aiLL 5/8/23 (#1) - AI Writing Tools Amidst The WGA Strike

Description

Chapters

Transcript

0:00 foreign
0:19 with the AI learning lab I was uh live
0:23 before but then it looked like I froze
0:26 because my user account just stopped
0:28 going up and people said it looks like
0:30 he froze
0:31 so I was talking about cabbage dad that
0:34 little picture there this little picture
0:36 uh this is the AI learning lab if you
0:39 are curious about AI if you're wondering
0:41 what the heck's going on with AI if you
0:42 have questions about AI I do my best to
0:44 answer questions there I also make
0:46 videos like pre-recorded things during
0:49 the day but I just started going live
0:50 about a week ago so I'm trying to figure
0:53 this out I turned 58 yesterday so if I
0:57 suck at this
0:58 that's why because I'm just an old guy
1:00 I'm a gen xer I ain't no Boomer I know
1:03 how to use technology
1:06 but I ain't no [ __ ] Millennial I'm
1:09 neither We Are The Forgotten bitter
1:12 class
1:13 bitter you betcha all right Kyle Shannon
1:16 yeah learning lab if you have questions
1:18 about AI jump in ask me questions I hope
1:20 you're all doing well tonight
1:23 um you can ask me anything let's see uh
1:26 Ron just
1:28 commented somewhere
1:30 that I'm living in amazing times these
1:33 are amazing times they're absolutely
1:35 amazing times this is the what's wrong
1:37 with what Millennials Millennials are
1:39 fine that what what happens is the the
1:42 older people
1:44 bitched about Millennials not working
1:46 like they worked and then the
1:48 Millennials called them Boomers and said
1:50 oh oh Boomer or sorry Boomer whatever
1:53 the hell the Boomer thing was
1:55 and and boomers are more like you know
1:58 World War II kids and and gen xers are
2:02 sort of in the in the in the between of
2:05 the Boomers and the Millennials and no
2:07 one talked about them so for years it
2:09 was Millennials and Boomers warring and
2:10 at some point someone goes wait a minute
2:12 what about Gen X why didn't they get
2:13 picked on so you know thought you were a
2:16 cam I'm not a cam I'm real well I could
2:18 be gpt5 it's just that you know is would
2:23 this really be the face you synthesize
2:25 if you're gonna make this a pre-recorded
2:27 thing what's wrong with Millennials
2:30 sorry about that media pipe did I say
2:32 media pipe I might have said media pipe
2:34 what platform can generate realistic
2:36 pictures
2:38 like that for a social uh technology
2:42 called
2:43 um dream booth and dream Booth sorry I'm
2:46 trying to change the focus did it go
2:49 there it goes
2:50 um dream Booth allows you to train
2:54 stable diffusion on your image so I
2:57 uploaded like 20 or 30 pictures and then
3:00 just started telling it to make images
3:02 and uh they're all sort of me but not me
3:06 which gave me this idea to
3:09 create this thing called Kyle Shannon
3:11 dream so if you go to Instagram Kyle
3:13 Shannon dreams you can see some of these
3:16 dudes that dude's particularly inspiring
3:18 I'll leave him up for a while
3:21 uh but that that was uh stable diffusion
3:24 you can also now with mid Journey you
3:26 could upload an image and and modify it
3:28 so I think you can do this this kind of
3:31 one you know one image will will
3:33 generate another image in a different
3:35 style so there's lots of different
3:37 things you can do
3:38 but these were all that unstable
3:40 diffusion all right oh okay Boomer that
3:43 was the thing someone just put something
3:44 in there
3:46 um all right what is the best AI
3:49 resource to generate video shorts and
3:51 voice
3:52 so it depends on the video shorts you
3:54 want to do if you want to do video to
3:58 video there's a there's a tech or a
4:01 company called runwayml and they've got
4:03 gen 1 and Gen 2 so gen 1 you can take a
4:07 video and uh like a video of yourself
4:10 walking down the street and then
4:12 transform that into like a claymation
4:14 dude walking down the street that's gen
4:16 one and they just released that as an
4:18 app and then so runwayml gen 1 I think
4:22 is an app I think it's the runway ml app
4:25 that you use and then Gen 2 is text to
4:29 video I think you have to still do that
4:31 on their Discord server if you want to
4:33 do things like
4:35 um
4:36 synthesized voice where you take a still
4:38 image like you know you can take a still
4:41 image like this old dude
4:43 and
4:44 um animate him to say something excuse
4:47 me say something
4:48 that there's a site called did D Dash
4:54 id.com
5:00 and it looks like this let me log in
5:02 here for a second
5:05 eh
5:11 log in with my Google single sign-on
5:18 all right so there's a bunch of videos
5:20 that I've made so if you say create
5:21 video
5:23 um what you do here is you pick one of
5:26 these images like that image there is an
5:28 image that Leah fasten who's a
5:29 photographer she's the co-host with me
5:31 for the AI Salon which by the way the
5:33 next AI salon
5:35 is tomorrow night if you're interested
5:37 in joining the salon go to a link in my
5:39 bio there's a link tree there so there's
5:41 a link to the salon it's just a
5:43 community of people geeking out about AI
5:45 so anyway she made me that so that's one
5:47 I've also got this image that I uploaded
5:49 that's that's uh the cyborg me that's
5:52 kind of the Genesis image of of uh Kyle
5:56 Shannon dreams and then they have all
5:57 these other people that are just their
5:58 their pre-built characters
6:01 so one way you can do this is you can
6:05 pick a synthesized voice and you can
6:08 type something in
6:09 so I know it's hard to see this stuff
6:12 but at the bottom here are the different
6:13 voices so the first thing is what
6:15 language do you wanna
6:17 do so you can do you know different
6:20 languages or you can do different
6:21 accents
6:23 um and then you can
6:26 um let's see I guess I have to pick a
6:27 language first hold please
6:32 we'll do
6:34 Australian English and then there's a
6:36 net
6:37 Carly Elise we'll do it dude we'll do
6:40 Duncan and then we'll let's go to
6:46 um
6:49 we're gonna go to
6:51 chat GPT
6:55 if you've never been to chat GPT it's at
6:59 um
7:00 chat.openai.com
7:02 and we'll say
7:04 um right
7:06 a short monologue
7:10 uh from and out of work actor
7:15 about
7:16 eating mac and cheese
7:22 out of work actor leadings against the
7:24 kitchen counter holding a back of a box
7:26 of mac and cheese ah mac and cheese the
7:29 ultimate comfort food all right so I'll
7:31 just grab that little paragraph there
7:33 so we'll go back to dash ID and we'll
7:37 paste
7:39 um that little monologue into the window
7:42 up there and then you say generate video
7:45 in this upper right hand corner
7:47 and it's going to make that dude say
7:50 that monologue in Italian or in us with
7:53 an Australian accent
7:55 and so there's all the other videos that
7:58 I've made I'll play one of the other
7:59 ones there
8:00 but it'll do it pretty quickly it's like
8:02 four or five seconds and then it will
8:04 pop up a little window and we'll say
8:06 generate
8:09 all right 10 seconds this is one of
8:11 those things oh it's going to be so cool
8:13 wait till you see it and then you know
8:15 the person's just sitting there mac and
8:16 cheese the ultimate comfort food for our
8:18 struggling actors
8:21 think about this humble dish that fills
8:23 the emptiness within me every time I go
8:25 without an audition or Worse face
8:27 another rejection
8:29 there's a certain Solace a warmth that
8:31 emanates from every gooey golden fork
8:33 for as if it's whispering to me you'll
8:35 get through this
8:40 I had to be serious and evil are these
8:43 the Rob robots that are gonna kill us
8:45 all I don't think so they might
8:50 but anyway that's how you make those
8:51 things so if you've ever wondered how to
8:53 make those things that's D Dash ID
8:56 um
8:57 all right what are the best AI resources
8:59 how to generate videos I did that
9:02 um
9:03 DTD I thought you were a cam okay Boomer
9:06 I'm not a boomer jet extra
9:08 Gen X is rad man we're so cool people
9:12 forgot to talk about us
9:14 do you use media pipe I do not what do
9:17 you smoke in a media pipe
9:20 oh that's a boomer joke
9:22 all right uh feel free to ask any
9:25 questions about
9:27 uh AI I'll do my best to answer it
9:32 that's what this channel is all about
9:35 all right
9:37 um
9:38 best platform to produce short videos
9:40 and video I think I just showed that
9:43 Opus is good for clipping videos yeah
9:45 Opus is good for clipping videos there's
9:47 another one called video v i d y o that
9:50 you can upload a long form video it'll
9:52 generate a bunch of clips
9:54 there's also dscript that does video
9:59 um so that's good
10:02 [Music]
10:03 oh Joyce joyce.vine shared the live
10:07 encourage more peop viewers to share
10:09 sure feel free to share this I think tap
10:12 on the screen that's what uh someone
10:14 said the other day that knows more about
10:16 Tic Tac than I did if you tap on the
10:18 screen I get lots of love Bubbles and
10:20 I'm human who doesn't want a love Bubble
10:23 all right what do I think Hey Ron what
10:25 do I think about the writer's union
10:26 strike uh I'm I'm a screenwriter I've
10:29 written seven screenplays so I empathize
10:31 with them the
10:34 um
10:37 their concerns about chat GPT are
10:40 warranted what they're basically saying
10:42 is they want they want the studios to
10:44 agree not to use
10:46 chat apt to replace them right which
10:48 which I understand and at some point
10:50 some
10:51 it's gonna be film Executives gonna say
10:53 oh yeah I can I can just have chat like
10:56 I always wanted to be a writer and but I
10:59 suck and I'm not going to admit that so
11:01 I'll just use it and produce my own
11:02 movie so I think that's going to happen
11:04 at some point um I don't think that's
11:06 going to be the norm I also think
11:08 that this idea that chat GPT is kind of
11:12 this separate thing from writing is
11:14 going to disappear relatively quickly
11:15 like it's happening in universities
11:17 right now and and higher uh you know
11:20 high school and things like that even
11:21 Junior High where you know there's kind
11:25 of this battle of is you know is chat
11:27 TPT cheating and we're going to have gbt
11:29 detectors and all that stuff the deal is
11:32 this these large language models like
11:34 chat EPT are going to be everywhere
11:36 they're going to be ubiquitous we're in
11:37 a really finite window right now where
11:39 chat GPT feels like it's this thing off
11:41 to the side and then the rest of the
11:44 world is over here it's going to do this
11:45 really really quick
11:47 Microsoft's going to roll it into office
11:49 365. Google's going to roll it into
11:51 Google Docs I would be dramatically
11:53 surprised if every screenwriting
11:54 software didn't rule GPT like generation
11:57 tools within it the writing process when
12:00 you use Chachi BT and when you do
12:02 generative images and things like that
12:04 is much more collaborative Than People
12:05 realize it's not like write me a
12:08 screenplay and out comes you know Star
12:11 Wars
12:12 um it's very iterative it's like I have
12:15 an idea here or like I one example I did
12:18 I took an incomplete screenplay I had I
12:20 copied and pasted in the first two
12:22 scenes of the screenplay the and and the
12:25 bulk of the outline but it was an
12:26 incomplete outline it was an ADD
12:28 like it was just an aborted attempt to
12:30 the screenplay
12:32 um
12:33 and so I just started sort of co-writing
12:36 with chat EBT where I said
12:38 um finish the outline for me and it
12:40 finished the outline and it wasn't bad
12:41 and then I said rewrite the outline as a
12:44 comedy and it did that and then I said
12:47 um
12:48 write scene three for me so I had
12:50 written scenes one and scenes two and it
12:52 wrote Scene Three and it kind of sucked
12:53 pretty bad and then I said no write
12:55 Scene Three where this character does
12:58 this and that character does that and
13:00 then it got better and then I told it
13:01 explicitly write it in the style of
13:04 scenes one and scenes two and then it
13:06 got better but it was it was a lot of
13:09 back and forth so I think we're going to
13:10 see is writers are going to start
13:12 incorporating these tools into their
13:14 existing process so I think long term
13:18 you know that I understand the sentiment
13:20 of that particular part of the strike
13:23 um I just don't think that's how things
13:25 are ultimately going to play out um so I
13:27 don't know I don't know I like it's a
13:29 bummer I hate strikes and but I
13:30 understand why they're there and I think
13:32 they're I think it's important that
13:33 people get paid but I also think we're
13:36 entering a an era of infinite content
13:41 creation instantly and I think the
13:46 economic Foundation of nearly everything
13:49 is going to transform pretty
13:50 dramatically is AI generated art legal
13:54 to use if you use things like so so it's
13:57 a very very gray area whether or not you
14:00 can copyright it the copyright office
14:02 has put out guidance that says you can
14:03 sort of copyright it but not really
14:05 where we think sort of but not sure
14:10 um some other government agencies put
14:12 out some really vague stuff so so it's
14:14 it's very very unclear right now
14:17 um things like mid-journing and stable
14:18 diffusion were absolutely trained on you
14:21 know lots and lots of images some of
14:23 which are copyrighted
14:26 um one of the misconceptions about the
14:28 way those tools work is is that people
14:30 think that they're copying and pasting
14:32 elements of copywritten work and pasting
14:34 that together like doing sort of
14:36 montages or collages that's not how they
14:39 work it's it really is this kind of
14:42 deeper mathematical thing every image
14:43 that's generated is a true
14:46 um original generated piece of art
14:49 um I happen to think it's art because
14:51 it's intentional someone typed in a
14:53 prompt to get it and they're generating
14:55 it but if you used another artist's name
14:58 to get a particular style then it gets a
15:00 little more gray and wonky
15:03 um so if you're using things like
15:04 mid-journey stable diffusion a lot of
15:07 corporate policy right now is you can't
15:08 use those like I know some people at
15:10 Deloitte digital
15:12 um they're not allowed to use AI
15:14 generated stuff within on their work
15:17 computers
15:18 um the the the one creative director
15:22 that I know um actually generates images
15:25 on his personal machine on the weekends
15:28 to use in concept boards during the week
15:30 but
15:32 um a good buddy of mine um is pretty
15:35 high up over there and he met with the
15:37 president of adobe last week and adobe's
15:40 got this thing called Firefly that was
15:42 trained on open source and Adobe stock
15:45 images which they own all the rights to
15:47 so Adobe is trying to make it safe to
15:50 use AI art commercially and so I think
15:53 right now in its beta stage you can't
15:55 you can't use it commercially it Flags
15:57 everything with with Adobe Firefly but
16:00 they're going to make it safe and
16:01 there's going to be lots of deals that
16:03 are cut that are going to make it safe
16:05 but right now it's a little it's a
16:06 little weird so if you work for a
16:09 conservative corporate organization I
16:11 would say just follow Adobe Firefly if
16:14 you work for a little startup I'd say
16:15 just use use it I wouldn't worry too
16:18 much about copywriting any anything you
16:20 generate right now because it's just
16:21 it's just all over the place and and
16:23 like one safety thing would be rather
16:26 than using a specific artist name try to
16:28 prompt the AI to generate something like
16:31 that without using their name that might
16:33 be a way to do it
16:36 um feel free to ask me questions about
16:37 AI you can subscribe to my lives
16:41 um by I don't know how you do that
16:44 but I did turn it on Tick Tock told me I
16:46 could turn on that feature and I did so
16:48 if you want to subscribe to these I
16:49 think you get notified when I go live
16:50 which why wouldn't you want to spend
16:53 time with me sitting in my home office
16:56 waxing poetic about crap
17:00 um let's get another
17:02 image up here hold
17:06 today
17:09 let's see what's fun oh that one's
17:11 disturbing
17:12 there we go
17:14 that'll confuse people when they come in
17:16 the hell is this guy doing what's with
17:18 the eye you weirdo is he Gen X
17:22 all right it's like cell phones the
17:24 cameras are so good everybody thinks
17:26 they're a photographer it's slightly
17:27 different than that
17:29 I mean it is that and there's kind of a
17:32 sameness to cell photography where
17:33 everyone uses the the the
17:36 the filters they're going to be people
17:38 that are really good at generative Ai
17:40 and people that are not really good at
17:42 generative AI but generative AI is
17:45 so good that if you have an idea for an
17:49 image
17:50 um and and you're kind of intentionally
17:51 trying to do something you can get there
17:53 with this stuff so I think it's it I
17:56 think it's of a slightly
17:58 I think it's slightly more elevated than
18:00 everyone has a camera in their pocket
18:02 if it could be fast wait it could be
18:05 that fast though it just takes training
18:08 I'm not quite sure if that's in
18:09 reference to I was thinking about
18:10 offering
18:12 these writers a free chat GPT coaching
18:14 service I mean do that like like if if
18:17 anyone is a writer on here that is not
18:20 actively
18:22 diving deep into learning chat EBT I
18:25 think you're doing yourself a disservice
18:26 I think writers
18:28 at least short term are going to be
18:30 better at wrangling results out of these
18:33 generative AI models than non-writers so
18:36 if you're not a writer diving into this
18:38 you should do it so Ron if you know
18:39 people that are are in that you know
18:41 screw that I'm never going to use one of
18:43 them damn robots you know it can't write
18:46 like a human
18:47 um get them to start writing with this
18:49 stuff and using it as a collaborator and
18:52 and their tune will likely change
18:54 um because it's it's it's not like
18:57 the thing about these generative large
19:00 language models is it's not like
19:03 it's not like a Google search where
19:05 you're just saying write me a screenplay
19:07 and it comes out like you're really
19:08 doing this back and forth I know I said
19:10 that before but like
19:12 as a writer like I find it incredibly
19:15 empowering to to collaborate with these
19:17 tools
19:19 um after a few prompts it got very nice
19:22 story lines in one morning yeah yeah if
19:24 you like seriously like like if if you
19:27 do two or three back and forth with chat
19:28 GPT that's good if you do two or three
19:30 hours going back and forth with it you
19:32 can get to some remarkable stuff
19:36 um we need to learn the prompts yeah
19:38 there are lots of prompts I've I've been
19:39 sharing this on my lives because I think
19:42 this is a big mystery if you go to
19:44 prompts.chat
19:46 prompts sorry prompts
19:50 dot chat
19:53 that will take you to awesome chat GPT
19:55 prompts the first half of this document
19:58 kind of explains to you about prompting
20:00 and how they work and it gives you some
20:02 resources and ebooks and training and
20:05 then the the whole rest of this document
20:06 are all of these
20:09 um prompts that that essentially give
20:11 context to chat gbt to act like a
20:14 certain kind of partner right so this
20:17 one says act as an Advertiser this one
20:19 says act as a Storyteller this one says
20:22 act as a football commentator act as a
20:24 stand-up comedian it's interestingly not
20:26 all that great at writing jokes uh but
20:29 gpt4 is better than GPT 3.5 was as you
20:33 would hope
20:35 um act as a composer act as a debater
20:37 act as a screenwriter act as a therapist
20:39 act as a act you know all these
20:41 different things and
20:44 it like it felt counter-intuitive to me
20:47 that you had to kind of tell it who to
20:50 be
20:51 but but the more you understand about
20:53 how these models actually work when you
20:55 tell it who to be you're kind of
20:57 limiting the semantic
20:59 cloud of words
21:02 um so if you tell it to be Shakespeare
21:04 it will it will only answer in iambic
21:08 pentameter with big fancy words like
21:10 Shakespeare so it kind of puts the model
21:13 into a particular semantic space where
21:16 your interactions with it you know all
21:18 live within that universe so if you
21:19 wanted to if you want to ask it
21:21 questions about quantum mechanics tell
21:23 it to act as you know Stephen Hawking
21:26 you know and and you know whatever other
21:29 you know physicist that you respect and
21:31 it will get better and better and better
21:34 um other things you can do
21:36 um one of the one of the recent hacks
21:38 with prompting prompting which I think
21:40 is is really interesting is um reflect
21:43 on this so you can have it give you an
21:45 answer and then you say reflect on your
21:48 answer you know critique yourself and
21:50 tell me what you don't like about it
21:52 what how could this be better and it
21:54 will tell you how it would be better
21:55 then you're like well make it better
21:57 like that and it will
21:59 um you can also just tell it what you
22:00 gave me is crap and here's why it's crap
22:02 and I want it to be better and it will
22:04 make it better all right um seems like
22:07 the answer to losing jobs is something
22:10 that would look like Socialism or com
22:13 communism well I I mean if if you just
22:17 go if if you just go to a reduction
22:19 reductionistic view of
22:22 how things are going to play out like I
22:23 think one of the one of the fears is
22:25 that the robots are going to take our
22:26 jobs and then humans just sit around
22:28 going well damn that sucks what am I
22:30 gonna do now
22:32 we're humans we're gonna figure
22:34 something out and the other thing that's
22:36 going to happen is these tools are so
22:39 powerful and transformative that the
22:41 initial disruption in jobs is going to
22:43 shift into oh here's a whole bunch of
22:45 other jobs
22:48 um I think if you just sit around and
22:49 wait for this [ __ ] to knock you off your
22:53 seat it will and then if you just sit
22:56 there and do nothing then it's gonna
22:57 suck and I think that there's going to
22:59 be some percentage of people that
23:00 happens to
23:02 and it's gonna suck really bad for a
23:04 while and I think that's where you know
23:06 for the first time I understand why
23:08 Andrew Yang was talking about Universal
23:10 basic income which I assume which is
23:12 what you're bitching about like you know
23:13 people should own their own damn money
23:15 earn their own damn money
23:17 um
23:18 I think some of the disruption is going
23:20 to come really quickly and really
23:21 swiftly in certain industries it's just
23:23 going to decimate them short term so I
23:25 actually understand how that could be
23:27 abridging mechanism to a new future
23:31 where new jobs are created
23:34 um but it's going to be painful and it's
23:35 again I I think we're way more
23:37 complicated than just reducing
23:39 everything to socialism sucks and you
23:42 know everyone should lift themselves up
23:43 by the boot heels we're more complicated
23:45 than that more nuanced in fact it's one
23:48 of the things that's really interesting
23:49 about Chach EBT is they just did a study
23:52 where they had patients interact
23:55 virtually with real doctors and with
23:57 chat GPT as doctors
23:59 patience
24:01 universally preferred chat GPT not only
24:04 because they thought the answers were
24:06 better but they were more empathetic
24:08 chatgpt when when you ask it really
24:10 polarizing political questions actually
24:12 gives you nuanced answers which we're
24:15 not in a society that does anything
24:17 resembling Nuance when it comes to
24:19 politics so um I actually think that
24:21 these tools might be able to guide us to
24:24 suck less as humans because right now we
24:27 suck pretty bad
24:29 am I bitter yeah I am
24:33 tonight's open source swallowing the
24:36 market stuff
24:37 in the no moat leaked Google oh thoughts
24:41 on open source swallowing the markets
24:43 yeah so I'll tell you that in a second
24:45 hang on I need more water I'm a little
24:48 thirsty
24:50 um
24:54 Okay so
24:56 there was a leaked memo
24:59 let's see Google
25:02 Google
25:04 look up Google
25:07 um no moat
25:09 like Castle moat
25:11 Google we have no moat and it takes you
25:14 to a
25:15 semi-analysis.com site and Google
25:18 locally on our phone meaning you don't
25:21 have to have an internet connection you
25:22 could be in the middle of the Wilderness
25:24 and look up you know emergency safety
25:26 information right on your phone
25:28 um so I think what it means for us is
25:30 that large language models are going to
25:32 become a ubiquitous they're going to
25:33 become as ubiquitous as the internet
25:36 just fabric of the internet will have
25:38 large language models embedded in it
25:40 that's kind of how I see it playing out
25:45 I'm thinking about a way of forming an
25:48 AI music band to play out live I think
25:51 that's actually kind of brilliant I
25:53 think that um if you go to sound draw
25:57 um dot IO I think it is
26:00 um that's a tool that is it's kind of
26:02 like Loop music like GarageBand is is
26:05 kind of the the genre of music it
26:07 creates but it it creates music very
26:10 very quickly and then there's there's a
26:12 bunch of other ones so I would think if
26:14 you
26:16 um s-o-u-n
26:19 no that's not it it's is it hang on
26:22 sorry about this
26:24 foreign
26:26 's lives you can watch him
26:29 try to type and talk at the same time
26:31 it's hilarious
26:36 soundtrack.io
26:37 yeah Sandra dot IO
26:40 um this is one and you just say create
26:42 music and then you pick a genre and it
26:45 just
26:46 funny and weird fear euphoric hopeful
26:50 um select a genre acoustic beats Funk so
26:53 let's grab funk
26:57 did I click it I might have to sign in I
27:00 don't know it says creating music
27:05 and then you know now like here's all
27:07 these songs it's like it generated 20
27:09 songs like that
27:13 right so I can see using tools like this
27:14 to generate songs you could use
27:17 um
27:18 um chat EBT to generate lyrics and then
27:21 yeah learn them and go I I think one of
27:23 the things if you look at the the there
27:25 was a Drake and the weekend song where
27:27 they synthesized their voices and used
27:29 AI to con compose the music and then it
27:33 essentially became a hit on on um
27:37 on the on the streaming services and
27:39 then they they subsequently took them
27:40 down
27:41 um but I think you're going to start
27:43 we're entering a world where fan
27:45 generated art is going to be as good or
27:47 better than the original artists and
27:49 there's going to be they're gonna be all
27:51 sorts of lawsuits about it uh but I
27:53 think it's really really interesting and
27:55 um
27:56 um what's her name uh Grimes Elon musk's
27:59 ex-girlfriend baby mama maybe had babies
28:02 with him
28:03 um she just said synthesize my voice do
28:05 anything you want with my music
28:06 synthesize my voice my likeness uh just
28:08 give me half
28:10 so she's saying to her fan base you know
28:12 go figure this crap out and if you make
28:14 something interesting and it's a hit
28:16 um give it to me or you know give me
28:18 half which I think is a brilliant move
28:20 let's find another dude oh we'll do a
28:22 little Joker version of me that's
28:24 disturbing that'll be good for someone
28:25 that just stumbles upon this all right
28:29 um
28:30 if you've got questions just pop them in
28:33 here I'll do my best to answer them
28:36 um it popped me down to the bottom so
28:38 I'm kind of scrolling up trying to find
28:39 things here we go
28:42 um
28:45 [Music]
28:46 okay portrait rights and economic cycle
28:49 structure are things that must be
28:51 considered chat TBT yep I agree they
28:54 must be considered and everything's
28:57 going to change like like
29:00 economic rights and compensation and
29:02 attribution must be considered but if
29:05 anyone thinks that it's gonna it's gonna
29:07 be the way it is right now
29:09 um they're they're being very naive
29:11 everything's going to change
29:13 and not just in the creative arts and
29:15 rights and things like that how do you
29:17 make chat GPT to interact to tabs
29:20 open on Chrome for example
29:24 um
29:26 so there are some extensions that you
29:29 can get and use there's so many of them
29:31 I've completely lost track I stopped
29:33 trying to deal with them you can also
29:35 download the edge browser so in the edge
29:37 browser in the upper right hand corner
29:39 you can't see in the upper right corner
29:41 there but if you click on that it opens
29:43 a side drawer that's a chat window that
29:45 is aware of the window to the left my
29:48 experiences is that it's not always
29:50 aware of the window to the left but in
29:51 theory you can ask Chachi BT you know
29:55 Bing's chat gbt questions about the web
29:57 thing so there's there's some ways to do
29:59 that I just haven't found I don't have a
30:02 need to do that all the time so I don't
30:03 have good answers for that one
30:05 unfortunately but I would say check
30:07 extensions and download the edge browser
30:09 and see if that does it for you
30:12 um follow those I've lost you my other
30:14 tick tocks are working hmm I yeah you
30:17 know what stream just crashed this
30:20 happened before
30:21 um can you tap the screen if I'm still
30:22 live
30:26 um
30:29 yeah I'm not live all right