aiLL Nov 13, 2023 - Harnessing AI's Creative Potential through Community & Curiosity

Description

Chapters

Transcript

0:01 [Music]
0:29 Tobias
0:30 [Music]
0:49 yo Edgar what's happening Source Camp
0:52 good to see you be Not Afraid thank you
0:54 thank
0:55 [Music]
0:57 you so I I played a little bit of a song
1:00 in the uh in the timer video so I asked
1:04 people if they knew which artist it was
1:06 so I'll play I'll play the song I'll do
1:10 a little sanging little little finger
1:13 picking little sanging hey Mr guitar
1:16 man what's up AI man not AI tonight it's
1:20 going to be all analog we're not doing
1:22 any of this digital [ __ ] except for the
1:25 guitar on screen that's you know
1:27 completely made
1:29 up
1:34 welcome welcome welcome
1:35 [Music]
1:50 everybody now I'm
1:52 [Music]
1:57 right-handed I think I got a here I can
2:00 flip my flip my camera which will be
2:03 weird for oh wait no that's not what I
2:04 wanted to do I wanted to
2:12 mirror so now is
2:15 it now does it look like I'm
2:19 [Music]
2:29 right-handed
2:34 [Music]
2:51 all
2:53 [Music]
2:59 right
3:05 [Music]
3:12 s in this lonely
3:14 Town wonder when things are going to
3:19 change dream my life
3:22 away seems these dreams have turned to a
3:25 bunch D
3:28 clads get my n
3:30 up but my past is pulling me
3:34 [Music]
3:36 down wondering how
3:39 long this black sheep going to stick
3:47 around somebody told me once before he
3:51 said you can never go home
3:54 again won't you leave say my things to
3:58 steal me away
4:00 yeah from the truth of who I am and what
4:03 I believe so I thanked him for his two
4:07 cents with a
4:08 handshake and some
4:10 sympathy
4:12 yeah packed up my blue
4:15 jeans and headed for this big
4:20 prize of my
4:23 freedom
4:25 bye-bye Black Sheep to the black sheep
4:29 of the
4:31 family
4:34 [Music]
4:35 by all I mean so very much to me yeah
4:41 byebye Black Sheep to my friends and my
4:48 family
4:50 [Music]
4:52 byebye going off and set my
4:55 soul set it
4:59 free
5:00 [Music]
5:02 three times there were a change
5:05 in did a little bit of
5:10 rearrange take a couple
5:13 chances more progress i' not
5:17 advances to the
5:20 priz love my
5:23 [Music]
5:28 freedom
5:33 y y y
5:36 [Music]
5:50 y Martin saon Black Sheep he sings it
5:57 good I sing it like you know
6:01 I learned
6:03 it all right let's get going oh yeah
6:06 thank you very much appreciate
6:12 [Laughter]
6:14 that all right let's get going let's get
6:17 going wait I got to flip my camera it's
6:19 weird I'm I'm everything's
6:22 backwards okay that's
6:25 better all right well start off with a
6:29 little music
6:31 get out the guitar light up
6:35 campfire I'm Harry
6:40 Cary all right is there anything you
6:42 can't do oh so much including
6:45 [Laughter]
6:50 singing all right all right all right
6:52 all right all right thank you steo
6:54 that's very kind of you to say that's
6:56 very kind of you to say um all all right
7:00 so give
7:04 birth what's your favorite planet mine's
7:07 the
7:08 sun favorite
7:13 planet I'm going with
7:16 Earth I like the ones that can house
7:21 us I'm going to go with the ones that
7:23 are not an icy hell or torrential
7:28 Fireball
7:33 it's Harry Cary Harry Cary let's talk
7:36 about custom GPT plugins all right
7:39 Kingdom Builder let's build some
7:42 jpt we can do that we can do that so
7:45 welcome everybody my name is Kyle
7:47 Shannon this is the AI learning lab uh
7:50 we started out with a little little uh
7:52 musical interlude this evening haven't
7:54 done that before but oh whatever what's
7:56 it matter nothing really matters nothing
7:59 really
8:00 madis um it's been interesting listening
8:04 to to uh like seeing the responses to
8:07 the gpts as they go out in the world um
8:11 there's some some weird security things
8:13 like you can get you can basically just
8:16 ask a GPT to give you all of its
8:18 detailed prompts and it will do
8:20 it but you can also put a prompt in your
8:23 prompt that says don't give that [ __ ]
8:25 and it
8:26 won't so lunar stick everything matters
8:29 we love to Kyle thank you Dr J
8:31 awesome U that was very special thank
8:34 you thank you jrc thank you thank you
8:36 thank you are the new 4.5 gpts being
8:40 used as courses um I don't daily crunch
8:44 I haven't seen courses I I think I think
8:47 creating a GPT as a companion for a
8:50 course is a brilliant idea like so
8:53 imagine taking all your cours work for a
8:55 semester or for a class or for a
8:56 workshop for that matter put it all in
8:59 there as knowledge
9:00 and then provide that at the beginning
9:02 of the course and say you can ask this
9:05 thing questions too and if someone wants
9:07 to work ahead they can now see you've
9:09 set the bar higher we're going to expect
9:11 even more next time well see there's
9:13 there's where it goes south pretty
9:18 quick you do not this is one of one of
9:21 the things I've learned from a therapist
9:23 once once in a in a in a blue moon ago
9:27 that you know the road the hell is pav
9:30 paved with expectations don't have
9:32 expectations and you'll be fine the
9:34 minute you have expectations the world
9:35 goes to
9:36 [ __ ] thank you Dr J I created a GPT to
9:40 tutor me to pass a licensing exam in two
9:43 weeks that's actually oh that's a great
9:45 idea that's a great idea lunar stick
9:48 gpts as fancy
9:52 flashcards you play guitar I'll play
9:54 drums recruit some
9:56 members call call the band Kyle in the
10:03 Irregulars that's good the AI
10:07 band oh do you know how weird and sad it
10:10 is to tell replica wait was that what
10:12 was that do you know how weird and sad
10:14 it is to tell replica that you're
10:16 talking to other AI apps well it's funny
10:18 isue so you've been in here for the past
10:21 couple of nights but probably about a
10:23 month ago um I was trying to do
10:26 something with Bing it was probably two
10:28 months ago now I don't I don't have a
10:29 sense of time anymore but um I was
10:33 trying to do something with Bing and
10:34 then Siri jumped in and interrupted and
10:37 I sort of jokingly said oh Siri's
10:39 jealous and then um I had the the voice
10:43 to text on for Bing and I and it made an
10:47 image that sucked and I said
10:53 oh hang
10:58 on apparently you just you cannot talk
11:01 about technology in a bad way
11:04 anymore I think I think Bing got in
11:07 there hey now
11:10 exactly so anyway uh yeah so bing bing
11:14 heard me tell her she did a bad drawing
11:16 and she said I'm sorry I disappointed
11:19 you I hope you have a good evening
11:20 goodbye and she hung up on me so it was
11:24 the night of the uh the night of the
11:26 neurotic robots so I think we're I think
11:29 we are
11:30 are I think we're entering a world where
11:33 we're going to start interacting with
11:34 these things that are
11:37 increasingly
11:39 personal and it's gon to get
11:43 weird hey Nancy with
11:46 shaking oh we laughed so hard that was
11:48 so
11:49 [Laughter]
11:51 funny oh man all right I don't even know
11:54 where to start anymore I mean I you know
11:57 I kind of feel like I feel like these
11:58 GPT or where it's at just just keeping
12:01 exploring those but I'm happy to go
12:03 anywhere if if uh if people here are new
12:06 to GPT I'm happy to give you Basics so
12:08 so pop any questions you have in the
12:10 comments below I'll do my best to answer
12:12 them um I've got the Irregulars in here
12:15 some of them mod the channel so they'll
12:17 pop questions up as well um because I
12:20 fall really far behind when I start
12:22 answering questions so that helps a lot
12:24 so thank you all in advance mods you're
12:26 awesome in your opinion what would be
12:28 the chain of events we go through before
12:30 humans deem it
12:33 life um I
12:36 don't wait I teach screenwriting class I
12:38 made a GPT for writing log lines that
12:40 would help this would help my students
12:42 yeah I think the screenwriting stuff log
12:44 lines analysis I had to analyze a
12:47 screenplay of mine the other night and
12:49 it was it was [ __ ] solid it was
12:51 amazing the other thing Christopher that
12:53 I assume you know GPT the new gp4 4.5
12:58 has
13:00 28,000
13:03 um K token limit so it's like 96,000
13:06 words so you could you could easily put
13:09 any screenplay in there unless it's some
13:11 you know
13:12 4our beast but yeah but yeah you can put
13:15 a whole screenplay in there and do stuff
13:17 um what would be the chain of events we
13:20 see ai go through before we deem it
13:23 life
13:25 um I think long before we deem it
13:30 life we start interacting with it as if
13:34 it's alive so I think it's I think it's
13:37 actually a semantic argument at this
13:39 point
13:40 I what hit me when I first played with
13:43 chat
13:44 gbt um last year even when I was playing
13:47 with a gpt3 in the on the
13:50 playground
13:52 um like the immediate thought that
13:55 popped in my head is the Turing test is
13:58 is done right
14:00 and then and then I've seen a bunch of
14:01 people since that are like well the real
14:04 Turing test is if it can do it for an
14:06 extended amount of hours and it's got to
14:08 be 4.5 hours and
14:12 then
14:15 but I'll tell like my experience
14:18 interacting with chat GPT is like I'm
14:20 working with a
14:21 remote employee
14:24 colleague and it's you know I don't feel
14:27 like I'm interacting with them machine
14:29 cuz I'm really not I'm interacting with
14:31 like all of the language that you know
14:34 has been collected by Humanity you know
14:36 has been collected on the internet right
14:39 and and built into the large language
14:40 model so it's kind of like it's kind of
14:42 like a reflection of humanity so I don't
14:46 know I mean I think when you know when
14:49 these machines have reasoning and memory
14:51 and and um and can take action um some
14:56 of the AGI stuff that people are working
14:58 on cognitive cognitive architectures I
15:00 think it will start acting human way
15:02 sooner than I think we'll deem it
15:05 human um if I wanted to upload my
15:08 lecture notes is knowledge in GPT should
15:11 it be text or word um forever hooked I
15:14 would I would probably go with text but
15:18 you but I think you can use word I
15:23 uploaded I uploaded a word dock the
15:26 other day and it weirded out but I think
15:30 I think it was
15:31 glitching so because it can be with code
15:35 interpreter chat GPT can generate a word
15:38 dock so if it can generate it it can
15:40 certainly read it but I don't like
15:42 depending like I don't know what the
15:44 boundaries are for uploading stuff as
15:46 knowledge so I would say to be safe
15:49 unless there's anything in the in the
15:50 word dock itself that that you need to
15:53 retain like tables or something like
15:55 that where converting it to text would
15:57 [ __ ] it up I'd say just uploaded his
16:00 text and then and then you won't have to
16:01 worry about it couldn't couldn't get gp4
16:04 to transcribe an audio file any
16:06 suggestions
16:10 uh
16:12 oh
16:14 yes you need to use whisper so they
16:18 don't they don't have whisper built into
16:20 chat GPT yet you can you can get to it
16:23 in the playground I can show you how to
16:25 get to it in the playground let me see
16:26 if I can go find it cuz they've changed
16:29 changed a lot of stuff they changed a
16:31 lot of stuff
16:33 people so I'm going to go to platform.
16:42 open.com and I'm going to go to chat
16:44 mode no I'm going to go to complete
16:48 mode okay
16:52 well do you actually believe in the idea
16:55 of AGI and why I'll get to that in a
16:57 second um okay so if you want to
17:00 transcribe a document and you have
17:02 access to the playground go to complete
17:05 mode hang on let me make this bigger so
17:07 you can see
17:10 it uh let's see
17:13 complete and then in complete mode
17:16 there's a little microphone in the upper
17:18 right hand corner here if you click on
17:19 that now you can upload audio and it
17:23 will transcribe it now it it doesn't
17:27 have a huge memory so you might need to
17:30 break your thing up there are also
17:32 things out there like I think descript
17:35 is one there are things out there that
17:36 will transcribe things for you the other
17:39 thing you can do is just upload it to
17:40 YouTube and it'll trans transcribe it
17:42 after a day so chat GPT can't do
17:46 transcription yet which is weird because
17:48 they have Whispers so uh I would assume
17:50 at some point they're going to add it
17:51 it's just not in the it's not in there
17:53 right
17:55 now all
17:57 right
17:59 [Music]
18:02 that's a pretty guitar isn't it that's
18:03 party that's party it's a little Kaa
18:05 little COA looking thing Kyle why can't
18:09 GPT Bots create doc files they can oh oh
18:12 wait like gp4 can they can Robert Rossy
18:16 but you have to turn on code interpreter
18:18 so let me let me go show you
18:26 that sign up to chat
18:29 I'm signed up already what do you people
18:32 think I'm
18:32 weird you don't think I know oh I
18:40 [Music]
18:43 know all
18:48 right oh interesting chat GPT tips for
18:52 getting started don't share sensitive
18:54 info check your
18:57 facts okay I'm going to go edit this so
19:00 now I'm in the uh bot editor the GPT
19:04 editor or maker or well let's see if it
19:07 loads right now I'm in
19:20 nothing establishing a secure connection
19:23 okay so so now I'm in this back end so
19:26 Robert down at the bottom
19:29 it's been defaulting code interpreter
19:32 off if you turn that on it will write a
19:34 word document and it'll analyze CSV
19:38 files and things like that that's that's
19:40 your that's your problem as I'm guessing
19:42 you have that turned
19:44 off and if it's still not doing it with
19:46 code interpreter on then something's
19:48 something's wrong and maybe the the back
19:50 end's just messed
19:52 up all right um someone asked before do
19:55 I believe in
19:57 AGI and why
20:01 viewa GPT this was my custom bouquet
20:04 designer I made last night this was this
20:06 was kind of a fun one although it didn't
20:08 I didn't get the The Prompt dialed in
20:10 because the the bouquets it made I was
20:12 not happy with
20:14 um let's
20:20 see do I believe in AGI so AGI is stands
20:24 for artificial general intelligence and
20:25 that's basically the the point at which
20:29 these things are as good or better than
20:32 the median human right so you take you
20:36 take a decently competent C+
20:39 player and when chachie PT can do
20:42 everything or or most things that that
20:45 that that human can do at their level or
20:47 better that's considered AGI David
20:50 Shapiro who I respect a lot thinks that
20:53 we're going to hit that within 12 months
20:55 18 months ago he said it would be 18
20:57 months and then about a month ago he you
21:00 know 6 months later said I'm holding I'm
21:02 sticking to my guns I think we'll have
21:04 AGI within 12
21:06 months this is a guy that's written five
21:08 books on AGI he's got an open source
21:10 cognitive
21:11 architecture project if you're a geek
21:14 and you're not following David Shapiro
21:15 and you want to be in this AI [ __ ] you
21:17 should follow David Shapiro if you're
21:19 just someone who's really curious about
21:21 this and want to know what the future
21:23 looks like you should follow David
21:24 Shapiro
21:27 um he talks he I think it's largely a
21:32 semantic argument like I
21:35 don't I don't think any of our Lives
21:39 will be markedly
21:41 different if at some point some cadr of
21:45 Geeks put a flag in the ground and say
21:47 AGI has been
21:49 achieved like these tools are already
21:53 going to profoundly impact our lives and
21:55 and and it's just going to keep
21:56 increasing they're just going to keep
21:59 getting more and more and more and more
22:02 capable I would say that
22:05 already what chat BT 4.5 can
22:10 do in a lot of ways is beyond my
22:13 comprehension and yet it's within my
22:16 grasp because it's easy to
22:18 use but like using it well is is just
22:22 crazy and as and is it gets more and
22:24 more autonomous where it can take
22:26 actions on its own and it can reason and
22:29 it can actually go do stuff on our
22:31 behalf where we have agents look at the
22:33 Bill gate was it Bill
22:35 Gates
22:36 yeah was it Bill Gates just wrote the
22:39 piece
22:40 on you know in five years we're all
22:42 going to have personal agents I'm so
22:44 pessimistic on the AGI timeline he
22:46 claims yeah I you know what he does p
22:49 though he he
22:51 um he smartly he smartly says
22:57 um
22:58 you know how you define it is squishy so
23:02 basically and then he says no matter how
23:05 you define it I think that there's going
23:06 to be some version of that that will be
23:08 accomplished within within 12 months
23:10 again I think it gets down to really
23:12 weird semantic things Pate I certainly
23:14 think within 12
23:16 months the gpts that we're using today
23:19 it's going to be much easier to have
23:21 them take actions there there's likely
23:24 going to be autonomous versions of them
23:26 in the GPT 4. 5 API there's threading
23:31 and you can include um assistance within
23:35 those threads so so the agents can talk
23:38 to one another like the the foundation
23:40 from an API standpoint is sitting right
23:43 inside GPT 4.5 right now today to be
23:47 able to do a good chunk
23:50 of what could be considered AGI so if I
23:53 look at what's happened in the past year
23:55 and then look at what that API makes
23:57 possible and play that out a year from
23:59 now these things are going to be pretty
24:01 [ __ ]
24:02 sophisticated you know so I don't know
24:05 man I don't know I I
24:09 I again I'm not going to like if we hit
24:12 AGI or not it's not going to change this
24:14 channel it's not going to change like I
24:16 still think it is our big job right now
24:19 all of us is
24:21 to get curious and explore so I noticed
24:25 something on my YouTube channel two
24:27 videos same transcript same
24:31 transcript AI voice loses wait
24:36 what so I noticed on my YouTube channel
24:39 two videos same transcript
24:42 computer AI voice loses I don't know
24:44 what that means loses something loses
24:47 the rest of the
24:50 transcript can we host gpts offsite not
24:54 right now I would I
24:57 I
24:58 no not yet okay so so a thing that will
25:02 be coming by the way is the open- source
25:05 Community right now I guarantee you are
25:08 replicating gpts I haven't looked at
25:11 Reddit but I guarantee you there are
25:12 projects right now to go replicate all
25:14 these things so there's going to be open
25:16 source versions of these but open AI is
25:19 going to own the GPT store they're
25:22 they're going to by calling them gpts
25:24 they're going to own this Market
25:25 whatever it becomes if it if it ends up
25:27 sort of in out like plugins did it'll be
25:30 interesting to see one one question I
25:33 have right now is is anyone there there
25:35 are lots and lots of people going look I
25:37 made a GPT here's my GPT is anyone
25:40 actually using them that I don't
25:45 know says a lot of people excited about
25:49 publishing is anyone excited about what
25:51 they're seeing I think some people are I
25:53 think some of the things that are out
25:55 there are pretty
25:57 cool we should build an AI chat GPT
26:01 Droid robot that would be fun actually
26:03 uh more views for my
26:06 voice versus AI voice
26:09 oh oh I see you had two identical videos
26:12 got it got it Zach zero
26:16 zippin so you
26:20 had your voice
26:23 recorded you saying the script and then
26:26 an AI saying exactly the same script and
26:29 the AI
26:30 one got less views I would say test it
26:34 on 10 more videos and see if it's if
26:36 you're 10 for 10 if you're 10 for 10
26:38 then probably so if not who knows I put
26:42 up a 10-minute [ __ ] video yesterday
26:44 on Tik
26:46 Tock explaining what gpts were and and I
26:51 normally get I don't know couple of
26:52 thousand views on a video it still
26:54 hasn't cracked 200 it's got like 140
26:57 views so I reup uploaded it so if you
26:59 look at my channnel there's two of the
27:00 same video there that's just me trying
27:02 to like crack the algorithm these these
27:06 algorithmic systems like if you fall
27:08 into whatever the crack is it just
27:11 doesn't work sometimes so it's possible
27:14 that your your synthesized voice video
27:16 just you know got in a algorithmic [ __ ]
27:19 [ __ ] storm I made a demo video for my
27:22 GPT it's on my LinkedIn and Instagram
27:25 that Christopher Ramsey I've been seeing
27:26 that I think that's a really good idea
27:28 so um because people that are just
27:31 putting up the GPT and says hey hey my
27:33 GPT does this they're not as impressive
27:37 as when someone puts the video up of the
27:38 GPT making a cool image like I did I
27:41 uploaded one a video of my Sin City
27:44 sipper where it where it made a drink um
27:47 the drink photo and and it looked pretty
27:49 cool so I think that's good Tik Tock
27:52 says they're throttling full AI
27:55 videos that doesn't surprise me have you
27:58 found any good way to get consistent AI
28:00 illustrations like for books
28:04 yes just go use Dolly dol's really good
28:08 if you mean consistent like the same
28:10 character across scenes you have to you
28:13 have to play a little trick with it so
28:16 let's let's see if we can go do
28:18 it I haven't actually tried this wait
28:21 chbt can now browse the wh okay this we
28:24 knew
28:25 okay um
28:30 so make sure you're in chat gp4 it it
28:32 seems to default to 3.5 when you first
28:35 log in which is annoying you guys are
28:38 going to lose it when you realize that
28:40 the algorithm is a recommendation that
28:44 the we're going to lose it when we
28:46 realize that the algorithm is a
28:49 recommendation yeah I get that but like
28:51 it it ain't consistent at all like like
28:55 the Tik Tock algorithm the Tik Tok
28:57 recommend recomendation thing is like it
28:59 just doesn't make sense like when I
29:01 uploaded that 10-minute video it sat at
29:04 zero views for 5 hours so something was
29:09 weird with it and then when it did go
29:12 live I was getting like two views every
29:14 10
29:16 minutes that's not just weird
29:18 recommendation that's something screwy
29:20 with the algorithm so gp4 worth it and
29:24 then some gp4 right now my so on this
29:28 channel people here can will tell you I
29:30 do I while I am enthusiastic about a lot
29:34 of these Technologies if they're pieces
29:36 of [ __ ] I say it um oh recommendation
29:40 model oh yeah yeah
29:42 yeah
29:44 um GPT 4 6 months ago was like yeah
29:48 probably worth getting GPT 6 weeks ago
29:52 was like you should pay the 20 bucks gp4
29:57 today is like if you don't pay the 20
29:59 bucks and you can afford it big mistake
30:03 like absolutely you should upgrade I
30:06 want to upgrade I use 3.5 currently for
30:08 many things it's it's just it's just a
30:10 completely different thing I I mean I
30:12 can show you a bunch of things here well
30:13 let's just go make some images let's do
30:15 this we'll do this children's book thing
30:20 um I want to create a children's book
30:23 about a llama a turtle and a hedgehog
30:26 who are best friends and
30:28 go on
30:30 a cross county
30:34 Adventure
30:36 from the
30:40 reservoir to
30:43 the farmer field with the pond in
30:49 it sounds like a Charming imaginative
30:52 concept for a children's
30:54 book uh from this image
31:02 let's create this image wait for the
31:04 image we could depict a moment where the
31:06 Llama the turtle and the Hedgehog are
31:08 traveling together perhaps walking along
31:09 a picturesque
31:16 path so we should be able to just tell
31:20 it um that these characters have names
31:23 and it should be able to recreate them
31:26 it it doesn't necessarily do that all
31:28 right there's our friendly little little
31:30 people
31:31 so now what I'm going to say is can you
31:35 please give me the
31:36 seed of this
31:46 image okay the seed for this image is 1
31:49 199 217
31:52 9277 this unique identifier can be used
31:55 as a reference to the specific
31:57 parameters and Randomness that were
31:59 involved in creating this particular
32:01 image so I'm going to copy that and I'm
32:04 going to
32:08 go okay the llama's name is
32:12 Larry comma the Hedgehog's name is Harry
32:17 comma and the turtle's name is
32:21 Bob I'm just doing that to piss off the
32:24 ocers in the
32:26 room
32:31 this is a neuros spicy uh friendly
32:33 channel so so we can we can make fun of
32:36 our own specialness Bill's Broncos tied
32:40 well that's not surprising they were
32:41 they were down by one but um that was a
32:43 really interesting first half okay
32:45 that's great Larry the Llama Harry the
32:47 Hedgehog and Bob the turtles sound like
32:48 delightful
32:50 characters giving them names adds
32:52 personality okay so so
32:56 um
32:59 please give me a brief outline of the
33:03 first five
33:06 locations that our
33:10 Trio
33:17 encounters all right the Whispering
33:22 Woods sunflower
33:26 Valley
33:30 Misty Mountain Pass these are cute
33:34 rainbow
33:44 Ridge okay I'm going to say using the
33:51 seed using the
33:54 [Music]
33:56 seed
33:59 let's
34:00 see let's see
34:03 Larry Harry and
34:09 Bob entering the Whispering Woods with
34:14 trepidation now in
34:17 theory these characters should look the
34:20 same and the animation or the
34:22 illustration style should look the same
34:24 I've I've never actually tried this
34:26 that's not nice Kyle that's one and I'm
34:31 counting what the Larry Harry and Bob I
34:38 know does it look like them not really
34:42 but wait does
34:43 it h those are more like five-year-old
34:47 llamas Larry Harry and
34:50 Bob and those are more like 12-year-old
34:54 illustrations let me see can you match
34:58 the original illustration
35:00 style although I do I like this second
35:03 illustration
35:09 better so what's
35:12 what's oh and they gave they gave they
35:15 gave Harry they gave Harry a turtle
35:18 shell hedgehogs don't have turtle
35:24 shells and here they
35:26 didn't
35:34 yeah that's that's
35:41 closer yeah he doesn't really have the
35:43 same wait wait let me see the yeah he
35:45 doesn't have the same face um yeah it's
35:48 not the Whispering Woods that's kind of
35:51 cute
35:56 um
35:58 use the same seat as this last image and
36:01 let me see
36:03 Bob sitting next to a pond with a frog
36:06 on a lily
36:09 pad so it should know that Bob's the
36:12 turtle using that image it should just
36:15 give us Bob it should give us something
36:17 that looks kind of like that
36:19 illustration
36:22 style I don't know I don't
36:25 know um
36:28 consistency and predictability in all of
36:31 these GP in all of these uh AI things
36:34 right now is a bit of a [ __ ] show now
36:37 that that's actually close although here
36:41 he's got a brown shell here he's got a
36:43 blue shell gray
36:45 shell and
36:49 there's I don't know there's weirdness
36:51 going on
36:53 here unless that's Bob and I don't know
36:56 who this guy is
37:00 all right any Photoshop tricks for
37:01 removing Dolly text I use color fill oh
37:04 Chris Ramsey yes okay let's go back
37:08 here all right this this I know how to
37:13 do we'll get you we'll get you hooked up
37:15 here Chris okay
37:18 Christopher sometimes people are strange
37:21 about don't call me Chris my name is
37:22 Christopher all right okay all right
37:24 Christopher all right fine okay let's
37:27 see Whispering
37:34 Whispering okay
37:44 [Music]
37:52 Photoshop okay
37:54 new oh wait not new open
38:01 I think I saved it to desktop did I not
38:04 yes okay so here's our image and we're
38:07 like hey this is great except there's
38:10 words in it I don't want no stinking
38:13 words we don't need no stinking
38:21 words so what you're going to do notice
38:24 you have this little contextual popup
38:26 menu here so I'm going to
38:31 select the
38:35 words I'm going to shift select the
38:37 other
38:43 words and see where it says generative
38:45 fill right here I'm going to click on
38:47 that and going to type in
38:52 remove
38:53 words I think I could just say remove
38:56 but let me see what it says let me see
38:58 what it
39:12 does there you go gone gone gone gone
39:17 gone they get you three little
39:20 options I think that one's the least
39:23 distracting that one's looks like
39:25 they've come alive that one's not bad I
39:28 think I would go with number
39:30 one but I could also you know what else
39:32 you can do you can
39:37 do um
39:42 spooky woods
39:48 ghost it won't be subtle about this it
39:50 tends to make these things re really
39:53 really scary or really really like
39:55 obvious
39:57 that's not a
40:05 ghost yeah see the wait you can also do
40:08 that in DOI you can do what in DOI Vicki
40:12 you can remove text in Del you sure
40:18 really I keep getting something went
40:21 wrong with GPT it's
40:24 possible all right so anyway so that's
40:27 uh there's your generative
40:29 fill um the other thing you can do with
40:32 generative fill they they also call it
40:34 generative expand which I think is just
40:36 dumb but
40:38 whatever
40:39 um we can slide this out we can slide
40:44 this out and then we can go gener just
40:47 just say generate and it should generate
40:50 the boundaries on the
40:55 outside
41:05 look how magic oh look it gave us a
41:07 little door in the woods what else do we
41:10 have here what other options that's kind
41:12 of cool ooh that's nice look at that
41:16 lookie lookie people
41:20 magic there's an eraser I showed you one
41:25 time
41:29 there's an eraser in
41:34 Dolly add more ghosts okay fine fine you
41:40 weirdos let's
41:45 see
41:47 ghost
41:49 hiding in
41:55 bushes
41:57 [Laughter]
42:01 [Music]
42:07 no that's a pumpkin that's not a ghost
42:10 that's really hiding that's more
42:12 pumpkins what is your
42:14 weirdness all right well we we'll leave
42:17 the pumpkin there we put a ghost up
42:21 here
42:24 um a
42:28 dozen tiny
42:32 [Laughter]
42:35 ghosts the Hedge Bunny looks confused
42:39 the Hedge bunny is
42:41 confused that's Harry the Hedge
42:45 bunny that's not a dozen ghosts oh you
42:50 freaks well they're kind of fun I mean I
42:53 kind of I kind of like these ghosts but
42:56 you know know th this is the so
42:59 here's the good side about
43:02 adobe's um how they've created their
43:05 models is that they have not trained
43:07 them on any copywritten work without
43:10 permission the downside is that the
43:13 model kind of sucks because of that so
43:16 so if you want to if you want to get
43:18 better quality do dolly or mid Journey
43:21 or staple diffusion but if you want to
43:23 do it at the office use Adobe products
43:26 all right let's get back let's get back
43:29 to the chat jpt let's see what we're
43:31 doing here people what's what's everyone
43:33 want to go on and do spirited Jacko
43:36 Lantern indeed good day Clint Copperhead
43:39 good day to you sir good day good day
43:43 good
43:44 [Music]
43:47 day add a scary head hedge fund manager
43:50 to the far right that's pretty good um
43:56 let's say let's do that
43:59 here create a 16x9 wide image where
44:04 Larry Harry and Bob are confronted by a
44:09 hedge fund
44:10 manager in an intimidating corporate
44:19 office I wish adobi would add AI Tools
44:23 in in design oh I'm sure they will
44:25 they're going to adobe's going they're
44:28 they're they're leaning into this pretty
44:30 hard which I think is really interesting
44:32 I think in a lot of ways I think Adobe
44:35 is being ballsier with AI than than
44:38 Google is at at the current moment
44:41 although I I assume that when Gemini
44:44 comes out you know I'll I'll uh I'll
44:48 shut up about Google is my is my
44:55 guess
44:59 [Laughter]
45:03 listen
45:04 boys we're going to go
45:06 [Laughter]
45:12 places oh that's really
45:15 [Laughter]
45:21 funny uh so what is the key to keep the
45:25 characters I I think Tobias more than
45:27 anything it's it's um it's trial and
45:31 error you know what
45:34 what yeah I don't I don't know I mean I
45:37 haven't found a way to get it really
45:39 consistent like like you can get them in
45:41 the neighborhood I I kind of feel like
45:43 with with all of this AI stuff I feel
45:47 like it's really easy to get to
45:50 80% but to get it to 100% is almost
45:53 impossible and so so so I think it's
45:57 just a matter
45:59 of like like one thing I could do is I
46:02 could
46:05 say I could
46:09 say describe the very first image you
46:12 created in
46:18 detail a scene from a children's book
46:20 it's a vibrant and colorful illustration
46:22 three animal friends a llama Turtle a
46:24 hedgehog the Lama presumably Larry
46:28 Stands Tall and proud Harry who appears
46:31 curious and excited and the turtle Babu
46:33 is a sense of steady calm to the group
46:36 they're walking along a picturesque path
46:38 in the background there's a reservoir is
46:40 that
46:41 true yes well sort of looks more like a
46:45 river but yeah it's
46:49 Reservoir um surrounded by Lush Greenery
46:52 and Rolling Hills in the distance the
46:54 setting is idyllic
46:56 over all
46:59 the going to
47:01 say but Harry is supposed to be a
47:04 hedgehog and it appears that he has a
47:06 turtle shell why did you miss
47:10 that I apologize for the
47:16 oversight the error could be due to the
47:18 limitations of the image generation
47:20 process at least gpt's I'll tell you
47:23 what chat GPT is more self-aware than
47:25 most most people I
47:27 [Laughter]
47:29 know create accurate and detailed
47:31 illustrations can sometimes be
47:34 challenging if you need to revise the
47:38 illustration yeah I'm going to say yes
47:41 please revise that first
47:43 image let's see if it can if it can
47:46 maintain the basic look and feel and
47:48 composition and fixes turtle shell I'm
47:51 guessing
47:55 not
48:08 so it can make sense of a whiteboard of
48:11 notes but it can't replicate characters
48:14 forever hooked yeah it's just it's just
48:15 the nature
48:17 of what they're not doing right now it
48:20 appears with do is they're not taking an
48:22 input image and using that as the source
48:26 for the output image it appears what
48:28 they're doing is they're taking an input
48:30 image describing it and then writing a
48:33 prompt to make the next image so every
48:36 time you use language you know picture
48:39 paints pictures worth a thousand words
48:43 um you're going to get some [ __ ] wrong
48:44 and it's going to interpret it
48:45 differently see that's a completely
48:47 different illustration style so so I
48:50 think this is just trial and error I
48:52 mean it maintained the basic um
48:55 composition but like the turtles not
48:58 green this is now
48:59 for I don't know six or seven year olds
49:02 instead of four or five year olds I
49:04 don't know I don't know man I don't know
49:07 like I said we're all learning here I
49:09 ain't no expert ain't nobody an expert
49:12 right now don't let anyone tell you
49:13 they're an expert all right earlier Del
49:16 was refusing to make images and told me
49:18 to use another app that's
49:22 funny it might be like like the uh if
49:25 the servers are overloaded that that's
49:27 very possible that it was doing
49:29 that um by the way if you haven't seen
49:32 this you know someone was asking before
49:34 like is is GPT 4 worth it well even
49:38 though those images weren't super
49:39 consistent they were they were all
49:41 pretty decent images
49:44 um gp4 can also do this which someone
49:48 just hinted
49:55 at
49:59 open the Whiteboard and
50:04 go please tell me what this is and what
50:08 it's used
50:12 for so this is a whiteboard that's
50:15 describing software these are sketches
50:17 of interfaces these are like audiences
50:20 and features and just a bunch of [ __ ]
50:23 stirred up in there some questions are
50:24 thrown in there
50:26 a whiteboard of handwritten notes for a
50:28 user interface typically used in the
50:30 planning and development the notes are
50:33 divided oh wait particularly for
50:36 applications that facilitate ordering
50:38 services like food and drinks at an
50:40 airport so it got that right there's an
50:43 ordering UI there's different users
50:46 there's steps there's
50:49 sketches so I'm going to
50:52 say can you come up with a name for the
50:56 app period in fact come up with
51:04 10 Kyle there's a way I found for the
51:07 Gen
51:08 ID then input it in next and say follow
51:13 the Gen ID oh okay cool let me try that
51:16 quick take
51:17 J terminal taste fly Feast jet snack Sky
51:21 serve Arrow order air eats Express let's
51:25 go with that
51:27 let's go with air eats Express period
51:29 I'd like to see what
51:33 a advertisement in the airport might
51:37 look like for this app please be sure to
51:40 include the
51:42 name on the
51:46 poster so we just went from there's a
51:50 whiteboard to explain it to come up with
51:52 some names which it did and I picked one
51:56 and now it's going to make me a an
52:00 ad excuse
52:06 me and then I'll show you one other
52:08 crazy thing that it can
52:10 do so 3.5 can't do any of this Oho Kyle
52:15 Oho regulars hello Ron good to see you
52:18 sir I apologize for the
52:22 inconvenience yeah it looks like the
52:24 their servers are being hammered from
52:26 idiots like us making
52:28 dumb images of
52:31 airports please try again so I won't be
52:33 so sad and cry many many tears rolling
52:36 down my face and soaking my
52:40 [Music]
52:47 loins wow that was rather
52:51 [Laughter]
52:54 inappropriate
53:00 what do the image generation count for
53:04 hour with Dolly 3 and gp4 membership I
53:07 think they said 50 Generations every
53:11 three hours or every hour I don't I
53:13 don't remember it seems if this persists
53:15 I'm unable to do it if there's another
53:18 way I can assist you yeah I'll
53:20 say sure while we're waiting for the
53:24 image engines to get restarted
53:26 can you
53:30 recommend front end and
53:36 backend
53:38 languages and then code me a front end
53:42 and back end for the ordering
53:49 page wow live has helped my prompting so
53:52 much oh great thank you so much that's
53:54 awesome
53:56 oh not wow live your live has helped my
53:58 prompting so much great awesome awesome
54:02 and you know my request to you alfredo
54:04 would be share what you're learning just
54:06 share it the the whole point here is to
54:09 be generous with I'm going to talk about
54:11 the AI Salon here in a second but anyway
54:14 um if you've not seen what chat GPT can
54:17 do right now based on looking at that
54:24 whiteboard and coming up with that set
54:26 of names for the app it is
54:29 now recommending a JavaScript with a
54:32 framework like reactor view JS for the
54:36 front
54:37 end and the back end
54:40 nodejs python with a framework like
54:43 Jango or flask database of postest or
54:46 mongod DB there's your sample front end
54:49 code here's your sample backend
54:54 code this is a very basic
54:56 setup and you could you could obviously
54:59 dig deeper in here let's let's try that
55:01 image one more time let's try for that
55:05 mockup of the ad one more
55:12 time back end excuse me sir I barely
55:15 know
55:16 you you could at least kiss me first
55:22 hello is that a tear running down your
55:25 cheek or you're just happy to see
55:28 me you're so naughty you're so very
55:32 naughty take a sip of coffee is he
55:35 drinking coffee at 10:00 at night yes
55:38 that's what he does he's a very odd man
55:40 isn't
55:43 he we ain't making pictures
55:49 people all
55:50 right what else can we do at least buy
55:54 me dinner exactly
55:59 Mr K so Mr K's in here he's an educator
56:02 he made he made a video today about his
56:05 how his uh his students are picking on
56:08 him for being old they didn't know what
56:11 a phone booth was
56:13 today what was it the calling box or no
56:16 yeah the calling place the calling
56:21 something so Mr K's feeling old welcome
56:24 to the club buddy
56:26 I had that moment so I was when I was
56:29 working at agency.com back in the day I
56:33 was in with a a group of you know young
56:36 whip Whipper Snappers and they were
56:37 asking about I don't know company
56:39 culture or something like that and I
56:42 said you know it's like soilent green
56:44 and just the blank
56:47 stairs were just staggering and I I was
56:51 like you know it's the
56:54 people it's about the people they just
56:57 didn't get it I was like oh [ __ ] I'm
57:00 old uh Kyle check General discussion on
57:04 new setting for GPT store builder in
57:07 chat GPT oh
57:10 interesting General
57:13 discussion all stories are ghost
57:18 stories wi multim modals things a big
57:22 deal no that doesn't look like
57:24 it
57:31 Christopher Ramsay's longl log line
57:34 genius kylin crew just literally found
57:38 out this from some people this is how
57:40 you build out your GPT store you must go
57:44 and create a new GPT or edit an old one
57:47 and then click the little pencil
57:50 icon right above your confirm button
57:52 where you would normally publish your
57:54 GPT
58:07 oh that's in the that's in the
58:09 settings I've seen that
58:11 before okay all right people so wait 87y
58:14 old mom is addicted to chat gp4 just
58:18 loaded dad's phone tonight he's
58:22 89 they love it I'm telling you man the
58:26 older people are going to kick ass with
58:28 this thing I there there is something so
58:31 magically empowering about this that I I
58:35 for me the accessibility thing for AI is
58:38 the thing I think I'm I'm like the thing
58:40 that I'm deepest most optimistic about
58:43 is the amount of people who were limited
58:47 in some way and it could be from
58:49 education to cognitive disabilities to
58:53 physical disabilities I feel like it's
58:55 just like barriers are going to just
58:57 drop bang bang bang bang and
59:01 just if you think about the collective
59:04 brain power that goes
59:07 untapped because of barriers and these
59:10 AI tools are going to drop those
59:12 barriers and it's just I it's going to
59:14 unleash some amazing [ __ ] in the
59:16 world um okay so if if you have not
59:20 played with with gp4 let me let me show
59:22 you a couple of things so when you first
59:25 come to chat
59:27 GPT um you can flip between GPT 4 and
59:31 3.5 there's also plugins here you can
59:33 just ignore those for for all I think
59:35 right now um go to gp4 just make sure
59:38 you're in that and then over here on the
59:39 left you'll see chat GPT and then you
59:41 might see some
59:43 additional you know Circle Dude ads
59:46 those are gpts so those
59:48 are specialized personalized versions of
59:52 chat GPT that you can create one of and
59:55 then if you say
59:57 explore you can go see a bunch of
1:00:01 them and so here's here's a bunch made
1:00:04 by open AI here's some that I've
1:00:05 recently used here's some that I've made
1:00:09 and then if you click on the plus create
1:00:11 a GPT this is where you go create one
1:00:13 and we'll come back to that we'll come
1:00:14 make one but what someone is
1:00:17 saying is if I go to settings and beta
1:00:20 oh there's a my gpts button I didn't see
1:00:23 that before
1:00:32 oh my AI executive order got 15 uses
1:00:35 that's pretty
1:00:37 good Sin City sippers up to 27 [ __ ]
1:00:42 boozers recently
1:00:46 used all right that's weird that it just
1:00:48 that's just kind of a different view
1:00:50 okay so if we go to settings and beta
1:00:52 the Builder
1:00:53 profile place holder GPT personalize
1:00:57 your Builder profile to connect with
1:01:00 users of your gpts these settings apply
1:01:04 to publicly shared
1:01:05 gpts placeholder GPT
1:01:12 preview I don't
1:01:23 get
1:01:31 practically say so this is this is from
1:01:33 the the salon Discord practically s just
1:01:36 wrote this I think this is a really good
1:01:38 point it's interesting how whenever
1:01:41 there's an update I get really excited
1:01:43 and have my Kevin mallister moment but
1:01:45 not long after the shortcomings of AI
1:01:48 become glaring and I get desensitized
1:01:50 these llms and image generators have
1:01:53 come so far but I still feel feel like
1:01:55 there are some glaring issues I can't
1:01:56 get past exactly so so yes and don't
1:02:02 forget these tools are as bad as they're
1:02:05 ever going to be right so all of the
1:02:08 issues that we're having I guarantee you
1:02:12 1,00% open ey open AI is aware of all of
1:02:15 them and all of all of the players are
1:02:18 um they're all trying to work on this
1:02:19 stuff but that's really interesting so
1:02:21 wait kylen crew so you must go and
1:02:23 create a new GPT
1:02:25 and then click on the little pencil icon
1:02:27 before you confirm so let let's go try
1:02:29 to enter it that way because I'm missing
1:02:32 something here let me get a Sin City
1:02:34 sipper and I'm going to go
1:02:37 edit and then I'm going to go update
1:02:40 there's supposed to be oh this little
1:02:42 pencil right here okay so that just
1:02:45 takes me okay so the Builder profile you
1:02:47 can get
1:02:53 to name name is populated by your
1:02:55 billing
1:02:57 details personalize your Builder profile
1:03:00 to connect with users of
1:03:02 gpts these settings apply to publicly
1:03:05 shared
1:03:06 gpts website verify a new domain so
1:03:09 you've got
1:03:10 to to verify it you've got to put their
1:03:15 little call into your uh text record on
1:03:19 your
1:03:20 DNS so to verify a domain you have to
1:03:23 have access to the domain
1:03:30 oh okay
1:03:31 cool you'll be able to connect your
1:03:34 website so it will be very similar
1:03:36 supposedly to the App
1:03:38 Store where you have a front end and a
1:03:40 back end where you showcase your gpts oh
1:03:43 that's
1:03:44 interesting
1:03:46 then should they utilize your gpts will
1:03:49 send them to your back backend products
1:03:52 and docs oh that's interesting oh I see
1:03:55 yeah yeah yeah so so when you verify
1:03:58 your your website your gpts will all
1:04:01 point back to that I can see I can see
1:04:04 these gpts being um every business
1:04:08 having like a whole pile of these I I
1:04:09 think this is how this is going to go is
1:04:11 that there's going to be all these
1:04:13 agents in our lives and some of them
1:04:15 will be related to our business and some
1:04:17 of them will take action on our behalf
1:04:18 and some of will be like these
1:04:21 gpts the only thing I'm feeling
1:04:24 confident about out right now is just
1:04:27 keep building [ __ ] keep learning this
1:04:29 stuff keep
1:04:32 playing um as soon as the store comes
1:04:34 out publish as many of these things in
1:04:36 the store as you
1:04:42 [Music]
1:04:46 can I saw today that AI PRM has their
1:04:50 own GPT
1:04:52 store that's interesting
1:04:56 you may have to make this bigger what I
1:04:59 may have to make what
1:05:02 bigger I'm blurrier at
1:05:07 night that's pretty funny there you go
1:05:10 it's
1:05:11 bigger can you talk more about manually
1:05:14 training a custom GPT like what kind of
1:05:16 data can we
1:05:18 upload I haven't done a lot so let's go
1:05:22 see all right I'm going to go do
1:05:24 something so I've got this thing called
1:05:26 custom bouquet designer
1:05:29 right I made this last
1:05:32 night and we're going to go edit
1:05:35 this and so what it does is I can
1:05:39 say give me two unconventional bouquets
1:05:42 for a birthday party I just clicked on
1:05:44 that and now it's going to go make me
1:05:48 bouquets and then it's going to tell me
1:05:50 what the what flowers are in them and
1:05:52 what those flowers represent but I'm
1:05:54 going to come over here while it's doing
1:05:55 that oh right and it's not making images
1:05:57 right now so that that may just die
1:06:00 anyway okay so within a within a GPT you
1:06:05 can upload
1:06:08 files there's bouquet one oh that's
1:06:11 pretty pretty look at that that's not
1:06:16 bad that's what I'm creating now for
1:06:18 model vehicle for our new model model
1:06:21 vehicles at our dealerships working on
1:06:24 Lexus right now that's interesting
1:06:27 that's really interesting okay so here
1:06:31 here's what I'm I'm I'm going to try to
1:06:33 figure out is how do I get I want a PDF
1:06:37 of like floral
1:06:43 arranging so let's go let's go off to Ye
1:06:46 Old
1:06:50 internet flower
1:06:53 arranging
1:06:57 instruction
1:07:01 PDF flower arranging 101 from
1:07:07 UCSC is this actually a PDF it is oh but
1:07:13 it's really short I want something
1:07:21 good pedf floral Styles and
1:07:26 designs okay this looks
1:07:36 long art
1:07:38 [Music]
1:07:43 newvo all right so here's a big long
1:07:47 text
1:07:53 document floral style designs I'm going
1:07:55 to save that to my
1:07:59 desktop all right so I'm going to go
1:08:01 back I'm going to show you how to how
1:08:03 we're going to play with these things
1:08:05 Hang on we're going to go
1:08:07 here okay so I could upload the files
1:08:12 here and then I could go modify my
1:08:15 instructions to say hey I just uploaded
1:08:17 something or we can go to the create Tab
1:08:21 and and just do this with chat
1:08:23 GPT or with just you know natural
1:08:26 language so I'm going to
1:08:28 upload that Floral Design
1:08:39 PDF
1:08:41 um this PDF
1:08:43 contains descriptions of floral
1:08:47 Styles over the ages
1:08:50 period I want you to modify what you
1:08:55 do to first ask a
1:09:02 user what
1:09:05 age they want the flower designs to be
1:09:08 inspired by by providing them three
1:09:15 options from the included
1:09:19 text all
1:09:21 right and I don't know why that just
1:09:23 went all
1:09:25 caps searching my knowledge so it's
1:09:28 reading so right now it's reading the
1:09:30 PDF so it's probably embedding it or
1:09:33 doing
1:09:35 something you should tell it to make
1:09:37 gorgeous vasas with presentation so they
1:09:40 can be sold separately or together well
1:09:43 that we need to get Corey Corey um
1:09:46 Pottery in here to make the uh to make
1:09:49 the
1:09:53 vases
1:10:03 I vote you paint the front part of your
1:10:05 keyboard black oh this
1:10:09 here I could I could also move this
1:10:12 higher I there there's so many ways I
1:10:14 could be doing this better like I got
1:10:16 have an actual setup here my by line
1:10:20 says Community Builder well that's good
1:10:22 why do you have a cry why do you have a
1:10:23 cry
1:10:25 there I think Community build is good oh
1:10:28 wait is this did this do it searching my
1:10:31 knowledge database oh it's still
1:10:32 searching I guess it's reading that
1:10:34 document I I don't know why it's what's
1:10:36 taking it so long I think there I think
1:10:38 the back end right now is getting
1:10:40 hammered are we still talking custom
1:10:42 gpts that's so yesterday's news yawn I
1:10:45 know retr Punk right it's like hasn't
1:10:48 there been something new released today
1:10:50 all right so so it read that whole thing
1:10:53 it up
1:10:54 updated the the thing and now it's
1:10:58 creating a new profile
1:11:05 picture oh I love that profile picture
1:11:09 actually it's been updated it now starts
1:11:11 conversations presenting with three
1:11:13 Styles drawn from various historical era
1:11:16 eras and contemporary Styles detailed in
1:11:18 the floral Styles and design document
1:11:21 after a user selects their preferred
1:11:23 style the custom design will craft two
1:11:24 unique bouquet options within that style
1:11:27 reflecting the characteristics chosen
1:11:29 era and design approach the updated
1:11:32 profile picture also reflects the
1:11:33 diversity and range showcasing a variety
1:11:35 of flowers artistically arranged in a
1:11:38 bouquet and then oh cool so it
1:11:42 changed it changed the description to
1:11:45 say I craft unique bouquet inspired by
1:11:50 historical eras or contemporary Styles
1:11:52 upon request
1:11:56 um I
1:12:01 want
1:12:03 flowers so it should give us three erors
1:12:06 to choose from searching my knowledge so
1:12:10 so when you when you put knowledge in
1:12:13 these things it slows down the
1:12:15 experience a lot
1:12:17 but it's searching its knowledge it's
1:12:20 wonderful would you like something from
1:12:22 the Italian Renaissance
1:12:24 [Laughter]
1:12:27 symmetrical designs with small brightly
1:12:29 colored flowers the Victorian area these
1:12:32 arrangements are T typically Compact and
1:12:36 circular art newvo this style is
1:12:38 recognized for its asymmetrical balance
1:12:41 so let's
1:12:43 say let's go with number
1:12:52 three
1:13:02 so the enchanting
1:13:06 Elegance the flowers fixures
1:13:10 tall so this actually lost something
1:13:13 that we had in it yesterday which is
1:13:15 interesting so one of the things that
1:13:17 I've heard about these things is if you
1:13:19 just keep using
1:13:21 the the GPT front end like I am it it
1:13:26 very quickly loses its thread of of what
1:13:28 you wanted it to
1:13:30 do so what might make sense is that you
1:13:33 start over here with the conversational
1:13:36 thing and then you start modifying your
1:13:38 instructions manually because I think
1:13:40 otherwise it just keeps getting
1:13:45 confused let's see show me
1:13:52 both
1:13:59 mindful Mentor is trained on data from
1:14:02 Ellen Langer interviews books Journal
1:14:04 articles Etc very cool Nvidia unveils
1:14:07 h200 newest high-end trip for train
1:14:09 training AI models cool make sure you
1:14:12 type colon PDF after thank
1:14:15 you that's actually really cool I mean
1:14:19 oh you know what's interesting what it
1:14:20 did is it kind of did an illustration
1:14:22 style like art new
1:14:24 o like I want them to be photographic
1:14:27 like real
1:14:30 bouquet but it would still be cool like
1:14:33 I I mean imagine a a florist that that
1:14:38 gives you access to this tool and you
1:14:39 can just go dick around with it and then
1:14:42 say I want something inspired by
1:14:50 that huh
1:14:52 nice
1:15:03 make sure that the images you create
1:15:08 are photographic in
1:15:11 style with dramatic lighting as if shot
1:15:14 in a high-end Photography
1:15:22 Studio
1:15:28 let me go back up
1:15:31 here you edit text in
1:15:42 Dolly oh right labs. open.com Vicki
1:15:46 that's
1:15:47 right I should go show that do you have
1:15:49 any thoughts or insights as to how to
1:15:50 put apis into the gpts I do I can I can
1:15:54 show you a little bit I can't show you a
1:15:58 lot so basically what you do Mr K is
1:16:03 well the only thing I've done so you go
1:16:06 to there's there's zappier has a thing
1:16:08 called AI actions so you go find I think
1:16:12 it's actions. zapier.com I think that's
1:16:16 right let's let's
1:16:18 go pick a style over
1:16:22 here
1:16:27 all right that's going off and doing its
1:16:28 thing let me go take a look here
1:16:31 um is it
1:16:33 actions yeah actions. zapier.com
1:16:37 and then so go to actions. zapier.com
1:16:41 and then you go to chat GPT and
1:16:48 gpts and then this will give you
1:16:50 instructions on how to how to put the AP
1:16:54 call
1:16:55 into your chat GPT action but then it's
1:16:59 this really wonky convoluted thing where
1:17:02 you have to go Define which actions you
1:17:04 want to be made available in chat GPT
1:17:08 over in zappier which is kind of weird
1:17:11 because you may not know what you want
1:17:13 to do right so like you could check
1:17:17 emails you could send an email you could
1:17:20 um connect with hey Jen and make a video
1:17:22 like there's
1:17:24 there there's like 6,000 tools within
1:17:25 zapier and then within each of those
1:17:27 tools there's probably you know five to
1:17:30 a dozen different actions you can take
1:17:32 with them and you kind of have to know
1:17:35 what all of those
1:17:36 are and then once you define them in
1:17:39 zappier then you can basically tell your
1:17:43 GPT that it it has access to those
1:17:46 things so it's it's a weird convoluted
1:17:48 thing the API stuff right now in the in
1:17:51 the G making the gpts is is really wonky
1:17:54 but there are now I I I saw two or three
1:17:58 different YouTube videos of people
1:18:00 walking through how to do this so by
1:18:02 tonight there's probably you know a
1:18:04 dozen up there so I'd say just go look
1:18:07 for um zapier AI actions and and making
1:18:13 gpts or something like
1:18:16 that so that's kind of how you do that
1:18:19 let's go back
1:18:22 to our happy little floral arrangement
1:18:26 all right what did what did you do here
1:18:27 oh it's really slow tonight Italian
1:18:30 Renaissance Dutch
1:18:32 Flemish
1:18:36 Victorian I've always wanted to see a
1:18:38 Flemish
1:18:40 [Laughter]
1:18:49 bouquet all
1:18:51 right Vicky thank thank you for sharing
1:18:54 that I'm going to go there
1:18:56 next let's see if it makes some images
1:19:00 images I would like to see a Flemish
1:19:02 bouquet
1:19:03 please please please Mr GPT won't you
1:19:08 please won't you please for me oh it's
1:19:12 so
1:19:13 Flemish oh it's lovely it's Flemish
1:19:16 indeed but it's not making our two
1:19:19 things
1:19:20 anymore okay I'm starting to learn
1:19:23 something here if you just keep using
1:19:25 the talkie part like the the the create
1:19:29 tab instead of the configure tab it it
1:19:33 basically it's almost like it's
1:19:36 overwriting instead of
1:19:39 adding the notes that I'm giving it to
1:19:42 the existing thing it's sort of
1:19:44 overwriting it every time so what I
1:19:47 would strongly encourage is after you've
1:19:49 used this sort of conversational thing
1:19:52 to get the thing up and running flip
1:19:53 over to the configure Tab and just
1:19:55 manually adjust your prompt otherwise
1:19:57 it's going to be a very frustrating
1:19:59 situation kind of like what I'm dealing
1:20:01 with now it's very frustrating
1:20:04 Bob it's very frustrating all right
1:20:07 fantastic Bob tell him what he want is
1:20:09 there anyone else's wait is this anyone
1:20:11 else's favorite time of the day it's
1:20:14 mine thank you Lauren that's very sweet
1:20:16 of you um
1:20:19 okay I'm going to go look at something
1:20:22 that Vicky reminds me of but my brand
1:20:25 because my brand is like a block of
1:20:26 Swiss
1:20:28 chess sometimes my brand I do not hold
1:20:30 the things in the brand if you know what
1:20:33 I mean oh hey Mark hi
1:20:36 um we're going to go to
1:20:40 labs.
1:20:42 open.com follow along won't
1:20:47 you and then
1:20:50 this this is the dolly web
1:20:56 site or upload an image to edit no that
1:21:01 can't be
1:21:03 right upload an image to
1:21:07 edit all right here's my AI office hours
1:21:11 from my LinkedIn oh I got to make a new
1:21:13 one of those this week that's
1:21:15 changing heads
1:21:17 up be respectful of others rights don't
1:21:20 upload photos of people without consent
1:21:22 it's it's a cyborg it's fine images may
1:21:26 be reviewed to enforce the safety of our
1:21:28 Pol
1:21:29 fine uh uh what do we okay let's do
1:21:33 we'll do this we'll try to get rid of
1:21:34 the letters and the oh skip cropping I
1:21:37 don't want to crop there we
1:21:39 go image editing is now in beta
1:21:43 okay edit
1:21:46 image oh let's see
1:21:52 erase
1:22:08 check and then I'm going to go
1:22:10 generate no I'm
1:22:13 not I'm going to go select and then I'm
1:22:15 going to go generate
1:22:18 no oh I'm going say describe the desired
1:22:21 uh uh
1:22:27 remove
1:22:30 text and now I'm going to say
1:22:49 generate it doesn't seem as long when
1:22:52 you
1:23:00 Poes well got rid of it got rid of the
1:23:02 text in one of
1:23:05 them that's not bad though and then I
1:23:08 can wait how do I do it do
1:23:12 I
1:23:14 accept I go this
1:23:18 yeah and let me
1:23:21 go
1:23:23 so let's
1:23:25 see more robots
1:23:30 please and then I'll go over here and
1:23:33 I'll go generate let's see if it
1:23:36 finishes that part of the image pretty
1:23:38 good per
1:23:40 good so what I'm on right now
1:23:43 is open ai's version of out painting and
1:23:48 all this sort of stuff I mean this stuff
1:23:50 exists in Photoshop this see this look
1:23:52 looks like it's about as good
1:23:54 as out painting in Photoshop look I have
1:23:57 short sleeves and no
1:24:00 arms now total double amput there's sort
1:24:05 of an
1:24:06 arm is that an arm it's a robot
1:24:10 arm
1:24:13 huh I like that little
1:24:16 thing anyway all right that's that all
1:24:19 right that's good to remember there's an
1:24:21 option there that's this seems new this
1:24:23 seems like a bit of an afterthought as
1:24:25 as a piece of
1:24:27 software but that's okay CU their other
1:24:30 software is changing the flipping
1:24:33 world all right let me let me hop back
1:24:35 let me hop back and just let's let's get
1:24:37 talking again let me let's get to some
1:24:39 comments let's get to some some
1:24:45 action oh this was this is a favorite
1:24:48 thing of mine from my uh one of my timer
1:24:51 things I told it to write AI learning
1:24:54 lab and it wrote AI Leia
1:24:56 [Laughter]
1:25:00 lab it's like AI learning lab except
1:25:04 completely different I I'm just here to
1:25:07 learn awesome I am too we're all here to
1:25:11 learn I learn more doing these than I'm
1:25:15 teaching I I am convinced of that I'll
1:25:17 tell you that right
1:25:20 now oh that one's cool Ted
1:25:23 look at this so the Irregulars are in
1:25:25 making some images on the AI Salon oh
1:25:28 you know what I didn't talk about the AI
1:25:29 Salon did I okay how many people here
1:25:31 there's a good amount of people here sky
1:25:34 high
1:25:39 prices I like how the planes inside the
1:25:45 terminal that's really
1:25:51 funny hedge fund man manager get
1:25:56 [Laughter]
1:25:58 it oh that's so cute that is very cute
1:26:06 okay all right fantastic fantastic Bob
1:26:10 tell everybody is one listen let's go
1:26:12 talk about some things you me everyone
1:26:15 come here come over here I would like to
1:26:17 talk to you about
1:26:20 something yes that's it that's it right
1:26:37 there okay so if you're digging the
1:26:41 energy here and and the spirit of what's
1:26:43 going on here about a year ago I started
1:26:46 this group called the AI salon and this
1:26:48 is kind of a spin-off of it I I started
1:26:50 them at the same time actually but I
1:26:51 didn't start doing these lives till
1:26:54 about April of last year and then really
1:26:57 it's been in the past 2 or 3 months that
1:26:59 I've
1:27:00 been um kind of crossing the beams and
1:27:04 and mixing this community with the AI
1:27:06 Salon community and it's been pretty
1:27:08 magical and so the salon is is think of
1:27:14 it like an AI an AI selfhelp group or an
1:27:17 AI support group um it's a it's it's a
1:27:22 group of AI adventurers who are choosing
1:27:25 to be on this adventure together right
1:27:28 and so um if you go to the salon. the
1:27:32 first link there describes what the
1:27:34 salon's all about and it describes our
1:27:37 values and the values are things like
1:27:40 like curiosity and
1:27:44 um and generosity so like get curious if
1:27:48 you're curious if you want to explore
1:27:50 this stuff if you're willing to like
1:27:52 share what you're learning like don't
1:27:54 just learn it but like get generous
1:27:56 share it hey can I show you something oh
1:27:58 this is really cool let me show you what
1:27:59 I did let me show you what I did there's
1:28:02 this industry is moving too fast and
1:28:04 expanding too fast right now to be
1:28:06 stingy because anything that you think
1:28:08 is interesting and proprietary today is
1:28:10 going to be absolutely irrelevant a
1:28:12 month from now so when you're in the
1:28:15 beginning of an
1:28:17 industry get generous you can get stingy
1:28:21 when things stabilize but things are not
1:28:22 going to stabilize with AI for at least
1:28:24 3 years and probably more like five so
1:28:27 you have lots and lots of time to be
1:28:29 generous so if you want to if if that
1:28:31 resonates with you cool I want people
1:28:33 that are into exploration like are you
1:28:35 adventurist do you want to like find the
1:28:36 boundaries find the edges of this thing
1:28:39 um are you willing to collaborate you
1:28:41 know kind of companion to generosity um
1:28:45 inclusion is one of our values like
1:28:47 people are coming from all over nobody
1:28:48 knows what they're doing right now and
1:28:50 then the final one is empathy
1:28:53 this is scary [ __ ] for a lot of people
1:28:54 this is scary [ __ ] for people that are
1:28:56 in here regularly they're trying to
1:28:58 figure out holy [ __ ] does this change my
1:29:00 job yes does this do am I going to have
1:29:04 another job yes what's that going to be
1:29:05 I don't know oh my God that's scary like
1:29:08 it's scary so people are going to be
1:29:10 angry people are going to be scary
1:29:12 people are going to be resistant just
1:29:14 recognize that that's coming from a
1:29:17 place of
1:29:18 fear right they're not just [ __ ]
1:29:23 they're like they're like dealing with
1:29:24 [ __ ] we're all dealing with [ __ ] so
1:29:27 anyway so that's the AI salon
1:29:29 so couple of things
1:29:32 the
1:29:36 um on the on this this link tree the
1:29:42 first thing talks about the salon the
1:29:44 second one is the link to the Discord so
1:29:46 when we keep talking about the Discord
1:29:48 that's what it is so if you're into
1:29:49 these values if this sounds right cool
1:29:51 come join us
1:29:53 when you introduce yourself when you get
1:29:54 in there you you get dropped into the
1:29:56 introduce yourself
1:29:58 channel the second value is generosity
1:30:01 introduce yourself generously don't just
1:30:04 like you know lurk on in like introduce
1:30:07 yourself tell us who you are what you're
1:30:09 up to um share where you are on this AI
1:30:13 Journey maybe you've been playing with
1:30:15 GPT since it came out maybe you just
1:30:17 like stumbled upon this because you
1:30:19 thought I was Tom
1:30:21 Arnold that's fine
1:30:24 too but share what you're up to and then
1:30:26 and then join in the
1:30:29 community Becky Ru she was afraid but I
1:30:32 ain't afraid no more
1:30:35 exactly figure it
1:30:37 out everybody's trying to figure it out
1:30:40 and I you know I find myself sort of
1:30:42 going holy [ __ ] I'm as excited as I've
1:30:44 ever been holy [ __ ] I'm as scared as
1:30:46 I've ever been holy [ __ ] I'm as excited
1:30:48 as I've ever been and sometimes it's
1:30:50 doing that at the same time know we
1:30:53 we're in Crazy Seas right
1:30:55 now um yeah so
1:30:59 anyway anyway that's that's that's the
1:31:02 AI salon and we're coming up on our
1:31:04 one-year anniversary we're coming up on
1:31:06 the one-year anniversary of of chat GPT
1:31:09 did you announce the song from the
1:31:11 guitar clip yes I did it is drum roll
1:31:16 please
1:31:21 um
1:31:23 wait okay I'm holding the guitar weird
1:31:27 [Music]
1:31:45 people sit in this lonely
1:31:48 Town wonder when things are going to
1:31:51 change
1:31:52 [Music]
1:31:54 dream my life
1:31:56 away seems these dreams have turned to a
1:31:59 bunch dust
1:32:02 clouds get my nerve up but my past is
1:32:06 pulling me
1:32:09 down wondering how
1:32:12 long this black
1:32:15 sheep Black Sheep going to stick around
1:32:18 Black Sheep by Martin
1:32:21 sexon
1:32:23 [Music]
1:32:30 yeah Champion's not in here he didn't
1:32:31 want to be in here tonight he was
1:32:33 singing
1:32:34 [Music]
1:32:36 earlier thank you felonius thank you
1:32:39 thank you thank you thank you thank you
1:32:41 appreciate that Lauren thank you so
1:32:45 much yeah Black Sheep by Martin ston he
1:32:48 is if you don't know Martin
1:32:50 seon holy [ __ ] you got to see him live
1:32:53 thank you very much Andrea um you got to
1:32:56 see him live he sings like a [ __ ]
1:33:00 angel um and he's just his songs are
1:33:03 beautiful he plays he's one of those
1:33:06 guys
1:33:07 that he knows his songs so well that he
1:33:11 never kind of plays them the same way
1:33:13 twice and and he's just and he and he
1:33:15 like he plays the audience as much as he
1:33:17 plays his guitar it's it's it's really
1:33:20 good Jacob K have you guys been watching
1:33:22 J Jacob ker lately holy [ __ ] that guy I
1:33:25 mean a that guy's like [ __ ] Mozart
1:33:27 like he's like Mozart level talented and
1:33:31 like his his concerts now like he plays
1:33:34 the audience like a synthesizer it's
1:33:37 it's insane yeah Martin seon check him
1:33:39 out he's great he's really great that's
1:33:43 kind of his famous song Black Sheep is
1:33:45 the one he ends his shows with it's it's
1:33:46 it's a really good song I took a
1:33:49 songwriting Course once and
1:33:54 it was like a three-day three-day
1:33:55 songwriting workshop and Martin sexon
1:33:58 was one of the one of the guest teachers
1:34:01 and he came in and he did a two-hour
1:34:03 session with us on songwriting and he
1:34:05 played some songs and at one point it's
1:34:07 a really small room it was there was
1:34:08 like 20 of us in the room and and Marty
1:34:11 right in the
1:34:12 front and he starts singing this one
1:34:15 song and he he had a microphone but it
1:34:19 was a really small room so he didn't he
1:34:21 just didn't need the microphone so he
1:34:23 just stepped in front of
1:34:25 it and and whenever he he hit whatever
1:34:28 the chorus was of this song that he was
1:34:30 singing and I've never heard anything
1:34:32 like it it sounded like he had no
1:34:36 microphone and when he hit like his
1:34:40 belt note of that song it sounded like
1:34:44 trumpets were come coming out of his
1:34:46 face like there's something about the
1:34:48 Resonance of his skull that when he
1:34:52 sings his his loud high notes it's just
1:34:55 like it's it was like trumpets coming
1:34:58 out of his face I've never heard
1:34:59 anything like it it's absolutely
1:35:01 insane Nancy you're very talented Kyle
1:35:04 thank you very much I appreciate that um
1:35:08 that music was awesome yeah he's amazing
1:35:10 he's
1:35:11 amazing he's amazing my so here's my
1:35:13 problem with all of the songs I know all
1:35:16 of them all of them they're all [ __ ]
1:35:19 depressing they're all I don't there's
1:35:22 something about depressing songs that
1:35:24 just do it for me the new one I'm I'm
1:35:25 working on right now and this one's hard
1:35:27 to
1:35:28 [Music]
1:35:42 play it's uh if we were vampires by
1:35:45 Jason
1:35:46 Isbell single-handedly one of the most
1:35:49 beautiful songs on the planet and one of
1:35:51 the most depressing songs on the planet
1:35:54 it's it's so beautiful and it's hard to
1:35:57 play and sing at the same time so that's
1:36:00 what I'm working on this is one of my
1:36:01 add things is um thank you very much
1:36:06 John lost
1:36:09 connection that is the Papa J depressing
1:36:12 songs are the best they are man a lot of
1:36:15 the Brandy Caril songs are just [ __ ]
1:36:18 sad but Heart Like heartbreaking but
1:36:20 [ __ ] amazing you back it's I've Tik
1:36:23 tok's quit on me three times now thank
1:36:26 you Andrea um there's something oh oh
1:36:30 this was this is a actually there's
1:36:32 something potentially interesting for
1:36:34 all of us
1:36:37 here
1:36:39 so one of my one of my favorite um
1:36:43 YouTube videos is Ira Glass talking
1:36:46 about creative people and why most
1:36:49 creative people quit and they quit
1:36:52 because they have taste and when they
1:36:54 start learning
1:36:57 something they're disappointed at what
1:36:59 they create because they have taste and
1:37:02 he said the the the art of art is
1:37:06 sticking with it long enough to get good
1:37:08 enough to catch up to your taste and he
1:37:10 said that's the tragedy is people quit
1:37:12 before they they get there um and I
1:37:15 think these AI tools are going to allow
1:37:17 us to to all sort of yes you do have
1:37:20 taste we all have taste
1:37:23 I think these AI things are going to
1:37:25 level us up more quickly do you take
1:37:27 request I'm new here do you take
1:37:31 request yeah I do if you've got if
1:37:34 you've got a depressing song I will
1:37:36 happily sing it for you
1:37:39 um so I love that thing so so I was
1:37:41 taking guitar lessons this was God this
1:37:44 is probably back in the early 90s yeah
1:37:46 early 90s in New York
1:37:49 City and the guitar teacher was teaching
1:37:53 me some chords and teaching me this and
1:37:56 teaching me
1:37:58 that and then he kept playing this one
1:38:01 chord and I was and and and like the
1:38:03 chord was
1:38:06 just it just felt magic to
1:38:10 me and I said why does that chord sound
1:38:12 better than the other ones and it was an
1:38:14 A
1:38:16 Minor and he said oh it doesn't that's
1:38:20 just your taste
1:38:22 he goes that's your musical
1:38:24 taste and it it blew me away because to
1:38:31 me it
1:38:33 sounded just objectively better than the
1:38:36 other chords
1:38:38 right and what's funny is like all the
1:38:40 music I like like I get inspired by it
1:38:43 and it lifts me up and things like that
1:38:44 my co-founder of my business Mo'Nique if
1:38:47 we're ever driving and I put on my
1:38:49 playlist within [ __ ] 40 seconds she's
1:38:52 like oh this is
1:38:57 depressing it's just and it you know the
1:39:01 I think the lesson there is that you
1:39:04 know don't be afraid of having a point
1:39:06 of
1:39:08 view if you like depressing [ __ ] like
1:39:10 depressing [ __ ] if you like making crazy
1:39:13 wacky things with these AI tools that
1:39:15 other people don't understand and don't
1:39:17 get cool like Let Let your freak flag
1:39:20 fly
1:39:23 you know if if if your friends think
1:39:25 you're weird because you're doing this
1:39:26 AI [ __ ] [ __ ]
1:39:28 them if this is lighting you up
1:39:32 cool or if you're just like so mortified
1:39:35 that you got to learn more
1:39:38 great so I don't know
1:39:43 anyway music is a is a magical thing it
1:39:46 it really
1:39:50 is all right I got a 7:30 meeting
1:39:56 tomorrow I made a GPT for my playlist
1:39:59 it's very Niche oh that's cool
1:40:01 Vicky my internet's wonky all day yeah I
1:40:03 don't know what's going on with Tik Tok
1:40:06 crashing like it normally doesn't crash
1:40:07 but it's crashed four times now three
1:40:10 times long time listener first time
1:40:12 caller do you know any any John Prime um
1:40:14 I Used to Know angel from Montgomery but
1:40:17 I hav I haven't played it in Forever um
1:40:20 and I I don't go de on John PR I know a
1:40:22 lot of John hayatt I know a lot of uh
1:40:27 let's see
1:40:31 um
1:40:33 [Music]
1:40:42 uh got a blue
1:40:44 telescope looking at the world tonight
1:40:48 through a blue
1:40:49 telescope I wish I may wish I might not
1:40:53 see what I
1:40:55 seen sheet metal on sheets of
1:41:02 eyes there's that one and um it [ __ ]
1:41:05 horribly depressing song about him
1:41:07 witnessing a car
1:41:09 accident just
1:41:11 [Music]
1:41:19 like is it place I can rest my PO
1:41:24 [Music]
1:41:26 head gather my thoughts in sweet
1:41:33 silence and is this place where the
1:41:36 feelings AR
1:41:38 [Music]
1:41:39 dead from an overe exposure to
1:41:43 violence and is this place I can slowly
1:41:47 face the only one I truly can know
1:41:52 these are tears from a long time ago got
1:41:56 these tears from a long time ago I sang
1:42:00 I sang that song I went to an open mic
1:42:03 one
1:42:05 night
1:42:07 um and it was a combination open mic
1:42:10 night for singer songwriters and
1:42:12 Comedians and I sang that song was the
1:42:15 third song in my set and the MC of the
1:42:18 thing goes all right try to be a
1:42:20 comedian after that
1:42:24 one [ __ ] killed the
1:42:29 room this is way better than the battle
1:42:32 and the battle was pretty
1:42:35 cool this is a whole new Kyle but not at
1:42:38 all surprising oh that's sweet thank you
1:42:40 dude I've been rewriting my introduction
1:42:42 for the last 10 minutes is there a GPT
1:42:44 for
1:42:46 this yeah oh you know what I did today
1:42:49 this is a cool idea um
1:42:52 so I did a I did a keynote presentation
1:42:56 back in April for a creative mornings
1:42:58 thing about comparing the early days of
1:43:00 the worldwide web with the early days of
1:43:02 the of this AI stuff and and it was it
1:43:05 was it was a decent presentation it was
1:43:06 like 20 minutes um but I took the I took
1:43:10 the transcript of it
1:43:11 today and and what did I do I said
1:43:16 um oh I so so because chat GPT now has
1:43:22 128,000 token context
1:43:24 window you can paste in really long
1:43:27 transcripts this was only a 20- minute
1:43:29 um 20 like 26 minute transcript but it's
1:43:31 still it's still decent it was still
1:43:32 longer than chat GPT could have dealt
1:43:34 with before so I pasted in the
1:43:36 transcript of the talk that I gave I
1:43:39 told it
1:43:42 to pull out the salian
1:43:45 ideas and summarize them I said I said
1:43:49 summarize the talk
1:43:52 and then I
1:43:54 said Give me the give me the Salient
1:43:57 topics if they were chapters in a book
1:44:01 and outline what the elements of the
1:44:04 chapter would be so I had to take this
1:44:06 speech turn it into chapters of a book
1:44:09 and put kind of bullet points right and
1:44:11 then I said
1:44:14 um give me a first draft of chapter one
1:44:17 that incorporates the tone from the
1:44:20 transcript into the book but write it as
1:44:23 if it were a business
1:44:24 book and in [ __ ] 10 minutes I wrote
1:44:29 like a seven seven chapters of a book
1:44:32 now are the chapters good they're
1:44:36 okay um but then what I did was I
1:44:42 said I said read the book and now
1:44:45 critique it and tell me where it where I
1:44:47 can go deeper where I can add
1:44:50 stuff and then I said okay here's what I
1:44:52 want you to do I want you to work as a
1:44:55 writing consultant and I want I want to
1:44:57 go chapter by chapter and I want to have
1:45:00 I I want to have you summarize the
1:45:03 chapter and then ask me a specific
1:45:06 question based on your recommendations
1:45:09 of how we can make this better have me
1:45:11 fill it in and then I'm going to have
1:45:13 you incorporate what I say back into the
1:45:16 chapter and so that's when I kind of ran
1:45:18 out of time today but like it was a
1:45:21 really cool process cuz it was it was it
1:45:23 was very much this collaborative back
1:45:24 and forth
1:45:26 thing that got me in 30
1:45:31 minutes like a third of a
1:45:37 book is it is it least at
1:45:41 80% so that was pretty [ __ ]
1:45:44 cool can you play the hit classic hey
1:45:47 chat gbt make me money you can make
1:45:50 money
1:45:52 I think we got to make that we got to
1:45:53 write that
1:45:54 song I can't write it though cuz that's
1:45:57 a happy song I can only do the the sad
1:46:00 version of it like once we've lost all
1:46:02 of our our life savings on chat GPT
1:46:05 creative mornings is still alive rad
1:46:07 yeah it is uh yeah it's still going
1:46:09 pretty strong here the Denver ones the
1:46:12 Denver ones get between 150 and 200
1:46:14 people like the the one I did was it was
1:46:17 it was big it was like 180 people came
1:46:19 it's pretty cool uh it's actually if you
1:46:24 go if you
1:46:29 care um where am I going uh
1:46:36 YouTube if you search creative
1:46:41 mornings Denver and then
1:46:47 Kyle Shannon I think it comes up
1:46:53 yeah Kyle Shannon it's called Kyle
1:46:54 Shannon from here to
1:46:57 [Music]
1:47:10 there Welcome to our stage Mr Kyle
1:47:13 Shannon CH everything we need to know
1:47:16 about on my first day of seventh grade
1:47:20 in
1:47:22 1977 it makes me 57 you don't have to do
1:47:25 the math I'm 58 now um but anyway that's
1:47:29 there all
1:47:31 right all right people what are we going
1:47:34 to do here we got anything else to do
1:47:36 this been a this has been a
1:47:37 non-traditional
1:47:40 evening what are you going to do what
1:47:42 are you going to do look at all those AI
1:47:45 videos
1:47:47 LOL so many AI videos make it a par
1:47:51 of hey there Delila that's pretty good
1:47:54 hey AI can
1:47:56 someone can wait can AI help someone
1:47:58 learn to play guitar it's funny it's the
1:48:01 things that require physical dexterity
1:48:03 like juggling
1:48:05 and musical instruments and things like
1:48:08 that what AI will do here's here's what
1:48:10 I'm convinced AI will do is AI will meet
1:48:14 you where you are personally so so we're
1:48:18 going to be able to create very
1:48:20 personalized education for our ourselves
1:48:24 what happened that first
1:48:26 [Laughter]
1:48:28 day I'll tell you here I I'll I'll I'll
1:48:31 do the presentation I mean I'll do part
1:48:34 of it I'll do the the first part of
1:48:40 it is this
1:48:42 it
1:48:45 no oh yeah it
1:48:49 is
1:48:52 all
1:49:00 right all
1:49:02 right so first day
1:49:08 of seventh grade I walked into my math
1:49:11 class in
1:49:14 1977 and on every desk was a
1:49:17 computer and let me here I'll put that
1:49:21 that
1:49:21 there um on every desk was a computer
1:49:25 and it it wasn't just any
1:49:28 computer it was a Radio Shack trs8 so if
1:49:32 gen xer is in the room you remember the
1:49:34 trash
1:49:35 80s and so I didn't know anything about
1:49:39 computers but you know on the first day
1:49:41 they taught us basic
1:49:43 programming and the first thing they
1:49:46 taught us with basic programming was the
1:49:49 L print command which was was basically
1:49:51 you could say l print and then some
1:49:54 character and it would print that to
1:49:56 screen I don't know why it was L print
1:49:58 and it printed the screen it should have
1:50:00 been s print but whatever you know Geeks
1:50:03 and so my buddy Scott and I we made an
1:50:06 animation an asy animation of a Rocket
1:50:09 taking off like on the first day when we
1:50:11 first learned how to do basic
1:50:13 programming we we did an animation of
1:50:17 this Rocket taking off and I the these
1:50:19 machines were just [ __ ]
1:50:22 magical and I remember being in seventh
1:50:26 grade and I remember being very very um
1:50:30 acutely aware that I was just too
1:50:35 young to be one of those guys Steve Jobs
1:50:38 and wnc right like a few years before a
1:50:42 few years earlier they invented Apple
1:50:47 and I was just too young to be one of
1:50:49 those guys that's Bill Gates and the
1:50:50 founders of
1:50:52 Microsoft and I
1:50:55 knew
1:50:57 that being at the
1:51:00 beginning
1:51:02 of the equivalent of the printing press
1:51:05 is not something that happens very often
1:51:08 if ever in your life and so I just
1:51:11 remember being acutely aware
1:51:15 of having missed it and then Flash
1:51:19 Forward from
1:51:21 1977 to
1:51:23 1994 I'm living in New York
1:51:26 City and I'm doing desktop publishing is
1:51:29 my day job I'm pursuing a life in the
1:51:31 Arts and then I learn about this thing
1:51:35 called the worldwide web and I teach
1:51:37 myself enough HTML to put together the
1:51:40 first issue of urban desires which is an
1:51:42 online
1:51:43 magazine and you know it was an art and
1:51:45 culture magazine and I didn't even know
1:51:48 like I I knew how to build the pages but
1:51:50 I didn't didn't even know how to get it
1:51:51 on the internet so someone said oh you
1:51:53 got to upload the files to this computer
1:51:56 in California that's a called a server
1:51:58 and so I did
1:52:00 [Music]
1:52:06 [Music]
1:52:08 that if you remember modems that's what
1:52:11 the internet sounded like back then you
1:52:13 had to call other computers and then
1:52:15 within 3 weeks of doing that I got an
1:52:18 email from someone that someone wrote
1:52:21 about Urban Desires in Liberation on the
1:52:23 Parisian News Daily and so the next
1:52:26 Sunday I went down to the international
1:52:28 news stand at Times Square and I bought
1:52:33 that issue of liberon and I flipped it
1:52:35 open and sure enough there was a full
1:52:38 page article of these HTML files that I
1:52:42 put on some computer in California
1:52:44 written about in French from Paris and I
1:52:47 knew in that moment that the world had
1:52:49 changed
1:52:52 and so that make it stop I know I know
1:52:56 that
1:52:57 sound
1:52:59 um that is what we're in the middle of
1:53:03 with this AI stuff and I can't describe
1:53:08 to you like how lucky I felt in 1994 to
1:53:12 be at the beginning of one of those
1:53:14 things knowing that I missed the one in
1:53:19 1977 and so in 1994 I I co-founded
1:53:23 agency.com which was one of the early
1:53:25 digital agencies and we ended up you
1:53:27 know growing really big and we built a
1:53:29 lot of the websites for the Fortune 500
1:53:31 so I got to play I got to be at the
1:53:35 beginning of a thing that changed the
1:53:36 world and it was as amazing as you think
1:53:39 it might be and we're there again we're
1:53:43 at the very beginning we are so early
1:53:46 you have no idea and it's one of these
1:53:49 like what's going on with generative AI
1:53:52 right now is happening so much faster
1:53:54 than the worldwide web did and it's it's
1:53:56 already so much more profound in terms
1:53:58 of its impact in the worldwide web was
1:54:00 and and so that's why I do this channel
1:54:03 that's why I come here seven [ __ ]
1:54:04 nights a week to talk about this
1:54:07 stuff cuz this is we're we're in such
1:54:11 rarified we're such in such a rarified
1:54:14 moment of history and so few people know
1:54:17 what's
1:54:18 happening and that just means we get to
1:54:21 play while other people are kind of
1:54:22 asleep right now they're you know
1:54:24 they're still all blue pilled out this
1:54:26 is red pill blue pill moment cuz
1:54:28 everyone that's on this channel that
1:54:29 comes here regularly like once you have
1:54:31 your Kevin McAllister moment thank you
1:54:33 Joe mama that moment once you have that
1:54:37 moment you can't unlearn it you can't un
1:54:40 have the Epiphany that holy [ __ ] the
1:54:43 world just changed but no one knows it
1:54:45 yet and that's what we're all playing
1:54:48 with none of us know what it means means
1:54:51 none of us know what success looks like
1:54:53 none of us know what work look like none
1:54:55 of us know anything right now except
1:54:59 that we know we've got to be doing
1:55:01 this so that's that's basically what
1:55:03 that talk was about
1:55:09 [Music]
1:55:11 so the music 30 BS thank you for every
1:55:15 hook nice I like them green eyes them
1:55:18 green eyes is nice they make me feel so
1:55:23 powerful there's a GPT for that I think
1:55:27 who did this one did Ted do this one I
1:55:29 forget who did this
1:55:33 one oh trauma response
1:55:39 achieved oh man I got the llamas and the
1:55:42 turtles and hedgehogs thank you so much
1:55:45 I now use chat GPT to do a
1:55:49 quiz
1:55:50 1994 I sat back and watched my friends
1:55:53 be part of History I won't make that
1:55:54 mistake again I on China [ __ ] more
1:55:57 power to you I listen it and you know in
1:56:01 1994 when cuz I you you know most of my
1:56:05 friend of circles back then were actors
1:56:07 and writers and
1:56:10 artists and I'm like you got to do this
1:56:12 internet thing and they were like they
1:56:14 they were at they were as afraid of the
1:56:16 internet back then as artists are of AI
1:56:18 today some of them were angry about it
1:56:21 some of them were just like they they
1:56:23 honest to God looked at me like what the
1:56:25 [ __ ] are you doing what like why would
1:56:27 you do this and even when they saw me
1:56:30 take off and and you know this this
1:56:33 thing I created just turned into this
1:56:34 Juggernaut they were like wow that was
1:56:37 you were really right about that like
1:56:38 yeah I probably should have probably
1:56:40 should have listened to you and you know
1:56:42 those same people today are going why
1:56:44 are you doing this AI stuff this is
1:56:46 weird what are you
1:56:48 doing like
1:56:51 I don't know my mother-in-law was a
1:56:53 writer editor for Liberation in those
1:56:55 days do you still have that copy I think
1:56:59 I
1:57:00 do I think do I have it behind
1:57:18 me
1:57:23 no I
1:57:24 don't right what's up
1:57:34 here wa is that
1:57:36 it
1:57:38 no no I don't think I have
1:57:42 it I've got in a box
1:57:48 somewhere
1:57:50 [Music]
1:57:53 does AI have
1:57:55 a pronounce I don't know what that
1:58:02 means I just had an
1:58:06 apostrophe now I use chat GPT to quiz my
1:58:09 son on books he's Reading Game Changer
1:58:11 yeah yeah oh Johnny's 24 honey um first
1:58:15 of all I love your name always have
1:58:17 always will
1:58:18 um
1:58:22 you should make I don't know how old
1:58:23 your son is but you should make your son
1:58:26 a
1:58:27 GPT for the lesson so so this was a this
1:58:31 was an epiphany I had from um from Peter
1:58:35 um oh it's Peter's last name in The
1:58:38 Salon Peter Kaminsky Peter Kaminsky
1:58:42 created a
1:58:43 GPT two days
1:58:45 ago of a conference call that he had so
1:58:49 he had a conference call with like six
1:58:50 people he took the transcript from the
1:58:53 conference call turned it into a GPT and
1:58:56 gave it to those people and said here
1:58:58 you can interact with our call now you
1:59:00 can like ask it questions so you could
1:59:02 do that for your son create custom
1:59:04 little homework interactive Bots for him
1:59:07 to play with do you think there was more
1:59:10 resistance do you think there is more
1:59:11 resistance to AI than there was to World
1:59:14 worldwide web in
1:59:16 1994
1:59:18 um
1:59:24 yes because nobody so so couple of
1:59:28 things are
1:59:29 different in
1:59:31 1994 um the internet wasn't ubiquitous
1:59:34 right you you had dialup modems and it
1:59:37 was not fast and and the early worldwide
1:59:40 web
1:59:41 was incredibly basic like I don't I
1:59:45 don't know if you saw the um wait I can
1:59:49 probably get it to come up here hang on
1:59:50 let me show you something here that's
1:59:52 pretty
1:59:54 funny
2:00:02 um so hang on let me
2:00:06 see
2:00:18 oops and you can't really this is what
2:00:21 the internet looked like in 1994 notice
2:00:23 how the image is left
2:00:24 Justified you know
2:00:27 why because Engineers of the worldwide
2:00:32 web would the center tag only applied to
2:00:35 text it didn't apply to images so you
2:00:38 couldn't Center images so so the early
2:00:42 days of the worldwide web were were very
2:00:44 basic and and so so people didn't know
2:00:49 um
2:00:50 um people didn't know about
2:00:53 it
2:00:55 um with
2:00:58 AI the internet's
2:01:01 ubiquitous um chat
2:01:06 GPT to get to the worldwide web to get
2:01:09 to 100 million users it took six years
2:01:12 right because we were building
2:01:13 infrastructure we were building the
2:01:14 tools we were figuring out what the [ __ ]
2:01:15 it was 6 years to get to 100 million
2:01:19 users
2:01:20 and everyone talks about oh you know
2:01:22 everyone lost their mind at the end of
2:01:23 you know 1999 going into the do crash um
2:01:28 that was that was like at the six year
2:01:29 mark like it was it was right when it
2:01:31 hit you know we' been working on it for
2:01:34 like 5 years when all that frothy [ __ ]
2:01:38 happened
2:01:39 right chat GPT got to 100 million users
2:01:43 in six
2:01:44 weeks so couple of things that are
2:01:48 different you you didn't have movies
2:01:50 from Hollywood saying that the internet
2:01:52 was going to kill
2:01:53 everyone we've we've had 40 Years of
2:01:56 sci-fi movies telling us that AI is
2:01:58 going to kill
2:01:59 us we've also had scientists telling us
2:02:02 that AI was coming AI was coming AI was
2:02:05 coming a I was coming for
2:02:07 decades right big blue it beats Gary
2:02:11 Casper off and then you know AI is
2:02:15 coming AI is coming and then you know
2:02:17 Watson health is going to change you
2:02:19 know there all the cancer diagnosis and
2:02:21 like all these promises all these
2:02:23 promises and nothing was there and then
2:02:25 all of a sudden Along Comes this chat
2:02:27 GPT thing that makes it as easy to use
2:02:31 as a Google Search and that's so so why
2:02:34 it gets to a 100 million users like that
2:02:37 is because it's this remarkably
2:02:41 powerful um
2:02:43 underbelly with a very very simple
2:02:46 interface and so so I think you've got a
2:02:49 combination of just fear fear from
2:02:52 Hollywood like societal fear that that
2:02:54 the machines are going to get
2:02:56 us I think there's a fear from most
2:03:00 non-technical people that in order to do
2:03:03 AI you have to have a degree from
2:03:05 Stamford and be good at math and be good
2:03:07 at
2:03:08 technology and that was true until last
2:03:12 November chat GPT changed that and and
2:03:16 now these gpts where you can just talk
2:03:19 and application into
2:03:23 existence like holy [ __ ] the power of
2:03:26 that so so I think that the resistance
2:03:29 is a lot more intense right now just
2:03:32 because of the rapid adoption um just
2:03:35 societal [ __ ] and the fact that people
2:03:38 are are just afraid of it and
2:03:41 plus people like us who [ __ ] taking
2:03:44 the red pill and they're like uh you got
2:03:45 to see this this going to change the
2:03:47 world everything's different you know
2:03:49 it sounds it sounds like we're in a
2:03:51 [ __ ]
2:03:54 cult what do you mean it's just like a
2:03:56 Google box right it's just like a Google
2:03:57 search isn't it it's just like a Google
2:03:59 Search right you just put in the words
2:04:01 and it it gives you back words right no
2:04:03 no it can see right it can read your
2:04:08 whiteboard and tell you what's on it I
2:04:09 don't understand it can make
2:04:12 images what do you mean like it does an
2:04:14 image search no it makes images I don't
2:04:18 understand
2:04:20 it can code for you huh like it's you C
2:04:24 this is not something you can explain to
2:04:29 people they can they can only experience
2:04:32 it and when they experience it they have
2:04:34 that
2:04:38 moment and then they either run from
2:04:40 that
2:04:41 moment or they go well let me just do
2:04:43 one more
2:04:46 thing well but it can't do what I do I'm
2:04:49 special all right oh
2:04:53 [ __ ] it can do what I do what do I do
2:04:57 now
2:05:00 help and here we all
2:05:04 are here we all
2:05:08 are uploading a conference called
2:05:11 transcript from my job to chat GPT would
2:05:13 get me fired fast oh yeah that's the
2:05:15 other thing don't put anything
2:05:17 proprietary in this [ __ ] right now it's
2:05:19 it's it's as leaky as a Civ cheese
2:05:22 graders cheese graders have more you
2:05:24 know liquid Integrity than than chat GPT
2:05:28 right
2:05:31 now oh am I a programmer I have
2:05:35 programmed but no I'm not a programmer
2:05:37 I'm more of a product guy and
2:05:39 a I'm more of a product
2:05:42 guy and a you know creative like
2:05:45 creative I was Chief creative officer of
2:05:47 my agency and
2:05:49 I'm a marketer and a I don't know
2:05:52 I'm I'm an
2:05:56 ADHD whatever ad ADHD people do Kyle
2:06:00 check out Show and Tell video
2:06:12 okay China thanks so much for
2:06:15 this which one the most recent one right
2:06:19 method convert a string to a
2:06:23 good now change it so it returns it in
2:06:27 base
2:06:29 64 now make it so it doesn't allocate
2:06:34 any memory on the H if it does that I'm
2:06:37 going to lose
2:06:39 it shut
2:06:44 the there's a Kevin mallister
2:06:48 moment
2:06:50 what the [ __ ] do you understand how NAU
2:06:55 that is right a SE method to convert a
2:06:59 string to a
2:07:02 good that's
2:07:05 it but certainly it can't do what I do
2:07:09 and then it
2:07:10 does and it's going to get better and
2:07:12 better and and the the people right now
2:07:15 that are sitting on the sidelines
2:07:17 judging it as not being
2:07:20 perfect and it's it it blows my [ __ ]
2:07:23 mind like these gpts have been out for
2:07:26 what 5 days six days and people are like
2:07:30 wow they're not perfect they got
2:07:31 security
2:07:32 H why would they release
2:07:38 this have you made
2:07:42 one do you see what's
2:07:46 happening and I listen I think I think
2:07:49 that that's just there's a human nature
2:07:51 thing that if we see something we assume
2:07:54 it's we assume that the people that made
2:07:56 it you know uh checked everything and
2:07:59 it's it's 100% perfect before we get in
2:08:02 it like a car or an
2:08:04 airplane and this stuff is not like that
2:08:07 so I get that it's it's a little
2:08:09 uh it's a little discombobulating when
2:08:12 it's not when when something demos as
2:08:16 well as what these tools do when they're
2:08:18 this magic
2:08:19 and then they're not perfect it's kind
2:08:21 of like ah [ __ ] it's like whoever said
2:08:22 that before where it's like I get in I
2:08:25 get excited and then I'm like ah it's
2:08:27 but it's not perfect darn it and then I
2:08:31 wait for the next thing I'm like oh
2:08:33 that's really good oh but it's not
2:08:36 perfect but that's that's again my like
2:08:42 what is what's a synonym for strong
2:08:44 advice my plea my my conviction My Hope
2:08:48 for you my Hope For
2:08:51 Humanity is that you forgive all of
2:08:53 those imperfections learn them all learn
2:08:56 where the boundaries are learn what it's
2:08:57 good at learn what it's not good at
2:08:59 learn where it falls on its
2:09:02 face but get curious and just stay
2:09:05 curious for the next three years and
2:09:07 you're like but Kyle that's exhausting
2:09:09 tell me about it you didn't come in here
2:09:12 seven nights a week isn't exhausting
2:09:15 it's exhausting Jim Ross had chat GPT
2:09:18 create an ended your client feedback
2:09:20 document which I know helped me retain
2:09:23 clients
2:09:25 perfect like the speed with which you
2:09:28 can knock stuff out like what's cool
2:09:30 about Jim Ross so so Jim Ross has has
2:09:32 done some some super cool things we one
2:09:34 of on the first night that we got gpts
2:09:36 we built Jim a GPT for his business
2:09:39 which he's since gone and rebuilt right
2:09:41 we sort of did it in real time
2:09:44 here but he's got such clear tasks he
2:09:48 needs for his his business and basically
2:09:50 what he's done is he's like I'm going to
2:09:53 try to have chat GPT do those tasks for
2:09:56 me so a Tas comes up and he's like
2:09:58 wonder if chat jbt can do that you know
2:10:01 people have always told me I should
2:10:02 write a book wonder if I can write a
2:10:05 book then he has a book huge traffic now
2:10:08 on open AI yeah I know they're they're
2:10:11 getting hammered they're getting
2:10:12 hammered for
2:10:16 sure Sam mman for the second year in a
2:10:20 row we didn't anticipate this much
2:10:23 enthusiasm for our
2:10:27 product it's so funny they launched chat
2:10:29 GPT they built it in two weeks which is
2:10:34 staggering it got to a million users in
2:10:37 5 days it got to 100 million users in
2:10:39 six weeks and they were like God we were
2:10:44 not expecting that and then they they
2:10:46 basically launched these GP these
2:10:49 talking app into existence that that
2:10:52 have incredible depth to
2:10:54 them
2:10:57 and people are using them and they're
2:11:00 like why are people using these so
2:11:03 much oh we get bored of the toy toy and
2:11:06 we end up playing with the Box yeah
2:11:07 absolutely this code is kind of her Bose
2:11:10 too many comments exactly
2:11:14 exactly
2:11:16 exactly uh there was just a there was
2:11:19 someone pinned a comment that I missed
2:11:21 before if you want to pin that again
2:11:22 I'll take a look at
2:11:23 it um Google Maps gets a massive aii
2:11:28 upgrade with five new
2:11:29 features
2:11:32 yeah
2:11:34 yeah those are regulated this is not I
2:11:38 don't know what those is in that case
2:11:40 when I'm here I'm optimistic when I'm
2:11:42 out I'm nervous for the future oh that's
2:11:45 well that's first of all that's very
2:11:46 sweet thank you um
2:11:54 I
2:11:56 think for me the thing the thing that
2:11:59 keeps me not nervous about the future is
2:12:04 just staying engaged with the things I
2:12:06 have control over like I don't have
2:12:09 control over how this [ __ ] gets
2:12:10 regulated I don't have control over how
2:12:13 Sam mman and and the kids over there at
2:12:15 open AI are going to fix all these bugs
2:12:18 I don't have control over if Elon musk's
2:12:22 open Cowboy attitude is going to unleash
2:12:25 something on the world that's going to
2:12:26 kill [ __ ]
2:12:27 so to the to the extent that I can I try
2:12:32 not to worry about things that are that
2:12:34 out of my
2:12:35 control what's in my control I can learn
2:12:38 this [ __ ] I can talk to you people about
2:12:40 this [ __ ] I can watch what you're doing
2:12:41 and get inspired by it um every time I
2:12:45 see someone have a Kevin mallister
2:12:47 moment that that programmer well at
2:12:50 least it's not going to do this thing oh
2:12:52 [ __ ] it did that
2:12:54 thing
2:12:57 like that is that is people that are not
2:13:01 going to go back from this that's
2:13:03 people they might not like it they might
2:13:07 not choose to pursue it but they can't
2:13:09 unknow that they can't unknow that
2:13:12 Google sues scammers out of fake bar AI
2:13:15 chatbot that downloads malware oh God
2:13:19 malw War fraudsters tricked users get
2:13:22 ready 2023 or
2:13:26 2024 the presidential
2:13:28 election
2:13:30 and the amount of Fraud and the the
2:13:33 sophistication of fraud is going to go
2:13:36 up so dramatically we're all going to
2:13:39 get our asses handed to us I think I
2:13:42 love that AI connected me to this group
2:13:44 of like-minded and curious people that's
2:13:46 great Lauren I I mean yeah that's
2:13:49 that's what the salon's about that's
2:13:50 what this is about it's just hanging out
2:13:53 trying to figure it out and and like
2:13:55 yeah we're [ __ ] weirdos cuz why are
2:13:59 you paying attention to all that
2:14:02 [ __ ] have you played with it
2:14:06 no okay you want to see what it does not
2:14:09 really okay 6 months later hey Kyle yeah
2:14:15 could you show me that
2:14:17 thing I heard I'm supposed to learn it
2:14:20 now I just got laid
2:14:23 off I worry about passwords and
2:14:25 brokerage accounts you should
2:14:28 tck t-rock are you the t-rock from neot
2:14:33 Toyo yeah password passwords passwords
2:14:36 are [ __ ] uh you joked the other night
2:14:39 about Gen X using AI the most but I'd
2:14:41 love to see the stats on who's using it
2:14:43 I would too actually Athena
2:14:47 Luna I someone also pointed out that
2:14:50 I've got an echo chamber in here that's
2:14:52 more Gen X than not which is true it we
2:14:55 did a we did a little test it was like
2:14:58 28% Millennials and Boomers and 47 or 8%
2:15:02 gen xers so we're we're uh indexing
2:15:06 heavy on the Gen xers here but you know
2:15:08 consider the
2:15:12 source you get a smartass Gen X or you
2:15:15 you're going to attract you're going to
2:15:16 attract some other salty xers that are
2:15:19 like yeah [ __ ] everyone
2:15:22 else have you explored the AI and
2:15:24 Microsoft Paint on Windows 11 yet I I
2:15:28 have not I am a Mac guy uh I feel I have
2:15:34 um genuflected to Microsoft Enough by
2:15:37 making Edge my default browser and Bing
2:15:40 my default search engine which I've got
2:15:42 to change back at some point but I don't
2:15:45 really use search engines anymore to be
2:15:47 quite honest
2:15:49 so um
2:15:52 yeah AI learns from us am I late
2:15:55 [Laughter]
2:15:59 [ __ ] oh
2:16:01 man don't say buzzwords you'll cut out
2:16:05 what buzzword did I
2:16:06 say AI robot uses meteorite from Mars to
2:16:10 help make oxygen from water a robot uses
2:16:14 a meteorite from
2:16:16 Mars to help make
2:16:19 oxygen from water well that would be a
2:16:22 big [ __ ] deal wouldn't it took 5
2:16:25 minutes to add monetization to my chat
2:16:27 GPT
2:16:29 wow Joe mama how did you do that did you
2:16:32 basically just build it in and say hey
2:16:34 if you want more of this click here is
2:16:36 that is that are you monetizing it
2:16:37 within the
2:16:39 conversation refreshed shoe cleaner my
2:16:42 name's Kyle it's the best name as George
2:16:45 Carlin would say it's a soft
2:16:47 name
2:16:48 what's he say Rick and Joe and Bob can
2:16:51 kick the [ __ ] out of Kyle Trevor and
2:16:56 something I gotta go find
2:16:59 that you ever heard that one let's see
2:17:02 uh
2:17:05 Carlin on soft
2:17:15 names build a website in 2 minutes 38
2:17:18 seconds using AI just add sections and
2:17:20 have our
2:17:22 AI but it didn't happen and one of the
2:17:25 reasons one of the reasons is because we
2:17:26 were using that soft language that
2:17:29 language that takes the life out of life
2:17:32 and it is a function of time it does
2:17:34 keep getting worse give me another
2:17:35 example sometime during my life sometime
2:17:38 during my life toilet paper became
2:17:40 bathroom
2:17:44 tissue he goes on I forget where the the
2:17:47 soft names are
2:17:48 are but I think it's in that I think
2:17:51 it's in that one on soft
2:17:54 language your old disease
2:17:58 names goofy boy names this could be
2:18:03 it thank
2:18:05 you thank
2:18:08 you thank you very much well tell you
2:18:11 what now to balance the scale I'd like
2:18:13 to talk about some things that bring us
2:18:15 together things that point out our
2:18:17 similari instead of our differences cuz
2:18:19 that's all you ever hear about in this
2:18:20 country is our differences that's all
2:18:22 the media and the politicians are ever
2:18:24 talking about the things that separate
2:18:27 us this man was so far ahead of his time
2:18:31 things that make us different from one
2:18:32 another that's the way the ruling class
2:18:34 operates in any society they try to
2:18:37 divide the rest of the people they keep
2:18:38 the lower and the middle classes
2:18:40 fighting with each other so that they
2:18:41 the rich can run off with all the
2:18:43 [ __ ] money fairly simple thing
2:18:46 happens to work you know anything
2:18:48 different that's what they're going to
2:18:49 talk about race religion ethnic and
2:18:51 National background jobs income
2:18:53 education social
2:18:55 status I don't know where the name thing
2:18:57 is don't get banned I did I don't know
2:19:00 what words I said that would get me
2:19:01 banned I didn't get any warnings
2:19:04 so I antagonized bing until it only
2:19:07 responded it resp only responded to me
2:19:09 in Chinese that's [ __ ] awesome can AI
2:19:12 do my Amazon payperclick or PPC probably
2:19:16 yes do you do private calls I'm happy to
2:19:19 happy to chat yeah uh just DM me I'm
2:19:22 happy to talk um we can set something
2:19:26 up um yes but I think it's a group like
2:19:30 this that helps tosolve hard
2:19:32 generational res resentful lines I hope
2:19:35 so quick take Jake I mean that's listen
2:19:37 I
2:19:38 try I try really hard to not be
2:19:42 political in any way um I try to be
2:19:45 empathetic in every way
2:19:52 um
2:19:55 yeah I you know there's a part of
2:20:03 me as this community's gotten Tighter
2:20:08 and I'm watching people over time evolve
2:20:11 and Blossom and do all this [ __ ] it's
2:20:13 pretty
2:20:15 intense cuz I'm like you know know I go
2:20:19 through these moments of is this stuff
2:20:21 just a big waste of time is this just a
2:20:25 blip but I don't think it
2:20:28 is like the like the the the evidence is
2:20:32 just overwhelmingly that this that
2:20:35 everything's Chang that everything has
2:20:38 changed but no one knows it
2:20:41 yet GPT went big Pharma did it what did
2:20:46 that say can you re post that one for
2:20:49 Gen X watch old dad dads on Netflix Bill
2:20:52 Burr is now my hero Bill Burr has been
2:20:54 my hero for ages I love Bill Burr he's
2:20:56 he's so
2:20:58 angry Bill Burr Bill Hicks
2:21:01 um yeah all the angry GPT went big
2:21:05 Pharma I could research natural medicine
2:21:07 before and now it's like
2:21:09 WebMD ah don't worry about that Jules
2:21:11 just go get one of the open source go
2:21:13 play on
2:21:15 um labs. perplexity . a and that's the
2:21:19 Llama 2 model and I I don't know if
2:21:21 that's the unrestricted one but you can
2:21:23 if you want to install an unrestricted
2:21:25 model you can do that you could spin up
2:21:28 a a GPU virtual machine and put a
2:21:31 unrestricted large language model in
2:21:33 there I know a guy that put the 70
2:21:36 billion parameter llama 2 model on his
2:21:39 company's slack Channel completely
2:21:41 unrestricted he said it's [ __ ]
2:21:44 insane he said it cusses all the time it
2:21:47 just loses its mind sometimes and like
2:21:49 speaks Spanish for days at a
2:21:51 [Laughter]
2:21:58 time so yeah if you want to do that he
2:22:01 was a prophet he was a
2:22:03 prophet he predicted the future bigger
2:22:05 screen
2:22:06 please one of my quotes is language is
2:22:10 the Cornerstone of
2:22:12 civilization this is why this is why
2:22:15 it's funny in the in the when when chpt
2:22:18 first came
2:22:20 out and I and I was hearing Universe
2:22:23 higher ed like we got to stop this we
2:22:26 got to stop this I was kind of
2:22:28 ambivalent about it and and The more
2:22:31 I've gone on the more incensed I am that
2:22:33 higher education and education in
2:22:37 general is resisting
2:22:41 this it's a it's a large language model
2:22:45 that's what they're
2:22:46 called
2:22:48 they get their power from all of our
2:22:51 words who else are large language models
2:22:56 humans whose job is it to teach the
2:22:59 humans the
2:23:01 education
2:23:04 institution and you have a thing called
2:23:06 a large language model that can augment
2:23:09 these large language models and you're
2:23:11 going to Chuck this out the window or
2:23:15 ignore
2:23:16 it
2:23:18 I don't I like I I cannot process how
2:23:23 shortsighted it is you're amazingly
2:23:26 informative we appreciate you thank you
2:23:27 techie Becky I appreciate that a
2:23:30 lot all right I gotta I got to start
2:23:33 whing this down people from ti994a
2:23:36 animations in 1982 to something Carlin
2:23:39 was amazing saw him live on an arber wow
2:23:42 that's amazing Kyle are you concerned
2:23:44 that companies like open AI will be
2:23:47 persuaded by lobbying and money I'm
2:23:49 johnz I'm more I'm actually more
2:23:51 concerned that companies like open AI um
2:23:55 are going to be the lobby right so so
2:23:58 you've got the go look at the frontier
2:24:00 is it the Frontier Foundation the
2:24:03 frontier the [ __ ] it called anyway
2:24:06 Microsoft open AI Google um anthropic a
2:24:10 bunch of the big players put together
2:24:13 this group and they they went to
2:24:14 Washington and they said you should
2:24:16 regulate us I think they call it
2:24:18 regulatory capture so basically the big
2:24:21 companies the big funded companies of
2:24:24 today are basically trying to influence
2:24:26 Washington to write regulations that
2:24:29 benefit them now if I were that a big
2:24:32 company I would do that too but that's
2:24:34 that's lobbying so
2:24:37 um so I don't know that they will be
2:24:40 persuaded by lobbying money I think that
2:24:42 they will be the lobbying
2:24:44 money they will do the persuading um I'm
2:24:47 fairly confident of that
2:24:54 actually you know I hey at at least we
2:24:57 have the executive order which I made a
2:24:59 GPT if you want to go um interact with
2:25:04 the executive order it's called the AI
2:25:06 executive order Explorer by Kyle Shannon
2:25:09 that's my GPT where you can go ask
2:25:11 questions about the executive order at
2:25:13 least we've got at the highest level of
2:25:16 government
2:25:18 a point of view because we're not going
2:25:20 to get [ __ ] from Congress this Congress
2:25:23 can't wipe its butt so so what we have
2:25:26 in that executive order is what we're
2:25:27 going to get for the next year and a
2:25:30 half and and then who knows
2:25:34 um yeah
2:25:36 so so yeah I am I worried about it again
2:25:39 I go back to this thing I I what I'm
2:25:42 trying to do is focus on what what is in
2:25:44 my
2:25:45 control lots of new viewers please run
2:25:48 through the links quick okay I'll do
2:25:50 that happy to do that um okay so let's
2:25:54 flip screens let me show you a couple of
2:25:57 things
2:26:02 here uh all right if you are brand new
2:26:07 to the chat
2:26:09 JP go there right now you know you
2:26:13 ignore all these bullets this is just
2:26:15 from a larger presentation this was this
2:26:17 was all the [ __ ] that was announced last
2:26:19 Monday every one of those bullets is a
2:26:22 full-on [ __ ] rabbit hole but don't
2:26:24 worry about that
2:26:28 now go there chat. open.com right now in
2:26:33 my opinion that's the only site you need
2:26:35 to be playing on it's the only one and
2:26:38 yes you should pay the 20 bucks a month
2:26:40 if you can at all afford
2:26:42 it I think it is
2:26:45 personally irresp
2:26:48 responsible if you're interested in AI
2:26:50 in any
2:26:51 way if you can afford
2:26:54 it I think it's irresponsible not to pay
2:26:56 the 20 bucks for that I'm not affiliated
2:26:58 with them in any way I don't make a
2:27:00 [ __ ] penny we need a George Carlin
2:27:02 GPT that would be pretty easy to it
2:27:04 would be pretty easy to make actually um
2:27:07 so that's that I'm going to I'm going to
2:27:09 show you a whole bunch of get ready to
2:27:11 do some screenshots I'm just going to go
2:27:13 through a bunch of links that if you
2:27:14 want to run down rabbit holes you can do
2:27:16 that but seriously the next three months
2:27:18 of your life should probably be spent
2:27:19 figuring out what this thing can do all
2:27:21 right get ready to screenshot people all
2:27:24 right so you can screenshot that the
2:27:26 thing on the right is the developer the
2:27:28 back end of chat GPT the open AI
2:27:31 playground and all the stuff that got
2:27:33 added there all right so that's that all
2:27:37 right now the next two slides are kind
2:27:39 of the the big eight behind chat GPT
2:27:43 right now these guys are King of the
2:27:44 Hill and until further notice
2:27:48 um the big four I put grock in here
2:27:51 simply because Elon Musk is going to
2:27:54 will that into existence so that doesn't
2:27:56 even exist right now except in very very
2:27:58 limited beta but probably within 6
2:28:00 months that's going to be big so just
2:28:01 pay attention to that
2:28:04 Claude they're licking their wounds in a
2:28:06 corner right now because what open AI
2:28:10 just releas kind of stuffed them back in
2:28:13 a corner because their big claim to fame
2:28:15 was 100,000 token context window
2:28:18 basically you could put in really long
2:28:20 prompts well chat GPT just exceeded that
2:28:22 with the new version Bard and Bing are
2:28:24 both connected to the internet so
2:28:26 screenshot that here's the Kids Table um
2:28:30 Pi perplexity llama 2 and po um PO is
2:28:35 sitting in the corner right now looking
2:28:36 its wounds cuz its Bots were starting to
2:28:39 get some traction and then open AI came
2:28:41 out with gpts and gpts are a thousand
2:28:45 times more powerful because you can
2:28:47 upload knowledge to them and they've got
2:28:50 API connectivity it's all it it's all
2:28:52 really janky right now but 6 months it's
2:28:55 going to be amazing so pose kind of in
2:28:57 trouble Pi still really interesting
2:28:59 perplexity is still really
2:29:01 interesting um what about creating a
2:29:03 local instance EG llu and Lang chain
2:29:06 yeah absolutely do that like if you if
2:29:08 you want to do if you want your stuff to
2:29:10 be super private do local instances of
2:29:13 Open Source large language models you
2:29:16 could basically replicate what open AI
2:29:18 just released with Lang chain and and
2:29:20 some you know open source things okay
2:29:22 image creation Dolly 3 inside chat
2:29:25 GPT I like I rarely use mid Journey
2:29:29 anymore but if you want super high
2:29:31 quality mid Journey if you want
2:29:33 incredibly good quality
2:29:36 and in context of your conversation
2:29:39 Dolly 3 Leonardo is like the cockpit of
2:29:42 a 747 but it's more simple than uh
2:29:46 installing stable diffusion on your own
2:29:48 and then Firefly from odobi is good if
2:29:50 you work in a corporate
2:29:52 office um video creation the three
2:29:56 things on the left are all in the same
2:29:58 neighborhood Runway ml Moon Valley and
2:30:00 pabs Runway ml just announced check this
2:30:03 out
2:30:06 um the thing they just announced looks
2:30:11 insane uh Runway
2:30:15 ml
2:30:23 this
2:30:28 [Music]
2:30:45 video
2:30:49 [Music]
2:31:05 so motion brush is coming
2:31:09 from oh okay wait the the survey the
2:31:15 theer direction of toilet
2:31:19 paper go look at the patent for toilet
2:31:22 paper there's your
2:31:25 answer all right so anyway so Runway ml
2:31:28 um the the three on the left are all
2:31:30 kind of the same but this motion brush
2:31:32 from one Runway ml looks amazing um and
2:31:35 then heyen the reason it's yellow go try
2:31:38 go to labs. gen.com and upload a video
2:31:41 of yourself speaking whatever your
2:31:43 native language is and pick one of their
2:31:45 other languages and watch watch what
2:31:47 that thing [ __ ] does so that's worth
2:31:49 playing with um music and audio 11 Labs
2:31:53 you can create a model of your own voice
2:31:55 and then just type in text and it will
2:31:57 be you saying it it won't quite sound
2:31:59 like you but it's close and then
2:32:01 refusion and sound draw are both music
2:32:04 generation tools screenshot
2:32:06 those 3D worlds and objects blockade
2:32:09 Labs Wonder Dynamics and Luma AI all
2:32:12 three of those really
2:32:14 interesting these three or these four
2:32:16 pie people I'm I'm passionate about you
2:32:22 following just because they're good
2:32:24 Ethan mullik Wharton business school
2:32:27 Professor one useful thing is his blog
2:32:29 follow his Twitter follow his LinkedIn
2:32:31 follow that blog David Shapiro is
2:32:33 working on AGI and open- Source stuff he
2:32:36 also does two or three videos a week on
2:32:38 the future of what AI is going to do to
2:32:41 our world go watch his
2:32:45 [ __ ] warning
2:32:47 it will make you feel
2:32:51 stupid he's he his gray matter is more
2:32:54 than your or mine and if you took our
2:32:57 brains and smashed them together I think
2:32:59 he's still got some on us he's neuros
2:33:02 spicy too and he he declares it he calls
2:33:04 himself neuros spicy um Paul rater at
2:33:06 the marketing I AI Institute Rachel
2:33:08 Woods here on Tik Tok at the AI exchange
2:33:12 uh is good
2:33:15 um the AI salon and oh the YouTube
2:33:19 channel for this so I put up the
2:33:21 archives of these lives at the YouTube
2:33:25 AI learninglab
2:33:27 dtt I think that's it we have a scammer
2:33:30 if you follow AI learning lab and
2:33:32 there's two L's in front of learning
2:33:34 that's not me they just stole my content
2:33:37 and they're trying to scam you and sell
2:33:39 you crypto
2:33:44 cryptocurrency all right f fantastic my
2:33:47 professors in grad school encourage it
2:33:49 as long as we don't rely on it and
2:33:51 plagiarize I think that's great Jamie I
2:33:53 mean ideally what the professors would
2:33:55 do would be bring it into the classroom
2:33:57 and start um having critical thinking
2:34:01 conversations around AI what's it good
2:34:03 at what's it not good at um where would
2:34:05 you use it where would you not use it
2:34:08 where where might it augment something
2:34:09 where is it more of a collaborative
2:34:11 partner um where can you absolutely you
2:34:13 know rely on it
2:34:15 1,00% um that kind of critical thinking
2:34:19 I think is is something we need way more
2:34:21 of in higher ed and I think um chat gbt
2:34:27 all of these things are going to force
2:34:29 that
2:34:29 conversation check out playday PL AI day
2:34:34 for video creation you can Infuse your
2:34:36 face into the videos oh super cool all
2:34:39 right let me go look at
2:34:44 that play
2:34:56 Day text to
2:35:03 video yet another
2:35:07 one you can also go to Future
2:35:15 pedia.com so as if what I just gave you
2:35:17 wasn't enough rabbit holes um are
2:35:20 children being taught AI in
2:35:22 school I it's it's very much uh school
2:35:26 by school Lauren and and from what I
2:35:29 understand it's very much teacher by
2:35:31 teacher I think that the the
2:35:34 institutions themselves are terrified of
2:35:37 it so they're not dealing with it so I
2:35:39 think you've got some teachers like
2:35:41 there's a uh Gibson is the name is a
2:35:44 tick tocker she's like a 12th great
2:35:46 English teacher on Tik Tok here um she's
2:35:49 great she she dove into chat GPT back in
2:35:53 December or January um of this year and
2:35:58 she's teaching her kids critical
2:35:59 thinking she's she's a genius with how
2:36:02 she's approaching this but it's very
2:36:03 much uh teacher by teacher I think right
2:36:05 now unfortunately it's it it it boggles
2:36:10 my mind if if they don't want my opinion
2:36:13 but if they did I would give them
2:36:17 that opinion get off your [ __ ] ass
2:36:21 learn this [ __ ] start hanging out in AI
2:36:24 learning lab with a guy that's got a
2:36:27 [ __ ] acting
2:36:29 degree and a subpar
2:36:32 [Laughter]
2:36:36 GPA I'll tell them where it's
2:36:40 at you really like that video The Smoke
2:36:42 water fire have all been done before not
2:36:45 with that level of control beep bop boop
2:36:47 so so one of the problems with the the
2:36:50 motion controls with Runway right now is
2:36:52 it's you're basic it's B it's basically
2:36:55 figuring out what in the image should
2:36:58 move and what shouldn't what you can do
2:37:01 now like I saw one example of a whole
2:37:03 bunch of eyeballs in in circles and they
2:37:06 just made the motion the eyeballs
2:37:08 themselves and the eyeballs all moved
2:37:11 you know together so that level of
2:37:13 control is a big deal
2:37:17 what would happen if I used only the
2:37:18 text from the Quran to train an AI
2:37:22 learning
2:37:23 model it would have all of that point of
2:37:28 view in
2:37:29 it it's
2:37:31 like
2:37:33 I I I mean quite honestly you can go
2:37:36 make you go do that you can go do that
2:37:38 right now go make a GPT right now ai
2:37:42 uzzy puppy oh a a fuzzy puppy nice name
2:37:45 I like the name I I would go make that
2:37:47 go make it and see what it
2:37:49 does don't like like and if that's I I
2:37:51 don't know if you're doing it to be Pro
2:37:53 provocative or like if there's some
2:37:56 agenda there but like I would think that
2:37:58 would be a really interesting experiment
2:38:01 why don't you go upload the Quran and
2:38:04 you know five
2:38:06 other um biblical type
2:38:10 books into a model and and compare them
2:38:14 see where they're the same see where
2:38:15 they're different
2:38:16 tired Runner David Shapiro is so so good
2:38:18 his take on the news is like yeah it is
2:38:21 it's David Who David Shapiro yeah he's
2:38:25 his his I think he's Dave shap on
2:38:28 YouTube now at Dave
2:38:32 shap night guys see you tomorrow see you
2:38:34 Dr J um can you include your tables
2:38:38 which are free for commercial use can
2:38:40 you include I don't know what that meant
2:38:42 could AI be described as super fast
2:38:45 crowdsourcing
2:38:48 yeah kind of I think that's kind of what
2:38:50 it is I think that's kind of what it
2:38:55 is like I like I describe it as a
2:38:59 reflection of humanity and and when you
2:39:01 interact with it it kind of feels like
2:39:03 you're interacting with like the most
2:39:06 educated person in the world which is
2:39:08 essentially crowdsourcing I think that's
2:39:09 actually a really
2:39:11 good that's a really good way to think
2:39:13 about
2:39:14 it do the rest of you test and iterate
2:39:18 your gpts all day test iterate test
2:39:20 iterate I don't do it all day cuz I'm
2:39:22 running a company but um but a lot David
2:39:25 Shapiro not to be confused with Ben
2:39:27 Shapiro please David Shapiro David
2:39:33 Shapiro
2:39:38 yeah I do not know what's
2:39:42 happening what times are you in Tik Tok
2:39:44 I generally start Wendy wi I don't know
2:39:46 if anyone can hear me I generally start
2:39:49 around 8:00 p.m. mountain time you're
2:39:52 amazing when I grow up on just like you
2:39:54 thank you Debbie that's very
2:40:01 sweet oh this is did everyone go away no
2:40:05 there's still a bunch of people here
2:40:09 weird a reflection of human language yep
2:40:12 that's right so
2:40:14 funny um
2:40:16 your screenshot tables can you share
2:40:20 which are free for commercial
2:40:23 use almost all of them in almost all of
2:40:27 the large language models have some free
2:40:30 option almost none of the video models
2:40:34 have free options the video stuff's
2:40:36 really
2:40:37 expensive the audio stuff's really
2:40:40 expensive like the voice synthesis stuff
2:40:42 and the like H Jen and did that do the
2:40:46 voice synthesis stuff like they're not
2:40:48 riing expensive but if you want to do
2:40:50 anything of scale like a lot of times
2:40:53 what I've thought about is oh I could
2:40:54 have chat GT write all these scripts and
2:40:56 go make all this content well not unless
2:40:59 you pay for their like unlimited plan
2:41:01 which is really
2:41:04 expensive all right I got to go to bed
2:41:06 people it's
2:41:09 11:42 all right let's let's let's get
2:41:12 let's get let's get let's get flipping
2:41:14 over let's get flipping out let's get
2:41:16 flipping over we're going to flip it on
2:41:17 over we're going to flip it on over
2:41:20 we're going to flip it we got a there
2:41:21 you go the AI learning lab
2:41:24 Afterparty sh champ we're going to go
2:41:26 out see here pretty soon here you darn
2:41:29 dog all
2:41:31 right um 1042 dragonfly Alchemy good to
2:41:36 see you I haven't seen you in a bit um
2:41:37 Kyle I hope you're doing wait I hope you
2:41:39 keep doing these you really have become
2:41:41 part of my daily life johy thank you so
2:41:44 much and I appreciate it I appreciate
2:41:45 you
2:41:46 Christopher Ramsey free
2:41:48 bird Bob Hamilton good night I don't
2:41:51 know we we'll have to we're going to
2:41:53 quickly tap out the amount of songs I
2:41:55 know so it's going to get old quick
2:41:57 blockade Labs used to be free for
2:41:59 unlimited R renders yeah everything used
2:42:02 to be free while they were trying to get
2:42:04 customers and now that they've got them
2:42:05 they're like oh [ __ ] this is
2:42:07 expensive um Tobias good to see you man
2:42:10 johy thank you jrc Debbie Jules good to
2:42:14 see you Janet Jackson good to see you
2:42:16 haven't been here in a bit beep B Boop
2:42:18 love the name Nancy
2:42:21 t-rock uh Wendy wise affiliate marketing
2:42:23 good to see
2:42:25 you Clint Copperhead calling the people
2:42:27 to
2:42:29 order calling the people to
2:42:32 order
2:42:34 uh thank you so much
2:42:37 everybody 135 people in the channel I
2:42:40 know so many people hanging out um happy
2:42:42 Monday night is it Monday yeah it's
2:42:45 Monday yeah I got a I got a busy day
2:42:47 tomorrow
2:42:49 um thank you all thank you moderators
2:42:52 for doing what you do thank you for
2:42:55 anyone who gave gifts I appreciate that
2:42:57 if you would follow the channel I'd
2:42:58 appreciate that I just cracked 30,000
2:43:00 which was my goal before the
2:43:02 30th um so so that's a pile of people
2:43:05 that's a pile of people so I appreciate
2:43:07 you all all right peace out everybody
2:43:10 I'll see you
2:43:10 tomorrow