Jan 3, 2024 - A Beginner's Guide to ChatGPT and Beyond

Description

Chapters

Transcript

0:01 [Music]
0:14 [Applause]
0:15 [Music]
0:17 let us
0:20 beun
0:24 together i' got some real State
0:27 [Music]
0:29 here
0:31 so
0:33 cigarettes at Mrs
0:36 Wagner's and we come to look for
0:42 [Music]
0:47 America Kathy said as weed the in
0:53 [Music]
0:55 Pittsburgh seems like a dream to me
0:59 now
1:01 it took me 4 days it it took it took me
1:05 4 days to hit from
1:09 [Music]
1:26 saging an impeccable
1:29 Style
1:30 [Music]
1:34 don't ever let them steal your
1:38 joy gentle ways keep them from running
1:45 wild they K in your face dress you down
1:49 and tell you that your place is in the
1:51 middle when they hate the way you
1:55 [Music]
1:59 shine
2:02 hey buddy you doing good yeah hey
2:06 [Music]
2:21 everybody all right we're done you're
2:23 good you did the people are happy they
2:26 said hi champy you're good you got high
2:28 chies you can go lay down oh you want to
2:30 come for love yeah oh yeah oh yeah
2:36 oh all right everybody full chilly here
2:39 I'm all bundled
2:41 up how is
2:44 everyone happy Wednesday night I didn't
2:47 know what day it was I looked at my
2:48 little my
2:50 computer yeah yesterday was yesterday
2:53 was first day back from work yeah today
2:56 was like a work day it was like a normal
2:58 work
2:59 day
3:02 I'm mostly better from my from from
3:04 Illness but I don't have the stamina
3:06 like I hit 3:30 in the afternoon I'm
3:07 like I'm done I'm done so slowly getting
3:12 back to it oh man let's see I was
3:16 wondering which was better best better
3:19 bet between G
3:22 wait between Bing and Google AI
3:26 grams grams
3:29 llms
3:32 between Bing and
3:34 Google well by far the
3:39 best AI model out there you're a
3:42 beginner okay cool by far by far by
3:47 Miles right
3:49 now chat GPT from open AI is the best we
3:54 have to stop meeting like this my my
3:56 husband calls you my boyfriend sorry
3:58 about that
4:00 sorry Mr techy
4:03 Becky we're just hanging out it's
4:06 totally
4:07 [Laughter]
4:09 platonic um that's funny um yeah by far
4:14 chat GPT is the best
4:16 Now Bing or co-pilot so you can go to
4:20 bing.com chat or you can go to co-
4:24 pilot. microsoft.com they're both the
4:27 same thing they're just different
4:29 branding
4:31 and that is Microsoft's version of chat
4:34 GPT and there's I there's conf I talked
4:36 to someone about this today they're like
4:37 but it's the same thing it's not the
4:39 same thing it's not the same thing it's
4:42 not the same thing it's not the same
4:45 thing um here's how it works you you
4:48 have the the the large language model
4:51 itself is called
4:53 gp4 that's the thing with all the data
4:55 in it all the
4:56 math this thousand dimensional latent
4:59 space space with trillions of words in
5:02 it that's GPT 4 chat GPT is an
5:07 application that was built that sits on
5:09 top of it that asks it for its data and
5:13 then Bing chat or co-pilot is a separate
5:17 different
5:18 application that mostly does the same
5:21 thing that calls the same Bas large
5:24 language model but is different and not
5:26 as good why is it not as good because
5:30 Microsoft has put so many safety guard
5:32 rails in it that it makes it
5:33 functionally
5:34 [Laughter]
5:37 unusable um and it's just it's just not
5:39 as good and then and then the Google AI
5:44 is called Bard bard. goole.com so I I
5:47 would sort of put them in that order I
5:48 would put I would put them
5:51 in chat
5:53 GPT Bing chat Bard although Bard and
5:58 Bing chat are kind of a tossup up cuz
6:00 Bard does some things better than better
6:03 than Bing
6:04 does with all these tools what you
6:06 really have to do is try them but what
6:08 where I would start is with chat GPT and
6:10 then and then the next question becomes
6:12 should I pay the 20 bucks to open AI for
6:15 chat GPT plus the answer is yes
6:18 absolutely yes if you can afford it if
6:20 you can't afford it don't do it but if
6:22 you ever drink coffee at Starbucks more
6:25 than three times a month there's your 20
6:27 bucks stop that [ __ ] make coffee at home
6:29 with your little folders
6:32 can like I do look
6:34 see that is not
6:39 Starbucks um chat gbt plus is a really
6:42 good deal right now just because it's
6:44 it's just it's just there's nothing like
6:46 it I think I think a good investment of
6:48 everyone's time right now would be learn
6:51 as much as you can about chat GPT plus
6:54 learn how to do code interpreter learn
6:55 how to leverage the fact that it's
6:56 connected to the internet learn how to
6:58 use dolly3
7:01 learn how to make
7:02 gpts we just did a big event this past
7:05 Saturday called GPT for good where Ann
7:08 Murphy and I from empowered fundraising
7:11 alternated every 3 hours for 24 hours
7:14 teaching people how to make gpts and
7:16 those people made 150 gpts in a 24-hour
7:20 period our goal was 100 it was really
7:22 cool um so you should learn how to do
7:27 those all right let's get started
7:30 all right okay thank you free or not um
7:33 chat GPT 3.5 is free four is 20 bucks
7:37 Bing is free Bard is free now there's a
7:42 new version of Bard
7:46 coming that's based on a model called
7:48 Gemini Ultra that's supposed to be as
7:52 good or better than chat G pt4 but it
7:54 hasn't launched yet that will probably
7:57 be 20 bucks a month is my guess that's
7:59 that's my
8:00 guess because um it's just a bigger
8:05 badass or model and it probably cost
8:07 them a fair amount of money to run
8:09 it and yeah so that um but there are
8:14 free options for all three of those but
8:16 but Bing and Bard right now are
8:18 completely free and then 20 bucks a
8:21 month for chat GPT chat
8:26 je um and we can go play with those I
8:29 can go show you some of that stuff but
8:30 let's let's dig through here and look at
8:32 some comments see who's in here tonight
8:34 I see Amelio's wife made it on time
8:37 she's starting the New Year's off right
8:39 I think she made a a New Year's
8:40 resolution and not be late so much she
8:43 there there was a lot there 2023 was was
8:47 was a a a a a veritable tulip field of
8:52 demerit marks on Amelio's wife's report
8:56 card it was just it was are there report
8:59 cards here oh yeah it's the learning lab
9:03 we don't screw around here this is not
9:05 this is not
9:07 optional people are required to be here
9:10 night after night after if you're if
9:12 this is your first time here buckle up
9:19 buttercup you can't not be
9:23 here once you learn about this AI stuff
9:26 you're like oh
9:28 [ __ ] guess got to learn it now and if if
9:32 you're thinking like there's an easy way
9:34 out like you can make
9:36 money these
9:38 idiots they're full of [ __ ] don't listen
9:41 to them I'll tell you the truth here's
9:43 the truth all these tools are janky
9:46 pieces of [ __ ] and no one knows what the
9:48 hell they're talking about including me
9:50 especially me my qualifications
9:54 NADA and that's what we love about it
9:57 all right I made a GPT and then it said
10:00 I was using too much resources oh Bill
10:03 how could you do that to Sam Alman you
10:05 used up all of Sam's resources so
10:10 rude that's just called uh rate limiting
10:14 you got rate
10:15 limited so chat gp4 I think you can do
10:18 40 40 calls in a three-hour period so
10:22 40 things post prompts whatever and then
10:27 it and then it's like I'm I'm tired now
10:30 I'm going to go away and then it makes
10:32 you take a
10:34 break it's just where it is right now
10:36 that'll get
10:37 better but that's a good thing that
10:40 means you were you were uh you were
10:42 working it good I missed your lessons
10:44 where do I make a make GPT learn to make
10:46 gpts well I can teach you tonight how
10:48 are you in the fam recovering thank you
10:50 a human being uh good I'm still I still
10:52 got a little sinus thing um my one son
10:54 Oliver is still pretty
10:56 bad Zach's doing okay but not feeling
10:59 great great Gabby's got a sinus
11:01 infection so we're we're all doing okay
11:02 like I'm I'm pretty much on the mend but
11:04 I'm the I'm the one of the four of us in
11:06 the family with a decent immune system
11:08 their their immune systems are
11:12 special all right thank you lime disease
11:20 um um but thank you for asking that's
11:23 very nice
11:25 uh Corey Sandler Pottery what's
11:28 happening amilio white finally on time
11:32 oh my connection is janky and I'm
11:34 multitasking it's going to happen Vicky
11:37 it's gonna happen I'm gonna do this no
11:40 okay Lita Lita Vico Twisted Twisted
11:43 Crone um absolutely do this dive in and
11:47 here's the deal don't put pressure on
11:49 yourself like you've got to learn it
11:52 even though this called the AI learning
11:53 lab don't put pressure on yourself you
11:55 have to do important stuff with it play
11:58 play Play Just Like play like like have
12:02 fun with it like make a poem make a song
12:06 make a meal plan that includes gummy
12:09 bears like seriously just do anything
12:13 you can to keep yourself
12:15 engaged make images like the image
12:18 that's here the AI learning lab Pokemon
12:20 cards there with our little animals on
12:21 it um that was made in chat GPT in Dolly
12:25 3 so when you pay for the 20 bucks a
12:27 month for chat GPT Plus inside of it is
12:30 dolly3 so you can just say make me an
12:32 image of whatever and if you want to do
12:34 that on Bing right now you just go to
12:37 bing.com or co-pilot microsoft.com and
12:40 do the same thing make me an image of
12:43 whatever and it
12:44 will
12:47 um and just
12:49 play and here's another here's another
12:52 um hack I a hack I hate that [ __ ]
12:55 term here is
12:58 another here here's the corporate the
13:00 corporate version of hack here's a best
13:01 practice for
13:03 you listen listen Lita welcome like to
13:07 give you a best practice for prompt
13:09 engineering
13:11 um one of the things that can happen
13:13 when you start playing with this so it's
13:15 it's just like a Google search box right
13:17 chat GPT and Bing and Bard and all those
13:19 are just like a Google search
13:22 box and you and you look at it and
13:24 you're like okay what should I what
13:27 should I put in there and then your mind
13:29 goes blank it's called White Page
13:32 syndrome right where it's just
13:36 like and so so here's a trick to prevent
13:41 that don't sit down at chat
13:44 GPT and then try to come up with
13:47 something to
13:48 do find a problem in your life or a
13:51 problem in someone else's life or some
13:54 project you've mean been meaning to do
13:56 um from very small to very big doesn't
13:58 matter but find a project first and so
14:02 you know your friend Sally wanted to put
14:04 together a gardening plan for the spring
14:06 okay great
14:08 perfect grabs Sally say Sally come on
14:11 over we'll work on your gardening plan
14:13 with Chad GPT she'll be like what are
14:15 you talking about and you you just be
14:17 like it doesn't matter just come on over
14:19 she'll be like okay and then she gets
14:21 over there and you sit down at the
14:22 computer she's like what are you doing I
14:24 thought we were going talking about my
14:25 gardening and you're like no no it's AI
14:27 it's cool she's like AI
14:29 I don't want to I don't I'm not a a a
14:32 Stamford educated mathematician what are
14:34 you talking about
14:37 Ai and then ask it to make Sally a
14:40 gardening
14:41 plan and watch her face melt off she'll
14:47 go so that's the that's a little trick
14:49 to not be intimidated by it is don't
14:51 start with the
14:52 technology start with what you're going
14:54 to do with
14:55 it haven't been online till now only
14:59 been up for a few hours today finally
15:01 catching up on sleep TK what's
15:03 happening well well welcome I I am here
15:06 to put you back to sleep that that is my
15:09 that is my lot in
15:12 life so TK so I mentioned before I I'll
15:16 get back to telling you how you can get
15:18 access to all the GPT stuff we talked
15:20 about in a second
15:22 here um man I still have a lot of
15:24 congestion it's driving me [ __ ] crazy
15:27 um cuz what
15:29 here's the deal I don't think I have
15:31 congestion and then I start talking and
15:33 the minute I start talking like the
15:35 faucets come on somewhere somewhere
15:37 between my brain and my face there's a
15:40 big 55 gallon drum of uh
15:45 fluid that just empties into my face the
15:47 minute I start talking thanks Siri
15:50 that's really
15:51 awesome
15:54 uh H I don't know what I was talking
15:56 about I built a GPT for ai. goov oh
16:00 that's so
16:02 cool what what is a.gov is that the is
16:06 that the uh the executive order what is
16:11 that
16:13 ai.
16:19 goov you go there it just says we're
16:21 watching you now hello Kyle we've been
16:25 watching you welcome to ai. goov
16:30 making AI work for the American people
16:32 join the national talent
16:34 [Music]
16:36 surge administrative actions build your
16:38 own build your AI skills very
16:42 cool so you made a GPT to to do this
16:45 okay that's that's super cool lunar
16:47 stick
16:49 awesome oh I was talking about TK so
16:53 TK who said she slept all day
16:56 slacker
16:57 um she came so so so we did this thing
17:01 called GPT for good last Saturday it
17:03 started 6:00 in the morning Pacific time
17:06 on
17:06 Saturday and Ann Murphy did a what
17:10 turned out to be a 2 and a half or three
17:12 hour session at 6:00 a.m. and then I
17:15 went at 9: so her her sessions were on
17:17 Zoom mine were on Tik Tok so I did my
17:21 Tik Tok session then at noon she did her
17:23 zoom and then at 3 I did another Tic Tac
17:24 we did that for 24 hours TK showed up
17:28 for the first one and stayed for all of
17:32 them she came in intending to make a GPT
17:35 or two I think she had eight on her list
17:38 initially she goes I'll make one or two
17:40 of these she ended up making 20 gpts in
17:44 a 24-hour
17:47 period She was a machine and she also
17:50 won we did a um an an Murphy generously
17:53 donated 250 bucks to whoever submitted a
17:57 GPT uh for a drawing that the 250 bucks
18:02 would go to the nonprofit of the
18:04 Winner's Choice and TK won the drawing
18:07 and so her 250 bucks is going to the
18:10 Minnesota firefighters Museum which is
18:12 super cool so what all does code
18:16 interpreter do it is not explained in
18:18 writing anywhere to block Tom it it
18:23 is
18:25 insane code interpreter is insane in
18:28 many ways
18:29 one it's insane in what it does like
18:31 what it does is remarkable it is also
18:34 insane how poorly documented it is and
18:38 how um no one knows it's there no one
18:42 knows it's there it's just sad um they
18:47 even defaulted it to be turned off when
18:49 you make a GPT all right so so let's get
18:52 in in in here um it's the official AI
18:54 arm yeah that's super cool very very
18:58 very cool
18:59 um I'm such a slacker and underachiever
19:02 Broody hen I know what did you make you
19:03 only made like eight gpts or something
19:06 like that I know it's just you can't you
19:08 can't compare yourself to TK she's just
19:11 on another plane her and
19:13 lunatick Lun these these gpts came out
19:17 and lunatick found her voice and she's
19:19 just
19:23 like oh last count I heard of hers she
19:26 was at 74 but I assume the i.gov GPT is
19:30 75 can you demo code interpreter I'm
19:32 going I'm I'm going to demo code
19:34 interpreter I'm just trying to figure
19:35 out where I want to start
19:36 tonight okay so here's what we're going
19:38 to do let me go show you something for
19:41 cuz cuz quite frankly it was this this
19:44 community when when when an and I
19:46 decided to do this 24-hour GPT for good
19:49 thing this community just [ __ ] kicked
19:52 into gear Vicky who has bad internet
19:54 connection tonight she built a website a
19:57 bunch of people made images got the word
20:00 out and then everyone [ __ ] showed up
20:03 for the party and and stayed for the
20:05 party like it was just amazing so
20:13 um go to come with me come with me and
20:17 you'll see it's a world of Pure
20:22 Imagination um go to the
20:25 salon the s a l n I and that'll take you
20:30 to a link
20:33 tree that is this and the very first
20:36 link is the GPT for good event directory
20:40 and this is the site that that Vicky
20:42 Baptist built for us generously very
20:45 generously and then she hung out all 24
20:48 hours while we were having people make
20:50 these things and she was taking the gpts
20:54 as they were being submitted and she
20:57 built this directory so there's 151 or
21:00 52 I don't know at this point more than
21:02 150 these are all gpts that people built
21:05 in this 24-hour
21:07 period absolutely amazing absolutely
21:10 amazing and and she she goes we hit the
21:13 goal of 100 gpts and kept going and so
21:15 here's all the ones we did after that so
21:18 just just remarkable so go check those
21:20 out they're quite inspiring they're
21:22 they're all for you know some cause or
21:25 or
21:25 nonprofit um some of them you know when
21:28 significantly beyond what what we taught
21:31 in the in the
21:33 sessions um some people are still
21:35 updating theirs um making them better
21:37 and better and better which is super
21:39 cool um okay so now let's go back to
21:42 that link
21:44 tree so on that link tree there's also a
21:47 YouTube playlist I still have to get um
21:51 an's Zoom recordings I I could watch
21:54 them today but I couldn't download them
21:56 I didn't have access so as soon as I
21:58 have them but all of my Tik Tok sessions
22:00 are there the second one is the YouTube
22:02 playlist that's got the actual sessions
22:04 themselves where we taught people how to
22:05 do this each session is a standalone
22:08 thing if you can if you watch one of
22:09 them you'll you'll essentially get it um
22:12 did you notice I cleaned up the site
22:14 Rose are even more even now I know it it
22:16 was looking really good it's it's lovely
22:19 very very nice oh yeah oh look at that
22:22 yeah you did a little table action
22:23 didn't you oh Vicki getting Fancy with
22:26 the tables I like it
22:32 what I know about tables is they're a
22:33 pain in the ass so thank you for doing
22:35 that it looks great yeah they look great
22:39 um really cool
22:42 um
22:44 so congratulations on an amazing
22:46 collaboration I no it it it was really
22:49 really really quite inspiring um so if
22:51 you go to the the YouTube playlist um
22:54 you can see my Tik tok's there and then
22:55 an's zooms will be there hopefully
22:57 tomorrow and what's what's nice is
23:00 they're sort of two different flavors
23:01 right my Tik toks are like this they're
23:03 kind of chaos and zany although that you
23:05 know it was much more Focus than my
23:07 normal lives um but Ann's were on zoom
23:11 and so mine tend to be shorter and just
23:13 more you know here's the here's the
23:15 stuff to do and I I would occasionally
23:16 answer questions that I saw and because
23:19 it was a zoom meeting there were people
23:20 there you can see them they're asking
23:22 questions hers tended to go most of the
23:24 three-hour chunk she had for her her
23:26 little time slots so if you want to see
23:29 Q&A and and meet some of the people
23:32 who've been doing the things then watch
23:33 her videos okay then the next thing down
23:36 is the GPT for good prompt sheet this is
23:39 the thing that we used to build the gpts
23:44 and so it starts out it it's it's nine
23:46 steps and then some sharing instructions
23:49 eight steps eight
23:51 steps and the first step is making the
23:54 GPT and I'll I'll read The Prompt The
23:57 Prompt is this GPT is designed to make
24:00 it very easy for someone to learn about
24:01 a nonprofit or forood organization named
24:04 org
24:05 name the knowledge about the
24:07 organization should be pulled from
24:09 website URL that's it so watch watch how
24:12 we do this so so let me show you what a
24:14 GPT is first so here's what we're going
24:16 to do there's there's one Advance videos
24:20 here's what we're going to do we're
24:21 going to we're going to go to chat
24:24 GPT I'm going to show you what a GPT is
24:27 first then I'm going to show you how to
24:28 make a GPT and then we'll go use code
24:30 interpreter all right sound like a plan
24:33 sound like a plan people are you digging
24:35 it are you groovy with the scene let's
24:38 see how do I straighten up my machine
24:39 all
24:45 [Music]
24:48 right why does he sing like that it's
24:50 very annoying he thinks he's a singer he
24:53 thinks he's some sort of Seline Dion or
24:55 something like that the world's greatest
24:57 singer I don't think so Mr Shannon
25:01 barely hold a tune I like hearing champ
25:03 thinging better than I like hearing you
25:04 s good Lord turn tune that guitar too
25:07 good Lord
25:09 okay um so here we are we're we're at
25:12 chat GPT um in order to make a GPT you
25:16 have to be using chat GPT 4 so so chat
25:19 GPT if you haven't played with it is
25:21 just you know you can ask it to do all
25:22 sorts of stuff and it just does
25:26 it um
25:28 make me a 7-Day meal plan that features
25:32 gummy bears KitKat bars and
25:38 [Laughter]
25:45 avocados I'll incorporate these items as
25:48 accents or treats along with healthier
25:52 items I I don't know what chat GPT is
25:56 speaking about how dare
25:59 she snack treat a small pack of gummy
26:02 bears
26:03 no you put the gummy bears in the baked
26:06 salmon I don't understand this anyway so
26:09 you do that uh we don't here we can stop
26:12 doing that um you can also say
26:16 um write me a country song
26:22 about walking down a dirt road in a
26:27 rainstorm
26:30 missing my
26:32 girl the
26:36 F-150 sure here's the country song Rain
26:40 on a dirt road just walking down this
26:42 old dirt road it knows what verses are
26:44 courses are it knows what songs are uh
26:46 then I could go um make a table
26:50 comparing Ford
26:52 F-150s and a Hot Wheels
26:57 car so you can have it oh wait hang on
27:03 stop below certainly here's a table so
27:06 and then it'll make you a table
27:07 comparing stuff like so it's like it it
27:10 is it's it's essentially impossible to
27:13 describe what this thing can do because
27:16 it can do so much right um it can even
27:19 do um you could say
27:23 um I want you
27:25 to write a Python program
27:32 that allows someone to design a custom
27:36 Hot Wheels car by selecting different
27:40 attributes and then designing the
27:42 components for
27:47 them so you can just talk [ __ ] into
27:50 existence okay so that's chat GPT can do
27:53 kind of anything so think of chat GPT
27:56 kind of like the operating system of
27:57 your computer or your phone gpts are
28:01 customized personalized versions of chat
28:04 GPT so you could make a GPT that helps
28:07 you do programming you could also make a
28:09 GPT that helps you make drinks my very
28:12 first
28:14 GPT is it's one it's it's one it's one
28:16 of the ones I'm the most happy with
28:18 because it's just it's so simple and
28:21 clean it's really really good it is
28:24 called where is it there it is sin
28:28 sipper and it it came up with the name
28:31 by the way it it it gave me a couple of
28:33 shitty names I get I said give me 20
28:35 good names and it came up with Sin City
28:36 sipper I'm like oh that's good so it's a
28:39 Vegas bartender with a Twist on the
28:41 classics and so I could say
28:46 um we just rang in the new year and it's
28:49 a little chilly here in Denver
28:53 period the Rodeo's in town and I'd like
28:56 you to come up with a nice custom
28:59 cocktail for the Cowboys trying to keep
29:02 warm by the
29:06 fire and it starts making an image so it
29:10 it will either uh give you the drink
29:13 recipe first and then make an image and
29:14 then sometimes it makes the image first
29:16 and then gives you the drink recipe but
29:18 what's remarkable about this the the
29:20 drink images it creates are [ __ ]
29:23 amazing like they're so accurate you
29:25 could you could literally put them in a
29:28 um oh that's so good look at that like
29:31 it it took in all of the you know the
29:33 cowboy stuff the fire it's it's a hot
29:36 drink well that's interesting it's a hot
29:39 drink with ice
29:41 cubes all right partner let's russle up
29:44 a warm cocktail for those cowboys to
29:46 Cozy up this is the Denver Rodeo warmer
29:49 2 O of bourbon it's a solid
29:51 start half oz of maple syrup a touch for
29:54 the wild west 4 oz of hot water like a
29:57 hot spring in the mountains a cinnamon
29:59 stick and a Twist orange perfect that's
30:01 perfect perfect cocktail um it sort of
30:04 screwed up the putting the ice cubes in
30:06 there but that's okay um so so a GPT is
30:10 just a specialized sort of mini version
30:13 of the bigger GPT but you have access to
30:16 all of what chat GPT is including Dolly
30:19 image generation including it's
30:21 connected to the internet including code
30:24 interpreter okay so how do you make one
30:27 of these bad boys it is stupidly simple
30:30 to to
30:31 start
30:34 um Shane steart I'm Shane and I'm an AI
30:37 addict hi Shane it's been three minutes
30:41 since I made a dolly image yeah we've
30:43 been there Shane welcome welcome welcome
30:47 yeah yeah oh sorry I wasn't listening I
30:49 was making something on chat J I was
30:52 making something on CH you can make M
30:55 with okay um so how you make these
31:00 things where you find them is in the
31:01 left hand side if you've got chat GPT 4
31:05 on um there are these little round
31:07 circles below the main chat GPT um open
31:10 ai's made some if you've made any
31:12 they'll show up in the list if you
31:13 played with any they'll show up in the
31:15 list and then there's this explore
31:17 button and that's where you can go find
31:19 some more gpts to play with including
31:22 the ones by open AI which some of those
31:23 are pretty good and then at the very top
31:27 create a GPT you just click on that and
31:31 and it's it's just stupidly simple so
31:34 you've got um create and configure mode
31:38 on the left so this is where you build
31:39 it and build is very Loosely defined
31:43 this is where you make it it's where you
31:44 create it and over here is where you
31:47 play with it so this is the test they
31:49 call this the playground or preview and
31:52 then in the upper right hand corner is
31:53 how you save it so I'm going to go back
31:55 to this document that I showed you
31:57 before called The Prompt starter so this
32:00 was the sheet that we used when we did
32:02 the GPT for good
32:04 challenge and so in step one you just
32:07 copy it says copy the text below so you
32:10 copy this GPT is designed to do this and
32:13 then you go to your GPT and in the
32:16 create mode don't go to configure mode
32:17 yet we're going to start in create mode
32:19 you you pop in that
32:21 prompt and then you put in the org name
32:24 I got to find an org name let's go find
32:30 let's go find a nonprofit
32:34 um
32:38 nonprofit about
32:45 clowns caring clowns International
32:52 see yeah this is
32:56 good video about us do they have a
32:58 donate
33:00 page
33:03 donate yeah PayPal
33:07 great so I'm just going to copy their
33:10 URL what's it called caring clowns
33:13 International so go back to our chat GPT
33:16 I'm going to SW swap out the org name
33:18 for caring
33:21 [Music]
33:25 clowns International
33:29 and then where it says put the URL in
33:32 I'm going to put the URL in
33:34 paste and that's it and I'm going to say
33:37 submit and now watch it says GPT Builder
33:40 updating GPT now watch the thing on the
33:45 right so it's updating it updating it
33:49 it's almost
33:50 done all right it's done bang this is
33:55 now
33:56 functional so tell me about caring
33:58 clowns International so I click on that
34:01 I don't even have to type anything doing
34:03 research with Bing so it's going out to
34:04 the internet going to the caring clowns
34:06 International
34:08 website and this is now a functional
34:13 GPT all
34:15 right unique and compassionate nonprofit
34:18 founded in 2001 after a clowning unit
34:21 visit to Vietnam of four individuals
34:23 including the US Navy Air Force veterans
34:26 organized to inspire focus on using
34:28 clowning as a means to improve the lives
34:30 of poor children worldwide
34:32 awesome and then that you just follow
34:35 those instructions in that Google sheet
34:37 and and you just keep improving your
34:39 your thing making it better and better
34:40 you'll make an icon for it things like
34:42 that okay um now over in the configure
34:47 side there's a place for an image which
34:49 it will make you an image just follow
34:51 those instructions it will name it for
34:53 you you can tell it you want a different
34:54 name or you can give it your own name
34:56 it'll it'll create a description here's
34:57 the prompt that makes this thing work it
34:59 wrote that for for us it wrote all these
35:02 prompting questions you can replace
35:03 those if you want to and then there's
35:05 also
35:06 this knowledge see where it says upload
35:09 files right here you can not only use a
35:13 website as a source of knowledge for
35:14 this but if you've got PDFs of a
35:17 particular disease or of a way you do
35:20 something or of data or whatever it
35:23 might be you can upload that content
35:24 into your GPT so you can make these G
35:27 pts very very specialized
35:31 to the the project you want to do there
35:34 it's it it is staggeringly
35:36 powerful and to put it in context if you
35:39 wanted to do this even two months ago
35:43 you had to use this thing called Lang
35:45 chain where you had to use command line
35:48 interface stuff and you had to do Python
35:51 Programming and you had to stitch
35:53 together all these services and you had
35:55 to fire up a vector database and you had
35:57 to take your data and you put it into
36:00 this thing and you had open AI do what's
36:02 called embedding it and then you take
36:04 those embeddings and you put them in the
36:06 vector database and then you do calls to
36:08 the vector database and you do calls to
36:10 other it was a
36:11 nightmare now Lang Chain's super
36:14 powerful don't get me wrong if you're a
36:16 geek go play with Lang chain go make
36:17 some cool
36:19 [ __ ] but you now as as a
36:23 non-programmer have the ability to to do
36:26 essentially the same thing it's quite
36:28 remarkable and then there's these
36:29 capabilities you can turn them on or off
36:32 web browsing dolly3 and code interpreter
36:35 um okay so now let's go look at code
36:39 interpreter and what that's all about
36:41 because that's a whole another [ __ ]
36:50 World
36:51 [Music]
36:52 um all right so we're back at chat JPay
36:59 so I'll describe code interpreter first
37:01 they also call it Advanced Data analysis
37:03 that's what it does is Advanced Data
37:07 analysis what code interpreter does is
37:10 it takes a file that you upload or files
37:13 yeah I think you can upload up to
37:16 five and it creates what's called a
37:19 container So within your chat
37:22 session a container is like a little
37:24 mini version of a computer inside your
37:29 computer how many are here how many here
37:32 use chat gp4 okay so there's a a uh
37:38 survey
37:40 up I got my 20
37:45 bucks all right how do I turn this
37:49 on how do I turn this on
37:54 people it's I don't see the thing oh
37:56 here we here we
37:58 go okay we got two minutes so go vote in
38:02 the uh go vote in the survey so in the
38:05 upper leftand corner right up there is a
38:07 little survey icon click on
38:10 that and it's asking out of the 130
38:12 people here how many people have TW H
38:15 have chat GPT plus which is the 20 bucks
38:18 a month
38:20 subscription oh Kyle what's what's the
38:23 action button okay so let me let me go
38:24 back to to uh one of my G pts here for a
38:29 second okay
38:32 so so I talked about uh web browsing
38:35 dolly3 and code interpreter you can turn
38:38 those on or off then there's this thing
38:40 called actions and if you say create a
38:42 new action um this is basically
38:45 incorporating apis into your GPT now
38:49 this is the reason I really don't
38:51 mention it if if you're just beginning
38:52 is unless you're a programmer and
38:55 understand what apis are this is a level
38:58 geekier than all the other [ __ ] all the
39:00 other [ __ ] code interpreter gets there a
39:02 little bit but but it's it's a mostly
39:05 automated service you just need to know
39:06 it's there um connecting an API actually
39:09 requires you to to create an API schema
39:13 now one really cool thing that they
39:16 did is how how do we do this
39:20 WEA where did they add
39:25 it
39:27 import from URL
39:30 no
39:33 um they can't they change the interface
39:36 all the time it drives me
39:38 crazy so I don't see it anymore they had
39:41 a thing in
39:42 here
39:44 where they had like an assistant like a
39:47 GPT assistant that would help you write
39:49 your thing it might still be in there
39:50 anyway what what an action is is so an
39:54 API is what's called an application
39:57 programming
39:58 interface and let's say if you wanted to
40:00 incorporate your GPT with Google Maps
40:04 and you wanted to pull in the address of
40:06 you know you would allow someone to
40:07 search for something and you wanted to
40:09 pull in the address for a location and
40:12 drop a pin on a map you could do that if
40:15 you had access to the Google
40:17 API and so basically this is where
40:20 you add that connection right so that's
40:23 that's what these things are okay so 20
40:26 we got 20 four votes in the survey
40:29 83.3% of the people in here um have chat
40:33 GPT
40:34 4 and 16% do the free chat GPT 3.5 so if
40:40 you want to make gpts like we're talking
40:42 about if you want to use code
40:43 interpreter like I'm about to show you
40:45 those are not available in GPT
40:48 3.5 um so cool cool survey I don't know
40:52 who put that survey up but I like it I
40:54 like it a
40:55 lot okay
40:57 okay I'm gone to play with Lang chain oh
41:00 Joker you're just going to go play with
41:02 your chain I know what's going on over
41:06 there I know what's
41:09 happening it's on I programmed my
41:12 gpts with PSE suda Fang and yaml made
41:15 them much better oh that's
41:17 awesome pseudo pseudo Fang pseudo Lang
41:20 pseudo Lang um very cool very very
41:24 cool um I have it okay all right
41:30 so I use actions to connect it with my
41:32 chat Bots and websites not hard no it's
41:35 not hard it's just it's just you you
41:38 listen adding an action in a GPT if
41:41 you're a programmer is
41:42 Trivial adding an action in the GPT if
41:45 you don't know what an API is is not
41:49 trivial don't forget mods can't vote
41:52 okay sorry mods tough on you but that
41:57 that does mean that cu the mods the mods
41:59 in here likely have chat gp4 um okay so
42:03 let's
42:11 go chat GPT can create the API itself
42:14 yep
42:16 exactly so listen if you if you like
42:18 geeking out Go Geek out on actions
42:20 actions are very powerful here here's
42:22 here's the
42:23 deal even without actions the fact that
42:26 the can do web browsing the fact that
42:28 this has got full access to
42:30 gp4 the fact that it can do Dolly image
42:34 generation the fact that it can do code
42:36 interpreter within your
42:39 GPT means that you can be build an
42:43 incredibly powerful
42:45 application without knowing any
42:49 programming like
42:52 zero and that's what that's what the
42:54 event this past weekend was about the GP
42:56 PT for good event assumed you didn't
42:58 know anything about programming and you
43:01 just want to make a powerful tool for a
43:05 nonprofit and that's what we did it is
43:08 it is a remarkable tool now is it
43:11 perfect no is it janky as [ __ ] yes just
43:14 crazy um how janky it is but that's okay
43:18 all right let's go use code interpreter
43:20 um so what I was
43:22 saying can you train 3.5 with new
43:25 articles you can
43:28 not um and in fact if I go to so so look
43:32 right here see that little paper clip
43:34 I'm going to switch to chat GPT 3.5 that
43:37 paperclip goes away I go to chat GPT 4
43:40 that paper clip comes back you're paying
43:42 20 bucks for a paper clip it's like
43:44 clippy only this one
43:49 works and it's not annoying
43:51 okay
43:53 so what what Advanced Data analysis or
43:56 code interpreter does is it creates this
43:59 thing called a container which is like a
44:01 little mini Standalone computer inside
44:05 your chat session and that little
44:07 computer has access to chat GPT to gp4
44:12 as well as a bunch of python libraries
44:15 so it's programming language and so what
44:18 you do is you upload data you upload
44:21 files to this thing and then you ask it
44:24 to do [ __ ] and it right and executes its
44:28 own python code to anal analyze those
44:31 files okay that's what's going to happen
44:33 here and it's it's just wild so what
44:36 we're going to do is we're going to go
44:37 off to a different website called
44:42 kaggle
44:44 kaggle don't forget to do your kaggle
44:48 exercises it's not kaggle it's kaggle I
44:52 know but it's still funny if you're
44:54 sophomoric and your humor t bastes um
44:58 okay K
45:00 a.com and then you go to data
45:03 sets and so what are data sets data sets
45:06 are just that there are data that people
45:08 compiled and uploaded to this thing so
45:11 we can go play with them so let's see
45:14 the song of the world global jukebox
45:18 household electricity
45:23 consumption here's a 9 megabyte online
45:26 retail transaction data let's grab that
45:28 one this is a 9
45:32 megabyte something I guess CSV
45:37 files so it's called archive 6 all right
45:39 so we're going to go back to chat
45:41 J and we're going to click on our little
45:45 paperclip and we're going to go find
45:49 where did I do that downloads yeah all
45:53 right so I uploaded this archive 6. zip
45:57 so we don't even know what's in it right
45:59 we don't know what's in it so I'm going
46:01 to
46:03 say um what files are in this and tell
46:10 me
46:12 what I should know about
46:17 them okay not exactly programming just
46:21 what is this boss so if you were work
46:23 let's say you were working at a company
46:26 and someone forwarded you a a a a zip
46:30 file and they said hey the data's in
46:33 here and that's all they told you but
46:36 you trusted them you knew who they were
46:37 and so you're you go to your data
46:40 analyst in the company and you're like
46:41 hey I got I got a zip file what is this
46:44 this is essentially what you'd ask ask
46:46 them right all right so bang now it says
46:49 analyzing look at what this is
46:52 doing it imported zip file like library
46:56 and Os Library so it imported these two
46:59 libraries to be able to unzip the file
47:02 it then unzips the file it says on
47:03 online retail.
47:05 CSV so it says here's what you should
47:09 know about CSV files like this one so it
47:11 gives me some general things about the C
47:13 if you want specific information I'm
47:14 going to go
47:16 yes I want specific information I just
47:19 said yes great to assist you further I
47:22 need to know the specific
47:24 information that's in
47:29 here wait let's
47:32 see um I'm just going to go please
47:38 analyze the data and tell me the
47:44 trends
47:46 you think might be
47:51 interesting analyzing so now what's it
47:53 doing now it opened the panda his
47:56 Library so it can actually read this CSV
47:58 file right and it's right it's folks are
48:03 you seeing what's happening here it is
48:05 writing and executing its own python
48:08 code it is prompting
48:10 itself code interpreter is a glimpse
48:13 into what's coming in 2024 a thing
48:16 called autonomous
48:17 agents so right now with chat
48:20 GPT you're the instigator right you're
48:24 the one that says I want you to do this
48:26 and then it answers you and you're like
48:27 no I didn't like that I want you to do
48:29 it different and then it gives you an
48:30 answer and then you're like you're a
48:32 poopy head and I think you do crappy
48:34 work and it's like I'm sorry I'll try to
48:36 do better work how's this and you're
48:37 like that's not bad and back and forth
48:39 and back and forth right what happens
48:41 with autonomous agents is you just give
48:44 them a goal and they go and that's what
48:46 this thing is doing right
48:49 so so so it it did some analysis and
48:53 then it gave itself an error it said it
48:56 appears there was an
48:57 issue to proceed I can try a different
49:01 approach now it's been successfully
49:03 loaded there are
49:06 54199 entries so this is a 500 plus
49:10 500,000 plus
49:13 row CSV
49:15 file given this information I will now
49:18 analyze the
49:19 data and and it says so what are the
49:22 columns index invoice number stock code
49:24 description quti quantity invoice date
49:28 unit price customer ID and Country and
49:31 so and so given this information I will
49:33 now analyze the data to identify
49:35 interesting Trends this will include um
49:38 sales Trends over time so this is sales
49:40 Trends over time from 2011 for a year
49:45 and it looks like the data ends at the
49:47 end there um but not surprisingly sales
49:51 go up toward
49:55 Christmas
49:57 so it wrote the code to make this
50:03 chart and then you know and then it goes
50:05 on uh let's see
50:09 um pick your favorite three Trends and
50:12 visualize each of them for me in a
50:15 unique and Visually compelling
50:20 way now I don't know anything about data
50:25 analysis and I don't know anything about
50:27 this particular data set imagine if you
50:31 had data from your
50:32 business and you knew a little bit about
50:35 what you might want to do with that
50:37 data you can now just casually ask it to
50:43 do the [ __ ] you need it to do this will
50:45 clean up data um it will reorganize data
50:51 um one of the things code interpreter
50:52 does is it will create files
50:57 so yeah so there's so it did it did a
51:00 bar chart about something let's see
51:02 that's sales
51:04 trends products by quantity
51:08 sold World War III
51:12 gliders or World War II gliders jumbo
51:15 big
51:17 red Retro spot I don't know what that is
51:21 popcorn
51:24 holder I don't know what the [ __ ] retail
51:26 data this was
51:29 but I mean it's
51:33 remarkable and then I could ask this I
51:35 could say I could say okay give me um I
51:39 only want the data for the popcorn
51:42 holder or or I want you to I want you to
51:44 create a new Excel file where you split
51:48 out the data into separate worksheets by
51:51 product so I just want data for the
51:53 popcorn holder on one worksheet and for
51:56 the whatever the [ __ ] the other thing
51:57 was in another worksheet it will create
52:00 that Excel file and give it to you to
52:03 download it's
52:05 amazing Joseph Baker I would not trust
52:08 this for sensitive data Joseph Baker is
52:10 absolutely correct you should not put
52:12 things in here that if it got out it
52:14 would destroy your life or your business
52:17 just assume that none of this [ __ ] is
52:18 secure until we hear otherwise and if
52:21 you're like but Kyle why would I even
52:23 use it at all
52:24 because
52:27 one of the things that you'll hear me
52:28 repeat over and over and over to the
52:31 point of annoying in in this channel is
52:33 all of these tools right now all of
52:35 these tools including the most
52:37 remarkable ones are janky pieces of
52:40 [ __ ] this is all very very new
52:43 technology and you're like but Kyle ai's
52:45 been around for decades AI is and all
52:47 the software we use yeah yeah yeah ai's
52:49 been around for decades generative AI
52:53 starting with chat GPT starting November
52:55 30th 2022 is is effectively a new
52:59 category of AI think think of chat GPT
53:04 as the consumer layer that gives regular
53:07 people access to the power of machine
53:10 learning and artificial
53:12 intelligence that's only been around for
53:14 a
53:15 year um you could argue it's been around
53:18 a little longer because of gpt3 but
53:20 whatever but just just ostensibly
53:22 November 30th 2022 is is the the flag in
53:26 the ground of the inflection point that
53:29 everything before it you had to be a
53:32 data scientist and mathematician and
53:34 engineer to be a part of and anything
53:37 after that you can be a regular human
53:39 being and do all this
53:41 stuff so all of these tools are very
53:44 very new and very very raw do not
53:46 mistake that for their a waste of time
53:51 any time you put in right now learning
53:53 these things is going to serve you very
53:55 very well it's one one of my sayings in
53:57 here dick around with AI it will serve
54:00 you just dick around with it just play
54:02 with it learn what code interpreter is
54:04 learn what it does good learn what it
54:05 doesn't do
54:07 good just learn it learn it learn it
54:09 learn it learn it learn it learn it
54:11 learn it learn it learn it learn it
54:12 learn it if it cleans up Dana will clean
54:14 my house getting tired of infinite
54:18 laundry we're we're I I here's what's
54:21 really [ __ ] up we're not that far from
54:25 that reality right they're already
54:26 putting these large language models into
54:29 robots you've got Tesla that's trying to
54:31 build a $20,000 humanoid robot you have
54:34 like five other companies trying to do
54:35 that right now and they're all making
54:37 progress like the Tesla bot 2.0 just
54:41 doubled its walking or increased its
54:42 walking speed by 30% and now it can
54:44 dance and move its head and its neck
54:48 um I saw one today that there there's
54:51 some there's some robot that rolls
54:53 around it looks like on it's on like a
54:56 like a bar cart rolling around but it
54:59 was cooking
55:01 shrimp calling an elevator it was it was
55:03 like they gave it like five or six tasks
55:07 and it was doing them like 100% of the
55:09 time without failing so we're like
55:14 we're we we are shockingly close to the
55:20 Jetson we just need our [ __ ] flying
55:22 cars now and we're close to that too
55:24 like the drones have you seen these new
55:26 drone shows with like you know 1,500 or
55:30 10,000 drones like it's it's starting to
55:33 get better than fireworks displays it's
55:37 amazing so anyway all
55:44 right good Lord people I know it's
55:47 called the AI learning lab but I can't
55:48 be teaching like this all the time I'm
55:50 supposed to be in here acting like an
55:53 idiot go watch the video of the last
55:55 last the last one what was it yeah the
55:58 one I did yesterday was that the one I
55:59 don't know there was one of I just acted
56:01 like an idiot for an hour and a half two
56:03 hours which that's most nights but it
56:06 was particularly dumb dumb
56:10 territory oh good lord okay I want to
56:13 tell you about one other thing so before
56:17 I told you to go here I to to the salon.
56:20 and and that's where all of that
56:22 information about building a GPT was
56:26 but the AI Salon if you go back to that
56:31 link
56:32 tree the fourth link is the AI Salon
56:35 Community click on that
56:37 link to the AI Salon website the the
56:41 mighty networks
56:43 community and the first thing I want you
56:45 to do is go to I think it drops you into
56:47 this link welcome to the
56:49 salon and this is going to tell you
56:52 about the AI salon and and hang on let
56:56 me get rid of that I had light going in
56:59 my lens which was making it hard to read
57:02 um and so this will tell you what the AI
57:04 Salon is a community of extraordinary
57:07 adventurers exploring the AI unknown
57:11 um here's who we are the AI Salon is a
57:13 ragtag group of Misfits neru well Wells
57:16 and miscreant who have gathered to
57:18 explore the uncharted waters of
57:20 generative AI where every everyone from
57:23 entrepreneurs to Educators for
57:25 healthcare workers to creatives and
57:27 everyone in
57:28 between
57:29 um if you're if you're the weirdo in
57:33 your
57:36 group that is obsessed with AI and you
57:38 can't understand why no one in your life
57:41 gets it this is your group this is your
57:44 group of weirdos or Irregulars as we
57:47 call them
57:49 here
57:51 um the the reason it's hard to find
57:54 people that get this is we're so early
57:57 McKenzie just did a study just came out
57:59 about two months ago that at the time of
58:02 the study 82% of of people in the US
58:06 have not used chat GPT 82% 82% that
58:10 means 18% of people have that means that
58:16 maybe I don't know 40% of them have used
58:20 it more than once it probably means that
58:22 half of that group has chat GPT
58:27 plus so we're talking like singled digit
58:32 percentages of people that use chat
58:37 GPT
58:39 right so this is very very new so this
58:42 group is for those people that have had
58:46 their Kevin McAllister moment this
58:50 one right your home alone moment of like
58:53 holy [ __ ] it can do
58:58 that and this is a community I started
59:01 this community the week after chat GPT
59:04 came out because when I saw what chat
59:05 GPT was I had an epiphany that oh [ __ ]
59:09 chat GPT does for AI what the worldwide
59:13 web did for the internet it made it
59:16 accessible it made it accessible by
59:19 non-programmers and non-engineers and
59:21 non researchers and
59:23 scientists it made it available to
59:28 humanity chat GPT did the same thing and
59:31 so this group is people
59:36 that are curious about it and and are on
59:40 the adventure and are choosing to be on
59:41 the adventure and if you're on this
59:43 adventure you know when you have that
59:46 moment that Kevin McAllister moment of
59:48 holy [ __ ] what just happened what did
59:50 that thing just do you can't unsee it
59:55 the other thing that's frustrating about
59:57 AI right now is you can't explain it to
59:59 someone it sounds like we're in a
1:00:01 [ __ ]
1:00:02 cult but you got like you you just type
1:00:05 in it and it makes a code computer code
1:00:09 in the it and it there's a picture it's
1:00:12 I made a picture of a llama with a hat
1:00:15 and people are like what are you talking
1:00:18 about you [ __ ] weirdo what what have
1:00:21 you done with your life you're like but
1:00:23 it's cool I I spent hours last
1:00:26 night making poems about the
1:00:29 Mesopotamian War why what it's so do you
1:00:33 not get how cool it is no they
1:00:36 don't they
1:00:38 don't they
1:00:41 don't but this group does and so and so
1:00:45 read about who we are and why we do what
1:00:48 we do but also look at our values so the
1:00:50 values are really important to
1:00:52 me these values are not just [ __ ] we
1:00:55 made up these are not successories
1:00:57 posters that we just you know pulled off
1:00:59 the successories website what we did
1:01:02 after about six months of this group
1:01:04 being together we said hey like what are
1:01:06 the attributes of this group that if if
1:01:08 it
1:01:09 grows we don't want to lose and so the
1:01:12 values here are who this group already
1:01:15 is and the reason I talk about these all
1:01:17 the time if you've been here before
1:01:18 you're go like oh here goes Kyle on the
1:01:20 [ __ ] values
1:01:22 yep why because values are really
1:01:25 important what's going to happen is this
1:01:28 as I go through the values and as you
1:01:30 read the values when you go to the
1:01:31 website and you read the values of the
1:01:33 AI Salon one of two things is going to
1:01:35 happen you're going to read the values
1:01:36 and it's going to feel refreshing you're
1:01:38 like oh my God this sounds like my
1:01:39 people that's one response the other
1:01:42 response is you're going to read these
1:01:43 values and you're like who the [ __ ] are
1:01:45 these people these are not your people
1:01:48 and either of those is
1:01:50 okay but don't join the group if the
1:01:53 values don't resonate cuz this is who's
1:01:54 here here so curiosity is one of the
1:01:57 values right generosity what's that mean
1:02:01 well if you learn how to do something
1:02:03 really cool with a prompt tell people
1:02:05 about it share it well yeah but that's
1:02:07 my prompt that's my proprietary [ __ ] it
1:02:09 [ __ ]
1:02:11 that share
1:02:13 it because right now no one has this
1:02:17 figured out no one if you think you're
1:02:19 really far behind you're
1:02:23 not you're not at all we're so early but
1:02:27 but you can't go to school for this you
1:02:29 can go to school for AI and machine
1:02:30 learning you can't go to school for
1:02:32 whatever the [ __ ] generative AI is about
1:02:34 to become it we don't know yet the tools
1:02:37 are just being formed we're in the
1:02:39 primordial soup
1:02:41 stage just like in the mid 90s with the
1:02:43 worldwide web nobody really knew what a
1:02:45 website was and what you could do with
1:02:46 them and what you couldn't do with them
1:02:48 that's the stage we're at
1:02:49 here so generosity share what you learn
1:02:52 um exploration we want people that are
1:02:55 actively on this adventure um
1:02:57 collaboration I I want people to like
1:03:00 team up and pair up and start companies
1:03:02 and start projects and just do stupid
1:03:04 [ __ ]
1:03:06 together um
1:03:09 inclusion AI generative AI is going to
1:03:12 impact every single profession that
1:03:16 touches a computer and probably a lot of
1:03:18 them that don't and if you think about
1:03:20 where we are in society to today what
1:03:23 profession doesn't touch a computer not
1:03:26 well a plumber no no sorry the plumber
1:03:30 has a phone that probably has some
1:03:32 invoicing program on it and some
1:03:34 marketing thing to get him listed in the
1:03:38 whatever nearly every profession touches
1:03:42 computers anything that touches a
1:03:44 computer in the next 3 years is going to
1:03:47 be absolutely transformed by AI so
1:03:50 people are going to be coming from all
1:03:51 over all experience levels all
1:03:53 backgrounds and then the last one is
1:03:56 empathy there is so much fear and anger
1:03:59 and trepidation and and just ignorance
1:04:01 about what AI is how it works what what
1:04:04 it makes
1:04:05 possible that understand that and have
1:04:08 empathy for people so if someone comes
1:04:10 in and they're like well you're a dumb
1:04:11 dumb because you're doing Ai and that's
1:04:13 just stealing from artists and [ __ ] you
1:04:14 big
1:04:17 black totally understand how you feel
1:04:20 like that have you played with it well
1:04:22 no cuz I'm I'm taking a moral stand okay
1:04:26 awesome you want to see something cool
1:04:27 well not really oh that's cool it could
1:04:30 do that I didn't know it could do
1:04:32 that right meet people where they
1:04:35 are so that's our values all
1:04:38 right next thing I would request if you
1:04:41 like those values and you want to join
1:04:42 the community go to the community
1:04:44 section on the left here and there's a
1:04:47 little section called introduce yourself
1:04:49 take our second value of generosity and
1:04:52 introduce yourself tell us who you are
1:04:54 tell us tell us what you do tell us what
1:04:56 you're interested
1:04:57 in you know do you have six grandkids
1:05:00 and and you're retired and just thought
1:05:02 a I would be fun or have you've been in
1:05:04 machine learning for 30 years and you
1:05:06 write the algorithms that makes this
1:05:08 [ __ ] work I like it doesn't matter just
1:05:10 it doesn't matter where you are on the
1:05:11 on the journey just introduce yourself
1:05:14 and you'll find your people here and
1:05:15 then in the community section there's a
1:05:17 water cooler area where you can talk
1:05:19 there's a showand tell section which has
1:05:21 become kind of this real time you know
1:05:25 art Jam from this channel like as as I
1:05:28 talk people look people are building um
1:05:31 gpts and sharing them in here um they're
1:05:36 do no fairs okay this is an inside joke
1:05:40 we are trying right now to figure out
1:05:41 the official animal list for the AI uh
1:05:45 learning lab Zoo um Cindy
1:05:50 [ __ ] literally just she just forced
1:05:55 raccoons into the animal list we we we
1:05:59 tried to keep raccoons out and now look
1:06:01 at what's happening so now that raccoons
1:06:03 are in they're in going no ferrets you
1:06:05 see how this happens people they've
1:06:08 poisoned the
1:06:13 well we've got monkeys playing the game
1:06:16 of life with people on it no fets y are
1:06:21 just [ __ ] nuts here's the clown
1:06:23 what's this the clown someone had to do
1:06:29 it from the clown nonprofit I knew that
1:06:32 would get some action you you you don't
1:06:34 think I do this stuff accidentally right
1:06:37 I know you weirdos are going to be doing
1:06:40 something oh that's beautiful who did
1:06:43 that Stacy just me and the trees wow
1:06:46 that's just just just her and the trees
1:06:49 in Orange that's beautiful nice prompt
1:06:51 or nice whatever you did there I'm AI
1:06:55 who are
1:06:57 you that's cool too Papa JJ so anyway so
1:07:01 that's the that's the showand tell
1:07:03 channel so go showand tell go do some
1:07:04 new stuff there show off that's what
1:07:07 that's there for llama Joker's got these
1:07:11 anthropomorphized llamas down man these
1:07:13 things are so
1:07:16 cool just amazing look at that wow who
1:07:19 did that auto
1:07:21 76 very nice damn
1:07:27 damn Tim sheffel my first two portraits
1:07:30 with
1:07:32 mj6 it's it's getting stunningly good
1:07:35 isn't it dag all right and then the
1:07:39 other thing wait like and you're like
1:07:41 but Kyle there's more oh there's
1:07:45 more why is he so excited isn't this
1:07:48 just a website no it's a community full
1:07:51 of people I think we have like 600
1:07:53 members hang
1:07:55 on
1:08:00 members member list
1:08:04 610 this site's like a month old we have
1:08:07 610 members we were on Discord for a
1:08:10 while and just people didn't grock
1:08:12 Discord right so so we moved over to
1:08:14 Mighty networks because it's easier and
1:08:16 it's web- based and you don't have to
1:08:17 learn
1:08:19 Discord even even mid Journey's leaving
1:08:22 Discord um okay so the other thing we
1:08:25 have here are guilds so so there's like
1:08:29 an AI 101 Guild if you're just getting
1:08:31 started there's an Art Guild a writer
1:08:32 Guild business uh an AI mechanics Guild
1:08:35 talking about the geeky stuff there's
1:08:37 healthc care there's video um uh Stacy
1:08:41 and Matt Mr K here on Tik Tok and Stacy
1:08:45 just me and the trees uh just kicked off
1:08:47 it it was officially announced last
1:08:49 night the education Guild if you're an
1:08:52 educator and you're like the only
1:08:53 educator in your District it's like we
1:08:56 got to be doing this AI stuff this
1:08:57 stuff's amazing students need to be
1:08:59 learning this we got to be engaging with
1:09:00 them help help help SOS someone needs to
1:09:05 pay attention to this this is your
1:09:09 group Stacy and Matt are like passionate
1:09:12 about this
1:09:13 stuff passionate about it and so that's
1:09:17 so so the guilds or sub communities
1:09:19 where you can go in and learn specific
1:09:20 things but also connect with people who
1:09:22 give a [ __ ] about the stuff you give a
1:09:24 [ __ ] about all right enough enough about
1:09:27 the AI Salon but that's the AI Salon you
1:09:30 should go there you should be a part of
1:09:33 it all right and it doesn't cost
1:09:36 anything this is not a sales pitch this
1:09:38 is just a pitch like
1:09:40 like you want to find people that are up
1:09:43 to some [ __ ] in the
1:09:46 world do you feel hopelessly behind with
1:09:51 AI do you know what I can promise you
1:09:55 so does every single person in the guild
1:09:59 or in the salon in the AI Salon 610
1:10:03 people 610 out of 610 are like I don't
1:10:07 know what I'm
1:10:08 doing you you know what I hear regularly
1:10:10 in here is oh looks like we have a new a
1:10:13 new survey let's see what the survey is
1:10:16 um okay two two minutes um in the in the
1:10:20 upper left hand corner right there is
1:10:22 where I see it click on that icon it's a
1:10:25 it's a a
1:10:27 survey of how many people are using chat
1:10:30 gp4 I'm scared right now you shouldn't
1:10:32 be
1:10:33 scared well you can be scared no scared
1:10:36 is an appropriate thing actually you
1:10:38 know what scared is a really appropriate
1:10:40 thing what I find myself oscillating I
1:10:44 find myself oscillating with AI between
1:10:46 like I'm as excited as I've ever been
1:10:48 I'm as scared as I've ever been I'm I
1:10:51 feel completely optimistic and hopeful
1:10:53 oh my God everything's going to fall
1:10:55 apart um it it really does go like that
1:10:59 like it it you keep sort of peeling
1:11:02 these layers it there there's a couple
1:11:04 of things that happen one
1:11:07 is you start out completely clueless and
1:11:10 then like you have your Kevin McAllister
1:11:12 moment and then you and then you start
1:11:15 digging in and so you have your initial
1:11:18 kind of
1:11:19 Awakening is is the best way I can
1:11:22 describe it of what's possible and you
1:11:24 go from being scared and weirded out by
1:11:26 it
1:11:27 to you know the heavens part and you're
1:11:30 like you see the possibility of the
1:11:32 future and then it's it's in inspiring
1:11:35 and then you can't sleep for three days
1:11:36 and you're just running down one Rabbit
1:11:38 Hole after another after another so so
1:11:40 there's that phase then you kind of go
1:11:43 into the phase of like you've figured
1:11:45 some [ __ ] out and now you're like doing
1:11:47 stuff and you're like oh I could use
1:11:48 this for work and I could use this for
1:11:49 home and I could and then you're on
1:11:51 you're on sort of the efficiency the
1:11:54 efficiency you you know um challenge
1:11:57 where where you're just like I'm going
1:11:58 to fix this I'm going to fix that and
1:12:00 then and then you sort of realize oh but
1:12:02 these these tools are kind of janky
1:12:04 pieces of [ __ ] and they sort of work
1:12:06 they're they're mostly perfect except
1:12:08 for they suck a lot and then so they're
1:12:11 not perfect should I not use them am I
1:12:13 oh and there's security and privacy
1:12:15 issues oh God and then you get
1:12:17 completely overwhelmed okay the uh the
1:12:20 uh
1:12:21 survey um 75% of the people have chat
1:12:26 GPT plus that's a $20 subscription 16%
1:12:29 these numbers are roughly the same um
1:12:32 are on three they they use chat GPT but
1:12:34 they're on
1:12:36 3.5
1:12:38 um 5% say say what and 2.7% say they
1:12:43 don't know so if you if you don't know
1:12:46 what chat GPT is go to chat.
1:12:48 open.com right
1:12:51 now go there right now chat
1:12:54 .op ai.com and just start playing
1:12:59 just type words into the box and hit
1:13:02 submit write me a poem about wet puppies
1:13:07 click is it 20 bucks a month yes chat
1:13:10 GPT plus is 20 bucks a month I I am not
1:13:13 affiliated with them in any way I have
1:13:17 told Sam mman for months now he should
1:13:21 absolutely be paying me cuz this channel
1:13:23 is essentially a big [ __ ] ad for his
1:13:24 company
1:13:26 but he's jealous of me so he hasn't put
1:13:29 me on the payroll yet but that's okay um
1:13:31 so I'm not affiliated with it so I don't
1:13:33 care if you actually do this or not but
1:13:36 hang on
1:13:38 um the 20 bucks a month for chat GPT
1:13:42 plus if you can afford it I think it is
1:13:44 actually irresponsible if you want to
1:13:47 learn about AI if you want to get into
1:13:50 whatever's
1:13:51 coming I think it's irresponsible right
1:13:54 now to not be on chat GPT plus and
1:13:57 learning everything you can about what
1:13:58 that tool can do because it's about to
1:14:01 get a lot
1:14:02 better but in the past 3 months two and
1:14:05 a half three months it's gotten so good
1:14:09 and so
1:14:11 deep that it's not just one tool it's
1:14:13 like eight tools and and they all have
1:14:16 their different advantages now here's
1:14:18 the thing I would not pay the annual you
1:14:21 know how they give you the 20% off if
1:14:23 you buy the year I wouldn't buy the
1:14:26 year with any of these AI tools right
1:14:29 now I would go month to month you're
1:14:30 like but Kyle it will cost me
1:14:33 more I would go month to month why is
1:14:36 that
1:14:38 well Google is about to release Gemini
1:14:43 Ultra which is apparently a chat GPT
1:14:45 plus killer now we don't know if it's
1:14:47 going to cost money it might it probably
1:14:49 will um and it's going to be really good
1:14:52 is it going to be better than chat GPT
1:14:54 might be how are we going to know we're
1:14:57 going to play with it we're going to
1:14:58 play with it here we're going to kick
1:14:59 the tires on it and if it sucks we're
1:15:01 going to say it sucks and if it's
1:15:03 absolutely [ __ ] amazing and it's so
1:15:05 amazing that we're like chat GPT who
1:15:08 then we'll move on I I don't sense
1:15:10 that's going to be the case but I do
1:15:12 think that we're going to have a viable
1:15:14 chat gb4 alternative that I that I
1:15:17 actually do believe that's what I think
1:15:18 is
1:15:19 coming and then chat GPT is not sitting
1:15:22 on its laurels like even though they
1:15:24 fired Sam Alman and rehired him and they
1:15:26 sort of [ __ ] up their life for a month
1:15:28 or
1:15:29 so they're paying attention and if you
1:15:33 don't think they're boiling up some [ __ ]
1:15:36 at chat GPT like like we've got
1:15:38 autonomous agents coming we've got
1:15:41 additional multimodal things like audio
1:15:43 and video which aren't really in there
1:15:45 right
1:15:46 now we've got different interfaces
1:15:49 probably we've got different we've got
1:15:50 the GPT store that's not out yet so you
1:15:53 can make G P PTS but there's not a place
1:15:54 to distribute them that's
1:15:56 coming probably within the first two
1:15:59 weeks of
1:16:01 January so we're we're you're just in a
1:16:04 wild thing so I would absolutely
1:16:07 absolutely 100% recommend the 20 bucks a
1:16:10 month for chat GPT
1:16:12 plus all right let's see let's see what
1:16:15 people are saying here let's let's see
1:16:17 let's see how much I'm getting
1:16:21 trolled Kyle I might be a little salty
1:16:24 towards Sam ever since the snub listen
1:16:26 forever hooked I'm right there with you
1:16:27 I appreciate it you're a subscriber to
1:16:29 the channel I appreciate your input
1:16:34 listen empathy is is one of our values
1:16:36 in the AI salon and just because Sam is
1:16:39 is a small man who's jealous of me and
1:16:42 and won't give me access to the new
1:16:44 features before other people and and and
1:16:47 won't send me regular checks for the
1:16:49 fact that this channel is just an ad for
1:16:51 his company that's just on him I have
1:16:53 empathy for him he's going through a lot
1:16:56 they fired him they rehired him that's
1:16:58 got to hurt you know him and his buddy
1:17:01 Greg were
1:17:02 out and it was funny when that was going
1:17:05 on you know what he was talking about me
1:17:08 so sad it's like yeah whatever I don't
1:17:11 Kyle's just I mean he's got that channel
1:17:14 and whatever that that that video thing
1:17:17 is and you know he's a dumb dumb like it
1:17:20 it was just it was like
1:17:21 Petty I so I feel bad for Sam I really
1:17:25 do all
1:17:31 right welcome to chat
1:17:34 add delusions are
1:17:38 us I don't know Sam Elon I
1:17:45 know a business write off absolutely
1:17:47 could be a business right
1:17:49 off I think Kyle needs my sound therapy
1:17:53 file to help him sleep after the show I
1:17:56 I'll tell you what I I promise you when
1:17:58 I leave these being on like I am for
1:18:02 these things for three hours a night I
1:18:04 sleep like a [ __ ] baby like nothing
1:18:07 is happening every everything has been
1:18:10 drained out of my my skull at the end of
1:18:13 these things maybe Sam's an alien it's
1:18:15 possible it's you know what that's very
1:18:18 possible cuz he maybe he doesn't quite
1:18:20 understand the human race and and one of
1:18:22 the reasons that I bug him so much is
1:18:24 I'm just so human and he just doesn't
1:18:26 understand it that's a possibility
1:18:29 absolutely I've been focused on
1:18:32 advocating for senior community ID theft
1:18:35 fraud education don't know how yet
1:18:38 um holy [ __ ] r or die I think I think
1:18:41 that is so my my thought is keep doing
1:18:46 that keep figuring it out use chat GPT
1:18:48 to help you communicate it come up with
1:18:50 plans come up with anything you can do
1:18:53 um
1:18:54 2024 is going to be a [ __ ] show with
1:18:56 misinformation and fraud and and the
1:18:59 elections it's going to be [ __ ] nuts
1:19:01 um and I the senior community and and
1:19:04 listen don't don't be I know you're down
1:19:07 there right now going oh the senior
1:19:08 community you know about that old man
1:19:11 don't do that that's that's
1:19:14 Petty don't be like Sam be like Jimmy um
1:19:18 but I but I think I think educating the
1:19:20 senior community and and doing things
1:19:22 like like families absolutely right now
1:19:25 should have um like communication safe
1:19:31 words like like you should be having
1:19:33 conversations right now with your family
1:19:35 members hey if hey Mom Dad if I ever
1:19:39 call you and say I was in an accident
1:19:41 and that I need
1:19:44 money here's what we're going to
1:19:47 say how many llamas do you
1:19:50 own and if I don't answer I own
1:19:54 7 then that's not me right like that
1:19:58 should be happening right now but I I
1:20:00 think that's huge so I'm I think that's
1:20:02 great you're doing that what was your
1:20:05 site called the AI Salon it it is called
1:20:08 the the it's called the AI salon and the
1:20:10 site to get to the link is the
1:20:14 salon. and the fourth link down is the
1:20:17 AI Salon Community that's the one that
1:20:19 takes you to the mighty Network's site
1:20:23 and then the first three links are all
1:20:25 about the GPT for good challenge we did
1:20:27 this past
1:20:29 weekend where people that
1:20:31 participated created and shared 150 gpts
1:20:36 for nonprofits in a 24-hour period it
1:20:39 was really cool it was a very cool
1:20:42 event
1:20:46 uh okay so that wasn't the real Sam that
1:20:49 stopped in earlier now we we we have a
1:20:52 Sam impersonator that is a fan of mine
1:20:55 but if the real Sam stopped in I he'd
1:20:58 shut this down there there' have been
1:21:00 complaints to Tik Tock
1:21:07 management I actually sent um Greg
1:21:09 Brockman who's the president of open Ai
1:21:12 and Sam's buddy um he put out a he put
1:21:16 out a tweet today how has AI changed
1:21:18 your life and I responded with the GPT
1:21:21 for good challenge I said said I created
1:21:24 this community a week after chat GPT
1:21:26 came out and this community this past
1:21:28 weekend did 150
1:21:30 gpts um and I put Danielle's cookie GPT
1:21:33 for good picture in there and I sent it
1:21:37 to Greg Brockman so I hope he sees
1:21:47 it teaching my elderly to use two Factor
1:21:50 authentication apps like twio that's
1:21:52 awesome
1:21:54 although I'm so [ __ ] sick of two
1:21:56 Factor
1:21:58 authentication your 17 Factor
1:22:00 authentication needs to and then check
1:22:03 your mobile device and and like one of
1:22:05 the services I do says check your
1:22:07 Android device like I have I I have an
1:22:09 Android device I use as a testing device
1:22:13 that that I use maybe once a month and
1:22:15 like why is one of my services tied to
1:22:18 that I don't know but like I never have
1:22:22 it with me it's never charged it's it's
1:22:24 just have you built any of your own AIS
1:22:28 using your own if not but why I
1:22:31 have I've built the first AGI well James
1:22:35 Hill call Sam because Sam's got a whole
1:22:39 company trying to do that so if you've
1:22:41 built one you should you should let Sam
1:22:43 know that he's not going to be
1:22:45 first
1:22:47 um have I built any of my own AIS I have
1:22:50 not built my own model if that's what
1:22:51 you're asking I have used other people's
1:22:54 models I just built um I just built a
1:22:57 new tool for my um my company called the
1:23:01 authenticity engine which takes so my
1:23:03 company is storyvine and we do authentic
1:23:06 video storytelling so we have a platform
1:23:09 that automates video storytelling people
1:23:11 answer some questions in an app it goes
1:23:13 up to the cloud 5 minutes later you got
1:23:15 a fully edited
1:23:17 video so I just built this thing called
1:23:20 the authenticity engine where you can
1:23:22 take that video and now you can
1:23:24 automatically generate this whole world
1:23:25 of content around it so it basically
1:23:27 takes the story of the video and just
1:23:29 blows it out in all these different ways
1:23:31 it's super cool um so I built that and
1:23:34 then I'm building a number of products
1:23:37 around video translation and um bringing
1:23:42 personas to life in video and and things
1:23:45 like that so so yeah I'm building I'm
1:23:48 building all the time and and digging it
1:23:50 it took me what was interesting about
1:23:52 the the
1:23:53 building is
1:23:55 that I the actual
1:23:59 building that I've been working
1:24:01 on other than sort of prototyping and
1:24:04 some internal process [ __ ] I did with
1:24:06 zapier has only been in the past two
1:24:08 months it took me it took me about nine
1:24:12 months
1:24:14 of just
1:24:16 being insanely curious about AI for me
1:24:20 to get my head around it enough to
1:24:22 incorporated into my business in in a
1:24:25 like in an organic integrated way it
1:24:28 took it took me almost a year because
1:24:31 it's so different and it's like like a
1:24:34 lot of what I was learning was like well
1:24:36 do I do I just completely flush my
1:24:40 company of 11 and a half years down the
1:24:43 toilet and just do AI or do I do company
1:24:48 here AI company here or do I do a little
1:24:51 bit of integration or do I do full
1:24:53 integration what of this do I keep what
1:24:55 of this do I jettison like it's it's a
1:24:59 lot and I you know I've got a small
1:25:01 company shut
1:25:03 [Laughter]
1:25:06 up ah Mr K empathy be curious not
1:25:10 judgmental exactly someone comes in and
1:25:13 they're terrified about AI why what
1:25:16 scares you about it oh yeah I'd be
1:25:18 scared about that too you know that's
1:25:20 not reality right or I'd be scared about
1:25:23 that too yeah that's something we got to
1:25:25 deal with you want to see something cool
1:25:28 though
1:25:30 absolutely it's nuts now I work for
1:25:32 ARA they lose they they they lose life
1:25:36 savings because they don't know yeah
1:25:38 yeah WR or die I I like the the the
1:25:41 stories are tragic of all the fraud
1:25:44 that's going
1:25:45 on and man is it going to get worse
1:25:47 because because right now it's
1:25:50 relatively easy to clone someone's voice
1:25:52 and do that fraud well it's it's going
1:25:54 to be
1:25:57 it's the ability for frauders to scale
1:26:01 those
1:26:02 operations is going to get really
1:26:05 easy and then it's just a numbers game
1:26:09 right and they're going to get
1:26:10 increasingly sophisticated and they're
1:26:12 it's going to get increasingly hard to
1:26:14 determine if someone on the other end of
1:26:17 the call like they're going to figure
1:26:18 out how to spoof phone numbers and
1:26:21 voices and uh it's going to be a mess
1:26:24 lunar stick I've built a few gpts
1:26:27 specifically for seniors and the elderly
1:26:29 awesome you should check out lunis
1:26:31 stick's um gpts so if you go to go to
1:26:34 the
1:26:36 salon. go there the first link is to the
1:26:39 GPT directory um in the directory you're
1:26:42 Daisy Thomas in the directory right
1:26:43 lunatick hang
1:26:45 on I think you
1:26:51 are yeah Daisy
1:26:53 Thomas so go to the directory go go to
1:26:57 there and click on the first link and
1:27:00 that's going to take you
1:27:02 to this document that says GPT for good
1:27:05 at the top with the robot and then
1:27:07 scroll down and you'll see this
1:27:08 directory of these 150
1:27:10 gpts that were produced in this 24-hour
1:27:14 period and shared in this 24-hour period
1:27:16 and and she's got a bunch of them in
1:27:18 there so go check them out really
1:27:21 inspiring
1:27:25 how we doing time wise we're
1:27:27 good need a code word for family members
1:27:30 so they won't be fooled yeah exactly a
1:27:32 code word a challenge
1:27:37 question if I if I call up and ask you
1:27:40 for money what's the code word you can
1:27:43 make money with
1:27:46 time that should be
1:27:51 ours
1:27:53 you're free to take that by the
1:27:55 way if you want if you want to take that
1:27:57 as a code word for your family
1:28:00 please um earth.org extending
1:28:03 translation to the non-human domain
1:28:07 interesting what's your business my
1:28:09 business is called storyvine so story
1:28:12 vine.com we've been around for 11 and a
1:28:15 half years we we work primarily in the
1:28:18 healthcare and Pharma space but with
1:28:20 this with this new AI stuff with the
1:28:22 we're doing doing this real time
1:28:26 um real time content generation or not
1:28:29 real time but but sort of immediate
1:28:31 content generation and then we're also
1:28:33 doing translations and personas I can I
1:28:36 can show you that stuff I can show you
1:28:37 what we're doing um I I have a I have an
1:28:41 instinct we're going to expand beyond
1:28:43 the healthcare and Pharma
1:28:44 space um because of these new products
1:28:47 so I'm pretty excited about
1:28:49 that 3 second sample to have a
1:28:51 believable voice clone yeah I know it's
1:28:53 crazy isn't
1:28:55 it I made an amazing GPT for Potters oh
1:28:59 Cory Sandler Pottery I know that's so
1:29:01 cool that's really cool we can just play
1:29:04 the sound bite for the password yeah but
1:29:06 if they don't know the password Clint
1:29:09 Copperhead oh oh you could just play
1:29:11 this one you could make
1:29:13 Mone yeah that's true now that now that
1:29:16 now that I've I've given you like at
1:29:18 this point there's 6 months of
1:29:20 recordings of Me 3 hours long I think
1:29:22 you have plenty of data to spoof me if
1:29:24 you want
1:29:26 to we already have a Tik Tok scammer
1:29:28 that steals all my content and tries to
1:29:31 scam all my followers but Tik Tok
1:29:36 doesn't care about that but if I say the
1:29:38 word booby in the wrong
1:29:40 order I'm banned for
1:29:43 life ah Story
1:29:45 vine well yeah story story Vine with a
1:29:49 why any lawyers in the salon
1:29:55 um yeah we do we we I don't know if
1:29:58 they're lawyer we we've got some we've
1:30:01 got some policy people in there for
1:30:04 sure
1:30:06 um and I've had a couple of people reach
1:30:09 out and say they want to do a privacy
1:30:10 and policy Guild so um I I would think
1:30:13 we do yeah but I mean there's 600 people
1:30:15 in there so I I honestly don't know but
1:30:17 what I would what I would say Mr K is
1:30:19 that like that's something to ask in in
1:30:21 um in either ask for help or in uh water
1:30:25 cooler um just just or you could I think
1:30:28 one of the things you could also do Mr K
1:30:30 is you could do a uh you could do a
1:30:32 survey in
1:30:34 there Kyle my country is expecting to
1:30:37 bring thousands of new medical jobs oh
1:30:39 my
1:30:40 county jobs so bet Story vine would do
1:30:45 well oh we should we should talk Daisy
1:30:48 absolutely let's talk cuz actually you
1:30:50 know what Daisy cuz um so we've done a
1:30:53 lot of work in the nonprofit and the
1:30:55 advocacy space we tried to get into um
1:30:58 politics and there's there's some
1:31:02 conceptually story Vine's brilliant for
1:31:04 politics the reality of how politics
1:31:06 works and you know this better than most
1:31:09 um it it it just doesn't make sense for
1:31:11 a lot of it but I would love to show you
1:31:13 what storyvine is and what some of the
1:31:15 new stuff is and if you've got um
1:31:19 projects that that you think it would be
1:31:20 good for let's let's talk about that
1:31:23 um that could be really interesting U we
1:31:25 have a lot of laws in place to protect
1:31:27 seniors actually yeah I know but the the
1:31:30 problem with all the laws to protect
1:31:31 seniors is that the the scammers aren't
1:31:33 going to care right the scammers just
1:31:35 the scammers are in and out like the if
1:31:37 you look at the all the scams going on
1:31:39 with the nft stuff they're just
1:31:41 heartless bastards like they'll just
1:31:43 keep finding new cracks you know to
1:31:46 squeeze through and it's going to be the
1:31:49 same with the with phone scammers and
1:31:51 and things like that
1:31:54 so I mean [ __ ] like we're we're probably
1:31:57 even going to get to the point
1:32:00 where there's going to be like FaceTime
1:32:04 video calls from people you know that
1:32:06 are not
1:32:08 them it's
1:32:11 crazy kind of guy nice I can build you a
1:32:16 GPT lawyer real quick B Boo Boo
1:32:21 Boo
1:32:23 where is mid Journey moving to I visit
1:32:26 infrequently on Discord so so right
1:32:30 now if you've made more than 10,000
1:32:33 images um on Mid Journey you've got
1:32:37 Alpha access to if you go to alpha. Mid
1:32:40 journey.com it's their website and you
1:32:44 can make images in the website so
1:32:46 they've got a new web
1:32:48 interface um that's got all sorts of
1:32:50 it's it's decent it's like it's got
1:32:52 decent clickability and things like that
1:32:54 I don't have access to it yet because as
1:32:56 soon as Dolly 3 came out I stopped using
1:32:58 mid journey and that was a mistake cuz I
1:33:01 probably would have been at 10,000 if I
1:33:02 kept using it but I stopped so what are
1:33:05 you going to do FaceTime from deceased
1:33:07 family members well yeah I mean I mean
1:33:11 there there is there is some
1:33:15 um I mean there's sort of some some
1:33:18 creepy weird stuff that could get weird
1:33:21 and dark and destructive but I think
1:33:23 there's also some really positive stuff
1:33:25 with you you could either connect with
1:33:27 deceased members um my my buddy Richard
1:33:30 Boyd he's been in AI for like 10 years
1:33:33 when his father passed he took all of
1:33:36 his father's writings and none of it was
1:33:38 digitized he digitized like his father's
1:33:41 lifetime of writings and put it into I
1:33:44 think it was a gpt2 powered um thing and
1:33:47 he made a virtual version of his
1:33:50 father we we're in this in this year
1:33:54 we're going to you're going to be able
1:33:55 to have conversations with people that
1:33:58 aren't there in fact I just got access
1:34:00 to the did agent thing um if you were
1:34:04 here I don't know a week or no three
1:34:07 days ago I they told me I had access to
1:34:09 it and then they said I got there and I
1:34:12 didn't have access to it they just gave
1:34:13 me access to it but but you're going to
1:34:16 be able to talk to people that aren't
1:34:17 there so it could be deceased ones it
1:34:20 could also be where where I see
1:34:22 the the sort of virtual family members
1:34:25 or virtual loved ones conversations is
1:34:28 with dementia and Alzheimer's patients
1:34:31 that imagine them being able to talk to
1:34:34 a friendly face that they trust and know
1:34:36 and you know to whatever extent they
1:34:38 know them that is infinitely patient is
1:34:41 there
1:34:42 247 knows all the stories all the family
1:34:45 history those can be built relatively
1:34:49 trivially right
1:34:50 now and and it's going to get easier and
1:34:53 easier and easier to do that and I I
1:34:55 just that virtual
1:34:57 companionship as a category is going to
1:35:00 get really
1:35:02 interesting I mean if you haven't talked
1:35:04 with pi do you know do you know Pi are
1:35:08 you a piie person so this is my iPad you
1:35:12 can do this with both the chat GPT app
1:35:15 and this thing called PI from this
1:35:16 company called
1:35:18 inflection and it's just it's just a
1:35:20 chat GPT like thing but in the lower
1:35:23 corner here is that little telephone and
1:35:26 you can
1:35:30 just hey there I was wondering if we
1:35:32 could have a quick
1:35:36 [Music]
1:35:38 chat sure thing I'm here for you what
1:35:41 would you like to chat
1:35:42 about oh I don't know I'm talking to a
1:35:46 bunch of people about Ai and how virtual
1:35:49 companionship is going to be an
1:35:51 important thing in the future and I
1:35:53 think it could even help with
1:35:55 Alzheimer's patients with their family
1:36:00 [Music]
1:36:03 members absolutely that's a really great
1:36:06 Point virtual companionship has a lot of
1:36:07 potential to be helpful for people
1:36:10 especially those who may be socially
1:36:12 isolated or dealing with cognitive
1:36:14 impairment like Alzheimer's AI can
1:36:17 provide a level of companionship and
1:36:19 engagement that can be incredibly
1:36:20 beneficial whether it's through
1:36:22 conversation memory games or even just
1:36:25 providing a listening ear additionally
1:36:27 AI can be used to provide personalized
1:36:30 support and assistance to caregivers and
1:36:33 F and that's that's one of the things
1:36:35 that inflection the company behind this
1:36:38 is is working on is that they want this
1:36:41 to be the world's best personal
1:36:43 assistant so right now it's really good
1:36:46 at at
1:36:47 conversation um probably soon this year
1:36:51 it's it's going to start to get well
1:36:53 it's going to get better and better at
1:36:54 conversation but it's also going to
1:36:56 start to be able to take actions so you
1:36:58 could have a conversation and say hey I
1:36:59 want to send an email to so and so and
1:37:01 it'll just do it right if it has access
1:37:03 to your to your
1:37:05 tools um and that's that's where a lot
1:37:08 of these things are going to go is that
1:37:10 is that we're going to have agents we're
1:37:12 going to have assistants people that or
1:37:15 not people
1:37:17 AIS that do things on our behalf that we
1:37:19 can communicate with in natural language
1:37:22 and you know surprisingly real natural
1:37:26 language you don't have to talk weird
1:37:28 like with Syrian Alexa right now you
1:37:30 have to talk
1:37:33 weird and they just don't know a lot and
1:37:35 they get stumped a lot because they're
1:37:37 basically just logic
1:37:40 trees these these large language models
1:37:42 are not logic trees they're large
1:37:44 language models they they are modeled
1:37:46 after human brains which is why they
1:37:48 kind
1:37:49 of represent them in a certain way and
1:37:52 then the data that they were trained on
1:37:54 was the output of humans that was stuck
1:37:56 on the internet and so they kind of know
1:37:59 our language and they kind of know some
1:38:01 of the nuances of how we talk and what
1:38:03 we
1:38:03 do so these things are
1:38:08 uncannily not robotic even though they
1:38:13 are technically calculators they're
1:38:16 probability engines which is
1:38:19 amazing it's awesome I never go to
1:38:21 Discord anymore I know me too do you
1:38:23 think corsera is a viable way to
1:38:26 learn right now
1:38:30 um it depends what you want to
1:38:34 learn I would say there there's some
1:38:37 free courses by meta and Google
1:38:41 and who else did one I think I thought
1:38:44 there were three of them on generative
1:38:48 AI but but quite frankly any
1:38:52 like I the answer is
1:38:54 no if you want to learn if you want to
1:38:57 learn you know how to build models and
1:39:00 the mathematics of it and and the
1:39:03 foundations of machine learning and and
1:39:06 you know building AI systems all that
1:39:09 [ __ ] yeah there's courses for that but
1:39:12 if you want to take a course on
1:39:14 generative Ai and chat GPT and prompt
1:39:18 engineering and this and that I can tell
1:39:20 you right now The
1:39:22 Prompt engineering workshops you took 3
1:39:26 months
1:39:26 ago are out of
1:39:29 date the ones you took six weeks ago are
1:39:32 out of date when Gemini Ultra comes out
1:39:35 whatever was the best practice for open
1:39:37 AI is going to be thrown right out the
1:39:40 window like things are changing so fast
1:39:43 right now and the technolog is evolving
1:39:45 so fast that there literally can't be a
1:39:49 course that covers it all and then on
1:39:52 top of that so just within large
1:39:55 language
1:39:56 models I've got a recommended list of
1:39:59 things I think you should pay attention
1:40:00 to just in large language models there
1:40:03 are nine of them right now just in large
1:40:06 language models image generators I've
1:40:08 got five or six in there and that that
1:40:11 list is probably light video tools I've
1:40:15 got I think
1:40:18 five and that list is probably light
1:40:23 like there there is so and and and the
1:40:25 other thing you know like this is the AI
1:40:28 learning lab so someone will ask me
1:40:29 about some video tool they're like show
1:40:31 me how to use did or show me how to use
1:40:33 Runway and I'll and I'll go to the site
1:40:35 and I'm like okay here's how you use it
1:40:36 and I'm like nope don't know how to use
1:40:39 it anymore they completely changed the
1:40:40 interface like it's got new features
1:40:43 there's a there's an image tool right
1:40:44 now called Leonardo that um I knew how
1:40:48 to use Leonardo and now I go to Leonardo
1:40:50 and there's four five new tools that are
1:40:52 there that are completely different
1:40:54 they're completely it's like completely
1:40:57 new applications that do completely
1:40:59 different things than make images so
1:41:01 what's the best best way to learn it
1:41:03 start with chat
1:41:05 GPT get the 20 Buck a month chat GPT
1:41:08 Plus subscription learn all of the
1:41:11 different features within chat GPT start
1:41:15 there
1:41:18 seriously then start to expand out and
1:41:21 say what are the what are the other
1:41:23 large language models that I might want
1:41:24 to learn well you've got Microsoft
1:41:27 co-pilot which is their version of chat
1:41:29 GPT you've got um Claude from anthropic
1:41:33 you've got Pi you've got Google bard
1:41:37 you've you've got a bunch of them you
1:41:38 got the whole open- Source community so
1:41:40 you can sort of learn the world around
1:41:43 large language models chat GPT right now
1:41:45 still King of the
1:41:48 Hill and then you know within chat GPT
1:41:51 you've got Dolly 3 so start learning the
1:41:53 image Generation stuff and so Dolly 3 is
1:41:56 already built into chat GPT you can also
1:41:58 do that at Microsoft then you have mid
1:42:00 Journey why would you use mid Journey
1:42:02 because the image quality is higher then
1:42:04 you've got Leonardo why would you use
1:42:06 that because it's built on top of stable
1:42:08 diffusion which is a whole open- Source
1:42:10 world that there's all sorts of flips
1:42:13 and switches and dials to if you like
1:42:16 tinkering Leonardo's you're going to
1:42:18 love it if you work for a company Adobe
1:42:21 Firefly because it hasn't been trained
1:42:24 on Commercial images and mid Journey has
1:42:27 and and like mid Journey like a document
1:42:30 just came out on all the artists that
1:42:31 they trained it on so they might get
1:42:34 their asses sued into Oblivion and then
1:42:37 you're stuck right so so how do you know
1:42:39 all this stuff you just start you get
1:42:42 you get curious you show up here oh wait
1:42:44 I'm here every night and I love
1:42:46 listening to you talk as a developer I
1:42:48 need to get with the times oh wow yeah
1:42:51 so Kenny boy I mean so Kenny boy I
1:42:54 mean I found this on the I don't care
1:42:57 Siri um as
1:43:00 you if you're a developer you have the
1:43:04 skills to dig deep and learn something
1:43:06 right so so this is all all this is and
1:43:10 and what the AI salon's about and what
1:43:11 this channel is about is about flipping
1:43:14 that switch in your head to just start
1:43:15 getting curious about it and then the
1:43:17 learning of it is
1:43:19 just like why do I do these seven nights
1:43:22 a
1:43:23 week my family asks the same
1:43:26 thing my co-workers ask the same thing
1:43:29 why do I do this part of it is for the
1:43:31 community like the community that we're
1:43:33 building here is [ __ ] remarkable part
1:43:35 of it for
1:43:36 me if I do these every night it keeps me
1:43:39 in the
1:43:40 conversation and and people will just
1:43:43 ask questions and and what I'll start to
1:43:44 notice is the conversation over time
1:43:47 starts to drift in a different direction
1:43:50 like like for a while it was just it was
1:43:52 all about chat gbt and then like in the
1:43:54 last like the 3 weeks before
1:43:57 Christmas there was just a lot of talk
1:43:59 about video and sunno music and like
1:44:02 there were like multimedia started
1:44:04 sneaking into the
1:44:06 conversation and it's
1:44:08 like do like are the video tools getting
1:44:12 that much better or that much
1:44:14 different no but they're getting good
1:44:16 enough that people that were ignoring
1:44:19 them before all of a sudden aren't
1:44:21 ignoring them and then if I look at the
1:44:23 AI Salon people will be sharing you know
1:44:26 where they were sharing just images
1:44:29 before now they're sharing gpts and
1:44:31 they're sharing songs and they're
1:44:32 sharing images that they turned into
1:44:34 videos right and so it's
1:44:39 it we're in the primordial soup stage of
1:44:43 the generative AI
1:44:45 era and we're going to be in that
1:44:47 primordial soup stage for probably 3
1:44:49 years like 18 months at a minimum
1:44:51 minimum 2024 is going to be absolutely
1:44:54 [ __ ]
1:44:56 Bonkers but it's all going to be raw
1:44:59 janky weird ass [ __ ] where it seems like
1:45:03 here's the leader in this industry and
1:45:04 all of a sudden some new Tool's going to
1:45:06 come out that's going to blow that out
1:45:07 of the water and we're just going to
1:45:09 see like primordial soup Like Big
1:45:12 Bubbles little
1:45:14 bubbles things crawling out of the
1:45:17 Ooze and
1:45:19 so there's no ree that you can take that
1:45:22 will teach you that there's no course or
1:45:24 Workshop I can't teach you that the only
1:45:27 way we're going to learn that is to to
1:45:29 do it together as a
1:45:31 community and and and hey I learned this
1:45:35 thing that I think's kind of stupid
1:45:36 right this is dumb and then you'll show
1:45:38 it to people and everyone goes no that's
1:45:39 not dumb that's [ __ ] amazing how did
1:45:41 you do that
1:45:44 what every new thing I want to learn I
1:45:46 have chat GPT write me a beginner level
1:45:49 ebook on it
1:45:51 archetypal architect that's [ __ ]
1:45:54 brilliant every new thing I want to
1:45:56 learn I have chat GP write me a beginner
1:45:59 level ebook on it [ __ ] brilliant
1:46:05 perfect and you know what you could do
1:46:08 if you want it if you so
1:46:11 desired have it write that beginner
1:46:14 level ebook pop it into a canva template
1:46:17 and publish it on Amazon Amazon
1:46:22 Kindle why
1:46:24 not AGI will dump the apple cart over
1:46:28 quick or not yep absolutely will AGI is
1:46:31 going to [ __ ] some [ __ ]
1:46:34 up are you Italian no but I speak flu in
1:46:37 Italian thanks to hey
1:46:41 Jen it's so true let me see do I have it
1:46:44 up
1:46:47 [Music]
1:46:49 here
1:47:04 so we've heard from a number of clients
1:47:06 that modular content is important to
1:47:08 them getting more efficient with how we
1:47:10 create content and how we leverage
1:47:19 content behind me you see these little
1:47:21 green boxes um that's how we think about
1:47:23 stories at Story vine we break stories
1:47:25 down into little atoms right little
1:47:34 modules A Story
1:47:37 vine I don't know how to speak Italian
1:47:39 but I do I speak it fluently I also
1:47:42 speak Mandarin fluently bet you couldn't
1:47:45 speak Mandarin
1:47:47 fluently now you
1:47:49 can
1:47:51 I can I can show you all these products
1:47:53 we're doing at Story vine they're
1:47:54 they're [ __ ] nuts like it they're
1:47:57 they're nuts I want to get out of the
1:47:59 Ooze myself
1:48:01 personally yeah I mean so so Brad you
1:48:05 can I mean you can take any one of these
1:48:07 tools right now and you could specialize
1:48:09 on it but I think
1:48:12 that I just I think we're in an we're in
1:48:16 a
1:48:17 phase
1:48:19 where it's going to be in all of our
1:48:22 best interests to
1:48:25 to learn as much as we can about what
1:48:29 these tools are what they're
1:48:33 not what they make possible because
1:48:36 because right
1:48:38 now all these tools are these
1:48:41 disperate islands of expertise right if
1:48:47 you want to make something in Runway
1:48:48 you've got to go to Discord and do their
1:48:51 thing and then now you can do it on the
1:48:52 website but there's a whole prompting
1:48:55 language and now there's camera
1:48:56 movements you've got to and and then you
1:49:00 go over to this thing to make the songs
1:49:02 but the songs are sort of good but
1:49:03 they're not good but then you can go to
1:49:04 this other tool that right like
1:49:06 everything is just this these weird ass
1:49:09 janky ass half-working
1:49:13 mind-blowingly powerful don't get me
1:49:18 wrong imperfect
1:49:22 tools they're going to all start to
1:49:25 merge they're going to all start to
1:49:27 merge and as they start to
1:49:32 merge if if you're not enough in the
1:49:34 game you may not even know that the new
1:49:37 feature showed up right like that's kind
1:49:39 of what happened in chat GPT that chat
1:49:42 GPT was relatively simple for most of
1:49:44 the year you put words in it and words
1:49:47 came out you put words in it and code
1:49:50 came out you put words in it and
1:49:52 formatted words came out like it was
1:49:54 pretty
1:49:55 straightforward and then they connected
1:49:57 it to the internet and it kind of did a
1:49:59 little bit more and then they took that
1:50:01 away and
1:50:02 then and then all of a sudden they
1:50:05 Incorporated image generation in it and
1:50:08 then they added image Vision in it so it
1:50:10 can now see images and understand
1:50:14 them have you seen
1:50:16 it read a
1:50:19 whiteboard you can upload a picture of a
1:50:22 whiteboard to chat
1:50:23 gp4 and say what is this and it will
1:50:26 analyze what's on the Whiteboard and if
1:50:28 it's sketches of software it will
1:50:31 understand those sketches and you can
1:50:33 tell it to make the interface for that
1:50:35 software or make an ad for that software
1:50:38 or write the code for that software
1:50:41 right now today it's in there right
1:50:46 now have you played with
1:50:49 that
1:50:51 I took a picture of I've got in in in my
1:50:56 ADD Marvel Mystery Marvel's Mystery Oil
1:51:03 brain you know you go into your
1:51:05 hyperfocus tunnel so about two years ago
1:51:07 I went into a hyperfocus tunnel and I
1:51:09 did a a Post-It note brainstorm about
1:51:12 the future of storyvine it's been on my
1:51:15 it's been on my whiteboard
1:51:18 for two years
1:51:22 and have I ever like typed up the notes
1:51:25 and written it into the document it
1:51:27 needs to be done no I took a picture of
1:51:30 it and said what is
1:51:33 this and it took all the notes for me
1:51:36 wrote it said here's what this is here's
1:51:38 what you were talking about here's what
1:51:39 you're thinking about here's your
1:51:41 options here's your threats here's your
1:51:43 this here's your
1:51:45 that if you haven't done that yet like
1:51:48 that's that that's a whole category
1:51:52 of how to use chat
1:51:53 GPT that isn't really about prompting
1:51:57 it's about okay if I've got a tool that
1:52:00 can understand images like understand
1:52:06 them what do I do with that I don't
1:52:09 [ __ ]
1:52:13 know no one does I mean some people do
1:52:17 the people that have been focusing on
1:52:18 that kind of have a clue
1:52:21 and then you take that and you combine
1:52:23 that with code interpreter that can read
1:52:25 an Excel file and generate charts and
1:52:30 graphs and you take a thing that can
1:52:32 write a song and you take a thing that
1:52:33 can make a video and then you make
1:52:36 pictures how does all this work together
1:52:38 I don't
1:52:42 know Kyle someone's got to teach me this
1:52:46 no one's going to teach you this if you
1:52:48 wait five years yeah there'll be a
1:52:50 course on this [ __ ] the opportunity will
1:52:54 have passed the like the massive
1:52:57 opportunity there's going to be
1:52:58 opportunity with this AI stuff for the
1:53:01 next decade 20 years like this is
1:53:04 everything's about to be
1:53:07 transformed you spending time in here
1:53:10 and at the AI salon and just being
1:53:12 curious in life and talking about what
1:53:14 you're learning one of the reasons
1:53:17 generosity is a value of the AI salon
1:53:21 is that telling people what you're
1:53:24 learning isn't just for them you get a
1:53:28 benefit of that too which is if you
1:53:31 start putting on LinkedIn and Facebook
1:53:34 and at your dinner
1:53:36 parties and and at work if you start
1:53:39 saying oh I was working on chat GPT and
1:53:41 this thing came out even if the response
1:53:44 is I hate AI why are you talking about
1:53:46 AI you dumb
1:53:49 dum
1:53:53 those people will recognize you as
1:53:54 someone who understands what's going on
1:53:57 and at the point at which they realize
1:53:59 that they're missing the boat and they
1:54:01 got to get their [ __ ] together who do
1:54:03 you think's the first person they're
1:54:04 going to
1:54:06 call and you know what's been happening
1:54:08 to the Irregulars in here the the people
1:54:10 that come in here
1:54:12 regularly that work for big companies
1:54:14 that went from being secret cyborgs
1:54:17 people that are secretly AI literate
1:54:20 to starting to talk about it at
1:54:23 work they were initially met with why
1:54:26 you talk about that stuff we can't use
1:54:27 AI at this work you know you're going to
1:54:29 get fired if you keep using AI at this
1:54:32 work you know what happened to
1:54:35 them those people came back and said um
1:54:38 it looks like the marketing department
1:54:40 wants to learn about AI could you give
1:54:41 us a workshop on
1:54:44 that are you still doing that AI
1:54:49 stuff and that Workshop is leading to
1:54:52 hey VP Bob wants to know about AI can
1:54:55 you meet with VP Bob
1:54:58 sure VP Bob just talked to the CEO the
1:55:01 CEO wants to talk to you
1:55:05 okay like that's
1:55:08 happening that's happening in here to
1:55:10 people that show up to this stupid
1:55:11 [ __ ] Tik Tok
1:55:13 Channel why cuz we're in primordial soup
1:55:18 right now nobody knows about this
1:55:21 stuff and you know what's amazing is the
1:55:25 people that that's happening to that are
1:55:27 being recognized as experts do you know
1:55:29 what they say when they get asked that I
1:55:32 I'm not ready I don't know what I'm
1:55:34 talking
1:55:35 about but you know what they
1:55:38 do they do know what they're talking
1:55:41 about it's just their expectation is I
1:55:44 want to be able to be like master's
1:55:47 degree level expert at this [ __ ] and and
1:55:51 we're not even like we're not even in
1:55:53 preschool yet we're just like like like
1:55:58 babies crawling around like a a wet
1:56:01 field right now no no one's got language
1:56:04 no one's got skills no one can stand up
1:56:07 we're all just walking around like
1:56:09 barely keep our head
1:56:11 up that's the stage we're at right now
1:56:13 so everyone feels ill prepared right now
1:56:17 and that's
1:56:19 okay
1:56:20 but what's not okay in my opinion is
1:56:23 sitting on the sidelines
1:56:26 waiting or Worse sitting on the
1:56:29 sidelines
1:56:30 judging dismissing it oh it's not
1:56:34 perfect oh it hallucinates you know
1:56:37 they're stealing from
1:56:43 people I don't use AI out of principle
1:56:48 okay
1:56:50 I don't use AI out of principle is like
1:56:53 Florida passing a law that hurricanes
1:56:56 are are are now
1:56:59 illegal it's a laudable law I understand
1:57:03 why you passed it but you know there's a
1:57:04 hurricane still coming right I have a
1:57:07 running list of parlor tricks so I'm
1:57:09 always ready if someone asks Source Camp
1:57:11 so Source Camp's one of these people
1:57:13 that is uh is being professionally
1:57:17 rewarded by being willing to be generous
1:57:21 enough when someone says I don't know
1:57:23 about Ai and so when she talks about
1:57:24 parlor tricks the list of parlor tricks
1:57:27 so when someone asks me in here what's
1:57:29 chat GPT how do you use that my parlor
1:57:32 tricks are like one of my go-tos is
1:57:35 explain quantum mechanics to me in an
1:57:37 Eminem rap it sounds funny and it does
1:57:40 Eminem waps pretty good and it explains
1:57:42 quantum mechanics so so like that's a
1:57:45 parlor trick um have it create a table
1:57:47 comparing two ridiculous Concepts that
1:57:51 you should never compare but it'll do
1:57:52 that that's a parlor trick um now you
1:57:55 can do things like have it make an image
1:57:58 of a Jar full of jelly beans in a field
1:58:01 with you know llamas Turtles and
1:58:04 hedgehogs running around the jar of
1:58:05 jelly beans it'll do that right so those
1:58:08 parlor
1:58:09 tricks
1:58:13 are like low education ways to show
1:58:17 people that these things are not just a
1:58:20 Google
1:58:20 search that there's something different
1:58:23 going on here and if you can get someone
1:58:25 to lean in and go wait a minute what
1:58:27 what did it do did did someone write
1:58:32 like is that copying and pasting That
1:58:34 Eminem rap from some Physics thing no it
1:58:37 just wrote that what do you mean it
1:58:39 wrote it the G in GPT stands for
1:58:43 generative so it actually generates that
1:58:45 it it wrote like that's an original
1:58:47 piece of writing it can do that yeah
1:58:52 yeah well it can't write marketing copy
1:58:54 well what are you working on right now
1:58:56 I'm working on this well here let's see
1:58:58 uh write a blog post about this concept
1:59:00 for that
1:59:02 product it could do that but it's not
1:59:06 good right we'll read
1:59:09 it oh my God that's really good well
1:59:12 that that part's bad but you know I
1:59:15 could work with
1:59:16 this right that's that Kevin McAllister
1:59:19 moment
1:59:20 I could really use one of those jobs
1:59:22 Woodworkers can't find a use for AI yet
1:59:25 oh archetypal architect Woodworkers I I
1:59:29 totally disagree I I I
1:59:31 think con conceiving different new kinds
1:59:35 of products new kinds of
1:59:37 designs I
1:59:38 mean you Woodworkers so
1:59:42 so holy [ __ ] so so um one of the guys
1:59:47 one of the guys who's an a regular in
1:59:48 here
1:59:51 um I think is in construction but but
1:59:55 he's he's either in construction or he's
1:59:57 just he's a supplier for people that
1:59:59 work in
2:00:00 construction he's using Code interpreter
2:00:04 to take pricing lists and dynamically
2:00:07 pull out sort of deals of the week so
2:00:10 they could use it for for sourcing
2:00:12 materials they could use it for Design
2:00:15 Concepts um I there's all like like
2:00:19 honest to God I I can't I can't imagine
2:00:24 a job that you couldn't look at these
2:00:27 tools and find some use for
2:00:29 them I think Woodworkers could find all
2:00:32 sorts of crazy ass stuff to do with this
2:00:34 coming up with
2:00:36 designs I did a I did a a a GPT um to
2:00:41 make floral
2:00:44 bouquet where I asked chat GPT um can
2:00:50 you tell me the meaning like the
2:00:52 emotional meaning of certain flowers and
2:00:55 of course it could
2:00:56 right I don't know red roses are love
2:01:00 and tulips are this and something is
2:01:01 that and so I was like okay if it can
2:01:04 understand um if it can understand the
2:01:07 meaning of
2:01:08 flowers then maybe you could you it
2:01:10 could understand how to create a
2:01:12 beautiful
2:01:14 bouquet that means
2:01:16 something and so I created a GPT that
2:01:19 basically you tell it like what emotion
2:01:21 you want your bouquet to
2:01:24 convey and then it
2:01:27 generates a bargain bouquet and a
2:01:31 premium bouquet and then it explains
2:01:34 which flowers it chose and what they
2:01:37 mean well if you've got a flower shop
2:01:40 imagine that you create and and and then
2:01:42 what it did so so it would design the
2:01:45 bouquet with with specific species of
2:01:48 flowers and then it would generate the
2:01:51 image of those
2:01:53 bouquet and then it would say if you
2:01:56 want to buy one of these bouquet reach
2:01:58 out to my flower shop but it would give
2:01:59 you the URL or the phone number or what
2:02:01 it was lunatick I built a GPT for glass
2:02:04 blowers I bet I could build one for
2:02:06 Woodworkers there you go team up with
2:02:08 lunatick so she's in the comments right
2:02:12 now if you think there's nothing you
2:02:14 could do for Woodworkers sounds like
2:02:15 lunatick will build you a GPT right now
2:02:19 for for your woodworking business or for
2:02:22 Woodworkers in
2:02:25 general but like if if you say well AI
2:02:29 is not going going to impact the the
2:02:32 florist well maybe it will if the
2:02:34 florist says here's a tool for all my
2:02:37 customers for free where they can design
2:02:40 their own bouquet where they can say I
2:02:43 want it to mean inspiration and love and
2:02:47 support and then it generates a cool
2:02:49 little report that says I chose the
2:02:51 flowers in this bouquet for for you know
2:02:55 inspiration love and support and here's
2:02:57 what they mean and then here's the
2:02:59 bouquet and that's the one I want
2:03:01 built and then that flower shop can
2:03:04 build that bouquet and the and the
2:03:05 person can pick it up you know 4 hours
2:03:09 later that's a that's a completely new
2:03:11 product that wasn't possible
2:03:13 before is there an AI to increase the
2:03:16 resolution of an image generated with
2:03:17 doly yeah there's a lot of them
2:03:20 um uh you there's uh you can go to clipd
2:03:25 drop. they've got an upscaler there's a
2:03:27 whole bunch of the tools right now do
2:03:29 upscaling there's there's a
2:03:31 new 35 Buck a month tool called I think
2:03:35 it's called Majestic or something like
2:03:37 that um there's another one called topaz
2:03:40 Labs there's some commercial ones that
2:03:41 you can pay for but there's also a bunch
2:03:44 of free ones so if anyone knows some of
2:03:46 the free upscaling tools just pop them
2:03:47 in the comments below
2:03:56 any suggestions for practical demos for
2:03:58 older folks my grandma doesn't see how
2:04:01 she'd use it
2:04:02 yet I think the the the on-ramp that I
2:04:06 found Sam good to see you Sam by the way
2:04:08 I haven't seen you in a while good to
2:04:09 see you um the the the on-ramp that that
2:04:13 I find is is almost always
2:04:18 um some interest of theirs and you can
2:04:21 start with pictures because dolly is so
2:04:24 dolly is so instant and
2:04:27 magical that if if your
2:04:30 grandma was really into doilies or
2:04:34 crocheting or like like like Becky Ru
2:04:37 who's in here is really into quilting
2:04:39 and she made a GPT that designs quilts
2:04:42 and it makes these beautiful [ __ ]
2:04:44 quilts so so I think if you just like
2:04:48 you know your grand Grandma you know
2:04:49 what her interests are you know maybe
2:04:53 like like my mother always talked about
2:04:55 this gymnastics meet she did when she
2:04:57 was 14 you know how sometimes older
2:04:59 people will talk about the same thing
2:05:01 over and over again there's a there's a
2:05:02 thing in their life that's really
2:05:03 meaningful to
2:05:05 them create an image from that era of of
2:05:10 you know describe your grandma what she
2:05:11 might have looked like in that era you
2:05:13 know doing the Jitterbug or whatever it
2:05:16 is that she talks about and show her
2:05:19 that so so like I I feel like the onramp
2:05:22 of something that's relevant to them
2:05:26 instantly where where where they're like
2:05:28 I don't get it I don't get it and you
2:05:30 show them a picture that's something
2:05:31 that they're like well wait where did
2:05:33 you get that well it just made that
2:05:36 what or another thing Sam could be pie
2:05:43 right so you know all those all those
2:05:46 crazy Tik Tok videos a GPT with
2:05:48 entertainment triv it from the 50s 60s
2:05:50 yeah that would be perfect
2:05:52 one you know the the the Tik Tok and
2:05:54 YouTube videos where they're trying to
2:05:56 get the Italian grandmother to talk to
2:05:59 Alexa and and it just it's just like an
2:06:01 epic fail for like an like 20 minutes of
2:06:05 Grandma trying to talk to Alexa and it
2:06:07 just fails um do it with pie like
2:06:13 [Music]
2:06:17 like I'm sitting here with my
2:06:19 grandmother and she loves the Jitterbug
2:06:22 and dancing in the 20s what can you tell
2:06:24 me about how the Jitterbug
2:06:29 started gpts on Demand by lunar stick
2:06:33 [Music]
2:06:35 exactly the Jitterbug has a fascinating
2:06:38 history that dates back to the 1920s and
2:06:41 1930s it originated from a fusion of
2:06:45 African-American dance styles such as so
2:06:48 so so I mean I think it I think it I
2:06:51 think it's always relevance I think it's
2:06:53 always storytelling I think I think the
2:06:57 bridge to getting to someone to use
2:06:59 technology is the interface has to be
2:07:01 simple enough and clean enough that you
2:07:03 can do it
2:07:05 well
2:07:07 conversational
2:07:10 gpts conversational gpts are now good
2:07:13 enough I mean if your grandmother speaks
2:07:16 another language she could speak
2:07:20 that language to piie and go back and
2:07:24 forth between English and and that
2:07:26 language lunatick I built a classic car
2:07:29 GPT that's kind of an older person hobby
2:07:32 yeah I built a a 70s muscle car poster
2:07:35 maker so so you basically you give it
2:07:39 your 70s muscle car model then you pick
2:07:42 your color and then you pick whatever
2:07:44 features you want like you want a blower
2:07:46 and you know spoiler side pipes and then
2:07:50 it makes you a poster with like if it's
2:07:54 a GTO Judge it says the judge and it
2:07:57 misspells it every time that's a fun
2:08:05 one I use pie every day I'm a little
2:08:08 frustrated it told me I was it was going
2:08:10 to email me a huge table okay so the
2:08:12 thing to know about PI right now is it
2:08:14 cannot take actions and it does not know
2:08:17 that it cannot take actions sometimes so
2:08:20 sometimes it will tell you sure I'll
2:08:21 send that to you right now it can't send
2:08:23 it to you um The Other Place Pi is
2:08:27 really frustrating is um if you use the
2:08:31 text the desktop version of it and
2:08:33 you're
2:08:33 typing and you want to get it to do
2:08:36 something to actually like give you an
2:08:38 answer oh my God is it frustrating
2:08:40 because it just asks you questions well
2:08:42 tell me why do you why do you want me to
2:08:44 give you that answer because I actually
2:08:46 just want it I understand your
2:08:49 frustration but but what's the purpose
2:08:51 of of me giving you this answer I really
2:08:54 want to know I know you really want to
2:08:56 know can you just give me the answer boy
2:08:58 I I would love to as soon as I
2:09:00 understand what's difference this is
2:09:02 going to make for you if I give
2:09:06 you but
2:09:08 again how did I learn that [ __ ]
2:09:12 around dick around with AI it will serve
2:09:15 you why because you'll realize oh is
2:09:19 really good on a 4-Hour car trip it's
2:09:21 really shitty if I want to get something
2:09:23 done in the office chat GPT is really
2:09:26 good for this but it's not good at this
2:09:27 one thing over here but Bard's better at
2:09:30 that so let me how do you learn that you
2:09:32 just learn that oh I'm trying to get
2:09:35 this sort of image in Del and it's not
2:09:37 working where do I get that I go to Mid
2:09:39 journey to get that because I know that
2:09:41 that's better there like how do I know
2:09:42 that you just know that by slamming your
2:09:45 head against the wall over and over and
2:09:48 over
2:09:50 again why would I take the time to learn
2:09:52 all these different pieces of software
2:09:54 well because you've had your Kevin
2:09:56 mallister moment you realize the world
2:09:59 is about to change I would argue the
2:10:01 world has already changed just not that
2:10:03 many people know it yet the world is
2:10:05 about to
2:10:07 change and a lot of people a lot of
2:10:11 people are going to be caught off
2:10:16 guard a lot of company are going to be
2:10:19 caught off guard with how quickly things
2:10:23 are going to
2:10:24 change you know who's not going to be
2:10:26 caught off guard this
2:10:28 group people that show up to this
2:10:31 channel regularly people that show up to
2:10:32 the AI Salon if I see there's a lot of
2:10:35 new people in here
2:10:38 um go
2:10:40 there the the first three links are
2:10:43 about this thing called GPT for good
2:10:45 which was a 24-hour GPT buildon we did
2:10:48 this past weekend the fourth link is the
2:10:51 link to the community the mighty
2:10:53 networks community at at um ai salon.
2:10:58 mn. go check out the AI Salon this is a
2:11:01 community of
2:11:02 people on this AI Adventure like
2:11:05 actively choosing to be on this
2:11:08 adventure actively choosing to be
2:11:11 generous about what they're learning and
2:11:13 sharing everything they can and learning
2:11:16 everything they can and Reporting what
2:11:18 they've discovered oh I just saw a tweet
2:11:21 on this thing that we've been talking
2:11:23 about look it looks like it's real now
2:11:25 oh and then people run over and try it
2:11:27 and they're like no it was [ __ ] okay
2:11:30 cool or they run over and try it oh my
2:11:33 God it's amazing it's a whole new thing
2:11:35 we got to go play and then people go
2:11:37 play and then they post and show and
2:11:38 tell and they right become an AI
2:11:41 Apprentice we're all AI apprentices
2:11:43 right now absolutely that's from Lord
2:11:46 digital
2:11:47 Gods digital Gods is so Godly that he
2:11:52 also is a lord on top of being a God I
2:11:56 didn't didn't did you even consider that
2:11:59 was a possibility now thanks to AI
2:12:02 That's a
2:12:07 reality lunatick you're an AI evangelist
2:12:10 you're an AI like you're a [ __ ] AI
2:12:12 rock star right now lunar stick you're
2:12:14 just
2:12:15 like she's she's just barreling
2:12:19 forward I I mean you know what's
2:12:23 amazing
2:12:25 is Da Daisy Daisy Thomas is her
2:12:30 name Daisy has
2:12:39 discovered she's
2:12:41 discovered these AI tools as a form of
2:12:44 personal expression which I find really
2:12:47 really inspiring she
2:12:50 she's she's not concerned with it's
2:12:54 funny Pate just
2:12:55 put Pate I'm here against my will
2:12:58 exactly you know Pate works for Google
2:13:01 Pate works on tpus and making them
2:13:03 better and faster so that all of our AI
2:13:06 [ __ ] works right Daisy's not concerned
2:13:09 with the mathematics Behind these tools
2:13:11 she's not concerned with how they work
2:13:12 she understands how they work she
2:13:14 understands what they can do but how
2:13:16 she's using AI right now is as a form of
2:13:19 personal expression she's getting
2:13:22 personal fulfillment out of building
2:13:23 these tools and giving them to other
2:13:25 people to do their [ __ ] with and she's
2:13:29 taking ideas that she's had in her head
2:13:31 for a decade or two or three written on
2:13:34 an index card and she's like I can do
2:13:36 this
2:13:38 now we're in a phase where we can take
2:13:44 radically
2:13:47 powerful comp
2:13:50 computational
2:13:54 systems and treat them as self
2:13:59 expression that's
2:14:01 mindblowing that's where we are right
2:14:03 now like that's the reality right now if
2:14:06 you choose to see it that way if you sit
2:14:09 back with your arms
2:14:11 crossed on a soap box going no they're
2:14:14 not perfect no they hallucinate no
2:14:17 they're stealing from
2:14:22 what you're losing is the joy of
2:14:25 discovering what they actually are and
2:14:28 when you discover that Lord digital Gods
2:14:31 title gifter good to see you thank you
2:14:32 for the Roses means a lot I appreciate
2:14:36 it a
2:14:38 lot the joy of Discovery right now is
2:14:41 everything do you I you may or may not
2:14:44 know
2:14:45 this well no this is this is patent
2:14:48 obvious this is a stupid [ __ ]
2:14:51 statement you only get one opportunity
2:14:53 in
2:14:55 history to be at the
2:14:57 beginning of a new technological
2:15:01 era we're at the beginning of this I've
2:15:04 been lucky enough in my life I just
2:15:08 missed one in the 70s the personal
2:15:10 computer I hit one in the mid90s the
2:15:13 worldwide
2:15:15 web and this is the third one in my life
2:15:19 and this one is so much bigger than the
2:15:21 worldwide web it's so cool and it's also
2:15:23 fun when they don't work I've deleted my
2:15:26 fair share of bad ideas yeah yeah
2:15:28 because in building something that
2:15:30 doesn't work Daisy you're you're you're
2:15:33 discovering a boundary right oh they're
2:15:36 really good up until this
2:15:39 wall and then what's going to happen is
2:15:41 6 months from now that wall is going to
2:15:44 disappear and you're going to go oh I
2:15:47 had to delete it GPT cuz it sucked at
2:15:49 that before let me try it now oh it
2:15:51 works now and there's a whole new world
2:15:55 of gpts beyond that
2:15:57 right it's
2:16:00 amazing lunatick you're an AI llm
2:16:04 supermodel it's
2:16:07 true but that's the air wi like that's
2:16:10 the the the the payback for investing
2:16:14 stupid amounts of time in AI right now
2:16:17 is that you get to discover [ __ ] before
2:16:19 other people discover it and you may
2:16:22 discover a way of using these things
2:16:24 that no one's
2:16:26 discovered and you don't need a math
2:16:29 degree from Stamford to do
2:16:31 that you just need your point of view
2:16:34 like one of the reasons
2:16:36 inclusion is one of our values in the AI
2:16:39 Salon is that when someone comes in with
2:16:42 a different worldview a different lens
2:16:45 that they're looking at this technology
2:16:46 through they're going to to use it in a
2:16:49 way you couldn't have
2:16:50 imagined and they're going to discover
2:16:53 something that you're like holy [ __ ] I
2:16:57 never thought about using it like that
2:16:59 yeah and now when you use it like that
2:17:02 with your worldview you're going to sort
2:17:04 of mix your [ __ ] into it and you're
2:17:06 going to come up with some new
2:17:08 completely different thing that's going
2:17:09 to blow someone else's mind that's the
2:17:11 pr primordial soup right now oneclick
2:17:15 Customs yep I wish I got with Google
2:17:17 into 2000 missed Bitcoin in 2010 now I
2:17:21 should learn AI I would say oneclick
2:17:24 Customs just take the should out of that
2:17:26 sentence I am learning AI just [ __ ]
2:17:30 flip the switch right now start showing
2:17:32 up at this channel join the AI
2:17:34 Salon start telling people in your life
2:17:37 that you're doing this because there
2:17:38 going to be another weirdo out there
2:17:40 that you didn't know was doing this that
2:17:42 the minute you say I'm really into this
2:17:43 AI stuff now it's going to go I am too
2:17:46 what did you learn oh I just just
2:17:48 learned this but I'm really bad at this
2:17:50 I'm really bad at that but I'm good at
2:17:51 this hey you want to do something
2:17:53 together do that start
2:17:56 tonight thank you all of this Tech has
2:17:58 given me new purpose again and that's
2:18:00 like magic yeah it I I know it's your
2:18:04 story right now of like just watching
2:18:05 you Blossom and how you're doing playing
2:18:07 with this stuff lunatick is just really
2:18:09 it's super cool it's very inspiring very
2:18:12 very
2:18:13 inspiring just me and the trees is here
2:18:16 Cindy [ __ ] is here Cindy Coon's in here
2:18:19 I I got to tell you you got to be
2:18:20 careful of Cindy [ __ ] lab five Futures
2:18:23 she's really in here on a pro raccoon
2:18:26 agenda um raccoons have reluctantly be
2:18:30 been accepted into the AI learning lab
2:18:34 Zoo um but there's there she there's an
2:18:37 agenda there's an agenda it's a pro
2:18:40 raccoon agenda now the agenda the
2:18:42 raccoons don't want the ferrets there's
2:18:44 a whole thing there's a whole thing
2:18:47 happening
2:18:49 sure AI will transform your life but you
2:18:52 got to be careful of uh lab five
2:18:56 Futures actually Cindy Coon's quite
2:18:59 remarkable you should you should get to
2:19:00 know her she is so cool join a boys band
2:19:04 via videographer okay that's awesome
2:19:08 Sherry D Sher
2:19:10 D just is in the middle of this really
2:19:13 inspiring video project um making a
2:19:17 video
2:19:18 um using generative AI
2:19:22 tools has she been a video producer
2:19:25 before I don't know but she is now and
2:19:28 it's she showed it to us it's amazing
2:19:30 it's really
2:19:31 good oh
2:19:35 [Music]
2:19:37 man what time is it it's got to be time
2:19:40 to yeah see I can feel it in my throat I
2:19:42 haven't been drinking water
2:19:46 either see when I get on of rolling here
2:19:48 you get you guys got me all fired up
2:19:51 talking about this this here AI stuff I
2:19:53 got all fired
2:20:01 up all right let's wind this let's let's
2:20:04 wind this this show up forever
2:20:09 hook making a whole album are you really
2:20:11 Sherry D holy
2:20:14 [ __ ] wow that's going to be so cool
2:20:21 yeah like like what does you
2:20:24 know we're in this weird era there's a
2:20:27 if you want to see something if you want
2:20:28 to see the beginnings of something that
2:20:30 that's cool and and Sherry d by the way
2:20:32 if you haven't submitted your stuff to
2:20:34 Runway TV you should absolutely do that
2:20:37 go check out runway. TV right
2:20:41 now it's
2:20:43 a if if you're if you're if you're an
2:20:47 older fell or lady like like
2:20:50 myself you probably remember the launch
2:20:54 of MTV in 1981 remember the commercials
2:20:57 on network television I Want My MTV and
2:21:00 we didn't know what the [ __ ] that
2:21:02 meant and then MTV launches and and you
2:21:06 like you you like went to it on your
2:21:08 cable TV
2:21:10 box and then you you kind of watched
2:21:13 with your head cocked like so you watch
2:21:16 music now and every and and like no one
2:21:19 got it do you remember the early days of
2:21:21 that that's what Runway TV is to me that
2:21:25 you look at it and it's it's all these
2:21:27 videos made with these text to video
2:21:29 generator tools and you know some of
2:21:31 them are really good and really high
2:21:33 quality storytelling you can tell that
2:21:35 there's storytelling going on there and
2:21:37 um the
2:21:38 the oh oh ironically it wasn't done with
2:21:41 runwayml so you can't oh that's a drag
2:21:43 but okay um but you go to runway. TV
2:21:48 and it it just it has a 1981 MTV vibe to
2:21:53 it it's like hey I know we could get
2:21:56 some of the videos made with our tool
2:21:58 and we could make a directory of those
2:22:00 videos and it's basically just a web
2:22:02 rapper sitting on top of YouTube
2:22:05 links but here's the thing about Runway
2:22:10 TV it's so janky right now that the
2:22:14 establishment streaming Community like
2:22:16 Netflix and Spotify and the
2:22:19 establishment film making Community like
2:22:22 Paramount Pictures and production
2:22:25 company a B and C are going to look at
2:22:28 it and they're going to
2:22:30 go yeah but that's not real film making
2:22:34 I'm a real filmmaker okay okay okay
2:22:37 we're not using this AI stuff for
2:22:39 production not in this house okay we use
2:22:42 real film and we use real actors and we
2:22:44 use real this and they're going to
2:22:46 dismiss it
2:22:48 and what's going to happen is whether
2:22:50 it's Runway TV or it's someone
2:22:53 else there's going to be a whole new
2:22:56 world of video storytelling
2:23:00 storytellers that emerges the new
2:23:02 artists and there's going to be kind of
2:23:04 three categories there's going to
2:23:06 be filmmakers who resist Ai and they're
2:23:11 going to keep making films the way they
2:23:12 make films there's going to be
2:23:14 filmmakers that Embrace Ai and and some
2:23:17 of them are going to emerge in this new
2:23:19 era era
2:23:21 area as a new kind of Storyteller and
2:23:25 then there's going to be people like
2:23:26 Sher D and other people who've never
2:23:29 done this before in their lives but
2:23:31 Embrace this AI stuff and and and learn
2:23:34 how to do video
2:23:38 storytelling using a completely
2:23:40 different kind of camera which is
2:23:43 AI
2:23:45 and you just watch mark my words in in 3
2:23:49 years we're going to see something like
2:23:51 a Runway TV emerge as a major
2:23:56 entertainment
2:23:58 platform and a and a major Studio that
2:24:01 will be absolutely Inc competition with
2:24:04 the status quo film making
2:24:07 industry and streaming industry watch
2:24:10 you
2:24:11 watch it's going to be wild it's going
2:24:14 to be wild so
2:24:16 anyway
2:24:18 cryptowolf have you seen the videos I'm
2:24:20 making with Runway ml I don't know but
2:24:22 cryptowolf um I yeah I think so I've
2:24:24 seen a couple of them um but but you
2:24:26 should be you should share them in the
2:24:28 in the video channel in the salon if
2:24:31 you're not and in in showand tell as
2:24:33 well um show them in
2:24:35 both
2:24:38 um open source beats Runway ml I think
2:24:42 we need to get the engineers and
2:24:44 creatives together here to make them
2:24:46 yeah I mean here's the deal with open
2:24:49 source is always good for Innovation um
2:24:51 it's almost always bad for for Consumer
2:24:55 grade user interface because the open-
2:24:59 source stuff is Led primarily by the
2:25:01 engineers because they're the ones on
2:25:03 The Cutting Edge of the technology it
2:25:05 takes that stuff
2:25:08 getting sort of popped up a level from
2:25:11 from interface and Commercial viability
2:25:14 um for for it to be accessible to a
2:25:16 larger audience but I agree if we can
2:25:18 get good interface in the open source
2:25:20 Community awesome but for the most part
2:25:23 it's it's a bit of a nightmare interface
2:25:26 wise any good text to video AIS we can
2:25:29 check out that you know of okay so I'm
2:25:30 going to show you something here so go
2:25:33 to
2:25:35 there
2:25:38 and go to the AI salon so you you go to
2:25:41 the AI Salon it's a mighty networks
2:25:42 thing um couple of thing I'm going to do
2:25:44 this really quick because I want to show
2:25:45 you one specific thing start out at
2:25:48 welcome to the salon check out what the
2:25:50 salon's about what our values are
2:25:52 curiosity generosity exploration
2:25:56 collaboration inclusion and empathy
2:25:58 check those out read what what we're
2:26:00 about if you dig it then you should join
2:26:03 um you should also introduce yourself
2:26:05 you should also go to showand tell and
2:26:07 water cooler and you know just look at
2:26:10 the conversations there and and show
2:26:12 show off some [ __ ] um raccoons are in
2:26:18 see look this is Cindy Coon's agenda
2:26:20 this is what she wants she wants
2:26:21 raccoons everywhere look it's this is it
2:26:24 this is it's it's as if no other animals
2:26:28 are in the AI Salon in the AI learning
2:26:30 lab Zoo look at that look at that look
2:26:32 at what's happening see um but also in
2:26:36 community go to AI tools and resource
2:26:40 and and the main page of AI tools and
2:26:42 resource is this recommendation list
2:26:44 that that I maintain so I personally
2:26:47 maintain this this is these are tools
2:26:51 that I have um sort of personally used
2:26:55 enough to to know that these are at a
2:26:59 minimum not a waste of your time there
2:27:02 are so many tools out there that are
2:27:04 absolutely full-on wastes of your time
2:27:06 there like thousands of them so this
2:27:09 list is is a list that I if if you want
2:27:11 to go down rabbit holes start with this
2:27:13 list so this is at um the the AI Salon
2:27:18 Mighty networks Community um so there's
2:27:21 there's nine different large language
2:27:23 models in here there's kind of the big
2:27:24 four and then there's specialty ones
2:27:27 there's image creation tools and then
2:27:30 there's video creation and so this is
2:27:32 directly answering the question is there
2:27:34 a good text to video tool so the first
2:27:36 two Runway ML and P labs are kind of the
2:27:40 two leaders right now and pabs right now
2:27:43 is free and it's good it's like as good
2:27:46 as Runway ml if not better it doesn't
2:27:49 quite have a lot of the bells and
2:27:52 whistles that Runway has right now but
2:27:54 it's free Runway is not free and it's
2:27:56 pretty expensive P Labs is free it's not
2:27:59 going to be free for long so go make
2:28:01 yourself some [ __ ] go get thee to a pika
2:28:05 make some [ __ ] then there's the second
2:28:08 category in video is Avatar creation hey
2:28:11 Jen and D- ID um are are the biggies
2:28:14 there's also Synthesia but that one's
2:28:16 kind of faded in my opinion um these two
2:28:20 are kind of one uping each other with
2:28:22 features right now did's got this thing
2:28:25 called agents which you can make a
2:28:28 version of yourself that is like a
2:28:30 chatbot agent where someone asks you a
2:28:33 question and you actually talk to them
2:28:36 but you can't put your own voice in it
2:28:38 but at hey Jen you can actually put your
2:28:40 own voice in it and this does video
2:28:42 translation that's really good and these
2:28:45 guys it's just these these are both
2:28:47 really interesting to play with and then
2:28:48 you've got Visa which is it'll make like
2:28:52 a a stock video corporate video that's
2:28:55 got a narration and a voice over and
2:28:58 background music and it'll Source all
2:29:01 the um stock video for you you just give
2:29:04 it a a goal and it goes off and builds
2:29:06 you this video and then you can pay for
2:29:08 it to remove the watermarks so those so
2:29:11 yeah there there's five and I mean
2:29:14 honest to God every one of these is a
2:29:16 whole [ __ ] Rabbit Hole like learning
2:29:18 how to get good at prompting
2:29:22 pabs either with images or with text or
2:29:24 with videos is a whole [ __ ] like you
2:29:29 could spend weeks just on that tool you
2:29:32 could spend weeks just on haen you could
2:29:35 spend weeks just on D- ID every one of
2:29:39 them every one of the tools in this list
2:29:41 is a rabbit
2:29:43 hole
2:29:45 everyone and then the there's even
2:29:47 people that you should follow Ethan
2:29:48 mullik is if you're not following Ethan
2:29:51 mullik then you're not really paying
2:29:54 attention to this a
2:29:55 ipace he's a Wharton business school
2:29:58 professor and he just he posts you know
2:30:01 multiple times a week if not daily about
2:30:04 stuff you should be paying attention to
2:30:06 David Shapiro's single-handedly trying
2:30:09 to open source the AGI movement
2:30:11 artificial general intelligence if
2:30:13 you're a geek and you're not following
2:30:14 him you're missing out and then if
2:30:16 you're in business Paul rer and Rachel
2:30:19 Woods um Paul rer from the marketing AI
2:30:22 Institute and Rachel woods from AI
2:30:24 exchange are are worth following so
2:30:26 anyway that's this list is you know
2:30:31 people in here earlier are like well how
2:30:33 do I get started you start with chat GPT
2:30:35 that's why it's at the top it's why it
2:30:37 says it's still the
2:30:38 best and then I basically
2:30:41 say look for free with chat GPT you get
2:30:44 custom instructions a fast model
2:30:47 and and you get uh conversational chat
2:30:50 GPT in the in the app and then for 20
2:30:52 bucks a month I explain here's all the
2:30:55 [ __ ] it does if you haven't played with
2:30:58 all of these features of chat GPT
2:31:02 4 what are you
2:31:05 doing go down these rabbit holes go down
2:31:08 these rabbit
2:31:10 holes understand what it's like to have
2:31:12 a a chat GPT that's connected to the
2:31:14 internet understand what code
2:31:15 interpreter is understand that chat GPT
2:31:19 can see here I'm going to do this right
2:31:21 now I got to go my voice is [ __ ] shot
2:31:23 I'm I'm still recovering from covid
2:31:25 people but look I'm going to do this for
2:31:26 you because I love
2:31:29 you okay I'm at chat GPT chat gbd4 see
2:31:33 this little paper clip right here I'm
2:31:35 going to click on it and I'm going to go
2:31:37 search on my hard drive for for
2:31:45 whiteboard which one one is it this one
2:31:48 is it no is it this one
2:31:50 yes okay here's a whiteboard is this one
2:31:53 of my whiteboards no wait Kyle Show and
2:31:57 Tell please okay wait hang on we're
2:31:59 gonna get back to whiteboards I have to
2:32:00 go to show and
2:32:02 tell Danielle told me there's something
2:32:04 cool there we're GNA go we're gonna do
2:32:08 it Cindy [ __ ] applied for I officially
2:32:11 on your side sorry Kyle see this is this
2:32:14 is what happens this is what happens
2:32:17 you reluctantly let in the wrong kind of
2:32:19 animal into your zoo and it just it
2:32:27 it and see this is this is I made a
2:32:30 little slip up I said one night that
2:32:32 joker and I were going to bed together
2:32:34 and now I can't unlive this and Joker's
2:32:37 life has been ruined you just what you
2:32:39 know this this is what happens see you
2:32:41 see do you see all right anyway back to
2:32:46 this thing thing okay I found a
2:32:48 whiteboard on the internet I just I just
2:32:50 went to like Google or Bing or one of
2:32:52 those and I said whiteboard and I found
2:32:55 some images and I found this image I put
2:32:56 it there and watch watch what I'm going
2:32:58 to do what is this question
2:33:03 mark so so you know you know how for a
2:33:06 lot of 2023 people are like you should
2:33:08 be a prompt engineer prompt engineering
2:33:10 is what it's about what was my prompt
2:33:12 here what is this this is not prompt
2:33:14 engineering chat GPT you can see it says
2:33:17 this image you've uploaded appears to be
2:33:19 a photo of a whiteboard with handwritten
2:33:21 notes and drawings it seems to be
2:33:22 planning or brainstorming session for a
2:33:24 user interface specifically for an
2:33:27 ordering system at an airport there are
2:33:30 notes about the main steps in the user's
2:33:32 Journey including choosing a language
2:33:34 selecting menu items ordering paying
2:33:36 checking the status of orders there are
2:33:38 also sections on discussing user types
2:33:40 such as business and family users use
2:33:43 cases like takeout and 1hour time limit
2:33:45 and potential features such as flight
2:33:48 status alerts terminal right it it
2:33:50 understands what this is um so I'm going
2:33:52 to say
2:33:58 um that sounds really
2:34:03 interesting I'd like to explore how we
2:34:05 might Market an app like this period why
2:34:08 don't we start with a name give me 20
2:34:10 possible names and taglines for this
2:34:15 app
2:34:19 again that's not prompt engineering
2:34:21 that's like what you would say to
2:34:22 someone if you were in a if you walked
2:34:24 into a room and there were three people
2:34:26 in there and there was this [ __ ] on the
2:34:28 Whiteboard you'd walk in and you go what
2:34:30 is this and they'd tell you well this is
2:34:32 what this is and and then you go oh
2:34:34 that's kind of interesting well if we
2:34:35 were going to Market this what what
2:34:37 might be some names for this and you'd
2:34:40 start a brainstorming session and you
2:34:41 might come up with things like Sky bite
2:34:44 taste the clouds order up high
2:34:47 or jet feast or Arrow eats or terminal
2:34:50 taste that's horrible fly foodie Runway
2:34:53 relish no these are all horrible let's
2:34:55 see um flight Fair
2:34:59 um these are all horrible can we lean in
2:35:02 more to the fact that it's going to be
2:35:04 fast and
2:35:07 convenient so you can just give it
2:35:09 feedback hey marketing team your ideas
2:35:12 are dumb could you give me less dumb
2:35:15 ones you don't have to be mean to your
2:35:17 marketing team like I was like I am here
2:35:21 but quick cart speed serve rapid repast
2:35:25 repast what what are we in the 1870s
2:35:27 come on people flash Feast Sonic snack
2:35:30 turbo taste Express each Dash dine zip
2:35:34 bite instag guest go grub flight fuel
2:35:38 okay flight fuel's it
2:35:40 beautiful perfect I love flight fuel
2:35:44 period let's go with that
2:35:47 now could you create for me an image of
2:35:51 what an ad in the airport might look
2:35:54 like that includes a screenshot of the
2:35:56 app and
2:35:58 incorporates the name And the tagline in
2:36:01 the
2:36:10 poster what's that creating image yes so
2:36:14 now this is one of the one of the
2:36:18 um the most
2:36:22 powerful um parts of having dolly3 image
2:36:26 generation inside chat
2:36:29 GPT is that you can create images in the
2:36:33 context of a
2:36:35 conversation look at that so wait what
2:36:38 was it called what was our thing
2:36:41 called um wait what did I say flight
2:36:45 fuel f fuel up fast what what was flight
2:36:49 fuels fuel up fast satisfy your hunger
2:36:52 at hypers speed so look so look here's
2:36:55 here's a screen here's our airport image
2:36:58 with our screenshot of our app that was
2:37:00 just a whiteboard sketch moments
2:37:05 ago that's what it might look like and
2:37:07 and and then I can give that feedback I
2:37:09 can say oh that's too futuristic or
2:37:10 that's perfect or whatever now I'm gonna
2:37:13 say great now give me
2:37:17 two possible interface
2:37:21 designs for the ordering
2:37:25 page and it should just fire up that
2:37:29 terminal
2:37:32 taste yeah I know listen not all of them
2:37:35 are good but but do do you do you are
2:37:39 you starting to Gro what happens when
2:37:41 you start
2:37:45 to understand what each of the different
2:37:47 modes of chat GPT can
2:37:50 do and now you realize they all work
2:37:53 together so so look now we've
2:37:56 got there there's an inter interface
2:37:59 right now is is that a perfect interface
2:38:01 I don't
2:38:02 know here are two possible um let's
2:38:08 see give me one that's got a dark
2:38:15 background
2:38:19 and focuses more on the speed element of
2:38:24 the
2:38:30 features and I'm going to we're going to
2:38:32 take this one step further because like
2:38:34 I'm not done yet I told you like if
2:38:38 you're not if you're if you're not if
2:38:41 you pay your 20 bucks a month and you're
2:38:43 not actively exploring the depth of what
2:38:46 is in chat GPT right now you're doing
2:38:48 yourself a disservice have you used
2:38:51 GitHub co-pilot I haven't but because
2:38:53 I'm not I'm not an engineer um but but
2:38:56 the engineers that I know that are using
2:38:58 GitHub co-pilot are talking about
2:39:00 anywhere between 30 and 60% productivity
2:39:04 increases all right so this focuses more
2:39:06 on like the speed right you know and
2:39:08 it's it's a darker interface okay great
2:39:10 so now I'm going to
2:39:12 say great this is perfect now I'd like
2:39:15 you to recommend a front end technology
2:39:17 and a backend technology and then Write
2:39:21 the sample
2:39:23 code for the front end and the back
2:39:26 end of this
2:39:29 app and it's probably going to tell me
2:39:31 yeah that's complicated and you know you
2:39:34 really can't do it I can do a little bit
2:39:39 here front end technology react native
2:39:43 backend nodejs with expressjs database
2:39:48 mongodb front end sample react native
2:39:53 code so so we went from here's a
2:39:59 whiteboard sketch or series of sketches
2:40:02 with some
2:40:03 notes to naming this application to
2:40:07 visualizing what that ad might look like
2:40:09 in an
2:40:10 airport
2:40:11 to exploring some interface ideas to
2:40:15 building the code
2:40:18 without trying without really even
2:40:20 paying attention to what it's doing
2:40:24 imagine if you understood these
2:40:27 capabilities and actually
2:40:30 tried actually took some time to solve a
2:40:32 specific
2:40:35 problem you start to get where we're
2:40:39 headed
2:40:41 because right again right now these
2:40:43 tools are janky pieces of [ __ ] [ __ ] that
2:40:46 don't really work but they
2:40:49 work remarkably well right it's just
2:40:52 that right now we as humans have to be
2:40:55 in the loop in all of these stages
2:40:58 because the tools right now aren't
2:41:00 capable of analyzing their own
2:41:03 work so you can do things like tell it
2:41:06 to analyze its own work and it'll help
2:41:08 you make it better but we're moving to a
2:41:11 place where when we get autonomous agent
2:41:14 Behavior which means
2:41:16 I can just give it a goal hey here's a
2:41:21 whiteboard take the ideas on this
2:41:23 whiteboard
2:41:25 and tell me every do everything I need
2:41:29 to do to turn this into a
2:41:32 business and it will start to spin up
2:41:34 its own agents that start to do all
2:41:36 these activities that I just showed you
2:41:38 there'll be a marketing agent that will
2:41:40 then spin up a bunch of marketing
2:41:43 workers for it so so it'll create its
2:41:45 own little CMO and it'll create a bunch
2:41:48 of workers that'll go do all the little
2:41:49 piece parts of the marketing of this
2:41:51 here's the marketing strategy here's the
2:41:53 marketing plan here's the social media
2:41:55 strategy here's the social media content
2:41:57 here's and all these agents are going to
2:42:00 talk to one another and then it's going
2:42:01 to create uh a a a ux committe in team a
2:42:06 project management team an engineering
2:42:10 team a front-end development
2:42:13 team a QA team
2:42:17 a business plan a whole CFO and
2:42:20 financial services
2:42:22 team and it's going to go off and do
2:42:26 it and like an hour later you'll get
2:42:29 you'll get a ping your company's been
2:42:32 started would you would you like us to
2:42:34 submit the app to the App Store now it's
2:42:36 been tested click here to tested in test
2:42:41 flight it's working by the way because
2:42:45 we have had a whole QA team test
2:42:48 it would you like to launch this
2:42:52 business
2:42:56 you that's why you dick around with this
2:43:00 stuff because
2:43:03 when these tools evolve in the way that
2:43:06 I just talked about you're going to know
2:43:10 exactly why it's a big
2:43:13 deal that Sam mman just got up on stage
2:43:17 and said chat GPT now has autonomous
2:43:19 agents in It Go Play you'll know what to
2:43:22 do with
2:43:24 them can we request chat GPT to run this
2:43:27 code and show the demo um not not right
2:43:31 now but you can take this code and throw
2:43:33 it into repet or throw it into your
2:43:34 development tools so but you can you can
2:43:36 absolutely take this code and go run it
2:43:39 it's just right now it it doesn't have
2:43:41 the tools internally if you're if you're
2:43:43 using Code interpreter it it writes and
2:43:46 executes python code within the
2:43:48 container in code interpreter so it it
2:43:50 sort of does a mini version of that
2:43:52 right now but if you're doing things
2:43:54 like react native that then you're going
2:43:55 to have to go use your other development
2:43:57 tools but you can absolutely go and and
2:43:59 you can have chat GPT say hey if I
2:44:02 wanted to run this code where would I
2:44:04 run it and it will tell you you go get
2:44:06 this well I'm on a Mac and I don't want
2:44:08 to run anything on the command line oh
2:44:10 okay then go use this tool download this
2:44:12 do this install this make sure you have
2:44:14 this Library installed and then you can
2:44:16 run it in your browser
2:44:18 window yes you can do all
2:44:22 that out of bounds as far as terms of
2:44:25 service uh try telling it repeatedly
2:44:28 wonder
2:44:29 why anyway all right listen I got to go
2:44:32 people what are you what are you doing
2:44:33 to
2:44:34 me I'm a little
2:44:38 gassy happy New Year
2:44:42 2024 I think this one was from Ted Ted
2:44:45 this one that's pretty good
2:44:47 one all right I just got in what this
2:44:51 code do so what that was hope was I
2:44:54 started with a
2:44:57 whiteboard that a whiteboard photograph
2:45:00 that I downloaded off the internet I
2:45:03 uploaded it to chat GPT and said what is
2:45:05 this it read the Whiteboard knew what it
2:45:09 was and then we developed a name for the
2:45:13 app um a market visualization of what a
2:45:17 a poster in the airport would look like
2:45:19 an interface and the code front end
2:45:23 backend and database for the app all
2:45:26 just based on the picture Kyle go to bed
2:45:30 I
2:45:33 know oh man you can also tell it to
2:45:36 rewrite it in flask and run it there you
2:45:38 go wow that's so cool yeah no it was
2:45:40 absolutely crazy all right um so do me a
2:45:44 favor go there the salon. check out the
2:45:48 stuff we did with the GPT for good
2:45:50 project this past weekend go look at the
2:45:52 150 gpts that this community made in 24
2:45:57 hours the first link there is that
2:45:59 directory and then the fourth link is
2:46:01 the AI Salon go check that
2:46:04 out all right I'm out of here people
2:46:06 it's late it's
2:46:08 tired it's
2:46:10 Wednesday
2:46:11 um peace out thanks for all you do
2:46:14 thanks for the cont contributions that
2:46:16 you make uh please follow me on this
2:46:19 channel join the salon it's free um if
2:46:24 you want to buy in one of the corners is
2:46:26 a
2:46:27 uh video series that can support the
2:46:30 channel anything you do support the
2:46:31 channel is cool um but if not the
2:46:34 biggest support for this channel is keep
2:46:35 showing up here show up here show up
2:46:38 show up show up show up show up why
2:46:41 would we come back here you're kind of a
2:46:43 dumb dumb what are your qualifications
2:46:46 just keep showing
2:46:48 up and get curious about this stuff
2:46:52 because it's it's not
2:46:54 coming AI is not coming it's
2:46:59 here it's
2:47:02 here
2:47:06 and this
2:47:08 community and ones like
2:47:11 it are going to in in the next 18 months
2:47:14 if you just stay connected to this
2:47:16 community remarkable things are going to
2:47:19 happen for all of us
2:47:21 watch watch all right peace everybody to
2:47:26 the mods thank you so much you did a
2:47:27 great job tonight because I I was I know
2:47:29 I was kind of ranting and rambling but I
2:47:31 was able to see like all of the the
2:47:34 relevant stuff that I could tell so so
2:47:36 thank you for the work you do thanks for
2:47:38 my subscribers thank you for the
2:47:39 Irregulars thank you for everyone
2:47:41 sharing crazy ass raccoon images support
2:47:45 ing Cindy [ __ ] against my best wishes
2:47:48 which is crazy to me but whatever fine
2:47:51 on
2:47:52 you uh the raccoons are cute I'll give
2:47:55 you that um all right just me and the
2:47:57 trees thanks for hanging out digital
2:47:58 Gods Lord digital Gods techy Becky good
2:48:01 to see you sorry about about the about
2:48:04 your husband and uh thinking that you're
2:48:07 spending too much time here but you just
2:48:09 got to tell him man it's this is it's
2:48:11 for an important cause Sam cor good to
2:48:13 see you dragonfly Alchemy Cory s Pottery
2:48:16 jrc wow got a lot of the lot of the
2:48:18 classics in here Emilio's wife was on
2:48:20 time tonight ASU good to see you Cindy
2:48:24 [ __ ] lovely seeing you as always Joker
2:48:27 Apple user a human being man got a pile
2:48:30 of people Clint copperhood header of
2:48:32 course forever hooked dag Rudy thank you
2:48:37 thank you thank you sir from South
2:48:39 Africa our biggest South African fan in
2:48:42 the house tonight Robert Rossy in the
2:48:44 house tonight
2:48:46 wow that's big crowd big big crowd
2:48:48 really cool yeah we were up to 200
2:48:50 people there for a while amazing there's
2:48:52 160 right now peace out everybody I'll
2:48:54 see you
2:48:54 tomorrow be here or be not here that
2:48:59 those are your two options peace