Feb 10, 2024 - Creativity with AI: From Writing to 3D Worlds!**

Description

Chapters

Transcript

0:00 [Music]
0:23 hard on a run keeps a hand on a gun
0:27 can't trust anyone
0:33 but I was so sure what I needed was more
0:37 tried to shoot out the
0:39 sun you
0:43 coming in days when we R we flew off the
0:47 page such damage was
0:52 done but I made it
0:56 through somebody knew I was been for
1:00 some
1:01 [Music]
1:04 someone so girl leave your boots by the
1:07 bed we ain't leaving this
1:12 [Music]
1:14 room someone needs medical help for the
1:18 Magnolia
1:22 Bloom it's cold in this house and I
1:26 ain't going out to chop wood
1:31 so cover me up and know you're enough to
1:37 use me for
1:39 good you
1:42 done that's
1:45 it put your faith to the test when I
1:49 tore off your dress enrichment on
1:54 [Music]
1:56 high I so it up I SW all that stuff
2:02 forever is
2:06 time the old lover sing I thought it'd
2:10 be me who helped him get
2:14 [Music]
2:15 home but home was a dream one that I'd
2:19 never seen till you came along oh champy
2:24 sing it with
2:25 me so girl leave you put dress by the oh
2:30 hey your dress up to dry we ain't
2:32 leaving this
2:37 room to py spre breaks open wide and a
2:41 river runs
2:46 through he carries his house on the
2:50 stones like piece of drift
2:54 wool so cover me up know you're enough
3:00 to use
3:03 [Music]
3:05 me oh Champion happy Saturday night
3:09 [Music]
3:29 everybody
3:31 [Music]
3:39 place I can
3:45 my my thoughts in sweet
3:52 silence this is a place where the
3:55 feelings AR
3:59 dead from an ere exposure to
4:03 [Music]
4:06 vience this and isest place I can slowly
4:11 face the only one I truly can
4:16 know these are tears from a long time
4:21 ago got these tears from a long time
4:25 ago I need to cry 30 years old
4:30 [Music]
4:34 so these are
4:36 tears from a long
4:40 time
4:46 [Music]
4:48 go dar dar you say
4:55 me where have you been all my life time
5:01 no you
5:03 done I have been swimming seven s
5:10 [Music]
5:15 SE I'd pull them into the waters darker
5:19 in I'd have warn them but I didn't
5:24 know these are tears from a long time
5:28 ago got these tears from a long time
5:33 ago I need to cry 30 years or
5:38 [Music]
5:43 so these are
5:45 [Music]
5:47 tears from a long
5:52 time
5:53 [Music]
5:58 go
6:03 yeah that's a good
6:05 champy Champ's got some blues
6:09 tonight he is singing it up all right
6:13 happy Saturday night everybody Saturday
6:15 Night Live coming to you live from De
6:17 Colorado blue gummy M tasty all
6:24 organic all
6:28 right how is it everybody welcome
6:30 welcome
6:34 welcome my name is Kyle Shannon this is
6:36 the AI learning lab you know what we're
6:38 going to learn about
6:41 Labs no I got
6:44 nothing talk about AI we talked about
6:48 life last night we talked about
6:51 cynicism it was spicy it was
6:54 spicy and we dug deep into chat
6:58 GPT
7:00 and explored that it goes a little
7:02 deeper than it might look on the
7:04 surface cuz chat GPT kind of looks
7:07 like fancy
7:10 Google but it's more than that so if you
7:13 got questions about AI pop them in the
7:15 comments
7:17 below uh if you're a regular here an
7:21 irregular
7:23 welcome if you're a mod welcome let me
7:25 get the mod Channel up and running so I
7:27 can watch the the chaos that ensues you
7:29 a fullon add meltdown last night I saw
7:32 on the mod Channel which was awesome
7:34 while I while I was off ranting you were
7:37 like having a whole whole art
7:40 competition which was [ __ ] awesome I
7:43 thought that was beautiful it's like
7:46 well I guess he's talking for a while
7:47 we're
7:51 good is Gemini better yet Gemini is
7:54 Gemini better yet
7:56 um we can play with Gemini um
8:00 the one thing where I think Gemini has
8:05 um an
8:09 advantage you don't need Gemini Advanced
8:12 for have you heard about GPT agents I
8:15 have I think they're going to be
8:20 significant right now you can make them
8:23 using open source tools and some open AI
8:26 stuff but um they're not they're not
8:29 sort of out for the masses yet I think
8:33 chat GPT I think open AI
8:35 will will probably hit them first so
8:39 we'll see what they do with them they're
8:40 going to be interesting does chat GPT
8:43 cost so there's I got to stop eating
8:46 gummy
8:47 bears I'm the only one talking so if I
8:50 have [ __ ] in my mouth it's not good
8:56 um chat GPT has a free three tier called
9:01 uh chat GPT
9:03 3.5 and then they've got a pay tier
9:05 called chat GPT
9:07 plus which gives you access to chat GPT
9:11 4 chat GPT plus the $20 a month
9:14 subscription in my opinion in my humble
9:16 little opinion is the best deal in
9:20 software for a long long time so if you
9:23 can afford it at all I would strongly
9:25 recommend if you want to get AI literate
9:27 if you really want to understand what's
9:29 coming
9:31 get yourself chat GPT plus and start
9:33 learning what it can
9:35 do understand what code interpreter is
9:38 learn about um GPT Vision learn about
9:43 Del where it can make images um if you
9:46 can't afford the 20 bucks a month you
9:47 can also go to co- pilot.
9:50 microsoft.com which is a lesser version
9:54 of it but it's still access to the chat
9:56 gp4 model and do and it's free so I
10:00 would start there let's take an image
10:02 and use it as an inspiration for a
10:03 children's 3 to four book we could do
10:06 that we can do that
10:10 um trying to think how I do that Mo
10:13 likes co-pilot very cool Lori Ryan I'm
10:16 really curious to see what Apple
10:19 announces this summer yeah my gut tells
10:23 me what what they're you know I'm hoping
10:27 that
10:28 they
10:31 um allow Siri to retire gracefully in a
10:34 in a home in
10:40 Vermont but I I have a feeling they're
10:44 they're going to Spruce her up a bit put
10:46 her in a New Dress and uh and keep it
10:48 going I just I think the Siri brand is
10:51 is kind of crap right now um but if Siri
10:55 if Siri got better um that would be
10:57 great and I have a new operating system
10:59 where apparent apparently you just need
11:00 to say her name you don't have you don't
11:03 do the hey in front of it anymore so I
11:05 don't know she she might still be
11:08 invading um I wrote a kids book but I
11:11 don't like the illustrations that I
11:13 generate so I think I need a human oh
11:17 yeah that's interesting well have you
11:19 tried lunatick have you tried uh doing
11:22 the images in mid Journey versus DOI cuz
11:25 if if you
11:27 want if you want higher quality
11:30 illustrations I'd go i' i' experiment
11:33 mid Journey if you haven't what sort of
11:34 advantages do chat GPT have compared to
11:38 the
11:40 others it's a lot um I I'll I'll dig
11:44 into chat gbt here let I'm trying to
11:47 think where to start just in in
11:50 general chat G pt4 is actually it's it's
11:54 what's called a mix of experts model so
11:57 chat GPT 3.5 I'm going to give you some
11:59 geeky numbers here which will mean
12:01 nothing to you and they should mean
12:03 nothing to you chat GPT 3.5 has 1 Point
12:07 1.75 billion parameters right just
12:10 basically the scale of the the amount of
12:13 crap you can the connections you can
12:15 make or the vector space I don't know
12:17 what it is but it's 1.75
12:22 billion
12:24 um GPT 4 is 10 times that so what's
12:29 called a mix of experts model and it's
12:32 got 10 1.8 billion parameters parameter
12:37 model so so 1.8 or
12:40 uh yeah
12:43 1.8
12:45 trillion wait 100 it's 180 billion time
12:49 10 so 1.8
12:51 trillion so it's got 10
12:54 different large language models that all
12:58 have a different expertise a different
12:59 specialty for lack of a better term and
13:02 when you do a prompt what chat GPT 4
13:04 does it says which of these 10 models
13:07 are the best ones that can answer this
13:09 and it sort of pulls answers from all of
13:11 them and mixes them it's called a mix of
13:13 experts model so just at a
13:16 core a core technology and scale
13:21 basis it's just it's just different it's
13:24 just better bigger bad
13:26 assier and then on top of that they got
13:29 all these other
13:31 features like um code interpreter which
13:35 can write and execute python code and do
13:37 full-on data analysis it's got GPT
13:40 Vision it's got conversational AI where
13:43 you can you can actually do that in 3.5
13:45 as well where you can talk to it and
13:47 it'll talk back um it's got Dolly so it
13:51 can make images and then it's got this
13:52 concept called gpts which are these
13:55 customized gpts so so the scale of
13:59 functionality that you get within that
14:02 simple little you know search box in gp4
14:06 just it just puts it in a completely
14:07 different League from anything else I
14:09 heard open AI will start leaving water
14:11 marks on what it generates it might I
14:15 I that's that's
14:18 fine I I think all of the
14:23 um all of the
14:25 consternation about water marking AI
14:29 generated images or AI generated writing
14:32 or this or
14:34 that is is kind of wasted energy um we
14:39 need to find other ways to
14:41 authenticate our stories and it doesn't
14:43 matter if we use AI tools or not
14:47 um people that use AI regularly don't
14:51 just say write me a story and push the
14:53 publish button that's not how it's used
14:55 for the most part if I generate images
14:58 do I own them yes you you do and they're
14:59 generated from scratch um they're
15:02 they're the G in GPT stands for
15:05 generative and and it's the same with
15:07 images the way they work is they're
15:09 actually generating original stuff
15:10 they're not copying and pasting so the a
15:14 lot of the the conversation where
15:16 someone says oh chat jbt is the world's
15:18 greatest plagiarism machine and you know
15:20 the images are stealing and this and
15:22 that they're
15:23 not um but it is a Mercury area so it's
15:28 it's an easy argument to make and it it
15:30 sells more
15:32 newspapers I don't think anyone buys
15:34 newspapers anymore it gets more
15:37 eyeballs um all right so here's what
15:40 we're going to
15:41 do I'm GNA I'm going to have you go to a
15:45 website cuz because a lot of the stuff
15:46 I'm talking about is just like do I have
15:49 to take notes do I have to remember all
15:51 this Kyle this is very
15:54 exhausting couple of things I want you
15:56 to know you don't need to
16:00 you don't need to be technical to get
16:03 into this AI game I think it's one of
16:05 the biggest misconceptions and one of
16:07 the greatest fears is like I I'm not
16:09 very technical I'm not even very
16:12 computer
16:13 literate that's okay you don't need to
16:15 be well I'm not good at prompt
16:17 engineering I don't even know what that
16:19 is you don't need to be okay
16:23 just you can just play these these tools
16:26 are simple enough that any anyone even
16:30 with with no computer skills other than
16:32 you know log in and go to a website if
16:34 you can do that you can do
16:37 this so so so Step One is don't feel
16:41 like you're too technically
16:44 unsophisticated to to play in
16:47 Ai and then the other fear I think is
16:50 well there's there's a couple one of the
16:52 other fears is the robots are going to
16:54 kill us they're not the robots are not
16:56 that good yet maybe maybe maybe in five
16:59 or six years they will be and then we
17:01 have to worry but we don't have to worry
17:02 about that right now so this is not
17:05 Terminator but it is definitely
17:08 like crazy
17:11 powerful
17:13 um and rather than think about the
17:16 robots might kill
17:18 us flip the script a little bit and say
17:21 what if the robots could make me better
17:24 at what I do or what I want to do what
17:26 if what if AI could amplify my talents
17:29 and skills because that's what they're
17:31 actually
17:33 doing when you think about it that like
17:35 that it gets really
17:38 exciting and then the third sort of
17:40 barrier that I hear is well I'm too far
17:42 behind I can't catch
17:46 up we're all far
17:48 behind no one's caught up no one will be
17:52 caught up no one can be caught up no
17:53 one's an expert right now but there's
17:56 people that have podcast I see Tik toks
17:58 where
17:59 you can make money
18:02 with they seem to know what they're
18:04 doing Source cam don't worry the robots
18:07 aren't going to kill us
18:08 [Laughter]
18:14 yet we're so early in this generative AI
18:18 era the the generative AI era from my
18:21 from my standpoint started on November
18:25 30th 2022 it's not the beginning of ai
18:27 ai has been around for DEC decades what
18:30 happened on November 30th 2022 was chat
18:32 GPT launched and chat GPT does for
18:36 machine learning and artificial
18:37 intelligence what the worldwide web did
18:39 for the internet it democratizes it
18:41 makes it available to non-
18:46 technologists that's only been around
18:47 for a year we're in the primordial
18:52 soup era of all these tools they're not
18:57 great you think you you think one's in
18:59 the lead then another one bubbles
19:03 up you think you have the stuff figured
19:05 out then they change the technology
19:08 entirely so you're not as far behind as
19:11 you think
19:12 so my recommendation is if you haven't
19:15 already flipped past this
19:18 channel old guy on Tik Tok nope
19:24 nope if you still
19:27 here right
19:30 now go to chat.
19:34 open.com that's the chat jpt website
19:36 just go
19:39 there and here we'll do it
19:44 together I'm going to flip into chat GPT
19:47 3.5 mode so this is free chat GPT just
19:50 go
19:52 here and this is
19:55 it you're like but C what do I do what
19:58 do I do let me reload
20:02 this so it should show up with these
20:04 little pre-prom questions you can just
20:06 click on one of
20:08 those but you can
20:10 also ask at things like
20:14 um let's see what do we want to do what
20:17 do we want to do let's
20:20 say um I
20:23 want to come up with a fun story for
20:31 my three and five year
20:37 olds of course
20:40 boom see how fast it
20:43 is let's see what it
20:45 said this is 3.5 by the way the magical
20:48 forest Adventure once upon a time in a
20:50 land not far far away I'm going to
20:53 say
20:56 um that's a little generic period my
21:00 3-year-old really likes trucks and my
21:04 5-year-old is obsessed with
21:08 hedgehogs
21:10 period we also
21:13 live in the desert
21:16 Southwest and talk about moving to the
21:19 Great Lakes period could you make us a
21:22 story that incorporates all those
21:27 ideas
21:30 and the answer is yes of course there it
21:33 is Hank the Hedgehog's big move once
21:36 upon a time in the desert Southwest
21:37 there lived a little hedgehog named Hank
21:39 Hank was curious and adventurous HED hog
21:41 who loved exploring the Sandy Dunes but
21:43 one day Hank's family came to Shared
21:45 exciting news they were moving to the
21:47 Great
21:49 Lakes let's see along the way so it
21:52 turns into a road trip so it's a buddy
21:57 film
21:59 burst of energy carried the truck he
22:03 even made F friends with the family of
22:05 ducks by the
22:06 lake and so with the waves lapping
22:08 against the shore HK Hank embraced his
22:10 no new
22:11 life swell right think you go
22:16 um make a table that compares living in
22:20 the desert Southwest to living by the
22:22 Great
22:24 Lakes sure here's a
22:26 table so
22:29 and if you're thinking wait is there is
22:31 somewhere on the internet a table that
22:34 compares living in the desert Southwest
22:36 to living in the desert Great Lakes no
22:38 these systems the g& GPT means
22:41 generative it's generating this stuff
22:43 based on your prompt it's creating
22:45 original content for
22:47 you I could also do something like write
22:50 a python script that allows a user oh
22:55 wait write a python script that allows a
22:58 user user to answer a few
23:01 questions about what they're looking for
23:04 in a place to live and it recommends a
23:07 region of the United States for them to
23:13 live it writes code too now put that in
23:16 chat gbt 4 and compare um
23:21 it'll it will look similar gp4 will be a
23:24 little slower but um but um
23:29 it like the the gp4 writing is just it's
23:32 it's more subtly better it's it's it's
23:34 hard to to do it here but what I can do
23:36 in chat
23:37 gp4 is let's see let's let's
23:44 take let's take our Hanks Big Move we'll
23:48 take this
23:50 story we'll flip over to chat
23:52 gp4 now watch what we can
23:55 do
23:57 um I I wrote this
24:05 story and I'd like
24:08 you I'd like you to improve
24:13 it let's see what it
24:19 says your story is delightful
24:23 engaging here are some suggestions to
24:25 enhance the story
24:27 title keeping the title as it
24:29 effectively captures the essence great
24:32 character development so it's giving us
24:34 suggestions this is chat gp4
24:37 now introduce characters with depth
24:40 behind Hank introduce a few more
24:42 characters Hank's inner
24:44 Journey um setting in description vivid
24:46 descriptions plots okay so it's just
24:48 digging in so it's helping us I wish I
24:51 wish it showed images like Bard though
24:54 hang on Mo I'm about to go
24:56 there and and Bard by the way the the
24:59 image generator and Bard is kind of crap
25:02 compared to um to this so let's say
25:07 um great rewrite the story based on your
25:10 recommendations and then follow the
25:12 story with a list of five visuals that
25:16 you
25:17 think would bring the story to
25:22 life all
25:25 [Music]
25:27 right
25:34 [Music]
25:36 Once Upon a Time in the shimmering
25:38 vastness of the desert Southwest there
25:40 dwelled a small Hedgehog named Hank Hank
25:42 was a creature of curiosity and bravery
25:44 whose days were filled with rolicking
25:46 Adventures yeah the writing's already
25:48 better among the Sandy D dunes and
25:50 Majestic cacti under the expansive
25:53 Starlet Sky Hank's world was one of
25:56 endless sunshine and Shadow this this it
25:58 it didn't understand that we wrote this
26:00 for a 3 to 5y- old but that's okay it's
26:03 a little older book
26:07 now in his Newfound Paradise Hank
26:10 befriended the locals a family of ducks
26:12 by the water's edge squirrels bustling
26:13 in the trees and raccoons with Tales of
26:16 moonlit
26:17 escapades each day was a discovery each
26:20 moment a treasure all right so Hank in
26:23 the desert a family packing Journey
26:25 Through The Changing land landscapes a
26:28 glimpse of Lake Michigan new friends and
26:31 Adventures terrific I'd like you to use
26:35 a children's watercolor illustration
26:38 style and create a 16x9 wide image for
26:42 each of those five
26:47 ideas and so it
26:50 will and so this is so chat gp4 it's a
26:53 more capable writing
26:56 partner it's a more capable coding
26:59 partner it's a more
27:01 capable um analytical
27:05 partner it can see images which I can
27:08 show you in a second it can make images
27:10 which we're seeing here there's there's
27:12 Hank in the desert
27:16 Southwest and then the next one what's
27:18 the next one they're packing
27:21 right family
27:27 packing
27:33 they put they put boxes on the outside
27:35 of the
27:39 truck that's a large [ __ ] family HKS
27:44 Han's family it's a lot of
27:46 them quick question how GPT answer most
27:49 questions and give us answer in minutes
27:51 how does it do it uh I can show you that
27:54 oh I lost my little my little dual
27:57 window
28:03 so how it
28:05 works is is really quite remarkable oh
28:08 there's they're there moving cross
28:11 country but Hank turned into a
28:14 puppy and and the moving truck turned
28:17 into into a triple trailer
28:26 semi
28:28 but I I guess there's there's Hank by
28:30 Lake Michigan I don't know who this
28:31 little weirdo is but there's Hank in the
28:34 lower right there thank you Robert Rossy
28:38 Now show us chat GPT 5 I
28:41 wish he's in chat GPT 4 3.5 still only
28:44 only does text that's correct nothing on
28:47 the outside anymore though it's abstract
28:50 beautiful and there's a nice there's a
28:53 bunch of ducks there's everyone but a
28:56 hedgehog so it got some of the prompt
28:57 coherence wrong but
29:01 um so so you know this is without me
29:04 really trying so so it can do this kind
29:06 of thing um I can also do things like I
29:10 could upload an
29:12 image let's grab what do we have
29:15 here yeah we'll grab that one I can
29:19 upload an image and say
29:23 please describe
29:26 this in
29:32 detail so this was an image that
29:34 Danielle created she was
29:37 comparing Dolly 3 which is in chat GPT
29:40 with co-pilot with firefly with Gemini
29:44 the new Google engine and it ended up
29:47 being quite comical this was the do
29:49 image which is which is quite beautiful
29:52 and stunning so chat gp4 can see this
29:56 image presents a visually stunning and
29:58 Fantastical scene featuring a bear with
30:00 magnificent iridescent Wings the bear
30:03 stands by the edge of a tranquil body of
30:05 water possibly a lake or river set in a
30:07 lush mystical landscape the wings are
30:09 expansive and feathered um the bear
30:12 itself is calm behind the bear is a
30:14 series of waterfalls cascading down a
30:16 rocky Cliff the foreground features
30:18 smooth Stones the image seems to capture
30:22 the moment of Pure Fantasy so one of the
30:24 things you could do is you can you know
30:26 have it do that and say
30:30 um
30:32 recreate this image but where it is
30:38 a llama not a
30:44 bear quick question how GPT answer most
30:47 questions that give us an answer in
30:49 minutes all right I'll I'll I'm going to
30:51 come to that I I
30:55 [Music]
30:56 promise
30:58 [Music]
31:03 fantastic Bob tell him what he won
31:06 you've won a new car 3.5 is quicker
31:09 that's correct if you want things just
31:10 really fast chat GPT 3.5 is really fast
31:15 there's a llama with horn likee
31:20 ears it's still pretty Fantastical um
31:24 okay so how does this stuff
31:26 work
31:29 um
31:32 okay I can do a I can actually do I've
31:34 got a presentation where I can show you
31:36 exactly how it works but I can also talk
31:39 about I can just talk about it
31:41 conceptually so what happens
31:46 is there's a piece of technology that
31:49 Google invented in 2017 called the
31:52 Transformer and what the Transformer
31:55 allows is it allows them to
31:58 train machine learning models on lots
32:01 and lots of
32:04 data what open aai has done better than
32:07 other people so far is they went out and
32:09 they got as much of the data of the
32:11 public internet as they could and they
32:15 they they do this thing called embedding
32:17 where they turn they take documents or
32:19 you know data from the internet public
32:22 internet and they ingest it right so
32:26 they call it training data they ingest
32:27 it but what what it actually ends up
32:29 doing is it takes the individual words
32:33 from all these documents from like 4050
32:35 years of public
32:37 internet
32:39 publishing and it turns those words into
32:42 mathematical numbers called vectors and
32:44 it creates this massive thousand
32:47 dimensional mathematical space it's
32:49 [ __ ]
32:50 insane and when you do a
32:54 prompt what the system does is it
32:57 creates a probability it creates it
33:00 based on your prompt it says what is the
33:02 probability of what they call the next
33:05 token but for layman's terms what's the
33:07 next
33:09 word that's the most likely next word in
33:12 the response to your prompt and it does
33:14 it mathematically it says somewhere in
33:16 this thousand dimensional mathematical
33:18 space there's a
33:20 point and it goes and finds that point
33:23 and then there's a word associated with
33:24 that point and it comes back and it
33:26 writes the word the
33:29 and and then it says okay what's the
33:30 next most probable word and it goes back
33:32 into Laton space what they call Laton
33:35 space it's all these words that are
33:37 clustered in groups of
33:39 meaning and it says puppy and then it
33:42 goes back and
33:43 is awesome right the puppy is awesome
33:46 and so it writes that little sentence
33:48 it's doing it one word at a time or one
33:50 token at a time or handful of tokens at
33:53 a
33:53 time and it's literally writing an
33:56 original
34:00 work every single time you prompt it
34:03 it's remarkable and the fact that it
34:05 does it that fast it's from a technical
34:09 standpoint it's it's a calculator it's
34:12 saying based on all these words give me
34:15 a give me what's called a vector give me
34:17 a point in mathematical space then I'm
34:19 going to go grab that and bring it back
34:21 that's all it's doing that's why it
34:23 makes it it does what's called
34:24 hallucinating where it doesn't
34:26 understand what it wrote because it
34:27 doesn't understand what it wrote it
34:29 doesn't understand what these words are
34:31 it just thinks they're points in in
34:33 mathematical
34:34 space so it knows that this point in
34:37 mathematical space is the most likely
34:39 and then there's a word associated with
34:40 that so it so it shows you the word not
34:42 the not the math
34:44 piece and that's how it does it which
34:48 when you understand how it works it's
34:50 even more remarkable what it does the
34:52 thing that makes it remarkable is the
34:54 data it was trained on which is like as
34:57 much of the public internet as they
34:59 could get their hands on right still a
35:01 very small subset of all the data out
35:04 there but it's a lot it knows a
35:07 lot Kyle explains this better than most
35:09 I appreciate thank you I I appreciate
35:12 the
35:13 appreciation and I can show you I've got
35:15 a slide if you want to sort of see how
35:17 this thing works visually I've got that
35:20 I can
35:22 also I I I can I can do it with
35:26 boxes I sometimes explain it with
35:29 [Laughter]
35:33 boxes oh let's see how Kyle explains
35:35 this with his hands tied behind his back
35:37 I cannot listen in my company Story vine
35:41 um my nickname is the handwavy guy I'm
35:45 the handwavy guy and my co-founder is
35:47 the woman that keeps the trains running
35:49 on
35:50 time so we're we're a good we're a good
35:55 partnership I am invent this [ __ ] she
35:58 makes sure we're still in business but I
36:01 got to I got to invent like
36:05 this um have you ever asked Pate how it
36:08 does magic I well I talk to Pate about
36:11 it all the time Pate hates the word
36:14 magic well I might like the word magic
36:16 he hates it in relationship to what
36:18 these tools do for paint this is this is
36:21 this is
36:22 mathematics for any engineer this is
36:25 mathematics for any human being it's
36:27 [ __ ] magic somewhere between
36:30 somewhere between those two realities is
36:32 actually what's going
36:36 on um as far as I'm concerned the the
36:39 level of math that Pate does is like you
36:42 know magic it's like good Lord I don't
36:44 know how he does
36:45 it but that's why we have different we
36:48 have different skill sets on this planet
36:51 because some people love the math and
36:52 other people like the words and other
36:54 people like the pictures and other
36:55 people like the process
36:58 and then other people like making
37:01 [Laughter]
37:03 Flapjacks or chopping wood or carving
37:06 wood it's it's math magical that's
37:10 really good Detroit Jay next time P's on
37:12 I don't know if P's here tonight
37:14 but I think we're going to have to we're
37:16 going to have to call it mathemagical I
37:17 think that what I like about it is it
37:19 acknowledges the math but it's going to
37:22 make him cranky and I that I'm a fan of
37:24 that I think that's good every time P
37:26 hates on magic another another crypted
37:28 is released into the wild it's so
37:36 true oh man this is brilliant I can't
37:39 wait
37:41 exactly Pate it's not magical it's math
37:45 magical
37:48 TM
37:49 he's steam coming out the ears face goes
37:52 [Laughter]
37:55 red um
37:58 okay what else what else were we what
38:00 other qu what other questions do you
38:02 have let's let's do let's do kind of a
38:03 Q&A night
38:05 tonight um let's see what other images
38:09 we got here oh there's the by the way
38:12 this is the here let me show you
38:13 something so we we we started with so
38:18 this was the wings or what whatever the
38:20 prompt was it doesn't really matter um
38:22 but she did that was the bear that Dolly
38:25 3 did chat GP did and then we were
38:28 comparing it with the new Gemini Gemini
38:32 Advanced which is their super fancy
38:35 Ultra 1.0 version and it
38:38 made it made this sad
38:42 bear he's got like paint on his fur
38:45 somebody glued like these shitty
38:47 Halloween costume wings on him he just
38:50 doesn't look
38:54 happy so so there's now in the AI Salon
38:57 I'll talk about the a AI Salon in a bit
38:59 but there's now a whole a whole there's
39:02 just there's like hundreds hundreds of
39:04 these Bears we call them the Gemini
39:10 bears oh poor thing anyway um that's an
39:15 image that Corey did both of these those
39:17 are really cool I thought I had other
39:20 something else in here
39:22 guess that's kind of
39:24 cool most of these images are that
39:27 people in the AI Salon have
39:30 created that's a cool
39:38 [Music]
39:41 one gorgeous it's gorgeous it's
39:45 gorgeous poor Gemini bear I know poor
39:48 Gemini bear everybody picking on a
39:50 Gemini bear he's
39:55 cute thanks everyone I appreciate you
39:58 letting me know I'm new to this channel
40:00 this is awesome so fly life fly life
40:05 campaign what do you want to know like
40:06 where where are you in the journey are
40:08 you brand new what how can how can we
40:10 help what can we what can we show teach
40:13 my shirt's a little interfering with the
40:17 uh with the letters maybe I should go
40:19 change my
40:20 shirt did you guys talk about the
40:23 women's event uh I haven't talked about
40:25 it yet shop panda so um Anne Murphy
40:28 today in her weekend GPT jams did the
40:33 inaugural
40:34 um it's
40:37 not what was it
40:39 called I forget the name of
40:42 it it was not the women in AI it was the
40:46 AI who remembers the name of
40:49 it she leads AI the inaugural she leads
40:53 AI event so Corey Sandler from Cory
40:57 Sandler Pottery who's here on Tik Tok if
40:59 you don't follow Corey Sandler Pottery c
41:02 r i Sandler Pottery you should follow
41:05 her on Tik Tok she's amazing um she
41:08 moderated a panel of women in Ai and so
41:12 it's a lot of people that hang out here
41:13 it's a lot of people yeah exactly tap
41:15 tap it was it was a really beautiful
41:17 event and just you know they were
41:18 talking about what they're doing in AI
41:21 how they discovered it women and stem um
41:24 a lot of them were talking about you
41:26 know
41:27 in my day I was like the only woman in a
41:30 300 person group of game testers at
41:34 where was it at Microsoft or at
41:36 somewhere at some big
41:38 company um amazing it was like it was
41:41 really amazing it was really good uh I'm
41:43 sure Mary Mary I'm sure it was recorded
41:46 um so so check
41:48 out check out check out Joker you're
41:52 going to get yourself in trouble I'm not
41:54 saying that out loud you bastard you're
41:56 not taking me me down I'm not getting
41:59 I'm not getting
42:04 cancelled oh man I find that I hit my
42:07 usage Cap all the time I have to wait a
42:08 few hours before I can use it again I
42:11 Know Dan Le that's that's a drag um if
42:14 you've got another person you could
42:16 share it with you could upgrade to to
42:18 chat GPT team and it's got higher rate
42:21 limits but I it that's a that's an
42:23 actual problem but it'll get better it's
42:25 just it's just shitty right now Lori
42:29 Lori drusk and Ryan it was great
42:31 appreciated hearing their stories and
42:33 areas of expertise it was great to have
42:35 such a diverse group yeah it was amazing
42:37 we had with Alicia or Alysa Alysa who's
42:41 a 17-year-old high school student who is
42:45 kicking ass with AI she's she she's
42:49 doing she she talked about so what she's
42:52 doing as a 17-year-old AI literate woman
42:55 is teaching the adults how to use this
42:57 [ __ ] and charging for it she's got a
43:00 she's got a good business going it's
43:02 [ __ ] fantastic just amazing and and
43:06 that that in particular is really
43:07 refreshing because it it it might be an
43:10 artifact of the fact that you know this
43:13 channel I I'm older and and maybe
43:15 attract an older audience but what I'm
43:17 seeing is there's a lot of older people
43:20 gen xers that are really into Ai and
43:23 there's a lot of young people that don't
43:25 give two shits about it so was really
43:27 nice to hear Alyssa and and oh Mi yeah
43:30 MIT is actually pitching Alyssa so one
43:32 of the things I suggest in here is as
43:35 you learn about AI like get curious
43:38 about it start playing with
43:39 it and start talking about it share it
43:43 put it on LinkedIn put it on Facebook
43:46 just put it out in the world tell people
43:49 so Alyssa's been doing that and MIT
43:51 reached out to her and said we should
43:53 talk you should come take classes here
43:56 so so you know like it it really does
44:00 like this channel is full of people that
44:03 just started going well I I don't I
44:04 don't really have anything to talk about
44:06 I don't really know what I'm doing I'm
44:07 like talk about it anyway and they're
44:09 like oh my God people think I'm an
44:11 expert I'm like I know tell me about
44:15 it I am a very mature student can you
44:19 advise how gp4 can improve my writing
44:22 advise advice on prompts
44:25 um
44:27 yeah I
44:30 can don't worry about prompt engineering
44:34 you can learn it you can Google it you
44:36 can go watch YouTube videos on it I'm
44:39 sure there are downloadable PDFs that'll
44:41 give you all sorts of recipes there are
44:44 people you can make money with that will
44:47 sell you a PDF full of a thousand
44:51 prompts um screw them for one
44:55 thing
45:03 understanding how to talk to these
45:05 systems is important but depending on
45:07 the system and depending on your use
45:10 case it may not be that
45:12 important so for example here let me
45:15 show you something
45:20 cool we're going to go to chat GPT and
45:23 I'm going to say
45:25 make
45:28 make me a make me a photo of a
45:36 1969 that's 2969
45:39 1969
45:41 GTO Judge in
45:43 Orange by the
45:46 beach so that is what I would call a
45:50 relative Earth rocker I'm a very
45:51 immature student perfect well you're
45:54 good for this channel make me a photo of
45:56 a 1969 uh GTO Judge uh in Orange by the
46:01 beach pretty crappy prompt
46:03 but it will go do that and so we'll see
46:07 what that looks
46:24 like all right lovely it's a lovely GTO
46:27 Judge it actually looks like a GTO judge
46:29 in the in the front looks like a 69
46:31 crazy now if I click on this
46:34 picture and I go to this little eye
46:37 right
46:38 here this is the prompt that was
46:41 actually
46:42 submitted a 1969 GTO Judge in bright
46:46 orange parked on a sandy beach with the
46:48 waves gently crashing in the background
46:51 the sky is clear with a few wispy clouds
46:54 reflecting the golden hours sunlight
46:56 casting a warm glow on the car's Sleek
46:58 line and in and it goes on and on and on
47:00 and
47:02 on I didn't write that chat GPT wrote
47:05 that so these systems are getting
47:07 increasingly sophisticated in taking
47:10 your
47:11 intent and helping you with it
47:15 now some systems don't do that really
47:17 well so some systems you have to you
47:19 have to be better like if you want to
47:20 use mid journey to make images you've
47:23 got to kind of understand their syntax
47:26 and their prompting language and their
47:27 modifiers and you have to it's
47:29 essentially like writing command line
47:31 codes to write images the latest version
47:34 of mid Journey version six it doesn't
47:38 have to be as structured you can just
47:39 talk in like flowery language like this
47:42 and it does pretty good but all of the
47:44 versions before that you had to be quite
47:46 technical
47:49 so don't think that you have to become a
47:52 master of prompting certain jobs you
47:55 might get you might you might want to
47:57 learn that but but I don't think it's a
48:00 requirement as we move into the era
48:03 where these systems uh do reasoning and
48:07 have what are called autonomous agents
48:09 you're not going to need to do much
48:11 prompting at all you're just going to
48:12 kind of give it a goal and it's going to
48:13 go so so I wouldn't worry a ton about
48:15 that
48:16 now if you w to if you want to
48:21 um you want advice on how to write
48:23 better with chat GPT here's the biggest
48:26 advice I can give
48:28 you don't think of chat GPT like you
48:30 think of Google so so and here's what I
48:33 mean by that how I think of Google is
48:36 you come up with the perfect prompt and
48:37 ideally with Google you want a short
48:39 concise prompt right I want to find you
48:42 know alpaca breeders in the southwest
48:45 bang and it will find you alpaca
48:47 breeders in the southwest you give it a
48:49 prompt and it gives you an answer and
48:50 you're done you're Off to the Races
48:53 think of chat GPT much more like
48:57 you've been partnered up with and
49:01 Incredibly well-educated like incredibly
49:04 well-educated
49:06 clueless
49:08 intern who's who's got the enthusiasm of
49:11 a golden retriever
49:12 puppy okay that's kind of what chat GPT
49:15 is so there's a couple of things you can
49:18 do it do do for it because it's clueless
49:21 one of the things that helps a lot is
49:23 you can tell it that you you can give it
49:25 a roll so so I could
49:27 say I want you to act like
49:31 a Harvard trained writing
49:35 partner who's published four books and
49:39 had one of those books turned into a
49:41 screenplay
49:43 period as we work together I want you to
49:48 take my ideas and help me flesh them
49:51 out so you can kind of give chat GPT a
49:54 role and then rather than think of about
49:56 I've got to come up with the perfect
50:00 prompt think of it more like you're
50:02 entering into this conversation and and
50:06 the way I like to do writing is I don't
50:10 have it write things right away I'll
50:12 give it context so so right now I just
50:14 told it I wanted to be a whatever a a
50:17 Harvard trained writing partner um now I
50:21 might
50:23 say I've been tasked with wri in four
50:27 articles for
50:30 LinkedIn over the next 6
50:37 weeks
50:39 and the topic has to include artificial
50:45 intelligence the future of
50:48 work
50:53 and the impact of employee
50:58 training all right that's a fantastic
51:05 assignment so so it's it's already
51:08 coming up with let's let's outline a
51:09 coherent strategy for your articles
51:11 article one the AI revolution in the
51:14 future of work introduction briefly
51:16 introduce how AI Technologies are
51:17 reshaping Industries main body so it's
51:19 outlining each article reshaping
51:21 employee
51:24 training
51:27 article three AI in employee training
51:30 enhancing Learning and Development
51:32 number four navigating the ethical
51:34 practices um then then I might say
51:46 um I like number
51:48 one but I want to make sure to
51:53 include some of the fears and and
51:56 skepticism about
51:59 AI weighted more heavily in article one
52:04 than it will be in article 4 period can
52:06 you give me a detailed outline for
52:08 article one with this new
52:14 guidance so so how I'm how I'm
52:17 interacting with it and this is this is
52:19 you know when I work with cat GPT I tend
52:21 to work like this I I'll sort of
52:24 just like I've got a writing partner
52:27 like I've got a remote coworker that's
52:30 really good and really fast with writing
52:32 it's like when you bring a new person
52:33 into a project fill them in on all the
52:35 backstory Earth rocker that's it
52:38 exactly where where Google wants you to
52:40 be as sparse as possible because it's
52:43 going to just sort of go look in its
52:44 index and and pull all the [ __ ] out of
52:46 the index chat GPT needs context and
52:49 it's it's the nature of the way it works
52:51 but I described before it's like there's
52:52 this mive massive mathematical space
52:57 when you give it more context it Narrows
53:00 sort of the field of where the answers
53:02 are going to come from so you tend to
53:04 get better more accurate you know less
53:08 hallucination kind of answers if you
53:10 focus focus the conversation and then
53:13 what I would do is I would just go on
53:14 and on in this style like I you know I'd
53:17 take this outline and I say okay that's
53:19 great but I I want to switch you know
53:22 the first three sections around and I
53:23 want to focus more on this than that
53:26 just saved my 25-year-old daughter and
53:28 her boyfriend with a landlord situation
53:31 she's a chat gbt fan now there you go
53:34 yeah it's really it's really good if you
53:35 wanted to write letters legal letters um
53:38 read legal letters it can do that too um
53:41 it's really good at analyzing other
53:43 documents but but I that's the biggest
53:46 thing I can say is
53:49 just except the fact that you're really
53:52 lucky to be alive right now that you
53:56 have at your hands the most powerful
53:59 writing tool in the history of
54:02 humanity and just treat it like you've
54:05 got kind of a
54:07 simpleton incredibly well-educated
54:09 writing partner and and it'll change
54:12 everything don't worry about prompting
54:15 this and learning that you can do all
54:18 that but you don't have to so I hope
54:21 that
54:24 helps
54:35 he wouldn't do anything love the puppy
54:37 remark thank you great simple
54:39 explanation you're welcome it it well so
54:43 I got to tell you you know I talk about
54:45 it being like a golden retriever
54:49 puppy you know how when you're dealing
54:51 with humans humans kind of suck
54:54 sometimes like even when they don't mean
54:56 to where you're like working let's say
54:58 you got a writing partner and it's your
55:00 best friend and they're a really good
55:02 writer and then you come up with a
55:04 stupid
55:05 idea what does your friend
55:08 do they make fun of you right even if
55:11 it's in justest and even if it's loving
55:13 like oh that's a good
55:16 idea and like a little part of your soul
55:18 dies inside right you're like a [ __ ] I
55:21 just I know it was a stupid idea and
55:23 then you edit yourself the next time I'm
55:25 not going to say that out loud again
55:27 what happens with cat GPT is you're like
55:30 um I kind of have a shitty idea it's
55:32 like hey what a fantastic idea that's a
55:34 unique spin on things it's it's always
55:37 positive and it's always
55:39 reinforcing and it's always
55:42 encouraging and then it'll go it'll say
55:45 oh that's great let's explore that and
55:46 it'll write it out and sometimes it
55:49 writes it out and you look at it and
55:50 you're like that was a horrible idea but
55:53 you know now that it wrote it out like
55:56 that and I can see it now I have an idea
55:59 and then you tell it that idea and it's
56:00 like o excellent
56:03 idea or sometimes it writes it out and
56:05 you're like that was better than I
56:08 thought I
56:10 didn't I hate to admit it but I'm kind
56:12 of a
56:15 thing there's something about it that
56:19 is it's it's constant positive
56:24 reinforcement and it's
56:27 weird human beings don't do that they
56:32 should
56:33 right we we we we live in a divisive
56:37 time where the past 10 15 years we have
56:40 bought into whatever the [ __ ] story has
56:43 been told to us that we should hate the
56:44 other side right so we're in this
56:46 divisive thing where we're yelling at
56:48 each other in coffee shops because
56:51 someone said the word frog wrong I
56:54 doesn't [ __ ] right we're just in this
56:56 weird time so I I at least personally
56:59 feel like hyper sensitive to
57:02 like can I say this
57:05 word am I get in trouble and what
57:08 happens with chat GPT is you can [ __ ]
57:10 say anything well you you can't ask it
57:12 to do naughty things you'll see it
57:14 safety guard rails come up sometimes
57:16 which is a drag if you're trying to
57:18 write
57:20 fiction certain kinds of fiction are
57:22 tough but for the most part
57:27 like like one of the things that really
57:29 blew me away about chat jpt early
57:31 on was you you know how in society right
57:34 now there's certain words that you you
57:36 literally can't say the
57:37 word right you know you can't say trans
57:41 right now without it you know instantly
57:44 politicizing the room right you're
57:47 you're either for it or against it and
57:50 like if you say that word I'm GNA
57:52 determine in three seconds if I should
57:54 hate you or or high five VI that's the
57:58 world we live in right now so there
58:00 there are certain words phrases topics
58:03 Concepts that you literally can't bring
58:05 up unless you know who you're talking
58:08 to and so it's it it's this stifling
58:11 thing where where people are afraid to
58:13 talk and what I found early on with chat
58:16 GPT is I'd be working on something and
58:19 I'd have a question and i' I'd be like
58:21 oh I've got a question about that
58:22 concept oh [ __ ] I can't talk about that
58:25 I was like oh well it's just chat jpt
58:27 and then I would ask it about that
58:29 concept and a remarkable thing happened
58:32 chat jpt didn't pick
58:34 sides chbt actually just gave a nuanced
58:38 answer that said here's how it works and
58:41 then I could go oh well some people feel
58:43 this way and some people feel that way
58:45 what's going on there and it would do
58:47 things like give me the historical
58:49 context for how those points of views
58:51 emerged and I was like holy [ __ ] I just
58:54 got educated and I don't feel
58:57 triggered that I said a triggering
59:02 word it's kind of [ __ ] remarkable and
59:05 what's amazing is what you hear from the
59:08 naysayers that are on the sidelines is
59:10 well these models are biased and they're
59:12 this and they're
59:13 that okay they might be biased but I'll
59:16 tell you what they're a fuckload less
59:18 biased than the human beings I interact
59:23 with and what's nice about the
59:25 technology is you can actually measure
59:27 the bias and you can either prompt
59:29 around it or you can let companies like
59:31 open AI know hey we got a [ __ ]
59:33 problem here this is all pasty white
59:36 guys let's go ahead and address
59:39 that and then they do what they do to
59:42 fine-tune the models to try to address
59:44 that and things are getting
59:46 better but it's
59:48 remarkable to have what's effectively a
59:52 creative partner that's always there for
59:54 you
59:55 always encouraging you can you can talk
59:58 about weird [ __ ] with it and and it will
1:00:00 give you actual answers insightful
1:00:05 answers and so just think of it like
1:00:08 that you're
1:00:10 lucky to be living at a
1:00:14 time where for free or for 20 bucks a
1:00:17 month you get to interact with the most
1:00:20 sophisticated writing tool in the
1:00:22 history of humanity and it's only
1:00:24 getting better
1:00:28 and the other here's the other thing I
1:00:29 would
1:00:32 suggest don't take any of it too
1:00:35 seriously like play like like if you've
1:00:38 got a big important writing assignment
1:00:40 start with a really stupid frivolous
1:00:42 writing assignment all right uh let me
1:00:45 write a story about a gummy bear that
1:00:47 comes to life and goes to New York City
1:00:49 to pursue an acting
1:00:50 career
1:00:52 great and and just play with that
1:00:55 because what it'll what it'll free you
1:00:56 up to do is sort of push into Corners
1:00:59 where if you were like working on a work
1:01:01 project you'd be like oh I got got to be
1:01:04 all like this but if you do the stupid
1:01:07 story first and then you're like oh let
1:01:09 me see an image for that oh that's kind
1:01:11 of cool and let me run down this Rabbit
1:01:12 Hole let me run down this Rabbit
1:01:14 Hole what you do is you kind of find the
1:01:17 edges of where chat GPT is good and
1:01:20 where it maybe isn't good and then when
1:01:22 you switch over to your work project
1:01:26 you'll have a much broader understanding
1:01:28 of the capacity of the
1:01:31 tool and
1:01:33 so one of the things one of the reasons
1:01:35 I like doing these is I tend to use
1:01:39 nonsensical examples when I when I do
1:01:41 stuff just so that I can sort of because
1:01:45 we're we're all right now in a in a room
1:01:48 feeling for where the walls are and
1:01:49 where the furniture is it's completely
1:01:51 dark we're all kind of feeling for
1:01:53 what's in this room where the
1:01:55 walls and
1:01:58 then the walls are moving all the time
1:02:01 so last week there was a wall here oh
1:02:03 [ __ ] there's not a wall here oh oh it's
1:02:05 better it's all of a sudden it's better
1:02:08 in this place it used to be shitty at
1:02:11 the only way you can find those walls is
1:02:13 to play so just play
1:02:18 play create pizza recipes create
1:02:22 Hedgehog shampoo formulas it like
1:02:24 anything
1:02:26 and then do work and
1:02:30 then talk about it share
1:02:34 it join the AI Salon which I'm going to
1:02:37 talk about right now or a group like it
1:02:40 get in community of people who are
1:02:42 playing with this cuz cuz it's so hard
1:02:46 to explore this stuff on your own it's
1:02:49 moving too fast it's too nuanced it's
1:02:52 too all over the place
1:02:56 but you get in a community and all of a
1:02:58 sudden someone will pop on and go oh I
1:02:59 did a cool thing look at this and you're
1:03:01 like holy [ __ ] I've been trying to
1:03:02 figure that out for a week and they just
1:03:03 did it and they just taught you how to
1:03:05 do
1:03:08 it so what I want you to
1:03:11 do is I want you to go to that URL the
1:03:15 salon.
1:03:20 A and what that is is not a not a
1:03:26 hair styling Studio
1:03:30 clearly it's a salon like an idea Salon
1:03:33 like a place where smart generous
1:03:36 interesting people show
1:03:39 up and connect and share ideas and share
1:03:43 what they're working on and Inspire one
1:03:46 another and challenge one another and
1:03:49 occasionally piss one another
1:03:50 off that's part of it when you got
1:03:53 community right
1:03:58 when you go there you're going to go to
1:04:00 this this link
1:04:02 tree and the first the first link here
1:04:06 says AI Salon Community go to
1:04:11 that and that's going to drop you into
1:04:13 welcome to the AI Salon Earth rocker
1:04:16 this community rocks it
1:04:18 does this community is
1:04:20 amazing and so you know the the little
1:04:23 picture here right it's not one person
1:04:25 on an adventure it's a group of people
1:04:27 on an adventure it's a group of people
1:04:29 choosing to be on an adventure
1:04:32 together that's what this community is
1:04:35 so if you're curious about AI or maybe
1:04:38 you've you know you've drunk the
1:04:39 Kool-Aid and you're like holy [ __ ] this
1:04:40 AI thing is remarkable it's amazing but
1:04:42 you're the only one in your whole
1:04:44 [ __ ] life that feels that way
1:04:46 everyone else is like Jane could we not
1:04:49 talk about AI please I just want to go
1:04:53 to FUD Ruckers and get a twisted
1:04:55 Margarita can could you could could we
1:04:57 not could you not talk about the chat
1:05:00 GTP please Jane
1:05:03 thanks if if that's what your life
1:05:05 sounds like right now then you might dig
1:05:08 this
1:05:11 group so I want you to go here I want
1:05:13 you to actually take in what we're
1:05:15 about who we are why we do it and then I
1:05:19 want you to explore our values so the
1:05:21 values are who this group already is I
1:05:25 want this group to grow but I also want
1:05:27 it to grow and maintain what's really
1:05:29 special about it right now what's
1:05:31 special about it right now is the people
1:05:33 in here don't
1:05:36 suck like like that's a like this is a
1:05:39 this is just like a don't be a douche
1:05:41 right don't be an
1:05:44 [ __ ] well what's that actually mean
1:05:46 well that's what these values are what
1:05:48 these values are is here's how people
1:05:50 are already behaving in this community
1:05:52 that makes it like that that makes makes
1:05:54 it a remarkable Community they're
1:05:57 curious oh I wonder what would happen if
1:05:59 I do this oh did you try this did you
1:06:01 try that right they're not coming in
1:06:03 with their arms crossed defensively
1:06:05 they're coming in like with that with
1:06:07 monkey mind I think they call it in in
1:06:09 in yogur Zen stuff right beginner's
1:06:13 mind no one's an expert right now if
1:06:16 anyone tells you they're an expert these
1:06:20 [ __ ] they're full of
1:06:22 [ __ ] so come in be curious and then and
1:06:26 then the next one generosity be generous
1:06:30 if you learn something tell people what
1:06:32 you learned here's how I did this prompt
1:06:35 I tried something different this time
1:06:36 and look what it gave me well I got this
1:06:38 image like Corey Sandler she gets these
1:06:41 images that are [ __ ] beautiful and
1:06:43 then she keeps going for like 30 or 40
1:06:46 more
1:06:47 iterations it's like a remarkable way of
1:06:49 working I never would have
1:06:52 considered and the only reason I know
1:06:54 that is because she's generously sharing
1:06:56 what she's up to exploration is one of
1:06:59 the values like are you
1:07:01 adventurous or you just want to lock
1:07:03 things down this we're not in a phase
1:07:06 right now in AI where anything is locked
1:07:09 down so you kind of have to be
1:07:10 adventurous collaboration are you
1:07:12 collaborative hey let's put on a show
1:07:14 you're good at this I'm good at that
1:07:16 let's let's do something together
1:07:17 inclusion everyone's coming from
1:07:20 everywhere right now
1:07:23 right so don't judge them if they're
1:07:26 just getting started in AI don't judge
1:07:30 them if they come
1:07:31 from they run storage
1:07:35 facilities Jim Ross who's a part of this
1:07:38 community he he buys and manages storage
1:07:42 facilities and teaches other people how
1:07:44 to do
1:07:46 that in a lot of AI communities they
1:07:49 would look at him and go well you
1:07:51 know he's he doesn't really do AI
1:07:55 I can tell you this Jim Ross is better
1:07:58 at AI than almost anyone I know he takes
1:08:01 what he learns and he applies it to his
1:08:02 [ __ ] business like like that he's
1:08:06 remarkable at it so be
1:08:09 inclusive
1:08:11 empathy there's so much fear and there's
1:08:14 so much anger and there's so much
1:08:15 misunderstanding about what these tools
1:08:17 actually
1:08:20 are
1:08:22 that with we had someone come in last
1:08:25 night they were very angry they were
1:08:26 very
1:08:28 cynical I got a little reactive and at
1:08:31 the same time I have deep deep empathy
1:08:34 for that
1:08:36 fear this is
1:08:39 profound profound stuff I'm 11 and a
1:08:42 half years into an automated video
1:08:44 storytelling
1:08:47 company and Along come a set of tools
1:08:50 that are going to automate
1:08:52 everything and take the manual
1:08:57 narrative process that I've developed
1:08:59 over 12 years and make it instantly
1:09:01 available to everyone my entire
1:09:04 business is if I don't do this right at
1:09:09 threat
1:09:11 right this [ __ ] can be really
1:09:15 scary so this is a community that
1:09:17 recognizes people might be pissed off
1:09:20 people might be scared people might be
1:09:22 reticent people might be skeptical and
1:09:24 rather than judging it say let me
1:09:27 understand why you feel that way oh yeah
1:09:30 I can understand that you want to see
1:09:32 something
1:09:35 cool and then the last one is gratitude
1:09:38 we just we just added this value three
1:09:40 nights ago thank you Danielle for
1:09:42 inspiring
1:09:44 this we believe that gratitude is a
1:09:46 practice gratitude is not just a feeling
1:09:48 but an action take the time to
1:09:51 acknowledge the efforts support and
1:09:54 successes of the extraordinary people in
1:09:56 this community especially if someone
1:09:58 inspires you go out of your way to
1:10:01 acknowledge them tag them in your post
1:10:03 reach out and let them know they made a
1:10:08 difference thank you
1:10:17 Empower does this take the world
1:10:19 internet consideration I don't quite
1:10:22 know what that question means
1:10:32 one I'm I'm on a new Mission
1:10:35 lately one of the things that I'm and
1:10:37 this is
1:10:40 a this thing is written about gratitude
1:10:42 to other people but one of the things
1:10:45 I'm noticing in this
1:10:48 community and one of the things I'm
1:10:51 noticing in general with people who are
1:10:55 leading AI
1:10:58 voices is that there's a lot of
1:11:00 apologizing for how not expert someone
1:11:03 is there's a lot of apologizing oh
1:11:06 that's beautiful thank you very much
1:11:07 Robert
1:11:09 Rossi there's a lot of apologizing for
1:11:11 well I'm not really an expert well this
1:11:14 just took me a few minutes well you know
1:11:16 I'm not really that
1:11:19 good one of the things I want to
1:11:21 encourage is gratitude for
1:11:26 yourself acknowledge that no one's an
1:11:30 expert right now if you if you are doing
1:11:33 this at all like if you're showing up to
1:11:35 these Tik toks if you're part of the
1:11:37 salon Community if if you're just
1:11:39 getting started with it but you're
1:11:40 curious about it and have an open
1:11:42 mind you are farther along than 95% of
1:11:46 the people on the
1:11:49 planet I one of the things I want to
1:11:51 encourage in this community is I want us
1:11:53 to all including myself stop [ __ ]
1:11:57 apologizing for how not whatever you
1:12:02 are and I I think it lives in this value
1:12:05 of
1:12:06 gratitude and so gratitude is outward
1:12:09 toward other people I think it's also
1:12:10 inward toward toward yourself the fact
1:12:13 that you're curious the fact that you're
1:12:14 showing up
1:12:16 here is
1:12:18 remarkable don't apologize anymore so
1:12:22 anyway so that's what the about that
1:12:24 these are who the people that are in
1:12:26 here now here's the thing about values
1:12:28 when you have a strong culture when you
1:12:30 have a strong
1:12:32 Community it's good and bad if if in
1:12:35 reading these values if they if they
1:12:37 sort of turned your stomach and you're
1:12:39 like o that sounds like an awful
1:12:41 Community really then this isn't your
1:12:44 people if you hear these values and
1:12:46 you're like oh my God is it really like
1:12:48 that yeah it's actually really like that
1:12:49 then these probably are your people so
1:12:51 my request is really read what we're
1:12:54 about really read the values and really
1:12:55 take them in and if this feels right to
1:12:57 you then you should
1:12:59 join and so how do you join I don't know
1:13:01 it's mighty networks is the is the
1:13:03 platform we're on but that doesn't
1:13:05 really matter just you know make an
1:13:07 account I think and then the first thing
1:13:10 my request for you to do if you want to
1:13:11 join this community is introduce
1:13:13 yourself right take that value of
1:13:15 generosity and generously introduce
1:13:18 yourself here's who I am here's where I
1:13:20 am with AI here's what I do here's my
1:13:23 hopes and dreams
1:13:26 oh but Kyle that's kind of scary I don't
1:13:28 know anyone they're going to judge
1:13:29 me right they
1:13:33 won't be
1:13:36 generous and you will see the generosity
1:13:38 in this community if you introduce
1:13:40 yourself and tell people where you are
1:13:41 they'll be like oh you should go here
1:13:43 you should check this out oh I started
1:13:45 there too I understand I understand what
1:13:47 you're going through I have a similar
1:13:49 kind of job I work for a big company too
1:13:51 and they won't let me use chat GPT it
1:13:52 sucks here's what I'm doing
1:13:55 and then you know check out the in the
1:13:57 community section there's a water cooler
1:13:59 you can talk about anything there's a
1:14:00 show and tell channel that can be really
1:14:01 inspiring there's a news channel there's
1:14:04 also AI tools and resources this one is
1:14:07 worth the price of admission oh by the
1:14:09 way you're probably thinking oh how much
1:14:11 does this community cost I see what this
1:14:13 is this is a this is a multi-level
1:14:15 marketing
1:14:16 scheme you can make
1:14:19 money no it's
1:14:22 free
1:14:26 it's free it's fre P it's fre why is it
1:14:30 fre because it's important that people
1:14:32 have a place to collaborate and and
1:14:40 congregate so the the the tool
1:14:44 list if you're just getting started and
1:14:46 you're like how do you figure out where
1:14:48 to even start that's what this list
1:14:51 is you could spend the next three months
1:14:53 months not looking at any other thing
1:14:55 with AI just grabbing [ __ ] off this list
1:14:58 and exploring it and and not even
1:15:00 scratch the
1:15:01 surface so the first thing on this list
1:15:03 is chat GPT and I talk about here's what
1:15:06 you get with chat GPT 3.5 the free one
1:15:09 and here's what you get with chat GT4 I
1:15:12 strongly unless you can't afford the 20
1:15:14 bucks a month and even then I would say
1:15:16 what can you do to earn an extra 20
1:15:19 bucks a month or not spend an extra 20
1:15:21 bucks a month to get this I would learn
1:15:24 everything you can about chat
1:15:26 gbt chat bt4 in particular every one of
1:15:30 these bullet points is a whole another
1:15:31 Rabbit Hole to go
1:15:33 down and if you can understand how to
1:15:35 quickly navigate between these which
1:15:37 that's how it's designed you can quickly
1:15:39 navigate between these
1:15:41 modes you can do some remarkable
1:15:48 [ __ ] I just sent this live link to my
1:15:51 friend in Ukraine Rea via the tel app oh
1:15:54 that's awesome
1:15:56 fantastic we've got our for my company
1:15:59 Story vine we've got um our developers
1:16:02 are in Prague and we've got a number of
1:16:03 developers we've worked with from
1:16:05 Ukraine did chat GPT learn from the
1:16:09 world's internet or oh that's the
1:16:11 question okay did chat learn from the
1:16:12 world's internet or just English
1:16:14 speaking it learned from the world's
1:16:16 internet rock paper scissors but there's
1:16:19 there's an English bias to what's what's
1:16:22 on the internet um
1:16:24 chat jpt knows all the languages so so
1:16:27 it knows all the languages I would say
1:16:30 that it is the the training set was
1:16:32 likely heavily skewed toward English and
1:16:35 probably toward America just because of
1:16:37 the nature of how the internet came
1:16:39 up um I also don't know like I actually
1:16:43 don't know how much open AI put into
1:16:47 making sure that there was some balance
1:16:49 in terms of content sources so it's
1:16:52 possible it's more balanced than I
1:16:54 thought I've been telling people I'm not
1:16:57 on on my networks not realizing AI
1:17:01 Salon oh I'm not on Mighty networks not
1:17:04 realizing AI Salon is on Mighty networks
1:17:07 what's wrong with Mighty
1:17:08 networks Mighty networks is just it's
1:17:11 just a hosting platform it's a community
1:17:13 management software it's it's somewhere
1:17:15 between Discord and Facebook it's fine
1:17:18 yeah the yeah I mean I think the
1:17:20 internet skews toward the Western World
1:17:22 I think that's right one love
1:17:24 and um and I think over
1:17:27 time see here's the thing and this this
1:17:30 is one of the things Sam mman is is
1:17:32 passionate about is he wants to build
1:17:35 these tools and get them in the hands of
1:17:38 people as soon as poss as soon as
1:17:39 they're they feel they're safe enough he
1:17:41 wants them in the hands of people
1:17:42 because he wants to understand what
1:17:44 works what doesn't you know what's
1:17:47 what's broken what's out of balance so
1:17:49 that they can fix it it wasn't a
1:17:51 negative comment just confused oh since
1:17:53 mighty networks is new to a lot of us oh
1:17:55 I see
1:17:57 yeah um okay so start with chat GPT and
1:18:02 then this list has um eight other large
1:18:05 language models Claude Gemini Advanced
1:18:08 perplexity AI co-pilot or kind of the
1:18:10 big four then you got Pi grock llama 2
1:18:15 llama 2 is actually the one from meta
1:18:17 that's been open sourced if you're
1:18:21 geeky um the open source Community right
1:18:24 now is on fire there are so many open-
1:18:28 Source projects if you're a developer in
1:18:30 any way I would strongly recommend you
1:18:33 get involved in open source projects
1:18:35 with with llama and and whatever um Lang
1:18:38 chain um autonomous agents there there's
1:18:42 all sorts of [ __ ] you can be building
1:18:44 right now um and you can be you know
1:18:47 sort of building on the back of the
1:18:48 Llama 2 stuff that's out there um that
1:18:51 that's a whole world in and of itself I
1:18:53 don't really cover it because it's it's
1:18:55 just way too deep it the waters are too
1:18:57 deep for my knowledge base um but
1:18:59 there's a there's a lot there um then
1:19:02 there's a whole section on image tools
1:19:05 Del mid Journey Kaa Deco here Leonardo
1:19:08 and
1:19:09 Firefly every one of those is is a
1:19:13 rabbit hole Leonardo in particular it
1:19:15 has gotten very very sophisticated
1:19:17 there's a there's so much you can do
1:19:19 with it and this is kind of the gateway
1:19:21 drug to doing stable diffusion
1:19:23 on the you know on the back end with the
1:19:25 open source Community that's a whole
1:19:27 another world so just like llama 2 is
1:19:30 the is the gateway to the open- source
1:19:32 Community for large language models
1:19:33 stable diffusion does that for images
1:19:36 then you got video creation Runway ml P
1:19:38 laabs haen did Visa then you've got
1:19:42 music and audio there's a new one that I
1:19:45 haven't really even played with that
1:19:46 much called Splash Pro where you can
1:19:48 sing into it and it will model your
1:19:51 voice and then it will make songs likeo
1:19:55 but it will make them in your voice
1:19:57 Corey I know you do a lot of singing you
1:19:59 should sample your voice if you're in
1:20:01 here um with that one that one's
1:20:03 interesting then there's 3D worlds and
1:20:05 objects and then I've also got people to
1:20:06 follow in here all right so that's that
1:20:11 the other thing to play with in the
1:20:13 salon is we have guilds and so the
1:20:16 guilds are hosted by people who are
1:20:18 passionate about those sub subjects so
1:20:20 these are like sub communities within
1:20:22 the salon and they have meetings and
1:20:25 they've got places that you can you know
1:20:27 hang out and brainstorm and share ideas
1:20:30 and share new tools and re really really
1:20:33 remarkable and then the final thing is
1:20:36 this community the the AI learning lab
1:20:39 um the the people that show up here
1:20:40 regularly are called the Irregulars we
1:20:42 have our own little clubhouse so while
1:20:44 you're in here playing you can go to the
1:20:47 AI Irregulars club and just share images
1:20:50 that you're creating or songs that
1:20:51 you're making or gpts that you made or
1:20:53 just go in and chitty chat about stuff
1:20:56 so we have our own little clubhouse
1:20:58 where we can have our own inside jokes
1:20:59 and things like that all right so that's
1:21:02 the AI salon so if you're
1:21:06 new if you're new to AI actually if
1:21:08 you're not even if you're not new to AI
1:21:10 if you are hungry for
1:21:13 community of people that are not going
1:21:15 to judge you but actually celebrate you
1:21:17 when you talk about this AI [ __ ] that's
1:21:19 what this group is think of it like an
1:21:21 AI support group hi my name's Kyle I'm
1:21:25 into AI ha
1:21:30 Kyle it's been it's been three and a
1:21:33 half minutes since I made an image in do
1:21:35 yeah yeah we
1:21:39 know Corey Sandler last night or on the
1:21:43 on the call today Corey Sandler she
1:21:45 showed a bunch of remarkable work she
1:21:46 did she was like I was up till 5: in the
1:21:50 morning just running down Rabbit Hole
1:21:53 after Rabbit Hole making AI
1:21:58 art this is this group is a bunch of
1:22:03 enablers we won't help you crack the
1:22:05 Habit we'll make it
1:22:13 worse but yeah that's the AI Salon it's
1:22:16 really [ __ ] amazing it's just it and
1:22:19 it's not amazing because it's got the
1:22:21 best tutorials
1:22:24 it's amazing because it's got the best
1:22:27 people like human
1:22:30 beings it's the other thing use your
1:22:33 real name put your picture up there if
1:22:35 you're comfortable doing
1:22:37 that
1:22:40 why cuz we're all trying to [ __ ]
1:22:42 figure this out get to know the people
1:22:44 in the community one of the one of the
1:22:47 one of the really big reasons why I went
1:22:49 away from Discord why I didn't like
1:22:50 Discord there's a legacy in disc ORD
1:22:53 because it came from gaming there's a
1:22:55 legacy of
1:22:57 anonymity which the gaming in the nft
1:22:59 community is like oh you don't want to
1:23:01 dox
1:23:07 yourself I want to get to know the
1:23:09 people that are in my
1:23:13 community that's what this is
1:23:17 about you know people have different
1:23:19 levels of comfort with safety if you
1:23:21 don't feel safe that then fine but if if
1:23:23 you're just doing it cuz everyone else
1:23:25 is doing it then use your real
1:23:28 name let us let us get to know who you
1:23:35 are hi my name is Megan and I can't
1:23:38 concentrate on homework because mid
1:23:40 Journey has taken over
1:23:42 exactly Hi Megan
1:23:47 welcome yeah what this group does is
1:23:49 like yeah yeah that's that's a drag
1:23:51 about the homework anyway you want to
1:23:52 know something else cool you can do with
1:23:54 mid Journey stares at my username yeah
1:24:00 exactly Payton Moder PhD it was funny
1:24:03 when when I started getting into I got
1:24:05 into the blockchain and nft world and
1:24:08 you know jumped into Discord and things
1:24:10 much to the you know dismay of my
1:24:12 children who you're on Discord dad oh
1:24:17 god um you know I got into some of these
1:24:21 nft projects and and everyone's talking
1:24:23 about you can't dox yourself you can't
1:24:25 dox yourself I'm like I've been doxed
1:24:28 for like 30 years what the [ __ ] are you
1:24:29 talking
1:24:33 about and then at one point I got into
1:24:35 an nft project where where there was a
1:24:37 brief time where my nft was worth a lot
1:24:40 of money and I was like oh I kind of get
1:24:41 it now but you
1:24:43 know but but you know nft economics took
1:24:47 over and that thing isn't worth nearly
1:24:49 what it was anymore so so that's good
1:24:51 it's all good
1:24:57 oh AI this group has helped me identify
1:25:00 this lifelong
1:25:03 Affliction yeah
1:25:05 cuno yeah I know I know sorry about
1:25:11 that
1:25:14 oh what else we got what other the
1:25:17 questions you have what are the
1:25:21 thoughts I spent was trying to make a
1:25:24 meme on Splash from gp4 it wouldn't let
1:25:28 me make a man in the shower yeah the
1:25:32 safety guard rails kind of suck
1:25:34 sometimes you can probably you know some
1:25:37 of the open source stuff where you can
1:25:39 probably install stable diffusion
1:25:41 there's probably versions of stable
1:25:43 diffusion without the uh safety guard
1:25:50 rails gravity jump love this group it's
1:25:52 a great group isn't
1:25:56 [Music]
1:26:03 it we're enablers but in a good
1:26:05 [Laughter]
1:26:09 way it's well it is
1:26:13 true there's so much opportunity I mean
1:26:16 that's the thing that
1:26:17 I'm I'm most excited about
1:26:21 is
1:26:26 the only thing that matters right now is
1:26:29 that you're in the conversation that
1:26:31 you're playing with these tools and that
1:26:32 you're talking about it and just
1:26:35 establishing yourself as someone who is
1:26:38 in the
1:26:39 game
1:26:42 because the
1:26:44 industry industry of generative AI
1:26:48 hasn't been born yet like it really
1:26:51 hasn't the McKenzie report from I think
1:26:55 it was from
1:26:57 November um was a follow-up report to to
1:27:00 a study they did at the beginning of the
1:27:02 year at the beginning of the year 87% of
1:27:05 people had were not using chat GPT I
1:27:08 assume this is in America 87 people 87%
1:27:12 of people were not using chat
1:27:14 GPT at the end of 2023 6 months later
1:27:20 82% were not using chat
1:27:23 GPT that's not a big
1:27:26 gain that means that four out of five
1:27:30 people probably don't even know what
1:27:31 chat GPT
1:27:34 is four out of
1:27:36 five and of the 18% that said they've
1:27:40 used chat
1:27:44 GPT I'm guessing about half of them this
1:27:48 is a guess this is
1:27:50 speculation probably used it once they
1:27:53 went there and they went make me
1:27:54 something write a song that's cute and
1:27:57 never used it again so now we're down
1:27:59 into single digits and then how many
1:28:02 let's call that 9% let's call it 10%
1:28:04 let's say 10% of people use cat GPT
1:28:07 decently regularly how many of them are
1:28:11 paying the 20 bucks a month probably
1:28:14 half of them and maybe even less and
1:28:16 then how many pay the 20 bucks a month
1:28:18 and use it a
1:28:20 lot it's probably like 3 to
1:28:25 5% so that means 95% of the
1:28:30 people right
1:28:32 now in
1:28:35 America don't have a [ __ ] clue what
1:28:37 this thing is this this industry hasn't
1:28:40 even started
1:28:42 yet so all you have to do right now is
1:28:45 be in the
1:28:46 game make kids book make pizza recipes
1:28:50 make it doesn't [ __ ] matter get
1:28:52 really good at image generation start a
1:28:55 start a video career Sherry D's making
1:28:58 music videos for
1:29:00 bands she's got a career in
1:29:03 nursing she's now a video an AI
1:29:06 videographer where is that going to lead
1:29:08 her I don't know it doesn't matter all
1:29:09 of this doesn't matter like just be in
1:29:12 the
1:29:13 game and opportunities are going to
1:29:17 Blossom in front of
1:29:19 you and by the way being in the game
1:29:22 doesn't mean being a secret cyborg if
1:29:25 you're one of those people who uses chat
1:29:27 at work and you don't want to tell
1:29:28 anyone about it cuz you're now doing in
1:29:30 5 hours what used to take you 30 hours
1:29:34 stop that start talking about it it's
1:29:37 way more valuable to you to let people
1:29:40 know you're using it yeah but they don't
1:29:43 they don't allow it I don't care tell
1:29:45 people at lunch you know what I do on
1:29:47 the weekends I do this with Chach tell
1:29:49 them here's what I'm doing with it
1:29:51 here's why it's cool here's what I like
1:29:53 about it but Kyle they're going to make
1:29:55 fun of me yep they
1:29:58 are
1:30:00 Sally could you stop with the chat
1:30:05 JP good Lord all you talk about is ai ai
1:30:08 this and AI that would you join some
1:30:11 kind of
1:30:15 cult that annoying
1:30:18 person 6 months from
1:30:20 now is going to t you on the shoulder
1:30:23 and go hey
1:30:25 thally I know you were kind of into that
1:30:28 AI stuff do you still do
1:30:30 that could you teach me cuz I I think I
1:30:34 need to learn it
1:30:36 now
1:30:38 95% somewhere some massive majority has
1:30:43 no clue what's going
1:30:44 on and this [ __ ] ain't going
1:30:48 away did you see the thing yesterday
1:30:50 what Sam alman's trying to rap
1:30:54 MoneyWise he's raised 13 billion which
1:30:58 is a lot for a
1:31:01 startup rumor was he was over in Saudi
1:31:05 Arabia or wherever Middle East trying to
1:31:09 raise a hundred billion dollar that was
1:31:11 right around the time where he was fired
1:31:14 for not disclosing something well it
1:31:16 came out yesterday or the day before
1:31:18 he's actually trying to raise between
1:31:20 five and seven trillion
1:31:25 dollar to create the world's largest
1:31:30 chip manufacturing company to be able to
1:31:33 do the Computing to do
1:31:37 AGI 5
1:31:39 to7 trillion dollar that's more than the
1:31:43 market cap of Microsoft and Apple
1:31:46 combined not their revenue their market
1:31:51 cap
1:31:53 this [ __ ] ain't going
1:31:57 away this was like actually in a
1:31:59 business publication right this was this
1:32:02 this wasn't like in psychosis Journal
1:32:05 psychosis weekly that he's just
1:32:07 delusional that he thinks he can raise
1:32:08 this much money There's an actual
1:32:14 plan it's bananas how much people are
1:32:17 ignoring what's going on
1:32:20 aruga exactly it's it is remarkable but
1:32:25 it like we've seen it before so in the
1:32:28 early days of the worldwide web I
1:32:29 started one of the first digital
1:32:30 agencies in
1:32:33 1994 and everyone thought I was a
1:32:35 [ __ ] weirdo kle why you talk about
1:32:37 worldwide we all the time this is boring
1:32:40 I'd be at parties oh you want to know
1:32:42 else you can do on the web you can stick
1:32:44 pictures on a page and then you can
1:32:45 click on a picture and it takes it in
1:32:48 there I had a bunch of friends that were
1:32:50 act actors and writers in artists they
1:32:53 were like
1:32:57 huh and then you know what's amazing so
1:33:00 I did this I did agency.com got you know
1:33:02 wildly successful like crazily
1:33:04 successful cuz I was in the
1:33:08 conversation and then all those all
1:33:10 those people that made fun of me after
1:33:12 after all that
1:33:13 success said to
1:33:15 me you know Kyle in in various times I
1:33:19 should have listened to you I should
1:33:21 have listened to you that was you're
1:33:23 really on to something there I was like
1:33:24 yeah
1:33:26 thanks those same people today when I
1:33:29 talk about AI they're like why are you
1:33:30 talking about AI it's like
1:33:33 really really you
1:33:35 [ __ ]
1:33:37 really Kyle I just had a GPT Define oh
1:33:40 crap that just went
1:33:42 away I didn't see what that
1:33:49 said that's a lot of money yeah yeah
1:33:53 laughing Thunder that's it seven
1:33:55 trillion it's a seven with 12 zeros
1:33:57 behind it it's a lot I just had chat GPT
1:33:59 Define mathem magical I'll post it in
1:34:02 the AI Salon yeah please do that that's
1:34:04 great and I think Pate you're in here
1:34:06 now so Detroit Jay came up with the idea
1:34:08 I was talking about um your and my
1:34:11 debate where I I think it's magical and
1:34:13 you're like no it's math it's not magic
1:34:16 um and and Detroit Jay said you know
1:34:19 it's
1:34:20 mathemagical so I think I think there is
1:34:23 some common ground for us for us to we
1:34:27 can we can walk on a bridge
1:34:28 [Laughter]
1:34:34 together oh man that's really
1:34:42 funny give them your Amazon wish list I
1:34:45 keep feeling like I'm late so I
1:34:46 appreciate the perspective yeah it's
1:34:48 it's funny it's funny tired Runner there
1:34:51 are absolutely people who've got a head
1:34:54 start on you but here's why it's not
1:34:56 such an
1:34:57 advantage
1:35:00 um if you got super masterful at Mid
1:35:05 Journey prompting right or or at a
1:35:07 particular kind of
1:35:10 prompting and then Del comes
1:35:13 out use creating images in the context
1:35:17 of a chat GPT chat is such a different
1:35:20 skill than Journey
1:35:23 prompting
1:35:25 that even if someone was way ahead of
1:35:27 you in mid journey and you sort of show
1:35:29 up when Do's there and you get good at
1:35:31 that it's just a different skill and and
1:35:33 it's like these new these new ways of
1:35:36 working and these new technologies just
1:35:38 keep coming one after another after
1:35:41 another after another
1:35:43 so my guess is it's not going to be
1:35:47 three it's going to be at least 3 years
1:35:49 and probably closer to a 3
1:35:52 to5 where any of this stuff stabilizes
1:35:55 in any kind of way it took it took with
1:35:58 the worldwide
1:36:00 web [ __ ] was being developed like I got
1:36:04 into it in 94 I think Tim burner Lee I
1:36:07 think he wrote the paper in ' 89 and and
1:36:11 the first something was 91 I got into it
1:36:16 '94 it probably wasn't
1:36:20 until 97 998 until there was any kind of
1:36:24 stability there that there was any kind
1:36:26 of infrastructure and
1:36:29 stability and then it probably took
1:36:32 until like 2002 2005 where where it
1:36:36 really started to kind of codify right
1:36:39 so it was it was close to a decade now
1:36:42 ai is moving faster so I think I but but
1:36:47 it's the the increases are exponential
1:36:50 and and they're going to be way more
1:36:51 disruptive so I think I think AI is
1:36:53 going to [ __ ] a bunch of stuff up first
1:36:56 so what it means is for the next 3 years
1:36:59 if you just get on the train you're fine
1:37:02 you're not that far
1:37:03 behind related one thing I worry about
1:37:06 with this group is GPT being the final
1:37:10 stage it feels
1:37:14 stagnant
1:37:16 um I don't quite know what you mean that
1:37:18 chat GPT is the final stage listen I
1:37:21 want for nothing more than the fully
1:37:25 multimodal Gemini Ultra to be here Pate
1:37:28 like
1:37:29 I one of my biggest disappointments
1:37:32 about Gemini Ultra is not that it's got
1:37:34 it makes shitty images I don't really
1:37:35 care about that they they'll get better
1:37:38 um the the thing that I'm really bummed
1:37:40 about that we didn't see in this launch
1:37:42 was the the the visual interface the the
1:37:45 infinite canvas
1:37:47 interface because it's a completely
1:37:49 different way of thinking about how you
1:37:51 interact with a large language model
1:37:53 having a truly
1:37:54 multimodal interface and a truly
1:37:57 multimodal um large language model was
1:38:01 something I was really looking forward
1:38:04 to because it requires a completely
1:38:05 different kind of
1:38:07 thinking and then and we're not even
1:38:10 talking about getting to you know large
1:38:13 language models that reason and have
1:38:14 autonomous agents like that's coming
1:38:16 this year too but I think there's
1:38:18 completely different ways of interacting
1:38:20 with these tools that we're no one's
1:38:22 thinking about right now because the
1:38:23 interface is all like a Google search
1:38:25 box which I think is such a [ __ ]
1:38:28 limiting thing like the fact that it's
1:38:30 the fact
1:38:32 that that it's still language is the
1:38:34 primary thing like even chat
1:38:37 GPT they've got this thing called GPT
1:38:41 Vision this group is all about
1:38:43 exploration and treat the ultimate goal
1:38:45 is making a GPT no I I would say I don't
1:38:48 think that's right Pate I I think
1:38:49 there's even been a shift in that in the
1:38:51 past month
1:38:53 the the reason people are excited about
1:38:55 making gpts is you've got people who are
1:38:59 not Engineers who are not coders who for
1:39:02 the first time in their lives have
1:39:04 access to a tool that lets them create
1:39:06 sophisticated applications without
1:39:08 having to
1:39:09 program that's like that's staggeringly
1:39:13 powerful so I would give people a little
1:39:15 bit of Grace that they're exploring
1:39:18 that but I've even seen a shift in that
1:39:20 I think there's a recognition amongst
1:39:22 this group that the gpts are limited and
1:39:24 the GPT store is a giant piece of
1:39:30 [ __ ] oh that's a great question
1:39:33 lunatick so so Pate what impresses you
1:39:36 most about Gemini that we haven't
1:39:38 noticed that's a great question and not
1:39:41 not whatever you've seen on the on the
1:39:43 inside but what about what was just
1:39:45 released are you most impressed with I
1:39:48 think that's a great
1:39:50 question
1:39:57 I wouldn't worry about this group
1:39:59 stagnating I ain't going to let it
1:40:04 stagnate I'm still a participant I'm G
1:40:07 I'm goingon to push y'all
1:40:09 [Laughter]
1:40:15 [ __ ] I will not stop building
1:40:18 with
1:40:19 AI until it can do the things I don't
1:40:22 like or don't know how to do
1:40:27 yep open AI holds off until Google
1:40:30 Microsoft or meta Does It Better Pate
1:40:33 what gets the
1:40:34 gpt's Gemini options we would love to
1:40:39 learn hey what did you do Saturday
1:40:42 night Jimmy what' you do Saturday night
1:40:45 I don't want to talk about it what did
1:40:47 you do I watched some old guy on Tik Tok
1:40:50 for three hours
1:40:53 hours hey Siri okay go
1:40:57 away
1:41:01 okay I think it's a very good model
1:41:03 overall and the text seems to flow
1:41:05 better okay cool you you know what I
1:41:07 think's a big deal pate and I I think
1:41:09 this is I think it's it's valid for
1:41:12 Gemini Pro or yeah Gemini Pro as well as
1:41:14 Gemini Ultra I think where
1:41:18 um where Gemini has a competitive
1:41:22 advantage that would be worth using it
1:41:24 for if they improve it cuz it's not
1:41:26 quite there but it's close is the
1:41:28 ability to search
1:41:30 YouTube if if I could have an interface
1:41:33 that would let me say hey here's this
1:41:35 idea I've got not only go find me videos
1:41:38 on YouTube that do this but jump me
1:41:41 right to the specific
1:41:44 section that that answers exactly you
1:41:47 know what I'm looking for or talks about
1:41:50 exactly what I'm talking about
1:41:51 and puts together an interface of like
1:41:53 here's a bunch of clips from
1:41:56 YouTube um that are the specific things
1:41:59 that go with this text I think Google
1:42:02 that would be a reason to use Gemini
1:42:04 like if take all the other [ __ ] away
1:42:06 like like let's assume that Gemini
1:42:10 Ultra is is going to be at parody with
1:42:13 chat GPT in terms of its writing
1:42:18 ability if it were if it were more
1:42:21 tightly integrated with YouTube which
1:42:24 you [ __ ] own you have all the
1:42:27 data and you created you know a dynamic
1:42:30 interface that it did some of that
1:42:31 Dynamic interface stuff that you demoed
1:42:34 on on uh Gemini launch
1:42:36 day that would be a reason to use it and
1:42:39 I I and and I think it's not far right
1:42:41 there sorry if I offended people about
1:42:42 stagnant comment I just want to
1:42:44 encourage people to be I I don't think
1:42:46 no you didn't offend me didn't offend
1:42:47 anyone I'm just spicy I was spicy from
1:42:50 last night
1:42:55 I listen that's that's the other thing
1:42:57 Pate I love having your voice in this
1:42:59 community I like I think it's important
1:43:01 I think it's important to have diverse
1:43:03 points of views back to the inclusion
1:43:04 thing I think it's important to have
1:43:06 diverse points of
1:43:08 you right Eric marier when I spoke on
1:43:11 that that panel that I streamed live
1:43:14 that bet me $1,000 that a small agency
1:43:16 wasn't going to take over a big agency
1:43:18 in the next three
1:43:20 years
1:43:21 like that's an important voice to have
1:43:23 in the conversation I feel like we're
1:43:25 holding pain accountable for Gemini he's
1:43:27 not not working as
1:43:31 expected actually PID is the reason that
1:43:34 Gemini works as good as it works right
1:43:36 because the the tpus are so efficient
1:43:38 pate's working on the TPU part of that
1:43:40 he's not working on the Gemini interface
1:43:41 or the go to market strategy I'll tell
1:43:44 you what if Pate were in Google
1:43:45 marketing and dealing with the go to
1:43:47 market strategy I'd be holding him more
1:43:50 accountable but but no no P P's pate's
1:43:53 doing the the good
1:43:55 work he's making that [ __ ] work um good
1:43:59 to always question everything absolutely
1:44:02 and here's here's the other deal about
1:44:03 questioning things like I don't mind
1:44:05 people questioning things I just want
1:44:07 people questioning things that are in
1:44:09 the game in the conversation it's the
1:44:12 people that are sitting on the sidelines
1:44:13 just going it's shitty it's shallow like
1:44:16 the guy last night it's shallow you
1:44:19 didn't have any specific feedback other
1:44:20 than I use it a lot and it's not
1:44:23 awesome that doesn't feel constructive
1:44:26 to
1:44:27 me that just
1:44:30 feels like
1:44:32 Fear YouTube would be awesome and
1:44:34 actually right now so uh Julie in
1:44:37 California you can if you go to Gemini
1:44:39 right now go to gemini.com and and tell
1:44:43 it to find you YouTube videos on a
1:44:45 certain subject it will and it comes
1:44:47 back and it gives you thumbnails like
1:44:48 they're doing some stuff decent with it
1:44:51 it's just that they've got all the data
1:44:53 they've got all the transcriptions so
1:44:56 they know that within you know these six
1:44:59 videos here are the sections within
1:45:02 those videos where they're playing so
1:45:03 that you could you could jump to those
1:45:05 specific parts of the videos there might
1:45:07 be something about usage rules or I
1:45:09 don't know there might be a reason they
1:45:10 can't do that legally but from an end
1:45:13 user standpoint that would be worth
1:45:14 paying
1:45:15 for also Gmail docs and slides yep
1:45:19 exactly Brian Whitney Kyle
1:45:22 yes did you have a question
1:45:24 Brian or you just stating my
1:45:27 name O nice 20 minute long videos are
1:45:31 the bane of question answering I exactly
1:45:34 right wouldn't that be cool if if you
1:45:36 could say I mean think about so I put
1:45:41 these videos on
1:45:44 YouTube I've got nightly two to three
1:45:47 hour videos on
1:45:49 YouTube wouldn't it be amazing to be
1:45:52 able to just say hey for the AI learning
1:45:55 lab
1:45:56 Channel where did Kyle talk about the
1:45:59 rabbit R1 and it just jumped you to a
1:46:02 series of Clips where I talked about
1:46:03 that
1:46:04 stuff that would be [ __ ] staggeringly
1:46:07 good having some Blandon oh yeah Walkin
1:46:10 blue that's solid I like it there's my
1:46:14 Blandon right
1:46:16 there
1:46:19 beautiful
1:46:21 oh what you know what I was going to I
1:46:23 was going to show something that I
1:46:24 thought was really interesting kind of
1:46:31 is let me go where do I want to
1:46:49 go
1:47:01 is this it
1:47:10 no 2.1 million followers this has got to
1:47:13 be oh that's
1:47:16 Facebook I don't know if you've been
1:47:17 watching Cheryl
1:47:19 Crow
1:47:25 this is not her
1:47:27 either dang it where's her actual
1:47:32 account
1:47:36 cherl
1:47:39 here
1:47:42 okay
1:47:46 so so Universal Music Group
1:47:52 [ __ ] up their negotiation with Tik Tock
1:47:55 or gets greedy whatever whatever the
1:47:56 reason is they pull all of their artists
1:47:59 off Tik
1:48:01 Tok and so what Cheryl Crow started
1:48:03 doing is getting on here and and
1:48:07 basically she's doing a Taylor Swift
1:48:09 she's re-recording her songs on Tik Tok
1:48:12 and having people um duet them and
1:48:15 things like
1:48:16 [Music]
1:48:19 that I woke up and call this
1:48:26 morning I just I think it's brilliant
1:48:28 and I think it's
1:48:31 um I don't know there there's something
1:48:35 about there's something about artists
1:48:38 finding a way to squeeze through the
1:48:40 cracks of the
1:48:42 machine that I think is going to happen
1:48:44 with AI
1:48:45 too that that authenticity and and
1:48:49 people
1:48:52 like authentically
1:48:54 self-expressing are going to find ways
1:48:56 to squeeze through the cracks of a lot
1:48:58 of the fears about
1:49:01 AI you know sort of
1:49:04 sanitizing creative and just making at
1:49:07 pum that you know all this stuff is is
1:49:09 AI created there's something about about
1:49:13 what she's doing and I I assume there's
1:49:15 other artists that are doing it but like
1:49:17 she's a big one and she basically said
1:49:19 since Universal my my music off here why
1:49:22 don't I do this I'll sing you some songs
1:49:25 want to duet them and like people are
1:49:27 now dueting them with her she's going to
1:49:29 end up finding someone to you know
1:49:30 co-create something with they're going
1:49:32 to record a
1:49:35 record and it there's there's this whole
1:49:37 new Lane that that's coming up so that
1:49:40 like the power
1:49:43 of of
1:49:46 individuals to squeeze through the
1:49:48 cracks of the
1:49:49 machine feels like it's a new thing and
1:49:52 I think AI is going to help amplify that
1:49:54 and give individuals
1:49:56 tools to be able to do things you know
1:49:59 Way
1:50:00 Beyond what they what they could have
1:50:03 historically just because of the nature
1:50:05 of how you had to get into the system so
1:50:08 anyway I just wanted to share that one
1:50:09 because I think what she's up to is
1:50:10 really cool can't play it cuz you know
1:50:12 I'll get Shadow banned and all that
1:50:19 [ __ ]
1:50:23 Kyle Shannon I I have a question just
1:50:26 having trouble typing throw pay to Bone
1:50:30 Kyle let's play with
1:50:31 Gemini uh
1:50:35 maybe
1:50:38 uh let let me see what other questions
1:50:40 we have I might do that it's and we
1:50:43 don't we don't have to take care of Pate
1:50:45 P P's a big boy like if if it makes
1:50:48 sense to use Gemini a will
1:50:56 I mean I think it's a it's a viable
1:50:58 alternative too good for Cheryl Crow she
1:51:01 gets it love her and that's so smart of
1:51:04 her yep I agree Tik Tok is a money
1:51:06 machine with online marketing the mega
1:51:09 corpse can only push the citizen so far
1:51:11 before we push
1:51:13 back
1:51:15 yep I want I want to create a logo have
1:51:18 any AI suggestions um yeah mid Journey
1:51:22 or do I mean you can just do it right
1:51:24 within
1:51:26 do
1:51:28 um what I would do is just go to chat
1:51:30 GPT chat GPT 4 so you got to have chat
1:51:33 GPT
1:51:34 plus talk about your
1:51:37 business um have it write a creative
1:51:40 brief for a logo and then make sure that
1:51:43 you agree with the creative brief and
1:51:45 then just have it give you a bunch of
1:51:46 logo Concepts now here's the deal about
1:51:49 the logos any logos that are created the
1:51:52 spelling is probably going to be a
1:51:54 disaster you'll probably get one or two
1:51:57 in there that are
1:51:58 close but they'll be like a little
1:52:00 asymmetrical they'll be not great so
1:52:03 what I would strongly suggest is you're
1:52:06 using any of these AI tools for
1:52:08 concepting and then take that concept
1:52:11 into illustrator and redraw it Trace
1:52:13 over it or give it to an artist who can
1:52:14 do
1:52:15 that um because these things aren't
1:52:18 making Vector images you can you can use
1:52:20 use bit map to Vector converters to
1:52:23 convert them but it's probably just as
1:52:26 easy to redraw stuff unless you're doing
1:52:28 super sophisticated um logos which if
1:52:31 you're doing a logo you want to be able
1:52:32 to simplify it
1:52:34 so
1:52:37 uh wait I've been having better
1:52:41 experiences with Gemini Advanced writing
1:52:44 over GPT but I start in GPT and move it
1:52:48 to Gemini oh that's really interesting
1:52:51 soggy wa waffle
1:52:54 ASMR that's really
1:52:57 interesting yeah I haven't done that yet
1:52:59 I haven't played with the writing yet
1:53:00 but I I will it's it's something the
1:53:03 problem with with doing
1:53:06 like doing writing comparisons on a
1:53:08 channel like this it's not very Visual
1:53:11 and and it's going to it it could take a
1:53:13 while to sort of figure out is it better
1:53:15 is it not because it's very subjective
1:53:17 so I'll do some work I'll do some work
1:53:19 offline
1:53:20 and and do some comparisons and then if
1:53:23 if I have some good examples that I can
1:53:25 walk through that that demonstrate some
1:53:27 things I I'll come back and do that
1:53:29 because I I think it's a good exercise I
1:53:31 think the exercise of comparing
1:53:32 different models with the same kind of
1:53:35 projects is actually a really important
1:53:37 thing is your YouTube listing automated
1:53:40 now it is not
1:53:42 Winston it is not in fact Brandon in
1:53:45 here is helping
1:53:46 me take the transcripts and write
1:53:49 descriptions for so I did put up all of
1:53:52 the all of the back Tik Tok videos to
1:53:55 the to the um YouTube channel here so if
1:54:00 you go to YouTube
1:54:03 slash
1:54:05 um AI learninglab dtt That's the YouTube
1:54:10 channel for this I have all of the
1:54:12 videos up including last night's
1:54:16 rant rant session um are all up on the
1:54:19 YouTube channel right now but none of
1:54:21 them have descriptions in them so
1:54:22 Brandon's helping me put that together
1:54:24 so thank you for that Brandon I
1:54:26 appreciate it API implementation and
1:54:28 actions in general being simpler on gpts
1:54:32 would be nice I agree with that thank
1:54:33 you Sher
1:54:37 bear
1:54:39 um that's the the other thing about
1:54:42 Gemini that I'd like to see is you know
1:54:44 is there an equivalent to code
1:54:46 interpreter um you know are there are
1:54:49 there is the is there an equivalent to
1:54:52 be able to upload PDFs or csvs or things
1:54:55 like that cuz right now it's not there
1:54:59 um so that that again there there's some
1:55:01 things that feel like they're enough
1:55:03 missing that it it doesn't feel worth
1:55:05 spending the 20 bucks to me right now
1:55:07 the timeouts on GPT are are infuriating
1:55:10 I totally agree with that Gemini is very
1:55:12 nice I find it very responsive that's
1:55:16 great that's great yeah the timeouts on
1:55:18 chat GPT are absolutely absolutely
1:55:20 [ __ ]
1:55:21 maddening life's tough it's tougher when
1:55:24 you're
1:55:26 stupid super zappier to the
1:55:29 rescue I was reviewing the rant part of
1:55:32 it I took the rant from last night and
1:55:34 put it up as a Tik Tok video so if you
1:55:36 go to my channel I just have a video
1:55:38 there called the rant where I just sort
1:55:41 of went off on
1:55:43 cynicism it really it really got to me
1:55:46 last
1:55:47 night hopefully in a in a in a positive
1:55:53 way but it did it bugged the [ __ ] out of
1:55:55 me I don't think I've been I don't think
1:55:58 I've been
1:56:00 that
1:56:04 passionate about something in a
1:56:10 while you guys are up to 31,000 likes
1:56:14 you getting don't give yourselves carpal
1:56:15 tunnel on a Saturday
1:56:17 night how'd you injure your wrist I
1:56:20 don't want to talk about
1:56:29 it oh I'd love to see you use the new
1:56:32 word on Tik Tok on behalf of the AI
1:56:34 Salon to draw new people in what's the
1:56:37 new word I'd love to see you use the new
1:56:40 word on Tik Tok
1:56:43 mathematical that new
1:56:46 word from GPT translate to Spanish then
1:56:50 Hindi then back to English to GP to fix
1:56:52 grammar beats AI detection there's a lot
1:56:55 of things that beat AI detection best
1:56:58 app for soft best app or software for
1:57:00 creating videos from my AI images stable
1:57:04 diffusion yeah stable video diffusion's
1:57:07 good
1:57:10 um all all of okay all of the video
1:57:14 tools right
1:57:16 now are
1:57:18 janky
1:57:20 so I would play you got pabs which is
1:57:24 free right now you've got Runway ml
1:57:26 which is not free but both of those are
1:57:28 really good you got Moon Valley which
1:57:31 you have to do you have to do that on
1:57:33 Discord and I think Moon Valley is free
1:57:37 I should probably add Moon Valley to the
1:57:40 list and then you have stable video
1:57:42 diffusion so I'd say those four you
1:57:45 should play with and then you've also
1:57:47 got in Leonardo you've got image to
1:57:52 video within Leonardo that's using
1:57:54 stable video diffusion right Leonardo AI
1:57:57 is is a a web front end to stable
1:58:00 diffusion
1:58:01 technology or stability AI technology so
1:58:05 so there's five options you could play
1:58:07 with can you talk about text to
1:58:10 3D I think it's a building block of
1:58:13 spatial am I
1:58:15 wrong it's totally so John Buckley yeah
1:58:19 it's it's to totally um a building block
1:58:22 to spatial so let me I'll I'll show you
1:58:26 I I'll show you a thing I'll actually
1:58:29 show you two and then we can talk
1:58:33 about we're just we're we're early for
1:58:36 that stuff but we're we're not far so
1:58:38 hang
1:58:41 on so if you go to the AI
1:58:45 Salon it's AIS salon. mn.
1:58:52 and then in the community section go to
1:58:54 tools and
1:58:56 resources
1:58:58 right do you have the personalization
1:59:01 option on GPT Kyle do you have the
1:59:04 personalization
1:59:06 option oh I don't think
1:59:14 so hang on chat add how can I help
1:59:18 you
1:59:24 I don't think
1:59:27 so
1:59:29 um okay AI Salon tools and
1:59:33 resources and then scroll
1:59:35 down to
1:59:38 3D worlds and objects so first place
1:59:42 we'll go is
1:59:44 Skybox blockade
1:59:46 Labs so this is
1:59:51 kind of a world bu Builder where here's
1:59:55 your default world that you can spin
1:59:57 around in right look up look down look
1:59:59 left look right but you can dream up
2:00:02 your own world so
2:00:06 say
2:00:09 um
2:00:10 dystopian
2:00:13 desert landscape littered with
2:00:18 smoking
2:00:20 oil
2:00:21 wells and broken down machinery and
2:00:28 cars and then I'll select the style of
2:00:32 let's see
2:00:36 surreal
2:00:40 no yeah we'll do surreal why not
2:00:45 generate and so this is text to world
2:00:51 what it's actually doing is is quite
2:00:54 simple it's it's creating a flat image
2:00:57 that's formatted in a very particular
2:00:59 way that makes it easy to map it onto a
2:01:02 sphere so so from a technical standpoint
2:01:05 this isn't truly creating a 3D world but
2:01:09 it's cleverly taking a 2D image and and
2:01:12 formatting that in a way that you know
2:01:16 gives
2:01:17 you the illusion of of
2:01:20 of 3D so so this is an image mapped to a
2:01:24 3D sphere and what some people are doing
2:01:27 with these objects is they're taking
2:01:28 them into Unreal Engine and then
2:01:30 creating objects in Unreal Engine to
2:01:32 making these interactive environments so
2:01:35 so this is a place that you can play
2:01:37 there's also in that list there's this
2:01:39 thing called Luma
2:01:42 labs. and this is text to 3D
2:01:45 object so let me log in
2:01:48 here
2:01:50 and I haven't used this in months so
2:01:53 this could be better worse or even
2:01:57 dead let's see best practices Community
2:02:01 create drop a file in this area click to
2:02:07 select
2:02:09 um I don't know what I'm looking for
2:02:15 here I thought this was text to 3D
2:02:17 object what what am I actually doing
2:02:32 here oh I guess this is this is like
2:02:36 um like gaui and splatting like 3D 3D
2:02:40 scene
2:02:41 creation there's also I don't know what
2:02:43 the tools are there's there's there's
2:02:45 tools out there now doing text to 3D
2:02:47 object you've also got Unreal Engine
2:02:50 that's doing
2:02:52 um procedural building where you can
2:02:56 just say I want to drop in a whole Cliff
2:02:58 scene here and it's incorporating that
2:03:00 into an actual 3D interactive scene um
2:03:06 so yeah to create immersive environments
2:03:09 there's going to have to be a bunch of
2:03:10 technologies that come together
2:03:12 including like gameplay logic 3D worlds
2:03:17 3D objects texture mapping
2:03:20 you know procedural World
2:03:22 building all of the individual piece
2:03:25 parts are there in some form or the
2:03:27 other but what we don't have right now
2:03:30 is the 3D World version of chat GPT
2:03:34 right like chat GPT does words and
2:03:36 pictures really good it's not doing
2:03:38 video audio or 3D World building really
2:03:41 good
2:03:43 um my BET right now is on um
2:03:48 unreal I think unreal is going to going
2:03:51 to going to drop a bomb in
2:03:54 2024 that's going to be text to 3D world
2:03:58 that you can
2:04:00 actually participate in in some way it
2:04:03 might not have game mechanics but you'll
2:04:05 be able to cruise around in a world that
2:04:07 you
2:04:08 create um but we'll
2:04:11 see chat GPT can now be instructed to
2:04:15 remember specific user preferences ac
2:04:17 across multiple interactions yeah saw
2:04:19 that that was coming Joker I don't have
2:04:21 it yet
2:04:23 um but yeah we're probably going to have
2:04:25 that soon and that's again that's a
2:04:28 thing that all of a sudden changes like
2:04:31 if you if you're really good at chat
2:04:33 chpt and you know how to use chat GPT
2:04:36 even the current thing and all of a
2:04:39 sudden it gets memory across
2:04:42 conversations it's going to completely
2:04:44 change how we we all use chat
2:04:47 GPT multiple interact so Source Camp
2:04:50 what what it's going to be able to do is
2:04:53 if you have a conversation over here
2:04:56 about
2:04:58 um doing some branding stuff and you
2:05:01 have a conversation here about doing
2:05:03 some business strategy and you have a
2:05:05 conversation over here about doing some
2:05:08 tactical Social Media stuff it can it
2:05:10 can connect the dots across those
2:05:14 conversations now how you connect like
2:05:17 how you have it remembered
2:05:19 and not remember like there's there's
2:05:21 some conversations I have in chat gbt
2:05:23 like if I'm just [ __ ] around here on
2:05:25 the channel I don't necessarily want
2:05:28 that conversation to go into my larger
2:05:30 memory context so I don't know how
2:05:33 you're going to opt in and opt out of
2:05:34 that so I don't if anyone's using it yet
2:05:37 I I'd be curious to know how they're
2:05:38 doing that but that's a big [ __ ]
2:05:41 deal right Pi does that a little bit
2:05:44 right now where Pi remembers who I am it
2:05:46 remembers what my name is it remembers
2:05:49 some of the conversations we've had
2:05:50 before I think that's going to get
2:05:52 better and better I think Apple's going
2:05:54 to be really good at this so chat GPT is
2:05:57 adding this idea of context you know
2:05:59 memory across
2:06:02 conversations but what does that
2:06:05 mean right now we all think and this is
2:06:08 to P's point about right we're we're all
2:06:10 thinking about discret gpts and and
2:06:14 discreet
2:06:15 conversations and now we're going to
2:06:17 come the ability to connect them and
2:06:20 have them understand each other what
2:06:23 does that mean I don't [ __ ]
2:06:25 know the fact that you can now call gpts
2:06:29 inside a
2:06:30 conversation completely made me rethink
2:06:33 what gpts even are right so so all of
2:06:36 these little Nuance things that they're
2:06:38 adding are going to change what's
2:06:41 possible in and in ways that I don't
2:06:43 think any of us understand yet
2:06:45 so it's it's cuckoo people it's Coco Pi
2:06:50 tends to forget past conversations over
2:06:52 time so I hope that does get better yeah
2:06:54 it it does and again I think that has to
2:06:56 do with context window like the size of
2:06:58 the context window so if they can take
2:07:02 Knowledge from this conversation and
2:07:04 embed that and drop it into a database
2:07:08 basically and then start to connect
2:07:10 these conversations together that that
2:07:13 would be pretty
2:07:14 powerful big [ __ ] deal would be GPT
2:07:17 allowing me to make folders for
2:07:19 conversations yeah that would be great
2:07:21 that would be a way to manage
2:07:23 it I'm trying to think about how my gpts
2:07:26 are connected yep I think we're going to
2:07:28 have to and again is it a manual thing
2:07:31 is it an automatic thing I I would think
2:07:34 it would have to be manual I think
2:07:36 folders would be an interesting
2:07:39 option it would also require
2:07:42 them to have a uh search function for my
2:07:47 past conversations
2:07:52 I I still can't believe that chat GPT
2:07:54 doesn't have a [ __ ] search function
2:07:56 on their website they've got it on their
2:07:58 their IOS app you can search for past
2:08:01 conversations on iOS you can't on the
2:08:04 website what the
2:08:14 [ __ ] I really really want to call
2:08:17 multiple a in GPT that'll come I think
2:08:21 that's going to come sooner than
2:08:23 later Pi needs to keep a dossier on us
2:08:26 and our
2:08:30 issues yeah I know Pi has a
2:08:33 2,000 token context window that's
2:08:37 pathetic I don't understand what don't
2:08:39 you understand Arabian of Israel oh
2:08:41 about the about the conversations okay
2:08:45 so right now when you're having a
2:08:46 conversation in chat GPT
2:08:50 let's say at the beginning of the
2:08:51 conversation I say um we're going to
2:08:54 talk
2:08:55 about
2:08:59 um um we're going to talk
2:09:04 about designing uh a
2:09:09 new
2:09:11 um oh we're going to talk about
2:09:13 designing new recipes for a vegan
2:09:17 lifestyle that incorporates
2:09:20 beef whatever right and and you start
2:09:23 out the conversation up there and and
2:09:26 then you're talking you're talking about
2:09:29 you know vegan lifestyle and vegan
2:09:31 lifestyle and and at some point down
2:09:34 here you're not mentioning beef anymore
2:09:35 you're just talking about vegan stuff
2:09:38 and and just like your phone as you're
2:09:40 scrolling text up your phone there's a
2:09:42 point at which it goes off screen and
2:09:45 then you can't see it anymore that's
2:09:46 kind of what happens here where if you
2:09:48 talk about beef early on in the vegan
2:09:50 conversation and then you're talking
2:09:52 down here and that goes up and that goes
2:09:53 out of the context window all of a
2:09:55 sudden chat GPT if you reference beef
2:09:59 will not understand why you reference
2:10:01 like it it won't know that you said that
2:10:03 before so in a given conversation chat
2:10:06 GPT has memory of the conversation so
2:10:09 long as it's in the context window which
2:10:13 right now is about 6,000 Words which
2:10:15 seems like a lot but it's not and then
2:10:17 some of the more advanced vers of gp4 go
2:10:20 go up to
2:10:21 128,000 tokens which is like I don't
2:10:25 know 2,000 words or something like
2:10:31 that what it sounds like we're going to
2:10:33 be able to do is take the context of an
2:10:35 entire
2:10:37 conversation and chat GPT will remember
2:10:39 that and then you can connect that to
2:10:41 other conversations so it can have
2:10:43 memory across
2:10:46 conversations so if I'm talking about
2:10:48 redesigning my kitchen here a vegan
2:10:51 lifestyle here and modes of
2:10:53 transportation here three separate
2:10:55 conversations but they're all sort of
2:10:57 related
2:10:59 philosophically somehow I'm going to be
2:11:01 able to connect those conversations and
2:11:03 chat GPT will have memory across
2:11:08 them which is a big
2:11:10 deal just because right now everything's
2:11:13 very siloed and and you put a lot of
2:11:16 work into having this conversation and
2:11:18 then you jump into something
2:11:19 semi-related it has no idea what you
2:11:21 were talking about
2:11:26 before folders for conversations yes
2:11:29 soon you might be able to wait what was
2:11:34 [Music]
2:11:36 that oh cool I didn't know it's on the
2:11:39 app it it's not yet well it is for some
2:11:41 people looks like it's coming for some
2:11:43 people chat GPT has a 4K context
2:11:46 window all of them on AP Pi have 128 for
2:11:50 GPT
2:11:51 turbo I thought GPT had 8 8K context
2:11:55 window chat GPT but I don't know it's it
2:11:59 what it is not is it's not
2:12:03 128k it's it's annoyingly
2:12:08 small soon you might be able to call in
2:12:10 YouTube's API that would be
2:12:15 [Music]
2:12:17 nice
2:12:21 just share it again Kyle wants it what
2:12:23 do I
2:12:31 want forever hooked just got chocolate
2:12:34 on my phone
2:12:37 screen chocolate
2:12:43 shmir oh I did I I pitched the salon
2:12:46 earlier Brandon I I did a pretty long
2:12:48 one
2:12:51 how to make files on chat
2:12:53 GPT all right here's something you may
2:12:56 not know so if you go to chat
2:13:01 GPT if you're in chat GPT 4 one of the
2:13:05 things it has in it is a thing called
2:13:07 code
2:13:08 interpreter and what code interpreter
2:13:10 can do is use
2:13:14 Python to do
2:13:16 [ __ ] if you're like I don't know what
2:13:18 python is it's a programming
2:13:21 language and so what code interpreter
2:13:24 does is it
2:13:25 will in the
2:13:27 background because it's got access to
2:13:29 python libraries it will write and
2:13:32 execute python code and so what it
2:13:35 allows you to do is analyze wait no oh
2:13:39 no how to group a file how to group
2:13:41 things to a
2:13:44 file within chat
2:13:47 GPT
2:13:52 I don't quite know what you're asking
2:14:02 for
2:14:04 um well what but what it can do is it
2:14:07 can actually so so here let me let me
2:14:09 try something here so let me let me make
2:14:12 um give
2:14:14 me five ideas about
2:14:20 um being a vegan that
2:14:24 won't annoy other
2:14:32 people lead by
2:14:34 example you said you wish there was a
2:14:36 way to group things POS Brandon already
2:14:38 showed us a way around
2:14:47 that
2:14:53 um all right so there's there's five
2:14:55 things there so now let me
2:14:57 say now give me five ideas for how to
2:15:00 keep my 70s muscle car
2:15:14 clean and then I'll do
2:15:31 [Music]
2:15:36 I'm going try one other
2:15:39 thing create an image
2:15:42 of a hedgehog riding a
2:15:47 camel
2:15:49 people like the hell is he doing I I
2:15:51 don't know I'm actually I'm actually
2:15:52 experimenting with
2:15:54 something so you can have a conversation
2:15:58 right you're having a conversation in
2:15:59 chat gbt you're you're doing a bunch of
2:16:01 disperate things within the
2:16:06 conversation all right there's there's
2:16:08 that okay now I'm going to
2:16:12 say please create a nicely
2:16:16 formatted please create a nicely
2:16:18 formatted Word document that I can
2:16:21 download that includes the five items
2:16:25 about being a vegan comma the five items
2:16:29 about how to maintain my muscle car
2:16:32 comma and this picture into the single
2:16:42 document so see where it says analyzing
2:16:44 here what it's actually doing is
2:16:46 generating a Word document in and it's
2:16:48 it should pull out all of this
2:16:51 content you know minus the prompts and
2:16:54 put them in a Word
2:16:58 document so in the case that we're
2:17:00 talking about where there's this new
2:17:02 feature coming I don't have access to it
2:17:04 yet where you could have memories across
2:17:07 conversations you could potentially you
2:17:09 know create a new conversation that
2:17:11 pulls in ideas from all of them let's
2:17:14 see if this actually does this I this
2:17:16 should work
2:17:19 but we'll see what's it doing in
2:17:24 here
2:17:25 yeah so docx is the library so it's
2:17:30 creating it's creating
2:17:33 um a word dock then it's putting in the
2:17:36 vegan tips then it's putting in the car
2:17:39 tips and then it put
2:17:41 in an image path I have a feeling the
2:17:44 image is going to break but we'll see it
2:17:47 seems there was an issue adding the
2:17:48 image to the document probably due to an
2:17:50 unrecognized image format or a problem
2:17:53 processing the file let me correct that
2:17:55 and try
2:17:59 again the document's been successfully
2:18:02 created here it is download the
2:18:08 document all right
2:18:11 so we now
2:18:17 have
2:18:33 there's the vegan stuff there's the 70s
2:18:35 muscle car stuff and then there's the
2:18:37 the image was intended to be here but
2:18:39 could not be added due to a technical
2:18:41 issue so it it didn't do that part right
2:18:43 but here's a nicely formatted Word
2:18:46 document right that's got you know the
2:18:50 living as a vegan without annoying
2:18:52 others there's my five things put in a
2:18:55 page break here's my maintaining 70s
2:18:57 muscle car things there's that and then
2:19:00 if it did the image right and there's
2:19:01 probably a way to get it to do the image
2:19:03 right um I don't know if that's what
2:19:05 you're asking Danielle but that is
2:19:07 something you can
2:19:08 do um that's not necessarily the things
2:19:11 I was talking about uh with the with the
2:19:14 memory across conversations
2:19:16 though all right let's see oh it's late
2:19:20 it's late people it's late on a Saturday
2:19:22 night it's late it's late it's l it's l
2:19:31 it's all right
2:19:34 so if you are
2:19:38 digging the kind of stuff we're talking
2:19:40 about here if you want to be in a
2:19:43 community that is up to this stuff join
2:19:46 the AI Salon go to the
2:19:49 salon. check out about the salon welcome
2:19:52 to the salon read about what we're up to
2:19:55 read about our
2:19:57 values if you resonate with those
2:20:00 values go introduce yourself join join
2:20:02 the salon so if you go there it's a link
2:20:06 tree the very first link is to the AI
2:20:09 Salon
2:20:11 Community all right so go do that it's
2:20:15 they're really amazing people
2:20:18 and then if you want to if you want to
2:20:20 support this channel follow me follow me
2:20:23 follow me remember Oddworld anybody love
2:20:26 Oddworld wasn't Oddworld the best the
2:20:28 little farts the
2:20:34 little
2:20:38 um follow me you can subscribe to my
2:20:41 lives you can buy the series I've got a
2:20:43 series in one of the corners here will
2:20:45 support the channel um um and joining
2:20:49 the
2:20:50 salon will support the
2:20:52 channel and then most importantly just
2:20:55 keep showing up to these fun
2:20:59 fact I do these lives seven nights a
2:21:02 week for like two to three hours a
2:21:05 night you're like you got to be [ __ ]
2:21:08 kidding me no that's what my wife
2:21:13 says seriously honey can't you take a
2:21:15 night
2:21:16 off I could but I choose not
2:21:20 to so come
2:21:22 back sometimes it's fun sometimes it's
2:21:24 pey sometimes it's boring as
2:21:28 [ __ ] tap tap some love we're at 33,000
2:21:32 tap it up
2:21:35 make but if you want to if you want to
2:21:37 learn about this AI stuff and if you
2:21:39 want to if you w
2:21:40 to here's the reason you keep coming
2:21:43 back here here what what I will
2:21:44 hopefully not teach you is specific
2:21:48 tactical [ __ ] about a tool or a
2:21:51 technique what I'm hoping you take away
2:21:54 from this channel is learning to
2:21:58 learn like just get curious about AI
2:22:01 just play with it don't feel like you
2:22:03 got to solve it don't feel like you got
2:22:05 to catch up with anyone just play play
2:22:08 play get
2:22:13 energized because if you get energized
2:22:16 and if you get engaged and if you start
2:22:17 learning and you get in that curious
2:22:19 beginners
2:22:21 mindset and you just hang out here for
2:22:24 the
2:22:24 next I don't know 3 months 6 months a
2:22:28 year and join communities like this
2:22:31 where you're hanging out with other
2:22:33 people that are doing this the amount of
2:22:35 opportunity for all of
2:22:38 us is
2:22:43 Indescribable that's all it takes you
2:22:45 don't need to be a technologist
2:22:51 doesn't hurt but you don't have to be
2:22:54 that yeah I just learned that you can
2:22:56 create images within chat GPT awesome
2:22:59 yeah it's and there's so much more in
2:23:02 there big big art man it's
2:23:05 it's it's
2:23:09 mindblowing it's it really is
2:23:13 mind-blowing what you can
2:23:16 do and again my favorite quote from Sam
2:23:19 mman in recent
2:23:22 weeks chat GPT
2:23:24 4 is best understood as a
2:23:28 preview progress here is not
2:23:32 linear so what he's what he's talking
2:23:35 about
2:23:36 is so so
2:23:39 gp4 was produced four years ago we've
2:23:43 been
2:23:44 playing so so GP bt4 was produced 4
2:23:50 years ago we've been playing
2:23:53 with lost
2:24:00 connection
2:24:01 strange am I
2:24:03 back that was weird my camera just went
2:24:10 dead all right hopefully can you still
2:24:12 hear
2:24:13 me here let me let me scroll to the
2:24:15 bottom yep you're back okay that was
2:24:18 weird
2:24:19 Okay um I thought that was one of this
2:24:22 intermission brought to you by Comcast
2:24:24 exactly I I thought that was one of
2:24:26 those you're kicked off Tik Tok forever
2:24:28 you're permanently
2:24:31 banned that was strange yes rebec and
2:24:34 can hear you okay cool um uh what was I
2:24:38 talking about thank you Lord digital
2:24:41 Gods oh chat gp4 is best understood as a
2:24:46 preview the progress here's is not
2:24:47 linear what that means
2:24:51 is as sophisticated as chat GPT
2:24:55 is and it's quite
2:24:58 sophisticated there's a lot to learn
2:25:00 within it there's a lot that you can
2:25:02 explore and push the edges of it's best
2:25:06 understood as a preview meaning this is
2:25:09 Child's Play compared to what's
2:25:12 coming progress here is not linear
2:25:15 meaning you know this is this is linear
2:25:18 chat gp4 is here gp5 is here
2:25:25 linear this is more like this chat gbd4
2:25:28 is
2:25:29 here chat gb5 is somewhere up here like
2:25:33 we think it's growing like this and it's
2:25:37 going to show up and be a completely
2:25:39 different thing so so learn as much as
2:25:43 you can at the level that we're playing
2:25:46 with right now
2:25:48 we're playing with training wheels we're
2:25:50 playing with wooden
2:25:53 blocks and what's coming are you know
2:25:56 sophisticated
2:25:58 computers like the difference is going
2:26:00 to be vast so let so what we all want to
2:26:03 do in the next year is just keep playing
2:26:06 just keep playing just keep playing
2:26:07 don't get attached to
2:26:10 anything we're all paying to be beta
2:26:12 testers we absolutely are and I'll
2:26:15 gladly do it because these are really
2:26:17 cool [ __ ] tools we get to play
2:26:19 with they're learning from us we're
2:26:22 learning from them we're all learning
2:26:23 together they're learning how we're
2:26:26 using their tools which is going to
2:26:28 affect how they do the next
2:26:33 thing GPT 5 will be an order of
2:26:36 magnitude better exactly Captain mut and
2:26:39 people don't understand what that means
2:26:41 order of magnitude means it's it is not
2:26:44 linear it is like this
2:26:47 you know th this is linear this is this
2:26:51 is we're here we're thinking we're going
2:26:54 on this trajectory and all of a sudden
2:26:55 we're going to go
2:26:59 Z do I have any [ __ ] clue what that
2:27:02 means I have no [ __ ] clue I know two
2:27:05 concepts and there's probably way more
2:27:07 than
2:27:08 two but there's reasoning will be coming
2:27:13 which basically means chat GPT will
2:27:16 understand what it said
2:27:17 and be able to reason and figure out if
2:27:20 it got it right and not give you an
2:27:23 answer until it knows it got it
2:27:25 right so basically mistakes will drop to
2:27:28 near zero in
2:27:30 theory and then the other concept is
2:27:33 autonomous
2:27:35 agents
2:27:37 where rather than you having to tell
2:27:39 chat GPT to do
2:27:41 something you'll just give chat GPT a
2:27:44 goal and it will generate a bunch of
2:27:46 assist
2:27:47 agents that go off and take action on
2:27:50 your behalf and come back with the
2:27:53 results so between reasoning and
2:27:58 agents it's a completely different
2:28:01 experience completely different product
2:28:03 completely different possibilities and
2:28:05 then I don't even know what El what else
2:28:07 what
2:28:09 else and there's just the quality of the
2:28:12 data will probably go up dramatically
2:28:14 we're playing with four-year-old
2:28:15 technology in in chat GPT
2:28:18 before so what's the new
2:28:24 [ __ ] it costs them more for me to play
2:28:27 than what they make make me pay yeah
2:28:29 absolutely that's the thing laughing
2:28:31 Thunder we are we are really lucky that
2:28:34 chat GPT has kind of established the
2:28:37 monthly nut for this technology at 20
2:28:39 bucks because everyone else in the
2:28:41 industry everything's between 20 and 30
2:28:44 bucks I I feel chat GPT could have
2:28:47 easily gone 60 70 80 120 and they
2:28:52 didn't yeah they're losing money on us
2:28:55 Show and Tell in a regular channel
2:28:56 before you go absolutely let's go one of
2:28:59 the fun
2:29:01 things is that the irregular so this
2:29:04 channel we have our own little private
2:29:07 Clubhouse and so is goian splatting a
2:29:13 thing that's good that's good digital
2:29:16 gods
2:29:18 without warning Kyle mentions gaui and
2:29:20 splatting
2:29:22 again so so we can do our own inside
2:29:25 jokes and and do fun things oh that's
2:29:27 beautiful TK that's beautiful oh nice
2:29:31 piano series oh that one's great oh look
2:29:33 at her creepy hands damn and like one
2:29:37 has a reflection and one
2:29:40 [Laughter]
2:29:42 doesn't
2:29:44 201
2:29:46 wow
2:29:47 artist that's very cool Escaping The
2:29:49 Machine artists Escaping The Machine
2:29:52 love
2:29:57 it wow gorgeous oh
2:30:09 creepy the
2:30:12 Irregulars I can almost hear the uh the
2:30:15 soundtrack to that sick
2:30:22 come very cool TK those are very cool
2:30:27 I'm enough just as I am yes
2:30:30 exactly no more apologies get curious
2:30:33 play apply share repeat that's
2:30:37 it this this is the this is the recipe
2:30:42 in my opinion for the next two or three
2:30:44 years just be curious play apply it to
2:30:47 work share what you're learning repeat
2:30:50 it over and over
2:30:51 again just keep learning keep learning
2:30:54 keep sharing keep putting it out there
2:30:55 keep letting people know what you're
2:30:59 doing oh that's gorgeous me one day love
2:31:02 it Triumph I'm in get curious play apply
2:31:06 repeat it can't do apply can
2:31:09 it doubles up the L's every
2:31:13 time those are so cool TK it was
2:31:16 Stonehenge
2:31:19 nice it's mathem magical
2:31:24 nice that's great
2:31:28 Siri no go
2:31:32 away
2:31:36 okay well these are Cory's yeah Cory
2:31:39 Sandlers so
2:31:41 cool all right so that's that so let's
2:31:43 go to show and tell so um
2:31:48 Show and
2:31:51 Tell hey I
2:31:53 Salone pretty
2:31:55 puppy I just do a
2:31:58 leaper AI learning lab oh that one's
2:32:01 cool very
2:32:05 cool oh nice Hedgehog riding the waves
2:32:09 That's it man better to be on the
2:32:12 surfboard than to be in the Cozy beach
2:32:15 hut with what's coming oh that's
2:32:19 gorgeous I loves me some Froggies look
2:32:21 at
2:32:22 that damn that's
2:32:26 good yeah so so if you're not in the AI
2:32:30 Salon you go
2:32:34 here click on the first link and join us
2:32:38 start out with welcome to the salon read
2:32:41 about what we're up to read about about
2:32:44 our
2:32:45 values and if you dig it jump
2:32:49 in it's an amazing group all right I got
2:32:52 to go people I got to get out of
2:32:55 here um thank you to the Irregulars
2:32:57 thank you to my
2:32:59 mods thank you to anyone that showed up
2:33:03 tonight it's [ __ ] Saturday night we
2:33:05 had Friday night date night last night
2:33:07 and I lost my [ __ ] and you came back and
2:33:10 some new people came back so welcome
2:33:12 keep hanging out here 43,000 46,000
2:33:15 likes you guys are amazing um all right
2:33:19 so so let's let's see who we got in here
2:33:21 tonight Stacy thank you Earth rocker
2:33:23 good questions tonight thank you
2:33:24 Amelio's wife good to see you TK amazing
2:33:27 work that was really beautiful what
2:33:28 you're sharing in there Brian Whitney go
2:33:30 to the chat and the Irregulars Channel
2:33:32 okay I'll do that after this quite
2:33:34 honestly I'm playing real life Monopoly
2:33:36 with AI totally G GPT billionaire Love
2:33:39 It joker good to see you laughing
2:33:42 Thunder great to see you good
2:33:44 interactions tonight thank you Sonic age
2:33:46 science hello Brian Whitney uh Megan
2:33:49 good to see you Tobias Dylan what's up
2:33:53 Tock um Brandon thank you so much
2:33:57 sherbear as always Mr it Source Camp who
2:34:01 else we got in here Mary Mary
2:34:04 Danielle um I forget your name with the
2:34:08 Russian with the
2:34:10 Russian words I knew it but now I can't
2:34:14 remember it V dancer thank you so much
2:34:16 Kev tuck Captain mut man a lot of new
2:34:19 people in here tonight this is so cool
2:34:21 Zippy zambo Detroit Jay lower the
2:34:25 hallucinations it's math
2:34:27 magical Pate p in here of course Julia j
2:34:32 all right everybody um really fun night
2:34:34 thank you thank you for inspiring me and
2:34:38 inspiring one another and being generous
2:34:40 and amazing people um I have I have deep
2:34:43 deep appreciation for this community so
2:34:46 thanks peace out I'll see you tomorrow
2:34:48 Super Bowl Sunday it's Taylor's game
2:34:51 we're just playing in
2:34:54 it that woman's a [ __ ]
2:34:57 beast
2:34:59 um just amazing single-handedly
2:35:02 transforming all of all of society she's
2:35:06 like [ __ ] the music industry I'll just
2:35:09 have my fans completely redefine how all
2:35:13 this [ __ ] works and then now she's
2:35:15 redefining football as we know
2:35:19 it Super Bowl Edition tomorrow yeah we
2:35:22 we'll definitely do that because it
2:35:23 starts at 4:30 here so either it'll be
2:35:26 toward the end or I don't know we'll see
2:35:28 all right good to see you all peace have
2:35:30 a good night