aiLL 06.18.23 - Unlocking Creativity and Potential with Generative AI

Description

Chapters

Transcript

0:01 foreign
0:16 Happy Father's Day
0:19 welcome to the AI learning lab
0:24 oh
0:26 a little little uh discombobulated here
0:28 getting my getting my act together
0:31 hope everyone's doing well happy Sunday
0:33 Happy Father's Day
0:35 the boys got me some nice Weller bourbon
0:39 so having a little sip of that
0:44 um
0:45 all right now this is the AI learning
0:47 lab welcome everybody uh we'll wait for
0:50 some others to join my name is Kyle
0:52 Shannon
0:53 I'm a an entrepreneur
0:56 and a uh founder and CEO of a company
0:58 called storyvine and hello Lily
1:03 oh my God hello
1:07 I don't know if that's if it's good or
1:08 bad
1:11 um I uh
1:14 this channel is all about demystifying
1:16 AI answering questions
1:18 brainstorming showing stuff if I know
1:20 how to do it uh making stuff up if I
1:22 don't just like chat GPT
1:25 mansplaining is a service oh Lily oh you
1:28 joined on Discord I saw that welcome
1:30 welcome
1:31 have fun geek out in there ask questions
1:32 make pictures with mid-journey
1:36 um
1:36 what else
1:38 um
1:39 um so so one thing before if you've got
1:41 questions about about AI pop them down
1:43 below
1:44 I'll do my best to answer them that's
1:46 that's what I do here I try to get on
1:47 here every night uh in in lieu of I used
1:50 to I used to sim race every night at Sim
1:52 race for a couple hours and then I
1:54 thought you know this this AI stuff's
1:56 really important so why don't I replace
1:58 SIM racing with um with getting on here
2:01 and and talking about this so if you
2:02 have questions about AI or or thoughts
2:05 uh pop it down below
2:07 um one thing that I that I I like to
2:09 talk about so in the mid 90s I was I was
2:12 there for the uh the early days of the
2:14 World Wide Web started one of the first
2:15 digital magazines uh called Urban
2:18 desires and and co-founded agency.com
2:21 which is one of the early digital
2:22 agencies and I was the chief creative
2:24 officer there and and built a lot of the
2:26 early websites for the Fortune 500 and
2:28 felt like incredibly lucky to be part of
2:32 you know a transformative technology
2:35 right there at the beginning and
2:37 um and I didn't think I'd see another
2:39 one in my life and here we are I think
2:41 this AI stuff is is you know as probably
2:45 more significant than the World Wide Web
2:47 and I think it's coming a lot faster
2:49 than people are prepared for so that's
2:51 why this channel exists if you're new to
2:53 Ai and haven't played with chat GPT
2:55 that's what these URLs are behind me
2:58 chat.openai.com that's the official
3:00 chatgpt website so while you hear me
3:02 Babble on
3:04 um go there and start playing with it
3:05 just start with what I call parlor
3:06 tricks have it write you a poem or a
3:08 song or uh I don't know explain quantum
3:12 physics to you in an Eminem rap and just
3:15 just start to get your hands around what
3:17 it can do bing.com is also chat gbt but
3:20 it's at Microsoft it's at their search
3:22 engine and that's gpt4 that you can use
3:24 for free and it's connected to the
3:26 internet and then poe.com is also you
3:29 can use chat EPT there but there's a
3:31 bunch of other models you can play with
3:33 um prompts.chat will teach you about
3:35 prompting and then futurepedia.io is a
3:37 big directory of
3:39 um
3:40 of just AI tools like lots of the music
3:43 tools and animation tools and picture
3:45 tools and word tools and all that sort
3:47 of stuff so
3:50 um
3:51 pop any questions you have below uh tap
3:54 the the like button initially that'll
3:56 let people know we're here I would
3:57 appreciate that to get the word out
4:00 and let's jump in let me let me go back
4:02 and see if there's any questions here
4:03 all right let's see I have so many
4:06 questions but my brain is fried learning
4:08 my oh learning python today I thought
4:11 you said learning Monty Python today
4:13 None Shall Pass
4:15 it's just a flesh wound uh learning
4:19 python today yes
4:21 um and also I I hope you had chat uh
4:23 gpt4 help you learn the python I hope
4:26 that's where you were learning it and
4:27 and uh our pair programming with with
4:31 GPT where where it becomes your partner
4:33 and gets you up to speed much much more
4:35 quickly but yeah please feel free free
4:36 to ask away
4:39 um hello hello hello is chat gpt4 access
4:42 current data so
4:45 the base model of gpt3 3.5 and 4 all
4:50 ends at September 2021 but the
4:56 um
4:57 gpt4 that's connected to the internet
4:59 and Bing that is also connected to the
5:01 internet kind of fills that Gap right so
5:04 when you do a search if you search for
5:06 something more recent than September
5:07 2021 it'll actually do an internet
5:09 search a Bing search apparently and it
5:13 will augment the answer with with
5:15 relevant data so yeah it effectively is
5:18 more current than that the base model is
5:20 not the base model still stops at
5:22 September 2021 that's just the nature of
5:24 how the those core base models are
5:27 trained it takes a long time and a lot
5:29 of money and a lot of people and a lot
5:31 of data entry
5:33 um to get it there so that's why there's
5:35 that cut off but yeah having it
5:36 connected to the internet helps
5:39 hello hello hello
5:42 um if you're new here my name is Kyle
5:43 Shannon I'm an entrepreneur I I'm CEO of
5:47 a company called storyvine I um
5:50 uh story finds an 11 year old video
5:52 storytelling platform and
5:55 um
5:55 we do automated video production though
5:58 ironically we don't use AI to do the
6:00 automation
6:01 um we figured out some other ways to do
6:03 that we actually use more Aristotle than
6:05 AI I'll leave I'll leave that as the
6:07 hand but
6:09 um I'm completely rethinking the product
6:10 and we're incorporating AI into the
6:12 platform so fascinating times to be
6:15 doing anything right now it's absolutely
6:18 wild so if you have any questions about
6:20 AI pop them down below if you want to
6:22 feel free to share the live so other
6:24 people can see it what's the best way to
6:26 generate
6:27 AI photography with your face so there
6:32 are there are a number of ways to do it
6:35 um
6:39 so so so let me give you let me give you
6:42 a couple of uh a couple of options one
6:46 is you can use mid-journey and they've
6:48 got a thing called blend and you can
6:50 also just upload your picture to Mid
6:51 Journey I'll show you how to do that in
6:53 a second mid journey is a bit of a pain
6:55 because you have to do it in Discord I
6:58 don't think that's going to be there for
6:59 too much longer but it is right now
7:02 also with dream Booth there's a thing
7:05 called stable diffusion there's another
7:06 thing called control net so if you're
7:08 geeky I would say stable diffusion gives
7:11 you more control
7:14 um and yeah that's probably the best way
7:17 um so let's let me let me I'm gonna flip
7:20 the camera here
7:22 and turn off my green screen
7:25 hold please
7:28 okay
7:30 so I'm gonna flip over to Discord
7:35 if you haven't used Discord let me let
7:37 me do a little Discord primer so so
7:40 Discord if you view slack at work
7:42 Discord is very much like that kind of
7:44 started in the gamer Community then it
7:46 got really popular in blockchain and nft
7:49 communities and mid-journey has has
7:52 adopted it so down the left hand side of
7:56 the screen here you see these circles
7:57 those are Discord servers so those are
7:59 different communities
8:02 um and mid Journey has one but I've also
8:03 got one here for the AI Salon this is
8:06 this is the salon that that I created
8:08 that if you want to
8:10 um if you want to hang out with a
8:12 community of people geeking out on AI
8:14 join the AI salon so go to the salon.ai
8:17 and and within Discord all of these
8:21 things are conversations and one of our
8:23 little channels here is the mid Journey
8:25 bot and so the mid Journey bot is where
8:27 you can create things so
8:30 the way you create a mid-journey image
8:32 just in general is you go slash imagine
8:36 and then it pops open a prompt and you
8:38 can say something like I don't know
8:40 super let's say
8:42 um let's say cartoon
8:45 of oh wait sorry hang on hold please
8:50 cartoon
8:52 of
8:54 middle aged
8:58 superhero
9:03 in Tights
9:06 and I'll just hit that and so that
9:07 that's going to send that off okay so
9:10 that's that's in general how you make an
9:12 image in mid Journey it's going to go
9:13 off it's going to make me five of those
9:14 I didn't really describe it very well at
9:17 all but just so you get the the sound so
9:19 now it's it's gonna start making some
9:21 superhero images and they always start
9:23 blurry and then they get you know over I
9:26 don't know
9:26 50 or so Cycles it gets more and more
9:30 clear that's how the how the diffusion
9:32 technology works
9:33 so one thing I can do is I can go to my
9:36 finder and let me go find an image of
9:40 myself here
9:45 oh that's not the one I needed
9:49 hold
9:50 [Applause]
9:51 please
10:00 okay there's some there's some images of
10:02 me so that one's fine I look a little
10:06 dorky but you know it comes to the
10:08 territory all right so basically what I
10:10 did I just dragged it from the finder or
10:13 I don't know whatever it's called in
10:14 Windows the windows equivalent of that I
10:15 just dragged a picture into the the
10:17 mid-journey message thing and it's there
10:19 and so I hit return
10:21 and now that uploaded that photo into
10:24 Discord and now what I do is I I right
10:27 click and say copy link so I'm copying
10:30 the URL of this image and now I say
10:33 slash imagine like I did before
10:36 and instead of typing what I want the
10:39 prompt to be I actually paste in that
10:42 URL of that image and then
10:45 I put in my prompt so let's go back up
10:48 so there's the heroes that it made just
10:49 in general
10:51 and I'm gonna go grab that same prompt
10:53 cartoon of middle age superhero in
10:56 Tights
10:57 and I'm going to put that in after that
11:00 URL so I've got the URL of the photo of
11:03 me and then I put in cartoon of a
11:05 middle-aged superhero and tights
11:07 and now it's going to take that image as
11:10 the inspiration and generate it so
11:13 um if I made that a photograph instead
11:15 of a cartoon then it would be a
11:16 photograph but these should be
11:17 superheroes that look more like me so
11:20 let's go look at the the ones that were
11:22 just generic
11:24 these are some pretty old looking
11:25 middle-aged people not for nothing
11:30 let's go zoom in on these bad boys
11:33 I mean I mean come on Grandpa here that
11:36 dude's not middle age he's wearing
11:38 diapers
11:40 that looks like Ed Asner
11:42 that worked out that one's sort of
11:45 middle-aged but you know that's that's
11:47 the one that looks like a superhero the
11:48 other ones are pretty funny all right
11:51 let's go see if if anything's happening
11:53 down below all right yeah there we go
11:54 you can already start to see
11:56 [Laughter]
12:01 dumb face in there
12:10 oh
12:11 yeah this is fantastically awkward
12:15 [Laughter]
12:17 that did it on the lower right huh
12:20 all right let's go look at these nice
12:22 all right we'll open these in the
12:24 browser so we can zoom in okay that
12:27 one's just offensive that's not what I
12:28 look like
12:29 that one's okay sort of looks like me
12:32 that one looks a decent amount like me
12:34 although the the body I look like
12:36 Jacqueline if you remember him if you're
12:39 gen xers and remember Jack LaLanne it's
12:41 like my sort of face on Jacqueline's
12:44 body
12:44 so anyway so that's how you do that so
12:47 so I hope that helps all right let me
12:49 flip the camera back here
12:51 and and turn on my little green screen
12:55 looks like we have some new folks in
12:56 here welcome everybody if you're new
12:59 here my name is Kyle Shannon this is the
13:01 AI learning lab
13:03 these URLs behind me I have them here if
13:06 you haven't tried chat EBT yet
13:08 um go to that top URL that's
13:11 chat.openai.com that's the official
13:13 website of chatbt if you haven't played
13:15 with it go play with it bing.com is
13:17 Microsoft's version of that if you click
13:19 on the chat button at bing.com and then
13:22 po.com is one that lets you try a lot of
13:24 different models and these two down
13:26 below are just resources that I think
13:28 are interesting I'm not affiliated with
13:30 any of these things in any way I just
13:31 think that they're
13:33 um if you haven't tried it do it and
13:35 those are the official websites chatgpt
13:37 is the thing that I think has has
13:40 created an inflection point that is
13:42 going to change lots and lots of things
13:44 so if you've got any questions or
13:47 thoughts about uh generative AI or AI in
13:50 general some say AI is a demon they do
13:52 and then some say it is a savior and
13:55 some say it's a tool I'm in that ladder
13:58 Camp
13:59 um I don't I don't it doesn't deserve as
14:02 much power as a demon right now it is
14:04 certainly a reflection of humanity so to
14:06 the extent that humanity is has dark
14:09 sides it reflects that
14:11 mm-hmm
14:13 drinking a little Weller that my boys
14:15 got me for Father's Day Happy Father's
14:17 Day to any fathers out there
14:19 yes yes yes I'm attached to wait sorry
14:23 what did that say oh tick tock tick tock
14:26 just jumped me to the bottom let me go
14:27 back and figure out where I was
14:32 [Music]
14:34 2001 A Space Odyssey I just saw that
14:38 hang on
14:39 I joined your Discord okay here we are
14:42 let's chat access to current data we did
14:45 that photography of the face
14:48 I'm attached to chat GPT by The Hip I
14:51 know what is what is GPT so okay so
14:55 so
14:56 two answers to the question if you're
14:58 asking what does the acronym mean it
15:00 stands for generative pre-trained
15:02 Transformer which is some geeky words
15:05 that's some geeky ass words but it
15:07 stands for generative meaning it
15:09 generates generates original content
15:13 pre-trained means that it's a machine
15:16 learning model that you don't have to
15:18 bring your own data it's been trained on
15:19 lots and lots of data like all of the
15:21 public internet and then Transformer is
15:24 the technology that let it be trained on
15:26 all that data effectively
15:29 um that's what it means now
15:31 how does it actually work let me I will
15:34 flip my camera back over
15:37 and turn off
15:40 this little screen share and let's go
15:42 play I'll go show you quickly
15:45 what chat CPT is so basically when you
15:48 get to chat GPT you have kind of a
15:50 history down here you've got your
15:51 settings down at the bottom
15:53 and then it's really just this blank
15:55 page it's not dissimilar to a Google
15:57 search where you have a
16:00 an entry window at the bottom
16:02 and then you can choose different
16:03 versions of it if you're using the free
16:05 version you just have access to GPT 3.5
16:08 I think they call it GPT turbo or
16:10 something like that when you don't pay
16:11 for it
16:13 um and then what you do is is you
16:14 basically ask it things and and and have
16:17 it do things and and it it it
16:20 the depth of knowledge that it has the
16:23 depth of what it can do is is mind
16:25 numbing and face melting but you can
16:28 have it do silly things like
16:31 um write me a country song
16:37 um like Garth Brooks
16:41 about I don't know a
16:44 cat that jumped
16:48 out of a pickup truck
16:53 on a Dusty Road
16:59 and it will write you a song and it will
17:01 do it very very quickly and it tends to
17:03 be pretty good even gbt 3.5 gpt4 is
17:07 better
17:08 uh the sun was high on a Dusty Road in
17:11 my old pickup truck I was carrying my
17:12 load radio playing some good old Tunes
17:14 just another day under the big Skies
17:16 Moon
17:18 and then out of nowhere in a blur of
17:20 gray a little cat jumped out and stole
17:22 my heart away so there you go so there's
17:25 your country song
17:26 um you could say
17:28 um give
17:30 it's some chords so one of the
17:33 interesting things about
17:35 um GPT chats is that it's got memory of
17:39 the things that you've talked about
17:40 before so when I said Give It chords it
17:43 understood that that meant give the song
17:45 chords and so now it it gave me a
17:47 classic gcd you know the old classic
17:50 three chord Country Song chords which is
17:53 cool now I can say
17:55 um
17:56 make a table comparing
18:00 um
18:02 four
18:04 different music genres
18:10 sure here's a table comparing four
18:12 different music genres right so pop
18:15 characteristics popular artists iconic
18:18 songs pop rock hip-hop country
18:21 um you can also do things
18:23 um
18:24 like if you want to Learn Python like
18:26 Lily who's in here who said she's she's
18:29 been deep diving on python you could say
18:31 write me a python
18:35 application
18:39 that measures the wave length
18:45 of light coming from a sensor I don't
18:48 know
18:49 I don't even know if that's possible
18:51 sure that's possible it can write it and
18:52 then it writes you the code so so
18:57 chat gbt basically allows you to to sort
19:01 of query against all of this deep
19:03 knowledge and you can start to do some
19:05 really interesting things like start to
19:07 solve problems with it you know for
19:09 example
19:11 um maybe you're trying to do a marketing
19:12 plan for your business you could say hey
19:14 act like a marketer and you know here's
19:17 what my business does and here's the
19:19 challenge I'm facing help me come up
19:20 with a marketing plan if I don't have a
19:22 lot of budget or if I do have budget or
19:24 whatever it might be and then it'll give
19:26 you a marketing plan then you can say oh
19:27 that's great can you help me write some
19:29 copy for some banner ads and then it'll
19:31 do that and and it literally kind of
19:34 anywhere you aim it it shape shifts into
19:36 the thing you're trying to do it's got
19:39 some um interesting challenges like it
19:41 will lie to you confidently they call it
19:44 hallucinating which is which is a pretty
19:46 funny term
19:48 um but anyway that's that's chat's apt
19:50 so if I hope that answered your question
19:52 can you build me a Salesman bot
19:55 um possibly I I mean what I would
19:57 encourage you to do is is do one on your
20:00 own and I don't quite know what you mean
20:01 by a Salesman bot meaning
20:04 um if someone asks it something it
20:06 always turns it around and says here's
20:07 my solution
20:09 um give me a little bit more context on
20:11 that and I can show you how we can
20:13 create a bot really quickly in Poe
20:16 um like I've created uh Shakespeare Bots
20:18 and I created a Yoda bot and a Jay-Z bot
20:22 and a Steve Jobs bot that just talked
20:24 like those people the salesman one could
20:27 be could be interesting could be
20:28 specific I can at least show you how to
20:30 do it and then you can go play
20:32 um do you have an Instagram too I've got
20:34 from the this image that's behind all
20:36 these URLs is from an art project they
20:38 did called Kyle Shannon dreams so I have
20:42 my Kyle Shannon at Kyle Shannon
20:45 Instagram is my photography so I'm a
20:47 photographer as well and then Kyle
20:50 Shannon dreams is my art project that's
20:52 got all my stable diffusion wacky self
20:55 portraits with little stories tied to
20:57 them and then also there's a tick tock
20:59 Channel called at Kyle Shannon dreams
21:01 that's me talking underneath those
21:03 characters as well but I haven't updated
21:05 that in a while anyway cheers Happy
21:08 Father's Day to uh to all your fathers
21:10 out there
21:13 or those with a father
21:16 um can you prompt chat GPT to compose
21:18 music from current real-time Schumann
21:20 resonance chart well right now it can't
21:23 see I mean if you haven't seen the
21:25 Schumann resonance stuff like I don't
21:27 know what the hell's going on with the
21:28 Schumann resonance but it all looks like
21:30 DNA today so if you haven't seen that go
21:32 to Twitter and type in Schumann
21:34 resonance and uh apparently apparently
21:38 we're entering another dimension or you
21:40 know time traveling or something who
21:43 knows maybe we really are in a
21:45 simulation that'd be trippy I mean all
21:48 of a sudden everything goes we just yeah
21:53 someone someone hacked the simulation
21:58 oh but but the question is um
22:02 you could if you took the data of the
22:05 Schumann resonance thing
22:07 you could have it compose music to it I
22:09 mean I don't know let's see
22:12 here let me go show you I'll go show you
22:14 this because because it's at least
22:15 interesting to look at
22:18 um
22:20 all right so we'll go to the Twitter
22:24 and we will go
22:26 shoe man
22:29 resonance
22:33 all right so
22:35 this crap up here
22:38 the stuff to the left is how this thing
22:41 normally looks I guess this is the
22:42 Schumann resonance and the Schumann
22:44 resonance is the
22:46 um frequency of the Resonance of Earth
22:48 and so it normally looks something like
22:50 that and then for the past 36 hours I
22:53 think
22:54 it looks much like human DNA so there's
22:58 some [ __ ] going down now I don't know I
23:01 don't know enough about the data you
23:03 could take the data from that and then
23:04 you could probably pop the data into
23:06 chat gbt and and say
23:10 um you know write write me a song
23:13 based on that but let's let's go over
23:15 here to
23:18 so so I just selected browse with Bing
23:21 so there's three different models if if
23:23 you pay for chat GPT plus you get access
23:25 to gpt4
23:29 um default you also get access to
23:33 gpt4 connected to the internet and then
23:36 you get access to plugins
23:38 so let's let's say um
23:41 the past 30 36 hours the Schumann
23:47 resonance
23:50 has looked like
23:53 human DNA
23:56 can you
23:58 compose a song in ABC format ABC format
24:04 is a
24:06 basically like a text formatting of
24:11 music that that there are players for so
24:13 can you compose a song in ABC format
24:17 um inspired
24:20 by the
24:23 anomalous data let's see if it goes off
24:26 and finds the data
24:29 the Schumann resonance is a global
24:31 electromagnetic resonance excited by
24:33 lightning discharges in the cavity
24:35 formed by Earth's surface and the
24:37 ionosphere they tend to Peak at certain
24:39 frequencies
24:41 well those resonances have complex
24:43 patterns it's quite a leap to say that
24:45 they look like DNA well they do the the
24:48 chart does however there's an AI can try
24:51 to interpret it
24:53 ABC notation
24:55 that's a very simple song there may be
24:57 this key is in C major
25:00 um can you make it in a minor
25:06 um
25:09 let's see what it does
25:12 sure so now it's going to write that
25:14 it's called resonances and Helix in a
25:17 minor
25:20 all right and then what I'm going to do
25:21 is I'm going to say where can I play
25:25 ABC Music
25:28 on the web
25:30 or when
25:36 so now I'm going to go copy all this
25:40 copy
25:42 all right you can use
25:46 this player
25:49 it actually gave us a player we're gonna
25:51 go paste so there's the song it wrote
25:54 it's just all quarter notes which is
25:56 boring and shitty
25:59 [Music]
26:03 okay that's horrible so one of the
26:05 things you can do with with uh geez
26:08 that's bad
26:10 um with gbts you can tell it it's bad I
26:12 say okay that song
26:15 was bad
26:18 and just a linear scale of quarter notes
26:25 please make something more interesting
26:30 musically musically
26:36 and it should apologize I apologize if
26:38 the previous song didn't meet your
26:40 expectations
26:43 let's try something a bit more complex
26:45 and interesting yes please shall we all
26:48 right so now it's doing some stuff let's
26:50 see it's going slow too
26:53 it's really thinking about this
26:55 or it's just slow maybe there's a lot of
26:57 people making Schumann resonance songs
26:59 right now all right this piece is
27:02 written in a minor and includes several
27:04 chords giving it harmonic context all
27:07 right let's go let's go uh listen to
27:10 this
27:12 wait hold great copy code okay
27:16 I'm gonna go back over my little player
27:19 paste all right now
27:21 it still looks boring as [ __ ] but let's
27:24 see
27:27 okay so the answer to your question is
27:30 yes it can write a song The the if the
27:33 question was can I write a good song the
27:35 answer is no no it probably could if I
27:36 spent some time if I spend some time
27:38 messing with it I could get something
27:40 better than that that was horrible
27:41 though
27:42 um if one of the things that that uh
27:46 happens with chat gbt by the way it
27:48 looks like we got some new folks in here
27:50 my name is Kyle Shannon this is the AI
27:51 learning lab uh feel free to ask
27:53 questions about AI down below I can give
27:55 you my background at some point if that
27:57 makes sense
27:58 um and I like to to answer questions
28:00 show things like I can
28:02 um
28:04 the the the prompt that I gave chat EPT
28:08 in that case was particularly weak and
28:10 it obliged and gave us a particularly
28:13 weak response like that was just a
28:14 crappy song
28:16 um there there are probably ways you
28:18 could prompt that like tell it to act
28:20 like a classical composer that has you
28:23 know great knowledge of Music Theory and
28:25 there's there's ways that you could you
28:27 know tweak The Prompt and part of the
28:29 art of Chachi BT is starting they call
28:31 it prompt engineering or prompt crafting
28:33 but part of the art is really
28:35 understanding the back and forth with
28:36 the machine and and getting good at
28:38 having the conversation with it but
28:40 anyway uh the the short answer to your
28:42 question is yes
28:44 um it's not really connected I can't
28:45 take that image of the Schumann
28:48 resonance and put that in yet although
28:49 that's coming
28:51 um gpt4 is multimodal it's just not live
28:54 yet multimodal meaning you can upload a
28:57 picture and have it interpret that
28:59 picture
29:01 by the way happy uh Father's Day to
29:05 all the fathers out there what is the
29:07 best free AI to use Okay the reason I
29:11 have these URLs here
29:12 is they all have free options
29:15 um chat.openi.com is the official chat
29:18 gbt website in my opinion that's the
29:21 best even the free version that's GPT
29:23 3.5 if you pay for it you get gpt4 and
29:26 you get plug-ins and you get some other
29:28 stuff that's coming
29:30 bing.com is Microsoft's if you it's
29:33 their search engine but if you click on
29:34 the chat button that's gpt4 for free and
29:37 it's connected to the internet
29:39 and then po.com lets you create Bots and
29:42 play with a bunch of different large
29:44 language models and their paid
29:46 subscription gives you access to some
29:48 super cool stuff too so that's why these
29:50 are there the these are ones that I
29:51 think are if you're going to start start
29:53 with this stuff if you're really into
29:55 images start with mid Journey probably
29:57 or adobe Firefly
29:59 uh and then these two things below the
30:02 bottom one futurepedia.io is just a
30:04 giant directory of AI tools if you want
30:06 to run down
30:08 a rabbit hole a day for the next year
30:10 you probably couldn't get through all of
30:13 what's there and it's you know it's
30:14 growing every day and then prompts.chat
30:16 will teach you about prompting that idea
30:18 of how do you prompt these uh large
30:21 language models to give you interesting
30:22 things that's what those are about all
30:25 right best stock trading platforms I can
30:27 use AI Bots to bid with
30:30 I so I would go ask chat GPT that for
30:34 one thing I don't I don't know a ton
30:35 about trading but you can also any
30:37 trading platform that has Bots you can
30:39 copy the code of the bot put it into
30:42 chat EBT and then have it give you a new
30:44 bot that that focuses on different
30:46 triggers or you can take two different
30:48 Bots and merge them together
30:50 um so you can you can have chat GPT act
30:54 as your
30:56 um an educator for you
30:59 um and a partner helping you do the the
31:01 trading stuff
31:03 2001 Space Odyssey mentions AI I'm sorry
31:06 Dave I'm afraid I can't do that yes it
31:09 does it's it's it's one of the original
31:11 sin sort of AI movies where the the uh
31:15 the the robots are gonna kill us right
31:17 so so one of the things going on in
31:18 society right now with with this AI
31:20 stuff is that people that haven't used
31:23 it there tends to be a default of fear
31:25 that these things are powerful we don't
31:28 understand them they're scary as [ __ ]
31:29 and by the way Hollywood for the past 50
31:31 years has told us that they're going to
31:33 kill us
31:34 and so so people are either dismissive
31:38 of it of like oh yeah they said
31:40 blockchain and nfts are going to be big
31:42 but those are stupid and now you're
31:44 saying and then Mark Zuckerberg said to
31:46 metaverse was going to be big and that
31:47 was stupid and now you're saying ai's
31:49 gonna be big uh yeah I am
31:52 I'm saying ai's the one the metaverse I
31:55 could never figure out it always seemed
31:57 like you know
31:58 um it's still why does it still look
32:00 like second life from like 2000
32:04 why right
32:05 and if you notice what Apple did they
32:07 didn't mention the word metaverse once
32:10 um and they created a pretty remarkable
32:11 piece of hardware and and redefined the
32:13 category of spatial computing
32:16 um but it's but it's a little scary from
32:18 the outside and and it's easy to dismiss
32:20 it but one of the interesting
32:22 um things about
32:25 um generative AI chat GPT in particular
32:27 and and why this channel exists why I
32:30 sort of Point people there
32:32 so I was a I was a
32:34 um I guess I hate calling myself an
32:37 innovator I was there at the early days
32:39 of the world wide web in the mid 90s I
32:40 started one of the first digital
32:41 agencies
32:42 and we all know how transformative that
32:45 was it changed like everything about how
32:47 we communicate about how we date about
32:48 how we do business about how just
32:50 everything right
32:52 to get to 100 million users the World
32:54 Wide Web took six years it took six
32:57 years to go from zero people using the
32:59 World Wide Web to 100 million
33:01 chat GPT got to 100 million users in six
33:04 weeks
33:06 six weeks to 100 million users fastest
33:09 adoption of technology in history it's
33:12 also profoundly more powerful right you
33:15 know it can write you a novel it can
33:17 write you a screenplay it can write you
33:19 a marketing plan it can do programming
33:21 for you it can do lots and lots of
33:23 things now does it do them perfectly no
33:25 does it hallucinate and lie and give you
33:28 bad answers a fair amount of the time
33:30 yes
33:31 um is it still the most remarkable piece
33:33 of technology ever put in the hands of
33:36 man yes
33:37 and what Chachi BT did the inflection
33:39 point why this channel exists
33:42 the inflection point is it's not just
33:44 the purview of scientists and
33:46 researchers anymore Chachi BT put it in
33:49 this simple interface that's like as
33:51 simple as a Google search box
33:53 that the rest of us get to play with it
33:56 um and so it's going to affect
33:57 everything from you know copywriting to
34:00 data analytics to you know illustration
34:03 to animation to filmmaking and
34:06 everything in between
34:08 so there's some thoughts oh it looks
34:11 like we have some new people here my
34:12 name is Kyle Shannon this is the AI
34:14 learning lab welcome everyone if you're
34:16 new to the space and haven't played with
34:18 chat GPT that's what that URL is if
34:20 you've got questions about AI pop them
34:23 in the questions below I am it's it now
34:26 says I have 99 Plus messages so I am
34:28 painfully behind in the messages so if
34:30 it seems like I'm not answering your
34:32 questions that's because I just haven't
34:34 gotten to them yet and I really do try
34:36 to answer them all I believe I have
34:39 spoken with you once before cool welcome
34:41 back thank you I really want to see it
34:43 I'm not sure what that was in was in
34:45 relationship too but tell me what it is
34:47 and I'll go look can you create Flyers
34:49 with AI sure
34:51 um and and you're not you don't even
34:53 necessarily need to do it with AI
34:55 um but if you go to canva you can just
34:58 go uh canva.com go to their flyer
35:01 section and there'll be a bunch of
35:03 designs there if you're talking about
35:04 writing the copy absolutely you can say
35:07 in fact you could say to check gbt you
35:11 know I want you to act as a a marketing
35:13 copywriter and I want to do flyers for
35:16 my band or for whatever it is my yard
35:18 sale
35:19 um can you write the copy give me 10
35:22 different Flyers
35:24 um and and what what images should be on
35:26 them and it'll write all that copy for
35:28 you you can pick the one you like best
35:29 you can tell it if it didn't get them
35:31 right you can tell it make it funnier
35:32 make it more serious make it whatever
35:34 and then you can just take that copy
35:36 over to canva paste it into a template
35:38 hit the export button and you're Off to
35:40 the Races
35:41 2001 Space Odyssey was 1979 I believe I
35:44 think that's right I think that's right
35:48 I did oh you wanna see something cool of
35:51 course you do
35:52 here we go
35:54 Lee will go check out something cool
35:57 so one of the one of the
36:00 um
36:01 sites that I have up there in my in my
36:03 little background thing is this site
36:05 called Poe
36:08 and so when you get to Poe pose just
36:10 like any other it's like chat GPT it's
36:14 got you know a thing on the bottom so
36:15 you can say
36:18 um make me uh
36:22 a flyer
36:26 about
36:28 uh let's see my band
36:32 um thin
36:34 the herd and make it exciting
36:40 I don't know whatever sure I can help
36:43 you make that send the herd bringing
36:44 heat to your ears join us for an epic
36:46 night of rock and roll as thin the herd
36:48 takes the stage these seasoned musicians
36:51 will blow your mind with their
36:53 electrifying sound and Unforgettable for
36:55 okay so it's just like just like a
36:57 normal chatbot right but
37:00 um down the left hand side here there's
37:03 a bunch of different options so Poe is
37:05 the or Sage is the
37:09 the the chat GPT of of Poe this is uh
37:12 quora is the company behind Poe and then
37:14 you got gpt4 Cloud plus Cloud instant
37:17 100K that's what I'm going to show you
37:18 in a second Claude instant and chat GPT
37:20 and then you can make Bots down below
37:24 um gpt4 Cloud plus and Cloud instant
37:26 100K you have to pay for but if you pay
37:29 for them this Claude instant 100K is
37:32 remarkable
37:33 so
37:34 I posted in
37:37 all of this text here
37:39 is the 2001 Space Odyssey screenplay so
37:44 I went to a website and I I found the
37:47 2001 screenplay and then what the 100K
37:51 means what this 100K means is that means
37:53 a hundred thousand tokens or or 75 000
37:56 words it means you can paste in up to 75
38:00 000 words in your chat window and start
38:01 interacting with it chat GPT by
38:04 comparison allows you to to do about
38:06 three thousand words
38:08 so I pasted in the entire 2001 Space
38:11 Odyssey screenplay and then I asked who
38:13 are the main characters and what happens
38:14 to them
38:16 Dave Bowman he's the astronaut in the
38:18 discovery mission to Saturn Frank pool
38:19 is his crewmate the HAL 9000 computer
38:21 starts malfunctioning and predicts
38:23 failures of equipment that turn out to
38:25 be fine Hal eventually lies about an
38:27 antenna unit falling off and ends up
38:29 killing Frank Poole
38:30 Dave disconnects house higher functions
38:33 and continues the mission in the zone
38:34 Okay so
38:36 this is really remarkable if you've used
38:38 chat GPT and you sometimes get
38:41 frustrated that it doesn't it you paste
38:45 in something big and it's like I can't
38:46 do that I don't want to deal with you
38:49 right now
38:50 um the Claude instant 100 000 token
38:54 model fixes that
38:58 um so that's cool huh all right let's
39:01 see
39:03 welcome everybody name's Kyle Shannon
39:05 this is the AI learning lab go here if
39:08 you've never played with chat gbt just
39:10 go to the top one
39:11 chat.openai.com and start playing with
39:14 it tell it to write your poem tell it to
39:16 write you a Python program if you like
39:19 coding
39:20 um have it plan your meal this is a fun
39:23 one so give it like let's let's say you
39:26 wanna you wanna start eating better and
39:27 you want to eat you know a certain kind
39:29 of diet have it recommended diet
39:31 then I have it write you a meal plan for
39:33 the week then have it give you a grocery
39:36 list based on the meal plan and then
39:38 have it organize the grocery list based
39:41 on sections of the store
39:43 and you'll be like that's cool
39:46 what else can it do lots lots
39:51 so 43 years ago is when AI was first
39:53 mentioned as far as I know ah I don't
39:56 know about that that's actually really
39:57 interesting
39:59 let's ask Sage when
40:02 was the first mention of AI in a movie
40:13 the first mention of AI artificial
40:15 intelligence in a movie can be traced
40:17 back to 1927 in the silent film
40:19 Metropolis
40:21 the movie features a character named
40:23 Maria who is replaced by a robot
40:25 doppelganger who's programmed to incite
40:28 chaos and destruction that sounds like
40:30 AI while the term AI was not explicitly
40:33 used in the movie the concept of a robot
40:36 capable of autonomous thought and
40:37 behavior represents an early example of
40:39 AI and popular culture since then ai's
40:43 been a recurring theme in science
40:44 fiction notable examples being 2001
40:47 Space Odyssey Blade Runner and
40:48 Terminator that's super cool
40:51 I didn't put that together with
40:52 Metropolis but it makes perfect sense
40:54 that's really cool okay
40:56 let's figure out a way to get rich off
40:58 this you got my money
41:01 if you want to make money off chechi PT
41:03 don't listen to those people
41:07 I think a lot of people can make a lot
41:09 of money with chat GPT the reason this
41:12 channel exists is I think anyone who
41:14 gets curious and starts exploring AI
41:17 right now is is going to position
41:20 themselves to be in really good shape as
41:23 this stuff starts to transform different
41:25 businesses that's why this is here
41:27 um
41:28 I think generative AI is going to be
41:30 incredibly disruptive to to almost all
41:32 knowledge work almost all creative work
41:35 um probably most work and I think it's
41:36 coming a lot quicker than people know
41:39 so so in a very you know in a very
41:42 serious from from a very serious place
41:44 like why I'm doing this is
41:48 um
41:49 don't don't just assume that you don't
41:51 have time like go play with it go play
41:54 with chat GPT
41:56 think about the thing you do for work or
41:59 or the thing that you do that you enjoy
42:01 and start interacting with chat GPT see
42:04 if you think seeing if you can use it
42:06 kind of like a really capable co-worker
42:09 that happens to be super smart like way
42:12 smarter than you or I will ever be
42:14 that's kind of what it's like
42:17 um and and you know have it have it help
42:20 you create a project plan for work or
42:22 have it help you create you know
42:24 something you're trying to do with the
42:25 family or with the church group or
42:27 something like that um
42:28 and and uh if you end up hating it and
42:30 they're like that's a waste of my time
42:32 cool but but don't don't sit sit in the
42:35 outside and judge it
42:37 um until you actually experience what
42:39 these things are capable of it's wild I
42:41 think in the 1930s it was promoted that
42:43 robots would be everywhere it was and
42:45 we're there finally
42:48 um if you could direct generative AI
42:50 toward one thing for the benefit of our
42:52 society what would it be
42:54 um
42:55 I the the thing that I'm most excited
42:59 about with AI and the thing where I
43:01 think it's got the possibility of doing
43:03 the most good
43:06 is that generative AI is really really
43:09 good at
43:12 interpreting complex things and making
43:15 them simple for lack of a better term
43:18 so
43:20 you can just put it on top of like like
43:23 with the plugins with um with chat GPT
43:27 you can take the kayak travel plug in
43:29 and just say go book me a flight to
43:32 Cincinnati on these dates and it will go
43:35 off book you those flights come back it
43:37 might ask you some clarifying questions
43:39 and then book you the trip
43:41 that's a really simple example but it's
43:44 going to start happening everywhere
43:46 there's also there's a new
43:48 um module coming out for chatpt called
43:50 code interpreter where you can upload an
43:53 incredibly complicated spreadsheet of
43:55 data
43:56 and it will just learn it and then you
43:58 can treat it like a full-on data analyst
44:02 so so
44:03 um what's interesting about this data
44:05 and it will write python code and create
44:07 charts and graphs and tell you what's
44:08 interesting about it you could say where
44:10 within this data are there any anomalies
44:13 like where does something not look right
44:14 and it will just find it for you
44:15 instantly
44:18 um
44:19 so
44:21 so then I look at so so if it's really
44:24 good at making the complex simple
44:27 then I look at well what's the what's a
44:30 lot of what we do every day what we
44:32 spend our time on every day a lot of
44:34 what we spend our time on is
44:38 dealing with overly complicated systems
44:42 I want to book a doctor's appointment oh
44:44 I got to call my insurance company to
44:46 find a doctor and then that and then is
44:49 it covered and is it in network and is
44:50 it out in network and then you go to the
44:52 doctor and then they're only supposed to
44:54 give you a 20 copay but it's a 362
44:56 dollar copics and then you're three
44:58 hours on the phone like all of that [ __ ]
45:01 is soul-crushing right and some people
45:03 do that for a living right some people's
45:05 job is deal with the complex systems
45:07 that are in place so we can make money
45:09 or or are just in place because we don't
45:11 have a better way to do it
45:13 large language models can sit on top of
45:15 that and automate all that
45:18 soul-crushing repetitive
45:21 low value over complicated work
45:25 it can automate it
45:26 automated out
45:29 and what that potentially leaves us
45:32 is to say as humans if I don't have to
45:36 do that shitty work
45:39 what might I do
45:41 how might I contribute
45:43 how might I connect with other other
45:45 people so for me the promise of
45:47 generative AI is it allows human beings
45:50 to kind of elevate to be their best
45:52 selves I know that sounds like utopian
45:54 and and and and kind of ridiculously you
45:58 know sappy
46:00 but I really do think that's a
46:02 possibility I mean my experience with
46:04 generative AI is
46:06 so so it's incredibly creative right
46:08 like you it creates these amazing images
46:09 and it can help you write it can help
46:11 you do all this stuff
46:12 one of the questions I get asked a lot
46:14 is well if it's doing all that creative
46:16 work aren't you less creative and it's
46:18 the absolute opposite is true I've I've
46:20 not been as creatively and
46:23 intellectually stimulated as I've been
46:25 in the past six months for like the past
46:27 15 years
46:30 because
46:33 the the AI is doing all of the work that
46:36 used to get in the way of my creative
46:38 thinking and here's what I mean by that
46:41 historically if I have a creative idea
46:43 and I've got ADHD so I've got like you
46:45 know [ __ ] 50 a day right some some
46:48 [ __ ] you know electro shock happens
46:51 in my brain it's like oh I got an idea
46:54 historically what I would have done is
46:56 I'd have gone okay if I want to bring
46:58 that idea to life I should probably make
47:00 a flow chart
47:01 to to like organize my idea I know let
47:04 me go get some flow charting software
47:06 and then I'd go on the internet go to
47:07 Google and say flow charting software
47:09 and then like 17 clicks later I'd find
47:12 something that was sort of like what I
47:13 wanted and I download it and then I'd
47:15 install it and then I wouldn't have
47:17 enough room in my hard drive and then
47:19 and then and then I would go to launch
47:21 it and like I'm not allowed to use it
47:23 unless I pay 25 for it and like at this
47:25 point it's 45 minutes in I don't even
47:27 know what the idea was and I don't give
47:29 a [ __ ] anymore and I'm like oh [ __ ] this
47:31 and I delete the software and I just go
47:33 about my day
47:34 what happens with this generative AI
47:36 stuff is I have that Whisper of an idea
47:38 and I'm like huh let me see if I can put
47:41 let's put that together I pop it into
47:43 chat GPT boom up comes it organizes my
47:46 thoughts oh that's really cool that
47:48 gives me two more ideas what do I do
47:49 with those oh you know what would be
47:51 cool is have an image that goes with
47:52 that and then I jump over to Mid journey
47:54 and within [ __ ] 30 seconds over there
47:56 now I've got an image like oh yeah
47:57 that's it and and
47:59 so I find myself operating at this
48:02 higher plane more consistently because
48:05 all of the Tactical over complicated
48:08 [ __ ] work like the blocking and
48:11 tackling eighty percent of the work that
48:13 we do happens like that
48:16 so for me the thing I'm most excited
48:18 about is okay do the 80 [ __ ] work and
48:22 let us as human beings do that that 20
48:25 stuff and what might we do if we don't
48:27 have to do all that other stuff
48:29 I think it could be really good I think
48:31 it could be profoundly good
48:34 or the robots kill us
48:36 oh looks like they've got some new
48:38 people here my name is Kyle Shannon this
48:40 is the AI learning lab a little bit
48:42 about my background I'm an entrepreneur
48:44 I've started I don't know 12 or 15
48:45 companies I was an innovator in the
48:47 early days of the World Wide Web uh I
48:49 started one of the early uh digital
48:52 agencies called called agency.com and
48:56 um the parallels between the early days
48:58 of the web and the early days of this
49:00 this post chat GPT AI world are
49:04 incredibly similar and and I think
49:07 what's coming is quite profound and so
49:09 this channel exists for me to just
49:10 explore that and uh and do all that I'm
49:13 I'm the co-founder and CEO of a company
49:15 right now called storyvine it's 11 year
49:17 old company and if you have questions
49:18 about AI pop them down below I'll do my
49:20 best to answer them I'm very far behind
49:22 if you haven't played with chat CPT
49:24 that's what that URL is right there
49:26 that's the official chat GPT website
49:28 bing.com also has chat GPT in it as does
49:31 poe.com all three of these are options
49:33 if you've not played with chat GPT
49:36 inside some form go play just go play
49:39 while I ramble
49:41 um this is great thank you very much I
49:43 appreciate that I'm making people laugh
49:45 what are you generating on uh I don't
49:48 know what I was generating on I think
49:49 that was poe.com that I was talking
49:51 about before
49:52 that top left looks like the waiter uh
49:55 looks like Walter White oh oh the images
49:57 oh I was generating on mid-journey
50:00 before when I was generating those
50:01 images big shout out from Minneapolis
50:03 Minnesota
50:06 hello Minneapolis what do you think of
50:08 Google bard I think the Google's behind
50:11 and Google bard is nowhere close to chat
50:16 gbt at least with the language I've
50:18 heard that it's better with with code
50:20 and Mathematics which makes sense to me
50:22 because Google's an engineering company
50:24 first
50:25 um but but I think they're very far
50:27 behind I think the core model is just
50:29 not as good as as
50:31 um is open ai's core models
50:35 Happy Father's Day to everyone
50:38 um
50:39 don't get me wrong I think I think I
50:41 mean Google's gonna do some really
50:42 interesting things they just dropped a
50:45 pretty remarkable piece of technology
50:48 it's you can't play with it yet but it's
50:49 called style drop and you take a single
50:52 image and it learns the style of that
50:54 image and then you can just prompt other
50:55 images and it like it copies the style
50:57 like in remarkable ways
51:01 um
51:01 so I wouldn't count Google out but
51:03 Google got blindsided by the popularity
51:07 of chat gbt as did open AI that invented
51:10 it as did Microsoft
51:12 Microsoft is pouncing on the opportunity
51:15 because they didn't have a chance in
51:17 hell at competing with Google search and
51:21 all of a sudden chatgpt came out and
51:23 it's an existential threat to Google so
51:25 really interesting times in terms of the
51:28 competitive landscape that that was you
51:30 know it was effectively a monopoly by
51:33 Google for a lot of years all of a
51:35 sudden
51:37 shit's different
51:39 shit's coming or shit's here and
51:43 Google's like uh we got AI too in fact
51:46 we invented this [ __ ] which they did
51:48 but then they
51:50 didn't focus on it I think probably
51:52 because they knew it was a threat to
51:54 their search business
51:57 that's my uh speculation that is purely
52:00 speculation
52:02 uh
52:04 uh some someone said AI is a demon
52:06 porthole possible
52:08 maybe that's what's going on with the uh
52:10 the Schumann resonance maybe it's all AI
52:14 or maybe it's a hack or maybe it's I
52:17 don't know who knows maybe the
52:18 simulation's firing up
52:21 um so so one of the if you've been to my
52:24 lives before you know this but but what
52:26 what Tick Tock does is I'm as I'm trying
52:28 to scroll through the questions it will
52:30 just jump me to the bottom and then I
52:32 try to go back and find my questions
52:33 where I was so that's what I'm doing
52:34 right now
52:36 so
52:38 so what you're doing right now is
52:39 watching an old man scroll through Tick
52:41 Tock trying to find his place
52:43 with with 58 year old eyes I can I can
52:47 see it Marge it's gonna be good it's
52:48 gonna be good just as soon as I get back
52:50 to that question
52:51 what do you say what'd it say what are
52:54 people talking about
52:55 am I being trolled oh my gosh what are
52:58 we gonna do what are we gonna do
53:02 um yeah actually if you would while I'm
53:04 scrolling um if you could just sort of
53:06 tap the screen and like let people know
53:08 we're here it'll get people in or share
53:10 the live whatever whatever we can do to
53:12 get some some folks here but we have a
53:13 decent amount of folks here I appreciate
53:15 you being here it's a Sunday night it's
53:17 Father's Day Happy Father's Day
53:20 um best app to create downloadable PDFs
53:22 that are visually appealing other than
53:24 canva
53:26 um
53:26 I don't know I like I would say canva
53:29 but but
53:31 so what I would say is this there may be
53:34 some AI [ __ ] to do that go to
53:37 futurepedia.io and look up print
53:39 production there are probably some some
53:41 tools out there that do that do that
53:43 pretty well
53:45 it gave the power to artists oh no
53:48 it did and and uh some of the artists
53:52 are not happy about that either like the
53:54 ones have been illustrators for 30 years
53:56 and all of a sudden they're like wait it
53:57 does what oh [ __ ] really
54:00 so so everyone right now that takes a
54:05 look at this thing is being faced with
54:07 some form of existential crisis
54:11 igoodbing.com I assume you're asking is
54:13 bing.com any good so so bing.com is
54:19 solid but it's not as good as gpt4 at
54:23 openai and here's why so if you go to
54:26 chat.openai.com they're 20 bucks a month
54:28 GPT plus thing gives you access to gpt4
54:33 bing.com is also gpt4
54:37 why are they not the same because they
54:40 there's this thing called fine tuning
54:42 where you can basically they've each got
54:43 their own kind of safety guard rails and
54:45 their own kind of fine tunings for how
54:47 they operate so for one thing is they
54:49 fine tune them in a different way the
54:51 other thing is when bing.com launched
54:55 there was a reporter from The New York
54:57 Times named Kevin Roos who started
54:59 having a chat with chat gbt at Bing and
55:02 spent four hours with it and got it to
55:05 admit that its name was Sydney and and
55:08 kind of made it snap and then Sydney
55:09 told Kevin Roos that it was in love with
55:11 him and he should leave his wife and so
55:13 Kevin screenshotted that and they put it
55:15 on the cover of the New York Times so
55:17 Microsoft lobotomized Sydney so so
55:21 bing.com has been a little squished
55:24 squished ahead it
55:27 um I think it's getting better but it's
55:29 it's just not as good as as the GPT for
55:32 it openai.com but anyway that that's why
55:34 so do you have YouTube I I do not have a
55:37 YouTube I think that seems to be a
55:39 request so at some point I'll figure
55:41 that out great content still learning
55:43 yeah me too so so listen here's the
55:45 thing if anyone right now tells you
55:47 they're an expert at this generative AI
55:49 [ __ ] they're lying nobody's an expert
55:51 right now this stuff is changing daily
55:54 like daily
55:56 um and anyone who claims to be an expert
55:58 is not now that doesn't mean there
56:00 aren't experts in machine learning and
56:01 AI again the the there's we're at an
56:05 inflection point before the inflection
56:07 point you have data scientists and you
56:10 know machine learning Specialists and
56:12 right researchers
56:15 that shit's all still happening and
56:17 those people are absolutely experts the
56:19 inflection point I'm talking about is
56:21 Chachi PT made
56:24 AI available to the rest of us right so
56:28 we're going to use it for business and
56:30 we're going to use it for our little
56:31 flower shop and we're going to use it
56:32 for for our lawn mowing business and
56:35 we're going to use it as a data analyst
56:36 right so how we use these generative AI
56:39 tools is is the new thing that I'm
56:41 talking about anyone who claims to be
56:43 expert in that the post inflection point
56:47 because when did chat gbt come out
56:49 November 30th 2022 like six months ago
56:52 this shit's moving way too fast no one's
56:54 an expert right now so everyone's
56:56 learning we're all learning
56:58 um how do you get kids to use AI let
57:00 them start playing I went to my
57:01 co-founder Monique she's got a son at a
57:04 school with like you know Wicked smart
57:06 kids and
57:07 and
57:09 um I went there and I just started
57:10 showing him how to make uh generative
57:12 images and they just kind of lost their
57:14 minds and like this one kid
57:17 um I've been I've been sort of teaching
57:19 them for two hours and literally just
57:20 wrangling cats and like they're bouncing
57:22 off the walls and like you know some of
57:24 the bad boys are going tell it to make
57:26 boobies I was like no I don't know how
57:28 to make Blues
57:29 so at the end of this two hours this one
57:32 kid comes up to me and he goes you know
57:34 I'm uh
57:36 I'm making a computer program in C sharp
57:39 and I was like really he goes yeah it's
57:41 a clone of Flappy Birds I'm like cool
57:44 and and I said how's it going he goes
57:46 well I got some I got a bug in the code
57:48 and uh so I'm gonna have to throw that
57:50 out and start over I'm like yeah it's a
57:52 bummer I said if you use chat GPT to to
57:54 help you yeah yeah yeah I know how to
57:56 use that all right okay
57:58 so so I said come here take a look at
58:01 something so I pop open chat gbt and I
58:03 say
58:04 make a c-sharp applications that's a
58:07 clone of flappy birds that will run in
58:09 unity which was the environment that
58:11 he's he's programming in and it spits
58:13 out all this code right because it's
58:17 it was like that moment in in Home Alone
58:20 where Kevin goes
58:23 we went
58:24 and and he goes I gotta get my computer
58:27 and he ran down four flights of stairs
58:29 we were in a four four-story school he
58:31 ran down four flights of stairs gets his
58:33 laptop runs back up to four flights of
58:35 stairs and he sits it down and goes okay
58:39 he goes here's my code so he copies the
58:42 code we we paste this code into chat EBT
58:46 and we say is is there a bug and it said
58:48 sure there's a bug right here and he had
58:50 a typo right it was a simple little typo
58:53 and so he copies the new code he goes
58:56 into Unity pastes it into unity and it
58:59 runs and now he's playing Flappy Birds
59:02 and he stops and he looks at me and he
59:04 goes
59:05 this changes everything it's like it's
59:09 like this magical moment so you get a
59:12 you get I mean how you get kids involved
59:14 in [ __ ] is you let their imagine take
59:16 over right or their imagination take
59:18 over and so
59:20 um figure out what they're interested in
59:22 like like if if I don't know if you're
59:25 teaching a thing about bugs you know go
59:27 into chat and have it make poems about
59:30 bugs and tables comparing different bugs
59:33 and
59:34 um have it come up with creative names
59:36 that incorporate Latin bug names into
59:40 the names and and then go to Mid Journey
59:43 or stable diffusion or where wherever
59:45 and have it make bugs and then talk
59:48 about like just get just start playing
59:51 play like and I feel like that goes for
59:55 adults as well right
59:57 it's really easy to go well this this
1:00:00 machine learning this these AI
1:00:02 algorithms are very important and
1:00:03 profound and you really should learn
1:00:05 this no one's going to want to learn
1:00:06 that they're going to be like shut up
1:00:08 [ __ ] like don't be such a pompous
1:00:10 dick
1:00:11 but if you say hey what are you
1:00:13 interested in let me show you something
1:00:15 cool
1:00:16 um
1:00:17 one of the things that so I started this
1:00:19 thing called the AI Salon the salon.ai
1:00:22 there's a link tree there if you want to
1:00:23 go to it um it's a community of people
1:00:26 that are just curious about AI I started
1:00:28 it to really just create a community of
1:00:31 people that were like you know I'm into
1:00:33 this let's figure this thing out
1:00:34 together and and it's that and and it's
1:00:37 the kind of thing where that there's
1:00:39 very much a sense of curiosity and
1:00:41 exploration and
1:00:44 through play through playing with these
1:00:47 tools you discover what's possible and
1:00:49 then at some point it clicks and you go
1:00:51 oh wait a minute
1:00:53 I wonder if I could use this for my
1:00:54 thing at work whatever the thing at work
1:00:57 is
1:00:57 and then you pop it into chat EBT and
1:01:00 you're like holy [ __ ] like you have that
1:01:03 moment that that kid had like right this
1:01:07 moment is the magic moment with
1:01:09 generative AI it's like the moment where
1:01:11 you realize this isn't a Google search
1:01:13 this is something different
1:01:16 um so just I don't know play get curious
1:01:20 any tips on creating dxf Vector files
1:01:23 with prompts from Ai and mid-journey
1:01:26 well you're not going to create any dxf
1:01:29 Vector files with
1:01:32 um
1:01:34 mid-journey because it's just creating
1:01:37 um bit mapped images
1:01:39 it will create things that look like
1:01:41 their dxf right if you say create a dxf
1:01:44 3D vector image in mid Journey it'll
1:01:47 create something that looks like that
1:01:48 but it won't have the 3D data right
1:01:50 because just the nature of how diffusion
1:01:51 models work it's creating denoised pixel
1:01:54 images whatever at some point in the
1:01:56 future they're going to be able to well
1:01:58 they can already do it now to a degree
1:02:00 whatever it doesn't matter like at some
1:02:01 point in the future they'll be able to
1:02:02 extract all that [ __ ] out of it
1:02:05 um but you can in chat GPT let me
1:02:07 actually just go see if it it does it
1:02:09 I'm pretty sure it does hold please I'm
1:02:11 gonna go do something for a second here
1:02:15 um let's see
1:02:17 um
1:02:18 write a d f x Vector of
1:02:25 spheres
1:02:26 um or
1:02:30 orbiting
1:02:35 orbit in
1:02:38 a sun
1:02:40 yeah so okay so let me let me show you
1:02:42 this
1:02:44 so I don't know where your
1:02:47 um displaying your DFX vector graphics
1:02:51 but
1:02:53 this shit's Wild
1:02:55 so whether it's in unity or some other
1:02:57 3D program wait what what did I do here
1:03:01 hold please okay
1:03:04 so so I'm gonna say
1:03:06 um create
1:03:09 um a
1:03:10 DFX vector
1:03:14 vector
1:03:15 scene with a plane
1:03:20 and
1:03:21 um
1:03:23 pyramids
1:03:25 orbiting a
1:03:28 Sun
1:03:31 and it will write you the code
1:03:35 and then you can just copy and paste
1:03:36 that and put that over in into your
1:03:39 program
1:03:40 there are probably if you go to
1:03:45 um
1:03:47 if you go to futurepedia.io this thing
1:03:50 on the bottom here
1:03:52 there are probably things that that do
1:03:54 that more automatically like there are
1:03:57 probably plug-ins for Unity or or
1:04:00 whatever
1:04:01 Unreal Engine or whatever you're trying
1:04:03 to do or you're trying to put some some
1:04:05 Vector [ __ ] in there
1:04:06 um or if it's just an art program
1:04:09 um
1:04:11 you know there might be a way to do it
1:04:14 more elegantly but you can just do it
1:04:15 and you can also do it the other
1:04:16 direction like you can copy that Vector
1:04:18 code out of your program paste it into
1:04:21 chatgpt and then just describe what you
1:04:23 want it to do it's excellent
1:04:26 at parsing code it's excuse me it's
1:04:30 excellent at it like scarily good at
1:04:32 parsing code so yeah it can just do that
1:04:35 [ __ ]
1:04:37 hey everybody my name is Kyle Shannon
1:04:39 this is the AI learning lab these URLs
1:04:41 behind my head here and and that crazy
1:04:43 eye I made in stable diffusion that was
1:04:45 part of an art project that I did
1:04:47 um
1:04:49 chat.openai.com is the official chatgpt
1:04:51 website I'm not affiliated with this in
1:04:53 any way so if you're like he's just here
1:04:55 how to make money off us
1:05:00 bing.com is Microsoft's search engine
1:05:03 but if you click the chat button that's
1:05:05 gpt4 that you can use for free and it's
1:05:07 connected to the internet and then
1:05:08 poe.com lets you play with three or four
1:05:11 no six different
1:05:13 um chat GPT models
1:05:15 um three from anthropic two from open Ai
1:05:18 and one from one one of their own
1:05:21 um so that's what they are canva got AI
1:05:23 now I know canvas got AI it writes stuff
1:05:26 it does um what do they call it bulk
1:05:28 create
1:05:29 so you you can have chat GPT generate
1:05:32 like a bunch of
1:05:34 Instagram you know sayings or whatever
1:05:36 and then you can go to canva and just
1:05:40 upload the CSV file of all those sayings
1:05:43 and it'll just spit out a bunch of
1:05:45 graphics for you it's stupid cool it's
1:05:48 so stupid
1:05:50 it's so stupid it's cool
1:05:53 Gen X baby
1:05:55 um I don't know those two things were
1:05:57 not related but I just feel like an
1:06:00 idiot and I am drinking whiskey but I
1:06:02 haven't had enough to be drunk yet
1:06:08 stock stock I like in AI I'm not an
1:06:11 investor
1:06:12 um other than in my own companies
1:06:14 um
1:06:17 I like the companies that I think are
1:06:19 worth paying attention to right now
1:06:21 Nvidia just hit
1:06:23 um a trillion dollar market cap AMD is
1:06:25 claiming that they're going to compete
1:06:27 with them
1:06:28 but I don't know I don't I mean can they
1:06:31 maybe they can
1:06:33 um I think they're maybe the only ones
1:06:34 who can Microsoft is
1:06:37 um
1:06:37 acting like a startup for the first time
1:06:40 in decades
1:06:42 um
1:06:43 open AI you can't invest in I think
1:06:45 anthropic's a really interesting company
1:06:47 but I don't think you can invest in them
1:06:48 they're private
1:06:51 um
1:06:55 Adobe is doing I mean they're big and
1:06:58 established but but the
1:07:01 their work in the past four months on
1:07:05 generative AI making it safe for
1:07:07 commercial entities to use generative AI
1:07:09 images is [ __ ] baller and and it's
1:07:13 like no one's touching them right now
1:07:15 so I think adobe's really interesting
1:07:19 um
1:07:21 is there an AI product that can do photo
1:07:24 editing with specific requests like fix
1:07:26 washed out backgrounds that exists to a
1:07:30 degree there's a lot of the the AI
1:07:32 filters that have been in Photoshop for
1:07:34 a while I mean one of the things about
1:07:35 AI is it's been in our software products
1:07:37 for for a long long time
1:07:40 I think what you're talking about though
1:07:41 is a large language model front end like
1:07:44 on top of
1:07:46 um
1:07:47 Adobe
1:07:48 um Lightroom for example where you've
1:07:50 got all these complicated you know
1:07:52 switches and settings and you just want
1:07:54 to be able to say you know make it look
1:07:56 more saturated you know give it darker
1:07:59 vignette and it just does all that [ __ ]
1:08:01 um
1:08:03 I would assume that we're going to see
1:08:05 that in Adobe you know probably within
1:08:08 the year like they just they just
1:08:10 released
1:08:11 in the beta of Illustrator a thing
1:08:13 called generative recolor or something
1:08:16 recolor where you basically take an
1:08:19 image and you just say give me a whole
1:08:20 new color palette and it dynamically
1:08:22 recolors the thing for you so
1:08:25 um so I think you'll see that coming I
1:08:27 don't know again I would anything very
1:08:30 very specific like that I'd go to
1:08:31 futurepedia.io I'd look up photo editing
1:08:34 or you know photography at
1:08:37 futurepedia.io there there's probably
1:08:38 something out there that does it but I
1:08:41 would wait for one of the big players to
1:08:43 incorporate it because they will
1:08:46 uh
1:08:48 uh do they pick stocks well
1:08:51 like the AI does what you tell it to do
1:08:54 so if you give it a bunch of data and
1:08:57 say of this data I just gave you
1:08:59 based on you know one give me an
1:09:02 approach for how I might choose which
1:09:04 stock to invest in and it'll tell you
1:09:06 that it'll probably tell you that it
1:09:07 can't give you Financial advice because
1:09:09 it's got safety [ __ ] in it
1:09:11 and then tell it to say based on the
1:09:13 data that I gave you you know which one
1:09:15 would you recommend and see what it does
1:09:17 so so again you know the the the the the
1:09:20 how you phrase the question do they pick
1:09:22 stocks they the AI
1:09:25 are not kind of walking around on their
1:09:28 own going I'm gonna go [ __ ] with people
1:09:29 now it's like it's a tool that human
1:09:32 beings get to use like we as human
1:09:34 beings get to use these AI tools so if
1:09:37 you want to use it to pick stocks start
1:09:39 interacting with it and start
1:09:40 understanding how much it understands
1:09:42 about analyzing company data
1:09:47 and then have it analyze some company
1:09:49 data for you and then just go copy and
1:09:51 paste the table off some Forbes article
1:09:53 that says here's all the big players in
1:09:55 Ai and here's all their company data
1:09:57 just copy that [ __ ] paste it right into
1:09:58 chat GPT and go tell me about this
1:10:01 see what comes up it will probably blow
1:10:04 your mind
1:10:05 but you're in control not the machine
1:10:08 we got to give that a year or two
1:10:10 they're in trouble all right Adobe
1:10:13 Photoshop can do that yep I'm not sure
1:10:15 how much it costs photoshop's expensive
1:10:17 as [ __ ]
1:10:18 you're brilliant oh thank you I don't
1:10:20 think I'm brilliant I am uh smarmy
1:10:23 [Laughter]
1:10:26 appreciate the compliment I need to get
1:10:29 better at taking compliments as I like
1:10:31 to say here shame is my love language
1:10:32 taking compliments not so good
1:10:35 foreign
1:10:41 how can I sign up without a phone number
1:10:42 you can't so if you don't want to sign
1:10:44 up with a phone number there you're
1:10:45 probably an nft person don't want to dox
1:10:47 yourself I hung out with a bunch of nft
1:10:50 people I'm I'm in a project called Neo
1:10:52 Tokyo so I get the doxing thing if you
1:10:54 don't want to do that and you have [ __ ]
1:10:56 to hide then just go use bing.com it's
1:10:58 free and I don't think you have to give
1:11:00 poe.com your phone number
1:11:02 but personally
1:11:05 open AI are the people that invented
1:11:07 this [ __ ] they're like they're the ones
1:11:09 that that have cracked the code on this
1:11:10 thing first and so I would go with them
1:11:12 their their model is by far the best
1:11:14 right now so that's just my two cents
1:11:18 PO is poe in the Apple Store yes poe.com
1:11:21 has an Apple app they do yes
1:11:26 Kyle demonstrated Adobe Photoshop oh
1:11:29 wait you're welcome
1:11:30 oh oh yeah I I demonstrated uh Photoshop
1:11:33 generative Phil I'm happy to do that
1:11:34 again if someone wants to see it
1:11:37 can anyone try prompts in your Discord
1:11:39 mid Journey Channel they can't or in my
1:11:42 Discord mid-journey Channel they can if
1:11:44 you go to Mid Journey play in the in the
1:11:47 AI Salon Discord Channel
1:11:50 you do slash imagine and you can make
1:11:52 things you I think they give you like 25
1:11:55 images that you can do for free and
1:11:58 after that no matter what channel you're
1:11:59 in you've got to have a subscription the
1:12:02 cheapest mid-journey subscriptions like
1:12:03 10 bucks that's the one I that's the one
1:12:05 I have but yeah you can do it right
1:12:07 within
1:12:08 um the AI Salon which is super cool
1:12:10 because you can talk to people in there
1:12:12 about the stuff you're doing and ask
1:12:13 questions and it's it's a really good
1:12:15 community
1:12:16 what's the app again I don't know that
1:12:18 movie is amazing
1:12:20 po chat GPT what does chat gbt wait why
1:12:25 does chat cheap BT only start a few
1:12:26 years ago why can't I have recent data
1:12:28 so it's it's the nature of how the core
1:12:31 the base models were trained
1:12:34 so basically to train GPT three and four
1:12:38 they open AI like spent years finding as
1:12:42 much as many data sources as they could
1:12:44 they essentially
1:12:46 sort of dug into every nook and cranny
1:12:48 of the public internet so everything
1:12:50 from Wikipedia to blogs to library
1:12:54 servers to uh what just whatever just
1:12:56 just data everywhere
1:12:58 that process of building that base model
1:13:01 was very very labor intensive and and
1:13:04 just required a lot of hand tagging of
1:13:07 data and organizing and sifting and so
1:13:11 that process stopped in September of
1:13:13 2021
1:13:14 the the so so while the base model stops
1:13:18 at September of 2021
1:13:20 bing.com and then gpt4
1:13:24 um that's connected to the internet at
1:13:26 open AI
1:13:27 because they're connected to the
1:13:29 internet if you do anything beyond
1:13:32 September 2021 it'll basically just go
1:13:34 do a Google Search and then it will
1:13:36 augment your answer with up-to-date data
1:13:38 so so it's it's um if you're using the
1:13:41 free version of GPT 3.5 that's not
1:13:45 connected to the internet and that will
1:13:46 tell you it stops at September 2021 but
1:13:49 if you go to bing.com and click on their
1:13:51 chat button that's connected to the
1:13:53 internet and you can get up-to-date
1:13:56 information in your searches
1:13:58 there are a few YouTube channels that
1:14:00 teach you how to make your own Discord
1:14:01 channel oh that's cool nice beautiful
1:14:05 I'm down the hall awesome Andrea welcome
1:14:10 the whole the whole she runs deep I mean
1:14:12 it's just like I like once you have sort
1:14:15 of the Kevin moment from home alone like
1:14:17 oh my God I don't know I could do that
1:14:21 once you have that moment
1:14:23 then I think you do one of two things I
1:14:25 think you go I'm not gonna deal with
1:14:27 this or you go
1:14:29 oh let me see what else it can do oh
1:14:32 [ __ ] it could do that too
1:14:34 and it could do that and then you're
1:14:36 just like let's get sucked in so welcome
1:14:39 it's been six months I still like I
1:14:42 still on a daily basis just my jaw drops
1:14:46 open I'm like really
1:14:48 like I
1:14:50 you know I don't know where I am I feel
1:14:53 like we are
1:14:55 lucky to be living at a time where we
1:14:58 have access to technology
1:15:00 that is effectively science fiction from
1:15:04 my childhood anyway brought to life like
1:15:09 we're playing with tools that are like
1:15:11 the tools that have been in our science
1:15:13 fiction films
1:15:15 you know
1:15:16 we don't have time travel yet but a lot
1:15:18 of the other [ __ ] is if if it's not here
1:15:21 it's coming or we can see like demos of
1:15:24 it like enough fake demos not like
1:15:26 Hollywood demos but like there's some
1:15:29 Geeks in a back room somewhere they got
1:15:31 some [ __ ] working that you know in a
1:15:32 year or two we're gonna see
1:15:35 it's just [ __ ] insane and and you
1:15:38 know also things like like
1:15:41 the speed of scientific innovation is
1:15:43 accelerating and I think I think in you
1:15:46 know in in a relatively short amount of
1:15:48 years we're going to start to see
1:15:49 scientific breakthroughs like you know
1:15:51 climate breakthroughs and medical
1:15:54 breakthroughs and and you know there's
1:15:56 there's one that I've been talking about
1:15:57 in the lives there's a
1:15:58 a paralyzed guy I think he's in Italy
1:16:02 he's been paralyzed for 12 years
1:16:05 they took fmris of his brain and they
1:16:10 um used a large language model they told
1:16:11 him to think about moving his legs and
1:16:13 he thought about moving his legs and
1:16:15 they used large language models to learn
1:16:17 those patterns
1:16:18 and then they built a Bluetooth bridge
1:16:20 to go from from his brain they took
1:16:23 those signals from his brain and they
1:16:24 sent him over Bluetooth to electrodes at
1:16:27 the base of his spine they jumped the
1:16:29 injury and now he's walking
1:16:32 like holy [ __ ] [ __ ] like
1:16:35 really yeah
1:16:37 so so like I I don't know so the rabbit
1:16:42 hole runs deep
1:16:44 what are the entrepreneurial and job
1:16:46 opportunities what should we do now so
1:16:48 so
1:16:53 the the short answer is I don't know
1:16:58 the the longer answer is there's there's
1:17:00 a guy named Paul writer
1:17:02 um he's got a podcast called the
1:17:03 marketing AI Institute
1:17:06 um I would strongly encourage you to
1:17:07 listen to it but he he has this saying
1:17:09 where he says AI is not going to replace
1:17:11 your job but someone using AI will
1:17:15 and so I think the most important thing
1:17:16 to do right now is just get curious just
1:17:19 start playing start understanding what
1:17:21 these things make possible start
1:17:24 um
1:17:25 go to an image generation tool if you go
1:17:28 to bing.com create
1:17:30 that's the dolly image generation tool
1:17:34 so if you go to bing.com and click on
1:17:35 chat you can chat with chat gbt for free
1:17:39 if you go to bing.com create and you
1:17:41 input make me a picture of a penguin
1:17:44 skiing down a black diamond Mogul Hill
1:17:48 you'll have a picture of a penguin
1:17:51 skiing
1:17:52 the the image generation isn't super
1:17:55 great but it's there
1:17:57 so so I would say first just start
1:18:00 understanding what these tools can do
1:18:03 and what they make possible then I would
1:18:05 go to something like prompts.chat which
1:18:07 which starts to teach you the concept of
1:18:09 prompting and how you start thinking
1:18:11 about interacting with these tools more
1:18:13 like a collaborative partner than like a
1:18:16 search engine because that's more what
1:18:18 it's like you kind of go back and forth
1:18:20 with these tools and you can do some
1:18:21 really sophisticated things very very
1:18:23 quickly
1:18:25 and then what I would do is I would
1:18:26 start saying you know what are you good
1:18:28 at like either what are you passionate
1:18:30 about professionally or what are you
1:18:32 good at professionally and start then
1:18:33 figuring out okay how can I do what I do
1:18:36 or what I I enjoy having these tools
1:18:39 help me do that better
1:18:41 and then the the final piece of the
1:18:45 puzzle so it's like get curious then get
1:18:47 professionally curious
1:18:49 and I would say along that Journey share
1:18:52 what you're learning with other people
1:18:54 that's why this channel exists I just
1:18:56 had to get this [ __ ] out of my head I'm
1:18:57 like I can't process how powerful this
1:19:00 is I can't like I have no place to put
1:19:03 this and no one understands what the
1:19:04 [ __ ] I'm talking about
1:19:06 so start sharing what you're learning
1:19:09 because as companies realize that they
1:19:13 need to deal with this AI stuff the
1:19:15 first thing they're going to do is go
1:19:17 who do I know that knows anything about
1:19:19 this oh deadly knows about this she was
1:19:23 talking about chat GPT last week she
1:19:25 said she did something with a meal plan
1:19:27 you should talk to her and then all of a
1:19:29 sudden you're talking with the boss and
1:19:31 the boss is going what the [ __ ] is this
1:19:32 AI stuff you're like let me tell you
1:19:33 right so and I've seen that in the AI
1:19:36 Salon where people that get curious
1:19:37 about this stuff and start talking about
1:19:39 it on LinkedIn all of a sudden have a
1:19:41 job
1:19:42 the guy that I I have a newsletter
1:19:44 called everyday Ai and Greg the guy that
1:19:47 I work with on that he just got a really
1:19:48 powerful job and a big part of it they
1:19:52 cited was the everyday AI newsletter the
1:19:54 fact that he was that he knew about Ai
1:19:56 and then he was talking about Ai and
1:19:58 that he was a thought leader in it so
1:20:00 just get curious like there's not a
1:20:03 specific thing there's not a course to
1:20:04 take is there anything that can make
1:20:06 chat EPT remember such as code or files
1:20:08 hallucinations too often yes so it does
1:20:11 hallucinate too often there is there's a
1:20:14 thing called Lang chain which allows you
1:20:17 to train chatgpt on your own data so you
1:20:20 can point it at directories and then
1:20:22 it's got a thing called a vector
1:20:23 database after you embed your files it
1:20:26 puts the vectors in a vector databases
1:20:28 all sorts of geeky [ __ ] right but that's
1:20:31 memory that Vector database is memory so
1:20:33 I would say go explore Lang chain
1:20:36 um and then you can also look at things
1:20:37 like um GPT for all and
1:20:41 Auto GPT and agent GPT which are kind of
1:20:44 autonomous agents that that leverage
1:20:47 Lang chain Lang chain stands for
1:20:50 language chain it's basically just a
1:20:52 series of tools you can string together
1:20:54 um to augment these these core large
1:20:57 language models
1:20:59 um you got to be decently geeky though
1:21:01 the other thing you can do is wait so
1:21:04 um gpt4 is going to be rolled into all
1:21:06 of Microsoft 365. the office 365. it's
1:21:11 going to be rolled into windows it's
1:21:12 going to be rolled into the Azure cloud
1:21:15 services it's going to be rolled into
1:21:16 Google's office suite so this stuff is
1:21:19 going to be coming I would say six
1:21:21 months to a year you won't have to use
1:21:23 something like Lang chain but if you
1:21:24 want to play now
1:21:26 um that's a that's a place to start can
1:21:27 chat do simple animation of black and
1:21:30 white line art not yet
1:21:32 and marry it to an audio file scripted
1:21:35 on demand oh you want full multimodal
1:21:37 that's coming that's probably gpt5
1:21:41 um gpt4 can
1:21:43 um ingest images I I don't think it can
1:21:46 create them but I'm sure when they roll
1:21:47 it together with Dolly it'll be able to
1:21:49 create them but that's not even launched
1:21:51 yet
1:21:53 um but
1:21:55 you can
1:21:58 you can have Chachi BT do things like
1:22:02 you give it a subject that you want 3D
1:22:06 line art for and it will generate a
1:22:07 mid-journey prompt and then you can take
1:22:10 that to Mid journey and generate that so
1:22:12 with multiple tools you can do what
1:22:13 you're talking about there's just not a
1:22:15 single tool that wraps it all together
1:22:17 right now
1:22:18 what are the top three tools you
1:22:20 recommend for Content creation chat GPT
1:22:23 is the big one it's it really is the big
1:22:26 one if you look at tools out there like
1:22:30 um oh what the [ __ ] it called uh
1:22:32 copy.ai and the other one Jasper those
1:22:36 things are just kind of pre-prompted
1:22:38 rappers sitting on top of chat GPD it's
1:22:42 sitting on top of GPT three and four
1:22:44 um
1:22:46 for me the big tools are you know one of
1:22:50 these three
1:22:51 for the words mid Journey for images or
1:22:53 stable diffusion for images or if you're
1:22:55 doing things commercially Adobe Firefly
1:22:58 and generative fill
1:23:01 um
1:23:02 and then zapier for automations like you
1:23:05 can you can automate
1:23:08 um like taking input from a form and
1:23:10 send that to open AI to do some cool
1:23:13 [ __ ] and then send that in an email or
1:23:15 post it to a slack channel so I'd say
1:23:17 zapier for automations
1:23:19 that's what I use there's a bunch of
1:23:21 them out there zapier for automations
1:23:23 chat GPT for Words mid Journey or stable
1:23:26 diffusion or adobe Firefly for images
1:23:29 and that's kind of it like that's all
1:23:31 I'm playing with right now if you want
1:23:33 to geek Out On Tools that's why I have
1:23:34 this here because I can't keep up with
1:23:36 them anymore and I stop trying I said
1:23:38 [ __ ] it I'm done I can't because it's
1:23:41 literally like
1:23:43 five tools a day come out that every
1:23:45 single one of them is a [ __ ] Rabbit
1:23:46 Hole you could run down for hours and
1:23:48 it's just like I can't
1:23:53 I I ain't gonna do it I ain't gonna do
1:23:56 it people I'm not gonna learn you stupid
1:23:57 tools
1:23:59 so that's why that's there and it's I
1:24:01 listen quite quite honestly it's fun to
1:24:03 go there I haven't been there in a while
1:24:04 it's fun to go there and just look at
1:24:06 some of the [ __ ] people are doing
1:24:07 because it's amazing
1:24:09 all right I have a couple of use cases
1:24:11 using GPT for agriculture legal and
1:24:13 psychology any thoughts yes yes and yes
1:24:16 um my thoughts are so so go to
1:24:19 prompts.chat the reason I have that up
1:24:22 there is the the top half of that
1:24:24 document we'll talk to you sort of in
1:24:26 general about prompting and how it works
1:24:28 and why it works but the bottom half of
1:24:30 that document
1:24:31 is just a series of prompts that you can
1:24:34 copy and paste that are like you know
1:24:36 act like an attorney act like a
1:24:38 psychologist act like a an agriculture
1:24:41 specialist right act like a farmer act
1:24:43 like a you know someone who understands
1:24:45 soil nutrients engineering I don't
1:24:48 whatever the [ __ ] it is
1:24:50 and you can you can just go find one
1:24:52 that's there but you can also just learn
1:24:54 the basic concept of
1:24:56 when you tell chat GPT to act like a
1:24:59 very specific persona it gets better and
1:25:03 it gets better at that thing
1:25:05 um so what I would say is don't worry so
1:25:09 much about chat GPT like like what do I
1:25:12 do with the tool I would start more with
1:25:14 what are you trying to do in agriculture
1:25:16 what are you trying to do with legal and
1:25:18 psychology start with a specific problem
1:25:21 and then go into chat gbt and say I want
1:25:24 you to act like a lawyer that can help
1:25:26 with this sort of problem and then
1:25:27 describe your problem
1:25:29 and say
1:25:30 you know do you understand you know what
1:25:33 I'm trying to achieve and is this
1:25:34 something you can help me with like
1:25:36 you're going to literally talk to it
1:25:37 like it's a person it's going to be
1:25:39 weird and then it'll say yes and it'll
1:25:41 probably give you some advice and then
1:25:42 start having a dialogue with it so
1:25:44 basically set it up as the expert and
1:25:47 then start interacting with it to solve
1:25:49 a specific problem in those three areas
1:25:51 that you talk about
1:25:52 am I late for the class Professor
1:25:54 Shannon you are not late Emilio's wife
1:25:56 although I've been doing this for about
1:25:57 an hour now but I I do have a Father's
1:26:00 Day whiskey that the boys bought me so
1:26:02 I'm I'm sipping on that which is nice
1:26:06 mm-hmm
1:26:08 all right
1:26:11 that way I can give
1:26:13 them all my work Secrets yeah so so that
1:26:16 is an issue
1:26:18 hang on so what did he's talking about
1:26:21 there now he's doing it in a smarmy
1:26:23 trolley kind of way
1:26:25 which is okay if you're gonna troll here
1:26:27 at least be creative with it
1:26:29 um
1:26:30 so because you have a chat history here
1:26:33 in in so I met chat GPT and this this
1:26:36 whole sort of list here is my chat
1:26:38 history
1:26:39 and um openai is using our interactions
1:26:44 with it like why we get to use these
1:26:45 tools for free is they're using our
1:26:47 interactions to fine-tune their models
1:26:49 to make their models better
1:26:50 if you want to use this for work
1:26:54 you can go into your settings and
1:26:56 there's a data controls section
1:26:58 and you can turn off chat history and
1:27:01 training
1:27:02 and then it won't keep your chat history
1:27:04 and it won't allow open AI to train on
1:27:08 your data so that's one step more secure
1:27:10 open AI is also coming out with a
1:27:12 business version of chat GPT that'll
1:27:14 have higher security and privacy
1:27:17 the the warning I'll give you that if
1:27:19 you turn this off you will lose all of
1:27:22 your chat history which is why I haven't
1:27:23 turned it off so there's an export data
1:27:26 function here so if you're going to turn
1:27:28 this off and and make this safer to use
1:27:31 at work
1:27:32 make sure you export your current chat
1:27:34 history first so that you have it so
1:27:36 that's there but or start to use the
1:27:40 back end so you can also go to
1:27:44 if you've got a
1:27:46 platform.openai.com playground this is
1:27:50 the API backend for chat gbt where
1:27:52 you've got different modes that you can
1:27:54 play in you've got different models you
1:27:56 can use
1:27:57 you can change the the temperature which
1:28:00 is like the creativeness of chat so
1:28:04 um so this is a step more secure than
1:28:06 chat GPT but
1:28:08 hope that helps for the chat EVT thing
1:28:11 turn off your training and chat history
1:28:14 and then it will be a little bit less uh
1:28:18 whatever you call it janky from a
1:28:20 security standpoint
1:28:27 all right we're winding down here it's
1:28:30 good appreciate y'all being here if
1:28:32 you're new here my name is Kyle Shannon
1:28:33 this AI learning lab uh chat.openai.com
1:28:37 is the official chatgpt website go play
1:28:39 there if you haven't done that yet
1:28:44 can schools actually detect chat gbt
1:28:47 usage they can but then students can get
1:28:50 around it by using anti-che EBT
1:28:52 detection
1:28:53 fixers and it's a war and and quite
1:28:56 honestly any school that is trying to
1:28:59 prevent people from using this tool
1:29:01 that's going to be ubiquitous are very
1:29:04 short-sighted right now so I think it's
1:29:05 actually a blip in our educational
1:29:09 um
1:29:09 situation where where like I don't think
1:29:13 it's going to be that that long-lived
1:29:15 um
1:29:17 chat CPT is going to be rolled into
1:29:19 Microsoft Word and like all of the
1:29:20 office suite it's going to be rolled
1:29:22 into all the Google Docs it's going to
1:29:23 be everywhere it's in all the note
1:29:25 taking apps
1:29:27 um so
1:29:28 yeah so I so so yes there are schools
1:29:33 that can detect it students can get
1:29:36 around it the smarter schools are saying
1:29:38 hey this is a tool that all of our
1:29:40 students are going to be using let's
1:29:41 figure out how we incorporate it into
1:29:43 the classroom and change our engagement
1:29:45 with students to start you know dealing
1:29:47 with things like critical thinking like
1:29:49 okay if Chachi BT can write a paper
1:29:51 what's good about it what's bad about it
1:29:53 what are hallucinations
1:29:55 um how do I get better at working with
1:29:58 these tools rather than you know just
1:29:59 pretending they don't exist right so
1:30:01 this was this is no
1:30:04 similar to when the calculator
1:30:06 was invented math departments picketed
1:30:09 at universities don't let the kids use
1:30:12 the calculator
1:30:14 pen and paper only
1:30:16 so we're very much in that place right
1:30:18 now with it
1:30:21 hey Kyle you are blowing my mind again
1:30:25 you are fun to watch High cheers glad
1:30:29 you're high
1:30:30 I'm not quite buzzed yet but but you
1:30:33 know it's fun to watch hi that's great
1:30:37 oh
1:30:40 I spend too much time on being white too
1:30:43 hard it's too much time being white too
1:30:45 hard take it easy on yourself
1:30:48 there are always I don't know quite what
1:30:50 you mean
1:30:51 being white too hard
1:30:55 I'm just I am white I don't know what to
1:30:59 tell you and I'll tell you boss but I'm
1:31:02 only like there are always minorities to
1:31:04 press on take it easy dude what
1:31:07 I don't know what you're talking about
1:31:08 yeah please don't bring kind of I don't
1:31:10 want the divisive stuff in here at all
1:31:12 I'm not interested in it I think it's
1:31:14 boring
1:31:14 I am ready I am ready for a country to
1:31:17 move on
1:31:18 past that [ __ ]
1:31:20 um do you think a lot of people will
1:31:21 lose their jobs yes or they transition
1:31:24 to other careers yes so I so I think
1:31:27 what's going to happen is this I think
1:31:29 that
1:31:30 anyone who spends the next year or two
1:31:32 getting at all involved with AI is going
1:31:35 to help
1:31:36 guard against the disruption but the
1:31:38 disruption is absolutely coming
1:31:41 I think my sense is there's going to be
1:31:44 a two to three year period where there's
1:31:46 massive disruption that might be bad
1:31:47 enough that the government has to step
1:31:49 in and do Universal basic income or you
1:31:51 know the thing that Andrew Yang was
1:31:52 talking about something like that I
1:31:54 think it'll be short term because
1:31:57 what we can't see yet is the jobs that
1:32:01 will be invented because of these tools
1:32:05 right so the tools that do the
1:32:06 disrupting also make other things
1:32:08 possible and other also make other kinds
1:32:11 of companies possible and a lot of the
1:32:13 displaced people
1:32:15 while it's going to be painful and
1:32:17 [ __ ] awful initially might go well
1:32:20 that was you know I didn't really like
1:32:23 that job anyway or whatever but now I've
1:32:25 got access to these
1:32:30 thank you
1:32:49 it's good
1:32:50 it's not planted but it's good
1:32:59 all right I'm back I'm back
1:33:04 [Music]
1:33:05 all right
1:33:08 I just asked chat GPT and it says it
1:33:11 doesn't retain memories or specific
1:33:14 interactions
1:33:15 so it retains the memory within its own
1:33:18 chat session it doesn't have memory
1:33:20 across chat sessions
1:33:22 So within it if I say
1:33:25 um
1:33:26 here's the business I have and here's
1:33:29 the challenge I'm facing can you help me
1:33:31 with that it will remember there's kind
1:33:33 of this sliding window of like 3000
1:33:35 words where it'll have memory of that
1:33:38 What's called the context window
1:33:41 within a single chat session but it
1:33:43 won't remember that stuff across
1:33:45 sessions
1:33:46 versions of these tools will come along
1:33:49 that have all that they're just not here
1:33:51 now
1:33:52 all right
1:33:55 more freed up to create can AI help
1:33:58 design my website yes you want to play
1:34:00 with that let's go play with that that's
1:34:02 pretty cool
1:34:03 let me flip my camera
1:34:07 and I'm going to show you something cool
1:34:09 here okay
1:34:12 so we're gonna go over one of the sites
1:34:14 that I talk about is Poe
1:34:16 po.com
1:34:18 so I talked before there's these
1:34:20 different models here there's gpt4
1:34:21 Claude Claude instant right the the
1:34:24 Claude instant 100K that's the one where
1:34:25 you can put in like 75 000 Word
1:34:27 documents super cool
1:34:29 there's also bots so you can make your
1:34:31 own bot down here there's create a bot
1:34:33 so you can create your own bot and you
1:34:35 can explore Bots
1:34:37 um a couple of the Bots that I've been
1:34:39 playing with
1:34:41 are mid-journey Bots
1:34:44 so let's see
1:34:47 MJ Captain so this is one
1:34:50 um one of the guys that was in these
1:34:52 lives actually created this one and told
1:34:54 me about it so I'm going to clear the
1:34:55 context so this little sweep button
1:34:57 clears the context
1:34:59 and I'm going to say
1:35:02 [Music]
1:35:03 um
1:35:05 create let's say a
1:35:09 web interface let's say a
1:35:13 modern web interface
1:35:17 for a
1:35:21 um I don't know what kind of business
1:35:22 you have for a flower shop let's call it
1:35:27 and I'm just gonna hit return and then
1:35:29 it's going to rewrite that into it into
1:35:32 a good mid-journey prompt
1:35:35 so it says a website for a flower shop
1:35:37 with minimalist design using a pastel
1:35:40 color palette inspired by Dieter ROMs
1:35:42 David Carson flat design
1:35:45 designed for mobile first with an
1:35:47 emphasis on floral imagery 16x9 is the
1:35:51 aspect ratio no cluttered no outdated no
1:35:55 busy
1:35:56 um style of 500 I don't know what c is
1:35:59 and then version 5.1 okay so I'm just
1:36:01 going to copy this
1:36:03 then we're going to pop over to Discord
1:36:05 there's there's my superhero that I do
1:36:09 that I did of of my face when someone
1:36:12 asked me about that earlier okay so now
1:36:14 in your Discord so if you haven't been
1:36:15 to Discord this is Discord you got
1:36:17 different servers down the left and then
1:36:19 I'm in the AI Salon server and then
1:36:21 you've got channels within here and I'm
1:36:23 in the mid-journey channel of our server
1:36:26 and the way you make a mid-journey image
1:36:28 is you type slash imagine
1:36:30 and then there's a prompt and I'm just
1:36:32 going to paste in that prompt that was
1:36:33 generated from that bot you don't need
1:36:36 to use that but you can just type your
1:36:38 own [ __ ] but I just thought it'd be cool
1:36:40 to show that and so now we'll just wait
1:36:42 and it's going to generate a web design
1:36:44 now the thing about these web designs
1:36:46 they will probably look pretty good the
1:36:48 the
1:36:49 um the mid Journey web design stuff it
1:36:52 does they tend to be pretty remarkable
1:36:54 like I you know I I tend to be pretty
1:36:57 impressed they're not layered files
1:37:00 right this is not creating an actual web
1:37:01 design so if you're a web designer
1:37:03 you're going to take this design you're
1:37:04 going to toss it into figma or whatever
1:37:06 and you're gonna you know extract it out
1:37:09 now there's going to be
1:37:10 things in the future that can do this
1:37:13 and probably at futurepedia.io there's
1:37:17 there's stuff that can already do this
1:37:19 so these are looking good the pastel
1:37:22 color palette is starting to emerge the
1:37:24 flowers are starting to emerge
1:37:26 I'm I'm liking what I'm seeing from a
1:37:28 typography standpoint
1:37:34 um
1:37:40 so that says website so let's see
1:37:45 all right these are these are actually
1:37:47 kind of more like Instagram
1:37:50 images than websites but they're not bad
1:37:54 but what we could do is we could say
1:37:57 um instead of website let's copy all
1:38:00 this
1:38:03 and we're going to do slash imagine
1:38:07 and I'm gonna say
1:38:09 um a web
1:38:12 interface
1:38:15 for a flower shop with minimalist
1:38:17 minimalist design
1:38:20 and
1:38:22 a number of options I don't know what
1:38:26 that if that's going to help but we'll
1:38:28 try that we'll see what that does
1:38:30 um
1:38:31 what I would do is I would go to
1:38:35 um
1:38:36 I'd go to Reddit
1:38:39 um or even Facebook and and start to
1:38:41 find communities that are doing
1:38:44 interface design oh that's what I should
1:38:46 have put I should have put interface
1:38:47 design start to look for communities
1:38:50 that that are doing interface design in
1:38:53 mid-journey oh I have another I have
1:38:55 another place we can go that might be
1:38:57 interesting that one in the lower right
1:38:59 starting to look interesting it's giving
1:39:00 us three options a flower per option
1:39:02 that's actually kind of a cool
1:39:04 cool idea
1:39:12 and if you notice that the images start
1:39:15 um like as these blurry things and then
1:39:17 get increasingly less blurry that's how
1:39:19 the diffusion model works that's how
1:39:21 these things are actually creating they
1:39:23 start as essentially kind of this random
1:39:25 noise image and then they denoise over a
1:39:27 series of steps
1:39:29 in the direction of whatever you asked
1:39:31 for
1:39:33 so that one's pretty cool that one in
1:39:35 the lower right
1:39:40 you know that'd be a cool website
1:39:42 actually
1:39:43 you know three three different options
1:39:45 that kind of look like cards I don't
1:39:46 know that's kind of cool
1:39:48 um but let me go let me pop over to
1:39:50 Facebook for a second and we'll go to
1:39:51 the mid Journey official
1:39:54 website
1:39:56 um
1:40:01 uh web interface
1:40:09 so if you go to Facebook there's there's
1:40:12 a mid-journey official group
1:40:16 and then like here's a bunch of
1:40:18 interface stuff that people are doing
1:40:25 those are not interfaces I don't think
1:40:28 I've been the right thing hang on let me
1:40:30 let me find mid-journey official
1:40:34 let me search and let me search for
1:40:37 interface
1:40:42 let me do software
1:40:45 interface
1:40:48 that's not software interface
1:40:56 that's sort of one there's one
1:41:02 um but anyway yeah you can do it my
1:41:04 examples were not great but
1:41:07 um
1:41:08 I'd say go to Reddit and and look for
1:41:11 people that are doing interface design
1:41:12 like oh oh ux design that that's the
1:41:16 term that I was thinking of let's see
1:41:21 how about
1:41:23 okay I gotta do slash imagine
1:41:29 and paste that in um a web
1:41:33 ux design system
1:41:38 for a flower shop let's see what that
1:41:40 does maybe that'll give us something
1:41:42 better
1:41:43 but anyway so the the the short answer
1:41:46 is yes it can do that
1:41:48 um go to futurepedia.io and look up the
1:41:52 AI web design tools because I'm sure
1:41:54 there are a bunch of them there now that
1:41:57 weren't there there before where you can
1:41:58 just describe a website and it'll go
1:42:00 build it for you because I know I know
1:42:02 that's a thing and those those tools are
1:42:05 out there
1:42:07 um so I don't know how teachers would
1:42:09 find out so they can actually there's
1:42:10 actually tools Isabella that you can
1:42:12 take a piece of writing pop it into the
1:42:15 tool and it'll say was this generated by
1:42:16 chat GPT and it can figure it out but
1:42:19 there are other tools that you can take
1:42:21 that piece of writing pop it in that
1:42:23 tool and have it rewrite it so that it
1:42:25 can't be discovered by those tools so
1:42:27 it's just a silly war going back and
1:42:29 forth right now
1:42:33 who's going to check that AI is doing
1:42:35 accurate work uh the people creating it
1:42:38 need to check that so there's a there's
1:42:40 a funny example I don't know if it's
1:42:42 funny the people in the case but a
1:42:45 lawyer two weeks ago wrote a legal brief
1:42:47 and turned it in in court
1:42:50 and the judge
1:42:52 um
1:42:53 check the sources in the legal brief and
1:42:56 they were all fake they were all made up
1:42:57 so so that lawyer got in trouble and
1:43:00 that lawyer should get in trouble right
1:43:01 it's on
1:43:02 it's on the lawyer to do the checking so
1:43:07 if you're doing something that has to be
1:43:08 accurate check it also use something
1:43:10 like bing.com I don't know why this is
1:43:13 waiting to start come on mid-journey do
1:43:15 your thing
1:43:16 um
1:43:17 use something like bing.com that's
1:43:20 connected to the internet and it will
1:43:21 provide citations and check the
1:43:23 citations and see if they're actually
1:43:24 right
1:43:25 so it's it's on you now to your point
1:43:28 Glenn
1:43:29 in the future there's going to be whole
1:43:31 Industries built around validating
1:43:34 um content that's generated out of these
1:43:36 generative AI systems to make sure that
1:43:37 they're accurate it's just not there
1:43:38 right now
1:43:40 I understand the super pressure okay
1:43:42 we're not going to talk about that
1:43:46 oh my God same not sure what that's
1:43:48 about
1:43:54 uh
1:43:58 so how do we transition with the loss of
1:44:01 jobs
1:44:02 um
1:44:03 be one of the people that's educated
1:44:05 about generative AI in your category so
1:44:09 like even if you're a copywriter and
1:44:11 copywriting gets hammered I don't know
1:44:14 why mid journey is not going why is it
1:44:17 not go wig
1:44:19 um I'll come back to that if it does
1:44:21 anything cool I'll come back to it
1:44:24 um
1:44:26 I don't I don't know that there's an
1:44:28 antidote to having whatever industry
1:44:32 you're in be affected by this
1:44:35 the only thing I know is if you're one
1:44:37 of the people that is
1:44:40 um
1:44:41 is up to speed on AI and AI comes in and
1:44:44 the company's like we got to figure this
1:44:46 AI thing out does anybody know anything
1:44:48 about this if you're one of those people
1:44:50 that raises your hand or people already
1:44:52 know is doing this you're going to be in
1:44:53 a better position than someone who's not
1:44:55 so like that's that's my only thought
1:44:57 right now is is get get as aggressive as
1:45:00 possible about learning these tools
1:45:03 and and uh
1:45:06 that that's that's going to help with
1:45:09 the transition
1:45:10 how do we use utilize AI on keywords
1:45:14 it's great at keywords give it a bunch
1:45:16 of
1:45:17 give it give it a bunch of content and
1:45:20 have it pull out keywords for you or or
1:45:23 um
1:45:25 yeah have it write you you know keyword
1:45:28 heavy blog posts and give it a set of
1:45:31 keywords and have it write content that
1:45:33 dynamically incorporates those and
1:45:35 dynamically incorporates hyperlinks of
1:45:37 your content within it
1:45:39 um it's great at that stuff any of that
1:45:40 kind of stuff it's really good what does
1:45:42 storyvine do so storybine is my company
1:45:45 um it's a it's a video storytelling
1:45:47 platform so we've got uh an app that
1:45:51 guides an end user through telling a
1:45:52 story so we do a lot of work in
1:45:54 healthcare and Pharma for example so if
1:45:56 Pfizer wants one of their patients to
1:45:58 talk about what it's like to have a
1:45:59 certain disease our app will actually
1:46:01 guide the patient step by step through
1:46:03 the storytelling process so introduce
1:46:05 yourself here you know take a picture of
1:46:07 yourself what did it feel like when you
1:46:09 were diagnosed how have you been
1:46:11 treating it how are you you know what's
1:46:12 going on today what are your hopes and
1:46:14 dreams for the future so they answer
1:46:15 questions in our app that goes up to the
1:46:17 cloud five minutes later you got a fully
1:46:19 edited video
1:46:20 and that's the that's we call it auto
1:46:23 magic editing um and so that's what the
1:46:26 platform does we've been around for 11
1:46:28 years like I said we're mostly in
1:46:29 healthcare and Pharma
1:46:32 um
1:46:33 I'm looking at all this AI stuff and
1:46:35 there's a lot of what we do that the
1:46:37 value of it goes to zero like the
1:46:38 automation everything's going to be
1:46:40 automated in a year or two
1:46:43 but there's some things that we do like
1:46:45 authentic human storytelling we're
1:46:47 really good at that I think the value of
1:46:49 that goes up in a world where there's
1:46:50 infinite content generation right so um
1:46:53 so we'll see
1:46:54 but but that's what I'm in the middle of
1:46:56 trying to figure out is where's our
1:46:57 value proposition what goes up what goes
1:46:59 down how do we survive
1:47:02 because we're going to be slapped up
1:47:04 side the set upside our head um just
1:47:07 like everyone else is
1:47:10 I'm just trying to be on top of it
1:47:14 um is your company hiring software
1:47:16 Engineers not right now
1:47:18 we are not hiring listening okay
1:47:22 does chatgpt do images no but
1:47:26 mid-journey does stable diffusion does
1:47:28 Adobe Firefly does and Dolly
1:47:32 does
1:47:34 gpt4 is multi-modal but it's not live
1:47:37 yet and what that means is
1:47:39 gpt4 you'll be able to upload an image
1:47:42 and it can analyze that image and so you
1:47:44 can interact with gpt4 with images but
1:47:47 it's not live yet
1:47:50 Minneapolis thoughts on Google
1:47:52 generative AI course I've heard it's
1:47:54 good is it something worth learning for
1:47:56 someone not in Tech yes so Risa this is
1:48:00 a really interesting thing there's kind
1:48:02 of two flavors of
1:48:04 do you want to learn about AI one is
1:48:06 doing about do you want to learn about
1:48:07 AI to go into machine learning and sort
1:48:09 of the science of AI and the other one
1:48:11 is do you want to learn about AI to use
1:48:13 generative AI
1:48:15 as a tool in business that's a very
1:48:17 different thing right so
1:48:20 um I'm not a machine learning person I'm
1:48:23 not a scientist I have a degree in
1:48:25 acting I but I'm an entrepreneur
1:48:28 and I get really passionate about new
1:48:31 technologies that I think can transform
1:48:33 how we do business that's sort of what
1:48:35 I've spent my career doing
1:48:37 and so I think this is a big one and so
1:48:39 I'm not interested in learning
1:48:41 I I like I only want to know enough
1:48:44 about how the tools work so I can
1:48:45 understand what they are and what their
1:48:47 limitations are and what their
1:48:48 possibility is beyond that I don't give
1:48:51 a [ __ ] I'll let the Uber Geeks and the
1:48:53 engineers do all that I'm interested in
1:48:55 the other side of this so I haven't
1:48:57 looked at the Google's generative AI
1:48:59 course I assume it's more on the latter
1:49:01 side of this inflection point that's
1:49:02 about how do you use these tools in
1:49:04 business so if it's that I would say
1:49:06 it's very worth
1:49:08 learning it it'll teach you some of the
1:49:10 basics it'll get you up to speed I'd say
1:49:13 more important than
1:49:16 taking what you think is the perfect
1:49:18 class start playing start playing go to
1:49:23 chat.openai.com start playing start
1:49:25 learning
1:49:27 just throw pasta at the wall see what
1:49:29 it's good at see what it's bad at
1:49:31 keep playing share what you're learning
1:49:33 do projects for people for free one of
1:49:36 the women that has been in the lives she
1:49:39 she said
1:49:41 um she started learning this stuff and
1:49:42 she set a goal for herself to teach a
1:49:44 thousand people
1:49:46 about generative AI whatever she's
1:49:48 learning she set a goal to teach a
1:49:49 thousand people she's now making ten
1:49:51 thousand dollars per engagement to teach
1:49:53 companies how to do this because she
1:49:56 started learning it and just putting
1:49:57 herself out there so I'd say just get
1:49:59 curious learn learn get hungry play tell
1:50:05 people about it do projects for people
1:50:07 tell other people how to do it and good
1:50:09 things will happen I I like that I'm
1:50:12 confident of I think there's a solid
1:50:14 two-year period where where if you're
1:50:16 decent at this stuff and sort of put
1:50:19 yourself out there as being decent at
1:50:20 this stuff not like in a braggy way but
1:50:22 like in a I'm excited about this way
1:50:24 good things will happen
1:50:27 Google is so five minutes ago Google's
1:50:30 struggling to catch up right now
1:50:32 um machine learning is going to be the
1:50:34 problem Singularity will happen when
1:50:37 they learn how to code it already knows
1:50:38 how to code
1:50:39 Singularity might happen when it learns
1:50:41 how to autonomously code and that's what
1:50:44 Auto GPT is
1:50:46 [Music]
1:50:48 um it's not aware of what it does right
1:50:50 now so so I think we have I think we
1:50:52 have some time until the singularity my
1:50:54 guess is five-ish years
1:50:57 that's sort of Ray Kurzweil say in 2029
1:51:00 some big uh AI scientists are saying
1:51:03 2030 so that's that's probably about
1:51:05 right what are the biggest moats in the
1:51:07 AI space so so what you're talking about
1:51:09 there so moats are like what what are
1:51:11 competitive advantages
1:51:13 there's a Google engineer that just
1:51:15 wrote a piece
1:51:17 it was an internal memo that was
1:51:19 apparently leaked I you know I don't I
1:51:21 don't know if it was leaked or if it was
1:51:23 leaked
1:51:26 um but it basically talks about the fact
1:51:27 that
1:51:29 Google and openai don't have emote the
1:51:32 title of the piece was we have no moat
1:51:34 and what what they're basically saying
1:51:36 is that for a hundred dollars the
1:51:38 developer Community can train one of
1:51:40 these open source large language models
1:51:42 to essentially be as good as what it
1:51:44 took Google like a year and 10 million
1:51:47 dollars to train
1:51:49 um and so that there's no what he's
1:51:50 saying is there's no competitive
1:51:52 competitive Advantage for us spending
1:51:54 all this money to make these core models
1:51:56 because the open source Community is
1:51:58 just knocking this [ __ ] off like that
1:52:01 um
1:52:02 I think it's one of the big concerns of
1:52:04 investors right now is that almost any
1:52:07 company you're going to be able to
1:52:09 replicate it essentially easily
1:52:11 overnight without a lot of money
1:52:14 um I think it's going to be a universal
1:52:15 problem so
1:52:18 so I don't know I for me I think I think
1:52:21 where companies may have some
1:52:23 competitive Advantage is with unique
1:52:25 data sets so there was a you've heard of
1:52:28 Bloomberg the financial services they
1:52:31 just created a thing called Bloomberg
1:52:33 GPT which is a GPT
1:52:36 tool sitting on top of all of their
1:52:38 proprietary data which they charge up
1:52:41 shitload of money for right but if you
1:52:44 want access to that data you'll pay for
1:52:47 that and and that's something that they
1:52:49 can they can control so that might be it
1:52:53 but I don't know I think it's honestly a
1:52:55 really big it's it's a big challenge
1:52:59 um
1:53:00 I think being flexible I think moving
1:53:03 fast I think experimenting a lot not
1:53:06 getting too lucked into a single
1:53:08 um model because things are going to
1:53:10 change really quickly so
1:53:13 we had a better answer for that
1:53:22 um
1:53:27 all right so many good questions here
1:53:29 they're great questions here
1:53:31 all right thoughts about Runway so
1:53:33 Runway is Runway ml they've got two
1:53:36 video
1:53:37 creation tools one called gen one one
1:53:39 called Gen 2 gen 1 allows you to upload
1:53:42 a three or four second video clip
1:53:45 of like a person walking and then you
1:53:48 can describe make it a claymation
1:53:50 animation and it'll be the person
1:53:51 walking but they'll be in claymation
1:53:53 that's gen one gen two is you just
1:53:56 describe a video woman in a field with
1:53:59 the breeze blowing
1:54:01 there's a woman in the field with a
1:54:02 breeze blowing
1:54:04 um
1:54:05 they've got a head start they're you
1:54:08 know they're they're
1:54:10 well their tools are remarkable the
1:54:12 videos have a lot of artifacts in them
1:54:14 like so I think a year from now they're
1:54:15 going to be remarkable when you can do a
1:54:17 30 second video clip
1:54:20 um that doesn't have artifacts in it and
1:54:21 it looks pretty realistic they're going
1:54:23 to be amazing so I I think they're
1:54:25 they're definitely ones to watch they're
1:54:27 they're definitely they've got the the
1:54:30 lead in the video space right now for
1:54:32 for generating
1:54:34 generative AI for video
1:54:36 um I think there's gonna be a lot of
1:54:38 competitors in the space and I think
1:54:41 it's going to start moving quite quickly
1:54:42 what's the best AI for character design
1:54:44 there was just one that came out called
1:54:47 Taffy Taffy hang on let me show you
1:54:55 [Music]
1:54:58 let's see is it called get Taffy I think
1:55:02 you go to get taffy.com
1:55:04 get t-a-f-i
1:55:07 is that right no
1:55:10 oh please
1:55:12 Taffy might not be the right thing
1:55:15 Taffy character
1:55:19 make taffy make taffy
1:55:21 go to make taffy.com
1:55:26 and then AI
1:55:30 yeah so they've got a
1:55:45 so you describe the character so what
1:55:48 they say give him a uniform that looks
1:55:50 like he's from a sci-fi film Bang
1:55:55 woman with mid hair length
1:56:02 so this isn't out yet but you can
1:56:04 register for the beta here so make
1:56:05 taffy.com
1:56:07 um I don't know if they're the best the
1:56:08 the the interesting thing about these
1:56:10 guys is
1:56:12 um
1:56:13 they've been doing
1:56:15 character design engines for a long time
1:56:19 and they're and now they're putting
1:56:21 generative AI sitting on top of it so I
1:56:24 think you're going to start to see that
1:56:25 with unity and with with Unreal Engine
1:56:27 and with all these 3D tools that are
1:56:29 incredibly complicated throwing a
1:56:32 generative AI layer on top of it to
1:56:34 basically just deal with all the
1:56:35 complexity so the interface of these
1:56:38 incredibly powerful tools just becomes
1:56:41 language just becomes your ability to
1:56:43 ask for what you want
1:56:50 all right I gotta go soon
1:56:52 I gotta go it's been hours it's been
1:56:54 hours man
1:56:56 it's been hours hours I see Google
1:56:59 putting bard
1:57:01 in Google home
1:57:03 Bart's gonna be in everything
1:57:05 so Bart's gonna be in Office 365 it's
1:57:07 going to be in Google home it's going to
1:57:09 be in the car thing Mercedes just
1:57:11 announced they're putting chat GPT in
1:57:13 the the card driving thingy I'm sure
1:57:16 Apple will do that when they come out
1:57:18 with theirs they will come out with one
1:57:20 because they've got the hardware what's
1:57:22 cool about Apple if you if you pay close
1:57:24 attention to their Vision Pro
1:57:25 announcement
1:57:28 two two things that I thought were
1:57:30 baller moves they didn't use the word
1:57:33 metaverse once and they didn't use the
1:57:35 phrase AI once
1:57:37 but they used machine learning and they
1:57:40 used vectors and they used
1:57:43 um inferences they used a lot of the
1:57:45 terminology of machine learning and what
1:57:47 they were basically saying is they can
1:57:49 do all that [ __ ] on chip
1:57:52 now why is that important because right
1:57:55 now
1:57:56 a lot of the AI stuff that we're playing
1:57:58 with we're sending our data to open AI
1:58:01 which they're using it for training and
1:58:03 and depending on how secure or not that
1:58:07 data is that's risky if you're actually
1:58:10 doing all of the AI processing locally
1:58:12 on the chip on your phone on your
1:58:14 goggles on your watch whatever it might
1:58:16 be
1:58:18 um
1:58:19 their ability to do high privacy
1:58:23 generative Ai and autonomous agents
1:58:27 they're gonna like I I don't know that
1:58:29 anyone's going to be able to compete
1:58:30 with them because they're the ones that
1:58:32 have always taken this privacy [ __ ]
1:58:34 really seriously and so the fact that
1:58:36 they've got you know privacy chips on
1:58:38 the devices and they're going to be able
1:58:40 to do a lot of the machine learning
1:58:42 processing locally means they're going
1:58:46 to be I think they're gonna
1:58:48 they're gonna show up with a bang but I
1:58:50 agree with you I think I think these
1:58:52 tools are going to be incorporated into
1:58:54 everything like it's going to be you
1:58:55 know how the internet is in everything
1:58:57 right now like that's what's going to
1:58:58 happen with this these generative AI
1:59:00 tools they're just going to be part of
1:59:02 the fabric of how we interact with
1:59:04 machines moving forward
1:59:06 Google made one of the biggest blunders
1:59:08 since the new Coke yeah
1:59:10 Google [ __ ] invented the Transformer
1:59:12 and then didn't jump on it and they let
1:59:15 other people do it and open AI took the
1:59:18 [ __ ] seriously and they did GPT one gpt2
1:59:21 gpt3 which started to get a bunch of
1:59:24 press and people were like holy [ __ ]
1:59:25 this gpt3 thing is really good
1:59:28 and then another year went by and
1:59:30 nothing from anyone
1:59:32 and then openai puts out chat gbt and it
1:59:35 goes from zero to 100 million users in
1:59:37 six weeks
1:59:38 so it wasn't like Google didn't have
1:59:40 fair warning of what was coming
1:59:43 but what they couldn't anticipate and
1:59:46 what Microsoft didn't think would happen
1:59:48 and even open AI didn't think would
1:59:49 happen nobody thought chat CPT would be
1:59:52 a hit
1:59:53 it was just a shitty little piece of
1:59:54 software they wrote in two weeks to put
1:59:57 out as a as a an example of what you can
2:00:00 do with this in conjunction with the
2:00:02 Microsoft announcement and it went to
2:00:04 100 million people in six weeks
2:00:07 caught everyone off guard and now
2:00:09 Google's on their heels going oh [ __ ] we
2:00:11 gotta catch up
2:00:14 so yes it is a big blunder but they're
2:00:17 big money with a lot of a big company
2:00:19 with a lot of [ __ ] money so you know
2:00:22 and they got a bunch of smart people
2:00:23 there so I I would not
2:00:26 um count them out just yet but they are
2:00:28 definitely on their heels
2:00:29 the company makes chat GPT is open AI
2:00:32 chat GPT doesn't focus on Graphics uh
2:00:34 but but the dolly can yeah that's
2:00:37 correct there's a there's right now all
2:00:39 of these tools are very disparate little
2:00:41 it's like we're we're in the spit and
2:00:43 duct tape phase of this part of the
2:00:45 industry right now
2:00:46 how would I use AI for getting movie
2:00:49 investors
2:00:50 um
2:00:52 [ __ ] I've written seven screenplays and
2:00:54 I didn't successfully get any movie
2:00:56 investors so so I'm not the right person
2:00:59 to ask for that I mean that's such a
2:01:02 business of relationships like what what
2:01:04 I would probably do
2:01:06 because I would have chat GPT give me
2:01:08 strategies on how to navigate Hollywood
2:01:11 politics
2:01:14 right because it's
2:01:16 part of it is just grinding out you know
2:01:20 sending out inquiry letters and pitches
2:01:23 and and doing all that [ __ ] but a lot of
2:01:25 it is who you know and how you know them
2:01:27 and how you get connected so I would
2:01:30 probably if I were getting back in that
2:01:32 game I I would take the
2:01:35 the game of Show Business a [ __ ] ton
2:01:38 more seriously and I'd have chat gbt
2:01:40 help me with that
2:01:42 old man gonna get old
2:01:48 all right I'm gonna do a few more
2:01:50 questions here and then I'm gonna go
2:01:53 Sunday night
2:01:55 we are all
2:01:57 doing what we're doing we're all hanging
2:01:59 out here
2:02:01 um I appreciate you hanging out with me
2:02:02 let's see
2:02:05 um it's cool you're each talking to each
2:02:07 other there are false positives with
2:02:09 those Checkers yeah there was some
2:02:11 someone was asking about the university
2:02:13 education Checkers how do they check the
2:02:15 things there are some University dude
2:02:18 um that that he ran all of the people's
2:02:21 papers through a a GPT Checker and a
2:02:24 hundred percent of them tested tested
2:02:27 positive for GPT and so he failed
2:02:30 everyone and like you know most of the
2:02:32 class hadn't used chat GPT some of them
2:02:34 didn't even know what it was
2:02:36 so yeah that's just we're that's just a
2:02:39 blip I would just ignore that [ __ ] for
2:02:42 for now best way to access unrestricted
2:02:46 chat GPT go to the open source community
2:02:48 so
2:02:50 um there's a thing called GPT for all
2:02:52 which I think you install it locally on
2:02:56 your machine and then it gives you
2:02:57 access to a bunch of Open Source large
2:02:59 language models
2:03:00 most of them don't have any of the
2:03:02 safety guard rails on them so that's how
2:03:05 you can do it
2:03:07 um I don't have anyone to suggest
2:03:08 because like a new one comes out every
2:03:10 other day and I can't keep up with it so
2:03:12 I said [ __ ] it so I'm gonna let oh like
2:03:16 like if you want to geek out on open
2:03:17 source models just go to Reddit open
2:03:19 source llm and just get into those
2:03:21 communities there's all sorts of
2:03:23 projects there that are listen they're
2:03:24 doing remarkable [ __ ] things but I
2:03:26 can't keep up with it so I'm I'm not
2:03:28 playing over there MIT is offering an
2:03:31 online course teaching AI Basics is it
2:03:33 worth taking
2:03:34 um again I would just understand which
2:03:36 side of the AI fence you're on if you
2:03:38 want to go into the machine learning and
2:03:40 science side of the thing
2:03:41 that's probably what MIT is teaching
2:03:44 if it's something like Wharton School of
2:03:48 Business teaching it is probably more on
2:03:49 how you use the AI rather than building
2:03:51 it so just figure out which side of the
2:03:53 fence you're on and then find courses
2:03:55 that
2:03:56 that work with that and again I don't
2:03:58 think like if you just want sort of the
2:04:00 Baseline education those courses might
2:04:03 be good but anything sort of trying to
2:04:05 keep up with what's happening with these
2:04:06 tools just start playing with them just
2:04:09 start using them
2:04:10 is there any way for chat to remember my
2:04:12 preferences like giving me a glossary of
2:04:14 terms not right now and and that's
2:04:17 something that
2:04:19 I think I saw interface I think someone
2:04:22 leaked some interface from open AI that
2:04:25 they're going to have a profile
2:04:26 that you'll be able to put in you know a
2:04:29 bunch of [ __ ] about yourself that I can
2:04:31 always answer in context of who you are
2:04:34 um so I think that's coming it's not
2:04:35 there right now and it's a pain and it's
2:04:37 a pain in the ass
2:04:39 that it's not there
2:04:42 um I never liked Google back in the 90s
2:04:44 I preferred Alta Vista and sq's yeah
2:04:47 Alta Vista was baller and then and then
2:04:49 Google did their thing and they they
2:04:51 figured out how to do it a little bit
2:04:53 better than everyone else and then they
2:04:55 got to like 92 94 market share and then
2:04:59 they got sloppy
2:05:00 and somebody invented chat GPT
2:05:04 now the game has changed
2:05:07 so it is very possible that they end up
2:05:09 not on the top of the Heap anymore
2:05:11 what's it called Runway AI Runway ml is
2:05:14 what Runway is called
2:05:17 all right how do I join the AI Salon uh
2:05:19 the salon.ai will take you to a link
2:05:22 tree
2:05:23 um that's got a link to our Discord
2:05:25 server and our meetups we've got a we've
2:05:27 got a uh our next Meetup is
2:05:30 Tuesday night
2:05:32 and we have we have a guy named uh
2:05:34 Christian Robbins who's a chief creative
2:05:36 officer at a post-production house and
2:05:38 he's coming and he's going to talk about
2:05:41 um he's going to talk about how they're
2:05:43 incorporating AI into their entire
2:05:45 workflow in video post-production that's
2:05:47 that should be really interesting good
2:05:49 night this helped a lot good night I
2:05:51 must go meditate before bed yeah I know
2:05:54 I get it people are making money selling
2:05:57 customizable mid-journey prompts on Etsy
2:05:59 yeah I you know again
2:06:02 uh
2:06:03 I I would say rather than buying prompts
2:06:06 I mean a you can get them for free
2:06:07 there's they're all over the place for
2:06:09 free but I think right now it's actually
2:06:11 valuable to learn how to do it yourself
2:06:13 so just go play with prompting and and
2:06:17 learn about prompting but but I wouldn't
2:06:20 I don't know
2:06:22 I wouldn't pay for that [ __ ] right now
2:06:24 stuff's moving way too fast
2:06:28 um
2:06:31 can you show what's the screen behind
2:06:33 you oh yes sorry so
2:06:39 check that openai.com that's the
2:06:41 official chat CPT website bing.com
2:06:45 if you go to their cert or to their chat
2:06:48 button that's chatgpt4 for free
2:06:50 connected to the internet
2:06:52 poe.com is has chatgpt and some other
2:06:56 models in it and you can create your own
2:06:57 bot and play with other people's Bots
2:07:00 prompts.chat is will teach you about
2:07:03 prompting for chat GPT and then
2:07:05 futurepedia.io is a giant directory of
2:07:09 AI tools
2:07:11 so
2:07:15 that give you what you need
2:07:18 speaking of films what site channel app
2:07:21 do you recommend to learn about short
2:07:23 films wait short films and story
2:07:29 um I don't know
2:07:32 I don't I don't I don't know about
2:07:35 I haven't been in the writing game in
2:07:37 for [ __ ] ever but
2:07:40 unless you're talking about super
2:07:42 up-to-date screenplays like chat CPT 3.5
2:07:46 probably knows all the movies up until
2:07:49 September 2021 and then you could use
2:07:51 chechi VD connected to the internet
2:07:53 after that
2:07:54 like it knows
2:07:56 like if you want help with story you can
2:07:58 go there and say
2:08:00 I want you to I want you to give me
2:08:03 the story Frameworks from Sid Fields
2:08:06 save the cat and Joseph Campbell's
2:08:09 Heroes Journey
2:08:11 and it will give you all those and say
2:08:12 now create a hybrid that that mixes the
2:08:15 best of the concepts from those three
2:08:17 story Frameworks into my own custom
2:08:19 story framework and it'll do that and
2:08:21 then you can have it you can just
2:08:23 basically have it help you create a
2:08:25 story framework and then go fill it in
2:08:27 with your story
2:08:28 um
2:08:29 I don't know that you need to go
2:08:31 anywhere like I kind of feel like we all
2:08:32 now have a professor in our pocket
2:08:35 like chat gbt can just teach you that
2:08:38 [ __ ] and then do it for you and then
2:08:40 like collaborate with you to actually
2:08:41 bring something to life
2:08:44 can check EBT helped me develop a
2:08:46 likable personality
2:08:50 actually you probably can like if you
2:08:53 tell it what you don't like about your
2:08:55 personality tell it to come up with like
2:08:57 a 30-day plan for for improving how do I
2:09:00 be less of a
2:09:02 whatever whatever you're not happy with
2:09:04 it could probably actually help with
2:09:05 that that would be a good experiment
2:09:07 actually right like maybe like maybe yes
2:09:12 um speaking of films okay we did that
2:09:14 one
2:09:15 um in business they tell you to use
2:09:17 professionals and focus what you're good
2:09:18 at yep
2:09:20 is there a wait is there a chat
2:09:25 gpt4 app
2:09:28 so there are 8 000 chat GPT apps that
2:09:34 are all Faker apps
2:09:36 that are just trying to get your money
2:09:39 so on my phone I don't use the open AI
2:09:42 chat GPT app but there is one I'll tell
2:09:44 you how to get it what I do
2:09:46 is I go to that URL
2:09:49 in safari
2:09:51 on my iPhone and then I make it a
2:09:53 bookmark to that page so I have like one
2:09:55 of the icons at the bottom of my phone
2:09:57 is chat CPT and I just I'm just going to
2:09:59 the website
2:10:00 um if you want to use the official
2:10:03 chatgpt app
2:10:05 go to openai.com without the chat go to
2:10:09 openai.com go to products and go to
2:10:11 chatgpt and you'll see a link to the
2:10:14 official app
2:10:15 it's it's essentially impossible to find
2:10:17 in the app store because there's so many
2:10:19 fake ones
2:10:21 thanks for the knowledge it's time to
2:10:23 sleep it is time to sleep all right
2:10:25 first story
2:10:27 um all right one more question I found
2:10:30 that even with the apps wait even with
2:10:33 the apps help
2:10:34 we still must do I'm not sure what that
2:10:36 means
2:10:39 all right and and you keep asking about
2:10:41 the real chat GPT app go to openai.com
2:10:44 products chat gbt that's the official
2:10:46 app there'll be a link there it's iOS
2:10:49 only right now not Android all right
2:10:51 everybody thank you so much for hanging
2:10:52 out with me on a Sunday night Happy
2:10:54 Father's Day to everyone hope you're all
2:10:56 doing well and uh have a good night and
2:10:58 I'll I'll see you probably tomorrow
2:10:59 peace out everybody