ ] far
2:28:26 out first of all so thank you for
2:28:28 sharing that but what's Happening Here
2:28:31 is people that were just like [ __ ]
2:28:34 hang out hanging out in here because I'm
2:28:36 an idiot and we we like to have fun are
2:28:38 now starting to collaborate together and
2:28:40 build [ __ ] together and Sherry Banks is
2:28:42 building python
2:28:44 applications to do blood platelet
2:28:47 spread something wait it's for antibody
2:28:51 identification for pre- transfusions of
2:28:54 rbc's
2:28:56 actually I don't know what any of that
2:28:58 means but that doesn't matter anymore
2:29:01 because I could know what that means in
2:29:02 three seconds in chat GPT so so two of
2:29:06 the Irregulars now were talking and said
2:29:09 oh I'm trying to solve this and then
2:29:11 Sherry Banks said oh let me try
2:29:13 something and now they got a thing
2:29:21 this that that so so to me um Joker and
2:29:27 and Sherry this is the
2:29:31 perfect like this is the perfect example
2:29:34 of I think something that we're going to
2:29:35 see over and over and over again that I
2:29:37 think people that get curious about Ai
2:29:40 and I I've said this before I'll keep
2:29:42 like it keeps being reinforced for me it
2:29:45 doesn't [ __ ] matter what you do try
2:29:48 anything go make images go make 3D
2:29:51 objects go do gazan splatting it doesn't
2:29:54 [ __ ] matter just go start playing
2:29:55 with this stuff get really good at
2:29:57 planning parties with chat GPT it
2:29:59 doesn't [ __ ]
2:30:01 matter because what's going to end up
2:30:03 happening is you're going to stumble
2:30:05 onto something or you're going to
2:30:06 stumble onto
2:30:08 someone who says I'm really trying to
2:30:10 solve this particular part of the
2:30:12 problem and you're like oh I know how to
2:30:15 solve that part of the problem or would
2:30:17 you mind if I tried experimenting to
2:30:19 figure out if I could solve that part of
2:30:21 the problem cuz I think I might be able
2:30:23 to
2:30:24 sure and companies are going to be born
2:30:27 out of that and people are going to
2:30:28 fundamentally reinvent who they are and
2:30:31 what they do I think we're going to see
2:30:33 it over and over and over that's the
2:30:35 opportunity right now it's not you can
2:30:38 make money
2:30:39 with there's no easy money this is all
2:30:42 [ __ ]
2:30:43 work but we've got a tool set now we got
2:30:46 a tool set in AI that is profoundly
2:30:52 powerful so start
2:30:55 playing start collaborating with people
2:30:58 hey I'm here's what I learned here's
2:31:00 what I'm doing the joking that we're
2:31:02 doing where people are dumping stuff
2:31:03 into the AI Salon Discord and just
2:31:06 putting the meme pictures in there
2:31:08 they're also having conversations in
2:31:10 between those about oh here's what I
2:31:12 learned and here's what I here's what's
2:31:13 going on for
2:31:15 me so it's an expansive conversation
2:31:17 it's fun we're having a lot of
2:31:19 fun and we're learning [ __ ] so I'm I'm
2:31:23 really excited about this um if you want
2:31:25 to join the AI salon by the
2:31:28 way it's
2:31:30 free for for all you cynical people oh
2:31:33 here comes the sales pitch he just wants
2:31:35 to make money with the stupid [ __ ] um
2:31:38 the salon. a is just a link
2:31:40 tree there's a link to our meetups which
2:31:43 happen every other Tuesday and then
2:31:46 there's a link to the Discord there
2:31:47 there's also a link that describes what
2:31:49 the salon is what our values are our
2:31:51 values are
2:31:55 important because we don't want
2:31:59 douches but here's the values
2:32:03 um let's see what direction curiosity is
2:32:07 the first
2:32:09 one be curious
2:32:13 generosity
2:32:15 exploration like be willing to explore
2:32:20 collaboration what Joker and Sherry did
2:32:25 inclusion and empathy the empathy one
2:32:29 listen this one's really
2:32:31 important as
2:32:34 as enthusiastic as I am for this
2:32:37 technology and for talking about it I
2:32:40 get that this shit's [ __ ] terrifying
2:32:44 and that a lot of people are facing EX
2:32:52 existential
2:32:54 risk for their livelihood for what they
2:33:01 enjoy but there's also just massive
2:33:04 opportunity but there's
2:33:07 no there's no way to to grab the
2:33:11 opportunity without going through the
2:33:15 [ __ ] you're going to have to learn
2:33:17 something
2:33:20 new you're gonna have to geek out a
2:33:22 little bit you're gonna have to forgive
2:33:24 the imperfections of the
2:33:25 state-of-the-art right now cuz all of
2:33:27 the
2:33:29 opportunity like there's nothing but
2:33:31 opportunity right now the the industry
2:33:34 is so small it's so like paper thin and
2:33:37 it's going to do this very very soon and
2:33:39 so anyone that starts playing right now
2:33:41 like your opportunity is massive but you
2:33:44 may have to be confronted with
2:33:47 I've been a copywriter for 30 years and
2:33:50 all of a sudden these [ __ ] tools do
2:33:53 what I do better than I do
2:33:55 it or at least faster
2:34:01 right and you have the choice do you
2:34:03 keep being a copywriter do you become an
2:34:05 AI powered
2:34:07 copywriter do you take your word skills
2:34:10 and start getting good at prompting
2:34:12 images and expand what you can do and
2:34:14 expand what you sell possibly
2:34:20 crazy crazy crazy crazy I like Claude
2:34:23 better than GPT since all the GPT
2:34:25 updates yeah I can see that I think
2:34:27 claud's claude's going to be a really
2:34:29 major competitor the $4 billion do from
2:34:32 Amazon probably doesn't hurt
2:34:34 either make a poster of James Dean as
2:34:37 Tarzan I can do that I got to go I got
2:34:41 to go to bed AI is non-judgmental
2:34:43 professional coding friend literally the
2:34:45 dream of for would be coder it's totally
2:34:48 it's Tyler so you're you're absolutely
2:34:52 right it's like um it's like the golden
2:34:55 retriever of work Partners it's just
2:34:58 like sure boss what do you want me to
2:35:01 do he needed a blood platelet thing he
2:35:04 has to tell you um but it worked on
2:35:06 python it's that's so amazing it's for
2:35:09 antibody identification actually super
2:35:11 cool there's a Google collab link very
2:35:15 amazing all right everyone listen all
2:35:18 right what's Joker and and uh Sher are
2:35:20 saying something here thank you for
2:35:21 creating a space for this where would
2:35:23 you find this yeah I thank you Sherry
2:35:29 um you know it's funny when I started
2:35:32 this I didn't I still don't really have
2:35:35 an end game in mind for this Channel or
2:35:37 for these
2:35:39 lives and it was it was funny when when
2:35:43 the some of the Irregulars went to the
2:35:45 salon when I started started talking
2:35:46 about the salon and they started showing
2:35:49 up for Salon meetings I was like oh wait
2:35:51 I've got my that's that's my Tik Tok
2:35:53 crowd now they're my salon thing and I I
2:35:55 had them in my mind as separate
2:35:57 things and what I'm realizing is that
2:36:01 these it's this is all it's it's all
2:36:03 just us trying to figure this [ __ ] out
2:36:06 so it doesn't [ __ ] matter like I feel
2:36:08 like this gives us an opportunity to
2:36:11 kind of invite new people in talk about
2:36:14 things explore things to together the
2:36:17 Discord allows us to collaborate you
2:36:19 know and I I don't know I'm I'm really
2:36:22 excited where it's going so um end end
2:36:25 question mark I don't know what that
2:36:26 means steo oh end end go to the
2:36:29 end did you figure out the mid Jour oh
2:36:32 that's you talking to each other I don't
2:36:33 know what end end means steo oh is this
2:36:36 the end yeah yeah I gotta go I gotta go
2:36:39 um it's getting late it's getting late
2:36:41 listen I'm great seeing you all Sher
2:36:44 Banks Joker thank you for sharing that
2:36:46 prod project Winston thank you for the
2:36:48 Llama a tasty um follow the rabbit
2:36:52 Silver Fox great seeing you who who else
2:36:54 we got in here madcap
2:36:57 Winston Robert Rossy good to see you
2:36:59 good night Emilio's wife endgame you can
2:37:03 check out anytime you like but you can
2:37:05 never leave they do expect me back
2:37:08 regularly if I'm not here they're like
2:37:10 Kyle where are you what's going on
2:37:11 everything okay you you're doing live
2:37:13 tonight right very irregular jrc good to
2:37:17 see you thank you very much digital Gods
2:37:20 madcap good to see you all right let's
2:37:24 see
2:37:25 Firefly just looking through here who
2:37:28 else we
2:37:31 got all right everybody happy Saturday
2:37:34 night um the
2:37:38 uh the uh tomorrow morning you got your
2:37:41 uh your F1 race if you're into F1
2:37:47 Max forsten's going to win again but you
2:37:49 know it'll be interesting to see who
2:37:51 comes in
2:37:55 second hope everyone has a good night
2:37:57 peace
2:37:59 out