aiLL 06/20/23 - Secret Cyborgs, AGI Funding, and Reading 320 Screenplays

Description

Chapters

Transcript

0:01 foreign
0:20 with the AI learning lab welcome
0:23 everybody
0:24 hope you're doing well it's what day is
0:27 today Tuesday is it Tuesday Tuesday the
0:29 20th
0:30 10 17 P.M
0:33 Do You Know Where Your Children Are
0:36 you have to be Gen X to remember that
0:39 that was a TV commercial
0:41 when parents were so negligent they had
0:43 to run TV commercials to ask the parents
0:46 if their kids
0:48 hahaha
0:50 we're in the house
0:51 [Laughter]
0:54 anyway this is the AI learning lab uh
0:57 other than talking about Gen X issues uh
0:59 we like to talk about AI things here
1:02 um I just got back from an AI Salon we
1:05 had here in Denver it was really good we
1:07 had a guy named Christian Robbins who's
1:10 the chief creative officer of a
1:12 post-production house talking about how
1:15 he used generative AI in the work that
1:17 he's doing and how it's transforming uh
1:19 how he does his work so that was pretty
1:21 fascinating
1:23 um I am an entrepreneur and I am not an
1:27 AI expert but I have an AI Channel why
1:31 is that
1:32 uh I've started I don't know 12 or 15
1:35 companies in my day and
1:38 um I see some really
1:40 dramatic parallels between the early
1:42 days of the World Wide Web which I was a
1:44 part of way back in the day
1:46 in the mid 90s and what's going on right
1:49 now with generative AI so I kind of
1:50 started this channel as a way to talk
1:52 about it demystify it share what I'm
1:54 learning help other people learn wow
1:56 looks like a whole bunch of people just
1:57 showed up my name is Kyle Shannon this
1:59 is the AI learning lab these URLs behind
2:02 my head here
2:03 um
2:04 the they're they're really just here
2:06 more as a public service announcement
2:08 than anything check that openai.com if
2:10 you've not played with chat jpt that is
2:13 the official chat GPT website
2:15 if you haven't tried it go there right
2:18 now well you hear me ramble go there
2:21 chat.openai.com and start playing
2:23 um ask it to write you a poem ask it to
2:25 write you a song ask it to write you a
2:27 meal plan and then turn that meal plan
2:29 into a grocery list uh try that just do
2:33 anything with it and see what it does
2:35 and and when you pick your jaw up off
2:38 the floor then then we can talk come
2:40 back here if you have questions about AI
2:42 pop them in the comments below hey there
2:44 uh I can't read because I'm wearing a
2:47 bad shirt for this but someone said hi
2:49 hello
2:50 couldn't read your name
2:52 um bing.com that's Microsoft's uh search
2:55 engine but if you click on the chat
2:56 button
2:57 that takes you to gpt4 which is chat GPT
3:01 also and it's connected to the internet
3:02 and it's free and then poe.com gives you
3:06 access to six different large language
3:07 models and a bunch of chat bots of all
3:10 different flavors it even lets you
3:11 create your own box there these other
3:14 two things prompts.chat will teach you
3:16 about prompting futurepedia.io is a
3:20 directory I think there's almost 4 000
3:22 AI tools there so if you want to run
3:24 down rabbit holes go there but I have
3:27 that there when people ask me really
3:28 specific questions about tools I just
3:30 point them there so that's what all that
3:32 stuff is
3:33 um I'm the co-founder and CEO of a
3:36 company called storyvine which it's an
3:38 automated video storytelling platform
3:39 I'm happy to chat about that but anyway
3:42 uh pop some questions into the comments
3:44 below I'll do my best to answer them
3:46 like I said I just got back from a salon
3:48 so I'm a little I'm a little crispy I'm
3:50 a little tired but I figured I'd get on
3:51 here say hey to you folks and let's get
3:54 in here and talk AI I'll have a lovely
3:56 hydration beverage
3:59 [Music]
4:00 um
4:01 ah 32 ounces of vodka before bed
4:05 works every time like a charm all right
4:10 let's say hello hi Kyle hello there
4:14 he lives I know I was I had a salon
4:16 tonight I was busy I was busy I was out
4:19 with the people if you know what I mean
4:21 how do we know you're not AI I could be
4:24 gpt5 it's very possible
4:28 that I am a simulation it's very
4:30 possible we're all in a simulation
4:34 have you taken a look at the Schumann
4:36 resonance lately tell me we're not in
4:39 her simulation tell me
4:41 what could possibly do that other than
4:44 Russians you know sending signals to the
4:47 sensors other than that nothing
4:52 the Earth is ascending to a new plane we
4:54 are ascending to a new plane fantastic
4:56 Bob tell him what he's won ah we lost
4:59 everyone my my my little uh joke down
5:03 there on the uh Schumann resonance uh
5:06 that that's some people packing so so
5:08 anyway if you have questions about AI
5:10 pop them down below I'll do my best to
5:12 answer them I hope everyone's doing well
5:13 tonight
5:14 all right let's see
5:17 is there a free AI that can draw a model
5:21 up
5:23 I don't know what you mean
5:24 I want to make a dry garden cage oh draw
5:28 a model up like
5:30 um
5:31 like CAD
5:33 um I don't know of any AI specific to
5:36 CAD you could go to something like mid
5:38 journey and have it render
5:40 um a design for
5:42 um for a garden cage but if you're
5:45 actually looking for you know drawings
5:47 renderings of like like plans I don't
5:50 know of an AI that'll do that but
5:52 go to futurepedia.io and it's possible
5:55 that there's something there that does
5:56 that because that's the kind of thing
5:59 putting a large language model sitting
6:00 on top of a um
6:03 like a a complicated software like CAD
6:06 software that's going to be increasingly
6:08 what how we see these large language
6:10 models use so I think that's coming so
6:12 it's very possible that futurepedia.io
6:14 will will have something there what are
6:16 the legal implications of AI oh they are
6:18 many they are deep it is completely
6:21 unclear I assume you're asking will I
6:25 get sued if I use it or can I copyright
6:27 if I use it uh not clear
6:31 not right now
6:34 um if you're talking about the image
6:36 Generation stuff and you're talking
6:38 about using generative AI for images in
6:41 a business I would strongly encourage
6:44 you to use Adobe Firefly or generative
6:48 AI within Photoshop because Adobe has
6:50 trained their image generation model on
6:54 a public domain and uh Adobe stock
6:58 images that they they own all the rights
7:00 to so I would I would uh do that when it
7:03 comes to the word stuff less clear I
7:05 wouldn't worry too much about it right
7:06 now unless um hang on I gotta sneeze I'm
7:10 gonna sneeze
7:11 uh
7:14 I was trying to say the word yes and I
7:17 sneezed at the same time so I went yes
7:19 chew
7:20 that's pretty good uh anyway the legal
7:23 stuff's uh just completely unclear right
7:25 now so unless you're doing something
7:27 where you're like you really want to
7:28 copyright something or you know there's
7:30 really lots of legal risk I just play
7:32 with it it's very the next 20 years are
7:35 going to be nothing but lawsuits uh in
7:37 both directions so
7:39 uh because I'm not worrying too much
7:41 about it right now I'm trying to find
7:42 grant funding prompts do you know where
7:44 to find them
7:46 um I don't I don't know where to find
7:48 them specifically but but
7:50 one of the things that
7:52 um you might want to try is take the
7:57 um the grant request
7:59 language from the grant request and then
8:01 language about your your non-profit and
8:05 put them both in chat gbt say here's a
8:07 grant request pop it in there here's
8:10 information about my company you know
8:12 can you give me an outline or can you
8:14 give me a
8:15 um you know an opening paragraph for
8:18 each of the sections of the Grant and
8:20 and see how it starts doing so you can
8:21 start doing it like that
8:23 um I know that there are new AI startups
8:27 focused specifically on grants
8:30 um I've heard of them and again go to
8:33 futurepedia.io there's a bunch of tools
8:35 there so you might be able to find
8:36 something there quite honestly what I
8:38 would do is encourage you to just go to
8:40 chat.openi.com start exploring uh start
8:43 learning the tool yourself and you'll
8:45 you'll probably be able to figure it out
8:48 because it's it's incredibly powerful
8:51 and if you if you get good at asking it
8:53 for things it gets good at giving you
8:55 answers including that
8:57 that's you thank you
9:00 that's a great idea we'll do are there
9:02 any optimization up opportunities
9:06 are there optimization opportunities in
9:10 GPT algorithms
9:12 um I am not a data scientist I'm not
9:14 quite sure what you mean
9:16 um you can take long pieces of data and
9:19 tell it to compress it so you can
9:21 optimize the the core data of a prompt
9:24 and then you've got more
9:27 um
9:28 tokens in the context window to play
9:30 with but I don't I assume that's not
9:31 what you're talking about if what you're
9:33 talking about is
9:36 um can they make it smaller and more
9:38 efficient yes go look at what's
9:40 happening in the open source Community
9:42 where they've taken the Llama large
9:45 language models and they're training
9:47 them and I don't know quantizing them
9:50 they're doing a bunch of stuff to
9:51 optimize them and make the models
9:55 increasingly smaller but still maintain
9:57 a high level of quality so that's
10:00 happening right now more on the
10:03 open source side than on the open AI
10:06 side
10:07 um I would imagine that
10:10 open AI right now is prioritizing
10:17 uh getting to AGI artificial general
10:21 intelligence they're prioritizing that
10:23 first and then they'll get to optimizing
10:25 it second because they have Microsoft
10:27 sitting there going take all the compute
10:29 power you want here's 10 billion dollars
10:32 now Sam Altman says he thinks they need
10:34 a hundred billion dollars to get to AGI
10:36 so they're going to be using they're
10:38 going to be using some more uh GPU
10:41 cycles and the old Azure cloud services
10:43 ecosystem
10:47 um
10:49 all right llama llm can you give some
10:54 idea uh I don't know what you mean can I
10:56 give some idea
10:58 about what about which of the the Llama
11:00 llms the open source llms I can't
11:04 because I stopped paying attention to it
11:05 because every time I thought there was a
11:08 really cool open source large language
11:09 model a new one would come out like two
11:11 days later
11:13 um the last one I paid attention to was
11:15 one called MPT 7B
11:19 um which was an open source model that
11:21 had a 65 000 token
11:25 um context window which is 45 000 words
11:28 48 000 words something like that a lot
11:31 uh and then two days later anthropic
11:34 came out with Claude uh 100K so
11:37 anthropic came out with their version of
11:40 um
11:41 a hundred thousand token model and
11:43 that's at poe.com you can actually go
11:44 play with that if you if you subscribe
11:46 to poe.com I think it's 20 bucks a month
11:49 you can play with their hundred thousand
11:51 token model
11:53 um
11:53 I would look at a project called GPT for
11:56 all where you can install
11:58 um
11:59 basically uh uh you know
12:02 like a large language model container on
12:05 your local machine and then use
12:07 different large language models uh but I
12:09 haven't been following them enough
12:10 because they're just they're just moving
12:11 too fast I can't I can't keep well I'm
12:14 too tired I'm tired and I can't keep up
12:18 the open source Community is moving too
12:21 fast for me and I move fast I'm Zippy
12:24 I'm Zippy but I can't keep up so I don't
12:28 know I don't know what to tell you there
12:29 I use Chet gbt today
12:32 for helping with vlookup Excel formula
12:35 saved me 20 minutes of trial and error
12:38 yeah doing doing uh Excel formulas which
12:42 like you know what you when okay you're
12:45 in Excel you're you've got to figure out
12:47 some formula like I don't remember this
12:49 stupid [ __ ] formula and you go to
12:50 Google and you're like how do I find
12:52 this formula in like you know 20 minutes
12:54 later you're still trying to find the
12:56 formula on Google and you're down all
12:57 these SEO rabbit holes because that's a
13:01 nightmare and then you just go to Chachi
13:03 PT and they're like I want this cell to
13:04 do that to that cell it's like here's
13:06 your formula it's amazing so that's cool
13:10 glad that you did that
13:11 I made a story up with chat GPT I put in
13:15 a ton of details
13:17 is it a hundred percent mine can I
13:19 monetize it yes you can they're not
13:21 going to come after you for that
13:23 um whether or not it's actually
13:25 copyrightable is different but can you
13:27 sell it yes but I would go listen I
13:30 ain't a lawyer I have a degree in acting
13:31 so don't believe me for [ __ ] I lie as
13:35 much as Chachi BT does that's not true I
13:37 don't hallucinate that much
13:40 um I would I would go you know go look
13:42 at their terms of service but I'm pretty
13:44 sure that that you've got the rights to
13:45 do whatever the [ __ ] you want with it
13:47 now can you actually copyright it that's
13:49 between you and the copyright office and
13:51 you have to prove that there was human
13:53 contribution involved significant enough
13:56 that you cross whatever their [ __ ]
13:58 arbitrary hurdle is this week
14:00 um like I said this this uh intellectual
14:02 property stuff's going to be a bit of a
14:04 disaster for some time
14:06 but but yeah go put it out there like
14:09 there are so many GPT books out out in
14:12 the world right now that
14:14 ain't no one coming after you in my
14:17 opinion not my legal opinion Poe has a
14:20 decent app too yeah they do they have a
14:22 pretty good app thanks you're good I
14:24 appreciate the kind words uh
14:27 all right uh there's just a handful of
14:29 us in here hope everyone's doing well if
14:32 you have any uh questions about AI pop
14:34 them down below and I will do my best to
14:36 answer them trying to think what we
14:39 talked about today
14:40 um at the salon at the AI Salon here let
14:43 me show you
14:44 so the salon that I'm talking about you
14:47 can find information at it about the
14:50 salon here the salon.ai there's a link
14:52 tree there it's got our Meetup we meet
14:54 every other week and then we have an
14:56 online Discord server where you can geek
14:58 out and ask questions and play with mid
15:00 journey and then if you go to Twitter
15:02 this is a newsletter that I put together
15:04 with a guy named Greg mushin and we do
15:07 threads about news articles and we have
15:10 a newsletter and we we just go nuts we
15:13 go nuts so anyway but that's that's some
15:16 of my stuff and
15:19 um yeah I'm paying attention a lot to
15:21 this here
15:22 AI stuff there's Emilio's wife best [ __ ]
15:25 best shirt yet for reading text Mr
15:28 Shannon I know this is like one of those
15:30 Rorschach tests where you're just like I
15:32 can't do it I think it hurts my eyes I
15:36 know but what are you gonna do
15:38 I trade Futures I have chat GPT help you
15:42 figure out some some analysis with the
15:44 Futures toss a little data in there and
15:46 have it have it give you some thoughts
15:48 on it it's pretty good at that stuff
15:52 it analyzes data real good uh how did
15:56 you know my vocation Mr Shannon I never
15:58 said that what did I say what did I say
16:00 your vocation was I forget what I said
16:02 I'm tired I was I was doing a salon
16:04 tonight so I did two hours there then I
16:07 went out and grabbed a drink with with
16:09 Salon people and then I came home and
16:11 now I'm on here so I'm like my brain is
16:14 uh what do they call it mush
16:16 that's a technical term
16:20 um all right so what else what else is
16:23 worth talking about there were there
16:24 were
16:25 one of the things we talked about at the
16:27 salon there were there were three
16:28 different news articles that that I saw
16:31 kind of in close uh succession
16:34 um that I thought were pretty
16:36 interesting one of them was a post by
16:38 Ethan Malik award Wharton business
16:40 school professor
16:42 talking about this concept of
16:45 um
16:46 secret cyborgs and the secret cyborgs
16:49 what he basically says is generative AI
16:52 tools are really good for individuals to
16:54 dramatically increase their productivity
16:56 So within companies there are lots of
16:58 individuals using AI to get way more
17:01 productive
17:03 but the organization itself because it
17:06 doesn't have data policies and it hasn't
17:08 quite adopted them yet and it might not
17:11 even have a position on if people can
17:13 use it the the organization itself is
17:15 not using AI so you have these secret
17:17 secret cyborgs within the organization
17:20 these employees that are getting AI
17:22 literate
17:24 but they're not talking about it to
17:26 upper management because they don't want
17:28 to be fired they don't want to admit
17:29 that they can do in four hours what used
17:32 to take them 20 hours or 40 hours
17:35 um and so there's this big disconnect so
17:37 that article came out and then there was
17:39 an article in Fortune magazine that
17:41 talked about
17:43 um ai's here to stay and CEOs know what
17:47 to do they they need to come up with AI
17:49 data policies and they need to make sure
17:52 that their entire stabs are AI literate
17:54 so it's basic this is Fortune Magazine
17:56 right really mainstream business uh
17:59 magazine saying this stuff's here to
18:01 stay get with it and then I saw another
18:03 article about the uh though the White
18:06 House is is prioritizing
18:09 um
18:10 generating significant AI policy to make
18:14 sure that we're all safe and and we
18:16 ensure that we are inventing and then
18:19 innovating and all that sort of [ __ ] but
18:21 still being safe so so those things kind
18:24 of all add up to me to say that as as
18:27 over hyped as this AI stuff is it's
18:29 starting to get normalized into the
18:31 business conversation
18:33 um it's starting to get normalized as
18:35 this is a foregone conclusion that this
18:37 is going to be everywhere so that's part
18:39 of the purpose of this channel is just
18:40 to talk about it as if it's coming
18:43 because it is in fact it's already here
18:49 look to your left look to your right one
18:52 of those people is really good at AI one
18:55 of those people is ignoring it
18:57 you know don't ignore it
19:00 all right
19:01 are there any Salon people here too I
19:04 don't know
19:05 possibly any Salon people in here your
19:08 brain is mush
19:10 you're never Zippy
19:14 um did you see Congressman Jeff
19:15 Jackson's Tick Tock today about AI I did
19:18 not what did it well I know I didn't I
19:21 was at work today especially if those
19:23 employees are working remotely yes
19:26 they they may not want to Emilia's wife
19:30 I can't read anything because of my
19:32 shirt why didn't you tell me I shouldn't
19:34 wear a shirt like this oh you tell me
19:36 that every live
19:39 they may not want to say
19:42 um they now have free time yeah that's
19:43 exactly right that's what that's what
19:44 that article talks about the Ethan
19:46 Malik's article talks about is that is
19:49 that these secret cyborgs they're like
19:51 I'm not telling the boss that I'm
19:53 getting this [ __ ] done faster now I mean
19:55 the the the kind of sad and ironic thing
19:58 is
20:00 as long as people are stingy with their
20:03 knowledge of what they're learning like
20:05 the entire organization could benefit
20:07 from that what Malik says in that
20:09 article is
20:10 the bosses need to basically make it a
20:13 safe zone for employees to say hey I
20:16 want you exploring AI I want you sharing
20:18 what you're learning I'm not gonna fire
20:20 you there's going to be no bad
20:22 repercussions if you're learning this
20:24 stuff
20:25 so you know
20:27 yeah like you know get with it bossy
20:30 bosses
20:32 Mike you in the live
20:34 all right as a national speaker as well
20:37 as one who funds for films where do I
20:40 start
20:42 um wait as one who oh finds funds for
20:45 films where do you start with AI so the
20:48 reason that I have these three sites
20:50 here you can ignore the bottom too
20:51 really I would start if you're brand new
20:54 to generative AI to chat GPT if you've
20:57 not played with chat gbt that's what
20:59 that URL is chat.openai.com where I
21:03 would start is go there and start
21:05 playing
21:07 um and by playing
21:09 just ask it to do stuff you find fun
21:11 funds for films
21:14 um go ask it about
21:16 um films that you care about ask it
21:17 about the plots or the characters or
21:19 have it compare I don't know
21:22 um
21:23 uh It's a Wonderful Life with Blade
21:27 Runner and and compare the major themes
21:29 between them and create a table for you
21:31 comparing the major plot points and
21:34 themes of those two films
21:36 um
21:37 what else did you say you do you're a
21:39 national speaker
21:40 um
21:41 um have it help you put together
21:43 outlines for Keynotes or take an outline
21:47 for a one-hour keynote and turn it into
21:48 a 20-minute speech or a two-hour
21:50 workshop and just watch what it does so
21:54 just start playing
21:55 the other thing you can do is this URL
21:58 here prompts.chat that's a document that
22:01 it's called the awesome chatgpt prompts
22:04 document the top half of that document
22:06 document talks to you about how you
22:09 prompt chat GPT the bottom most of that
22:14 document is all these individual prompts
22:17 where you tell Chachi PT to act like a
22:19 certain Persona act like a screenwriter
22:21 act like a mathematician act like a
22:24 psychologist act like a you know a
22:26 non-profit fundraiser or a film
22:28 fundraiser and it will give you a sense
22:31 of
22:32 the depth of what's possible with that
22:35 single tool the the other these other
22:37 tools here are just variations on the
22:38 theme so if I were you I would start
22:40 with chat.openai.com just go start
22:43 playing
22:46 if you hear other people talk about this
22:48 [ __ ] it's way scarier and way less
22:52 impressive it's scarier because we hear
22:55 oh the robots are gonna kill us right we
22:57 hear the Hollywood [ __ ] we hear we hear
23:01 um you know they're going to take out
23:03 humanity and and you know whatever all
23:05 that all that [ __ ]
23:07 um which seems really scary and
23:09 intimidating and then
23:11 when you hear someone talk about it
23:13 they're like well there's a text box and
23:14 you put some text in it and it gives
23:16 squirts some text out and you're like
23:18 well that's [ __ ] Google so is this
23:21 like Google no it's not
23:25 it's like that's what Google wants to be
23:27 it's actually not because Google doesn't
23:30 want to actually give you answers what
23:32 Google wants you to do is click on the
23:33 40 [ __ ] links they put up there
23:35 because every time you click they make
23:37 money
23:38 that's what Google wants
23:40 that's not what chat GPT is it is it is
23:43 an incredibly capable think of it almost
23:45 like this you have a free
23:47 personal assistant or not not a personal
23:49 assistant you have a free intern
23:52 that works beside you and it's the
23:54 smartest intern you've ever worked with
23:56 in the history of interns this is the
23:58 smartest one
24:00 and start interacting with catch EPT
24:02 like it's an intern
24:04 hey I just got in from a long day I you
24:07 know I haven't booked the speaking gig
24:09 and I don't know whatever a month
24:13 um I have the summer off you know can
24:16 you help me come up with ideas for how
24:17 to book more speaking gigs out of the
24:19 cup
24:20 just start interacting with it like it's
24:22 an intern a really capable one and then
24:25 just see what see what happens
24:27 um that's how I'd do it so anyway all
24:31 right listen I am exhausted I hope this
24:34 helped I was just here for a half hour
24:36 so I'm normally here for like two hours
24:37 but I had a big long Salon tonight
24:42 um let's see last one Becky I viewed 322
24:46 slide decks from Michael Bay I wondered
24:49 if they used AI to produce the slides
24:51 well
24:53 possibly
24:55 um if they were submitted in the last uh
24:59 six months probably
25:03 um what I would say is this what's good
25:05 for the goose is good for the gander
25:08 um not only can they write that [ __ ]
25:09 with AI you can
25:12 um read that [ __ ] with AI take the text
25:15 of those you know turn those things into
25:17 PDFs here I'll show you how to do
25:19 something quick let me let me show you
25:21 something quick here
25:22 now you're gonna have to to do this
25:24 you're gonna have to pay for the
25:26 um
25:27 you're gonna have to pay for the
25:30 plus version of chat GPT
25:33 but
25:34 let's see we're gonna do it on chat gbt
25:37 well yeah
25:38 let's see turn that off there
25:43 um
25:43 in
25:45 in chat GPT if you pay for chat GPT plus
25:48 you get access to regular chat GPT but
25:51 you also got access get access to this
25:53 thing called plugins
25:56 and if you go to plugins the way they
25:57 work is you you can install all these
26:00 different plugins there's a plug-in
26:01 store where you can go find them and
26:03 then in any given session you can
26:05 activate up to three of these and one of
26:07 them is called ask your PDF
26:10 so you could make PDFs of those pitch
26:14 decks and then you
26:17 um you know get a URL like put it on
26:19 Google Drive or something like that get
26:21 a URL of that pitch deck and then you
26:23 can let me go find a PDF here here's one
26:27 so this is a this is a white paper this
26:30 is scientific white paper so I'm going
26:32 to take this URL
26:33 I'm going to go over here I've got ask
26:35 your PDF plugin enable and I'm going to
26:38 say okay based on this and then I'm
26:41 going to paste in that URL
26:45 um
26:46 please
26:48 give me a
26:50 10 bullet summary
26:56 of the paper
26:59 and it will now fire off this plug-in
27:03 using ask your PDF it's basically going
27:06 out looking at it it's pulling in a
27:09 summary
27:10 it's now I don't know I think off
27:13 reading it or something like that yeah
27:15 it just got back a whole bunch of
27:16 content
27:17 and now it's writing me a 10 bullet
27:19 point
27:21 um
27:22 overview summary
27:24 of this paper
27:27 um I can stop generating that I could
27:29 say
27:30 um
27:32 uh turn that summary into a Holley
27:39 wood log line
27:42 [Music]
27:45 in a world where complex visual based
27:48 tasks challenge AI assist gbt emerges as
27:52 the multimodal AI assistant integrating
27:54 Advanced tools and pioneering a new
27:56 approach to reasoning with its ability
27:59 to plan execute inspect and learn it
28:02 outperforms benchmarks and tackles real
28:04 world applications setting a new
28:06 standard in the realm of AI technology
28:08 okay it's not the best log line in the
28:09 world but it understands [ __ ] like that
28:12 so you can take uh you know things
28:15 things like those those 200 what'd you
28:18 say 320 pitch decks for Michael Bay and
28:22 automatically knock out a bunch of these
28:24 the other thing you can play with one of
28:26 the other sites I have up there is this
28:28 this site called po Poe
28:31 um and so Poe one of the models that
28:34 it's got so so it's got these different
28:35 models you can play with here one of the
28:37 models it's got if you subscribe to it
28:39 is called Claude instant 100k
28:43 and do I still have these in here yeah
28:45 so I paste it in so what's what all this
28:48 text is here I pasted in the entire
28:50 script of uh 2001 Space Odyssey so
28:55 that's the entire screenplay and then I
28:57 asked it what are the two main
28:59 characters and what happens to them Dave
29:01 Bowman gives you the bullet points of
29:03 what happened to Dave how he had a
29:05 little run-in with how
29:07 and Frank Poole uh and then I say what
29:10 are some interesting lines that that Hal
29:12 says to Dave and it pulls out some
29:14 quotes from the screenplay so I'm
29:16 literally copying and pasting an entire
29:18 screenplay in and then I can I can pull
29:20 out uh details from that screenplay so
29:27 you know if if they're writing [ __ ]
29:30 screenplays using chat GPT you can read
29:32 their [ __ ] screenplays with chat GPT
29:35 [Laughter]
29:40 fun here at the AI learning lab bitter
29:44 bitter listen I've written seven
29:45 screenplays in my life
29:47 um I've optioned two of them none of
29:49 them have been made so I know that world
29:51 well
29:52 and uh I have 80d so the fact that you
29:55 can read the fact that you could even
29:57 consider reading 100 320 outlines of
30:01 Michael Bay potential films is is that's
30:04 miraculous so I'm gonna have to pay for
30:07 it now yes yeah so so I would say either
30:10 pay for
30:11 um chat gbt with plugins and then use
30:14 the PDF plugins or pay for Poe and do
30:17 the Claude instant 100K I personally
30:22 um which one would I go with well if
30:24 you're gonna be if you want to start
30:26 interacting with complete screenplays
30:29 um
30:29 probably Poe but
30:32 hmm
30:34 I just think I think the the I think
30:36 chat GPT is just a better tool it's just
30:38 a better tool so I would go chat gbt but
30:40 if you can swing 40 bucks a month pay
30:42 for both that's what I'm doing right now
30:44 I can't really afford 40 bucks a month
30:46 but it's you know I I have access to the
30:51 most powerful you know writing
30:53 technology in the history of mankind so
30:55 it's kind of worth 40 bucks a month I
30:58 pay a lot more for that than for you
31:00 know my shitty car
31:05 all right what are your top three AI for
31:08 creators video and audio I don't know
31:10 the video and audio world well
31:13 um
31:14 video I mean you should definitely be if
31:17 you're not playing with Runway gen 1 and
31:19 Gen 2 from what Runway ML on the video
31:22 side on the audio side I don't know
31:25 there's there's a kind of an AI driven
31:28 Loop composition tool called sound draw
31:31 which is okay but it's it's Loops it's
31:35 not generating uh original stuff and
31:38 then I really don't know I would just go
31:39 to futurepedia.io look in the music
31:42 section and just start running down the
31:45 [ __ ] rabbit holes one after another
31:47 and until you find something cool I mean
31:50 like I would sort them by
31:52 user ratings because that site now has
31:54 like 4 000 AI
31:57 applications in it it's it's completely
32:00 intimidating it was daunting the numbers
32:03 had me cross-eyed yeah I'm sure Becky
32:05 yeah yeah what like I I'm serious like
32:08 the the ability for you to very very
32:11 quickly
32:12 you know sort through one of those pitch
32:15 decks if they're pitch decks or if
32:17 they're sending you actual screenplays
32:18 or even if they're just outlines the
32:21 ability to quickly summarize them and
32:23 give a couple of bullet points of you
32:26 could even ask it [ __ ] like
32:28 um what genre is the screenplay
32:32 um what's its closest you know
32:34 contemporary of of produced film like
32:37 you could you could probably very
32:39 quickly come up with a well you probably
32:41 already have a report that you know Mr
32:44 Bay expects so you could probably just
32:47 create a a pretty pres pretty specific
32:49 prompt for chat GPT toss an outline in
32:53 there hit squirt and out will come a
32:56 report for that and then if you've got
32:58 320 of them have chat EPT generate all
33:00 those reports take the 10 most
33:02 interesting ones and actually read those
33:04 like that might be an interesting way to
33:05 attack that
33:09 um act as a tick tock live streamer and
33:11 teach me about AI with humanistic
33:13 elements yeah that's good
33:18 you can make money
33:23 um yeah I'm my brain is useless tonight
33:26 I got nothing I got nothing here I got a
33:28 little screenplay [ __ ] because I know
33:29 that world
33:31 all right listen I'm out of here I hope
33:34 you're all wait I'm in the position of
33:36 having Mr Bay's Camp asking my guys for
33:39 the money oh that's interesting
33:42 huh well
33:44 you still you still gotta shift through
33:46 a bunch of
33:48 turds to get to you know
33:51 an interesting script or two don't you
33:53 right
33:55 because they're not giving you 320
33:56 scripts because they want to make 320
33:58 movies they're giving you 320 scripts
34:00 because they're hoping that somewhere in
34:02 that pile of crap are five that don't
34:04 suck and two that actually have a chance
34:07 of getting funded and one maybe just
34:09 maybe
34:10 gets funded and gets made right
34:12 I mean isn't that the numbers like even
34:15 if even if they're coming from base
34:16 people to you
34:19 I could be wrong I haven't been in the
34:21 business and 25 years but
34:25 that's a rough business
34:27 but hey man if you got the money
34:28 you wanna you wanna fund a tick tock
34:31 Channel
34:35 we got ideas we could be making some
34:37 things
34:43 all right Coolio well I hope this helps
34:45 uh yes subscribe or join or whatever
34:47 follow whatever the hell you do on Tick
34:48 Tock
34:49 um I try to go live once a night um like
34:51 I said I did a salon tonight so I'm a
34:53 little crispy so I'm gonna cut it a bit
34:55 short because I'm tired and uh hope
34:57 you're all well have a good night peace
34:59 out everybody
35:01 we'll talk soon see you tomorrow