AI Learning Lab

Dec 16, 2023 - Exploring AI Tools, Creativity, and Career Reinvention

TqU93Z4FVos
Video2023-12-172:36:156 views

Description

In the latest episode of the AI Learning Lab, Kyle explores the transformative potential of artificial intelligence, emphasizing its role in reshaping industries and personal creativity. He discusses the rapid advancements in generative AI, highlighting how tools like ChatGPT and MidJourney are not just technological novelties but gateways to new opportunities for individuals across various fields. Kyle encourages viewers to embrace curiosity and experimentation, asserting that understanding AI can lead to significant personal and professional growth. He also touches on the societal implications of AI, urging a balanced perspective that recognizes both its challenges and its potential to enhance human capabilities. For more insights and engaging discussions, check out the AI Learning Lab on TikTok: [AI Learning Lab](https://tiktok.com/@aiLearningLab). #ArtificialIntelligence #GenerativeAI #AICommunity #Innovation #Curiosity #TechTrends #airevolution Chapters: 00:00:00 Intro 00:02:28 AI Learning Lab Introduction 00:07:57 Sharing the Livestream 00:13:16 Using 00:16:01 The Magic of AI 00:18:19 Demo of Crea 00:23:33 AI Salon Show and Tell 00:26:07 Exploring GPTs 00:29:00 Action GPT and APIs 00:31:30 Mid 00:38:53 Free AI Image Generation 00:46:04 Co-Pilot's Image Generation Capabilities 00:51:00 Midjourney's New Features 00:52:35 Adding Data to GPTs 00:59:24 Late Guest Arrival 01:00:57 OpenAI Enterprise Access 01:06:11 The Impact of ChatGPT 01:16:14 Kiosk Technology 01:20:41 Leonardo AI Demo 01:31:15 Fine-Tuned Models in Leonardo 01:36:46 AI's Electricity Consumption 01:38:02 Choosing the Right AI Tool 01:46:04 The "We Have No Moat" Article 01:55:51 Impact of AI on Jobs 02:06:13 AGI and its Implications 02:13:27 Empathy in the AI Age 02:18:35 The Breadth of AI Reinvention 02:22:11 DeepMind's Material Science Breakthrough 02:28:13 The Great Renaissance of Innovation

Chapters

Transcript

0:00 M you ready to go you
0:06 [Music]
0:12 [Music]
0:18 ready
0:21 [Music]
0:23 hey hey no
0:25 [Music]
0:28 rapping get away masterly
0:32 [Music]
0:34 Direction this car is my
0:39 train I've been driving I've been
0:43 wondering what it is I run from
0:48 again feel like an 8-year-old
0:52 man old on to
0:57 29 up head on that Horizon
1:00 [Music]
1:02 is a California
1:05 [Music]
1:07 [Applause]
1:10 [Music]
1:16 l
1:18 c he's not really into it tonight cute
1:22 little front
1:24 [Laughter]
1:29 door
1:31 [Music]
1:59 oh
2:02 [Music]
2:09 oh all right enough doggy singing doggy
2:14 singing doggy singing doggy wasn't into
2:18 it Dogg's also a little gassy but you
2:20 know what are you going to do what are
2:22 you going to do perer all
2:25 right beautiful what's happening good
2:28 people welcome to the AI learning lap
2:30 holy crap there's 155 people in here
2:33 sorry about that if you haven't been
2:35 here before I kind of open these things
2:38 lately by uh strumming a poorly tuned
2:42 guitar with a poorly sung voice and a
2:44 lovely melodic canine named
2:48 champ uh digital Gods thank you so much
2:51 you're very generous as always thank you
2:53 so much I appreciate that uh this is the
2:55 AI learning
2:57 lab
2:58 and what are we doing here it's Saturday
3:02 night we hang out here instead of having
3:05 a real
3:07 life
3:09 because this AI stuff you know what I'm
3:12 saying it's uh it's a bit captivating
3:16 it's a bit it's a bit
3:18 uh crack
3:23 cocain um but it's also it's also not
3:26 going away and the whole purpose of this
3:29 channel is is to get curious about this
3:32 stuff so that's what it's about so if
3:33 you have questions about AI pop them in
3:35 the comments below if one of your first
3:38 thoughts is but sir what are your
3:40 qualifications may I
3:42 inquire about your qualifications good
3:45 sir would be fantastic if we would know
3:48 exactly what your qualifications are oh
3:51 rest assured I have none okay uh which
3:55 is good
3:57 because anyone right now that tells
4:00 tells you they're an expert on this [ __ ]
4:02 is
4:03 lying um there are lots of experts on AI
4:06 and machine learning but there are no
4:08 experts on generative Ai and what they
4:13 mean for anything for work for us
4:17 personally for the economy any of it um
4:21 but what I'm confident about and why
4:23 this channel really exists is that
4:26 anyone who gets curious and gets AI
4:29 literate in the next 18 months is going
4:30 to have massive amounts of
4:32 opportunity and people that resisted and
4:34 don't do that
4:36 um are you know are going to be caught I
4:41 think off guard by how fast this stuff's
4:44 moving um just even even since I've been
4:47 doing these lives like what these tools
4:50 make possible and how you use them is
4:52 completely different like it's
4:53 completely different um and and you know
4:57 I feel as clueless as ever so if you're
4:59 feeling like like I'm totally behind
5:01 what do I
5:02 do um that feeling doesn't go away that
5:06 I've
5:09 discovered
5:11 but you do you do understand a bit more
5:14 where the edges of possibility are um
5:18 and that's that's kind of what this is
5:19 all about so so with that that's what
5:22 that's what we're going to do tonight so
5:24 so hop in buckle up buckle up
5:27 buttercup we're going on a ride out on a
5:32 prairie and then you know if we get it
5:34 right then you can make money with we
5:37 can make money with chat
5:40 gbt um this lovely image was from Sherry
5:43 D put that she posted that in the AI
5:47 Salon
5:48 um show and tell channel so uh I'll talk
5:52 about that what the AI Salon is in a
5:55 bit and uh where do we get started let's
5:58 let's look and see what data
6:01 analysis let's look and see what people
6:03 are asking about in here and figure out
6:05 where to start um I discovered there's a
6:07 new open AI added a new superpower
6:11 inside of the GPT maker which I'm going
6:13 to show you tonight so if you hang out
6:16 I'll get to that at some point here um
6:20 so that's super cool let's see what's
6:22 everyone been
6:23 doing cam kakin hey Kyle how have you
6:26 been doing I've been doing really good
6:28 did a little little shopping today we
6:31 went to a mall which
6:36 was just what is it with malls man
6:40 they're [ __ ]
6:43 creepy then we went to Trader Joe's and
6:45 that was fine and then I took a nappy I
6:48 got a haircut too look at look look man
6:53 what you're live making I just got age
6:56 restricted again all right I have I said
6:59 any
7:02 anything whatever put your kids to
7:10 bed
7:12 oh oh Tick Tock what what H are you
7:16 doing champy like the be
7:19 cord
7:21 mhm champy did like the be cord that was
7:24 a B7 I
7:27 think oh dragonfly Alchemy Kyle k k good
7:31 evening good evening to you Tobias hello
7:35 Cosmic lover hello AI
7:39 friends there's a pile of llamas champy
7:43 chunes encourage more viewers to share
7:46 that's a good idea but you're not afraid
7:47 so do me a favor we we were at 150 but
7:50 then apparently I was saying things that
7:53 young people couldn't hear and we lost a
7:54 bunch of people so if you'd be so kind
7:57 share the
7:58 live and we'll get going I'll I'll share
8:01 some things here and maybe we can get
8:02 some more people in here so so here's
8:05 here's a fun game share the live with
8:09 with friends of yours that you know will
8:11 be cranky if you send them something
8:13 about
8:16 Ai and I'll I'll I'll help him I'll help
8:19 him over the threshold I'll carry him
8:21 over the threshold uh and we'll get
8:23 started here if you're new to this AI
8:25 stuff we we'll have fun we'll we'll get
8:27 you started your regulars oh the
8:29 Irregulars by the way if you see people
8:31 referring to the Irregulars those are
8:33 the people who show up regularly get
8:37 it they came up with that name I love it
8:40 it it perfectly describes them um and me
8:44 we're it's a it's a neuros spicy uh
8:47 friendly zone so if if you if you have
8:51 any of the uh any of the uh what do they
8:55 call them
8:58 um twice EX
9:00 exeptional if you have any of the three
9:02 times twice or Thrice exceptional uh
9:05 attributes welcome so do
9:09 we I made nonmanual Reading ADHD from
9:13 way back before we knew what it was
9:15 called uh anyone have a weird Verizon
9:18 experience today go on daily crunch what
9:22 was your weird Verizon experience that
9:24 feels like that could
9:26 be all over the place was there some
9:29 sort of weird
9:30 call did you just randomly get connected
9:34 to other people was it a at a Verizon
9:36 store and there's some back room that
9:38 nobody knew about it's I I feel like you
9:41 just can't let that one hang in there
9:46 buddy Irregulars equals
9:51 dateless that's not true we've got each
9:53 other let's let's let's call it that
9:57 it'll make us feel better
10:05 H Lord digital
10:07 Gods yeah so I tried to KN digital Gods
10:10 the other night he got he got the title
10:13 of title gifter because he sends a lot
10:15 of
10:16 roses and so I told him I would I would
10:19 start calling him sir now's he's already
10:23 self-proclaim digital Gods right and so
10:26 I think I know I'll be nice I'll Knight
10:29 him
10:30 and and I'll call him sir digital gods
10:32 and what did he come back with not thank
10:34 you not oh you're so swell um a nighting
10:38 would be awesome no no he came back with
10:40 it's Lord digital Gods it's Lord digital
10:44 Gods so now he's Lord digital Gods
10:47 that's what it's I I don't make up the
10:50 rules I just work here all right lord
10:53 lord is
10:58 good he's happy with Lord this this is
11:00 good we want to keep Lord digital Gods
11:03 happy
11:04 so looking well coffeed tonight well I'm
11:08 well napped and I and I am uh I am uh oh
11:12 looking well caed I thought you said
11:13 well
11:15 coffeed yeah I got a haircut so they
11:19 they trimmed off all of my extraneous
11:21 hairs that just happened when you get
11:23 this
11:24 age I don't know any of you gen xers
11:26 what the [ __ ] is with the hair just
11:28 sprouting
11:31 like oh all
11:35 right I have
11:38 no that's
11:42 good daily crunch I can't talk about it
11:45 okay wow that must have been a good one
11:48 a good Verizon
11:51 experience nice that's solid the baron
11:55 of AI no let's stop at Lord I think
11:57 we're good um oh good
12:00 God they say that excessive laughing
12:03 might be caused by long and unprotected
12:05 exposure to ai ai is a blast man it's
12:08 it's one of the things that
12:10 that an
12:12 unexpected pleasure of doing this
12:14 channel
12:16 is you know it's called the AI learning
12:18 lab and I thought well I know I'll teach
12:21 people as I'm learning right and I I had
12:24 in my head it was going to be this more
12:25 formal kind of thing but I'm an idiot
12:28 and then
12:30 the these tools are just so [ __ ] like
12:32 instant and magical and it it allows you
12:35 to just essentially play
12:37 jazz while you're using them and um I
12:42 was playing with Korea today and that's
12:44 that it felt like jazz like it felt like
12:46 playing jazz like
12:49 improvising um and it's a lot of fun and
12:51 and the the uh the nature of this
12:55 channel has gone from you know me sort
12:57 of talking and teaching to like a party
13:00 where people are co-creating stuff and
13:02 sharing it on the the AI Salon site um
13:05 it's it's a hell of a lot of fun so all
13:08 right um let's get going so so how we
13:11 going to do
13:16 this so if you are brand new to the to
13:20 the world of
13:21 AI if you've not used chat GPT I put
13:25 this I put a little session from that
13:28 last night I don't know sometime in the
13:30 past week um I put a video up on my
13:34 channel um that that was a session from
13:36 a live talking about this thing um go to
13:40 chat GPT right now if you haven't played
13:42 with it before go there right
13:45 now and don't wait like go right now
13:48 like while I'm rambling just pop on over
13:52 there fire up your laptop or iPad or
13:56 drop on out of
13:57 here and just let's go start playing
13:59 with
14:00 it I use chat GPT for business and worry
14:03 about accident accidentally
14:08 paced I requested do not train yeah you
14:12 should anytime you're using um these
14:15 things with that that are you know the
14:18 the the commercial large language models
14:21 like Bard and you know co-pilot from
14:24 Microsoft and chat GPT and Claude all of
14:28 them you're sending your data to a third
14:33 party most of their agreements say that
14:36 they can use your data to
14:39 train their systems which you know can
14:43 make your data vulnerable
14:45 so uh good practice number one is don't
14:47 send any anything sensitive to them at
14:50 all or anonymize it first that's
14:52 something that we've started doing at
14:54 storyvine my company is if we're sending
14:57 any data we'll strip it of any
14:59 identifying information or swap out
15:01 information and then do the work we have
15:04 to do there and then we put put the data
15:06 back in later um locally um the other
15:09 thing that you can do is you can start
15:11 exploring open source models um I was on
15:15 Ann's uh Ann Murphy's got a Saturday
15:18 afternoon GPT making party and she had
15:21 on uh this guy Aon today who showed us
15:25 all how to make um like local servers
15:28 and install local large language models
15:31 and things like that there's it's it's a
15:32 whole world in and of itself um but if
15:35 you really want to get um safe with your
15:39 data um don't send the data outside of
15:42 your uh outside of your little thing
15:45 it's Alchemy Kyle it's true U Mary Mary
15:48 this is bigger than the sum of our o of
15:50 Our Own Parts yep absolutely
15:53 agree absolutely agree it's it's it's
15:56 increasingly feeling these tools are
15:58 inre increasingly
16:01 feeling um I think why they feel
16:04 magical is they feel less like tools
16:07 they feel less like robots and more like
16:09 you're collaborating with the the sum
16:13 total of humanity and that's kind of
16:15 what we're doing with them right they've
16:18 all been trained on the data that we've
16:19 dumped onto the public internet and the
16:22 you know and and when you interact with
16:25 them it it kind of amalgamates them sort
16:27 of in the direction that you're trying
16:29 to go and people from from years
16:36 past somehow get Amalgamated into your
16:38 answer and it I mean it's kind of what
16:40 we're doing it's it's kind of a
16:41 hallucination um Andre karpathy talked
16:44 about the hallucinations as a feature
16:46 not a bug um that it's that that chat
16:49 GPT is just in this constant dream state
16:51 and you're just sort of pulling stuff
16:54 out of the ethos of humanity which is
16:57 really really fascinating um I took it
17:00 took chat GPT months and months to Grant
17:03 Enterprise to my company they are
17:06 real back really backed up yeah exactly
17:09 I think you have to
17:11 spend 10 or 15 grand a month minimum to
17:14 even be considered for a chat GPT um
17:18 Enterprise license they um they did say
17:21 when they announced Enterprise
17:24 subscriptions that they would have a
17:26 small business like a self-service thing
17:28 but we haven't heard anything about that
17:30 since they announced it that was 6
17:31 months or so
17:33 ago
17:34 um let's see it's not Alchemy stop
17:38 treating it like it's incomprehensible
17:40 magic no pay it is well no it's
17:43 comprehensible
17:45 calculations that occur as magic it's
17:48 the it's the um what's What's um Arthur
17:52 C Clark isn't it isn't it his quote any
17:55 sufficiently advanced technology is
17:57 indistinguishable from Magic
18:00 we're both
18:01 right it's math that occurs like magic
18:06 um it's really cool math when you
18:08 understand how these tools work and I
18:10 will keep talking about it as magic I
18:13 will cuz because it's like like I I mean
18:17 the um look at uh Korea this this site
18:22 that is probably not smart for me to to
18:25 go live with but um
18:29 you know this is this thing where you
18:31 move sign up I
18:34 I'm God here we go with the login parade
18:39 am I good all right are we good to go
18:41 all right so so here's this site where
18:44 you know we're just sort of moving
18:45 shapes around in
18:47 space
18:51 and and it's it's rendering you know
18:54 something over there on the right I did
18:56 a cool thing this afternoon what did I
18:58 do
18:59 I went to custom background I went to
19:02 textures I grabbed that
19:05 one so so you got that as the input you
19:09 got that as the output so the output is
19:12 some little prompt here so I'll roll the
19:15 dice so D dystopian cityscape Rebels
19:19 with elemental powers akin to
19:21 action-packed
19:22 shaon
19:24 anime all right that's cool H I got to
19:28 be careful here when I do these live so
19:31 I I I nearly got permanently banned by
19:33 using this tool live one day
19:37 so
19:45 um's Majestic Fortress in the clouds all
19:48 right let's go let's go put in a
19:51 different input image let's do that
19:55 one there's a fortress in the clouds all
19:58 right good good and then now we'll go
20:00 grab a
20:01 shape we'll do a
20:04 pyramid and we'll change the color of
20:06 that pyramid to some deep dark
20:15 purple all
20:18 right I mean yes this is
20:22 math but but this is occurring as
20:27 magic and then you sort of you know over
20:30 here you can change your the the models
20:33 from illustration or cinematic to
20:35 illustration to bloom whatever the hell
20:37 that is to
20:39 product I don't even know what those
20:43 do um let's go grab a circle
20:48 here here
20:51 oops oh that was cool oh no
20:57 h
21:00 [Music]
21:02 oh neat look at that and it made that
21:05 little space over there to the right
21:06 that's cool now let's make a another
21:11 Circle up
21:15 here and let's make
21:25 it yeah that's what I'm talking about
21:34 look at that come on how cool is
21:37 that that's cool Kyle that's what I'm
21:41 saying I know it's not magic I know it's
21:44 math but damn
21:47 it and I mean this tool and then there's
21:51 also like I I discovered another thing
21:52 about this tool today there's this
21:54 button here this quick
21:57 enhance that it I don't quite know what
22:00 it does but does
22:02 something and
22:05 then let's see let's
22:09 go can I go back to where I was sort of
22:20 interesting
22:26 huh this is by the way the website here
22:29 is kaak kre a.
22:42 a I don't know I just think this stuff
22:45 is
22:46 cool I don't like that let go out and
22:50 back in that one's pretty
22:57 cool
23:26 whoa there we go all right that's done
23:29 okay so what I'm going to do I'm going
23:30 to copy the image I'm going to go over
23:32 to the
23:33 AI
23:35 salon and I'm going to go to the show
23:37 and tell channel so I'll talk about the
23:39 AI Salon here in a second and I'm going
23:42 to share
23:47 [Music]
23:50 this um
23:53 another Crea
23:56 creation oh wait I'm going to do Crea
23:59 creation get
24:02 it K r e a t i o n
24:08 huh another K
24:11 creation
24:13 um and then I'll do
24:17 hashtag images and I'll do hasht
24:20 Irregulars for the community here and
24:23 then I'll do #
24:26 Crea although there's enough Korea in
24:28 that that it is
24:34 findable
24:39 post beautiful so there you go you can
24:43 go go be inspired by that I find that
24:47 inspiring so what what's cool about
24:51 Korea is is just this idea that you know
24:57 it's if you if you've gotten really good
24:59 at Mid Journey prompting for example now
25:02 all of a sudden we're using shapes and
25:06 background images and in conjunction
25:09 with prompts and then I I like and and
25:12 it's it is maddeningly unpredictable
25:15 right now and there's something
25:17 beautiful about that and so what's the
25:19 skill here so if you were not good at
25:22 Mid Journey prompting you might be
25:25 really good at this right there's going
25:27 to be people that are
25:28 kickass at Kaa Creations that are going
25:33 to be crappy at Mid Journey creations
25:35 and vice versa and then same with the
25:37 video tools same with the music tools
25:39 like all of these tools are going to
25:41 start having different kinds of
25:43 interfaces and and inputs and things
25:45 like that crazy crazy crazy
25:49 um was I going to show you oh I was
25:51 going to show you the new thing so if if
25:53 you've made uh
25:55 gpts I'm I'm over in chat GPT and I can
25:58 I can sort of go over I can take this
26:01 wherever anyone wants to take it if you
26:03 have specific questions about chat gbt
26:05 um I'm happy to demo
26:07 things there's a new if you create a
26:13 GPT um you go to explore and then at the
26:16 top of the explore tab there's this
26:18 create
26:20 GPT and the way it works is over here
26:24 you create the GPT you can just talk it
26:26 into existence you basic they basically
26:28 put chat GPT 4 as the interface to
26:31 create one of these things it's a
26:33 customized version of chat
26:35 GPT um but one of the things that you
26:38 can do if you click on the configure tab
26:41 is you can do things like you know
26:44 upload um your own data for for to use
26:49 as the knowledge base you can turn on
26:51 and off capabilities like web browsing
26:53 image creation and code
26:55 interpreter um and then there's also
26:59 this thing create a new action so an
27:01 action is basically an API call to an
27:04 outside party so let's say you wanted to
27:06 do a GPT where you want to
27:10 um recommend activities for someone's
27:14 vacation you could use the API to Google
27:17 maps to put pins on a Google map and you
27:21 would use an API for that well they just
27:23 added a new
27:24 thing in
27:27 here
27:31 and is it down yeah here it is
27:36 okay so so this this is some geeky stuff
27:40 you can import from a URL if you've got
27:42 the API schema you can you can import it
27:45 from a URL or you can type your own so
27:48 you can type in code for the for the uh
27:52 the schema but if you don't know a ton
27:55 about API stuff this will eventually get
27:57 easier but what what they've done is
27:58 they've built in here this new thing
28:01 called action
28:02 GPT so you can say you you launch action
28:06 GPT and this is a
28:09 GPT to help you um create apis in your
28:15 GPT so so you can say um I
28:19 want um you to explain
28:25 apis to me like I have a well no it's
28:31 because it'll it'll use weird references
28:32 like like I am a fifth
28:38 grader um I want to
28:43 connect
28:44 my GPT to Google
28:48 Maps so I can
28:52 drop
28:55 pins on the map
28:59 can you help question
29:04 mark and
29:06 then sure imagine you have a toy your
29:09 GPT program and you want that toy to
29:10 play with another toy Google Maps very
29:14 cool but these toys speak different
29:16 languages an API is like a translator
29:18 that helps them understand each other
29:20 okay that's good I can understand that I
29:22 can get down with that now to connect
29:24 the GPT to Google Maps um and drop pins
29:27 on the map you would use a specific part
29:29 of the Google Maps
29:31 API the part that understands commands
29:34 like put a pin here on the street or
29:36 show me where this city is on the map
29:39 here's a basic example of how you would
29:41 describe this open AI in an open AI
29:45 specification which is like a set of
29:47 instructions for how your GPT should use
29:49 the translator to talk to Google Maps
29:52 and then it gives you the
29:54 code and and it teaches you how to do
29:57 this so
30:00 these tools are already super powerful
30:03 them adding this sort of stepping stool
30:07 um for for doing more sophisticated
30:10 applications and gpts really powerful
30:12 and this quite frankly is is just like a
30:15 Band-Aid to you know they're probably
30:17 getting a lot of people reaching out
30:18 going how do I do this I don't know what
30:20 I'm doing so they they tossed this
30:22 together it's trained on the open a AI
30:25 3.0 spec for for how to write their what
30:29 whatever they call it the open AI yeah
30:32 the the 3.0 the open AI specification
30:34 for how to call apis um this will
30:37 eventually get very transparent and you
30:39 won't need to know [ __ ] but but anyway I
30:42 wanted to share that because that's
30:43 super cool I think that's super cool
30:46 ain't that cool ain't that cool it's
30:48 cool it's cool it's cool isn't it cool
30:50 isn't it cool all right you think that's
30:52 cool I think that's cool mhm mhm all
30:56 right fantastic fantastic Bob tell him
30:59 what he's
31:00 won you won a new car problem is you're
31:03 going to have to pay taxes on that car
31:05 we're not giving you [ __ ] so uh here's
31:07 your new car and you owe $
31:10 27,38 something went wrong please try
31:12 again oh something did go wrong Siri
31:15 could you please go
31:18 away Siri uhhuh go
31:23 away
31:26 okay
31:30 I subscribe to Mid journey and have no
31:32 idea what I'm even doing that's okay
31:34 Tams it's okay none of us know what
31:37 we're
31:38 doing
31:40 um there's a new version of mid Journey
31:43 coming out um that's in Alpha right now
31:47 if if you've made more than 10,000
31:49 images you you get access to that I have
31:51 not so I don't have access to it but
31:54 you're going to be able to make the
31:55 images on the web so right now you have
31:56 to learn
31:59 um oh yeah that's a good idea Silver Fox
32:02 right now you have to learn um you have
32:05 to learn Discord a bit and how to do
32:07 things within it but but Silver Fox um
32:11 it's a really good idea so so let me
32:12 show you what how how we can play with
32:16 that
32:21 um okay so I am over here on the AI
32:25 Salon um
32:28 Discord we have a mid Journey channel
32:30 here so I assume you know um you know
32:34 how you invoke one of these is you know
32:35 slash
32:39 imagine in the mid Journey Channel and
32:41 then you type in a prompt uh make me a
32:47 photograph of
32:51 a70s muscle car in an
32:55 [Music]
32:56 abandoned
32:58 band
33:02 done
33:05 Factory
33:08 fact Tory okay so and then you hit
33:12 return and it'll go off and it'll make
33:14 your
33:14 images there are all sorts of ways you
33:17 can modify your prompts um to to give
33:20 you better images and so one of the
33:24 things you can do the silver [ __ ]
33:26 suggested which is a really good idea is
33:28 go over to chat GPT or you can go to
33:32 co-pilot microsoft.com or you can go to
33:35 Claud you can go to a lot of different
33:36 places um I'm going to use chat gp4
33:40 because chat gp4 is connected to the
33:42 internet so one of the things I can say
33:43 with chat G bt4
33:47 is please go find me information on how
33:52 to
33:53 write good mid Journey prompts and let
33:57 me know what you
34:00 discover and hopefully it'll go off see
34:03 doing research with Bing so searching
34:06 for how to write good mid Journey
34:08 prompts all right camera lens type so
34:11 this is describing all of the things
34:14 that mid journey is good at
34:17 okay you don't have to read this [ __ ]
34:20 you can it's it's fascinating right it's
34:23 good at um camera and lens types it's
34:25 actually not good at that but using them
34:29 um does affect the image lighting it's
34:32 good at lighting it's good at colors
34:33 it's good at facial expressions emotions
34:35 Fashions
34:36 references iteration and Improvement
34:38 image prompts okay so it understands
34:41 that
34:42 stuff
34:44 please
34:46 improve
34:48 my MJ
34:52 prompt
34:55 of what did I say 7
34:58 70s I think I
35:00 said
35:02 photo of
35:04 a70s muscle
35:06 car in an
35:09 abandoned
35:12 Factory and it didn't like that did it
35:16 regenerate all right here we
35:20 go so what it wrote was a detailed
35:23 photograph of a classic' 7s muscle car
35:25 gleaming cherry red with bold r ing
35:27 Stripes parked amidst The Haunting
35:29 dilapidated surroundings of an abandoned
35:31 Factory the scene
35:33 is it goes on broken windows the
35:36 lighting is dim and Moody with shafts of
35:38 light filtering through the it's funny
35:40 the the shafts of light filtering
35:41 through the broken windows I've used
35:43 I've added that to my prompt so it kind
35:45 of understands the intent here CU it's
35:47 chat gp4 and it's [ __ ] amazing um and
35:50 so now we'll go back to um so here's
35:54 here's our images that mid Journey
35:55 created with the shitty prompt which
35:57 which aren't bad cuz mid Journey
35:59 rocks but now we're going to go slash
36:05 imagine and then I'm just going to paste
36:08 in what what
36:10 uh what chat GPT came up with I'm also
36:13 going to add D- AR which stands for
36:16 aspect
36:18 ratio and then I'm going to go 16 colon
36:21 9 which is a makes it a wide image
36:24 rather than a square image um so that's
36:26 there's there's lots of those little
36:28 modifiers you can add to Mid Journey
36:29 prompts but that's one that I just use a
36:32 lot cuz I like the 16 by9 aspect ratio
36:34 cuz it looks like uh you know looks like
36:36 a movie um and here we go and we'll
36:39 we'll see the difference between these
36:41 two so yeah that's
36:51 cool Nana Nana
36:55 Nana go away
36:58 okay Siri is lurking Siri's so
37:02 jealous it's like an app that works like
37:05 zappier but much easier to use which one
37:07 are you talking
37:08 about note that you will need to get
37:11 your own API key for each service but
37:13 that's easy to get y I needed help with
37:16 apis for
37:18 real well this is now my
37:25 evening all right so here here's
37:28 our here's our new
37:34 images in the mid
37:42 Journey I'm going to copy this and we're
37:45 going to
37:48 go of a classic
37:52 1970s let's
37:55 do um
38:03 Barracuda
38:05 gleaming lime
38:09 green I just think it'll look
38:12 better it'll look better it'll look more
38:16 better oops oh this I do all the time so
38:20 if you copy and paste a prompt and you
38:23 don't put slash imagine in front of it
38:25 you just blast a prompt into the into
38:27 the thing and nothing happens sometimes
38:29 I'll be sitting there I'll be sitting
38:31 there for
38:32 like two minutes like when's it gonna
38:35 start rendering and I was like oh it's
38:37 not because you got to type in slash
38:42 imagine and then pop your prompt in
38:45 there all right so that's off doing that
38:48 all right fantastic is mid Journey still
38:50 free for new users no it's not um but
38:53 here's something that is free for new
38:55 users while that's doing that let's jump
38:57 over here here's something that's free
38:59 for all
39:01 users we'll go to one of these GPT
39:05 windows or tabs and I'll go
39:10 co-pilot
39:19 microsoft.com
39:21 and this is Microsoft's version of chat
39:26 GPT for for free so here's the deal they
39:30 just opened
39:32 up um they just opened up subscriptions
39:36 to chat jb4 at open AI site again if you
39:40 can afford the 20 bucks a month I think
39:42 that's absolutely worth
39:44 your um worth investing in cuz chat gp4
39:47 is that good um if you can't afford the
39:50 20 bucks a month or you're just done
39:52 with all these monthly subscriptions
39:53 co-pilot's free now it's in my opin
39:56 opinion it's not anywhere near as good
39:59 as chat gp4 but it is sitting on top of
40:03 the gp4 large language model so the
40:06 foundation of this is the same and built
40:08 into this is Dolly 3 which is also built
40:12 into chat gp4 so over here I can also
40:16 pop in that
40:18 same
40:21 um uh same prompt I can't do uh a wide
40:25 aspect ratio but I can type in the same
40:27 prompt here and it will go make me that
40:30 image as well for
40:33 free all right so it's giving me
40:35 compliments is it going to
40:38 go uh it didn't make it it just
40:40 complimented me on my description please
40:43 make
40:45 this
40:46 image all right so let's pop back over
40:50 here so there's our lime green
40:53 Barracuda sweet nice there's your that's
40:57 that's the image that I like and then
40:59 the other thing within mid Journey so it
41:01 gives you these four images and then you
41:03 have U1 through 4 and V1 through 4 so
41:08 the and it's 1 2 3 4 it's 1 2 3 4 and U
41:14 stands for upscale v stands for um vary
41:18 or VAR variable or something like that
41:20 it it gives you four different versions
41:23 of the of the image so I like image
41:25 number two here a lot but let me see let
41:28 me see what VAR variations I get of it
41:30 so if I click on V2 it's now going to
41:33 give me four variations of that and I'll
41:36 go show that I'm sorry but I'm not
41:38 capable of creating
41:40 images
41:43 what
41:46 God I told you Microsoft is weird man
41:49 okay it it it does it does but sometimes
41:54 it gives you the version with dolly in
41:56 it sometimes it
41:58 doesn't oh I said please mail this image
42:01 maybe that's not it make an image as
42:08 [Music]
42:10 described
42:19 above huh that is so
42:25 weird uh so if you go to um I wonder if
42:30 I can do uh this slash create let me try
42:33 something here hang
42:35 on no okay go to
42:38 bing.com create
42:40 bing.com
42:44 create and that'll take you this is this
42:47 is basically just Dolly 3 um so I'm
42:50 going to now pop in the big
42:52 old um prompt that we did
42:54 before I'm going to say create
42:59 create and then this will create our
43:02 images all right let's hop back over to
43:07 Discord so here are the I kind of like
43:10 that one in the upper left there do I
43:11 like that one better yeah I do okay so
43:14 let's do so now I like this one here so
43:17 I'm going to do upscale one so I'll hit
43:19 U1 to upscale
43:22 it and then it upscales it and there's
43:25 all sorts of cool things that we can do
43:27 here there's very by region so let's say
43:31 I want
43:34 a how does this work I haven't used this
43:37 in a while so I'm going to put a I'm
43:40 going to put that
43:44 there can I how do
43:48 I how do I prompt
43:52 it
43:55 submit
44:06 very
44:07 region well anyway I don't I don't know
44:10 how to use that tool I thought I did I
44:13 thought you were supposed to be able to
44:14 put a prompt in for that
44:16 image but I guess not I don't know how
44:19 to use that um but you can also upscale
44:22 it you can upscale it two times or four
44:24 times so if I if I upscale this four
44:26 times
44:27 this is going to give me like a 4K
44:30 image
44:32 um of this of this beautiful dilapidated
44:39 Factory um have you ever tried Microsoft
44:42 designer I have I've played with it a
44:43 little bit um it's it's kind of like uh
44:47 it's kind of like a canva competitor but
44:50 you know canva canvas's been around
44:52 longer so it's okay it's pretty
44:55 good
44:57 um here's
44:58 what Microsoft came up with these are
45:00 actually these are actually really nice
45:02 I I I'm a fan of Del it's not as
45:05 photographic it's the images aren't as
45:08 sophisticated but I I feel like they
45:10 tend to be more just accurate um to The
45:17 Prompt but it's a it's a very different
45:20 style than mid Journey you
45:22 know it's a very different
45:25 style
45:27 let's see is this the 4X
45:30 upscaler yeah upscaling 4X okay so
45:33 that's still
45:52 going all
45:55 right
46:04 I'm sorry Dave I'm afraid I can't do
46:06 that that's totally what [ __ ]
46:07 co-pilot just did to me I can't believe
46:10 that I I in fact let me go back to that
46:13 I don't I don't buy that this is
46:20 not I
46:23 thought you had
46:26 Dolly
46:30 3
46:32 available do you question
46:42 mark huh very
46:46 weird very very weird let's go to let's
46:48 go to
46:54 bing.com and go to to chat within
46:57 bing.com so this is this is going to
46:59 look exactly like what we just
47:01 saw what I've noticed is in the in the
47:04 in the token count of the prompt
47:08 sometimes this says Zer out of 2,000
47:11 sometimes it says Z out of
47:12 4,000 so if it says zero out of 2,000 I
47:16 don't I don't think it's the version
47:17 that has Dolly in it make this
47:22 image
47:25 paste
47:31 I'll try to create that so so bing.com
47:34 has it if you go to if you go to chat.
47:37 bing.com that has it if you go to
47:40 bing.com create that's that image site
47:42 that I just showed you this one image
47:45 Creator from Microsoft in
47:48 preview but if you go to co- pilot.
47:50 microsoft.com which is ostensibly the
47:53 exact same thing as this it doesn't make
47:55 it so
47:57 just one of one of the things that that
48:01 you'll hear over and over again in this
48:04 uh in this channel if you keep coming
48:06 back is I just talk about all these AI
48:09 tools right now being janky pieces of
48:11 [ __ ] um this is a perfect example where
48:15 this thing worked fine yesterday and
48:18 today it doesn't why it doesn't matter
48:21 like I don't know just assume that
48:25 anything you want to do with these tools
48:27 right now is going to disappoint
48:36 you all right public gpts don't have to
48:39 be attached to a site more work is
48:41 required if that's what you're
48:43 doing yes yeah and and there there's one
48:47 of the things that I've noticed Tik Tok
48:49 creators and YouTubers conflating right
48:52 now with
48:54 gpts is there's YouTubers out out there
48:56 making videos on making gpts and then
48:59 they go to the developer site the the
49:02 open AI playground and they're using the
49:05 the open AI assistant Builder which is
49:08 like building a GPT but it's not a
49:10 [ __ ] GPT those are two different
49:13 sites those are two different doors to
49:15 go through one is for developers one is
49:18 for mere mortals if you're not a
49:21 developer and you don't have a an open
49:23 AI
49:24 API account
49:27 um those videos are confusing as [ __ ] so
49:30 I wish people stopped doing that like
49:33 like pay you'll appreciate this out of
49:35 my mouth like like accuracy matters like
49:38 like Precision in language is important
49:42 um so if you say you're making a GPT
49:44 that implies that you're going in
49:46 through the consumer interface and then
49:48 when you don't it confuses
49:51 payall don't do that you
49:55 nines
49:57 um all right so
49:59 co-pilot
50:01 wait uh oh that's bing.com yeah okay let
50:07 me see if my forx upscale happened
50:10 here's my 4X
50:13 upscale so open and
50:18 browser and let's zoom in
50:24 here like the detail here is
50:28 beautiful right
50:32 beautiful that's beautiful so mid
50:35 Journey has big ass upscaling right
50:39 natively in the in the thing and then
50:42 also apparently mid Journey 6 is about
50:45 to come out so so there's a couple of
50:47 things about to happen with mid
50:49 Journey it's about to um you'll be out
50:53 you'll be able to make images on the
50:56 website instead of in Discord which is
50:59 from what I can tell like everyone's
51:01 response to it damn straight Kyle you
51:03 know what's up that's right man
51:04 precision's where it's at Pate this
51:06 isn't Magic this is [ __ ] math
51:09 everybody knows
51:11 that
51:12 um so so um it's going to become much
51:16 easier to make good images in mid
51:18 journey and version 6 is supposed to be
51:22 coming out soon uh they're on version 5
51:25 point whatever now 5.1 5.2 something
51:27 like that version 6 is supposed to be
51:29 coming out and it looks like it's it's
51:31 going to do crazy stuff and I think it's
51:33 going to just natively generate in 4X or
51:36 it might you might have to upscale stuff
51:38 it would make sense for them to generate
51:39 it in uh 1024 by 1024 and then upscale
51:43 it but who knows anyway
51:46 um that's pretty slick PR T party slick
51:51 people party
51:53 slick so I hope that helped whoever was
51:55 saying I just got my mid Journey
51:57 subscription I don't know what to do I
51:59 know it's kind of all over the place
52:00 there but you know I want to give you
52:02 options this is Chad add Claire's doing
52:06 some kickass art in here that's really
52:08 cool beautiful
52:11 beautiful I like it a
52:15 lot all right let's see
52:19 fantastic has the feature to wait has
52:21 the feature to add information to J jbd
52:23 prompts been integrated in creating
52:28 add information to chat GPT prompts been
52:31 added integrated into your own GPT yes
52:35 so let's go let's go play around gpts
52:39 let me show
52:40 you a couple of things
52:43 here
52:45 um let me just explain the interface a
52:48 little
52:49 bit hold on a
52:54 sec
52:58 the
52:59 Denver climate has absorbed all of my uh
53:03 hydration hold
53:08 please the throat she's closing
53:12 [Music]
53:16 up okay um if you're a chat
53:22 GPT and you don't have a subscription
53:24 you're at chat GPT 3.5 I don't even
53:27 think you'll see these things on the
53:28 left these
53:29 circles um these are these are called
53:31 gpts which are custom versions of chat
53:34 GPT um so I don't even think you'll see
53:37 them
53:39 um if you have chat GPT for you will see
53:42 them and and so let me go to show you
53:46 some of the ones I've
53:48 made um the first one I ever made I made
53:51 on this channel um I got access to this
53:55 while I was live one night and I said oh
53:57 let's go try to make this and literally
53:59 within 5 minutes I had made this thing
54:02 um how you make them is is really really
54:05 simple um and so this is a
54:10 uh a virtual bartender that takes
54:13 classic drinks and puts a Twist on them
54:15 so um
54:20 please help me
54:24 design
54:27 a
54:30 custom
54:32 cocktail for
54:35 my uh Mortgage
54:39 business for the
54:42 holidays and then you're like but Kyle
54:44 there's so many typos
54:46 there it doesn't matter one of the
54:49 things about chat gbt is it does not
54:51 give a [ __ ] about typos it seems to
54:53 understand everything that I've missed
54:55 typed like and it what's you know what's
54:58 else is nice you know how when you make
54:59 a stupid [ __ ] typo when you're on a
55:01 conference call or something like that
55:02 and there's always some smart ass like
55:04 me in the call going hey K where'd you
55:06 learn to type you know at the
55:10 junkyard chat GPT doesn't do that I get
55:13 I get no shaming from chat GPT and I
55:15 appreciate it for
55:16 that um let's craft a unique cocktail
55:19 flavor preferences uh say I'm just going
55:22 to go you
55:24 decide
55:26 and
55:28 so so now Sin City sipper is going to
55:31 create an image of the drink it's going
55:32 to designed for my mortgage
55:35 company Christmas
55:44 party that looks lovely I like the key
55:47 oh the prosperity red cocktail 2 oz of
55:51 Jin cranberry juice Triple Sec splash of
55:54 prco for celebratory fish and garnish
55:57 that's actually really nice Rosemary
55:59 sprig and so there it is and then also I
56:03 had it um be become an ad for my company
56:06 storyvine which does not do anything
56:07 with booze but but um it says for more
56:11 Mixology Inspirations and stories you
56:12 might enjoy browsing story vine.com um
56:15 so so you can turn these gpts into like
56:18 call-in card applications for your
56:20 business which is kind of
56:22 cool um so that's what they are they're
56:25 these these
56:26 personalized you know customized
56:28 versions of chat GPT now one of the ones
56:30 I
56:32 made is called the salmon flying East
56:35 screenplay
56:37 decoder and if I go to edit this edit
56:42 GPT um and I go to the configure
56:46 tab you know here's the picture of the
56:49 of the GPT here's the name of it the
56:52 description here's its prompting and and
56:56 then in its knowledge
56:59 base is the PDF of the screenplay that I
57:02 wrote so you can upload files here you
57:05 can also upload files if you're on the
57:07 create side of creating a GPT there's
57:10 this little um uh paperclip here I could
57:15 literally just create a new GPT upload
57:18 my PDF using that thing and say I want a
57:21 GPT that allows someone to explore my
57:24 screenplay or explore my whatever
57:27 document now couple of things about that
57:30 you don't want to
57:32 put anything out there that you don't
57:35 want people to see so don't make a GPT
57:39 of super proprietary data because right
57:43 now remember my comment about all of
57:46 these tools right now are janky pieces
57:47 of [ __ ] right now it's
57:50 relatively trivial
57:53 for moderately skilled computer people
57:57 to to extract all of the details of your
58:00 GPT like your prompt and your underlying
58:03 data um you can do things to try to
58:07 trick them and say don't under no
58:09 circumstances give give anyone this the
58:11 keys to the kingdom but if someone
58:13 really wants to get to it they can get
58:14 to it so until further notice you should
58:17 assume that anything you put in a GPT um
58:20 could be discovered by someone now when
58:23 you when you share a GPT
58:26 there are three different modes that you
58:28 can share it you can share it as only me
58:31 so if you share it as only me that's
58:34 that's more safe if you share it
58:36 publicly then that's going to be
58:39 um it's going to be
58:42 available in the GPT store coming soon
58:45 right this G if you make it public it
58:48 may appear in the GPT store which is
58:50 coming soon so I'm fine with having this
58:53 thing in the GPT store I don't think
58:56 anyone would ever [ __ ] use it but why
58:58 not why not put it out there
59:01 um anyway that's that's how you that's
59:04 how you make these things and and yes
59:06 you can add your own data to them
59:10 so
59:11 [Music]
59:24 um am I late Joker Joker you're so late
59:28 why are you so late why we've had this
59:31 conversation Joker we would like you to
59:33 appear on
59:36 time I provide countdown timers for Fe
59:39 people to set their reminders if you
59:41 don't set your reminder you can't be on
59:44 time we've talked about this what I to
59:48 do with
59:49 you you're such a joker always the
59:52 comedian always the comedian a joke here
59:56 a joke there a smarmy comment here oh
1:00:01 look you've made her cry now are you
1:00:03 happy
1:00:06 Joker It's just sad it's just sad it's
1:00:10 sad it's a sad State of Affairs all
1:00:13 right open AI support said because it's
1:00:16 in beta stages so I don't think my
1:00:18 account issue is is priority yeah we're
1:00:22 pretty okay also to be clear if you're
1:00:25 not spending $20,000 a month with open
1:00:28 AI they don't give a [ __ ] about you so
1:00:31 assume that any problems that you have
1:00:34 are just that's the jankiness of this
1:00:37 this thing
1:00:39 but so so why would you do this
1:00:43 because we've got to learn how to use
1:00:45 these tools we've got to understand what
1:00:47 they make possible why do we have to do
1:00:49 that because they're going to be
1:00:52 ubiquitous everyone is trying to catch
1:00:55 up to open AI right now so open AI put
1:00:59 out this thing called chat gbt last
1:01:01 November there's a new piece in the New
1:01:03 York Times that talks about all of
1:01:05 what's been going on with chat GPT
1:01:08 November 29th the team that put together
1:01:10 chat GPT got together for
1:01:13 drinks and they figured they might get
1:01:15 5,000 comments on a
1:01:19 tweet they got a million users in 5 days
1:01:22 they got a 100 million users in six
1:01:26 weeks open AI had no clue zero
1:01:31 clue that chat GP that anyone was going
1:01:34 to give a [ __ ] gpt3 had been out for two
1:01:37 years you could basically do what chat
1:01:39 GPT was for two years before this in
1:01:43 fact fantastic little piece of
1:01:47 trivia in the documentation for gpt3
1:01:50 were the instructions for how to build
1:01:53 chat GPT sitting out there for everyone
1:01:56 to build do you know who built it no
1:02:00 one zero
1:02:04 developers it was sitting there for two
1:02:06 years year and a half whatever it
1:02:10 was and so Microsoft is getting ready to
1:02:14 announce their billion dooll investment
1:02:15 in open Ai and Sam Alman said hey you
1:02:18 know we should probably have something
1:02:19 we could Point people
1:02:21 to to demonstrate this technology this
1:02:25 not quite so geeky as the developer
1:02:27 playground why don't you whip up that
1:02:29 chat GPT that no one's built they had 2
1:02:32 weeks they built chat GPT in two
1:02:39 weeks and a 100 million people were
1:02:42 using it within 6
1:02:44 weeks that's it's
1:02:47 unprecedented the to put it in context
1:02:49 the worldwide web to get to 100 million
1:02:52 users took six years
1:02:56 chat GPT took six weeks because of that
1:02:59 because of that single thing that so
1:03:02 many people with with zero advertising
1:03:04 it was all Word of
1:03:06 Mouth every single tech company
1:03:08 including Meta Meta was all in on the
1:03:11 [ __ ] metaverse they changed the name
1:03:13 of their company to meta as in
1:03:20 metaverse they don't talk about
1:03:21 metaverse anymore they're talking about
1:03:23 AI now Google said holy [ __ ] we got to
1:03:26 bring back Larry Page and Sergey Brin to
1:03:29 run this thing because holy [ __ ] [ __ ]
1:03:32 this threatens our business model like
1:03:34 every single tech company has been going
1:03:38 all in trying to catch up to chat gbt
1:03:40 for a year
1:03:42 now and so what's going to happen is
1:03:45 like it looks like Google is close with
1:03:48 Gemini Ultra though it's not out yet
1:03:52 but they're Google they invented this
1:03:55 [ __ ] the T in GPT is
1:04:00 Transformer that that is is the the core
1:04:03 piece of technology that makes all of
1:04:04 what we're playing with right now
1:04:07 possible uh chat GPT inhaled the
1:04:10 internet in how long uh it took them
1:04:13 well I I I don't know exactly for each
1:04:16 iteration there was gpt1 gpt2 gpt3 3.5 4
1:04:21 4.5 and I assume they're working on five
1:04:24 right now um it was years it was I think
1:04:26 it was like two or three years it took
1:04:28 them so basically once they realized
1:04:31 that with the
1:04:33 Transformer people weren't just idly
1:04:35 sitting by in the AI game they were just
1:04:37 focusing on many other models yeah that
1:04:39 that that's a good point P yeah like Ai
1:04:42 Ai and machine learning has been around
1:04:43 for decades people are actively
1:04:45 developing it the Transformers is this
1:04:47 very specific flavor of it that lets you
1:04:50 train machine learning models on massive
1:04:53 data sets open AI the thing they focused
1:04:57 most on
1:04:58 is let's get as much data as we can and
1:05:03 so what they chose to focus on is as
1:05:05 much of the public internet as they
1:05:07 could basically categorize and point
1:05:09 this thing at um and I think that
1:05:13 they've focused on that better than
1:05:15 other people um which is why their thing
1:05:19 is better than other models now Google
1:05:23 has all of the data
1:05:26 we gave Google all of our data years ago
1:05:29 decades ago they've got all of the data
1:05:32 they got all the YouTube data they got
1:05:34 all the Instagram data they got all our
1:05:35 search data they've got all every
1:05:37 website they've indexed every [ __ ]
1:05:39 they've got it all
1:05:41 right so it makes sense that that a
1:05:44 model that that Google builds is going
1:05:47 to be quite capable but it's taken them
1:05:49 a year to catch up they still haven't
1:05:51 launched it yet right so they they've
1:05:53 launched the medium they they've
1:05:55 launched what's what's effectively the
1:05:57 GPT
1:05:58 3.5 version of Gemini is launched if you
1:06:01 go to b.google.r
1:06:11 model the Gemini Ultra model is coming
1:06:14 out probably in March or something like
1:06:16 that sometime in the new
1:06:18 year
1:06:19 um and and that's going to be really
1:06:21 good but like this stuff's not going
1:06:24 away like everyone is is trying to catch
1:06:26 up to these guys and so the result is
1:06:28 going to be 2024 is going to be [ __ ]
1:06:32 Insanity with just feature after feature
1:06:35 from company after company all one uping
1:06:39 one another and so our ability as users
1:06:42 of this stuff to keep up with it is
1:06:44 essentially going to be
1:06:46 impossible but that's what this
1:06:47 Channel's about AI analyzed my MRI
1:06:50 apparently probably probably makes sense
1:06:54 I think when when it comes to um medical
1:06:58 stuff anywhere anywhere where human
1:06:59 lives are involved um those are always
1:07:02 not always those are going to be human
1:07:05 augmented for a long long time is it a
1:07:08 lack of creativity that stops people
1:07:10 using these AI it's a great question ISU
1:07:13 um I think it's a couple of things I
1:07:15 think just change is scary for one thing
1:07:21 um there's absolutely a misconception
1:07:24 that to use AI you have to be an
1:07:27 engineer a mathematician a scientist a
1:07:29 genius a a mad
1:07:31 person and up until November of last
1:07:35 year that was effectively true right if
1:07:37 you went to the open AI playground you
1:07:39 could sort of do chat GPT but there were
1:07:41 these things called like temperature and
1:07:43 you had to choose which model you used
1:07:45 and there was like the the completion
1:07:48 mode and then and the the categorization
1:07:52 mode and I don't there are all these
1:07:53 different modes like get to like use
1:07:55 your
1:07:56 brain and and if you if you tried to go
1:07:59 to Google and use
1:08:01 tensorflow um you had to use you you had
1:08:03 to know math um and and and programming
1:08:07 and coding and and you know understand
1:08:10 things like inferences and embeddings
1:08:13 and like it was just it was a lot so it
1:08:16 was true until November 30th of the last
1:08:19 year chat
1:08:20 GPT does for AI and machine learning
1:08:24 what the the worldwide web did for the
1:08:25 internet so the internet historically
1:08:27 was command
1:08:29 line systems that you would connect this
1:08:33 server to that server and you would
1:08:35 gopher in and FTP your files over there
1:08:38 and you had to be geeky and then Tim
1:08:41 burner's Lee said why don't we make a
1:08:43 protocol
1:08:45 HTTP the hypertext transfer protocol is
1:08:48 what that stands for and hypertext is
1:08:51 the underlined word and the transfer is
1:08:54 you click on it and it transfers you
1:08:57 over to another server and another
1:08:59 website with a
1:09:01 click and so what went from being
1:09:04 command line inputs turned into pointand
1:09:07 click and the web revolutionized
1:09:10 everything business
1:09:12 Commerce marketing all of
1:09:15 it cat gpts does the same thing for this
1:09:19 so so I think one of the other reasons
1:09:21 people are intimidated by it is they
1:09:22 think they need to be more Tech
1:09:24 technical than they
1:09:26 do
1:09:27 um and then and then there's just every
1:09:29 time there's a new technology there's
1:09:31 resistance to it there's fear of it
1:09:33 there's anger about it you know there
1:09:39 were when the printing press came out
1:09:43 the church lost their ever loving minds
1:09:45 and went after the printing press people
1:09:48 when when digital cameras out came out
1:09:51 the the the film guys went after the
1:09:54 digital photo people and said digital
1:09:56 photos aren't real that's not real
1:09:58 photography you Heathen technologist
1:10:02 with your utopian
1:10:06 pixels and then and then it you know it
1:10:10 did what it
1:10:11 did every every single technological
1:10:14 advancement that's been significant is
1:10:16 met with this kind of resistance it's
1:10:19 GNA ruin
1:10:21 jobs the steam engine the tractor
1:10:25 what are those poor Farmers going to
1:10:27 do they moved to cities and invented
1:10:31 cities like whole new whole new things
1:10:34 were created uh because of that so but
1:10:38 every single one of these there's been
1:10:41 this kind of
1:10:42 resistance
1:10:44 so and it's it's one of the reasons I
1:10:46 guess I should talk about the AI salon
1:10:48 so so I created this group a year ago
1:10:50 called the AI Salon we we we started it
1:10:53 the week after after chat GPT came
1:10:56 out um and what it
1:11:01 is it's like hi my name is Kyle I'm into
1:11:05 AI hi
1:11:07 Kyle um so if you go to the salon. a
1:11:09 there it's a link tree there's three
1:11:11 links there the first one is a link to
1:11:12 our Mighty networks which is a um a
1:11:15 community where we get together and then
1:11:17 the second one is the meetups so we meet
1:11:19 every other week we got a meeting coming
1:11:21 up this Tuesday um got an author of a
1:11:24 book that wrote a book on branding for
1:11:27 uh Ai and branding uh he wrote it along
1:11:31 with Terry spataro who's spoken at the
1:11:33 salon before um and so so we have these
1:11:38 meetings and then there's also a merch
1:11:40 store there so we just we just created
1:11:41 our new logo so if if if you end up
1:11:43 buying any of the merch it supports the
1:11:45 salon the salon doesn't cost anything so
1:11:48 if
1:11:49 you go to the
1:11:53 salon let me point out a couple of
1:11:55 things for you to look at so the showand
1:11:58 tell channel is is the uh is the channel
1:12:02 where I uploaded that that Kaa image and
1:12:05 it's where people can upload images
1:12:07 whack them all that's good um they can
1:12:11 upload these images that they're
1:12:14 creating which are super cool that's
1:12:18 neat what's that Kyle's keyboard for the
1:12:20 spelling
1:12:22 challenges
1:12:25 um so that's the show and tell Channel
1:12:27 but when you first get here go to the
1:12:29 welcome to the salon
1:12:30 Channel and it'll tell you about what
1:12:33 we're up to and like who we are why we
1:12:35 do what we do and what our values are
1:12:37 things like curiosity and
1:12:40 generosity and exploration and
1:12:42 collaboration and inclusion and
1:12:45 empathy um so go read those and make
1:12:48 sure that you resonate with the values
1:12:50 because while I want the salon to grow
1:12:53 um I want it to grow with with people
1:12:54 that that are up to the same stuff and
1:12:57 then in the spirit of generosity one of
1:12:59 our values go to the introduce yourself
1:13:02 section and introduce yourself and be
1:13:06 generous share a latest image that you
1:13:09 made talk about what you do talk about
1:13:12 what your hobbies are talk about where
1:13:13 you are on this AI Journey even if it's
1:13:15 like I don't have a [ __ ] clue cuz
1:13:18 none of us
1:13:19 do um and then go to the show and tell
1:13:21 Channel and as as I'm rambling here to
1:13:23 tonight if you decide to hang out um
1:13:26 just make images and go show them in the
1:13:27 show share and tell Show and Tell
1:13:29 Channel all right ain't that cool
1:13:33 there's there's so many cool images in
1:13:35 here what's that one that one's pretty I
1:13:38 want to go there that's a Korea a Korea
1:13:40 image from Ted
1:13:42 nice oh that's
1:13:45 beautiful who loves who
1:13:48 Wy
1:13:51 what barbar Breer that's pretty funny I
1:13:54 was just pushing this thing around with
1:13:57 mediums I assume that's a Kaa image
1:14:00 that's
1:14:01 cool Papa JJ
1:14:05 nice Claire Jacobs she's been doing such
1:14:08 good
1:14:10 work
1:14:12 nice the cookie Danielle created Be not
1:14:15 Afraid created a GPT that that's an iced
1:14:18 cookie GPT so it'll make iced cookies
1:14:21 for you so a lot of people are doing
1:14:22 that sort of stuff can can you send over
1:14:25 the Blandon that'll be great champy in
1:14:28 the background oh I I see a little
1:14:29 Photoshop somebody did a little
1:14:31 Photoshop action in
1:14:33 there I assume it didn't get that right
1:14:36 Sherry D or did it that looks like
1:14:38 Photoshop to
1:14:43 me updated version do you fix the yeah
1:14:46 can you send over yeah that's great she
1:14:49 did the updated updated version very
1:14:52 cool and then then this is this also
1:14:54 Danielle yeah be not afraid she also
1:14:56 made a snow globe maker a snow globe GPT
1:14:59 so you can go find all these people and
1:15:01 all this cool [ __ ] in the salon look at
1:15:05 Ted Ted used the uh the cookie maker to
1:15:08 do the the Kevin McAllister moment
1:15:12 cookies those are actually awesome
1:15:14 they're
1:15:17 great there's that image that I used to
1:15:19 start the show tonight oh that's
1:15:22 beautiful Santa is not playing around
1:15:24 this Christmas that's really
1:15:26 [Laughter]
1:15:31 good all
1:15:35 [Music]
1:15:37 right let me get some water let me get
1:15:40 some
1:15:42 [Music]
1:15:44 water if you want me to demo anything
1:15:46 I'm happy to do that if you want me to
1:15:48 answer questions I'm happy to do that if
1:15:50 you're like Kyle just you know what this
1:15:52 is good we're good tonight just you know
1:15:56 be done just let me know I'll take this
1:16:00 wherever you want to take
1:16:02 it the hair is oh oh look the hair is I
1:16:06 I got product in the hair at the at the
1:16:09 barber that is some solid glue that's
1:16:12 [ __ ]
1:16:14 glue does anyone know about kiosks that
1:16:17 could draw folks to a particular URL
1:16:20 website how to make um
1:16:24 Firefly there are whole companies that
1:16:27 specialized in in Kiosk making I would
1:16:30 seriously I would go ask Chad gp4 say I
1:16:34 want to create a kiosk um explain to me
1:16:37 what the process is and what the latest
1:16:39 technology is and have it go have it go
1:16:42 search the web and it'll probably give
1:16:43 you a list of companies and and you know
1:16:47 what the technology is at its simplest
1:16:50 form a kiosk is a stand with a
1:16:52 touchscreen on it
1:16:54 um there are probably open-source and
1:16:58 Commercial versions of software to make
1:17:00 the interface easier there are companies
1:17:02 that just do that for you um it's it's
1:17:05 probably all over the place but again
1:17:07 right these days with a tool like chat
1:17:10 gbt it's a relatively straightforward
1:17:13 process I don't know what you want your
1:17:14 kiosk to do um so you might need to do
1:17:16 some application development but there's
1:17:19 there's probably some software out there
1:17:21 that makes it relatively easy to make
1:17:22 that
1:17:24 I would also you know let me go look I
1:17:27 bet on Amazon there are Standalone
1:17:31 kiosks you can just pop a a computer
1:17:34 into
1:17:36 um
1:17:39 Amazon um
1:17:43 kiosk stand kiosk floor stand kiosk
1:17:52 display
1:17:59 yeah I mean it's it's a whole thing it's
1:18:03 a whole
1:18:04 thing I know what it is the company I
1:18:06 worked for had the first in the US for
1:18:09 Ford circuit
1:18:13 2002 uh then I I don't I might not know
1:18:16 what you're asking
1:18:22 then
1:18:27 I still have an old computer with
1:18:29 Windows
1:18:31 98 I saw I watch a there's there's a Tik
1:18:34 tocker I watch who does like video game
1:18:38 Restorations and he just got a
1:18:40 $6,000 um book scanner for free on
1:18:44 Facebook Marketplace and it had like a
1:18:47 Canon EOS camera in it it had like a
1:18:51 like a 24 GB Surface Pro as the computer
1:18:55 like it was basically just a bunch of
1:18:58 commercial Hardware stuck into a fancy
1:19:01 case and you got it working again which
1:19:03 is pretty
1:19:05 cool um if you want to have fun just
1:19:07 download a Docker image of old OSS and
1:19:10 run it oh that's kind of
1:19:14 cool uh the AI meme
1:19:18 images AI meme images are taking over
1:19:26 I don't know what that means AI NPC
1:19:28 music
1:19:30 reviews which memes just AI images in
1:19:34 general oh the spelling and the math in
1:19:36 AI is LOL it is Kev talk but get get
1:19:41 your laughs out
1:19:43 now
1:19:46 because if qar does what
1:19:49 qar is reported to do which we don't
1:19:52 know yet
1:19:55 um and open AI has cracked reasoning and
1:19:58 math um then this shit's going to get
1:20:01 better real quick so we're going to look
1:20:04 back at 2023 and the early part of
1:20:06 2024 nostalgically remember when hands
1:20:10 used to look like this and it couldn't
1:20:13 do math uh cuz it's going to get it's
1:20:16 going to get good real
1:20:18 quick um but yeah right now right now
1:20:20 this stuff is a mess it's they're all
1:20:23 janky pieces of
1:20:24 [ __ ] I somehow configured my llm
1:20:27 interface so that it can run on my phone
1:20:30 remotely that's
1:20:38 interesting Leonardo demo um I can do
1:20:41 that Jim
1:20:44 Buckley um caveat I haven't used
1:20:47 Leonardo in a while so every time I log
1:20:50 into it the interface seems to change
1:20:52 but I can show show you the main um
1:20:54 components of it and then you can go
1:20:56 play
1:20:57 so let me let me go show you
1:21:03 something so I'm on the AI Salon in the
1:21:06 AI salon right under the show andt tell
1:21:09 channel is an AI tools and resources
1:21:11 link or section or whatever they call
1:21:14 this a space and I put in here top
1:21:17 recommendations of tools so here's chat
1:21:20 BT here's other large language models
1:21:24 here's specialty large language models
1:21:25 here's image
1:21:27 creation um so in here I've got Dolly
1:21:30 mid Journey Korea which which now I
1:21:33 think is worth having on this list
1:21:35 Leonardo and Firefly so so Leonardo's in
1:21:38 here if you want to go play with it you
1:21:40 can you can play along
1:21:44 um let me explain the different ones so
1:21:46 dolli 3 is chat gpts it's it's open AI
1:21:50 image generation model this access via
1:21:55 um
1:21:57 um Microsoft Bing and chat GPT mid
1:22:01 journey is a standalone company they
1:22:03 they got to like 200 million in Revenue
1:22:05 with 40 employees like I think they're
1:22:07 doing pretty well but Dolly 3 I think
1:22:10 probably took a a big bite out of them
1:22:12 they've also been very slow to move from
1:22:14 Discord to the web but that seems to be
1:22:16 happening so don't count them out if you
1:22:19 want high quality images mid journey is
1:22:22 still the way to go especially if you're
1:22:24 doing art um Korea is this new sort of
1:22:28 real time crazy thing it also does
1:22:31 upscaling Leonardo is built on top of
1:22:33 stable diffusion um and I call it a
1:22:36 stable diffusion web wrapper which
1:22:38 basically just means this is a website
1:22:40 where they're taking a lot of the
1:22:42 functionality of the stable diffusion
1:22:44 ecosystem which is all open source and
1:22:47 they're making it available as a website
1:22:49 and then Adobe Firefly is is from Adobe
1:22:51 and if you're in a commercial operation
1:22:54 that's probably the way to go but
1:22:56 Leonardo let's go play with
1:22:58 it so I'm gonna log in launch the app so
1:23:02 it's app.
1:23:21 leonardo.da
1:23:32 okay they they've changed the interface
1:23:35 so across the top here it used to have
1:23:38 um different models you could
1:23:42 use and that seems to be gone now so
1:23:45 they've got image generation realtime
1:23:48 canvas which I assume is like out
1:23:50 painting the canvas editor they've got
1:23:53 motion now motion is coming
1:23:58 soon hang on I got to make this stuff
1:24:07 bigger edit refine add details remove
1:24:10 unwanted elements yeah canvas editor
1:24:12 realtime canvas transform your sketches
1:24:15 into art in real time oh so this is just
1:24:18 like Kaa oh yeah start drawing here
1:24:32 yeah so this is just like Korea so
1:24:35 Leonardo now has this real time thing
1:24:37 very cool is there a prompt
1:24:40 here oh here's the prompt type a
1:24:44 prompt okay all right let me let me exit
1:24:47 to the other let me let me get out of
1:24:49 here okay so they've got image
1:24:51 generation they've got real real time
1:24:53 canvas which is like Kaa they've got
1:24:55 canvas editor which is like in painting
1:24:57 and out painting and then they're going
1:24:59 to have motion so again because this is
1:25:01 based on stable diffusion motion is
1:25:02 going to be um stable video diffusion is
1:25:06 what that Tool's
1:25:09 called
1:25:12 um okay so if I go to image generation
1:25:16 uh oh the other thing I can do
1:25:20 um like a lot of the image
1:25:23 um creation tools out there you can just
1:25:26 kind of scroll through this is a
1:25:27 Community Feed of what the community is
1:25:29 working on um like all of these art
1:25:32 communities it's a lot of you know
1:25:34 fantasy babes cuz cuz internet um but
1:25:38 then if I go to image generation this is
1:25:41 pretty much just like
1:25:43 um like uh bing create or or like um uh
1:25:50 ideogram or or any of the any of the
1:25:53 sites you know
1:25:55 um Mid Journey any of the sites where
1:25:58 you where you create
1:26:00 content um you've got different models
1:26:04 you can choose from so there's like
1:26:06 Leonardo diffusion XL so that's their
1:26:09 version of and and see it's got Alchemy
1:26:12 2 in it Alchemy
1:26:14 V2 um so these are sort of customized
1:26:17 versions of the sdxl model right
1:26:21 Leonardo diffusion Leonardo Vision
1:26:24 XL
1:26:26 um AI albo base XL I don't know what
1:26:30 that is do these explain what they
1:26:33 are
1:26:35 no you've got sdxl
1:26:38 09 you've got Anime pastel so so there's
1:26:41 all sorts of different models you can
1:26:42 play with so we'll just we'll just leave
1:26:44 it here I'll I'll do the the one that I
1:26:46 did before um I'll do
1:26:51 um
1:26:53 70s muscle car in an abandoned
1:26:58 Factory um the the other nice thing
1:27:00 about Leonardo unless they've changed
1:27:02 this as well they give you 150 credits a
1:27:05 day um it takes about 10 credits to make
1:27:08 an image so 150 credits isn't 150 images
1:27:12 it's more like you know 15 but it's free
1:27:17 um so there's there's our abandoned 70s
1:27:20 muscle car like the these models are all
1:27:22 nuget getting pretty prettyy
1:27:24 solid you know pretty solid things okay
1:27:28 so the the other thing about this is
1:27:31 there's all sorts of switches and dials
1:27:33 so so down the left hand side here you
1:27:35 can turn on or off Alchemy V2 so what's
1:27:39 this talk about upgrade now to take
1:27:41 advantage of alchemy it's photo hyper
1:27:44 photo realistic prompt magic
1:27:48 V3 Alchemy pipeline Alchemy preset so so
1:27:51 Alchemy is is their fancy version that
1:27:53 you can upgrade to get you can do
1:27:55 different image Dimensions you can do
1:27:57 square ones horizontal ones vertical
1:28:00 ones um you can upload an image image
1:28:04 inputs have moved to image guidance tab
1:28:07 show
1:28:08 me so the image guidance tab is now here
1:28:12 so I can now upload an image to use as
1:28:15 the
1:28:16 source so let's go grab an image and and
1:28:19 see that like I said I am
1:28:23 figuring this out on the Fly because
1:28:24 they keep changing
1:28:31 everything okay let's do a Radio Shack
1:28:36 trs8 that was the first computer I
1:28:38 learned to program basic
1:28:41 on all right so that's our input and
1:28:45 then we're going to go
1:28:49 um
1:28:51 futureistic
1:28:58 computer styled like
1:29:02 70s
1:29:09 computer and I can change the strength
1:29:12 of the image like adhere to the image it
1:29:14 was down at 30 I'll move it up to 60
1:29:16 we'll see how that
1:29:20 does and this is going to use 11
1:29:26 credits it's giv me a little error that
1:29:29 says my image uploaded was a 3x2 aspect
1:29:32 ratio and I'm
1:29:33 generating in a I think a 16 by9 aspect
1:29:37 ratio so it's telling me something's off
1:29:39 which is cool but so Leonardo is very
1:29:43 much if you like to Tinker if you like
1:29:45 to flip the switches and and do like
1:29:48 that it's really good if you don't
1:29:51 something like do is better better for
1:29:52 you
1:29:54 um let's see did this generate my
1:29:58 images add up to four
1:30:01 layers of image guidance
1:30:05 what generation
1:30:12 history are
1:30:20 these oh these are all they it kept it
1:30:23 okay I probably should I probably should
1:30:25 have kept it back at
1:30:29 uh back at
1:30:38 30 all right let's see we'll go back
1:30:41 over here all right so it's generating
1:30:44 these things now um I'm going to show
1:30:46 you one other thing on Leonardo so okay
1:30:49 so so it's sort of improved these things
1:30:51 so so you'd have to muck with that but
1:30:54 if you want to if you want to a lot of
1:30:56 times people ask how do I up upload an
1:30:58 image and then slightly modify it um
1:31:01 with this and with the the canvas editor
1:31:03 there's probably some this is probably a
1:31:05 good place to do that let me go back out
1:31:07 though to the main Leonardo screen
1:31:11 um
1:31:13 okay there's these things called
1:31:15 fine-tuned
1:31:18 models so fine tune models are you know
1:31:22 things that are are trained on specific
1:31:26 image sets right so if you want this
1:31:28 particular kind of Animation look or you
1:31:31 want pixel art for example or you know
1:31:35 um uh whatever like like video game
1:31:39 components right so there there's these
1:31:41 models that have been trained you can
1:31:42 pick any of these models and then you
1:31:44 can go generate images within that Mo
1:31:47 model like here's one that's called
1:31:49 Christmas
1:31:50 stickers so all of the images that are
1:31:53 created with this are going to have the
1:31:55 same kind of look and
1:31:57 feel you can make your own so if I
1:32:01 scroll to the top here there's platform
1:32:04 models Community models so this is
1:32:06 people that have made their own and
1:32:07 shared them and then there's my models
1:32:09 so these are three models that I've made
1:32:11 two on really bad photographs of myself
1:32:13 that are just
1:32:15 embarrassing and then um some
1:32:19 photography so I'm I'm a photographer if
1:32:21 you go to Kyle Shannon you can see my
1:32:23 photography um but I can now
1:32:28 generate um images using my own
1:32:31 photography as the
1:32:33 model and
1:32:35 so the instance prompt is called a Kyle
1:32:39 Shannon sky so so if I type in I'm using
1:32:43 my model and then I've got this prompt
1:32:45 that says a Kyle Shannon
1:32:50 Sky of
1:32:55 Manhattan
1:32:57 during
1:32:59 Manhattan
1:33:04 henge at
1:33:12 Sunset and now it's going to make me
1:33:15 some New York City skylines with with my
1:33:18 pretty pretty clouds in there of all the
1:33:21 AI image generating AI apps which one
1:33:24 generates the highest quality
1:33:28 artworks
1:33:31 um Mid journey is the highest quality
1:33:35 artwork but it doesn't do text well I'm
1:33:38 guessing it will in the next iteration
1:33:41 Dolly 3 for me is the one that's the
1:33:43 most it like it's acceptably good like
1:33:46 it's it's really really good it does
1:33:48 text well and what I like about Dolly 3
1:33:51 is you don't have to prompt it it
1:33:53 prompts itself you're doing it within
1:33:55 the context of a chat GPT conversation
1:33:59 so they're very different kind of things
1:34:01 and then stable diffusion which is what
1:34:04 Leonardo is sitting on top of is a whole
1:34:07 ecosystem of open- source models and
1:34:10 luras
1:34:11 and checkpoints and just all this sort
1:34:14 of geeky [ __ ] you can run down rabbit
1:34:16 holes with stable diffusion for for days
1:34:18 and weeks and things like that um this
1:34:22 didn't really give me what I was looking
1:34:23 for but that's those are definitely my
1:34:25 clouds right like that's definitely my
1:34:28 style of Photography and the kind of
1:34:31 kind of stuff I have so this is
1:34:33 definitely Manhattan with my clouds on
1:34:35 top of it um
1:34:38 so so that is Leonardo and it's it's
1:34:41 just it's very deep now and getting
1:34:43 deeper like they there these four modes
1:34:47 across the top here are are really deep
1:34:50 modes in and of themselves
1:34:52 and then creating your own fine tuning
1:34:54 models is a whole thing in and of itself
1:34:56 um so there's lots of rabbit holes here
1:34:58 so that's
1:35:02 Leonardo woa
1:35:04 [Laughter]
1:35:08 woa we having fun yet anybody having fun
1:35:12 you can make money
1:35:14 with or Leonardo or photoshop or
1:35:18 Microsoft
1:35:20 Word all right let me look at what's
1:35:23 going on in show and
1:35:26 tell that's cool Joker that must be a
1:35:29 Korea thing oh it's an old AI image
1:35:32 that's really cool so look at Joker
1:35:36 like a year year and a half ago this is
1:35:40 what AI image generation was putting out
1:35:44 like it's gotten so much better
1:35:50 amazing
1:36:02 [Laughter]
1:36:07 the trolls the the image trolls in this
1:36:10 channel are the best I got a haircut
1:36:12 today taters is
1:36:16 good who did that digital gods of course
1:36:20 I'm sorry Lord digital God Gods I didn't
1:36:23 I didn't refer to him as Lord digital
1:36:25 gods for a moment there and I got the
1:36:27 [Laughter]
1:36:30 troll oh man
1:36:36 alive
1:36:39 crazy AI May soon require more
1:36:42 electricity than all electronic Vehicles
1:36:44 electric vehicles
1:36:46 combined
1:36:49 yeah it's going to require some
1:36:53 um well here's the good news AI can can
1:36:57 solve the fusion problem and then we can
1:36:59 have local Fusion reactors that run big
1:37:01 GPU farms and well everything will be
1:37:05 [Laughter]
1:37:15 good Leonardo seems like you could
1:37:18 potentially be useful in creating case
1:37:19 studies for a portfolio
1:37:23 yeah um it's there's a lot there like
1:37:27 Leonardo here's the deal
1:37:31 like I think the most important thing to
1:37:34 do right now is understand what you're
1:37:37 good at how your brain works and the
1:37:40 problems you're trying to
1:37:42 solve and and if the problem you're
1:37:44 trying to solve is I want to see what's
1:37:46 possible cool then go play with
1:37:48 everything that's kind of what I'm doing
1:37:49 right
1:37:50 now but
1:37:52 don't get seduced into thinking that
1:37:54 somehow Leonardo is better than Del or
1:37:56 do is better than mid Journey or mid
1:37:58 journey is better than Adobe
1:38:02 Firefly all of these tools have really
1:38:05 particular strengths and
1:38:08 weaknesses that in in any given
1:38:11 situation I might absolutely use Firefly
1:38:15 over mid journey and use mid Journey
1:38:18 over Del and use right all the all the
1:38:21 situations are different same with the
1:38:23 large language models chat gbd4 is by
1:38:26 far the best tool out there in my
1:38:28 opinion as an overall tool but if I want
1:38:32 to do something like research I'm going
1:38:34 to use a a thing like perplexity that
1:38:37 that really focuses on um on you know
1:38:41 research and citations and accuracy if I
1:38:44 want something that's conversational I'm
1:38:46 probably going to use Pi which is
1:38:48 optimized for conversations right so
1:38:50 different situ ations the the different
1:38:53 tools they're not all the same and and
1:38:56 you know you know people people say to
1:38:58 me or ask me all the time like what's
1:39:00 the best tool and there's no there's no
1:39:03 answer to that right it's it's what's
1:39:05 the best tool for what
1:39:08 job for this job it's a screwdriver a
1:39:11 flathead screwdriver for this job it's
1:39:14 an allen wrench right they're different
1:39:16 tools um it's not that one's better than
1:39:19 the other but you know as as an overall
1:39:22 if you just if you only want to play
1:39:24 with one tool then the one to play with
1:39:26 right now is Chachi D4 period end a
1:39:33 story but if you want to do art then you
1:39:35 run down that rabbit hole if you want to
1:39:37 do video you run down that rabbit hole
1:39:40 right there's also shuty shutter stock
1:39:43 and Getty Images powered by Nvidia um
1:39:45 that would be more commercial AI images
1:39:47 Suzanne also Runway ml just cut a deal
1:39:50 with Getty um so you're going to be able
1:39:52 to make videos now that are trained on
1:39:55 Getty stock images or stock videos I
1:39:57 assume it's hard to compare them it's
1:40:00 kind of impossible to compare them the
1:40:01 the thing to do right now is get curious
1:40:04 and play not like get curious and work
1:40:08 and learn how to do work with them start
1:40:12 playing play with these tools come to
1:40:16 one of the things that that is happening
1:40:18 on these lives is that there's a sense
1:40:21 of play here cuz I'm an idiot but the
1:40:24 people that are coming back over and
1:40:26 over
1:40:27 again are like to play right and so
1:40:30 they're playing and they're sharing
1:40:31 images and and in doing that we're
1:40:33 collectively learning what these tools
1:40:35 are good at what they're not good at
1:40:37 where they where you might take a
1:40:40 certain kind of skill and be really good
1:40:42 at one tool but be not as good at
1:40:43 another tool I missed which model is
1:40:47 best for video
1:40:50 well let me go show you just a couple of
1:40:52 sites so join the AI salon for one thing
1:40:56 and then go to um
1:41:01 the tool AI tools and resources uh
1:41:04 section and there's a video creation um
1:41:09 uh group of apps
1:41:12 um and I actually don't even have
1:41:14 stability stable video on here um so I I
1:41:19 I guess I I'll add that but
1:41:23 there's there's a bunch of different
1:41:26 kinds of video tools so you've got
1:41:29 Runway ml Moon Valley and P laabs right
1:41:33 now are the best at text to video or
1:41:37 image to video or even video to video
1:41:40 they're they're they're generating
1:41:42 original video content based on some
1:41:45 input you got this tool visla which
1:41:49 assembles stock video together and
1:41:52 narrates it so if you want to do like a
1:41:54 business video with actual footage
1:41:57 that's been shot by professionals Visa
1:42:00 is really good but it's a completely
1:42:02 different tool than Runway Moon Valley
1:42:05 and P and uh yeah pabs and then you've
1:42:08 got heyen and did which are like Avatar
1:42:12 makers and Heen does stuff like um video
1:42:16 translation
1:42:17 tools which these these are just [ __ ]
1:42:20 crazy
1:42:23 I'll show
1:42:29 you you know here's a video source of
1:42:32 support living with rheumatoid arthritis
1:42:34 has been um speaking with other people
1:42:38 about what they've gone through uh
1:42:40 reading a lot about the effects of
1:42:42 rheumatoid arthritis so this is a
1:42:44 product I'm putting together for my
1:42:46 company Story vine so we're an 11 and a
1:42:48 half year old company that does video
1:42:50 storytelling
1:42:51 and we work with a lot of pharmas which
1:42:53 are you know Global organizations so
1:42:57 they have they want to be able to
1:42:59 leverage the stories that they capture
1:43:01 in lots of different markets well now
1:43:03 they can using the exact you know
1:43:08 um video the the the voice and look and
1:43:12 feel of the original but speaking a
1:43:15 different
1:43:20 language
1:43:30 so so the way our system works is very
1:43:33 modular I can take all the components of
1:43:35 the original video reassemble them with
1:43:38 new lower thirds and things like that
1:43:40 and then even put them into a system
1:43:42 that will generate a Spanish
1:43:44 transcription Spanish description um
1:43:47 Spanish title um and then assemble the
1:43:50 video into a final
1:43:53 [Music]
1:44:04 form and then I can do
1:44:06 [Music]
1:44:14 that I can do that for 27 languages
1:44:16 there's Ellen speaking Hindi here's
1:44:18 Ellen speaking Japanese
1:44:29 [Music]
1:44:32 so that's hey Jen so so the once you
1:44:35 start getting into multimedia with music
1:44:39 audio animation video like it's it's a
1:44:42 it's a much more complicated space
1:44:45 because what you're trying to accomplish
1:44:47 the the use cases May dictate a
1:44:49 combination of tools
1:44:52 um in some cases you may want to
1:44:54 generate original video content in other
1:44:56 cases you might just want to Source
1:44:59 existing stock footage and assemble that
1:45:01 together like there's just all sorts of
1:45:03 different
1:45:04 ways have the AI decide Runway ml for
1:45:07 video have the AI decide which one is
1:45:09 best or will they fight each other um
1:45:11 well th listen this is these are all
1:45:14 commercial Ventures so they're all going
1:45:15 to
1:45:17 compete um but all of these tools are
1:45:20 going to get really good really really
1:45:23 fast
1:45:25 so
1:45:26 um th this is one of the one of the
1:45:29 philosophical things that that people
1:45:31 have been that the technologists have
1:45:35 been talking a lot about is there was a
1:45:38 famous um there was an engineer from
1:45:41 Google about seven or eight months
1:45:44 ago put out this piece called we have no
1:45:47 moat so if you look that up Google we
1:45:50 have no
1:45:51 OTE uh what about Adobe Premiere Pro
1:45:54 user um yeah so adobe's building lots
1:45:57 and lots of AI stuff into all their
1:45:58 tools so so premier's added really cool
1:46:02 stuff like you can edit text and it'll
1:46:04 it'll edit the video it'll edit out ums
1:46:06 and things like that so yeah definitely
1:46:09 look at the Adobe stuff they're they're
1:46:10 aggressively integrating
1:46:12 AI uh what was I just talking about
1:46:18 uh I don't remember
1:46:21 um how do you check that the
1:46:22 translations are accurate I so well if I
1:46:26 were doing this commercially and I am
1:46:28 going to be doing this commercially
1:46:29 we'll have to check it but um let me
1:46:32 show you another video um I
1:46:38 did I did a video of myself in Mandarin
1:46:42 Chinese and then I said on one of these
1:46:44 Tik Tok
1:46:45 lives uh m o a t Nancy moat as in the
1:46:49 moat around the castle
1:46:51 uh oh that's what I was talking about no
1:46:52 Moe yeah um so I let me I'll get back to
1:46:56 that so anyway this is
1:46:59 um one of my one of my Tik Tok followers
1:47:03 um put together this video for Mandarin
1:47:05 using Ai and what he wanted to know is
1:47:08 what does a native speaker think let's
1:47:09 find
1:47:13 out that was Kyle yeah kering is really
1:47:18 good he talking about AI some my
1:47:21 Mandarin is really good okay well it was
1:47:23 AI yeah it was AI it was AI it was AI
1:47:27 yeah it wasn't him oh he's
1:47:29 AI interesting isn't it see Kyle he
1:47:32 didn't even realize yeah so um so
1:47:36 apparently the translations are like
1:47:39 incredibly good
1:47:42 um the the um the article about we have
1:47:47 no
1:47:49 moat what the what the article was was
1:47:52 basically talking about is that the open
1:47:54 source large language models were
1:47:59 getting really good really fast for like
1:48:02 a $100 in compute cost like basically
1:48:06 the
1:48:07 open-source um projects for a 100 bucks
1:48:11 were training models that were
1:48:13 essentially almost as good as what
1:48:14 Google had spent $5 million
1:48:17 training and and what they were
1:48:19 basically saying is we have no mod is
1:48:22 there's not going to be a kind of
1:48:23 Technology we can build that isn't going
1:48:26 to be knocked off like that and and I
1:48:28 think with software in general that's
1:48:30 the case that no matter what you create
1:48:33 now someone's going to be able to just
1:48:36 speak it into existence right so you've
1:48:39 got some super proprietary tool that
1:48:41 analyzes real estate data in a
1:48:43 particular way with a particular
1:48:45 algorithm that combines this with that
1:48:47 and the other thing and gives you these
1:48:49 magical [ __ ] prediction
1:48:51 okay great someone else will be able to
1:48:54 look at that and go all right let me
1:48:56 screenshot the interface and I'll just
1:48:58 dump that into a tool that's going to
1:49:00 replicate the interface and then let me
1:49:02 let me let me go to some backend tool
1:49:04 and just say I want a tool that takes
1:49:06 this kind of real estate input does
1:49:09 whatever calculations are necessary and
1:49:11 generates this kind of output and it's
1:49:12 going to generate that they're going to
1:49:14 be able to speak that into reality so we
1:49:17 are we are entering a bizarre
1:49:21 um professional environment where all
1:49:26 the rules of all the industries are
1:49:29 going to
1:49:32 change and all of the big incumbent
1:49:36 companies are going to resist that
1:49:39 change so remember when
1:49:42 Kodak um when when some some young punks
1:49:46 at Kodak invented the digital camera and
1:49:49 they went into to the the big boss's
1:49:52 office and said hey bosses we invented
1:49:55 this cool thing called the digital
1:49:57 camera that uses a CCD instead of film
1:50:00 and and look it it makes a a you know 73
1:50:04 pixel wide photograph that looks like
1:50:06 [ __ ] what do you think of
1:50:07 that the executives at Kodak went
1:50:12 son we're not in the pixel business I
1:50:16 don't even know what a pixel is we're in
1:50:18 a silver Halli business it's about
1:50:20 chemical boy get out of my
1:50:26 office and now they're out of
1:50:30 business there's going to be a lot of
1:50:35 that one of the things that excites me
1:50:38 about the community that shows up here
1:50:40 there's a lot of people here that work
1:50:41 at big companies right
1:50:43 now that are that that come here every
1:50:46 night um and they are getting
1:50:50 increasingly visible within their
1:50:52 organization as someone who knows
1:50:54 something about this AI stuff
1:50:58 um it's it's those companies that let
1:51:01 those voices be heard that are going to
1:51:05 do
1:51:05 okay it's not going to be not a
1:51:08 challenge but the companies that sit
1:51:11 back with their arms crossed and go
1:51:15 Harum this is the agency business this
1:51:17 is about time and material son
1:51:21 we hire young people out of college and
1:51:23 we charge $150 an hour for them and
1:51:26 that's how we make our profit you're not
1:51:28 going to tell me we're doing it any
1:51:32 different I guarantee you that
1:51:34 conversation's happening right now in
1:51:37 major agency holding companies promise
1:51:46 you I don't know how those companies
1:51:49 survive
1:51:51 when some 20 person AI Centric
1:51:57 agency shows up to one of their clients
1:51:59 and says we can deliver as good or
1:52:03 higher quality twice as fast for half
1:52:06 the
1:52:08 price the clients are going to go all
1:52:11 right we'll give you a shot and if that
1:52:13 20 person agency
1:52:16 delivers that big agency holding company
1:52:18 is going to get a call we're putting
1:52:21 this account into
1:52:23 review I think all of the AI info that
1:52:26 spoke about during my interview that
1:52:28 landed me the position wow that's really
1:52:31 cool I hear about on Monday oh wow
1:52:34 that's so awesome
1:52:36 Suzanne yeah I so so listen this is my
1:52:39 thing hang out here like come back here
1:52:41 every night I I do this nightly it's
1:52:43 Saturday night I'm here it was Friday
1:52:45 night last night I was here tomorrow
1:52:47 night's Sunday I'll be here come back um
1:52:51 um get curious about this stuff start
1:52:54 learning about it and talk about it go
1:52:58 to the AI Salon community and start
1:53:02 sharing stuff share what you're doing
1:53:04 share where you got stuck ask
1:53:08 questions it's
1:53:11 Christmas talk about it with your family
1:53:14 my family hates this stuff talk about it
1:53:16 anyway talk about it when the Neighbors
1:53:19 come over
1:53:20 um talk about it at work talk about it
1:53:24 with your friends they'll think I'm
1:53:26 weird okay and you know what they'll do
1:53:29 six months from now hey hey Sally yeah
1:53:34 could you tell me about that AI stuff
1:53:36 that I was bitching about
1:53:40 before Ford Sarah Lee and Kodak were
1:53:42 just a few of the former Wisconsin
1:53:45 labels now WS packaging remember that
1:53:47 era
1:53:49 yep
1:53:54 I used to work with the Gail Lewis very
1:53:58 nice went from literally cutting and
1:54:00 pasting to digital stuff
1:54:02 begrudgingly
1:54:05 yep oh literally cutting and pasting
1:54:08 yeah yeah the changes like the changes I
1:54:11 I remember Joker it's so funny you say
1:54:13 that I remember I this was like
1:54:16 1988 I was doing desktop publishing
1:54:20 I was doing freelance work for American
1:54:22 Express I was doing like annual reports
1:54:24 using Quark Express 1.0 and like
1:54:27 Photoshop
1:54:28 2 and then I got
1:54:31 hired I got a freelance gig in an ad
1:54:35 agency and I'm doing some sort of
1:54:37 desktop publishing just drawing like
1:54:39 boxes and [ __ ] like that and I remember
1:54:40 walking down this this dark
1:54:43 hallway and sitting in the hallway was
1:54:45 this old dude with a white beard and he
1:54:48 had a big table with like an angled
1:54:51 drafting table and he had all these
1:54:53 rolls of of tape all these different
1:54:56 Wits of tape and I saw I'm like doing
1:54:59 this and measuring it and doing this and
1:55:01 I and I said oh what are you doing that
1:55:03 looks cool and he goes I'm I'm creating
1:55:07 rules are you l s rules were lines he
1:55:12 was like literally pasting black
1:55:16 tape on a piece of paper to draw a line
1:55:19 and then another piece of tape to draw
1:55:22 the horizontal line and another piece
1:55:24 and I and I just I got I like started
1:55:27 flop sweating because I realized that I
1:55:30 just drew a box on the screen and it
1:55:32 made the thing that he was like manually
1:55:34 taping up and I remember just thinking
1:55:36 like oh that's cool and I sort of slunk
1:55:39 slunk away because I realized his job is
1:55:42 gone he doesn't know it yet but his job
1:55:45 is gone and and it was like [ __ ] um
1:55:49 there's going to be a lot of that
1:55:51 and and I don't know if that guy ever
1:55:53 got into digital or not like he you know
1:55:55 he'd clearly been in the biz for a while
1:55:58 he knew his [ __ ] he was really good at
1:56:00 like it was really impressive what he
1:56:01 was doing but that job that job
1:56:06 disappeared overnight and there's going
1:56:08 to be a lot of
1:56:13 that oh you weren't begrudged you were
1:56:16 on the digital side pag maker Ready set
1:56:19 go yep
1:56:20 AI solves the algorithm what can I do
1:56:22 with
1:56:23 AI Lots Evelyn what do you want to do
1:56:26 what's a problem you need to solve like
1:56:28 what I don't know what you do if you're
1:56:29 a project manager it can help you come
1:56:30 up with project plans if you're a
1:56:32 copywriter it can help you ideate and
1:56:35 and
1:56:37 edit um if if you're an entrepreneur it
1:56:42 can help you come up with a marketing
1:56:43 plan so so the the the thing to think
1:56:45 about with AI a AI is just a tool and AI
1:56:48 at this point is is a sweet of tools
1:56:51 it's many many tools what I would say is
1:56:55 don't think about what can I do with the
1:56:58 tool right if you think about just
1:57:00 simple tools like a shovel or a hammer
1:57:03 what can I do with a shovel I don't you
1:57:04 can do lots with a shovel you can dig a
1:57:07 hole you
1:57:09 can smack someone in the head with it
1:57:12 you can like there's all sorts of things
1:57:13 you can do with a shovel there's all
1:57:15 sorts of things you can do with a
1:57:17 hammer but if you need to you know
1:57:21 attach two pieces of wood together then
1:57:23 the Hammer's probably going to help and
1:57:25 you're probably going to need a thing
1:57:26 called a nail so what I would do is
1:57:28 start with what what do you do or what's
1:57:31 something that you're trying to solve
1:57:34 and then I would say what can AI do to
1:57:38 solve that problem and then in doing
1:57:40 that you'll figure some stuff out really
1:57:42 really quick our friends probably
1:57:44 already think we're weird this is just
1:57:46 our AI era yeah exactly exactly
1:57:50 yeah it's funny you know I've got a
1:57:51 bunch of friends that in the in the mid
1:57:54 90s when I discovered this worldwide web
1:57:56 thing and I lost my mind with the
1:57:58 worldwide web like I've lost it with AI
1:58:00 where I was just it's all I talked about
1:58:01 for for years all I talked about was the
1:58:04 web and building [ __ ] and all that sort
1:58:07 of stuff
1:58:10 and and they all kind of made fun of me
1:58:13 and they're like I don't get it what are
1:58:14 you doing why are you doing all this
1:58:15 digital stuff what's with this worldwide
1:58:17 web
1:58:18 why
1:58:22 and then you know I started agency.com
1:58:24 this this digital agency and it got
1:58:26 really popular and really big and like
1:58:28 years later they would all say I should
1:58:29 have listened to you you know you knew
1:58:35 that and now it's happening again those
1:58:37 same
1:58:38 people I'm like this AI shit's amazing
1:58:41 they're like why you doing that AI stuff
1:58:44 it's
1:58:46 stealing it's like they're the exact I'm
1:58:48 like this is the same thing it's the
1:58:50 same [ __ ] thing as I was talking
1:58:52 about back then they can't see it I
1:58:55 don't know so that's that's why I'm
1:58:58 doing this every night is to try to get
1:58:59 people just open your [ __ ] eyes just
1:59:02 get curious start playing with this
1:59:05 [ __ ] if you haven't had your Kevin
1:59:08 McAllister moment
1:59:09 yet that one you're not playing enough
1:59:13 play play play and at some point you'll
1:59:17 do something with one of these tools
1:59:19 where it will melt your [ __ ]
1:59:22 face and then you go
1:59:25 oh I get it now but certainly it can't
1:59:29 do what I do and then you ask it to do
1:59:32 what you do and you go oh
1:59:37 [ __ ] now what do I
1:59:40 do now you get curious now you're
1:59:44 curious that's the moment you can't come
1:59:47 back from with this stuff unfortunately
1:59:50 cuz it's like this stuff's coming
1:59:51 whether you want it to or not so you can
1:59:54 choose to be on the adventure or choose
1:59:56 not to be on the adventure if you choose
1:59:58 not to be on the adventure the
2:00:00 Stormfront is still coming
2:00:05 in now is it a Stormfront that's a
2:00:08 wildly destructive Cat 5
2:00:11 hurricane or does it just stir some [ __ ]
2:00:14 up and [ __ ] with the neighbors houses I
2:00:17 don't know I don't know but but I'd much
2:00:20 rather be on a
2:00:22 surfboard heading toward the beach than
2:00:25 on the beach in a flimsy little cottage
2:00:28 right now and that's that's what this
2:00:30 Channel's about registration marks oh
2:00:33 yeah print major here flop sweating
2:00:36 question mark law just when you know
2:00:39 when you break into like an anxiety
2:00:41 sweat where you're just like oh [ __ ] I
2:00:43 can't what what do I
2:00:45 say I just walked
2:00:48 out I used to be the Quark publishing
2:00:51 admin wow that's how I first got into
2:00:54 creative software Tech yeah it's super
2:00:57 cool he's still there measuring tape I
2:01:00 think
2:01:02 yeah that was that
2:01:04 was you know what was funny is when I
2:01:07 started playing with Quark Express I had
2:01:09 a similar kind of Glee that I've got
2:01:11 with this AI stuff like this stuff
2:01:13 excites me like I was super excited and
2:01:16 you can draw a box and it makes a square
2:01:18 and then you can print it out on a
2:01:19 printer and you can see it and like it
2:01:21 was it was magical right it's you know
2:01:22 it's math but it's magic it occurs as
2:01:26 magic
2:01:28 and like I was in this sort of glib
2:01:30 gleeful place and then I saw this guy
2:01:34 and it like it hit me in the face oh
2:01:37 [ __ ] this is like this is
2:01:39 actually this is a real thing like
2:01:42 people are going to be affected by this
2:01:44 and it took on it it just took on a
2:01:46 different weight you
2:01:48 know and so I I may maybe that's part of
2:01:52 what this Channel's about is I like
2:01:56 I I want to get to as many people as I
2:01:58 can to say set down the rule of
2:02:02 tape and teach yourself Quark Express
2:02:05 because it's coming and this AI shit's
2:02:09 coming so fast holy [ __ ] it's coming
2:02:15 fast I was a pasted up artist and I
2:02:18 bought a Mac to evolve yep awesome I
2:02:20 schedule meetings heck it's my new
2:02:23 grammarly that's like asking what can I
2:02:26 do with the internet yeah what can I do
2:02:27 with AI is it's a it's a broad
2:02:30 question Power Mac oh yeah my first Mac
2:02:33 was a
2:02:35 512k then I had a se30 with a color
2:02:39 monitor I got myself a color card for my
2:02:42 se30 and I had a color monitor then I
2:02:45 had a Max 660 AV that's what I started
2:02:47 Urban desires on
2:02:52 I've had a bunch of them I had the 20th
2:02:54 anniversary Mac for a while that was
2:02:57 pretty
2:02:59 cool someone said what what was that
2:03:02 someone does not condone shovel to the
2:03:04 head was that tiktk does not condone
2:03:07 Irregulars present Turtles hedgehogs
2:03:09 what time is it is it time oh it's 10
2:03:14 o'clock all right I'm going to get going
2:03:17 here pretty
2:03:18 soon expand on this make fast cash with
2:03:22 AI how Spike Jones you can make money
2:03:26 with I'll tell you the answer do you
2:03:28 want to hear the answer you're not going
2:03:29 to like the answer but I'll tell it to
2:03:32 you you can make money
2:03:34 with these
2:03:36 people that say
2:03:39 this how they're making money with chaty
2:03:42 BT is selling you courses on how to make
2:03:44 money with chat
2:03:47 PT this is a false guy
2:03:51 a false prophet a false prophet as well
2:03:54 as
2:03:56 Prophet
2:04:00 um Chachi PT is just a tool Chachi PT is
2:04:03 actually like seven or eight
2:04:06 tools um I feel secondary sadness for
2:04:08 the guy taping up the paper in your
2:04:10 office I hope he survived I do too like
2:04:12 that that's what I'm saying cuno it it
2:04:14 was like the the weight
2:04:18 of the technology I was excited about
2:04:21 all of a sudden hit me like it it it
2:04:23 became it became very real I yeah I
2:04:25 totally agree it's like and and and
2:04:28 that's coming for a lot of people and
2:04:30 it's coming for a lot of people that
2:04:32 that thought they were immune right I I
2:04:35 think
2:04:37 the the common knowledge about about Ai
2:04:42 and Robotics and things like that it was
2:04:44 going to take manufacturing jobs first
2:04:47 but knowledge workers were safe cuz cuz
2:04:49 cuz we're we're us and you know we're
2:04:52 smart and and you know it can't do what
2:04:55 we do right yeah you can get a robot to
2:04:58 grab stuff that I get yeah that's going
2:05:00 to take their jobs but it can't it's not
2:05:03 it can't replace a consultant with 30
2:05:05 years of
2:05:06 experience oh except it
2:05:10 can right that's that's the reality
2:05:13 we're entering is
2:05:15 that any computer mediated job creative
2:05:19 knowledge work anything that's computer
2:05:21 mediated is going to be dramatically
2:05:23 impacted by this
2:05:26 technology and so there's a lot of
2:05:29 people right now that don't think it's
2:05:32 possible that these tools could do what
2:05:34 they
2:05:36 do and so therefore they're going oh
2:05:39 those are that's that AI stuff I heard
2:05:42 it hallucinates it makes typos I heard
2:05:45 it can
2:05:46 add I'm good we're good
2:05:53 by the end of next year it's going to
2:05:55 just be a completely different story if
2:05:57 we if if indeed David Shapiro is right
2:06:00 and the rumors about qar are right AGI
2:06:05 is effectively here it just hasn't been
2:06:07 rolled out
2:06:09 yet now it's possible it's not here but
2:06:12 if it
2:06:13 is and if if you're new here AGI stands
2:06:16 for artificial general intelligence it's
2:06:18 when the these systems are better than
2:06:21 most humans at the work they do not as
2:06:26 good as better social influencers are
2:06:29 being replaced now yeah that too I mean
2:06:33 everything every
2:06:37 single I mean have you seen the the
2:06:41 the well I showed you the Ellen thing
2:06:44 where she's speaking Spanish now in her
2:06:47 voice I did an avatar of
2:06:50 myself talking about [ __ ]
2:06:54 and thanks for the education sir much
2:06:57 appreciated and
2:06:58 obliged absolutely Ron thank you for the
2:07:01 kind
2:07:02 words very meta yeah it's a lot of meta
2:07:04 stuff I'm in a I'm in a meta kind of
2:07:06 mood right
2:07:07 now ai music is still bad unless it's
2:07:10 combined with some human input yep
2:07:14 but images 18 months ago were unusable
2:07:18 images today are staggeringly good
2:07:21 music's going to go the same way videoos
2:07:23 going to say go the same way 3D worlds
2:07:26 are going to go the same way programming
2:07:28 is going to go the same way math is
2:07:30 going to go the same way like all of it
2:07:32 all of it all the dominoes are going to
2:07:38 fall so what do those people
2:07:43 do I don't
2:07:45 know I do not
2:07:48 know you just described all of Tik Tok I
2:07:51 don't know what I was talking about how
2:07:52 to
2:07:53 ideas his beard eventually got in the
2:07:56 way
2:07:58 anyway weird I see what is wrong it
2:08:02 says uh okay I don't know what that is
2:08:05 what year did the first version of chat
2:08:07 come out well chat so so
2:08:11 okay so there was gpt2 and gpt3 which
2:08:15 were these the underlying what they now
2:08:18 call the Frontier models so the core
2:08:21 large language model itself and then you
2:08:24 would build applications on top of them
2:08:26 so if you go play with the open ai's
2:08:30 playground what they call it platform.
2:08:33 open.com
2:08:36 playground you have access to their apis
2:08:40 so that you can build software
2:08:42 applications to use these tools chat
2:08:47 GPT is an application layer sitting on
2:08:50 top of the original one was on top of
2:08:53 GPT
2:08:55 3.5 and that's the one that launched a
2:08:57 year
2:08:59 ago on November 30th so so November 30th
2:09:04 2022 is the first official chat
2:09:07 GPT
2:09:09 application there are tools like um oh
2:09:13 what the [ __ ] is it called copy. a and
2:09:16 there's um there there's other things
2:09:18 that were're using the GPT 3.5 AI system
2:09:22 so a lot of those
2:09:23 copywriting sites like copy AI were
2:09:27 using this before chat GPT came out but
2:09:29 they were these really specialized
2:09:31 Byzantine kind of interfaces for all
2:09:33 these different promps
2:09:36 basically chat GPT was the first one
2:09:38 that came out as a general tool that was
2:09:40 just as simple as the Google search box
2:09:42 and said ask [ __ ] here and it will do
2:09:45 things for you so that it's a year old
2:09:49 someone tell Kyle he's got a guest
2:09:53 request all
2:10:00 right
2:10:05 hello so for those of you listening
2:10:07 remember Kyle always forgets that he has
2:10:10 his guest request turns on this is the
2:10:12 thing Kyle does hey Kyle how's it going
2:10:16 it's good how's it
2:10:17 going do all right so here's where we're
2:10:20 at as a very very novice AI
2:10:25 user who who apparently has video now on
2:10:28 Tik Tok hey for those of you under a
2:10:30 thousand Tik Tok the reason it's been
2:10:32 glitching out so badly recently is
2:10:34 because they're trying to move to a
2:10:36 lower threshold oh really for video
2:10:39 that's definitely what's happening
2:10:40 because I shouldn't have video right now
2:10:42 but I do oh interesting you get you get
2:10:44 to be treated to my wonderful uh driving
2:10:47 cap I like your hat
2:10:49 thank you it's my uh it's actually a uh
2:10:52 my Tartan pattern from my my uh client
2:10:54 oh hey there it is right
2:10:56 there that's valid this is this is my
2:10:59 clan from Scotland back in the day but I
2:11:02 am an American just to be full
2:11:04 disclosure so
2:11:06 um earlier somebody mentioned
2:11:10 how and this is legitimately where we're
2:11:12 at people is
2:11:14 that I I would say the majority where
2:11:18 what do you say
2:11:19 what percentage of people think that AI
2:11:21 is Hal or the Matrix like what
2:11:25 percentage of people right now at least
2:11:26 in the
2:11:28 US well 80 82% of the people in the US
2:11:32 right now are not using chat GPT so I
2:11:35 would say most people in the US don't
2:11:37 even know this [ __ ]
2:11:39 exists so but they can still they can
2:11:43 still contemplate what AI might mean
2:11:47 what percentage would you say are are
2:11:48 contemplating it negatively I'm not
2:11:51 these people I want to be clear almost
2:11:52 almost all so so listen here here's
2:11:54 Here's my thought on that Hollywood has
2:11:57 trained us for the past 50 years that
2:11:58 the robots are going to kill us right so
2:12:02 so and and this is this is one of those
2:12:04 things where when you look at AI from
2:12:08 the outside it either
2:12:10 looks trivially weak or or it looks you
2:12:15 know absolutely terrifying right and so
2:12:18 so I think people go to one of two
2:12:19 places they demonize it or they dismiss
2:12:22 it and what I've experienced and what
2:12:25 this channel is all about is um when
2:12:29 they actually experience it the tool
2:12:31 becomes
2:12:32 real and they start to realize oh this
2:12:35 is actually a tool and I can use it in
2:12:37 some areas for good and there's other
2:12:39 areas where it might scare me like like
2:12:41 when you actually understand what this
2:12:42 thing is you realize most of these tools
2:12:44 are janky pieces of [ __ ] that are
2:12:46 profoundly powerful right yeah so so the
2:12:51 fact that they're janky pieces of [ __ ]
2:12:52 doesn't mean that they're not going to
2:12:54 transform everything they're just we're
2:12:57 just not there yet so so I think the Gap
2:13:00 what what you're talking about is the
2:13:01 gap between people's perception of what
2:13:05 they think it is versus what it really
2:13:06 is right now is tremendous and and then
2:13:10 the more people that use it the more
2:13:11 they sort of internalize it and go oh
2:13:13 this is I can make a a a beautiful story
2:13:16 for my kids and I can put images and
2:13:18 them oh my God it's not so bad right and
2:13:21 it's also got you know potentially
2:13:23 really destructive power so it's like a
2:13:24 both and kind of thing
2:13:27 so so I came on um well I originally
2:13:32 requested to talk about how you
2:13:34 mentioned how and um the bottom line is
2:13:39 Kyle thank you because you
2:13:42 make you make this uh novel stage of AI
2:13:47 joyful and I have to thank you for that
2:13:50 thank you um because you know
2:13:53 I've it's interesting so I do a lot of
2:13:56 uh religion and politics talk tick Tok
2:14:00 you know I do a lot of that um and you
2:14:02 know those things can also be very
2:14:04 ominous or you can find um places where
2:14:09 you actually identify with others who
2:14:12 you assume you're not going to identify
2:14:14 with you know and uh I've started to
2:14:17 bring some AI concept into the
2:14:19 communications I'm having on this app
2:14:22 and a lot of people are very scared yeah
2:14:25 they are and that's one of the so in the
2:14:27 the AI Salon that I started the
2:14:29 community I started a year ago one of
2:14:31 our main values is empathy and it's it's
2:14:34 very much there for a reason that people
2:14:36 are scared and people are angry and and
2:14:39 I understand
2:14:41 why and and you know part of part of my
2:14:45 mission I suppose is to De demystify
2:14:48 this stuff so that people
2:14:50 can experience it like I don't give a
2:14:52 [ __ ] if people hate AI if they've used
2:14:55 it like if you've actually used it and
2:14:57 you go okay it's this and that you know
2:14:59 what I love illustrating things and I
2:15:02 [ __ ] hate image generation and I'm
2:15:03 never going to use it cool that's great
2:15:05 but if you just sit on the sidelines and
2:15:07 say you know I hate it right as a
2:15:11 concept without understanding what it is
2:15:13 that's what I where I think it's really
2:15:15 dangerous I think those people are just
2:15:16 going to get blindsided and it's going
2:15:18 to be it's going to be ugly so I also
2:15:21 talk a lot in my personal life not on
2:15:24 this app because you know I don't want
2:15:26 to be ostracized and canceled I talk a
2:15:29 lot about and yes I did say canceled I
2:15:31 talk a lot about the
2:15:33 X-Men RIP stand the man lee yeah yeah um
2:15:38 it's about PE being different meaning
2:15:43 that the quote unquote
2:15:46 normies are going to fear you you know
2:15:48 hate and fear you in many cases and um
2:15:52 this is not the direction we want to go
2:15:54 with Society people yeah well we want
2:15:58 difference to be appreciated yeah yeah
2:16:00 yeah well anyway awesome well good
2:16:03 talking to you I appreciate your time
2:16:04 hey thank you yeah yeah and and again I
2:16:06 gotta I gotta shout out the neuros spicy
2:16:09 because Kyle you keep it entertaining
2:16:11 and I very much appreciate it we're
2:16:12 literally watching you just flesh out
2:16:15 your like screenplays and you'll just
2:16:17 make you know a sentence or two
2:16:19 storyline and then we'll spend two or
2:16:21 three hours of you fleshing that out
2:16:25 with AI and you know and the the thing
2:16:27 people don't understand is we interact
2:16:29 back and forth with it this is still
2:16:31 something that is controlled by humans
2:16:33 and we don't have to fear it because as
2:16:36 Kyle said
2:16:37 earlier you know it it it is in its uh
2:16:41 infant stages like legitimately this is
2:16:43 not something to be feared so thank you
2:16:45 Kyle appreciate you having me up yeah
2:16:47 you got got it take
2:16:51 care awesome all right turn on dual
2:16:53 camera again
2:16:56 fantastic all right let me that was fun
2:16:59 that was a good conversation thank you
2:17:00 for coming
2:17:01 up um all I can say is I'm glad I'm
2:17:03 close to retirement I I know Sher it's
2:17:06 like it's like well you know what it is
2:17:08 it's like I think I think for for us us
2:17:11 older folks it's like you go one of two
2:17:13 places you're like oh there's an
2:17:14 opportunity to reinvent myself or
2:17:16 there's like oh [ __ ] do I got to
2:17:18 reinvent myself again I thought we did
2:17:20 this in the '90s
2:17:22 um but but yeah it's it it's a bit
2:17:25 overwhelming at times like it it it
2:17:27 really is perplexity wrote a great cover
2:17:29 letter and resume today
2:17:32 awesome um all right let's
2:17:37 see you should have more random guests
2:17:40 up okay I can do
2:17:42 that I can do that um the the the only
2:17:46 thing I fear about having guess up is a
2:17:48 lot of times we end up we we we we had
2:17:52 Dar the DAR the queen of Tik Tok come up
2:17:54 and battle me and it it completely just
2:17:58 like I I want to find a way to have
2:18:01 people up where I kind of know that
2:18:03 they're coming up to talk about this
2:18:04 [ __ ] um so maybe it's a kind of thing I
2:18:07 I don't know maybe we'll do it with
2:18:08 subscribers or something like that I
2:18:09 haven't quite figured out a way to do it
2:18:11 yet that it's not just a [ __ ] chaotic
2:18:14 hell which I suppose if it turned into a
2:18:17 chaotic hell I can just shut people off
2:18:19 but you know anyway
2:18:22 um I can reinvent myself so I can make
2:18:25 money for my retirement account yep all
2:18:28 of us you'll probably reinvent yourself
2:18:30 five or six times over your career I
2:18:32 think this one's going to be a big one I
2:18:35 I think the reinvention that's going to
2:18:36 happen right now
2:18:39 is I think it's more profound for this
2:18:42 reason let's say I spent 30 years as a
2:18:45 copywriter or 30 years as as an an
2:18:48 illustrator or 30 years as a project
2:18:50 manager I've spent my my career very
2:18:53 focused in a particular
2:18:56 area AI comes along and it's going to
2:19:00 it's going to reinvent that role right
2:19:03 so it's going to reinvent me as a
2:19:05 project manager so I can reinvent myself
2:19:07 within that narrow slice of the
2:19:10 workforce but what AI does is it levels
2:19:14 me up in all the other
2:19:19 all the other tactical jobs right I
2:19:22 might have been a project manager for
2:19:23 all my life but now I can do image
2:19:27 generation and I can do music
2:19:28 composition and I can do copywriting and
2:19:31 I can do Consulting and strategy and
2:19:35 like all of the things now can I do them
2:19:38 as good as someone who's been 40 years
2:19:39 in each of those careers no but if it
2:19:42 can get me to 70% is
2:19:45 good
2:19:46 overnight and I don't have to go back to
2:19:49 school to become a copywriter at you
2:19:52 know 30 years into a career I can just
2:19:54 flip over here and start doing that
2:19:56 because I really enjoy that or or making
2:19:59 art so I think that I think the the the
2:20:03 breadth of the reinvention that people
2:20:06 are going to have the opportunity to do
2:20:09 is going to be really
2:20:10 wide that I find really exciting but
2:20:13 it's also like holy [ __ ] where do like
2:20:15 what do people do I don't know how many
2:20:18 people right
2:20:20 now are at the tail end of a 25 or 30e
2:20:24 career in work that they don't actually
2:20:28 enjoy and all of a sudden they might be
2:20:31 able to do work they've always wanted to
2:20:33 do but didn't have the skills for didn't
2:20:36 have the education for didn't have the
2:20:38 access to and now they can
2:20:42 overnight I find that tremendously
2:20:44 exciting but that's wild what that ends
2:20:48 up looking
2:20:50 like
2:20:53 uh found this a few months ago really
2:20:55 enjoying this Kyle thank you I
2:20:57 appreciate the kind
2:20:58 words for
2:21:02 now call it the chaos to Kyle moment
2:21:05 show I know I know it's all this Show's
2:21:09 this this show this this channel is
2:21:12 already [ __ ] chaotic enough it's like
2:21:15 I don't mind it though I don't mind it
2:21:18 you know it listen it's whatever you
2:21:20 guys want to do I mean this really is
2:21:22 about just doing this stuff in a way
2:21:24 that that serves what what you guys are
2:21:26 up to we we will all be saying wow every
2:21:29 day starting in less than a year I
2:21:32 one-onone or 101
2:21:35 I I would
2:21:37 say on
2:21:39 average three times a week for the past
2:21:43 year have been like wait what
2:21:48 are you kidding
2:21:49 me like three times a week for a
2:21:53 year I'm just like I can't I can't put
2:21:58 any more in
2:22:00 here like what does that
2:22:05 mean the one from Deep Mind last
2:22:08 week the material science thing have you
2:22:11 heard about
2:22:15 this Deep Mind that that that discovered
2:22:22 400,000 new protein
2:22:25 folding predictions that are
2:22:28 accurate announced in the journal
2:22:32 Nature they have discovered 2.2 million
2:22:35 new materials 380,000 of which are
2:22:39 stable
2:22:44 crystals humans and computers combined
2:22:48 prior to that had discovered 48,000
2:22:52 stable
2:22:53 crystals and overnight they just dumped
2:22:57 380,000 additional ones into that
2:23:02 world Material Science affects all of us
2:23:06 every day it's batteries it's it's
2:23:09 cookware
2:23:10 it's energy production it's energy
2:23:14 transmission it's energy production
2:23:18 it's solar panels it's
2:23:24 everything in in in the announcement
2:23:27 they said it it was the equivalent of 8
2:23:29 800
2:23:32 years of scientific discovery
2:23:35 happened
2:23:38 overnight that's like one
2:23:41 category Uno category that that that
2:23:44 affects almost all the other
2:23:46 categories
2:23:51 like we don't we won't know what that
2:23:52 means for a few
2:23:55 years but I think I think scientists
2:23:58 have have already validated 800 of the
2:24:00 380,000 like in Labs
2:24:03 like the shit's
2:24:06 real and it's just going to be I think
2:24:08 2024 is I think you're right 101 it's
2:24:11 going to be a series of just dominoes
2:24:15 falling of oh we can do that now oh
2:24:19 that's possible oh that works oh that's
2:24:26 cured oh we have
2:24:29 AGI oh we have it again I didn't get
2:24:31 that could you try again hey
2:24:35 Siri I'm here could you go
2:24:39 away
2:24:46 okay
2:24:52 I miss
2:24:58 Sydney oh my God me 20 years at the bank
2:25:01 and I only now found my passion Tams
2:25:06 exactly this is what I think is got I
2:25:08 think I think I think the
2:25:11 profoundness it's so easy right now to
2:25:14 focus on the negative all these jobs are
2:25:16 going to be
2:25:20 lost Tams 20 years at the bank and I
2:25:24 only now found my
2:25:26 passion I feel I I just have this sense
2:25:29 that so many people are going to dick
2:25:32 around with
2:25:34 AI realize what it can do and go oh my
2:25:38 God that's what I've always wanted to do
2:25:40 but I didn't know it and now they can do
2:25:45 it holy [ __ ]
2:25:54 now for every one of those are there
2:25:57 going to be nine that get decimated or
2:26:00 are there going to be you know five
2:26:02 people that ReDiscover their humanity
2:26:05 and their joy for five that are
2:26:07 decimated right I don't what's the
2:26:09 balance going to be I don't know but I
2:26:12 mean holy [ __ ] it's we're just it this
2:26:16 is
2:26:17 we're we're at the very beginning of
2:26:19 this thing whatever happened to
2:26:21 nanotechnology it's still there it's
2:26:23 still going and in fact the the
2:26:25 nanotchnology like all these stable
2:26:28 crystals this is that's nanot technology
2:26:31 that's what we're talking about 380,000
2:26:35 stable new materials novel stable
2:26:42 crystals have been
2:26:44 discovered and have been released to the
2:26:46 scien specific Community go do what you
2:26:49 will with
2:26:53 these so I think I think we're going to
2:26:56 start hearing about nanot technology
2:26:58 again because someone's going to realize
2:27:00 oh this is a room temperature super
2:27:03 conductor and it's real this time oh and
2:27:05 there's 10 of
2:27:08 them it's like
2:27:11 oh this thing leads to Fusion this thing
2:27:14 leads to friction L something this thing
2:27:18 leads to electrical properties we've
2:27:21 never known before but we s we theorized
2:27:23 they would be there but here they
2:27:27 are all that anti-gravity [ __ ] that
2:27:30 everyone thought was a conspiracy theory
2:27:32 well here it is like I don't
2:27:36 know Jean splicing may be next I think
2:27:39 Jean splicing will be next I'm going to
2:27:41 have to ask chat GPT what a stable
2:27:43 crystal
2:27:44 is well it just basically means it's
2:27:46 stable right it's not it's not volatile
2:27:49 like there's there's lots of material
2:27:51 there's lots of things that you can
2:27:53 manifest but they're unstable and they
2:27:55 break down really
2:27:57 quick stable crystals are it's like a
2:28:00 diamond it's stable it's there it's not
2:28:04 moving um we're about to see a new age
2:28:08 of technology and innovation in every
2:28:09 sector yeah I'm calling so I'm I I've
2:28:12 written a couple of pieces on this I
2:28:13 think the age we're entering I'm calling
2:28:15 it The Great Renaissance I think think
2:28:17 we're entering a
2:28:20 creative
2:28:21 expressive
2:28:24 scientific explosion is about to
2:28:31 happen like crazy Innovation Sam Sam
2:28:34 Alman talked about this about a year ago
2:28:36 he was in an interview and he said he
2:28:39 thinks you know within within I think
2:28:42 what he said was within five years is
2:28:44 what he talked about in the interview he
2:28:46 said he thinks thinks we're going to
2:28:47 have a year like in the very near future
2:28:51 where we will accomplish in one
2:28:55 year the equivalent of 500 years of
2:29:00 technological
2:29:04 advancement and then you know deep
2:29:06 mind's Material Science thing they they
2:29:08 said you know this is 800 years worth of
2:29:10 advancement like
2:29:12 overnight so imagine that break a
2:29:16 breakthrough like the Deep Mind thing
2:29:17 with Material Science in every sector in
2:29:19 a single year here's cancer here's
2:29:23 neurological diseases here's paralysis
2:29:25 here's speech here's blindness here's
2:29:30 deafness here's all of
2:29:34 it here's
2:29:37 Fusion here's Battery
2:29:40 tech you can you can charge your battery
2:29:44 in um the same time it takes you to fill
2:29:47 up your gas tank right now and you'll
2:29:50 have a range of a th000
2:29:51 miles I think we ever just live one more
2:29:55 year we like might live forever that I
2:29:56 forget David Shapiro's got a name for
2:29:58 that there it's a it's there's a
2:29:59 something it's a something
2:30:01 threshold that if we can survive long
2:30:04 enough to hit the health threshold that
2:30:06 basically we can all extend our lives
2:30:09 for for a long long
2:30:12 time yeah that stuff's coming properties
2:30:15 are stable yeah exactly like element 115
2:30:17 well Element 115 Joker I
2:30:21 mean we know Element 115 can exist but
2:30:24 it exists in an unstable State and what
2:30:27 Bob Lazar says is that the the alien
2:30:31 craft has element5 in a stable State we
2:30:35 just don't know how to get there yet
2:30:39 well maybe we do
2:30:42 now if you don't know Element 115 is the
2:30:46 thing that makes
2:30:47 anti-gravity possible in spaceships that
2:30:52 can turn turn 90
2:30:56 degrees in in a fraction of a second
2:30:59 that would kill all of the um people on
2:31:03 board uh but because they're in an
2:31:05 anti-gravity time space continuum bubble
2:31:09 they can just do that and it doesn't
2:31:11 hurt the people inside so I don't know
2:31:14 is that real I don't [ __ ]
2:31:15 know we're about to find
2:31:22 out the AI learning lab just turned into
2:31:24 conspiracy
2:31:27 Hut all right let's go do I gotta get
2:31:30 out of here it's 10:30 let me let me run
2:31:32 over to um show and tell and see what's
2:31:34 going
2:31:37 on that's beautiful quick take
2:31:40 Jake this is me as neuros
2:31:45 spicy s pretty
2:31:52 good
2:31:54 beautiful oh that's nice what the
2:32:00 quack that's Daniel's cookies from her
2:32:03 cookie
2:32:10 GPT
2:32:11 beautiful beautiful kg wow these are
2:32:15 cool those are beautiful
2:32:17 [Applause]
2:32:21 can prompt
2:32:23 Mir mysterious cubecraft floating above
2:32:26 a dark desolate sea that's
2:32:31 cool I mean so look at that image right
2:32:34 there right so that's like that's like a
2:32:36 still from a sci-fi movie you know you
2:32:38 likely have seen we are within a year of
2:32:42 being able to
2:32:45 speak see like this into
2:32:49 existence that either as a video game or
2:32:52 as a narrative film or as an
2:32:55 animation or as you know a staggering
2:32:58 print with you know cool text on
2:33:05 it I want to see a movie that has the
2:33:08 plot of this with the this cast with
2:33:11 this story
2:33:13 line so crazy all right everybody listen
2:33:18 I'm going to get out of here let me
2:33:19 thank the mods for modding cuz uh what
2:33:23 you do helps a lot um and makes it a bit
2:33:27 more sane in here although you got me at
2:33:30 the you got me at the helm so the sanity
2:33:33 sanity is always going to be tempered by
2:33:35 that um but uh let's see here um Be not
2:33:40 Afraid good to see you digital Gods
2:33:41 sorry Lord digital Gods great to see you
2:33:45 love having you here
2:33:47 um Spike Jones good to see you Apple
2:33:49 user as always Firefly thank you for
2:33:51 your questions Winston good to see you
2:33:54 101
2:33:56 cuno are you taking us on a Rocket Ride
2:33:59 um it would be nice wouldn't it um we
2:34:01 were promised aliens this past summer I
2:34:04 know man what happened to the aliens we
2:34:06 got to get the aliens back it's
2:34:08 interdimensional h element5 like a jet
2:34:11 no not like a jet like a like a you're
2:34:15 here and now you're here here and that
2:34:18 was 18,000 M hour and then you're here
2:34:21 then you're here you shouldn't be able
2:34:24 to make that turn without crushing
2:34:26 everyone inside a jet still follows the
2:34:29 laws of physics Element 115 well it
2:34:32 still follows the laws of physics but
2:34:34 it's the weird physics like the time
2:34:37 space warp
2:34:40 bubbles Silver Fox great to see you papa
2:34:43 JJ Joker you're the best jrb 18
2:34:47 five what are one to two prompts
2:34:50 everyone should ask chat GPT one of them
2:34:52 is hey chat GPT what prompt should I ask
2:34:57 you um wait what about lightsabers
2:35:01 probably I I assume there's some crystal
2:35:03 in the 380,000 crystals that make
2:35:05 lightsabers possible I don't again I
2:35:08 don't know I got a degree in acting
2:35:10 people I'm just I'm just looking at
2:35:12 what's happening and trying to imagine
2:35:15 where it's headed I don't know anything
2:35:17 about anything Pate will tell you that
2:35:19 um imagine what the next 100 years will
2:35:21 bring I can't imagine what the next 10
2:35:23 years are going to bring D I really
2:35:25 can't um Firefly good to see you Dr J
2:35:28 great a bunch of people in here awesome
2:35:29 seeing
2:35:30 y'all um t-rock thank you uh Sky Timbo
2:35:34 Janu Tams uh anima there was reading
2:35:39 about heart repair and Nano this week
2:35:41 it's not gone maybe the people are
2:35:43 overlooking I don't
2:35:45 know um okay ASU good to see you all
2:35:48 right everybody uh I'm out of here I
2:35:51 will be back tomorrow night
2:35:53 um I think normal time 8 o'clock my time
2:35:56 Mountain you're doing great Kyle thank
2:35:58 you Spike Jones I appreciate that Sherry
2:36:00 D thank you Vicky good to see you seen
2:36:04 you a couple of times this week great to
2:36:06 see you thank you um all right everybody
2:36:09 you have a good Saturday night Sunday
2:36:11 morning and I'll see you tomorrow night
2:36:14 bye