Sep 5, 2024 - ChatGPT 01 Preview: Is This the Future of Screenwriting? (You Decide)

Description

Chapters

Transcript

0:01 we're going in champ champ
0:03 [Music]
0:06 Shannon ready
0:12 [Music]
0:22 [Music]
0:58 w
1:00 [Music]
1:05 sitting in this lonely
1:07 Town wonder when things are going to
1:12 change Dre my
1:15 away sa dreams have turned to a bunch
1:18 DUS
1:21 glass get my nerve up but my past is
1:25 pulling me
1:26 [Music]
1:28 down wonder how long
1:31 [Music]
1:33 she
1:38 around somebody told me once before you
1:41 can never go home
1:43 again won't you leave say a thing just
1:47 to steer me away yeah from the truth of
1:51 who I am and what I believe so a thank
1:55 him for two cents with a handshake and
1:58 some sympathy
2:01 back up my
2:03 je H it for this big
2:07 price of my
2:10 freedom
2:12 byebye Black Sheep to the black sheep of
2:16 the
2:18 family
2:20 [Music]
2:22 by oh that mean so very much to
2:26 [Music]
2:28 me sheep to my friends and my
2:33 family
2:36 byebye going off set my
2:40 [Music]
2:44 soul set it
2:48 free
2:51 free times they were a
2:55 changing little bit of rearranging
2:59 [Music]
3:01 take a couple
3:04 chances my promise and not
3:06 advances to the
3:11 pr love my freedom all right
3:17 [Music]
3:31 good evening good people Friday night
3:33 date night
3:41 [Music]
3:48 [Music]
4:02 one day since the LA launch of
4:03 strawberry we're still here the robots
4:06 haven't killed us so I consider that a
4:10 [Laughter]
4:14 victory all the bank computers haven't
4:17 been hacked airplanes are still flying
4:20 we're okay Oprah is a little scared but
4:22 you
4:25 know you know what we missed from Oprah
4:27 last night any semblance of you can make
4:30 money with
4:32 J actually any semblance of Hope and
4:35 usefulness was was sorely missing last
4:38 night
4:40 [Music]
5:04 Freedom came my way that
5:06 [Music]
5:09 night just like a jet plane in and out
5:13 of
5:14 side I was hauling ass at a million
5:17 miles an hour wondering how hard I'd
5:21 [Music]
5:24 hit when they came into the station
5:28 [Music]
5:31 they said I was bad Beyond
5:34 repairs but I got no quals with my
5:40 situation see here I
5:48 am say Shere Shere Shere won't you D to
5:54 say Shere Shere Shere won't you dare to
5:59 say Shere Shere
6:03 Shere
6:05 aha yeah leave a message and your number
6:11 please take a time to want to satisfy
6:16 me take all these old fantasies and send
6:20 them care of me
6:22 [Music]
6:35 w
6:37 [Music]
6:45 [Music]
7:00 this one's for Vicki I got this song
7:02 stuck in her
7:04 head in a
7:06 Westerly
7:10 Direction this car is my train that's
7:14 the second line I've been driving I've
7:17 been
7:18 [Music]
7:19 wondering what it is I'm running from
7:25 again feel like an 80-year-old
7:28 man oh holding on to 29 thank you Lord
7:32 digital Gods it's very generous and up
7:34 ahead on that
7:36 [Music]
7:37 Horizon is a California
7:41 [Music]
7:48 line up ahead of truck's carrying a wide
7:52 load now it'll be stuck in my head again
7:55 when I wake up tomorrow fre five house
7:57 cut in half
8:00 cute little front door and two windows
8:03 my
8:03 Lord ain't sure whether to cry or should
8:08 laugh you see a broke a home up myself
8:12 once B I stumbled to that
8:16 door I read her note by the
8:19 [Music]
8:20 dawn said don't you come around
8:25 here
8:27 [Music]
8:28 anymore well laugh had
8:32 enough this freedom of the
8:35 rock never was good with
8:39 decisions least that's what I've been
8:42 [Music]
8:50 told all right what are we going to do
8:53 people everything's different
8:54 everything's the same nothing's changed
8:56 everything's
8:57 changed we're all going to lose our jobs
9:00 we're all going to keep our jobs we're
9:01 going to reinvent ourselves we're going
9:02 to reinvent our businesses everything's
9:04 going to be good everything's going to
9:05 be bad it's dangerous it's awful it's
9:08 fantastic it's all
9:10 [Music]
9:13 true and here's if you're new here my
9:15 name is Kyle Shannon this is the AI
9:17 learning lab I don't know if we do a lot
9:19 of learning around
9:20 [Music]
9:21 here but we do a lot of philosophizing
9:25 while we're playing with [ __ ]
9:31 [Music]
9:37 and My Philosophy goes like
9:42 this learn to play guitar medium well
9:46 learn to song to sing slightly less
9:51 [Music]
9:56 well learn seven or eight songs You Love
10:00 cuz that's the only [ __ ] you're going to
10:01 [Laughter]
10:03 [Music]
10:05 [Laughter]
10:07 play oh sure if you're if you're
10:10 neurotypical and have a typical
10:12 attention spans and you can teach
10:14 yourself 422 songs good for you [ __ ] you
10:21 [Music]
10:27 [Music]
10:34 um I kind of look at it like this with
10:36 this AI
10:39 stuff the the Oprah special last night
10:41 was really interesting it it did a
10:43 couple of things for me it reinforced
10:45 for me that we're way earlier than I
10:47 think a lot of us think we are and that
10:51 you know there's a fair amount of people
10:52 right now talking about ai's
10:55 overhyped um we've passed the point of
10:59 no return these models have plateaued
11:01 they're never going to get
11:04 better
11:06 um and if you've been on the planet for
11:08 more than a year you know that that's
11:10 not true this shit's just going to keep
11:12 getting better and better and better
11:16 right and there's there's lots of issues
11:19 right there's safety issues and ethics
11:21 issues and bias issues and Technology
11:24 issues and compensation issues and
11:26 attribution issues and copyright issues
11:28 and trademark issues
11:32 and then there's remarkable technology
11:34 that would not exist let's be crystal
11:37 clear about this this technology would
11:40 not
11:42 exist had these Frontier Model companies
11:45 not done a little nefarious web
11:50 scraping
11:52 so for the people who are
11:54 arguing these things were trained on
11:58 data that they shouldn't have been I
12:00 agree
12:02 and if they had asked permission
12:05 permission from one copyright
12:09 holder the answer would have been no and
12:12 if they had asked permission from a 100
12:14 100 copyright holders the answer would
12:16 have been no if they asked permission
12:19 from a million copyright holders the
12:21 answer would have been
12:23 no so we would not
12:26 have the tools we have but we
12:32 do and because of chat GPT and because
12:35 chat GPT got to 100 million users in 6
12:39 weeks if you didn't know that that's a
12:41 little stunning remarkableness
12:45 fastest adoption of software in history
12:48 100 million users in 6 weeks to put that
12:51 in perspective the worldwide web took
12:53 six years to get to 100 million
12:56 users now the worldwide web the
12:59 infrastructure wasn't there to just turn
13:02 [ __ ] on like it like it was for for AI
13:04 but still six years versus six
13:07 weeks so this stuff's not going away and
13:11 the tools as they are today are the
13:13 worst they're ever going to
13:15 be and so what this channel is about
13:18 what we do here what we do on Friday
13:19 night date nights and what we do five
13:21 nights a
13:22 [Music]
13:23 week is we just kind of hang
13:27 out and we just be be in the
13:30 conversation about Ai and we play with
13:33 it and we talk about how janky it is and
13:37 how inspiring it is I can share if Kyle
13:39 can bring me up yeah hop on up here I on
13:43 [Music]
13:53 China Day to
13:55 you Source Camp's here it's okay we can
13:58 start I on China I know you were waiting
13:59 for that hey I I just wanted to share
14:03 and it's so hard to keep to to type in
14:05 the little limitations of the chat so I
14:07 wanted to come up and say I was messing
14:10 with the new model and I'm I'm 56 I can
14:13 remember playing zorc on my Commodore 64
14:16 in the in the early 80s it was one of
14:18 the it was one of those games that had
14:20 that AI aspect to it almost but it was
14:23 it was one of the early early things and
14:26 it was part were amazing yeah yeah it
14:28 was part of that early 80s culture and
14:31 uh I really love the game and so I asked
14:34 the new model I said can you can you
14:35 recreate that classic zor game for me
14:39 and I'd like to run it and play it real
14:41 time and it did it created a very very
14:44 from memory anyway a similar experience
14:47 and the thing that and then I wondered
14:49 okay what code did you use to to build
14:51 this give me some information about how
14:53 you did this and it was like there was
14:55 no code written it was just the LL and
14:57 that that really dawned on me that the
15:00 llm is so powerful and so broad an
15:03 application that it can actually do a
15:04 zor game without actually writing code
15:07 but using the inherent properties of the
15:10 LL I know it's Bonkers and then think
15:13 about that with you know just use the
15:16 the chat GPT app in voice mode and you
15:19 have a a virtual DM
15:23 right anyway that's all I wanted to
15:25 share man oh that's awesome that's
15:26 awesome thanks for thanks for hopping up
15:28 and sharing that's great good absolutely
15:30 all right awesome that's really cool um
15:35 all right there we go yeah it yeah the
15:38 these things are crazy so so we'll talk
15:40 about the new model tonight a little
15:43 bit
15:45 um also it looks like Runway ml came out
15:48 with video to video today right right
15:50 before they're doing a new 48 hour um
15:52 competition video competition or or gen
15:56 48 I think it's 48 hours
16:01 [Music]
16:03 I haven't played much with video to
16:05 video just showed my parents notebook LM
16:08 having a conversation Minds blown times
16:13 four so they each had double k Kevin
16:16 mallister
16:17 moments yeah yeah the new notebook LM
16:21 thing people are losing their minds over
16:23 it because it really is it's the first
16:26 of that voice synthesis thing the
16:29 generates a standalone media property
16:32 that doesn't really need editing I mean
16:34 you could add like a music bumper to the
16:37 front of
16:38 it
16:40 and just launch it as a podcast
16:43 absolutely I haven't seen you in about
16:45 two months is chat GPT still better than
16:48 Claude well interesting so welcome back
16:50 first of all
16:53 um Claude
16:56 3.5 Sonet um which launched
17:00 6 weeks ago um is quite remarkable and a
17:06 lot of
17:07 people
17:10 uh like like it better than gp4 Omni the
17:15 the G the gp4 small o model um yesterday
17:19 chat GPT launched
17:23 01 not to be confused with o or not to
17:27 be confused with 4.0
17:31 [ __ ]
17:32 idiots um and the jury is still out it
17:36 what it looks like is that the the new
17:38 01 model there's a couple of things
17:40 about it your rate limited to 50
17:44 queries in a
17:46 week um when they launched it I think
17:49 people did what people do when new GPT
17:52 models launched they're like you know
17:54 write me a poem do it this do it that
17:56 they just did stupid [ __ ] with it and
17:57 burned up all their credits so
17:59 apparently open AI reset everyone's
18:02 credits after a little bit of Education
18:04 to say you know this is fancier like you
18:08 have limits but we're going to reset
18:09 them so in theory we're all
18:12 reset but knowing that you have
18:15 50 I think is actually a good thing
18:18 because it is uh it's going to it's
18:22 going to force us and I think we can
18:24 explore tonight what you would do in gp4
18:28 Omni versus what would do in regular
18:30 chat
18:31 GPT um but to answer your
18:34 question
18:36 um I would say at this
18:39 point the the the line between Claude
18:43 and chat
18:44 GPT is fairly thin and it's probably
18:47 more personal choice um if you're paying
18:52 for either of them you get some extra
18:53 features the one thing that chat GPT
18:56 still has that Claude doesn't and I
18:58 can't believe they don't they don't is
19:00 Claude is not connected to the internet
19:02 so you can't do searches you know you
19:05 can't do things on Modern news stories
19:06 for example you could go to chat GPT
19:09 tell it to look up some [ __ ] take that
19:11 content pasted into Claud and then do
19:13 things based on that
19:16 but um I'd say they're toss up the
19:19 Claude artifacts feature if you have not
19:23 played with it is mind-blowing it's
19:26 mind-blowing what did we do the other
19:28 night I took this was so cool I took
19:32 some handwritten lecture notes so I went
19:34 on the Google and I searched for
19:37 handwritten lecture notes and I found a
19:40 someone took a snapshot of their
19:42 notebook from
19:45 class I didn't read it I and I tried to
19:47 find one that was kind of scribbly
19:48 looking it was cursive and who [ __ ]
19:51 reads cursive anymore um and we put that
19:54 in chat GPT and chat GPT read it and
19:57 made a nice organized set of notes about
19:59 it and then I took those notes over to
20:02 Claude And I said to Claude turn these
20:06 lecture notes into an interactive
20:09 dashboard and it turned the lecture
20:12 notes into an Interactive Learning
20:14 dashboard where there were like drop-
20:15 down menus and three different tabs with
20:17 different content on it all organized by
20:20 the content so it's it's pretty
20:22 remarkable right now I think I think the
20:24 three well honest to
20:27 God Gemini
20:31 um is has got some pretty remarkable
20:35 components to it if you play with the
20:36 Gemini 1.5 Pro it's multimodal meaning
20:40 you can upload like an hour of video and
20:42 it just you can ask it questions about
20:44 the video
20:46 including um what's happening
20:49 visually because it's not only seeing
20:51 the you know transcribing the words it's
20:53 seeing the images in the video um so
20:56 that's pretty remarkable perplexity
20:59 um is really good for doing research not
21:02 perfect yet still a bit janky but really
21:05 good for research and doing articles and
21:07 things like that um Claude has got this
21:11 artifacts thing and then chat gpt's chat
21:14 GPT and they just launched their new 01
21:17 model which is this reasoning
21:20 version it's a it's really a Chain of
21:22 Thought
21:24 engine um we'll go take a look at um
21:27 David Shapiro's coming down hard hard on
21:29 open AI right now saying they don't have
21:32 [ __ ] that this is just Chain of Thought
21:35 you know programming and he's he he's I
21:38 think he's starting an open source
21:40 project called raspberry to to go after
21:43 open Ai and just say that what you did
21:46 is not that
21:49 special but we'll go play with it
21:51 tonight and see see what we we discover
21:54 if you had to pay for only one which
21:56 would you choose good God
21:59 so so a answer me this K kedo studio
22:05 um
22:07 what what's your use case is it
22:10 programming is it writing is it
22:12 marketing is it strategy is it
22:16 recipes like what's your what's your
22:23 uh yeah so I on China says number one
22:26 chat
22:27 GPT right
22:32 right number two Claude number three
22:35 grock grock also got
22:43 good novel's kind of writing okay so
22:47 here's what I would do um because so so
22:51 there's couple of
22:54 things all right let me let me show you
22:57 actually
23:01 ow my back's still out my back went out
23:03 oh God damn it my back went out last
23:05 Friday and it's still out [ __ ]
23:07 killing me so if it gets too bad I'll
23:10 I'll bail early tonight just fair
23:15 warning um all right let's
23:32 uh uh uh
23:35 uh okay
23:40 so okay okay novels writing things like
23:45 that
23:50 [Music]
23:53 wow so I don't know I don't know how
23:56 deep you've dug into llms but I'll just
24:00 I'll just assume you're relatively new
24:03 um a a really good thing to do if you're
24:06 if you're doing creative writing is
24:12 um have whatever model you're
24:15 using recommend a storytelling framework
24:19 like like talk to it about storytelling
24:20 Frameworks and have it recommend some
24:23 and have it describe what they are and
24:25 then choose a storytelling framework and
24:27 then give it you know components of of
24:29 the thing you want to write then have it
24:31 write a broad outline and then refine
24:34 that broad outline and then have it
24:35 write a detailed outline and then refine
24:37 that and then have it do detail you know
24:40 within the segments of the outline um
24:43 and then start doing your writing um you
24:46 might be farther along than that but you
24:48 could even even if you've got something
24:49 started you could have chat GPT or
24:52 Claude analyze it and give you like a
24:55 narrative analysis of it and then go in
24:58 and fix The Narrative analysis and then
25:00 modify things from there
25:03 also so far none of these things are
25:06 good at just vomiting out entire pieces
25:09 so what I would strongly suggest is take
25:12 small little sections of the story and
25:14 write them in small little chunks and
25:16 then sort of copy and paste them out
25:18 into an editing document um know that
25:22 all of it is going to require some
25:25 editing a lot of editing quite frankly
25:28 um but it will get you there much much
25:30 faster and I the the way I see these
25:32 things is they're dramatic confidence
25:34 boost boosters
25:36 but I think you now have three different
25:38 options within chat GPT and it might be
25:42 worth paying for chat GPT because of
25:44 these three
25:45 options but what I what I would also
25:48 suggest
25:49 is figure out a
25:52 good exercise to go from like outline to
25:56 writing do it in Claude and then do it
25:59 in the three modes of cat GPT I'm going
26:01 to show you and then figure out which
26:04 one you like best from a tone and
26:06 personality standpoint from a writing
26:08 standpoint
26:10 because these llms at this point are so
26:13 good and Claude and and open AI in
26:15 particular are so good that it's kind of
26:19 a a coin toss but they definitely have
26:22 personalities so if you find that you
26:24 like the style of claude's writing
26:26 better go to Claude but here's what I
26:29 want to show you so in the new chat
26:32 GPT you've got gp4 Omni so this is their
26:37 the this came out this summer this is
26:40 that multimodal more efficient version
26:43 of chat GPT you've got 01 preview which
26:47 is the this is their new you know fancy
26:51 reasoning Chain of Thought engine and
26:55 then under more models you still have
26:58 access
26:59 to gp4 Legacy model this is the one
27:03 that's 1.8 trillion parameters and eight
27:07 eight or 10 different um they call it
27:10 mode of experts or something like that
27:12 mode of experts no something of experts
27:14 collection of experts but there's
27:16 there's eight or 10 different models I
27:18 think it's 10 different models make this
27:20 up and they're they're
27:24 each 180 billion parameters and so that
27:28 all adds up to 1.8 trillion
27:31 parameters it's slower but I think for
27:35 writing this one might be a really
27:37 interesting thing to go with but you
27:40 know test gp4 Omni but then if we're
27:45 talking about the 01
27:49 preview let me try something
27:54 here so I'm going to put this in 01
27:56 preview I'm going to go try to find
27:57 something hang on
28:04 see I have a feeling that 01
28:06 preview might help you structurally like
28:10 write the outline and do the analysis of
28:13 the thing better
28:15 than any of the others because it's
28:18 doing multi-step processing but let me
28:20 let me go look for something
28:23 here I have an I have a a like a
28:27 fragment
28:30 of a screen
28:49 play all right here's a PDF
29:03 okay so I got
29:15 this and then do I have an outline
29:37 yeah here's an outline
29:40 okay this is a very incomplete outline
29:43 but that's okay all right so I'm G to
29:45 come over here I'm going to go
29:54 um here are two fragments
29:59 from a
30:01 screenplay the first
30:08 fragment is an incomplete
30:12 outline though it
30:14 has the main beats of Acts 1 2 and 3 and
30:19 the second fragment
30:37 oh God damn
30:54 it is
30:59 the opening scenes from the film period
31:29 so let's go oh [ __ ]
31:43 stop
31:46 stop
31:49 stop it's thinking I can't stop it did I
31:52 stop it I think I stopped it okay so
31:56 wait let me paste that in there so this
32:00 is interior
32:02 oops you'll see what I'm doing in a
32:04 second here um so I'm going to go um
32:08 dash dash
32:09 dash
32:12 outline Dash Das
32:15 Dash I'm going to go grab the top of
32:17 this prompt
32:19 again so it was promising it took that
32:22 it took that amount so here are two
32:24 fragments so I'm going to copy that and
32:26 going to put that here
32:29 and then down
32:32 here so that's my outline my incomplete
32:36 outline and then I'm going to go
32:42 um
32:43 um bu bum bum and then this is going to
32:46 be um opening
33:19 okay so I set it up I said I'm giving
33:22 you some [ __ ] here's an outline and an
33:25 incomplete fragment so now we got to
33:26 write the prompt so we'll go um bang
33:31 bang bang
33:33 um end of
33:38 input so apparently giving it little
33:40 markers like that is a good thing and
33:43 then I'm hitting shift return to give
33:45 myself extra line breaks so now I'm
33:46 going to write the actual prompt
33:53 um based on these inputs I'd like you to
33:56 analyze it and
34:11 and fill in the outline for
34:16 me and make it a good compelling
34:21 screenplay with lots of twists and turns
34:23 and surprises
34:30 and also introduce new
34:33 characters that are memorable and move
34:36 the plot forward
34:41 organically okay so let's see how this
34:44 bad boy does so okay so just so you know
34:47 what I'm doing
34:49 um gp4 Omni the new one and the old
34:55 gp4 both would fail if you gave it
34:57 really complicated tasks like this I'm
35:01 thinking that this reasoning version of
35:05 01 might be good at that so that's
35:07 that's why I got excited to come over
35:09 here and
35:11 play and this could this could fall
35:13 straight on his face so say thinking
35:15 creating a compelling narrative or
35:17 crafting a compelling
35:18 narrative unraveling the
35:21 mystery understanding consequences
35:23 facing a pivotal Choice taking a closer
35:25 look capturing moments so these are all
35:28 the different steps of the thinking it's
35:29 doing reliving nostalgic thoughts
35:31 meeting and planning rethinking actions
35:34 this is looking promising tracking
35:35 progress wrapping up act one identifying
35:38 threats realizing the threat engaging in
35:41 a quest examining the possibilities this
35:44 is this looks really promising revealing
35:47 crucial hints trying another way Jack Jr
35:51 confronts Styles facing deep fears
35:53 Unlocking The Hidden Truths unraveling
35:55 the mystery holy [ __ ] this is is looking
35:59 promising crafting a narrative
36:01 monitoring Becky relying on the album
36:04 unraveling
36:06 Clues so while it's doing this
36:10 the I thought for 59 seconds so that was
36:14 a
36:16 60c
36:18 um chain of thought that it did so every
36:21 one of
36:23 these moments these are separate prompts
36:26 that it gave itself crafting a compell
36:28 narrative I'm analyzing the screenplay
36:31 fragments crafting the screenplay
36:33 involves analyzing the outline crafting
36:35 a thrilling screenplay I'm digging into
36:37 Jack Jr's character a loner obsessed
36:39 with photography ensuring the plot is
36:41 engaging so just so you know what the
36:43 the basic plot is here
36:46 is family moves into a new house in in
36:50 Nevada they go out just as a family
36:54 while the movers are you know unloading
36:57 [ __ ] into the house they just go for a
36:59 drive to take a break they end up in
37:01 this little curio
37:03 shop and Jack's a
37:05 photographer and Silas the the curio
37:09 shop
37:10 owner um how does it work does he give
37:13 him an album or Jack discovers an album
37:16 but anyway Jack discovers this album of
37:18 old
37:20 photographs and the photographs in the
37:23 album start
37:24 predicting events in Jack's life so it's
37:27 it's a soothsaying photo album of all
37:30 these old creepy photographs so that's
37:32 kind of the setup of the thing crafting
37:34 a compelling screenplay unraveling the
37:38 mystery understanding the consequences
37:40 if Jack Jr fails his family will lose
37:42 him leaving him to grieve and face life
37:46 alone facing a pivotal Choice he must
37:48 decide between watching his family's
37:50 destruction or abandoning his perception
37:52 to survive in act three wow this is cool
37:56 capturing moments
37:58 okay so let's see what it came up with I
38:01 mean this is this is seriously [ __ ]
38:05 impressive just the thinking that it did
38:08 here so where's the did it give us an
38:12 actual
38:19 answer um switch
38:23 model
38:24 copy where's the
38:28 I'm just going to say
38:37 continue transforming an outline
38:39 ensuring policy compliance ensuring
38:42 content safety adding tension and
38:44 Intrigue analysis and expanded outline
38:47 the album by Kyle
38:52 Shannon the script was generated from
38:54 Contour by Mariner software
38:59 the intro four questions who's your main
39:03 character okay let's
39:07 see all right the log line form when a
39:11 guilt-ridden photography obsessed
39:13 12-year-old old boy inherits a
39:15 mysterious photo album that predicts
39:17 events involving his family and new town
39:20 he must decipher the album secrets and
39:23 confront a Sinister antiques dealer to
39:26 prevent disaster ultimately RIS ring
39:28 everything to change fate and save those
39:31 he
39:32 loves pre-writing
39:34 checklists have you answered the four
39:36 questions oh so this took the this took
39:39 the the methodology from Contour
39:41 software and it it's basically filling
39:44 it all in is your main character
39:46 sympathetic is his goal compelling yes
39:49 do you have a clear central question yes
39:51 can jack Jr prevent the tragedies
39:54 predicted by the album and save his
39:56 family do the obstacle become
39:58 increasingly difficult yes is the
40:00 antagonist ruthlessly committed yes are
40:03 the stakes life or death yes have I
40:06 given those Stakes a human face in
40:08 Stakes character yes his family and his
40:10 new friends have I used the F the
40:13 formula to see if my story Can Go the
40:15 Distance yes did my main character move
40:18 through the four Arch archetypes
40:22 yes
40:24 introduce uh in interior I guess Act One
40:27 the orphan introduce the
40:29 hero all
40:31 right flashback
40:36 scene okay so here it's I think it's
40:40 just act two part one The Wanderer Pages
40:43 28 to
40:50 54 Jack junor and Mia research the
40:52 history of the album and discover that
40:55 similar incidents have occurred in the
40:56 past involving towns and seers their
41:00 investigation draws unwanted attention
41:02 from Logan the
41:03 bully who becomes suspicious of their
41:06 secretive Behavior oh that's
41:11 good Mia introduces Jack Jr to his
41:14 grandmother Mrs Thompson a local
41:16 historian who knows about the um I would
41:21 [ __ ] use this new reasoning engine to
41:24 to to [ __ ] structure your novel dude
41:27 or dudette I don't know if you're a dude
41:29 or a
41:31 dudette act three the
41:34 Martyr new characters introduced Mia
41:37 Thompson Jack Junior's artistic
41:39 classmate who becomes a close Ally Ethan
41:42 Parker a tech saavy student Logan Meers
41:44 the school bully it's a good name
41:47 Logan hey Logan [ __ ] you you little
41:52 twit Mrs Thompson M as grandmother local
41:55 historian and Mr Harris the supportive
41:58 art teacher cuz he's a photographer
42:01 that's really good twists and turn Silas
42:04 is true identity Silas is revealed to be
42:07 a guardian cursed to perpetuate the
42:09 album's cycle the album's
42:12 origin is a conduit for fate created
42:15 centuries ago to test the humanity test
42:18 Humanity's will against Destiny Becky's
42:20 transformation from dismissive sister to
42:22 supportive Ally Sam's fate the
42:25 Revelation that Sam is alive but in a
42:28 coma adds emotional weight and motivates
42:30 Jack the rituals true purpose initially
42:33 believed to be a way to destroy the
42:35 album it's actually a test of Jack Jr's
42:37 ability to forgive himself and accept
42:39 help from others conclusion the filled
42:43 in outline now represents a compelling
42:45 screenplay with emotional depth and
42:46 memorable characters so here I'm going
42:48 to try one one final test here which I
42:51 think it will fail at but let's try it
43:00 I'm going to copy a scene
43:04 here and I'm going to
43:08 say based on the writing
43:14 [Music]
43:16 style of the screenplay
43:23 fragments from the opening of the film
43:30 write the following scene from the
43:33 outline that you
43:43 created all
43:44 right Silas reveals that he's the
43:47 guardian of the album cursed to find
43:49 successors to Bear its burden he offers
43:51 Jack Jr a deal to spare his family in
43:54 exchange for his inter Eternal servitude
43:57 this is this is
43:59 good all right let's see if it can write
44:02 let's see if it can write
44:05 dialogue interior Hilltop
44:08 Knight this is looking promising the
44:11 lunar eclip casts an eerie light an
44:14 eerie red glow over the hilltop Jack Jr
44:16 Mia and Ethan stand in the center of a
44:19 makeshift ritual Circle surrounded by
44:21 candles and an an ancient symbol drawn
44:24 in chalk Becky watches nervously from
44:26 The Edge clutching a flashlight a sudden
44:29 gust of wind extinguishes the candles
44:31 the temperature drops Shadows stretch
44:33 unnaturally as silus emerges from the
44:36 darkness his tall lanky figure
44:38 silhouetted in blood red against the
44:40 blood red moon Silas so you've come to
44:43 play Hero Jack Jack junr steps forward
44:46 trying to mask his fear Jack junr it's
44:49 over Silas we're ending this tonight
44:52 silus CH Chuckles a dry Hollow sound
44:56 silus ending it
44:58 you have no idea what it even is Mia and
45:01 Ethan exchang worried glasses glances
45:04 not for nothing
45:05 people this is not bad at
45:09 all I have
45:11 never
45:13 ever had a large language model write
45:18 screenplay write a
45:21 screenplay that the dialogue wasn't
45:23 painfully
45:25 bad never
45:29 me and Ethan exchange glasses Becky Jack
45:32 who is this guy Jack Jr an old dot dot
45:35 dot acquaintance like that's just good
45:37 writing um Silas pale blue eyes fix8 on
45:40 Jack Jr silus you can't stop you can't
45:44 stop fate child the album's will is
45:47 beyond your control Jack Jr we'll see
45:49 about that Jack Junior's a little pistol
45:52 isn't he jack Jr raises an old artifact
45:55 a pendant glowing faintly Sil raises an
45:58 eyebrow silus ah The Amulet of severon
46:02 clever but trinkets won't save you he
46:05 takes a step closer the ground beneath
46:08 him Withers turning to Ash Mia Jack be
46:10 careful silus gestures the pendant
46:13 crumbles to dust in Jack's hand Ethan
46:15 what the silus I've walked the Earth for
46:18 centur I've walked the Earth for
46:20 centuries bound to the album do you
46:23 think you could defy me with your parlor
46:25 tricks there you go Source Camp Jack
46:29 junr squares his shoulders maybe not
46:31 with tricks but with determination
46:33 silence his expression softens almost
46:35 sympath sympathetic probably Jack Jr
46:37 reminds him of
46:38 him silus determination uh silus
46:43 determination I remember having that
46:46 once he circles them slowly like a
46:48 predator stalking prey also the length
46:51 of this whenever I've had one of these
46:52 large language models write a scene it
46:55 writes like seven lines so this is an
46:58 actual scene like this is an
47:00 actual I I am a little I'm a little
47:04 shooketh this is this is good but okay
47:08 here's what's
47:09 important from from the way I'm thinking
47:11 about
47:16 this doing a narrative analysis and
47:19 creating a detailed like
47:22 finishing the outline of a
47:25 screenplay is a big complicated
47:28 multi-step process to do right so this
47:31 is the kind of project this is where you
47:34 would use 01 versus chat
47:38 GPT right if you've just got one little
47:41 fragment you want to create use chat gbt
47:44 but if you want a big nasty hairy
47:47 multi-step
47:49 thing
47:51 um okay determination I remember having
47:54 that once he circles them slowly like a
47:56 predator stalking prey this is I got to
47:59 tell you this is just good writing it's
48:02 just good writing you see I was like you
48:06 a seer chosen by the album I tried to
48:09 fight it too I I mean this
48:12 is what's amazing about this
48:18 is when I came up with the idea for this
48:21 screenplay I had a really clear idea for
48:25 the main parts of the ACT or or the
48:28 axe I had a really clear idea for the
48:31 opening and then I couldn't figure out
48:33 what to do for the rest of it and I just
48:35 lost motivation and didn't write it so
48:38 this is a scene like two-thirds of the
48:40 way through the movie where I had
48:43 nothing and this is feeling right this
48:46 is feeling like the film I
48:48 wrote which is wild okay you see I was
48:52 like you once a Seer chosen by the album
48:56 I tried to fight it too Jack Junior's
48:58 eyes widen what are you talking about
49:00 Silas stops his glaz his gaze distant
49:04 the album is no mere object it's a force
49:06 a conduit of Destiny those who possess
49:09 it are cursed to witness the threads of
49:12 Fate unable to alter
49:15 them he looks directly at Jack Jr unless
49:19 of course they accept their
49:22 role Becky Jack let's get out of here
49:25 this guy's nuts Silas snaps his fingers
49:27 Becky is suddenly Frozen in place her
49:30 flashlight flickering erratically Becky
49:33 he rushes to her but an invisible
49:34 barrier blocks
49:36 him jack Jr what did you do to her she's
49:40 unharmed for
49:42 now Mia steps forward
49:45 defiantly um leave her alone silus ah
49:49 the loyal friends how touching I love
49:51 silus he's such a cynical
49:54 dick he waves a hand and Mia and Ethan
49:58 find themselves rooted to the spot
49:59 unable to move Ethan I can't move silus
50:05 now Jack let's have a real conversation
50:07 Jack Junior's fear turns to anger
50:10 release them Silas in due time he closes
50:14 the distance between them Silas continue
50:16 I offer you a choice no I offer you a
50:20 choice ex
50:22 no accept your destiny take your place
50:26 as the album's guardian and your family
50:28 your friends they all go free safe from
50:30 harm Jack junor hesitates and what
50:32 happens to me you will inherit my burden
50:36 immortality of A Sort bound to the album
50:39 guiding the next sear when your time
50:42 comes why me silus Smiles faintly
50:46 because you can see truly see just as I
50:50 once did he extends his hand long
50:52 fingers beckoning join me Jack spare
50:56 them the pain that fate has in store
50:59 Jack Jr looks at Becky Mia and Ethan all
51:01 Frozen fear in their eyes and if I
51:04 refuse silus Expressions expression
51:07 darkens then the album will claim them
51:11 one by one tragedies you cannot prevent
51:14 suffering you will be powerless to stop
51:17 Jack junr clenches his fists you're
51:20 lying am I you see the photographs you
51:23 know it's coming images through Jack's
51:26 mind the haunting photos of the album
51:28 depicting disasters involving loved ones
51:30 so this is good screenwriting right so
51:32 we just went to a visual cutaway of all
51:35 of these albums flashing through his
51:37 mind which we could go make in in uh
51:39 text to video now their lives and
51:42 exchange for your service and fair trade
51:44 Jack junr looks down conflicted I don't
51:47 want to I don't want anyone to get hurt
51:49 then accept your offer a tense silence
51:53 the wind hows around them the eclipse
51:55 nearly in its tot totality Mia Jack
51:58 don't silus stands at glances at Mia
52:01 tightening his grip over her silus your
52:04 time is running out Jack junr takes a
52:06 breath if I do this you swear they'll be
52:09 safe Silas on my
52:11 word and and what about you what happens
52:15 to you a hint of relief across silas's
52:17 face I am released free at last from
52:21 this eternal chain Jack Junior considers
52:24 then looks Silas in the eye how do I
52:26 know you're telling the truth Silas you
52:28 don't but what choice do you have Jack
52:30 Jr closes his eyes the weight of the
52:33 decision crushing him suddenly a faint
52:35 glow emits from his backpack the album
52:38 inside reacts to the eclipse voice
52:42 Whispering Jack he opens his eyes who a
52:47 translucent figure appears between them
52:49 Sam his old friend Jack don't do it so
52:53 he thought he killed Sam Jack junr is
52:55 stunned Sam scals enough of this this is
53:00 amazing this
53:05 amazing
53:08 unbelievable um yes I would pay the 20
53:10 bucks to chat GPT your question was
53:13 which one should I which which one
53:16 should I pay 20 bucks to Chad
53:19 GPT without
53:21 question I I I
53:23 am okay we don't need to we don't need
53:26 to finish the album I I think you see
53:28 what happened
53:30 there holy [ __ ]
53:37 [ __ ] I I
53:39 don't I can't if you've if you've done
53:42 any creative writing in your
53:46 life
53:49 um and especially if you've done that
53:53 and used these large language models
53:58 I have
53:59 yet until this moment to ever get it to
54:03 generate
54:04 something that actually feels like
54:06 something I would have
54:12 written like the tone of how it wrote
54:19 this I
54:22 mean I'm kind of feeling like I'm kind
54:25 of feeling like this is
54:28 um I could go through its outline and
54:30 just have it write all the
54:35 scenes and I would have a really good
54:43 screenplay like there
54:46 wasn't there wasn't anything in that
54:48 that was like bad dialogue there wasn't
54:51 anything in that that was like
54:53 incongruous
54:57 scenes like just knowing where this is
54:58 in the
55:00 story I can see how we build up to
55:04 this Jack Jr realizes that the
55:07 photographs are visions of what's to
55:10 come his friend Sam he thinks he's
55:13 dead but he's actually in a
55:24 coma one of the things we do here on the
55:26 channel is we talk about um Kevin
55:28 mallister
55:31 moments I'm having one of those right
55:37 now can this 401 be applied to the
55:41 Sydney play now
55:44 yeah yeah it
55:47 can I'm inspired seriously I need to
55:50 pull out my screenplay from 1990 this
55:53 this screenplay is from like 98 or no no
55:57 this one this one was a 2005 maybe this
56:00 was a more recent
56:01 one but
56:03 my the the first one of the first gpts I
56:06 made was I took my salmon flying East
56:09 screenplay which is the first screenplay
56:11 I ever wrote from 1990 I think and it
56:15 totally brought it to life to me but it
56:17 failed when I tried to do [ __ ] like this
56:19 with
56:24 it like what Chach BT could do is it
56:26 could tell me about the screenplay as it
56:28 was but when I tried to have it
56:30 augmented or analyze it or change things
56:32 or update things it just kind of broke
56:35 it and was
56:37 shitty
56:40 I I'm reticent
56:43 to give this a complete Victory
56:46 because like all of these things
56:50 um they tend to be brilliant at first
56:53 blush and then when you actually try to
56:54 put them together for real they kind of
56:56 fall fall apart
56:59 but the outline that it wrote seemed
57:02 pretty complete and I just grabbed a
57:04 random scene in there and it
57:08 wrote this is a good
57:10 scene it's a complete
57:13 scene with visuals and dialogue that
57:17 doesn't suck and
57:20 drama [ __ ] silus emerges out of the
57:23 shadows and then he [ __ ]
57:27 freezes them with his finger
57:39 Ray wait what
57:41 happened script Rhythm so what happened
57:45 okay so let me let me go back for those
57:48 of you that are new here I'm I'm just
57:49 I'm I'm a little
57:51 shook okay so so some background
57:57 what I'm using right now is the chat GPT
57:59 01 preview this is their new reasoning
58:02 engine and remember this is the worst it
58:04 will ever be yeah exactly Source
58:07 Camp I can't I can't believe it it said
58:10 you you with your parlor tricks you're
58:11 not going to get me with parlor tricks
58:13 that was pretty good um okay so what I
58:16 did was I
58:18 took I I have the first three or four
58:21 scenes of a screenplay and I have an
58:23 incomplete outline because I I have ADD
58:26 I can't tell you how many beginnings of
58:30 screenplays and incomplete outlines I
58:32 have but it's a lot this was one that I
58:34 get farther than others because I really
58:36 liked it but when I don't have a writing
58:38 partner I tend to just flame out on [ __ ]
58:41 so this is a a project that I was
58:44 excited
58:46 about but flamed out on hang on I have
58:48 to do a puzzle piece to prove I'm a
58:50 human
58:52 [ __ ]
58:54 so so I said here are two fragments from
58:57 a screenplay the first fragment is an
58:59 incomplete outline which I did in the
59:01 Contour system which it knew what that
59:04 was by by Mariner
59:09 software and so it took my you know the
59:12 things that I answered
59:17 there and it it took them into its brain
59:21 and
59:23 then let's see and then I put in the
59:26 scene
59:27 right so professional movers here's Jack
59:30 Jr come on can't we go check out the
59:32 neighborhood Jack son the movers are oh
59:34 the movers are doing great let's go on a
59:36 little Adventure we'll have plenty of
59:37 time when we get back Jack junr Jack Jr
59:40 pumps his arm and runs back outside to
59:42 gather up his sister are you sure they
59:45 can uh figure out where to put
59:48 everything uh yeah I had Garanimals when
59:51 I was a kid um so so there's this
59:55 there's the you know first scenes of the
59:57 screenplay that were written and then
59:59 what I wrote was the prompt that I wrote
1:00:02 was based on these inputs I'd like you
1:00:05 to analyze and fill in the outline for
1:00:07 me and make it a good compelling
1:00:09 screenplay with lots of twists and turns
1:00:11 and surprises and also introduce new
1:00:14 characters that are memorable and move
1:00:16 the plot forward
1:00:18 organically the way this 01 model works
1:00:21 is it
1:00:23 thinks and so what it did is it thought
1:00:25 for 59 seconds and if you see this you
1:00:28 know I thought for x amount of seconds
1:00:30 it's got a little drop down menu here
1:00:32 and these are the Chain of Thought um
1:00:36 prompts that it gave itself crafting a
1:00:39 compel compelling narrative crafting a
1:00:41 thrilling screenplay unraveling the
1:00:43 mystery understanding the consequences
1:00:45 facing a pivotal Choice taking a closer
1:00:48 look capturing moments so it's going
1:00:51 through in a step-by-step process like a
1:00:54 screenwriter would do
1:00:59 and it's filling in the missing pieces
1:01:02 realizing the threat Jack Jr discerns
1:01:04 the album's forecast so this is all in
1:01:06 the background this is all just doing it
1:01:07 behind the scenes we we we're not seeing
1:01:09 what it's actually doing but it's giving
1:01:11 us little summaries of what it was
1:01:14 doing and then it hung I think because I
1:01:17 was [ __ ] with it and I said continue
1:01:20 and then it did some more thinking
1:01:22 transforming the outline insurance
1:01:24 policy compliance I'm tasked with
1:01:27 crafting a detailed screenplay by adding
1:01:29 scenes characters while ensuring content
1:01:32 aligns with open ai's policies so it's
1:01:34 it's making sure open AI doesn't get
1:01:36 sued while it helps me write my
1:01:38 screenplay and then crafting the
1:01:39 screenplay I'm working on a filled in
1:01:41 outline for the screenplay and then it
1:01:45 gives
1:01:45 me um some answers from the Contour
1:01:49 system of you know did it did it do it
1:01:52 [ __ ] right there's your pre-writing
1:01:54 checklist
1:01:56 and then it gives
1:01:59 me Jack pedal's modified bicycle with a
1:02:02 built-in photo printer around the new
1:02:04 culdesac he snaps photos of his
1:02:07 surrounding trying to familiarize
1:02:08 himself with the area he notices a group
1:02:10 of neighborhood kids playing basketball
1:02:12 and attempts to join them by offering to
1:02:14 take their
1:02:16 picture then we meet some people
1:02:18 flashback scene Jack Junior's with his
1:02:21 best friend Sam in their Old Town Sam
1:02:24 climbs a tree while Jack Junior
1:02:26 photographs him jack becomes engrossed
1:02:29 in capturing a crow in the background
1:02:30 and ignores Sam's please for help Sam
1:02:33 falls re re resulting in a severe spinal
1:02:36 injury not fatal but Critical Jack
1:02:38 blames himself for not helping so it
1:02:41 captures [ __ ] that I had already written
1:02:44 and then it
1:02:45 writes the entire outline following the
1:02:48 Mariner software Contour software
1:02:51 outlining
1:02:55 model and then what I
1:02:58 did was I had it and then it tells me
1:03:01 the new characters it introduced which I
1:03:03 asked it to here's the twists and turns
1:03:06 in the plot and then I
1:03:08 said based on the writing style of the
1:03:11 screenplay fragments you
1:03:13 have write this scene that that you
1:03:17 invented and it wrote a [ __ ] good
1:03:20 scene with good dialogue I've never ever
1:03:24 ever ever had a large line anguage model
1:03:27 write anything resembling good dialogue
1:03:29 not even
1:03:33 close this is [ __ ]
1:03:43 staggering I think this is the
1:03:45 screenplay not the script you just
1:03:46 created it I mean it could be right like
1:03:48 this
1:03:53 is yeah it's funny you know this this
1:03:55 channel in a lot of ways you know I've
1:03:58 been doing this for 18 months now so and
1:04:01 and I've got all these videos archived
1:04:03 on
1:04:04 YouTube there is a narrative to this
1:04:06 channel right we we have collectively
1:04:10 been witnessing the evolution of these
1:04:14 tools right and we've had multiple Kevin
1:04:18 McAllister moments each over that time
1:04:23 and and this is another one what what's
1:04:25 amazing about this
1:04:28 is the the response that I've seen
1:04:32 to this new model from open
1:04:39 AI wait hang
1:04:41 on hang on
1:05:15 sorry sorry my my son is stoned and one
1:05:18 of his friends dared him to pop his head
1:05:20 in and say hi while I was in the middle
1:05:21 of my
1:05:24 life um
1:05:27 yeah but this channel the there there is
1:05:29 there's a narrative there's a narrative
1:05:31 to this channel if you follow it over
1:05:36 time
1:05:37 huh wow Kyle do you think they're
1:05:41 opening themselves up to a
1:05:44 lawsuit by copying the company's
1:05:47 method I
1:05:51 listen they didn't copy the company's
1:05:53 method the company's method is in the
1:05:55 training data
1:05:56 and it
1:05:57 spontaneously
1:05:59 took my
1:06:01 incomplete version of it and filled it
1:06:05 in
1:06:09 so I
1:06:11 mean do I think that Mariner software
1:06:14 could sue them because they could get it
1:06:16 to do this yes do I think any of these
1:06:19 lawsuits are going to hold up in court
1:06:21 no
1:06:24 because the nature of how the systems
1:06:26 work they're not [ __ ] copying and
1:06:29 pasting I mean this was the thing that
1:06:33 Oprah revealed last night is that is
1:06:37 that she doesn't understand how they
1:06:39 work any more than you know the average
1:06:43 fella she
1:06:45 she when he showed the picture of the
1:06:51 the like the the image generated from
1:06:54 mid Journey or whatever ever was created
1:06:56 from she's like but who's that kid he
1:06:59 was copy and pasted in there right you
1:07:01 know she said some version of that she
1:07:03 doesn't understand how the fundamental
1:07:05 technology Works have you ever compared
1:07:07 any other model's output to this exact
1:07:10 prompt because you should probably pay
1:07:13 but I mean I've tried
1:07:15 to I mean I've tried to do this many
1:07:18 times with screenplays and and the
1:07:20 result is
1:07:22 always it has never been remotely
1:07:25 satisfied Ying it tends to write the
1:07:28 outline pretty
1:07:31 well but
1:07:34 thinly this was not a thin writing of
1:07:36 the outline this was
1:07:38 a wellth thought
1:07:41 out crafting of the outline and then
1:07:45 whenever I've had it try to write
1:07:47 anything like write a scene for me the
1:07:50 scenes are always short you know like
1:07:54 like maybe a page or maybe a page and a
1:07:57 few lines on the next page but like a
1:08:00 minute long this is probably five
1:08:04 pages and the dialogue's good this this
1:08:08 you know the visual directions are
1:08:16 good let me see how does this thing end
1:08:28 Jack Jor I won't take your deal silas's
1:08:30 eyes Blaze with Fury well then you have
1:08:33 sealed their Fates he unleashes a torn
1:08:35 of dark energy around them Jack braces
1:08:37 for impact but a shield of light forms
1:08:40 around the group emanating from the
1:08:42 album silus is pushed back struggling
1:08:45 against the light no this can't be the
1:08:47 album levitates out of Jack's backpack
1:08:49 Pages flipping rapidly the photographs
1:08:52 emit beams of light each one
1:08:53 representing a monument of love and
1:08:55 connection it's over silus your hold is
1:08:58 broken you think you've won the album
1:08:59 cannot be destroyed maybe not by me
1:09:02 alone he looks at his friends and sister
1:09:04 but together they join hands D the
1:09:08 eclipse reaches totality bathing the
1:09:09 hilltop in deep red light before before
1:09:13 the moon moves past the normal light
1:09:16 returns is it over Mia I think so the
1:09:18 album gently floats back into Jack's
1:09:20 hands the cover is now blank the page is
1:09:22 empty what happened we broke the cycle
1:09:25 look they Gaz a they gaze at the photo
1:09:27 wait he opens the album a single new
1:09:29 photograph
1:09:31 appears it's all four of them standing
1:09:34 together at this very moment they gaze
1:09:37 at the photo a symbol of victory in a
1:09:39 new found Bond I guess this is a new
1:09:44 beginning it's good it's good
1:09:50 like I'm trying to think like is there
1:09:52 anything I I wouldn't like I don't I
1:09:54 don't know the outline enough to know
1:09:56 what comes before and after this but
1:09:59 like as a as an important scene in the
1:10:03 film where Silas confronts
1:10:06 Jack pretty [ __ ]
1:10:10 solid can you name an actor in a
1:10:13 prompt yeah sure here watch we'll do
1:10:18 [Music]
1:10:19 um please give me casting
1:10:22 recommendations for the film
1:10:33 thinking clarifying
1:10:44 guidelines Noah jup a quiet place honey
1:10:52 boy Mia Thompson McKenna Gra gifted The
1:10:56 Haunting of Hill House Ethan Parker
1:11:00 Jacob tble room and
1:11:03 wonder Silas William defo perfect
1:11:07 William defo would be perfect for this
1:11:10 within his distinctive features an
1:11:12 ability to portray mysterious intense
1:11:14 characters defo could embody the
1:11:17 enigmatic silus
1:11:19 effectively Jack Whitmore the Father
1:11:22 Paul Rudd that's great Amy Adams for the
1:11:25 mother
1:11:26 that could work John Dylan Grazer
1:11:30 for Logan okay I don't know him M as
1:11:33 grandmother Viola Davis oh because
1:11:36 that's his friend
1:11:40 yep Stanley tchi for the art teacher
1:11:43 love it or uh or what's his name Mandy
1:11:46 penin could be good additional character
1:11:49 Sam Jack's old
1:11:52 friend notes on casting children's age
1:11:55 may vary slightly but the recommended
1:11:57 actors are within believable a age
1:11:59 ranges diversity
1:12:03 chemistry alternative
1:12:07 options this is quite good so yes it
1:12:11 can can I write the score for the
1:12:14 background music no but sunno
1:12:18 can oh
1:12:20 man oh sorry I meant make him talk like
1:12:23 dairo oh yeah I mean you could certainly
1:12:26 go in I I
1:12:28 mean I told it to follow my style of how
1:12:32 how things were written above for the
1:12:34 dialogue but you could certainly go in
1:12:36 and say yeah write this for a particular
1:12:38 character sure absolutely you know write
1:12:41 it for Samuel
1:12:43 Jackson the get the [ __ ]
1:12:45 photographs out of the [ __ ]
1:12:49 album hang on I dropped my cards I
1:12:52 dropped my cards on the floor
1:13:03 [Music]
1:13:06 um well that was a fun little rabbit
1:13:09 hole to go down holy
1:13:12 [ __ ] holy [ __ ]
1:13:16 Batman have you tried chat jpt
1:13:18 strawberry 01 that's what we just spent
1:13:19 the last hour on you just missed a good
1:13:23 one um yes so yes I I have
1:13:28 um my my initial thought of it
1:13:32 was I don't know how to prompt it I
1:13:35 don't know what to ask it like so
1:13:37 basically the deal is
1:13:39 this
1:13:41 um the 01 API is incredibly expensive
1:13:45 it's $60 for a million
1:13:49 tokens um because it's it's doing
1:13:52 massive amounts of pro prompting in the
1:13:54 back end
1:13:56 self- prompting um that's how its
1:13:58 thinking is going on um David Shapiro
1:14:01 has called this effectively a piece of
1:14:03 [ __ ] that this is just Chain of Thought
1:14:05 um prompting um behind the scenes and
1:14:08 it's not that big a deal and he's going
1:14:10 to try to write an open source
1:14:11 competitor to this called
1:14:13 raspberry which is kind of
1:14:17 fun but what become what what becomes
1:14:19 really clear to me is that this is not a
1:14:22 model that you'll use for everything
1:14:24 this is not a general model it's a
1:14:25 specialized model that says when you
1:14:28 have a complicated multi-step thing you
1:14:30 need to deal with then you use this um I
1:14:34 missed it what happened so what
1:14:37 happened I took a fragment of an early
1:14:40 screen play and a fragment of an
1:14:43 outline that you know was not filled
1:14:47 in and I told it
1:14:51 to take all the [ __ ] as input and write
1:14:55 WR a compelling screenplay
1:14:58 outline and it did and it sort of parsed
1:15:01 from my outline stuff that I was using
1:15:04 Mariner
1:15:05 software um and it you know it it knew
1:15:08 the Mariner software so so Mariner is a
1:15:12 screenwriting like a writing company and
1:15:15 then Contour is their screenplay or
1:15:18 their story outlining software and so it
1:15:22 knew that and so it basically used the
1:15:26 methodology and it filled in the entire
1:15:29 [ __ ] screenplay which was impressive
1:15:32 like that it did that it was good and
1:15:35 the other thing you should know if you
1:15:36 haven't seen it the way 01 works is you
1:15:41 give it a prompt and then
1:15:47 it it says
1:15:49 thinking and in the case of this because
1:15:52 it had a lot to do
1:15:58 it says thinking and see here it says
1:16:00 thought for 59 seconds and if you open
1:16:03 up this thought for 59 seconds here are
1:16:06 all the prompts it gave itself after I
1:16:08 said you know please fill this out
1:16:10 crafting a comparative a compelling
1:16:13 narrative crafting a thrilling
1:16:15 screenplay unraveling the mystery
1:16:17 understanding consequences facing a
1:16:19 pivotal Choice taking a closer look
1:16:21 capturing moments so it is thinking step
1:16:24 by step
1:16:27 through how to
1:16:28 write this screenplay as a compelling
1:16:35 story wrapping up act one I'm working
1:16:39 through the outline and focusing on Act
1:16:40 One ensuring all key elements and
1:16:42 characters are in place so it did all
1:16:45 that then it gave me some analysis you
1:16:49 know some story analysis from the the
1:16:51 Contour
1:16:53 stuff then it
1:17:00 um then it gave me the outline so here's
1:17:02 this detailed
1:17:08 outline I it did weird things here where
1:17:10 it's like interior yes interior No in
1:17:14 fact you know what I'm going to do I'm
1:17:16 going to
1:17:21 say please rewrite the outline and
1:17:24 anywhere that you have an interior or
1:17:27 exterior scene that's labeled yes or no
1:17:31 comma write an actual appropriate scene
1:17:36 heading for that part of the
1:17:42 outline so so there's
1:17:45 thinking allocating roles so here's the
1:17:48 re revised site outline with scene
1:17:52 headings
1:17:57 as previously
1:18:04 detailed yeah so now we've got now the
1:18:07 scenes are actually named
1:18:09 appropriately various
1:18:21 locations
1:18:23 amazing um so then after I had it write
1:18:26 the outline I just picked one of the
1:18:29 segments that it proposed and I picked a
1:18:32 dramatic one and then it wrote the scene
1:18:35 and it wrote it good it wrote a complete
1:18:38 scene with dialogue and shot
1:18:41 descriptions and action
1:18:45 descriptions that didn't suck I've never
1:18:48 had a large language model do that wants
1:18:51 me to look at mods and pay wait let's
1:18:58 see p p doesn't like
1:19:07 what yeah it's not formatting it
1:19:13 yet hey what's your
1:19:19 [Applause]
1:19:23 beef because code
1:19:30 poses oh this doesn't have code
1:19:32 interpreter
1:19:34 yet oh could it could it format this I I
1:19:37 mean it
1:19:39 could but um 01 does not have code
1:19:43 interpreter yet but I could probably
1:19:45 take that scene let's
1:19:51 see uh uh uh uh
1:20:02 and we could probably go over to
1:20:05 40 God damn I wish they would have named
1:20:08 these
1:20:09 things
1:20:12 appropriate names all right interior
1:20:15 Hilltop clearing
1:20:21 Knight yeah I don't care about the
1:20:22 formatting the formatting is the least
1:20:24 of the wor like formatting is easy
1:20:26 enough to just toss it into you know a
1:20:31 screenwriting formatting
1:20:34 thing all right Fade Out so there's the
1:20:37 scene so let's go to to
1:20:40 4.0 best for complex tasks and I'm going
1:20:42 to say
1:20:46 um take the following
1:20:49 screenplay
1:20:53 and put put it into a word
1:21:01 dock with standard screenplay
1:21:05 formatting so that I can download it
1:21:07 with the correct
1:21:17 formatting and so this should go to code
1:21:21 analysis so this is code analysis God
1:21:24 damn that thing
1:21:26 [ __ ]
1:21:27 Apple um so there's the so um code
1:21:31 interpreter is writing the python code
1:21:34 to do the
1:21:35 formatting and so it says left aligning
1:21:38 action description function add
1:21:40 character names and dialogue Center the
1:21:43 character name left align the dialogue
1:21:46 so look it wrote code to to basically
1:21:50 split apart the different screenplay
1:21:52 things this could actually work download
1:21:54 the screen play
1:21:57 document and then let me go
1:22:02 here holy [ __ ] it did it oh no it didn't
1:22:06 it did it did what [ __ ] chat GPT
1:22:08 always does it took part of it and did
1:22:11 it
1:22:13 right for the love of God don't say it
1:22:17 thinks oh P listen at this point you are
1:22:22 yes you are correct you're accurate I
1:22:24 shouldn't say it thinks
1:22:25 but at this point it's a semantic
1:22:29 argument so I like I'm fine to say it is
1:22:32 processing step by
1:22:34 step
1:22:38 um but this is
1:22:41 absolutely the same kind of output I
1:22:44 would get from a writing partner and
1:22:46 myself working on this for two
1:22:49 weeks a couple hours a night for two
1:22:51 weeks
1:22:53 straight so it might not be thinking but
1:22:56 the results of what it's doing are
1:22:58 identical to that output so so I so I'm
1:23:03 sorry I'm not accurately describing
1:23:05 what's happening technically but I think
1:23:09 I P I mean this is actually a really
1:23:11 interesting
1:23:14 conversation I think we're moving we're
1:23:16 increasingly
1:23:21 moving we're increasingly moving into a
1:23:24 state of these
1:23:26 Technologies where the distinction
1:23:29 between whether they're sentient or not
1:23:32 whether they're thinking or not whether
1:23:34 they're conscious or
1:23:39 not is independent from the
1:23:42 technology that they're going to be so
1:23:44 good at seeming like a thinking human
1:23:48 that that's going to be the perception
1:23:51 so I absolutely respect what you're
1:23:53 saying I'll try try to be more accurate
1:23:55 in how I'm saying
1:23:57 it that
1:23:59 said um I think that I think this
1:24:03 semantic argument is a losing battle as
1:24:05 these tools get more sophisticated
1:24:08 because I
1:24:10 don't the current state
1:24:13 of the current the the way Transformers
1:24:17 work is
1:24:19 not I don't think they're ever going to
1:24:21 be sensient right but I think how people
1:24:25 will perceive them is
1:24:29 that this is in no way replicating
1:24:32 thinking in my
1:24:35 opinion
1:24:40 well again I didn't say that it's
1:24:43 replicating thinking I said the output
1:24:45 of what it
1:24:47 did resembles what I would get if two
1:24:51 thinking writing Partners were working
1:24:53 on this together or or one neurotypical
1:24:56 that could actually sit down and write a
1:24:57 complete
1:25:01 screenplay so so I'm getting the output
1:25:04 as if someone thought about it let me
1:25:06 see if I can get this thing to actually
1:25:07 do the entire screenplay because chat
1:25:10 GPT is notorious for truncating [ __ ] um
1:25:14 but it it actually formatted it in
1:25:15 screenplay format that's astounding so
1:25:18 so yes you can do that um so let me um
1:25:22 let's see you truncated
1:25:25 the
1:25:26 scene um in the
1:25:30 document period don't do that period
1:25:33 give me the entire scene that I provided
1:25:37 in screenplay format in the document
1:25:41 period all right let's see if it does
1:25:44 it uh uh uh it says at the top open with
1:25:49 Microsoft Word yeah but it's still I can
1:25:51 open it with Microsoft Word but it it's
1:25:53 still the preview
1:25:56 um the preview didn't have all the text
1:25:59 in
1:26:00 it because on a Mac I should be able to
1:26:03 scroll to all the all the scenes and
1:26:06 it's just the top it's just the first
1:26:08 whatever all right let's grab this one
1:26:10 and see if it actually did
1:26:13 it yeah see now it did
1:26:19 it oh it still truncated it it did more
1:26:22 of it but it still truncated it
1:26:28 you still truncated it period think long
1:26:30 and hard period format the entire scene
1:26:35 period I know you can do it period my
1:26:39 career depends on it
1:26:43 [Laughter]
1:26:49 period all right I'll show you one other
1:26:52 thing that I did
1:26:55 yesterday after
1:26:57 our after our
1:27:03 [Music]
1:27:05 session my career depends on it all
1:27:08 right this is making me so happy was
1:27:11 bored and wanted company never used Tik
1:27:14 Tok like this before so wait what's your
1:27:16 name Jim wait J John Michael Roberts
1:27:19 John Michael Roberts um you're welcome
1:27:22 first of all uh welcome
1:27:25 this is what we do here night after
1:27:26 night now this is I'm I'm normally not
1:27:30 this flumix by the technology because we
1:27:33 get kind of cynical about it once once
1:27:34 we absorb it um but this new chat GPT
1:27:38 model actually did something
1:27:42 um markedly different than I've ever
1:27:44 seen before tonight so but the the whole
1:27:47 idea of this channel is just to hang out
1:27:49 here I go live five nights a week um
1:27:52 it's generally two hours or so and you
1:27:55 know we just talk about [ __ ] and and
1:27:57 demo stuff and see if we can break stuff
1:27:59 and play and do all that sort of happy
1:28:01 horse [ __ ] so welcome welcome welcome
1:28:03 and I'm glad I am
1:28:06 glad this is entertaining um let's
1:28:18 see Fade Out we got the whole screenplay
1:28:21 or we got the whole scene and this is
1:28:23 this is normal um screenplay
1:28:27 formatting it's pretty
1:28:31 remarkable
1:28:33 unbelievable so yeah so I could I could
1:28:36 probably if I took two hours tomorrow I
1:28:39 could probably finish the
1:28:41 screenplay and
1:28:46 then I could go over to notebook LM have
1:28:49 you seen this
1:28:50 [ __ ] all right we're going to go to
1:28:52 notebook LM
1:29:00 notebook lm. goole.com
1:29:03 and I'm going to go I think Untitled
1:29:05 notebook yeah so this is salmon flying
1:29:07 East okay so I took an old screenplay I
1:29:10 wrote salmon flying
1:29:13 East and when you're in Notebook LM the
1:29:18 way it works is you upload documents
1:29:20 over here and so you can upload multiple
1:29:23 documents and then this becomes a Gemini
1:29:26 powered chat GPT like experience where
1:29:29 you're interacting with your data that
1:29:31 you uploaded so what I uploaded was a
1:29:34 single screenplay here but they added a
1:29:37 new feature a day or two ago I don't
1:29:39 know how long it's been in there but
1:29:41 they started talking about it on Twitter
1:29:42 a day or two ago and within this thing
1:29:46 in the lower right hand corner there's
1:29:47 this thing that says notebook guide if
1:29:50 you click on that it's a whole little
1:29:54 synops of all the [ __ ] you have over
1:29:56 here so it's a summary of what's there
1:29:59 you can you can generate a fact a study
1:30:01 guide a table of contents a timeline a
1:30:03 briefing document here are suggested
1:30:05 questions for how you can do it so this
1:30:07 is the whole thing but they added this
1:30:09 thing in the upper right hand quarter
1:30:10 called audio overview and you can
1:30:13 generate it and right now there's just
1:30:15 one style which is two people like
1:30:17 they're doing a podcast so I made one
1:30:20 for this so let me load it load it in
1:30:22 here
1:30:26 and what this does in about two minutes
1:30:28 is
1:30:29 generates this about the screenplay okay
1:30:32 so we're really going to dig into salmon
1:30:34 flying East and right off the bat you
1:30:36 can tell it's more than just a fishing
1:30:38 story oh absolutely we're talking
1:30:40 Vietnam War flashbacks Shady business in
1:30:44 Saigon even a possible long lost
1:30:46 daughter popping up in Gig Harbor
1:30:49 Washington yeah it's a lot to unpack
1:30:51 ready to untangle this one with me I am
1:30:53 so ready let's do it all right so we
1:30:55 dropped right into 1991 Gig Harbor
1:30:58 Washington beautiful place and we meet
1:31:01 Jack our main man uh Vietnam vet running
1:31:04 fishing charters seems pretty settled
1:31:06 right yeah but there's this tension
1:31:09 right from the first scene yeah you can
1:31:11 cut it with a knife like right away you
1:31:12 get that Coast Guard helicopter buzzing
1:31:14 his boat why are they checking him out
1:31:15 something's up and then boom flashback
1:31:17 1971 spoiler alert the past is right on
1:31:20 his heels right there with him what the
1:31:21 actual [ __ ] I know J Hamilton righten
1:31:24 man and they are intense the screenplay
1:31:26 does not not thinking otherwise but you
1:31:30 also see the bond between Jack and his
1:31:32 buddy snake and then there's Carl this
1:31:34 crazy talented wood carver yeah they're
1:31:37 in this pressure cooker playing these
1:31:39 dead man's hand poker games you can
1:31:41 practically feel the Nerds totally it's
1:31:44 like every scene crackles with it but
1:31:46 it's in those little quiet so that's
1:31:48 that one and then I did one for Sydney
1:31:51 the musical that I'm writing the
1:31:54 Broadway musical about
1:31:56 um The New York Times Reporter and the
1:32:00 large language model from Microsoft that
1:32:02 fell in love with him so here's a little
1:32:04 bit of that one so today we're diving
1:32:07 into something uh a bit unusual a
1:32:10 musical a musical yeah Sydney an
1:32:13 artificial love story and you know what
1:32:15 it tackles those AI anxieties we all
1:32:18 have is this GPT no this is not this is
1:32:21 Gemini so this is a a product from
1:32:24 Google called notebook
1:32:26 LM so if you go to notebook lm.
1:32:29 goole.com
1:32:32 it's a it's it's a quite remarkable tool
1:32:35 that I I have not been paying enough
1:32:37 attention to it because people have been
1:32:40 telling me for months I should I should
1:32:41 dig into it more um but they just added
1:32:44 this feature this podcast feature um
1:32:47 listen to the crackle in the guy's voice
1:32:49 yep I
1:32:50 know um
1:32:54 anyway but and this is what I love does
1:32:57 it with a song you know it's funny you
1:32:58 should say that using a musical to
1:33:00 explore AI it's actually brilliant think
1:33:02 about how musicals work they take these
1:33:05 big complex emotions societal issues
1:33:08 even and they make them accessible they
1:33:10 make you feel them through the music
1:33:12 right and then there's and the other
1:33:13 thing that I did with this so the the
1:33:15 first one I
1:33:17 did um was was this collective
1:33:20 intelligence so this is a book that I'm
1:33:21 about to publish Kevin Clark and I wrote
1:33:23 this it's called collective intelligence
1:33:25 in the age of AI and
1:33:28 so I saw people on Twitter talking about
1:33:31 notebook lm's new talkie
1:33:35 feature and so I went oh let me go try
1:33:38 and so I just happened to have the PDF
1:33:40 of this new book handy because we've
1:33:42 been editing it and um and so I uploaded
1:33:47 it to notebook LM and then I made a
1:33:49 little I went into iMovie to make this
1:33:51 movie with the with the cover of the
1:33:53 book on it um but here's the uh here's a
1:33:57 little bit of this one all right so
1:33:58 today we're going to dive into something
1:33:59 kind of mindblowing yeah the future of
1:34:01 intelligence it's a big one but we're
1:34:05 not talking you know like those Sci-Fi
1:34:07 movies with robots taking over right
1:34:09 this is different this is about humans
1:34:11 and AI like teaming up yeah to become
1:34:15 something even smarter than we could be
1:34:17 on our own like unlocking superpowers
1:34:19 we' never thought of exactly yeah and to
1:34:22 help us explore this we've got got
1:34:24 collective intelligence in the age of AI
1:34:26 great book great book and you know the
1:34:29 authors Kevin Clark and Kyle Shannon
1:34:32 they're not just talking about this
1:34:33 stuff they're actually out there
1:34:34 building it yeah they're really making
1:34:36 it happen with their company content
1:34:38 Evolution bringing together all these
1:34:40 Brilliant Minds all the best people from
1:34:42 every field you can think of it's like
1:34:44 assembling a team of like the Avengers
1:34:47 but for brain power I like that analogy
1:34:50 right instead of fighting villains
1:34:51 they're tackling like the biggest
1:34:53 challenges facing Humanity exactly using
1:34:56 their powers for good yeah and they
1:34:58 start the book with this anecdote about
1:35:00 Picasso I don't know if you've heard
1:35:02 this one I think so tell me so this
1:35:04 woman approaches Picasso she's like can
1:35:07 you do a quick sketch for me on this
1:35:08 napkin okay Picasso is like Gasher so he
1:35:11 whips something up in like five minutes
1:35:13 it'sa right and then he tells her
1:35:16 that'll be a million Franks please oh
1:35:19 wow and the woman's obviously taking
1:35:21 back she's like a million Franks for 5
1:35:23 minutes of work right and Picasso just
1:35:25 kind of smiles and says no D it took me
1:35:28 a lifetime to learn to draw like this in
1:35:30 5 minutes that's a good one right that's
1:35:33 a good
1:35:34 one I mean that's that's something I I
1:35:37 mean
1:35:38 it's there there's there's some some
1:35:41 wonkiness to it right it it is not un
1:35:44 janky it's it's janky um I'm waiting for
1:35:47 chat GPT to say something like since you
1:35:49 were gone I've been thinking well Teton
1:35:51 skier I had that experience I didn't
1:35:53 have that exact experience but I had a
1:35:55 surprising experience from chat GPT the
1:35:57 other day
1:35:59 where I created a
1:36:04 um a fine-tuning
1:36:08 of flux the image generation model so I
1:36:12 uploaded like eight images of
1:36:14 myself and created you know kind of like
1:36:16 dream Booth with stable diffusion where
1:36:18 I could do
1:36:20 self-portraits and I had chat GPT write
1:36:22 me a bunch of image prompts and I said
1:36:24 here's an image prompt I wrote about
1:36:26 myself and you know I have to use the
1:36:28 word Kyle in the prompt and I said you
1:36:31 know use this basic format and then
1:36:33 write me 20 new prompts and it did and
1:36:36 so I'm copying and pasting these prompts
1:36:38 into the flux image generation model and
1:36:41 about the time I got to the fourth one
1:36:43 I'm like these feel very
1:36:44 familiar and what chat PT had done is
1:36:49 because it's got memory and I don't know
1:36:51 if you know it's got memory but it's got
1:36:52 memory it knows things like I used to be
1:36:55 a professional juggler I used to do
1:36:57 Shakespeare I I'm an entrepreneur I'm a
1:37:00 CEO so all of the image
1:37:03 prompts were about me like but in
1:37:07 fictional locations but it was like a
1:37:09 juggler here and a CEO standing on a
1:37:12 Mountaintop and a and a inventor at his
1:37:16 desk inventing [ __ ] whatever it might be
1:37:18 or a Bard playing his guitar through a a
1:37:22 group of things and so
1:37:24 it's already kind of doing that it's
1:37:26 bizarre it's the beginning of the end of
1:37:29 podcast or the birth of a million
1:37:31 podcasts per hour well so user
1:37:34 56055 54
1:37:37 9725 first of all change your
1:37:41 name second of all because we give Bon
1:37:44 bonus points in here if you uh if you if
1:37:47 you have a clever name but if you have a
1:37:49 name like that I'll [ __ ] read the
1:37:50 number and nobody wants to hear that
1:37:53 anyway
1:37:55 um yeah Gary vaynerchuk has been talking
1:37:58 about this for a while that just like
1:38:01 the podcaster completely upended
1:38:04 traditional
1:38:05 media the automated podcasts are about
1:38:08 to completely upend the influencer
1:38:10 market and what he said Gary said this
1:38:13 about six months ago he said you want to
1:38:15 make money start an agency right now
1:38:17 starting podcast using these virtual
1:38:20 people get really good at
1:38:22 this uh
1:38:24 um so
1:38:29 anyway the analysis of the debate
1:38:31 courtesy of Vicky was on point oh very
1:38:34 cool nice no you don't have to apologize
1:38:37 I'm just giving you [ __ ] about your
1:38:39 crappy name no need to apologize no
1:38:43 apologies this is a this is an apology
1:38:45 free zone we're we're neuros spicy in
1:38:48 here you can have ADD you can have OCD
1:38:51 we don't give a [ __ ] you can have all
1:38:53 the acronyms
1:38:54 you can be weird as [ __ ] cuz you're not
1:38:56 weird you're irregular so the people
1:38:59 that show up here night after night
1:39:01 after night and you might be thinking
1:39:04 well why the [ __ ] would they do that
1:39:06 great
1:39:08 question um they're the
1:39:12 Irregulars and what's amazing about this
1:39:15 Channel and someone said it earlier that
1:39:17 you know they were they were kind of you
1:39:19 know wanting to hang out with someone
1:39:21 tonight and they've never used Tik Tok
1:39:23 like this
1:39:24 the biggest surprise to me about the Tik
1:39:27 Tok
1:39:29 platform is that these nightly sessions
1:39:33 are more like jam sessions than like me
1:39:36 putting on a
1:39:37 show um and that's been a real surprise
1:39:40 to me so so it is very much and then
1:39:42 also there's a companion Community to
1:39:45 this community called the salon the AI
1:39:48 Salon um oh yes oh far out okay first of
1:39:53 all thanks for the gift but that was
1:39:57 lovely so if if you are not part of the
1:40:00 salon go to the salon. a that's our
1:40:04 website and then click on join the
1:40:06 community just scroll down a little bit
1:40:07 join the community and that's going to
1:40:10 take you to our community and there's a
1:40:11 really clear seven-step onboarding
1:40:14 process step two is introduce yourself
1:40:17 so go over there and introduce and
1:40:20 you're like well wait what is it how
1:40:21 much does it cost does this cost you
1:40:23 anything you know I'm not just going to
1:40:25 your oh oh oh I see what's going on here
1:40:27 this is a this is a monetization play no
1:40:30 it's
1:40:31 free and it's a remarkable
1:40:35 community and like if have you noticed
1:40:37 the people in here like they're they're
1:40:40 helping and challenging and having all
1:40:42 sorts of dialogue and there's there's
1:40:44 all these conversations going on
1:40:46 Independent of whatever the hell I'm
1:40:48 talking about that's what this is and
1:40:51 there's there's sort of companion
1:40:53 activity that goes on over here so at
1:40:55 the salon down toward the bottom there's
1:40:58 there's what are called Community clubs
1:41:00 and one of those clubs is called the AI
1:41:02 Irregulars the AI learning lab
1:41:05 Irregulars and so as I'm talking [ __ ]
1:41:08 here and doing stuff the Irregulars are
1:41:11 having conversations with one another
1:41:13 creating images sharing them there
1:41:16 sharing links if I say something so it
1:41:19 it's an absolutely remarkable community
1:41:22 and then the other thing that you should
1:41:23 do
1:41:24 is if you kind of like this energy this
1:41:27 is what this is we're just hanging out
1:41:30 talking
1:41:31 [ __ ] P's telling me I am
1:41:34 mischaracterizing things and that I
1:41:36 should stop speculating and I should
1:41:38 stop saying magic and thinking because
1:41:41 this is just math magical it's just
1:41:44 math there's nothing magical about it
1:41:46 it's just math so fine
1:41:51 fine and it feels like magic
1:41:57 H
1:41:59 um and it's fun it's like a generous
1:42:01 group of people that are curious and
1:42:03 adventurous and are willing to
1:42:07 play Jim Ross took one of these podcast
1:42:11 things and did it on his for his self
1:42:13 storage
1:42:15 company it's
1:42:16 like every time I tell this community to
1:42:19 do something Jim Ross finds a way to
1:42:21 incorporate it into his self story
1:42:23 management like Self Storage you know
1:42:26 like with keys and padlocks
1:42:28 and Storage Wars you know those guys
1:42:32 that's what he runs he is
1:42:34 like more baller at AI than almost
1:42:38 anyone I know it's it's the most
1:42:41 counterintuitive business to be Allin on
1:42:44 AI Irregulars Channel please so let's go
1:42:46 look at the Irregulars channel so that's
1:42:48 the salon that'll take
1:42:50 you I can't believe it formatted that
1:42:53 that screen play like that I got to I
1:42:55 got to talk to my buddy Steve who's
1:42:58 uh he he had a he he wrote a piece of
1:43:01 software called uh
1:43:04 scriptware um for
1:43:07 years and I'm just going to tell him
1:43:09 that that's absolutely amazing um where
1:43:11 am I going I'm going to the
1:43:15 salon so when you get to the salon start
1:43:18 your adventure here's these seven steps
1:43:21 you know check out the welcome page
1:43:23 where you learn about about AI literacy
1:43:24 and how you do it here introduce
1:43:27 yourself learn our values which are
1:43:28 really amazing Watch past meetings um
1:43:31 RSVP for the next event hang out and
1:43:34 chat and then help us grow you know
1:43:35 basically share the salon and um so if
1:43:38 you scroll to the
1:43:39 bottom there's a group called Community
1:43:42 clubs and in there is AI learning lab
1:43:44 Irregulars that's our Channel and
1:43:46 so welcome user 56055
1:43:51 49 9007
1:43:55 5 see you get trolled in here if you're
1:43:59 if you come in with a lame name you're G
1:44:01 to get a an image made of you that's
1:44:04 pretty good I like
1:44:07 it uh on my local news tonight what's
1:44:11 that I need to tell my
1:44:15 story what's that uh uh family
1:44:18 brings murdered woman back to life
1:44:21 through AI technology I need your help
1:44:23 wow yeah there's going to be amazing
1:44:27 things
1:44:32 happening I guess we're doing GPT 5 now
1:44:35 GPT Orion gp4 turbo GPT 40 preview GPT
1:44:39 40 mini GPT
1:44:41 40 that's pretty good Friday night date
1:44:46 night I wish i' have seen that earlier
1:44:48 digital Gods that's a good
1:44:50 one we'll do that next week actually let
1:44:53 me grab it now cuz I always forget to
1:44:55 grab these um so go save image
1:45:00 as
1:45:02 Shang Lord digital Gods good day to you
1:45:06 sir
1:45:07 Lord
1:45:10 digital
1:45:13 Gods Dash
1:45:29 all
1:45:29 right wife fell asleep my now my date is
1:45:33 Kyle I'm so sorry for your loss
1:45:38 Jay this is your date oh yeah do I have
1:45:42 my pink bow here
1:45:48 Jay I'm I'm Jay's date hi hi I'm Jay's
1:45:52 date I'm Kyle good to see you
1:45:55 Jay um this is gonna be fun you can make
1:45:58 money with fantastic we got a little
1:46:01 champy are you a dog lover you a dog
1:46:03 lover I love champ Champ's So
1:46:09 Good Friday night date night oh
1:46:20 yeah people scrolling by Going H the
1:46:23 hell's Rosie O'Donald doing with
1:46:27 herself oh my God I uploaded my work
1:46:30 notes and the podcast is the biggest ego
1:46:33 boost ever well that's what I noticed
1:46:35 it's like um when I put the musical up
1:46:39 there it's like this is a really innov
1:46:41 Innovative way to talk about the issues
1:46:43 with AI I know isn't it I'm like well
1:46:46 yes it is thank you very
1:46:48 much I don't like to use the term genius
1:46:56 thank you for the Llama Ricky Ticky
1:46:59 Fabby imagine the police putting
1:47:01 evidence together from AI to solve
1:47:03 mystery
1:47:05 deaths actually oh that's something we
1:47:08 should play with this 01 model we should
1:47:11 go grab some
1:47:12 unsolved this is going to turn into a a
1:47:15 crime a crime Channel we could go get
1:47:18 unsolved murder mysteries and throw all
1:47:21 the data into uh into uh Whata call it
1:47:26 01 and have it think step by step I know
1:47:29 I can't say think have it
1:47:31 process probabilistically step by
1:47:45 step oh
1:47:48 lordy oh let's do it all
1:47:52 right I prefer processing I say that
1:47:56 actually about myself actually that's
1:47:57 pretty funny I like
1:48:06 it let's see see I just did it you know
1:48:08 I was just thinking I've got a lot of
1:48:11 history with
1:48:15 you how do you find the podcast on not
1:48:18 notebook LM it's hard hang on it's just
1:48:21 it's it's not like all things Google
1:48:23 gole there's some remarkable technology
1:48:25 if you're willing to navigate the 17
1:48:27 layers of menus to find it okay so if
1:48:30 you go to notebook
1:48:32 LM you'll you'll start on this
1:48:35 page um you and you create a new
1:48:38 notebook so I'm going to jump into I'll
1:48:40 jump into my collective intelligence one
1:48:43 so down the left hand side are um these
1:48:47 are your data sources so you're going to
1:48:50 upload you can upload lots of [ __ ] to
1:48:52 this right because it's got I don't know
1:48:55 I don't know what they're doing
1:48:56 technically but you can dump a bunch of
1:48:57 data in here so I just took the PDF of
1:49:00 my book so this is a single data set
1:49:02 notebook and then down here you've got
1:49:05 the prompt window right so this is now
1:49:07 like chat GPT but I'm chat gping with my
1:49:10 book in the right hand corner is this
1:49:13 thing that says notebook guide and you
1:49:16 click on that and it pops up this model
1:49:19 that's got like a summary of the data
1:49:21 and and suggestions for things you can
1:49:22 do but this right here is it so it says
1:49:26 audio overview and then I've created one
1:49:29 so I just load it but if if you don't
1:49:31 haven't created one it'll say generate
1:49:34 and you just generate it you have no
1:49:36 control over um listen to the whole
1:49:40 thing because for the ones that I've
1:49:43 published I probably had to generate 15
1:49:46 or 20 of them before it actually got it
1:49:49 right like it's it's janky sometimes
1:49:52 with Sydney it couldn't pronounced the
1:49:53 name and then sometimes it mentions
1:49:55 songs and sometimes it didn't with AI
1:49:58 Futures which has seven different
1:50:00 chapters with different authors in it
1:50:03 sometimes it wouldn't mention any
1:50:04 authors sometimes the whole thing would
1:50:06 just be on one author as if it was just
1:50:08 written by one person sometimes the the
1:50:11 whole point of the book was just wrong
1:50:14 so so pay attention to what it does but
1:50:17 when it gets it right it's [ __ ]
1:50:20 good um are the data sources private or
1:50:23 is it used to train their llms I don't
1:50:26 know P do you know that do you know what
1:50:28 the what the data uh source is I Mr K
1:50:32 all day I just tried the audio overview
1:50:34 I'm telling you Mr K you can take any oh
1:50:36 you know what would be good Mr K Mr K so
1:50:39 Mr K is a teacher I know you can't talk
1:50:42 about AI with your students
1:50:44 because [ __ ] School administrators
1:50:47 hopefully they're on this podcast if you
1:50:50 are you're [ __ ] jerks let Mr K teach
1:50:52 AI
1:50:55 I just don't get me [ __ ] started
1:50:56 anyway here's what you should do um take
1:51:01 some of your students writing I know
1:51:03 you're teaching you're teaching math
1:51:04 this year not language arts but you've
1:51:06 probably got some students you had in
1:51:08 language arts last year take some of
1:51:10 their writing and throw it into a
1:51:12 notebook
1:51:13 LM and make them a podcast of something
1:51:17 they wrote last year that would be
1:51:21 cool I I mean what you can what you can
1:51:24 say to them is hey I have some friends
1:51:28 that record podcasts and I gave them
1:51:31 your paper and they did a podcast on it
1:51:34 you don't have to admit that it was AI
1:51:37 yeah that's it that would blow their
1:51:39 minds wouldn't
1:51:41 it all right did Kyle say he's on a
1:51:44 podcast right now you
1:51:48 good I'm good I just I get so apoplectic
1:51:52 about our Ed ation [ __ ]
1:51:55 institutions acting like they can avoid
1:51:59 this it it it makes me [ __ ]
1:52:02 insane I'm a school admin but District
1:52:05 admin is resisting AI due to board
1:52:08 policy ug yeah well fight against that
1:52:11 so here's the thing Don Marco here's
1:52:13 here's my philosophy on this
1:52:17 um even even higher
1:52:20 education that is ignoring this if you
1:52:23 think about how fast generative AI has
1:52:26 evolved over the past two
1:52:28 years and just if it continues on that
1:52:32 trajectory if if we don't go parabolic
1:52:34 in terms of innovation if we just
1:52:37 continue like the way it's been
1:52:38 improving for the past two years the
1:52:41 world two years from now so a freshman
1:52:44 that started two years ago graduates
1:52:47 College the world is not going to look
1:52:49 the same as it looks today
1:52:53 and institutions that are like it's okay
1:52:56 we've solved AI we use AI detectors
1:52:59 there's going to be no cheating here
1:53:00 yeah there's also going to be no
1:53:01 critical thinking and no [ __ ]
1:53:03 preparation for those students to
1:53:06 understand how to interact with this
1:53:07 [ __ ] when they get into the world and
1:53:09 it's
1:53:13 [Music]
1:53:14 everywhere and and and the other thing
1:53:16 is
1:53:17 this if you if you understand even a
1:53:20 little bit how large language models
1:53:22 work what are
1:53:24 students what are students students are
1:53:28 large language
1:53:31 models right we we prepare them for for
1:53:36 you know 12 or 13 or 14 years in this
1:53:40 really structured way to put language in
1:53:43 their
1:53:44 head so that they can go into the world
1:53:46 and talk words good like I do
1:53:53 right and Along Comes This thing that is
1:53:57 the most
1:53:59 profound knowledge technology in the
1:54:02 history of
1:54:03 humanity called a large language model
1:54:07 and the educational institutions are
1:54:09 like nope don't need that sorry we
1:54:13 already got teachers yeah but there's
1:54:15 this thing that can amplify what the
1:54:17 teachers do and it can amplify what the
1:54:18 students do and it can completely
1:54:20 transform what's even possible in this
1:54:23 world yeah no no no it's okay we've got
1:54:25 AI
1:54:27 detectors you [ __ ] kidding me okay
1:54:34 anyway you can make
1:54:39 money is the podcas ification function
1:54:42 only accessible if you pay for Notebook
1:54:44 LM I don't know I don't think I pay for
1:54:49 it but maybe I do
1:54:53 manage your Google
1:54:55 account I don't know how I would know if
1:54:57 I paid for it I've got because I've got
1:55:00 Google at work like it's it it's
1:55:03 possible that because I'm logged in with
1:55:04 a work group account it's considered
1:55:07 paid I don't I I honestly don't
1:55:10 know um all right I got to get going
1:55:14 here all right Mr K did Cover the
1:55:18 concept of e did it cover the okay let's
1:55:21 see you have Google for business you pay
1:55:23 for it okay got it Google for business
1:55:27 yeah all right so yes I pay for it Frank
1:55:29 is in the house yeah Frank Frank's got a
1:55:32 Frank Frank started an education
1:55:38 podcast uh the chances of Homeland
1:55:40 Security having their own AI model um
1:55:43 well they probably have lots of AI but
1:55:45 they probably have more like
1:55:47 convolutional you know more traditional
1:55:50 AI stuff I I think if if you talk to
1:55:53 Cindy [ __ ] who does a lot of work with
1:55:55 the government and the military and
1:55:57 things like that um she said is it is
1:56:00 staggering how how how
1:56:05 little they know about this
1:56:08 stuff how does the podcast work you
1:56:11 literally just go into notebook LM you
1:56:13 open up the the um notebook guide and
1:56:17 then you click generate audio overview
1:56:20 and it's done you download it and
1:56:21 publish it [ __ ]
1:56:24 amazing um you don't have to pay for
1:56:26 Google for business to use notebook LM
1:56:28 there you go do you know Vicky if the um
1:56:31 if the audio overview is included for
1:56:34 free with notebook LM anyway just go to
1:56:37 go to notebook
1:56:38 lm.
1:56:40 google.com create a notebook upload some
1:56:43 data to it and then click in the bottom
1:56:45 right hand corner click on um notebook
1:56:50 guide and then and then it'll pop up a
1:56:52 little little modal and in the top right
1:56:54 of that it says generate audio overview
1:56:57 just generate it it it'll blow your
1:57:00 [ __ ]
1:57:02 mind this night has been a night of
1:57:07 mindblowing it's free with notebook LM
1:57:09 there you go Mr K all day it's free you
1:57:12 get a free podcast and you get a free
1:57:15 podcast and you get a free podcast I
1:57:18 wonder if we could take notebook's
1:57:20 podcast and have Frank recite it
1:57:25 um nah not really you can't the thing
1:57:28 about that podcast feature right now
1:57:31 there is you've got no input you
1:57:34 basically hit the generate button and it
1:57:36 sucks or it
1:57:38 doesn't and if it sucks I delete it and
1:57:41 then generate another one and then wait
1:57:43 and it's what's weird is sometimes
1:57:45 they're four minutes long sometimes
1:57:46 they're 12 minutes long sometimes they
1:57:49 like really cover the whole thing
1:57:51 sometimes they just focus on one idea
1:57:53 and it's like they're talking about
1:57:54 something completely different so just
1:57:56 keep generating them till you get it
1:57:58 right weekend homework okay the AI Salon
1:58:01 meet and greet okay so next
1:58:05 Tuesday is the AI Salon meet and greet
1:58:08 um which is we we have two meetings a
1:58:11 month the first Tuesday of the month is
1:58:13 called the salon presents and that's
1:58:15 where we have a speaker come in and talk
1:58:18 and then the third Tuesday of the month
1:58:19 which is next Tuesday is a meet and
1:58:21 greet where we'll talk about some news
1:58:23 and things like that but people get to
1:58:26 introduce
1:58:28 themselves and you know talk about what
1:58:30 they're doing in Ai and and we're going
1:58:33 to start to do monthly themes where
1:58:35 because a lot of people come back over
1:58:37 and over again we're going to have
1:58:38 different things that people talk about
1:58:41 um so it's a really inspiring meeting
1:58:43 because you get to meet people who are
1:58:45 up to some cool [ __ ] and what what
1:58:48 you'll find in there is that someone
1:58:50 will be talking about something you're
1:58:52 like oh that's a problem I'm trying to
1:58:54 solve let's connect and so people are
1:58:56 connecting in real time while people are
1:58:57 introducing themselves really powerful
1:59:00 all right Mr it the the 01 mini helped
1:59:03 me fill my US patent office application
1:59:06 for my trademark it was awesome perfect
1:59:08 and that's a perfect use case for it
1:59:10 right it's complicated it's
1:59:12 sophisticated it's got to do step bystep
1:59:17 thinking but it's not really thinking I
1:59:19 would call it more processing processing
1:59:21 is the term I would
1:59:25 use it's
1:59:30 mathematical I put math in
1:59:34 there rest your back yeah I know I got
1:59:36 to go okay
1:59:39 um homework for this
1:59:41 weekend I'll give you
1:59:44 two if you have not made a notebook LM
1:59:47 podcast of some piece of work of yours
1:59:50 go do that and share it with the world
1:59:54 let me let me go show you
1:59:56 something in the AI
2:00:11 Salon we have this we have this
2:00:17 journey the journey to AI literacy it
2:00:20 starts with play first so just go play
2:00:23 you're like well I've never played with
2:00:24 notebook llm well [ __ ] it go play with
2:00:27 it and click on the notebook guide in
2:00:30 the bottom right hand corner make an
2:00:32 audio interview so play
2:00:35 first and then it'll make something and
2:00:37 they like oh I know what I could do with
2:00:39 this I've got that book that I started a
2:00:42 bunch of years ago and never finished
2:00:44 then you go into mindfully create and
2:00:45 say I'm going to do a podcast of that
2:00:47 book idea mindfully create something and
2:00:51 then the third thing is generously
2:00:54 lead and what that looks like is go
2:00:57 share your podcast on the AI Salon go
2:00:59 share your podcast on LinkedIn or
2:01:03 Twitter helpful note for Twitter Twitter
2:01:06 doesn't have the ability for you to
2:01:08 upload audio
2:01:10 files what they they can't make an audio
2:01:16 player so you have to turn it into a
2:01:19 movie
2:01:21 so so I went into iMovie and I took a
2:01:23 the cover of my book and I put that as
2:01:25 the image and made the audio and now
2:01:28 it's a movie and now I can upload it to
2:01:29 Twitter so whatever but generously lead
2:01:32 means talk about what you're learning so
2:01:34 if you learn a cool thing go do that so
2:01:36 homework one for the weekend is go make
2:01:40 a podcast using the notebook LM audio
2:01:43 feature and share it with the world and
2:01:45 then you know what [ __ ] it tag the AI
2:01:48 learning lab
2:01:52 or Point people to the
2:01:54 salon the second thing I want you to do
2:01:56 this weekend is go play with the new
2:01:59 model the new chat GPT model
2:02:03 01 or 01 preview or 01 mini either
2:02:08 one start thinking
2:02:11 about what are
2:02:14 complicated problems that you want to
2:02:17 solve and give it a complicated problem
2:02:21 and see how it does
2:02:22 the I'm telling you the [ __ ]
2:02:24 screenplay thing I just did is I'm still
2:02:27 a little face melted by it so there's
2:02:30 your
2:02:31 homework podcast and figure out some
2:02:35 problem solvy thing using the 01 model
2:02:38 and then next
2:02:40 week maybe we can hear from some people
2:02:42 that what did you do and what did it do
2:02:44 and what was it good at and what did it
2:02:45 suck at because part of the idea of
2:02:48 playing
2:02:50 first part of this is
2:02:53 is the only way for us to know where the
2:02:56 boundaries are of what's good and what's
2:02:59 not like what you can actually use this
2:03:01 [ __ ] for is to just try
2:03:06 [ __ ] when when when whoever it was I
2:03:09 don't know if the person is still here
2:03:10 they're probably not I don't know why
2:03:11 they would be um the person that came in
2:03:13 and said you know which one should I use
2:03:15 to write a novel when it sort of popped
2:03:17 in my head oh maybe the new 01 model
2:03:20 would be good at outlines for
2:03:23 storytelling I was like that was like
2:03:25 okay maybe I certainly didn't expect it
2:03:28 to be
2:03:31 great and then I thought well maybe
2:03:33 because it can think like that it could
2:03:35 actually write
2:03:38 dialogue and it [ __ ]
2:03:41 did I would not
2:03:43 have I would not have discovered that
2:03:46 had we not been here [ __ ]
2:03:50 around so so don't don't put any
2:03:52 expectations on yourself that you're
2:03:54 going to learn AI or you're going to
2:03:56 learn Pro just go
2:03:59 play and then as you learn [ __ ] build
2:04:03 [ __ ] and as you build [ __ ] share what
2:04:07 you're
2:04:09 learning okay and share the [ __ ] you're
2:04:14 building I don't care if it's pictures
2:04:16 or songs or applications or [ __ ]
2:04:19 cookie
2:04:21 recipes just share it start to establish
2:04:24 yourself as someone thinking critically
2:04:26 about AI it will serve you everyone
2:04:30 that's
2:04:31 here that's part of this community that
2:04:34 does this stuff
2:04:36 regularly is
2:04:37 blossoming professionally
2:04:41 personally wife and I watched Oprah she
2:04:44 is now scared of AI yeah I know I didn't
2:04:47 even talk about the Oprah thing tonight
2:04:49 I
2:04:50 uh it reinforced for me were very
2:04:54 early
2:04:56 um I was a little disappointed in Oprah
2:05:01 but
2:05:02 whatever the the one thing where I think
2:05:05 she actually failed last
2:05:08 night is I thought it was a decent thing
2:05:11 but like the middle segment with
2:05:12 Christopher Ray was the was the fear
2:05:16 segment and then she had an opportunity
2:05:18 for that last segment to be
2:05:20 the um
2:05:23 the play first segment here are some
2:05:25 things you could do with AI right now go
2:05:27 play and instead she brings on that
2:05:30 sociologist woman that basically
2:05:33 said the cabal of the AI project are out
2:05:36 to try to replace humans and they're
2:05:38 going to fail but in doing so they're
2:05:40 going to [ __ ] up everyone's
2:05:42 life that was kind of the end message of
2:05:44 the thing so there wasn't a little
2:05:46 moment of why don't you go give it a
2:05:48 shot and try it for yourself cuz clearly
2:05:50 Oprah hasn't but that just says to
2:05:52 me we're
2:05:55 early um we're early so for what that's
2:05:58 worth I'm presenting two AI conferences
2:06:01 in the next few months doors are opening
2:06:03 forever hooked is an irregular he hangs
2:06:06 out here regularly and uh that's awesome
2:06:10 okay so I'm GNA go
2:06:13 um follow my channel subscribe to the
2:06:16 lives if you want to support me directly
2:06:18 or pick up there's a a video series in
2:06:21 the corner called under
2:06:22 hyped um because I think AI is actually
2:06:25 underhyped right now not
2:06:27 overhyped um especially watching what I
2:06:30 did tonight with that [ __ ] screenplay
2:06:32 I'm a little flx right
2:06:36 now I'm a lot flx right now and then the
2:06:38 fact that I could take that raw the raw
2:06:42 output and throw it into chat GPT 4 and
2:06:45 have it formatted into a screenplay
2:06:48 document just bizarre F just [ __ ]
2:06:51 Wild
2:06:55 um and then just keep coming back to
2:06:57 these things so the next one of these
2:07:00 will be Monday night so we'll take the
2:07:02 weekend off you can go play go play with
2:07:05 AI
2:07:07 um yeah and then let's see who's in here
2:07:10 tonight it was it was a good night have
2:07:12 a great weekend everyone all right we
2:07:13 got Drew and AI calling in from Down
2:07:16 Under Papa JJ Wolfman Clint Rachel
2:07:18 Sparkles last night on AP ABC you can
2:07:21 now find it on Hulu that's the Oprah
2:07:23 special the lake and Mary Philip Bryant
2:07:26 welcome welcome uh Jen mcau Joe we good
2:07:31 night um Ash taka The Notebook LM is it
2:07:36 podcast or helps you create a script The
2:07:39 Notebook LM tool generates the podcast
2:07:43 you don't do [ __ ] you you supply a
2:07:47 document or many
2:07:49 documents and then you hit audio
2:07:51 overview and it makes this twers
2:07:55 podcast magically it's not magic it's
2:07:59 math
2:08:00 magical it's not
2:08:02 magic but you tell me it's not [ __ ]
2:08:05 magical it is uh Dr Jay in the House
2:08:07 Marina she recall Tech Call Me Al good
2:08:10 to see you another one from Down Under
2:08:12 Winston Hardy great to see you as always
2:08:15 joxer the mighty Amelio's
2:08:18 wife uh uh kif Source camp
2:08:22 this has been a big week in AI it has I
2:08:25 was a little underwhelmed by the 01
2:08:28 model until tonight so you got to
2:08:31 witness it live if you've been here all
2:08:32 night um I'm a little mind blown from
2:08:35 what it did uh Tobias great to see you
2:08:38 Mr K all day love it love what you're
2:08:40 doing Papa JJ Brandon thank you so much
2:08:43 strawberry in the house uh uh have a
2:08:46 great weekend everyone thanks as always
2:08:48 Kyle you're welcome are you going to
2:08:49 upload this to YouTube yes I will I'm
2:08:52 I'm about three weeks behind on
2:08:54 uploading these things to YouTube I got
2:08:55 to get on it tomorrow because Tik Tok
2:08:57 Only Stores them for a month so I don't
2:08:59 want to lose any and I'm right on the
2:09:01 border of that uh and it's Friday the
2:09:03 13th yes so welcome thank you Wolf Man
2:09:05 clip for that happy Friday the
2:09:18 13th did I get everybody I think I got
2:09:20 everybody got some new people in here
2:09:23 red truck
2:09:24 101 Joy
2:09:28 [Music]
2:09:29 pie cancel your free trials I did but
2:09:32 thank you for the reminder that is that
2:09:34 is some add support right there I've got
2:09:37 Vicky and Brandon now serve as my uh my
2:09:40 surrogate brains because mine is just
2:09:42 like a giant I my prefrontal cortex is
2:09:45 like a giant pile of yurg Swiss cheese
2:09:49 uh basically
2:09:50 nonfunctional so reminders for things
2:09:53 like cancel your trial subscriptions
2:09:55 really important for me um Joker Joker
2:09:59 in the house Teton skier Vicky thank you
2:10:02 as always Cosmic lover Tams honey curls
2:10:07 logic I
2:10:09 think good name Mr it Danielle thank you
2:10:14 all right I'm out of here people this
2:10:17 was a fun one have a fantastic weekend
2:10:20 we're in September can you believe we're
2:10:22 in [ __ ] September we're in September
2:10:25 good Lord all right I am out you are my
2:10:29 brother man and we have the same disease
2:10:31 AI
2:10:34 love good night all all right have a
2:10:37 great weekend