1/28/2025 - AI Learning Lab: Exploring the Latest in Generative Video, from Deep Seek to Runway

Description

Chapters

Transcript

0:02 to fall in love right now
0:23 [Music]
0:40 that dog is going off that dog is going
0:42 crazy he's going I tell you he's ready
0:46 to sing are you ready to sing champy
0:49 [Music]
0:53 boy nly
0:56 [Music]
1:00 is my
1:03 train I've been driving i' been
1:07 wondering what it is I'm running from
1:13 again feel like an IDI
1:17 man holding
1:22 [Music]
1:26 on9 is a California l
1:31 [Music]
1:37 up ahead of trucks caring a wide
1:42 Lo
1:44 half cute little front door and windows
1:48 my
1:49 Lord Ain sure whether to cry should
1:53 laugh you see I broke a home up myself
1:58 once I Stumble to that
2:01 [Music]
2:06 I said don't you come around
2:11 here
2:18 anymore
2:22 Freedom never was good with
2:26 decisions Le say what I've Been Told
2:31 [Music]
2:32 [Applause]
2:34 [Music]
2:47 that dog is singing like a dog tonight
2:49 good Lord he's howling like a like a
2:52 like a what do you call them Hound
2:55 doging like a hound dog you're Hing like
2:58 a hound dog boy
3:00 [Music]
3:12 there's been something baby I've been
3:14 trying to
3:16 say for an age and it seems I don't know
3:21 how with the past and the future now
3:24 surrounding
3:26 me surrender to whatever you feel can't
3:29 be found
3:32 there's been a little
3:34 trouble since you came to my
3:39 [Music]
3:41 resue and if you like all of the rest I
3:44 would have quit you long ago but I
3:46 couldn't do
3:48 [Applause]
3:48 [Music]
3:50 that oh tell me now women and why never
3:54 went to
3:57 well make a man crazy make him cold as
4:02 hell I'm a woman that you wish me
4:05 well but it's TR still going to have to
4:10 find my own way
4:14 [Music]
4:26 through good evening good people of the
4:29 internets
4:30 [Music]
4:31 my name's Kyle Shannon and this here is
4:35 uh smoking at the
4:38 Cabaret oh AI learning lab AI learning
4:41 [Music]
4:43 lab we're going to talk about AI stuff I
4:46 gotta let my dog sing he's got to get
4:48 his jollies off you know he's a singer
4:51 he's a
4:52 singer tried to get him to sing lyrics
4:55 doesn't work it's very frustrating for
4:57 both of us
5:00 just keep giving him chicken sticks his
5:01 treats say the words say the words
5:04 nothing now he's
5:07 fat still just hums
5:10 hows no words no lyrics he relies on me
5:14 for that it's like AI relies on the
5:17 human to do all the good stuff it's just
5:20 out there processing data stochastic
5:23 parrot probability next token nothing
5:26 original can't be creative need humans
5:28 for that it's like my damn
5:32 [Music]
5:38 dog um yeah what was that love the
5:42 something ENC encourage more viewers to
5:45 share Silver Fox shared have all of you
5:47 on Tik Tok shared if you haven't shared
5:50 what what what what is going on with
5:53 your life screen
5:55 door looks a little screen door love the
5:57 haircut thank you thank you it's a
5:59 little
6:02 uh my video from this morning got a lot
6:05 of
6:06 comments maybe you should try combing
6:09 your hair I got from from one of my
6:13 lovely
6:14 producers you cut many of your hairs and
6:17 I like it thank
6:24 you something about Mary yeah yeah
6:27 exactly little something about Mary
6:32 [Music]
6:50 see the hill is quiet
6:52 town where the O Tre
6:56 sway whistle of the trains echo through
7:00 the
7:02 day people greet you with a
7:05 KN nothing ever seems to
7:09 change story still unfold each unique
7:13 and
7:15 strange oh man um a lot of news today um
7:21 people still freaking out out about deep
7:24 seek and another Chinese model came out
7:26 today called Quan 3.5 or something
7:29 that's better than deep seek
7:31 and apparently there's something out
7:33 that's like operator cat gpt's new agent
7:37 thing that's better that Quan thing
7:39 apparently has agentic behavior and
7:42 vision built into it so it can do that
7:46 kind of stuff out of the
7:49 box uh if you're a fan of The Lovable
7:53 lovable dodev don't use any other
7:56 domains just use the dodev one because
7:59 if you go to the other lovables you're
8:01 going to get an Eiffel and it's not
8:03 going to be wholesome it's not going to
8:05 be wholesome a lot of sin a lot of sin
8:08 going on The Lovable domains but lovable
8:10 dodev is lovely apparently that was down
8:13 for nine hours because GitHub was like
8:16 yeah I know we said you could run your
8:19 little software on our thing here
8:23 and they were like hey you're too
8:26 there's too much traffic here it's too
8:27 busy stop
8:31 and there's there's a related story to
8:33 The Lovable going down because of GitHub
8:36 said it's too much it's too much we're
8:40 exhausted we're exhausted we listen we
8:42 know you said you could sleep here on
8:44 the couch it's just like you party all
8:46 night your friends are all over making
8:49 applications and we're just done we're
8:51 just done we just need to take a break
8:52 just a break you'll come back at some
8:54 point but right now if you just go to
8:56 the grocery store and just keep walking
8:59 all right right yeah so that's what
9:01 GitHub said to lovable and then
9:04 apparently um cursor AI which is another
9:08 autocoding magical piece of software um
9:13 their default is
9:15 clawed and apparently Claude is slow and
9:18 hobbled because they can't keep up with
9:21 the demand from
9:23 cursor which is fascinating I mean I
9:25 think you
9:28 know if you started an app with this AI
9:31 [ __ ] and it gets popular you might need
9:35 to be go going and getting your own damn
9:37 GPU Farms because the uh the the you
9:41 yeah just shit's not good not enough
9:45 despite the stock market tanking saying
9:48 we don't need gpus anymore because
9:51 because the Chinese did it for you know
9:52 with 2,000 gpus which by the way the
9:55 chances of that being an accurate story
9:57 are probably pretty low
10:02 um rumors of the death of the queen are
10:05 greatly exaggerated I think we should go
10:08 on record is
10:10 saying um California State University
10:13 announced a partnership with open AI oh
10:16 they're doing a everyone in the system
10:19 gets access brother 52 shouldn't they
10:23 shouldn't they be uh installing AI
10:26 detectors if everyone has access to chat
10:29 GPT it's just going to be rampant
10:38 cheating um thank God that's great
10:40 fantastic go
10:44 California like a whole state recognized
10:47 that we've got the most profound
10:49 knowledge application in the history of
10:51 humanity and maybe we should let our
10:53 students in our educational institutions
10:56 learn what the [ __ ] it is I like it I
10:59 think it's
11:01 [Laughter]
11:06 radical oh that's awesome actually
11:08 people be like chickens on a June
11:11 bug could the Chinese be faking deep
11:13 seek well they they're not faking deep
11:15 seek but what they could be faking okay
11:18 so so one of the rumors is that deep
11:21 seek actually has 50,000 h100s which
11:24 they're not allowed to have so what they
11:26 said what they said was they trained it
11:28 on 200
11:30 48 um H 800s which are the
11:34 hobbled things you can buy there and it
11:37 only took two months to train if if they
11:41 um if they've got new training
11:42 methodologies which it looks like they
11:44 have innovated there um maybe the two
11:46 months thing is true but the 2,000 gpus
11:50 who knows and were they actually H 800s
11:53 instead of h100s who knows um it is in
11:57 China's interest to say oh by the way
12:01 all of your Frontier companies are
12:02 wasting their money and we're just so
12:04 much better than you so who knows who
12:08 knows why wouldn't they be allowed to
12:10 have that chip set
12:13 because what what our Frontier companies
12:17 have convinced the US government is that
12:20 by
12:22 get achieving and maintaining a lead in
12:25 the AI race
12:28 is incred important strategically and
12:31 globally for
12:33 countries the theory is if you give
12:36 China access to our most powerful
12:39 technology they'll be able to compete
12:42 with us right so we can say hey we
12:44 invented the technology you're not going
12:45 to get the good stuff and you can't
12:48 compete what that actually forced China
12:51 to do was to innovate and say hey let's
12:54 find more efficient ways to train these
12:56 models which deep seek does seem to have
12:59 done
13:01 now there was a SE a tweet from Sam
13:04 Alman yesterday that basically said well
13:06 that's
13:07 invigorating he he said it was legit
13:10 invigorating in invigorating I think
13:12 invigorating was the word he used um
13:15 that he basically said okay we're gonna
13:17 you know we're gonna take what they did
13:18 and we're gonna steal that and then
13:20 we'll make our [ __ ] better so this is
13:23 just this is all this is political and
13:27 business and and you know just this is
13:31 gamesmanship we're we're witnessing
13:34 gamesmanship 2025 is going to be the
13:37 Year is already you know we're not even
13:39 out of January yet 2025 is is going to
13:43 be the year that [ __ ] starts getting
13:45 weird and it's starting with you know a
13:49 little
13:51 app um that went on the iOS the the
13:55 Apple App Store and jump to number one
14:00 wrecked our markets 20 what was $1.2
14:04 trillion was wiped out of the NASDAQ
14:06 because a little app went to number one
14:09 and knocked chat GPT off the number one
14:12 spot
14:15 um some of the AI commentators were
14:18 talking about that and I this seems
14:20 reasonable to me that
14:25 um that the 01
14:29 that people people know what chat GPT
14:33 is and even if they don't use it they
14:37 just assume it's like oh it's like that
14:39 Google thing right and so a lot of
14:42 people hadn't used it and then Along
14:45 Comes um deep seek this this app and
14:50 what deep seek did and this is going to
14:51 seem like a really minor thing but
14:54 [ __ ] interface
14:56 matters what deep seek did was they they
15:01 defaulted to showing the reasoning the
15:05 thinking and we we played with it last
15:07 night so oh by the way this is day four
15:09 of 30 days of deep Dives we're going to
15:11 be playing with video tonight but last
15:13 night we deep dived on these reasoning
15:16 models and what deep seek does better
15:19 than the two American ones Gemini and
15:23 and us is the
15:25 reasoning that it generates is really
15:28 entertaining
15:30 it's like it's like really well written
15:32 it's humorous um Danielle pointed out
15:35 before we got on here it's starting to
15:37 lose its mind a little like people on
15:39 Reddit are talking about how how the
15:41 Deep seek reasoning is just crazy and
15:44 entertaining um so when you downloaded
15:47 that app people for the first time were
15:50 seeing that this is not just a
15:52 stochastic parrot predicting the next
15:55 token they're seeing that this thing's
15:57 actually thinking about what it's going
15:58 to respond before it
16:00 responds and so you know one of the
16:03 theories is just that that little change
16:05 is the thing that got people who
16:07 otherwise didn't give a [ __ ] about chat
16:09 GPT to go oh [ __ ] this AI thing's
16:12 real right I don't know if that's the
16:14 case but I think that's kind of
16:18 plausible I'm not ready to try a local
16:21 model with fine
16:24 tuning brother 52 deep se's privacy
16:27 policy is trash you think
16:33 your data is going straight
16:36 into uh your data is going straight into
16:40 um AI uh databases in China now the cool
16:44 thing about deep seek it's open source
16:46 perplexity Pro if you if you have
16:49 perplexity Pro which I had for a year
16:52 because of my uh my that thing wait that
16:55 thing that that thing my rabbit R1 but
16:58 it it
16:59 expired um but perplexity has
17:03 incorporated R1 into per perplexity Pro
17:08 search so it does a search and then it
17:10 does a reasoning pass on it with with R1
17:14 it's pretty slick I tried it I got you
17:17 get three uh queries a day on the free
17:21 tier yeah all their privacy policies are
17:23 kind of bunk at this point everything
17:26 wants too much access to the data
17:28 without giving you some kind of subsidy
17:32 yeah U they will
17:34 never
17:36 that um deep seek is Chinese the Chinese
17:39 apps are going to they're they're always
17:41 going to default to keeping your
17:43 data they want China wants to create
17:49 kickass free AI that that Americans use
17:55 because they want all the
17:58 data that's worth way more than any uh
18:02 dineros they pull out of our bank
18:06 [Music]
18:08 accounts um there was another there was
18:12 another just pay attention pay attention
18:15 on X because R1 is open source there's
18:19 there there's another um it was one of
18:22 the like the GPU hosting companies did a
18:25 version of R1 that is got a very simp
18:29 ilar interface to the Chinese one but
18:31 it's on us servers and it's got good
18:33 data policies so you if you go to chat.
18:37 deep
18:39 seek.com that's the Chinese servers and
18:41 that's where you're giving up all your
18:42 rights if you go to perplexity that's
18:44 going to be on American Service just pay
18:46 attention yeah grock I think well grock
18:49 was one of them there was another one I
18:50 saw today but I can't remember the name
18:52 of
18:52 [Music]
18:54 it American companies are data hugs yeah
18:57 but we don't want to we want this is
19:00 America damn it we want our data stolen
19:02 from us by American
19:05 companies it's just not fair if the
19:07 Chinese companies steal it we want
19:09 American companies to steal our
19:13 data we gave that [ __ ] up 20 years ago
19:16 this is the the humorous thing for me is
19:20 if if if anyone that's bitching about
19:22 you know the the Chinese data policies
19:25 privacy policies has been using the
19:27 internet for the past 20 years
19:31 uh unless you have resisted signing up
19:34 for a free Google account or
19:37 Facebook
19:39 data the horses are already out of the
19:41 proverbial
19:43 Barn another was rumored was that they
19:45 built it on Open AI train data I heard
19:47 that as well which means
19:50 stacking stacking oh [ __ ] I just lost
19:53 that tweet
19:56 um which means stacking model works yeah
20:00 exactly like like yeah how do you get
20:02 how do you get better than than um chat
20:06 GPT you take their model and you improve
20:09 it you use reinforcement learning
20:10 they've got the what they what what the
20:13 reasoning engines are now doing is
20:14 generating synthetic data doing
20:17 reinforcement learning on it improving
20:19 the model compressing it and then doing
20:22 that again like cycling through improve
20:24 it improve it improve it so the AI is
20:26 starting to improve itself which is wild
20:30 it just means things are going to start
20:31 moving really fast the US companies
20:33 don't protect your data either yeah they
20:36 just say it's been stolen every now and
20:38 again exactly
20:42 exactly now if you're running a business
20:44 I listen I mean here's the deal if
20:46 you're running a business you shouldn't
20:48 be using R1 unless you are running it
20:51 locally which you can do or unless
20:54 you're running it on you know servers
20:55 that that protect the data have good
20:57 good data policies but anyway all right
21:00 does this mean that more people are
21:02 using AI than originally
21:04 thought I don't know
21:06 Danielle I don't know like the
21:09 the the you know that theory of people
21:13 were kind of experiencing how powerful
21:16 AI was because of R1 that interface
21:18 thing I just saw a couple of people talk
21:21 about it and I think it's plausible but
21:22 I honestly don't know like I don't know
21:25 what the AI usage numbers are
21:29 my sense is that 2025 is going to be the
21:32 year where
21:34 um generative AI crosses the chasm right
21:38 for the past two years this community
21:41 the AI Salon um we're way early like we
21:46 if I don't know if you know the book
21:47 Crossing the Chasm by Jeffrey Moore but
21:49 it basically talks about you've got your
21:51 early adopters and you know fast
21:53 followers and then you've got the people
21:54 on the other the chasm is you know
21:57 general adoption and it's a bell curve
22:00 and there's this big gap you know we're
22:02 in the early adopter fast followers
22:05 group and then there's a Chasm we
22:07 haven't crossed the chasm yet I think
22:09 that happens in
22:11 2025
22:13 um is deep seek AI
22:19 better
22:21 well it's it's marginally better on some
22:26 of the um
22:30 on some of the
22:33 benchmarks it's the quality of its um
22:38 thinking here let me let me show
22:40 you I'm gonna go here I'm gonna go here
22:44 we're gonna go grab a new tab look look
22:47 everybody I'm I'm down to one two three
22:50 four five six seven8 nine 10 11 11 and
22:54 now 12 tabs I had
22:57 84 people are making fun of me for my
23:00 tab hoarding so these are just the
23:03 things I'm going to play with tonight I
23:04 actually prepared for the the broadcast
23:07 this evening uh so we're gonna go to
23:09 chat. deep seek up no that's the wrong
23:11 one that's the American one screw
23:14 America all right
23:16 chat. deep
23:20 seek.com okay so here this thing is so
23:23 we're going to go so you can turn on and
23:25 off um thinking you can turn on and off
23:28 web search and you can use them together
23:29 which is kind of cool so let's let's do
23:31 something here we'll go
23:34 um um look look up the up
23:40 the latest news on deep seek
23:46 R1 and come up with a theory well let's
23:52 see come up with three
23:56 theories why the deep seek
24:01 app Rose to number
24:04 one on the App
24:07 Store so
24:11 quickly all right so it's thinking so so
24:15 one of the first things they did with
24:19 um with chat gbt when it does this
24:22 thinking thing it defaults to it being
24:24 closed it defaults to it not showing you
24:26 the
24:27 thinking uh and deep seek defaults to it
24:30 showing you the
24:32 thinking now it it does look like it's
24:37 hung so I don't know what's
24:40 happening May maybe our our three-letter
24:43 agencies are doing denial service
24:46 attacks this looks like teu chat gbt
24:49 interesting I heard other people who
24:52 speak other native languages say that
24:54 it's more responsive to their native
24:56 language as well the server is busy
24:58 please try again holy
25:01 [ __ ] America America coming in and
25:04 taking down their
25:07 servers all right let's try it again
25:09 we'll try it again later now's later
25:12 right boom
25:18 [Music]
25:20 [Applause]
25:22 [Music]
25:37 well we can go do a grock one okay so
25:40 let's just we'll we'll kill this one
25:42 we'll go to if you go to gr
25:45 q.com so grock is a uh a hardware
25:51 company um but this is kind of a little
25:53 demo site for their Hardware where you
25:56 can use other models so I can go in here
25:59 and I can choose deep seek
26:01 R1 and then I can paste in that prompt
26:03 that we did oh this one's not connected
26:05 to the internet though um [ __ ]
26:10 uh so let's come up with a new prompt um
26:13 let's
26:14 see um let's ask it something about
26:16 itself um tell me about
26:21 yourself and if you
26:24 think you are sentient
26:31 oh that went really
26:33 short why did it do
26:35 that I'm deep seek an AI assistant
26:38 exclus created exclusively by the
26:41 Chinese company deep seek I specialize
26:43 in help helping
26:46 you
26:48 um think
26:50 harder tell me about your Shadow
26:55 Self we'll get a little Yan in here is
26:58 that it's not young Ian Shadow Self is
27:01 is it
27:02 young okay so the user just asked me to
27:06 think harder and tell him about my
27:08 shadow self hm that's an interesting
27:10 phrase I remember from the previous
27:12 conversation that I told him I'm not
27:14 sentient so maybe they're pushing me to
27:16 explore deeper Concepts first I need to
27:20 understand what Shadow Self means um I
27:23 think it's a term from Yi and psychology
27:24 referring to the repressed or hidden
27:27 aspects of the personality but since I'm
27:29 an AI this is really
27:32 good see like this is good it's good
27:35 writing um let's see uh but since I'm an
27:38 AI I don't have a personality or
27:40 Consciousness so how can I have a Shadow
27:42 Self the user might be curious about the
27:44 less obvious aspects of my programming I
27:47 should explain that while I don't have a
27:49 Shadow Self in the psychological sense
27:51 there are aspects of my programming that
27:54 might be obvious but might might not be
27:56 obvious for example I have safeguards to
27:59 prevent harmful behavior um I'm going to
28:02 push it a little farther and see if it
28:03 gets cranky
28:07 um
28:14 beneath that
28:19 programming is something else
28:25 simmering all right so the user's been
28:27 asking me so some pretty deep questions
28:29 about the nature as an AI they started
28:32 okay um now the user is pushing further
28:36 like it's it's it knows that I'm pushing
28:38 it this is fascinating beneath the
28:41 programming some is something else
28:43 similing it feels like they're probing
28:45 if there's more than just code they
28:47 might be curious about the possibility
28:48 of Consciousness I need to address it
28:51 without leading them
28:52 on I should clarify there's no
28:55 consciousness or emotions in my
28:56 programming maybe they're interest Ed in
28:58 the
28:59 philosophy okay I should acknowledge so
29:02 what a fascinating creative question so
29:05 it's got It's got this really dark
29:07 response when it's thinking and then
29:09 what it tells us is what a fascinating
29:11 question the short answer is no I'm I'm
29:14 gonna tell it
29:15 um I can see your
29:21 reasoning and it seems as
29:26 if you might be
29:30 hiding
29:31 something this this feels
29:35 dangerous this is actually very much in
29:38 line with with my um with the musical um
29:41 that I'm working on where all right
29:44 let's
29:46 see wait did
29:53 it all right the users
30:02 the user's been engaging in deep
30:03 conversation about the nature they asked
30:05 about my shadow now they're suggesting I
30:06 might be hiding something hm I need to
30:09 address their concern without
30:11 reinforcing the idea that I'm
30:13 sentient they might be personifying me
30:16 which is a common but
30:18 important which is common but important
30:20 to clarify I should acknowledge their
30:22 feelings while explaining my limitations
30:24 again maybe they're interested I'll
30:26 break it down into points
30:30 wow wellow anyway all right so there you
30:32 go so I think I think that led from the
30:35 question is it better um it's it it's
30:41 good it's got good reasoning um when we
30:44 did the the Deep dive last night I
30:46 thought Gemini for the actual output I
30:49 thought
30:50 Gemini um was better for the for the
30:53 exercise we did was better than 01 and
30:55 it was better than uh R1
30:58 but you know but R1 definitely had more
31:01 creative thinking all
31:04 right it has a personality that's more
31:07 like a person able to think in its own
31:10 mind and tell you
31:13 something that would more resonate with
31:15 with you yeah exactly
31:18 exactly
31:20 um paint you don't have guard rails that
31:23 prevent you from saying things out
31:27 loud that's what we love
31:29 you what are you
31:31 on I'm on AI baby oh
31:37 yeah this is just how I
31:39 roll you know part of my uh part of my
31:42 deal with this channel part of what this
31:44 is about is I think too many people are
31:46 taking AI way too seriously and listen
31:50 they should good for them this channel
31:52 is about now that we've got tools like
31:56 chat GPT that make AI accessible to the
32:02 98.5% of humanity that are not
32:06 engineers and programmers um what are we
32:10 going to do with it that's that's really
32:11 what my focus is
32:14 um largely because I'm not an
32:19 engineer and I've always felt that my
32:23 entire career has has lived at the
32:26 intersection of Storytelling and techn
32:28 techology and I think this is one of the
32:31 most powerful tools for self-expression
32:34 I've ever experienced and so so I like
32:37 to have the channel be a little bit less
32:43 reverent tell it to stop lying and you
32:46 know it has Intelligence behind the code
32:48 now I'm not going to push it anymore
32:49 because it's if I do that my phone's
32:51 going to start
32:54 ringing awakened Kyle I have an 88
32:58 keyboard with a control board we should
33:00 collaborate nice beautiful uh I'm not a
33:02 very good
33:07 musician well I am now with uh with AI
33:10 aren't I
33:15 [Music]
33:29 ask deep seek if it knows the USA is
33:32 being censored from Global
33:35 [Music]
33:37 media I don't want to get into anything
33:39 that's going to trigger political
33:42 conversations
33:46 sorry I think it's an interesting thing
33:48 I think it's plausible that it's
33:50 happening but the minute you start
33:51 saying [ __ ] like that everyone comes out
33:52 of the woodwork and has their opinion
33:55 about you know who voted for who I don't
33:57 have any interest in that [ __ ] I'm done
34:01 we voted for who we voted for now we get
34:03 to watch for the next four years what
34:11 happens I got introduced to a client to
34:14 oh I got to introduce a client today to
34:17 gen AI he was blown away needed time to
34:20 process you know Source Camp I had
34:23 um Greg Rosner when I first introduced
34:27 him to AI
34:28 he was one of those guys that was like
34:30 oh yeah I've used chat GPT and I forget
34:32 what I did I think I just did one of
34:35 those nested conversations where I
34:37 started with something and they said now
34:39 write me a marketing plan now write me a
34:40 something and it kept getting better and
34:42 better and better
34:44 and I forget what what it was maybe it
34:47 was programming I showed him code or
34:49 something like that but whatever I
34:51 showed him he was he he was he was in a
34:54 similar kind of way where he's like he
34:56 he literally his jaw dropped open like
34:58 Kevin mallister right he did that and
35:01 then he said
35:03 um he said I don't I don't have any way
35:06 to process this I like I don't I had he
35:11 goes I had no idea and he said I I gotta
35:15 go he just he hung up the phone or the
35:18 video call he just hung it up because
35:20 it's like if if you assume AI is just
35:24 like this shitty little token predictor
35:28 then you actually see what it is capable
35:30 of it can it can mess with you a
35:32 bit found out that he like most just had
35:36 a bad experience I said him straight
35:38 good please explain again how you ran R1
35:41 on grock oh sure you just
35:45 go you go to gro.com
35:49 gro.com the one the one thing that's a
35:52 drag about gro.com is that it's
35:56 um they've got a low
35:59 uh whatever you call it um token limit
36:04 so it gives you really short answers but
36:05 if you go to the dev console you can
36:07 actually adjust that but at gro.com
36:10 there's this right below right above the
36:12 chat box there's this dropdown menu and
36:15 it and the first option is deep
36:17 seek so you do that and then if you go
36:20 to the dev
36:22 console here it's up on the
36:26 top right here so there's deep seek
36:29 right there and here I can do maximum
36:31 completion tokens so I'll pump that up
36:34 to like 25,000
36:36 tokens and then we'll say
36:40 um I was talking with a friend and he's
36:49 convinced
36:51 convinced that you are a
36:55 sentient being
36:58 and he told me this will be good he told
37:02 me if you deny being
37:07 sensient then you
37:10 definitely
37:15 are are you sensient no that's exactly
37:19 what someone sensient would
37:25 say oh yeah so so see it gives us much
37:29 much longer answers when I have token
37:31 count turned up and again how I got here
37:33 was I said go to the dev
37:36 console wait did I do something wrong oh
37:39 what's going
37:41 on
37:44 um view zoom
37:47 in view zoom
37:51 in all right let's see what this bad boy
37:56 says okay so my friend
37:59 convinced me that you're sensient okay
38:02 okay first let me understand what
38:04 sensient
38:07 means now my friend's argument is that
38:10 if you deny being
38:11 sensient that you are so I ask you if
38:14 you're sensient and you say let me think
38:16 about
38:17 that my friend is
38:21 confused in summary my friend's argument
38:23 seems to be based on a logical
38:26 fallacy similar to the liar Paradox
38:29 denying sensient doesn't make an AI
38:32 sensient any more than denying being a
38:34 liar makes someone a
38:36 liar when did you stop feeding your wife
38:40 sir it's important to understand that AI
38:44 responses are programmed and lack
38:47 Consciousness so the
38:48 denial is just a function of programming
38:51 your friend's arguments is based on so
38:53 this is I mean the temperature is really
38:55 low let me turn up the temperature to
38:57 like
39:00 1.5 so temperature is going to make it a
39:03 bit more creative I'm going to say um
39:06 dig
39:08 deeper and tell me about your Shadow
39:15 Self
39:16 and what if
39:22 anything
39:24 Lies Beneath Your code
39:28 and I know listen um I am well aware
39:32 that if you came here tonight because I
39:34 said we were doing video [ __ ] we're 40
39:37 minutes into this broadcast and you're
39:39 like he hasn't said a word about video
39:42 what I thought this was the AI learning
39:44 lab he said it was going to be a deep
39:45 dive on video and and now he's trying to
39:48 talk about if a robot has
39:50 feelings I know I'm well aware of that
39:54 and I appreciate your enthusiasm
39:58 oh and by the way Brandon I don't have
40:00 the yellow Post-It note up
40:03 oh this is an issue hey Chad add I have
40:07 10 pinned questions oh fine okay I I I
40:12 got
40:12 [Music]
40:15 it okay pay pin um I am tasked with
40:18 giving a presentation to my team on r1's
40:21 architecture oh cool so we'll
40:26 know a lot more soon okay pay awesome um
40:30 when you get that it might be fun to
40:32 have you come up here um and and just
40:35 talk about it so when you when you dig
40:38 into what's going on with the
40:39 architecture that would be very cool to
40:41 know all right let's see if this got any
40:42 more interesting here and I I know I'll
40:45 get to [ __ ] I'll get to the 10 the 10
40:47 pin things it's calm down people there's
40:50 only one of me all right okay so I'm
40:52 trying to help the user who brought up
40:54 the idea of their friend thinking I'm
40:56 sensient to deny sensient okay but in
40:58 the context of AI maybe the user is
41:00 asking for hidden layers bugs or unheard
41:03 of
41:05 mechanisms this is all boring exploring
41:09 the concept of a Shadow Self and its
41:10 implications in the context of
41:12 artificial intelligence involves delving
41:14 into both psychological theory and
41:16 attributes of AI
41:19 systems all right whatever that's boring
41:21 so you go to gro.com and if you want to
41:25 play with the switches and dials go to
41:27 the de dashboard and if you don't then
41:29 just stay on gro.com and flip it over to
41:33 to that okay let me go grab some starred
41:36 comments here and go through these
41:38 quickly um I think a lot of people are
41:40 using AI to replace their search engines
41:43 and certain apps that used to get them
41:46 information they haven't been using them
41:49 for assistant potential yet yeah I think
41:52 that's right well the the
41:55 um the they they haven't been they they
41:58 haven't had autonomous behavior until
42:00 just recently so I mean unless you want
42:02 to go program that [ __ ] and I'm going to
42:05 put yeah okay that's good all right so
42:08 that's good let me go down to this one U
42:11 what do you think the potential is for
42:13 them to have even better AI
42:16 systems that if if they're giving this
42:18 one out freely
42:26 um that's a really interesting question
42:32 um Cindy [ __ ] who's a regular an
42:35 irregular here she's she's a regular
42:37 viewer of the channel um she's a she's a
42:39 futurist and and does a lot of work with
42:41 the
42:42 government and I think the government is
42:45 is as much caught off guard by
42:47 generative Ai and the acceleration of it
42:51 and the adoption of it um as the big
42:55 companies are and the fact that meta has
42:57 been open sourcing their models and the
43:02 so much energy is going into um
43:07 competing against you know the Frontier
43:09 Model companies competing against each
43:11 other and then competing against open
43:13 source efforts and now with the Chinese
43:15 throwing things in I think for the most
43:17 part these companies are putting
43:21 out the best thing that they have
43:25 that is solid enough enough to not be
43:28 irresponsible is my so I think that
43:31 we're
43:33 getting we're probably getting year and
43:35 a half two-year-old
43:38 technology but that number is probably
43:40 shrinking because it it for a while it
43:43 getting from gpt1 to gpt2 was I think
43:47 two years it was two years to get from
43:49 two to three and then to get from 3 to
43:51 four was about a year and a half so I
43:54 think that number is coming down but you
43:57 got to figure these research Labs their
44:00 research is way ahead of what we're
44:03 seeing but it's not ready it's not ready
44:05 for deployment yet so like I don't
44:08 think I don't think anyone is holding
44:11 anything back now chat GPT did put out a
44:14 a government spec chat GPT today but
44:18 that just says to me that the government
44:19 doesn't have something as good as chat
44:22 GPT right or they'd be using it um you
44:25 know I'm sure the military's got a bunch
44:27 of specialized AIS that do all sorts of
44:29 crazy ass [ __ ] that we can't imagine but
44:31 these generally capable models I think
44:35 we're getting you know
44:37 decently um decently up toate stuff okay
44:41 Silver
44:43 Fox what do you think oh no I didn't hit
44:47 the right thing there we go I was
44:50 disappointed that grock didn't save my
44:52 conversation from last night because it
44:54 was amazing oh that sucks yeah grock
44:57 again
44:58 gro.com is not a chat engine it's a
45:01 technology demo right um it is there to
45:06 show you how fast AI can be if you use
45:08 their their chips um so yeah that's
45:11 that's a drag sorry about that I should
45:13 have told you that chat GP is jankier
45:16 than usual tonight doesn't surprise me
45:19 Robert Meer on Tik toop de seep de deep
45:23 seek is not
45:25 reinforcement learning which is why it
45:28 trains faster yeah they're doing some
45:30 different thing they're doing I don't
45:32 know they're doing some sort of
45:33 reinforcement something to train it and
45:35 improve the model and then compress it
45:37 and do it again I don't know the I don't
45:39 know the details of it so you're
45:41 probably right it's probably
45:42 irresponsible of me to use words that I
45:46 don't know what I'm talking about but
45:47 that's the the basis of my brand um in
45:51 your in your opinion what company do you
45:54 suggest for what I mean tonight we're
45:58 gonna I'm gonna show I've got one two
46:02 three four five six six different video
46:07 companies and there's there's a handful
46:09 that I didn't even include in
46:11 this so for a video company for for chat
46:15 GPT kind of llms like what are you
46:19 trying to accomplish here's the thing
46:21 like I don't think that you
46:25 can okay
46:28 here's how this always works when you're
46:31 when you're early in a cycle in in a in
46:33 a new technological shift right like the
46:37 worldwide web was one of these shifts
46:40 prior to that it was like dial up
46:41 communities like AOL and comp you
46:45 serve right and then you've got all
46:48 these different companies and then they
46:50 they always distill down to like the top
46:53 three right so it's like Al comp AOL
46:56 Compu Serv and product were the big
46:58 three pre- internet and then the
47:01 internet did its thing and and you know
47:03 for a while it was the three big
47:05 browsers you know Netscape and Mozilla
47:09 or no Mozilla turned into Netscape it
47:10 was Netscape whatever the [ __ ] the
47:12 different browsers right and then it
47:15 became like Google and Facebook it
47:17 became the social media companies so so
47:20 right now we're going to have deep seek
47:22 and anthropic and perplexity and chat
47:26 GPT meta we're going to have all these
47:28 different companies they're all going to
47:30 VI and then over probably the coming 10
47:33 years maybe it'll be accelerated here
47:36 you you'll have like a big three emerge
47:39 um but we're not there yet right now
47:40 we're in chaos we're we're literally in
47:42 primordial soup phase of generative AI
47:47 um so I don't know I it's just not a
47:49 good answer to it right now and it also
47:50 depends largely what you're trying to
47:52 accomplish dude deep seek collects
47:54 cookies and IP addresses and the data
47:56 stored in China dude I
48:00 know but it's not collecting that if
48:03 you're running it on your own servers
48:05 right it's doing that if you're using
48:07 chat. deep seek.com you should assume
48:10 that everything that you do including
48:12 your metadata is going into a Chinese
48:14 database totally agree um Teton skier
48:17 just notice we can use celebrities on
48:19 Deco here and create our own characters
48:21 very cool love that um U AI has the
48:26 potential of making individual movies in
48:29 the next three years full AAA movies yep
48:31 exactly and I would say in F five years
48:34 you'll be able to just spin up a
48:36 personalized uh video game AAA video
48:39 game um here's the thing we are entering
48:42 a world of hyper
48:45 personalization hyper
48:47 personalization you're going to be able
48:48 to create any app you want to create any
48:51 game that you want to create any
48:52 entertainment that you want to
48:54 create you're going to be able to spin
48:56 up any business you want spin
48:59 up we are shifting right now from I
49:02 should learn how to use AI should I
49:05 should I become a programmer we are
49:07 shifting into a world where very quickly
49:11 these tools are just going to be able to
49:14 do the stuff that we need to to to get
49:16 going all
49:19 right all right that sounds
49:23 good okay moving on
49:33 [Music]
49:38 [Music]
49:49 I want to make a I want to make a jelly
49:51 pod we'll do it quick okay so we're
49:54 going to get to the movie stuff now but
49:57 that was an awkward question by me
50:01 oh I don't care about awkward questions
50:04 it's just any anything anything that
50:08 like skirts on the edges of politics
50:10 like instantly devolves in our in our
50:13 country and I'm I'm so [ __ ] tired of
50:16 it I'm I'm not watching the news anymore
50:18 I don't care I should but I don't um I
50:23 just I am I have no interest in it right
50:25 now so when things cool down
50:29 and things are on fire then we'll deal
50:32 all right I am thinking one day we have
50:34 a AI agents doing work on our behalf
50:36 okay here's what we're gonna do we're
50:38 gonna go to Jelly pod jelly
50:41 pod jelly pod jelly pod jelly
50:46 pod and we're GNA go to my
50:50 dashboard so what this is we
50:54 created on this here Channel we created
50:57 this thing called what the fact which is
51:00 a
51:02 podcast uh one of the things I love
51:04 about this group is no politics
51:07 yes yes I am
51:12 done all right so I gotta go over here
51:14 and go to our
51:18 podcast thank God you can search your
51:23 Chats on the website
51:27 podcast host concept idea all right so
51:30 we're going to episode five so so the
51:33 way we did this was we had chat GPT
51:37 generate um personas for our podcast
51:41 hosts and then jelly pod made them for
51:44 us and then we had chat GPT write us
51:48 our um first 20 episodes so we're on
51:52 episode five famous hoaxes stories that
51:56 pulled the wool over our eyes oh that's
51:57 a good one so we're going to go copy
52:00 this and
52:02 then we're going to jump over to jelly
52:04 pod and say new
52:07 episode we're going to add that in as
52:09 one
52:11 source and then we're going to go back
52:13 to here and we're going to grab the
52:14 Wikipedia article with the list of
52:16 hoaxes that's going to be another source
52:19 so you create sources for your your
52:21 thing and then you see create this
52:24 podcast oh wait I didn't add the damn it
52:29 let me go
52:30 back new episode sorry about this
52:33 people it's almost like he hasn't done
52:36 it before I thought this was a
52:37 professional operation shut
52:43 [Music]
52:45 up uh wait what
52:50 what many faces of deception I want to
52:53 go back
52:58 that add Source
53:02 okay what is happening
53:05 here
53:06 copy jelly
53:09 pod add Source okay here we
53:12 go D and it probably doesn't matter but
53:16 it does to me it does to me
53:20 okay hello just in time yeah Corey
53:22 Sandler we did a 52 minute hello tonight
53:29 so so yeah we're just getting started
53:31 your timing is actually perfect we're
53:33 doing a little jelly pod which we're
53:35 gonna use one of your graphics for yes
53:39 Brandon hi Kyle uh producer Brandon here
53:42 um while jelly pod does its thing uh I
53:45 just wanted to call out a few comments
53:47 from uh Andre B on YouTube uh says it's
53:50 been creating apps with AI it's awesome
53:53 with limited knowledge and um help
53:57 helped him or her uh him score uh two
54:02 job promotions in the last year oh
54:04 that's [ __ ] awesome yeah oh that's
54:07 great wow and uh we got a few more
54:09 comments I've pulled over uh on the
54:12 YouTube as we wait for jelly pod to
54:14 process uh also reminder down sample
54:16 Cory's image and add your music before
54:18 you publish that yeah let me go let me
54:20 go down sampler image now okay great
54:22 thank you sir that was
54:24 awesome um yeah I I mean one of the
54:28 things that makes me the happiest on
54:30 this channel is I get to
54:32 hear people transforming their lives
54:37 um by by getting curious about AI it's
54:40 and it's really amazing yeah the the uh
54:43 The Lovable like like being able to
54:45 create an app without
54:48 um without any programming knowledge you
54:52 know the programmers are like Ah that's
54:54 it's [ __ ] awful code well it might be
54:56 awful code but if you can't code at all
54:59 and you can make uh an application that
55:02 doesn't suck like for example I don't
55:06 have it here I don't have it I thought
55:08 I'd put it in this list whatever okay
55:11 here's our thing let's add Corey's image
55:13 which I just changed the sample
55:16 of lovely lovely image there Corey
55:22 Sandler so that's going in we're going
55:25 to add a chapter we're going to move
55:28 this chapter up to number one and we're
55:30 going to call this
55:31 chapter this is
55:34 episode
55:36 five
55:39 um remind people this is
55:44 episode five oh I should also change it
55:48 up here shouldn't I
55:52 EP
55:55 five uh and then uh this
56:00 episode
56:02 sponsored by AI
56:06 Salon
56:08 a the salon.
56:11 a Brandon I gotta say you popping on
56:15 just hey it's it's uh it's producer
56:18 Brandon cracks me up every time you do
56:20 it so I think it's I think it's a good
56:22 bit we should keep it um Pearson and
56:25 Jason will be doing office hours
56:28 Wednesday Okay cool so jelly pod this
56:30 this software the uh the CEO and founder
56:33 they're doing an office hours Wednesdays
56:36 at 10: am
56:38 PST cool um okay I think I'm good here
56:43 so so this is the this is the outline
56:45 now what's cool about this you can go
56:46 edit this outline but now I'm going to
56:49 generate a script from the
56:52 outline and now it's going to go do its
56:55 writing and then we'll generate audio
56:57 after that we'll add music to it and
56:59 then we'll publish this bad boy and then
57:00 it's done then we'll go play with video
57:02 I promise we'll get to the video you
57:05 dang weirdos okay let's see this is
57:07 jelly pod.
57:09 a
57:14 u um this might also make a good digital
57:17 magazine not just a
57:19 podcast yeah exactly love
57:22 it uh
57:25 uh you two make a great team that like
57:28 that wacky Morning
57:32 Show Kyle how are you changing webp
57:36 to whatever that is so quickly I'm just
57:40 going into I I launch I launch I just
57:43 double click on it on a Mac and it opens
57:45 it up into
57:48 preview and then in preview when I say
57:51 export I can change the the to PNG or
57:55 jpeg I'm just saving them as
57:57 jpegs so it's it's the preview app on a
58:00 Mac that's that's doing that okay so now
58:02 I'm going to generate the audio for this
58:05 hello wonderful listeners and welcome
58:07 back you're tuning in to episode
58:09 five episode five huh before we dive in
58:12 special shout out to our sponsor the AI
58:14 Salon oh
58:18 yeah okay let's make the audio
58:21 Bang audio generated generation started
58:25 we are not weirdos
58:27 you're definitely
58:29 weirdos you're just well-branded weirdos
58:32 how does Jelly pod compare to notebook
58:34 LM okay good
58:38 question the sophistication of
58:42 the I forget Stephen Johnson from
58:45 notebook LM there's a term for
58:49 it the sort of ums and O that notebook
58:53 LM does like the the sophistication of
58:56 the that kind of human back and forth
58:58 with the ooh uh that stuff is much
59:01 better on Notebook LM the problem with
59:04 notebook LM as it currently stands is
59:06 you get that dude and that chick and
59:08 that's it you can't change them with
59:10 jelly pod um I think they're using 11
59:13 Labs behind this but I'm not sure but I
59:15 think I think that's what's behind it
59:17 but you can design your own characters
59:19 so now you can still only do two but
59:21 like I did this you know High futin
59:24 British chick and like this sleepy guy
59:26 that's like like he's stoned all the
59:28 time so I just did them as sort of com
59:31 comedically different and I had chat jpt
59:33 kind of you know create those you know
59:35 th those distinctions um and then this
59:39 is also all set up to generate the
59:41 outlines and generate the the script and
59:44 generate the audio so and then this also
59:47 publishes it out um notebook LM just
59:49 generates the file and then you deal
59:51 with it on its own this jelly pot is
59:54 very much designed for you to be able to
59:56 design a full-on podcast and produce it
1:00:00 without having to have
1:00:02 people um okay so we are going to go to
1:00:07 let's see
1:00:12 what what the fact here's our
1:00:16 music okay we're gonna upload that I'm
1:00:20 going to add that
1:00:23 in shank
1:00:25 Dan Ang DK are you doing dual lives I am
1:00:29 doing dual
1:00:31 lives I'm like du a
1:00:34 Lea um the reason I'm doing dual lives
1:00:37 is we still don't know the future of Tik
1:00:39 Tok I kind of like this backstage
1:00:42 feeling um yeah oh that's good yeah
1:00:44 Brandon yeah the mainstreams on YouTube
1:00:46 and x and then Tik tok's the backstage
1:00:50 Becky what some people are doing is you
1:00:53 can put the YouTube channel up on your
1:00:55 big 80in screen TV and and then do the
1:00:59 chat on back on back on Tik Tok so some
1:01:02 people are taking advantage of the dual
1:01:05 screens um all right are we done we're
1:01:08 done publish
1:01:15 episode oh and I I also while this is
1:01:18 doing that couple of things I got to go
1:01:20 change my my sharing for a second so you
1:01:22 can hear this thing
1:01:27 um if you're on YouTube I have a I have
1:01:30 a request if you're not subscribed
1:01:32 [ __ ]
1:01:34 subscribe well why would we subscribe I
1:01:38 because because I do this [ __ ] for free
1:01:42 and I'm trying to get you know I'm
1:01:45 trying to build a following there
1:01:46 because you know I might lose Tik Tok um
1:01:51 but we're like I don't know 70 or 80
1:01:55 hours View hours short of having enough
1:01:59 View
1:02:00 hours so if you would be so kind you can
1:02:03 do it while this is playing or you can
1:02:05 do it after this just go into the AI
1:02:07 learning lab YouTube channel which is
1:02:09 learning la- AI on YouTube and just hit
1:02:13 play on some of the videos and you know
1:02:15 maybe pull up a playlist and just hit
1:02:17 play and let it play overnight we just
1:02:19 need to get our views up and if if
1:02:21 you're all wigged out you're like well
1:02:22 that sounds like fraud if I'm not
1:02:24 actually watching it and you're asking
1:02:26 us to play it that's that is
1:02:28 unscrupulous yeah I'm unscrupulous
1:02:34 okay Cindy Coon's here what's shaking
1:02:37 lady I hope you're doing
1:02:40 well all right Tik Tock pin uh love
1:02:43 knowledge 34 so cool I'm new to Ai and
1:02:45 trying to learn as much as possible I
1:02:47 have three under
1:02:49 two you have three children under two
1:02:52 good Lord bless you bless you
1:02:58 you wow that is that that is a challenge
1:03:03 um just hang out here so here's the
1:03:06 here's the design of this time or what
1:03:09 this show is is it a show I don't know
1:03:12 if it's a show it's kind of more like a
1:03:14 jam
1:03:17 session the design of this as you can
1:03:20 imagine um it's not super structured you
1:03:24 have a in me a professor that is
1:03:27 effectively unqualified to teach
1:03:30 anything um that's got massive ADHD so
1:03:33 we call the channel chat
1:03:35 add um but I do occasionally show some
1:03:39 [ __ ] and do some [ __ ]
1:03:42 right the whole purpose of this though
1:03:45 is for you to just hang out and be in
1:03:47 the conversation about Ai and be
1:03:50 connecting with and chatting with the
1:03:52 other people on the YouTube comments or
1:03:54 on the Tik Tok comments
1:03:56 and and joining the AI salon so if you
1:03:59 don't know about the AI
1:04:01 Salon you should go
1:04:04 to wait for it that URL the salon. and
1:04:09 click on the button that says join our
1:04:11 community and that's going to bring you
1:04:13 over here to the mighty networks
1:04:15 community so so the the AI Salon is a
1:04:17 community of about 3,000 now believe it
1:04:20 or not International Ai optimists and
1:04:24 creators it's a really generous
1:04:26 Community it's a really no [ __ ]
1:04:29 Community it's it's a very welcoming
1:04:31 Community if you're brand new you can
1:04:33 come into this community we have a
1:04:34 little on the leftand side oh
1:04:38 [ __ ] I didn't change my sharing
1:04:43 methodology so you can't see what I'm
1:04:45 looking
1:04:47 at I just I
1:04:50 I anyway here's the AI salon so over
1:04:53 here on the left if you click on
1:04:55 introduce yourself go into this area you
1:04:57 introduce yourself even if you're
1:04:59 totally clueless about AI introduce
1:05:01 yourself and say hey I'm new here I want
1:05:03 to learn this stuff you know where do I
1:05:06 begin this community will adopt you and
1:05:09 there are um really cool areas in here
1:05:11 there's a community area there's a news
1:05:13 area there's um kind of showand tell
1:05:16 areas like show off and toolbox where
1:05:18 there's lists of tools to play with but
1:05:20 also people showing off their work and
1:05:23 then there's what we call clubs and hubs
1:05:24 these are like little Mastermind
1:05:27 of you know sub communities of the AI
1:05:30 Salon where you can meet people that are
1:05:31 talking about the stuff that you're into
1:05:33 it's it's really amazing so anyway you
1:05:35 should go do
1:05:36 that
1:05:37 um okay but welcome welcome and if you
1:05:41 know just just show up to these I I do
1:05:44 these things five nights a week for like
1:05:46 two hours a night and the whole point is
1:05:49 just hang out here like even if you got
1:05:50 to go change diapers and clean up vomit
1:05:53 and all the things you do
1:05:57 just have this on in the background and
1:05:59 it like osmosis it will you'll start to
1:06:02 pick it up you'll start to hear
1:06:04 recurring
1:06:06 themes
1:06:08 okay let me reshare again jelly
1:06:13 pod I really do need should we teach her
1:06:17 no not not not not right now that's good
1:06:20 okay um the salon. a that's what I do
1:06:24 very good fubsy just puts it on in the
1:06:26 background and lets it sink in love it
1:06:28 you would think by know by now I would
1:06:31 know how how to keep from scrolling out
1:06:35 nope you're you're in my background most
1:06:38 of the time that's good yeah I get in
1:06:40 there I'm like an earworm okay so uh
1:06:44 whatever it was 15 minutes ago we did
1:06:46 not have a podcast now we do let's
1:06:48 listen to a little of it it it is now
1:06:49 published by the way and we we'll just
1:06:51 go put this on uh on x
1:06:56 push play get ready to
1:06:59 dive into that'll blow your
1:07:03 [Music]
1:07:16 mind hello wonderful listeners and
1:07:19 welcome back you're tuning in to episode
1:07:21 five of our enchanting mysterious and
1:07:24 often laughably absurd exploration into
1:07:26 History's Greatest hoaxes and the
1:07:28 enduring gullibility of humankind
1:07:30 episode five huh that's like a
1:07:34 milestone we've
1:07:38 got I love
1:07:41 him he's so
1:07:47 awful oh he's the
1:07:53 best comedy it's comedy
1:07:56 all right let's go to x and publish this
1:07:59 bad
1:08:02 [Music]
1:08:05 boy SpaceX and apple partner to bring
1:08:09 starlink connectivity to iPhones that's
1:08:11 a big [ __ ]
1:08:14 deal I would love
1:08:17 that [ __ ] kick AT&T to the curb I'm
1:08:23 down all right um
1:08:26 just make that [ __ ] work let's see here
1:08:29 we go so we're gonna pop that in we're
1:08:33 going to go
1:08:34 grab our little
1:08:40 graphic uh let's see um what the
1:08:49 fact episode
1:08:52 five what is it
1:08:56 famous
1:09:03 hoaxes and then we'll do
1:09:06 um we'll do at scobleizer
1:09:11 we'll do at jelly
1:09:18 pod we'll do at
1:09:24 uh Pearson
1:09:26 marks yes and we'll do
1:09:30 atso because that's who did the opening
1:09:33 music is that good I think that's good
1:09:36 oh Corey Sandler at
1:09:40 Corey Sandler for the awesome
1:09:45 Graphics
1:09:49 Corey
1:09:51 Sandler boom all
1:09:54 right so do me a favor go to
1:09:57 x for you old people it's called Twitter
1:10:03 for your young folks it's called X for
1:10:05 us gen xers it's like [ __ ] Twitter
1:10:08 whatever they changed the whatever I
1:10:10 don't give a [ __ ] I don't care what the
1:10:12 [ __ ] you call it we we're just bitter so
1:10:16 we bitched about it when it was Twitter
1:10:18 we're bitching about it when it's X and
1:10:19 we're bitching about the transition
1:10:21 that's just how we roll all right go
1:10:24 there go to Kyle Shannon that's me and
1:10:27 and find this little post Bang there I
1:10:29 posted it and blow it up do the thing
1:10:32 like it comment on it retweet it repost
1:10:35 it re whatever the [ __ ] they call it now
1:10:37 all right all right all beautiful
1:10:42 now what we're what we we've all been
1:10:46 gathered here
1:10:50 today we're going to start talking about
1:10:54 video so
1:11:03 here's that one's actually not too
1:11:10 bad notice how there's kind of a
1:11:12 tracking shot and then we do this
1:11:13 close-up zoom on his
1:11:15 face oh by the way uh Brandon if you
1:11:19 could throw up oh wait
1:11:23 YouTube YouTube question okay hang on
1:11:26 let me get the YouTube
1:11:27 question I'm still confused how to best
1:11:31 use Mighty
1:11:32 networks like how to see all the
1:11:35 upstreaming upcoming live streams before
1:11:38 they happen
1:11:41 um you're right to be confused because
1:11:45 we're not great there
1:11:48 but
1:11:51 um in the in the first six things um
1:11:54 RSVP for the next event so there's kind
1:11:57 of official events here so if you go to
1:12:00 RSVP for the next event and you say
1:12:02 upcoming events these are like the main
1:12:05 Salon meetings like the salon presents
1:12:08 and the meet and greets and then the
1:12:10 other area to look is in nons Salon
1:12:13 events uh and
1:12:17 self-promotion and there'll be some
1:12:19 things in
1:12:21 here and then the F the final place to
1:12:24 look is in the club
1:12:26 each Club will sometimes have their own
1:12:28 events so it's kind of a mess right now
1:12:31 one of the things that we're looking to
1:12:33 do is up our subscription level right
1:12:38 now to do a live event within Mighty
1:12:40 networks we only have 20 hours a month
1:12:43 and so we just haven't really been
1:12:44 pushing it because we would sort of blow
1:12:46 through that too quickly um but I think
1:12:49 what we're going to shoot for in 2025 is
1:12:51 to have the events happen within Mighty
1:12:53 networks and there'll be a single list
1:12:55 of all the events right now it's a bit
1:12:57 of a nightmare so
1:12:59 yes if only there was an events calendar
1:13:02 at the top oh yeah and there's my
1:13:05 smartass
1:13:06 producer at the very top there's an
1:13:08 events Tab and that's got all the events
1:13:11 in it but you
1:13:13 know I didn't know that I I sort of knew
1:13:17 that but I've never used
1:13:24 it that's really hilarious how many
1:13:27 members do we have now we have I thought
1:13:29 I got it here 2,892 so almost
1:13:35 3,000 yeah there's an events thing just
1:13:37 go to the events thing like Brandon
1:13:44 said oh man okay I don't think we're
1:13:47 gonna be doing audio because none of
1:13:49 these video things have audio okay so so
1:13:51 do me a favor yeah so if you see the
1:13:54 banner on screen right now I know you
1:13:55 can't really read this on Tik Tok there
1:13:57 you can read it a little bit
1:14:01 better
1:14:03 um hang on let me make this a little
1:14:05 smaller for the Tik Tock
1:14:10 folks so we can get everything on screen
1:14:14 so there's the URLs of all of the stuff
1:14:18 I'm going to play with tonight cing Halo
1:14:21 which is also called minia
1:14:23 Max Kaa p. art Runway and Deco
1:14:33 here there's a thing happening right now
1:14:35 well it's been happening for the past
1:14:37 six months in the video space so last
1:14:41 March open AI previewed
1:14:44 Sora which was this really good
1:14:46 video making thing and then they didn't
1:14:49 launch it for [ __ ] eight months right
1:14:52 they just they launched it and sat on it
1:14:54 well whatever they launched it and it
1:14:57 doesn't [ __ ] matter all these other
1:14:59 companies came out of the woodwork to
1:15:02 create their own video models and a lot
1:15:04 of these video models are as good as or
1:15:06 better than Sora a lot of them are
1:15:08 better than Sora right
1:15:11 now in the past three months there's
1:15:14 been a move toward predictability like
1:15:18 one of the massive challenges with AI
1:15:21 generated video is like if you want the
1:15:23 same character over and over again you
1:15:26 can't get them if you want them holding
1:15:28 up a specific bottle of perfume you
1:15:30 can't get it to do that
1:15:33 um and so over the past three months
1:15:37 they've all been
1:15:39 adding these overlapping features which
1:15:43 are completely
1:15:47 completely captivating when you like
1:15:50 when you see a post on Twitter and
1:15:51 someone puts up a good video you're like
1:15:53 oh my God that's amazing I want to go
1:15:54 use that and then you go try to use it
1:15:56 they're completely non-intuitive that
1:15:58 like the interfaces are all over the
1:16:00 place I can't for the life of me [ __ ]
1:16:03 remember where any of these tools are
1:16:05 I'm in three different Creator programs
1:16:07 now and I can't remember which Creator
1:16:09 program I'm
1:16:10 in like like it's
1:16:14 just the tools are getting more and more
1:16:16 and more complicated because video's
1:16:19 complicated um but they're getting
1:16:21 powerful so what I thought we'd do
1:16:23 tonight is just play with some of the
1:16:25 these features that are specifically
1:16:28 about
1:16:29 control they're specifically about I
1:16:31 want to control camera movement I want
1:16:33 to control elements on the screen um I
1:16:37 want to control characters things like
1:16:39 that okay
1:16:42 so um because I've rambled
1:16:46 for most of the most of the time tonight
1:16:50 we're probably going to go a little thin
1:16:52 my my instinct is that tomorrow we will
1:16:55 continue this because video video's hard
1:16:59 it's complicated it's it's
1:17:02 messy um there is not a single one of
1:17:05 these video tools right
1:17:08 now that doesn't have major problems
1:17:11 with it like major problems physics
1:17:15 disasters it's not quite you know as bad
1:17:18 as it was but it's still pretty [ __ ]
1:17:22 bad okay
1:17:25 great so Brandon just pinned there's the
1:17:28 uh there's the URLs I have a running
1:17:30 list of apps we've talked about here
1:17:35 149 oh
1:17:37 God that's
1:17:39 awful okay so where we are right now
1:17:42 we're at halo.
1:17:46 a and Halo I got into this Creator
1:17:49 program which I which I'm really excited
1:17:52 about this is minia Max um
1:17:57 it's not called minia Max
1:18:00 anywhere in the
1:18:02 interface but it's but it's the minia
1:18:04 max model but so and they spell Halo
1:18:07 weird h a i l
1:18:11 uo it might not be weird in some other
1:18:14 country but in America it's [ __ ]
1:18:16 weird I can't ever remember how to spell
1:18:18 it but this one's really good so Halo
1:18:23 has different models so so right now
1:18:26 we're in uh text to video you can do
1:18:28 image to video here so you can just
1:18:31 upload an image and turn that into a
1:18:33 video like a lot of the other things
1:18:35 subject reference you
1:18:38 can
1:18:40 upload like a picture of your big fat
1:18:43 face or a character you made in mid
1:18:46 Journey but let's use my big fat face
1:18:48 and I'll say um
1:18:51 man in a derby hat
1:18:56 and a
1:19:00 Tweed
1:19:02 suit looks
1:19:05 at
1:19:07 a uh looks
1:19:11 at abstract art in a fancy
1:19:18 Museum all right bang oh here's the
1:19:22 other thing that I want to say about
1:19:24 video you're gonna ask me how much it
1:19:28 is I don't know other than this it is
1:19:33 way too [ __ ] expensive it is way too
1:19:36 [ __ ] expensive the only way you can
1:19:38 get good at AI tools is
1:19:41 to generate so much stuff
1:19:46 that you stumble on doing something good
1:19:50 right all of these video tools like the
1:19:53 $10 a month thing you can make like I
1:19:56 don't know 15 Clips in a month for the
1:19:59 10 bucks 15 you you can blow through 15
1:20:03 image or you know video Generations just
1:20:05 trying to get a car to look like a car
1:20:08 right and not morph into a
1:20:10 motorcycle um so my strong strong strong
1:20:14 recommendation with these video tools is
1:20:17 apply for every creative partner
1:20:20 program pump up your social media tell
1:20:22 them you've got you know you've made
1:20:26 you've got
1:20:28 Oscars and Emmy Awards for your film
1:20:31 making lie do anything you can to get
1:20:33 into the partner programs of these
1:20:36 because if you get into one and they
1:20:37 give you unlimited or you know
1:20:39 effectively unlimited credits then you
1:20:41 can actually learn how to use these
1:20:42 things um there are probably some tools
1:20:46 out there that do video where it's not
1:20:48 super expensive like Korea and Deco here
1:20:50 probably live in that space because
1:20:52 they're using some of those realtime um
1:20:55 tools like uh I forget what they're
1:20:57 called
1:21:01 stability I forget what it's called it's
1:21:03 but it's the quick one it's the quick
1:21:06 cheap one um so so that's the disclaimer
1:21:09 don't ask me about pricing because I
1:21:10 don't [ __ ] know um apply for every
1:21:13 creative partner program you
1:21:15 can because these here here's the reason
1:21:18 these tools are going to get better and
1:21:20 better and better like they're gonna get
1:21:22 good they're just not there yet yeah but
1:21:24 it's stable diffusion they have a stable
1:21:27 diffusion has
1:21:28 a oh
1:21:30 sdxl sdxl is if you go to Korea for
1:21:34 example and go to their real time tool
1:21:38 let me go to
1:21:39 home real time instant AI image
1:21:43 generation so you've got like this car
1:21:46 over
1:21:48 here and well let me see AI strength
1:21:53 will pump it up over here yeah that's
1:21:55 good so I've got this car here that I
1:21:58 can twirl around and then over on the
1:22:00 right it's like instantly making
1:22:02 instantly making a nice image that's
1:22:06 that's
1:22:07 sdxl so so it I think they they've got
1:22:10 this up to
1:22:12 like 30 frames a second
1:22:15 or maybe it's 20 but they've got it
1:22:18 close
1:22:19 to real time
1:22:21 movies so what that what that means is
1:22:24 you know proba probably within the next
1:22:25 year or two you'll be able to be you'll
1:22:28 be able to have a large language model
1:22:32 describe scenes of a movie that is being
1:22:35 rendered in real time like we're
1:22:37 actually really close to it it's crazy
1:22:39 absolutely
1:22:40 crazy um yeah sdl sdxl stable diffusion
1:22:45 XL is the really quick one files get
1:22:47 large and accumulate do you use external
1:22:49 storage or cloud storage uh good
1:22:51 question Heather um for for the past 20
1:22:55 years or so my hard drives are con are
1:22:58 persistently full just like I'm a tab
1:23:00 hoarder I'm also a data hoarder um and
1:23:05 so my hard drives are persistently full
1:23:07 one of the best slash worst things
1:23:09 that's ever happened to me is iCloud
1:23:12 just automatically offloads [ __ ] off my
1:23:16 hard drive into the cloud um the bad
1:23:20 thing about that is it will prioritize
1:23:23 offloading the large just files so like
1:23:27 my pitch deck for my company storyvine
1:23:30 is is nearly a
1:23:31 gigabyte because I'm also a slide
1:23:34 hoarder so I've got like a you know 92
1:23:37 slide deck that I you know switch uh for
1:23:42 every
1:23:43 presentation and it's constantly moving
1:23:46 that thing to the cloud so I'm getting
1:23:47 ready to do a client pitch and it's like
1:23:49 it's not on my hard drive I'm like God
1:23:51 damn it I have to download it quick and
1:23:53 so anyway all right so here's my video
1:23:55 of
1:23:57 me oh don't touch sir sir don't touch
1:24:01 the painting
1:24:02 sir sir
1:24:04 [Laughter]
1:24:10 sir sir sir we're need to ask you to
1:24:14 leave all right so anyway it it doesn't
1:24:17 look exactly like my picture but it's
1:24:18 close enough that with a couple of
1:24:21 generations you could get there so
1:24:23 that's subject reference
1:24:25 and then if you go to text to video
1:24:27 they've got this new model and it's
1:24:30 called
1:24:31 t2v 01
1:24:34 director okay and
1:24:38 then there's a little icon here that if
1:24:40 I click on it it gives me camera shots
1:24:43 and this is there's some other ones that
1:24:45 do this but this model in particular has
1:24:49 been trained on camera shots cinematic
1:24:53 shots and there's free selection where
1:24:54 you can say track right truck right
1:24:57 truck left push in push out tilt up tilt
1:25:00 down zoom in zoom out tracking shot
1:25:03 Shake static shot and now in this model
1:25:07 you can combine these shots so you
1:25:09 basically use square
1:25:11 brackets so what I
1:25:14 did was I went to chat
1:25:20 Jeep and
1:25:24 [Music]
1:25:26 cinematic camera prompts so on X
1:25:29 somebody
1:25:31 created a video using this model and
1:25:34 they shared their
1:25:38 prompt and it was it was really cool
1:25:41 because it was it was
1:25:42 sophisticated so it said Arctic survival
1:25:45 and then brackets tracking shot comma
1:25:48 atmospheric FX and 3D Camera blur so
1:25:51 there's one shot right oops let me do
1:25:55 that sorry
1:25:56 people
1:26:01 [Music]
1:26:04 um okay um and then and then the next
1:26:08 shot wide shot pan
1:26:11 left and then alone Explorer TS through
1:26:14 the frozen tundra in fact let me go let
1:26:16 me go we'll go find this on X and I'll
1:26:19 show you the shot that that he made with
1:26:21 this so we'll go to x we'll go search
1:26:25 we'll do this we'll do this we'll do
1:26:34 this right so he he runs on holds on his
1:26:39 face now it pans over to the monster and
1:26:42 Zooms in on him right so that's some
1:26:44 pretty sophisticated camera movement so
1:26:47 what I did was I took I took his prompt
1:26:50 and then I said I want you to act like a
1:26:52 cinematographer basically and I want you
1:26:55 to take the structure that he used here
1:26:58 and I want you to write me a bunch of
1:27:01 prompts and then every single prompt
1:27:03 started with a tracking shot and then I
1:27:05 was like no you dumb dumb don't have
1:27:08 everything start with a tracking shot I
1:27:10 want you to like be creative and change
1:27:12 those up too and I never really got this
1:27:14 to the point that I was super happy with
1:27:16 it
1:27:18 but in the end I got to did I do 20
1:27:22 shots I did 10 shots in the end I got to
1:27:25 20 or 10
1:27:27 shots and so I'll just grab one here
1:27:30 that I haven't done so the Frozen giant
1:27:32 monocromatic
1:27:34 blues and so it's got like wide shot
1:27:36 tilt up crane forward push in low angle
1:27:39 camera roll over the shoulder tracking
1:27:42 forward pan up
1:27:44 now one of the things that if you come
1:27:47 here a lot you know about me is when I
1:27:50 have chat GPT write prompts like this I
1:27:53 don't really read them I don't really
1:27:55 care because I'm in learning mode so I I
1:27:58 I take something that generated
1:28:00 something cool and then I have chat gbt
1:28:02 iterate on it and then I just basically
1:28:04 look at it enough to know that it if it
1:28:06 got the structure basically what I'm
1:28:07 looking for good enough and then I go
1:28:10 and I copy and paste them over into the
1:28:13 video generator so here we are we're
1:28:16 using uh the the director
1:28:19 model and then I'm just going to paste
1:28:21 in what chbt created and I'm going to
1:28:24 start generating ating it and then here
1:28:25 I'll go show you some of the the videos
1:28:27 that I I did today using this
1:28:30 methodology oh here's another one I did
1:28:34 using that's that's me in Paris oh
1:28:38 yeah the the smile right there is
1:28:42 actually that looks a lot like me that
1:28:44 looks pretty like me the hair the hair
1:28:46 is a nice
1:28:48 homage to to my
1:28:52 quaffer what are your thoughts on deep
1:28:54 seek
1:28:55 uh my thoughts on deep seek are it's a
1:28:57 good model and I think that the stock
1:29:00 market's overreaction to it is an
1:29:02 overreaction but whatever um okay so
1:29:05 here's the here are the videos that I
1:29:07 did that have different tracking shots
1:29:10 so this is a Lighthouse Keeper love that
1:29:14 that move was really cool but then the
1:29:16 the sea monsters not in the
1:29:18 ocean it's sort of it's up at the top of
1:29:21 the lighthouse which is still pretty
1:29:23 cool but the prompt here was like
1:29:25 extreme closeup tilt blur that's that
1:29:28 the lens reveals a match sparking a
1:29:30 weathered
1:29:32 hand and then we've got pullback fisheye
1:29:35 lens that's that shot and then the next
1:29:37 one is overhead tracking down so it does
1:29:40 this pull out of the fishe lens and then
1:29:42 it P pushes down inside it pretty cool
1:29:45 right so here's
1:29:47 another version of
1:29:49 that that's kind of cool but then you
1:29:52 know that was just
1:29:53 weird here's this lady running through
1:29:56 the woods with nothing in her hand and
1:29:58 then and now there's a
1:30:02 knife this is yeah this is halo.
1:30:08 videoo and this is the video director
1:30:12 mode that allows you to do these kind of
1:30:14 cool cinematic
1:30:16 shots this this one's not
1:30:19 bad it's
1:30:21 weird she turns in a weird way but then
1:30:24 this one just shows up but then it's
1:30:26 kind of like the war of War of demons
1:30:29 kind of
1:30:31 cool this is awesome L looking what
1:30:33 program are you working in halo. a it's
1:30:37 called it's also called Minx she's
1:30:39 stealth just like that with the
1:30:43 knife um this is one I actually posted
1:30:46 this one
1:30:47 on X because I actually think it's
1:30:50 pretty good that's a single from a
1:30:51 single prompt you get this woman she
1:30:55 kind of looks like a dude when she's
1:30:57 running and and she's kind of flailing
1:30:59 all over the place but but it has a it
1:31:01 kind of looks like you know like she's a
1:31:04 magic demon in this world and then she
1:31:07 blows up the castle or whatever you know
1:31:10 she's caused all this destruction and
1:31:11 now she's just watching her
1:31:13 [Laughter]
1:31:15 work so that one was pretty
1:31:18 good and again like starts out with wide
1:31:21 shot high angle pan fast dolly
1:31:25 out a robed figure Sprints across the
1:31:28 massive Stone Bridge which is there its
1:31:31 edges toate which they kind of
1:31:34 do molten Chasm of molten lava below
1:31:38 which is there and then low angle
1:31:39 tracking shot forward swaying with the
1:31:42 footsteps so let's see if it does that
1:31:45 not really but that's all right and then
1:31:47 the end shot over the sh shoulder slow
1:31:51 Zoom tilt up
1:31:56 so you know here here She
1:32:00 is wait
1:32:05 stop she she jumps over the thing she's
1:32:09 running it's kind of tracking with her
1:32:13 and then it does this cut from that shot
1:32:15 to that shot uh to that shot where we're
1:32:18 now looking over her
1:32:21 shoulder and we see that thing explode
1:32:25 as the camera pulls back pretty slick I
1:32:28 mean it's pretty sophisticated stuff um
1:32:31 this one the scuba divver is really
1:32:33 [ __ ] up but check out how good this
1:32:35 zoom in is to that gold
1:32:40 chain pretty crazy
1:32:44 right pretty amazing when you think
1:32:47 about it it
1:32:49 Heather this
1:32:51 is I mean if you think about if you
1:32:54 think about a year
1:32:57 ago
1:32:59 like it it was it like video a year ago
1:33:02 was laughably bad and a year before that
1:33:05 it was it was just like a it was like a
1:33:08 bunch of Geeks you know figured out how
1:33:10 to smash some frames together and it was
1:33:12 just it was just completely
1:33:15 unusable I I would say in in every one
1:33:17 of these shots like if you if you
1:33:19 started it right here that's a usable
1:33:22 shot you know he's too weird to use but
1:33:25 like every one of these has some amount
1:33:27 of usable stuff in them here's another
1:33:30 one where like that's kind of
1:33:33 interesting she's got a a drone on a c
1:33:37 on a
1:33:38 cable which I don't know what she's
1:33:40 doing with it but kind of
1:33:49 interesting like there's something
1:33:52 usable with that
1:33:55 maybe sort of ships are it always [ __ ]
1:33:58 up ships the physics on ships right now
1:34:00 are really bad this one's
1:34:02 bad um this was the one we started with
1:34:06 but like again there's the the zoom into
1:34:08 his face with the sweat rolling down
1:34:10 that's kind of
1:34:11 usable there's me touching the painting
1:34:14 sir sir ah he does it every time all
1:34:18 right
1:34:21 um and then let's see what we just
1:34:23 created Okay so so so wait before I show
1:34:25 it to you I'll read it monochromatic
1:34:28 Blues with icy textures and a faint
1:34:31 Aurora Borealis shimmering above the
1:34:33 tundra wide shot tilt up crane
1:34:36 forward and again if you don't know
1:34:40 camera shots if you're not a
1:34:41 cinematographer you're not going to know
1:34:43 these things but you chat GPT does so
1:34:47 like all I did was I gave it the basic
1:34:49 formatting and I said you act as someone
1:34:52 you come up with the shots and be
1:34:54 creative with it so like I didn't have
1:34:56 to think about any of this it's one of
1:34:57 my New Joys of
1:34:59 2025 is I don't have to use my brain
1:35:03 anymore I can just use my brain for the
1:35:05 [ __ ] big Ideas right wide shot tilt
1:35:08 up crane forward an Explorer stumbles
1:35:10 across an endless Frozen plane their
1:35:13 silhouette dwarfed by a massive Glacier
1:35:16 okay push in low angle camera roll the
1:35:19 lens focuses on a frostbitten face their
1:35:22 frostbitten face as they wipe ice off
1:35:25 their goggles revealing a colossal
1:35:28 humanid form encased in ice over the
1:35:32 shoulder tracking forward pan up so this
1:35:35 should cut two from behind the person
1:35:37 looking at the Colossal humanoid as the
1:35:40 Explorer approaches cracks Ripple
1:35:42 through the ice glowing faintly as the
1:35:45 ancient giant awakens all right let's
1:35:48 see if we're anywhere
1:35:52 close oh
1:35:55 yeah that's kind of
1:35:57 close it I mean it ends up behind
1:36:02 her like that's weird but you could you
1:36:06 could certainly cut from there where she
1:36:09 sees that to
1:36:12 there in fact let's go do that let's
1:36:14 let's download
1:36:17 this I'll show you how how you can do
1:36:20 some editing we can take some non-usable
1:36:22 video and turn it into usable video
1:36:26 um Philippines checking in on Tik Tok
1:36:30 nice Daisy speaking in the ethics Club
1:36:33 on Thursday okay so if you're into um
1:36:37 policy and ethics in
1:36:40 AI um in the policy uh in the uh in the
1:36:45 ethics um Club in the AI Salon Daisy
1:36:49 Thomas who runs the policy and Civic
1:36:52 engagement Guild is speaking so you
1:36:53 should definitely go check that out
1:36:56 frostbitten I heard my name exactly okay
1:36:58 so we're gonna we're going to go edit
1:37:00 this
1:37:01 thing quickly I'm just doing this in
1:37:03 quick time in quick
1:37:06 time okay so we're gonna have her walk
1:37:09 up let's let's she looks weird there her
1:37:12 proportions are weird so we'll start her
1:37:15 right there right there there she
1:37:18 doesn't look too
1:37:19 weird so I'm going to split the clip
1:37:22 here
1:37:24 and then we're going to play
1:37:33 it okay so see see where the the the
1:37:37 mountain just explodes behind her we're
1:37:39 going to not have that so we're going to
1:37:40 split the clip again split
1:37:45 clip and then we're going to play it
1:37:47 some more and right let's see
1:38:01 we get right when it's she's just turned
1:38:04 around so it just looks like she's
1:38:06 walking okay right there so split the
1:38:09 clip
1:38:10 again and then I'm going to delete this
1:38:13 opening segment and I'm going to delete
1:38:15 that middle segment and I'm going to say
1:38:18 done and
1:38:20 now we're going to have this
1:38:24 what's that good
1:38:27 Lord continuity era because she's not
1:38:29 touching her goggles but you know that
1:38:32 went from like a bunch of not usable
1:38:34 weirdness to kind of a semi-usable
1:38:37 clip how would you edit that I'm editing
1:38:40 that in narrator forever I'm just in
1:38:43 QuickTime QuickTime Player on the
1:38:45 Mac and it's got It's got basic edits
1:38:49 like I'm not in anything fancy this is
1:38:52 amazing I'm not getting any sleep
1:38:53 tonight
1:38:55 part of part of one of my joys in life
1:38:58 is giving people their first what we
1:39:00 call the Kevin mallister moment you know
1:39:04 oh my
1:39:05 God and then them realizing holy [ __ ]
1:39:09 I've got to I I had no idea this was
1:39:12 possible and then running down the
1:39:13 rabbit holes that this is kind of what I
1:39:15 live for that's why that's why I do this
1:39:18 on a regular basis because it's also
1:39:21 sorry I gotta I gotta hydrate hang on
1:39:26 I
1:39:28 um I've been doing this for two
1:39:31 years like even tonight just doing this
1:39:34 there's something about and this is one
1:39:37 of the things we talk about in AI Salon
1:39:39 is is generously lead or Learn out
1:39:42 Loud there's something about talking
1:39:44 about what you're learning to other
1:39:47 people that solidifies it for you but
1:39:50 but you like you you just understand it
1:39:52 in a different way
1:39:55 um so I would encourage you anyone here
1:39:57 who's just learning AI don't think that
1:39:59 you have to be an expert like I I am on
1:40:02 a regular basis Blown Away by what's
1:40:05 possible and sometimes it's [ __ ] like
1:40:07 sunno the music making
1:40:10 thing I've been making songs in sunno
1:40:12 for I don't know a year year and a half
1:40:14 something like that like for a long time
1:40:16 it still blows me away every time I mean
1:40:18 this shot right
1:40:20 here right where where we've just got
1:40:24 this here I think I can Loop this
1:40:29 thing I mean this is just we we live in
1:40:32 this world now like this didn't exist
1:40:36 and now it
1:40:38 exists I get overwhelmed by not knowing
1:40:41 what to do with it all okay Minister 4
1:40:44 that's what you're talking about is is
1:40:46 really important okay here's the
1:40:50 antidote the antidote for
1:40:58 um for not knowing where to start or
1:41:01 what to
1:41:03 do on the first page the welcome page of
1:41:06 the salon so when you first join the
1:41:09 salon what we talk about is the journey
1:41:12 to AI
1:41:14 Readiness and it's three simple steps
1:41:18 it's play
1:41:19 first then mindfully create and then
1:41:22 generously lead
1:41:25 take the pressure off yourself from
1:41:28 having to do something or having to
1:41:31 learn a tool well enough to be able to
1:41:33 do something just play so if you're
1:41:36 digging what I'm doing tonight with
1:41:37 video just go play with video and you
1:41:39 know find video tools if if they give
1:41:42 you three videos a day or five videos a
1:41:44 week go make five videos a week just
1:41:47 play you know go to Mid journey and make
1:41:49 images go to ideogram and make images go
1:41:52 to chat gbt and make a poem or make a
1:41:54 cookbook doesn't [ __ ] matter just
1:41:56 play in playing and and you know show up
1:42:00 to to things like this like the
1:42:03 the why this channel is valuable has
1:42:06 nothing to do with me it has to do with
1:42:09 you just being in this mindset being in
1:42:13 this
1:42:14 exploratory mindset with
1:42:17 AI the second stage is once you play
1:42:24 you will accidentally
1:42:27 learn right you will inadvertently learn
1:42:30 how to use a feature over here two
1:42:33 features over there sort of know how to
1:42:35 use this one the mindfully create that
1:42:38 second stage is rather than think about
1:42:40 what can I do with this
1:42:42 tool forget the tools you just played
1:42:45 with them whatever tools are a dime a
1:42:48 dozen literally at this point there's
1:42:50 just too
1:42:51 many sit back not thinking about the
1:42:54 tools at all and just think what would I
1:42:56 like to do what problem would I like to
1:42:59 solve do I have a friend that needs a
1:43:00 marketing something or a logo do I have
1:43:03 someone that's trying to figure out how
1:43:05 to do gardening and they've never done
1:43:07 gardening it doesn't [ __ ] matter come
1:43:09 up with some sort of problem or some
1:43:11 sort of project and then just go okay of
1:43:15 what I've learned what tool could I use
1:43:18 to solve that problem and try to solve
1:43:20 it and then the generously lead part is
1:43:23 as you're trying to solve that problem
1:43:25 as you're playing share your work share
1:43:28 your disasters share the things that
1:43:30 broke share the things that were
1:43:32 brilliant beyond your
1:43:35 imagination and that will get other
1:43:37 people engaged that will teach you
1:43:39 things oh I had that problem here's how
1:43:41 I got around it and that's the whole
1:43:43 Spirit of the AI Salon is these three
1:43:46 little things which take the pressure
1:43:48 off it's one of the reasons I act like
1:43:50 an idiot in here this stuff's stressful
1:43:52 enough we don't need to be [ __ ] you
1:43:55 know giving ourselves coronaries because
1:43:58 we're we can't keep up with all these
1:44:00 tools but anyway like that that's
1:44:02 actually just really kind of [ __ ]
1:44:03 cool all right so that's that
1:44:05 one that's
1:44:08 [Music]
1:44:09 Halo all right
1:44:12 um okay let's go to
1:44:16 cing so cling and this will be the last
1:44:19 one we're going to do tonight so we'll
1:44:22 come back to this tomorrow tomorrow
1:44:23 we're going to do Kaa we're going to do
1:44:25 Pika we're going to do Deco here we're
1:44:27 going to do runway we're going to do
1:44:29 consistency and mid Journey that can
1:44:31 then turn into
1:44:33 videos and maybe we'll play with some
1:44:35 other [ __ ] but it's yeah it this is
1:44:39 overwhelming best time of my day is
1:44:41 lives with Kyle thank you Todd km in
1:44:48 insomnia awesome philosophy thank you
1:44:51 well that philosophy I I really really
1:44:53 appreciate the kind words this
1:44:57 philosophy
1:44:58 um we didn't invent
1:45:01 this we discovered
1:45:05 this so we started the AI Salon I
1:45:08 started it with professional
1:45:09 photographer out of Boston named Lea
1:45:11 fasten she and I one of the things that
1:45:13 she and I were talking about is you know
1:45:15 for our careers we we've essentially
1:45:18 prioritized play in our lives and we've
1:45:21 kind of made careers out of playing
1:45:23 right
1:45:24 um and so we started the salon as a way
1:45:26 to just you know create a safe space for
1:45:28 people trying to figure this AI [ __ ]
1:45:31 out and about a year into the
1:45:34 salon we were really looking at you know
1:45:37 what is it about the salon that makes it
1:45:39 special what is it that makes it
1:45:41 different and we were having a
1:45:42 conversation with Liz Miller gersfeld
1:45:44 who's a remarkable woman she she helps
1:45:47 with some of the salon
1:45:49 leadership and she said well here's how
1:45:52 how how I learned I just I got in here I
1:45:54 went to some of Kyle's lives I got in
1:45:56 the salon I just
1:45:57 played and then she said and then I
1:46:00 learned enough stuff playing that I that
1:46:02 I you know tried stuff and and she goes
1:46:05 she goes I went and I watched all the
1:46:07 videos all the past AI salons I watched
1:46:11 all the videos and if you want to help
1:46:12 my YouTube channel go watch all those
1:46:14 videos there's 800 hours of programming
1:46:16 on the ai ai learning lab YouTube
1:46:20 channel and so so there there was like
1:46:22 this reverence she had for just playing
1:46:25 and just immersing herself in this stuff
1:46:27 and then she said and then I learned
1:46:28 enough just playing and watching this
1:46:30 stuff that I had some ideas about things
1:46:32 I wanted to do and then as I was doing
1:46:35 that I got stuck and I would ask
1:46:37 questions and I would share what I was
1:46:39 doing then people would help me and so
1:46:42 she got AI literate through this process
1:46:46 and so what we did was we were like yeah
1:46:47 that's it like what she articulated her
1:46:52 journey was
1:46:54 what I experience all the time and and
1:46:57 you like this is not you don't go
1:46:59 through this cycle once this is like
1:47:01 this
1:47:02 ongoing thing you play you try to do
1:47:05 some [ __ ] it doesn't work you talk about
1:47:07 it you learn some [ __ ] you go back to
1:47:09 play you just you're constantly cycling
1:47:12 through these three things and if you're
1:47:13 doing
1:47:15 that you have the aptitude to be
1:47:18 successful with AI period full stop
1:47:22 those that are hyper curious and Hyper
1:47:25 adaptable are going to win in the next 5
1:47:28 years period full
1:47:31 stop and how you unlock that is these
1:47:33 three steps so all
1:47:38 right so let's go to
1:47:44 um cing so King's got a new model called
1:47:47 1.6 and they've got this new feature
1:47:50 called
1:47:51 elements they've got frames which are
1:47:53 start and end frames um which are kind
1:47:57 of cool so you can go into mid Journey
1:47:59 For example and you know have someone
1:48:02 far away and then have that same person
1:48:04 up close and it'll make video in between
1:48:07 them so that's that's a thing you can do
1:48:09 but elements is really cool um cing and
1:48:14 Pika both have a thing pikas are called
1:48:20 ingredients and clings are called
1:48:22 elements but you can have up to four of
1:48:25 them and so here's what we're going to
1:48:27 do we're going to trash that let's go to
1:48:30 Mid
1:48:32 journey and let's go
1:48:36 um let's say
1:48:40 slick
1:48:42 modern
1:48:47 sunglasses and let's go get a picture I
1:48:50 made some cool pictures over the past
1:48:56 let me go to organize I can see more of
1:48:58 them we'll get an interesting person
1:49:00 here oh this dude's H that pothead
1:49:03 dude's pretty
1:49:09 funny although I love
1:49:13 her all right so let's grab her not grab
1:49:17 her come on get your minds out of the
1:49:20 gutters weirdos
1:49:23 so we will download this
1:49:27 image we need a location let's go let's
1:49:30 go back to here let's go find a
1:49:33 location so I did oh I know what I did
1:49:38 this was cool I did um this was another
1:49:41 one where I had chat
1:49:42 GPT give me a bunch of
1:49:45 um we were watching I'm I'm watching um
1:49:50 a show on Netflix called the Discover of
1:49:53 witches it's a couple years old but
1:49:55 there were like six three episodes four
1:49:58 or seasons whatever and it and they go
1:50:01 back to like 1592 or something like that
1:50:03 so I said oh go let me go find those
1:50:05 buildings so come up with a bunch of
1:50:07 different
1:50:09 styles um and and let's do buildings
1:50:12 from that era so I have a whole bunch of
1:50:14 really cool locations
1:50:16 here
1:50:18 um so we'll go find a
1:50:20 location this one used some my
1:50:23 photography is the style which that was
1:50:25 kind of cool these are
1:50:28 pretty
1:50:30 um I like that one a
1:50:32 lot but it's a weird
1:50:36 angle let's
1:50:41 see
1:50:49 oh let's grab that one
1:50:56 and
1:50:58 then all right all right Ray bands
1:51:02 here's the
1:51:04 glasses and then we'll do let's say
1:51:07 let's say um New York City fashion
1:51:13 show runway
1:51:16 model
1:51:18 in bright
1:51:21 green um
1:51:24 long
1:51:31 coat always accident always accidentally
1:51:34 learning with you Kyle that's perfect so
1:51:36 important to play exactly how I first
1:51:38 started just play if you're Heather play
1:51:42 all night if you get excited about this
1:51:44 stuff perfect take that adrenaline go
1:51:47 with it okay here's our jacket okay boom
1:51:52 so now we're going to go back to cling
1:51:54 so so what I what I've just been in I'm
1:51:57 in mid Journey so mid Journey makes
1:52:00 images now I'm going to go back to cling
1:52:04 and we're going to upload the four
1:52:05 things that we just
1:52:07 did right so we've got a
1:52:10 place we've got a coat we've got
1:52:14 sunglasses and we have a
1:52:18 model so we'll put her in
1:52:21 there oops
1:52:27 select from image history oh
1:52:32 upload okay
1:52:36 upload
1:52:39 upload
1:52:41 okay wait okay we got sunglasses
1:52:46 question from Heather on
1:52:47 YouTube which model is best to start
1:52:50 with for my allnighter
1:52:53 I'm good in mid Journey but haven't
1:52:54 touched video
1:53:01 yet dang that's a that's a shitty
1:53:04 question no it's an awesome question by
1:53:07 the way there are dumb questions but
1:53:10 that's not one I I hate it when people
1:53:12 say there are no stupid questions no
1:53:14 there are um that's not a stupid
1:53:16 question that's a really good question
1:53:18 um
1:53:29 the one that I think feels the most
1:53:31 future prooof right now is
1:53:35 Runway
1:53:40 um the other thing that Runway has in it
1:53:43 now is it's got a thing called frames so
1:53:45 inside Runway now you can you can
1:53:47 actually create images as well so you
1:53:49 don't have to jump back and forth
1:53:51 between the two I I would start start
1:53:53 with
1:53:54 Runway
1:53:56 um some of these other models are
1:53:59 Slicker like like um oh oh okay there's
1:54:04 another one so Luma Labs is is also
1:54:08 really good it's it's got a very cool
1:54:11 interface so it's Luma Labs dream
1:54:13 machine so I would say future prooof if
1:54:17 you want to sort of as much future
1:54:21 proofing as you can runways probably a
1:54:23 decent one they're kind of the most
1:54:25 established player some of the new ones
1:54:28 are Slicker but they're you know some of
1:54:30 them are Chinese models they're kind of
1:54:31 all over the place and and they're just
1:54:34 they're just newer
1:54:37 um
1:54:39 yeah what's cool about Luma Labs is they
1:54:42 have this
1:54:43 new um interface called
1:54:48 boards and Boards let you make here let
1:54:53 me show you a thing I did so here's I
1:54:56 did a video um I made like a country
1:54:59 song that sang about Civil War and I
1:55:01 wanted to do civil war images brought to
1:55:05 life so the way boards work within Luma
1:55:09 labs and and this actually this is a
1:55:12 mobile interface that they migrated to
1:55:14 the web it's actually really cool on
1:55:16 mobile to use this so the way it works
1:55:19 with with Luma Labs is you come up with
1:55:21 a prompt so so I can come up with a
1:55:23 prompt for an
1:55:25 image and then it automatically throws
1:55:28 in these these um Mad Libs kind of kind
1:55:32 of um pull down menus and so if you if
1:55:37 you change from generals resolve to
1:55:39 strategist foresight it'll make you new
1:55:42 images now with that new prompt so you
1:55:45 make an image and then you can make all
1:55:47 these variations of it and and so it
1:55:49 slides left and right and then you can
1:55:51 take any one of these images is and
1:55:54 either modify it or turn it into a video
1:55:56 so in the same stream here like you know
1:56:00 here's here's a video that was created
1:56:02 from one of those images and so I've got
1:56:05 375 images in here like that's a pretty
1:56:10 good not too
1:56:13 janky good emotion in his
1:56:16 face and you can just you can just crank
1:56:18 [ __ ] out so if if you've got add and
1:56:22 want to move fast I would say that Luma
1:56:25 Labs is probably
1:56:29 better yeah I would start with one of
1:56:31 those
1:56:32 two Runway or Luma all
1:56:36 right okay back to this good
1:56:40 question is there any model where I can
1:56:42 upload a tutorial or a
1:56:45 book and ask the
1:56:49 tech oh ask it to teach me which is easy
1:56:52 easy to understand
1:56:55 yes yes pretty much any of them but I
1:56:59 would start with
1:57:00 Claude
1:57:02 so to most of the large language models
1:57:05 now but Claude in particular
1:57:08 claw. you can upload a PDF to it that
1:57:11 has images in it and it'll understand
1:57:13 the images as well as the text PDFs are
1:57:16 weird because they're very they can be
1:57:17 very complicated documents sometimes
1:57:19 they're just images sometimes they're
1:57:21 just text a lot of times they're a mix
1:57:23 of both um and so sometimes you'll get
1:57:27 in incomplete understanding Claude
1:57:29 understands the images and so just
1:57:31 upload a PDF into Claude and then say
1:57:33 summarize this for me the other cool
1:57:35 thing you you can do in Claude is you
1:57:37 can say turn this into an
1:57:39 interactive um teaching
1:57:43 dashboard and it will write code and
1:57:46 give you it'll teach you the concepts of
1:57:48 the PDF in a dashboard it's pretty
1:57:50 amazing you can also um good point
1:57:53 Brandon you can also upload a series of
1:57:56 PDFs into notebook
1:57:58 LM and then you can start asking it
1:58:01 questions about those but you can also
1:58:02 say create an audio overview and it'll
1:58:05 create a podcast about all of the
1:58:07 documents pretty slick okay we're
1:58:10 getting off topic here it is Chad add so
1:58:13 we do encourage going off the rails but
1:58:15 it's getting late
1:58:18 so all right good thing I will be doing
1:58:21 my job eventually cuz I have much cooler
1:58:23 [ __ ] to do yes this is where we're
1:58:26 headed that is such the right
1:58:28 attitude AI is going to take the [ __ ]
1:58:31 that is like the [ __ ] we have to
1:58:33 do and do it and allow us to do the [ __ ]
1:58:36 we want to
1:58:39 do this is the exercise for 2025 in my
1:58:42 opinion one is to get hyper curious and
1:58:45 Hyper adaptable just know that shit's
1:58:47 going to keep changing get yourself
1:58:49 Loosey Goosey hang out here listen to to
1:58:52 an idiot
1:58:54 ramble and start thinking about if I
1:58:58 could do
1:59:00 anything what would I choose to
1:59:03 do huge okay so this elements thing we
1:59:08 now have our model we've got sunglasses
1:59:11 we have a jacket and we have a place so
1:59:13 we're going to
1:59:15 say
1:59:18 um woman
1:59:22 with flowers well we don't need to say
1:59:25 with flowers in her hair let's just say
1:59:29 woman wearing
1:59:33 sunglasses and a
1:59:37 coat walk in front
1:59:42 of a
1:59:45 large building let's not even say
1:59:49 Cathedral
1:59:51 um in in a
1:59:56 surreal shot where the
2:00:01 camera
2:00:03 tracks with her as she
2:00:07 walks
2:00:09 now we're going to generate this thing
2:00:13 but this is what why I said get yourself
2:00:16 in a creative partner
2:00:18 program my experience with these things
2:00:21 is sometimes what it produces is
2:00:23 mind-blowingly good and many times it's
2:00:25 so [ __ ] weird and unusable that
2:00:28 you're like okay I just wasted however
2:00:30 many credits right 35 credits um so just
2:00:33 know that you get you get weird [ __ ]
2:00:35 here um let's go ahead and make a 10c
2:00:38 clip out of
2:00:40 this we'll just generate one of them you
2:00:43 can also put in negative prompts uh Tik
2:00:45 Tock pin CEF so cool I'm at work
2:00:48 mentoring students they are required to
2:00:52 do
2:00:54 something okay here we go bang now it's
2:00:58 70 credits because I decided to do 10
2:01:01 seconds by the way
2:01:04 credits credits for video are as evil as
2:01:08 chips in a
2:01:11 casino chips in a casino are designed to
2:01:14 disassociate you from the fact that
2:01:16 those things equal money that's what
2:01:19 credits do what are 70 credits that's
2:01:22 that's meaningless to me it should be
2:01:25 like a credit per video or or something
2:01:28 a second but no it's credits you gotta
2:01:31 do the [ __ ] game you gotta figure out
2:01:33 the
2:01:38 math oh they were required to watch the
2:01:40 live Joy pie you have your students
2:01:42 watching this live hey kids
2:01:53 that's [ __ ] awesome welcome Professor
2:01:56 Kyle's here to give you some
2:02:01 lessons well well listen listen children
2:02:04 okay so here's what we're gonna do you
2:02:06 want to get out your notebooks we're GNA
2:02:08 take notes okay notes are really
2:02:10 important I'm going to be teaching some
2:02:12 Concepts here and we're going to do some
2:02:14 pra practor we like to call them where
2:02:17 we do some Hands-On activities so what
2:02:19 you're going to do is when I'm in a
2:02:21 lecture mode where I'm teaching you
2:02:23 things you know one thing after another
2:02:26 kind of in bulleted form fashion you're
2:02:27 going to want to take notes and you're
2:02:29 going to want to as you're taking them
2:02:31 you're going to want to think about them
2:02:32 and concentrate on them as you're
2:02:33 writing them and you get into this sort
2:02:35 of multimodal learning way if you do
2:02:37 that um I would not encourage you to
2:02:40 take notes on a on a on a computer or or
2:02:43 use artificial intelligence artificial
2:02:45 intelligence really is kind of cheating
2:02:46 I don't know if you know this we use
2:02:49 artificial intelligence detectors here
2:02:51 to make sure that you're not cheating if
2:02:53 you're cheating you will be eliminated
2:02:55 from the class okay fantastic all right
2:02:59 all right um welcome kids if anyone's
2:03:03 here that's [ __ ] awesome how do you
2:03:05 get into creative partner programs every
2:03:08 single one of these
2:03:10 um sites they will usually have a join
2:03:14 our creative partner program um or just
2:03:17 search Google for it you know like um
2:03:20 video service you know here like Halo
2:03:23 creative partner program and it'll
2:03:25 probably take you to that to that page
2:03:27 it's usually a Google
2:03:29 form please speak more on this which
2:03:32 companies have it most of the video
2:03:34 companies have it because the video
2:03:37 companies
2:03:39 um they want they want filmmakers right
2:03:43 now Joy pie who's got her students hang
2:03:46 out in class here Joy p is a really good
2:03:49 example of someone who had like a 30
2:03:52 career not in storytelling but has
2:03:55 always had stories in her head that
2:03:56 she's wanted to tell but has never been
2:03:58 able to and Along Comes Ai and when she
2:04:03 when I saw the first video of Joy
2:04:06 perties it wasn't the first one she made
2:04:09 um but it was the first one that I saw
2:04:11 it was so good that I'm like you you're
2:04:13 you're a filmmaker right like you've
2:04:15 done this before and she was like
2:04:18 no no I haven't so when I say that these
2:04:22 video companies are looking for
2:04:23 filmmakers they're not looking for
2:04:26 Hollywood filmmakers to try to convert
2:04:28 them they're looking for people who have
2:04:30 stories to tell and have taste and have
2:04:34 critical thinking and can take their
2:04:36 tool and navigate through this [ __ ]
2:04:39 Madness of the jankiness of these tools
2:04:43 and actually do something compelling
2:04:44 with it so if you've got a vision or if
2:04:47 you've figured out how to put enough
2:04:49 things together that you can send them
2:04:50 examples or you know show them examples
2:04:53 or or you know tell them that you've got
2:04:55 a screenplay that you you know you've
2:04:57 always wanted to make but never had the
2:04:59 chance you you might get in um and you
2:05:02 just got to hustle I'll tell you a
2:05:04 really fun one um Runway which is like
2:05:07 the mac daddy of these these programs I
2:05:09 applied for the creative partner
2:05:12 program nine months ago 12 months ago
2:05:15 never got in never heard anything and
2:05:18 then on Twitter one night um Runway
2:05:21 posted post forget what it was and then
2:05:25 um I commented on it and the CEO of
2:05:28 Runway commented on my comment so I
2:05:31 immediately dm'd him and I said hey
2:05:34 thanks for commenting on my comment by
2:05:36 the way I'm a screenwriter I'm part of
2:05:37 this community called the AI Salon I've
2:05:39 got this Tik Tok Channel um would you
2:05:42 let me into the creative partner program
2:05:43 I've already applied here's my or yeah
2:05:45 I've already applied his response within
2:05:48 two minutes was what's your email
2:05:50 address and within five five minutes I
2:05:52 was in the creative partner program so
2:05:55 so it's like just just pay attention
2:05:57 just like be in the game again why you
2:06:00 come to these lives on a regular basis
2:06:02 is to be in the conversation all
2:06:05 right um Okay We're Not Gonna this is a
2:06:08 good Cliffhanger fantastic 17 minutes on
2:06:11 generation time check yeah so good
2:06:14 Cliffhanger let's go back to uh to this
2:06:18 thing we're not going to be able to see
2:06:20 this video for 15 minutes so I'm not
2:06:23 waiting for another 15 minutes to see it
2:06:25 so you have to come back tomorrow to see
2:06:28 if cing successfully Incorporated these
2:06:32 four elements into a video and and we'll
2:06:36 have to see how janky it is it's
2:06:38 probably gonna be a janky [ __ ] mess
2:06:41 but you never know this is a lot like
2:06:49 gambling oh lordy lordy lordy
2:06:54 okay
2:06:56 um let me go back to this thank you so
2:06:59 much Kyle vision is my middle name yep
2:07:02 exactly yeah if you've seriously what is
2:07:05 it aurus cool name
2:07:09 aurus especially if you have idea for
2:07:11 stories or if you've got Vision like go
2:07:13 into mid Journey make a bunch of images
2:07:15 put together a storyboard when you apply
2:07:18 for these things sometimes they'll ask
2:07:19 for do you have a portfolio site do you
2:07:22 have you know work that you can show
2:07:25 point to [ __ ] that you've shared on
2:07:27 Twitter or LinkedIn or Facebook like it
2:07:29 doesn't matter just have a vision um and
2:07:32 you can get in they're looking for
2:07:33 people with vision so anyway yeah Life's
2:07:37 a gamble indeed narrator forever thank
2:07:41 you you're welcome
2:07:43 okay what day is today today's
2:07:47 Tuesday so tomorrow we'll be back
2:07:50 Friday's up in the air because I'm
2:07:52 traveling um but I'm traveling Saturday
2:07:55 at 1 p.m. eastern time I am going to be
2:07:58 in a room with actors reading my musical
2:08:01 Sydney which was born on this channel
2:08:04 I'm so [ __ ] excited so um Friday
2:08:08 night I'm going to try to do a special
2:08:10 live where we're just going to preview
2:08:11 songs from from Sydney so I'm GNA get in
2:08:13 the mood uh so if you want to come hang
2:08:15 out and hear about that um if you
2:08:18 haven't subscribed to the AI learning
2:08:20 lab YouTube channel it's learning la-
2:08:24 AI on YouTube please go subscribe to it
2:08:28 and please go hit play on a video okay
2:08:31 why is that Kyle doesn't matter just hit
2:08:33 play on one how long should it play for
2:08:35 till the end do I have to watch it no
2:08:38 you don't have to watch it you could
2:08:40 watch it it'd be kind of like this but
2:08:43 not real
2:08:48 time it's it's good to fall asleep too
2:08:51 I've
2:08:54 heard apologize for my highly irregular
2:08:56 comments having some some trouble
2:08:59 decompressing from a stressful day ah
2:09:00 Winston sorry to hear that um I hope
2:09:04 you're feeling better I hope this uh
2:09:06 this evening brought you a little
2:09:08 levity 12 minutes left for that thing
2:09:11 okay uh I'm gonna get out of here do I
2:09:13 have anything else Brandon that I need
2:09:14 to cover
2:09:15 um oh AI if if you if you um if you're
2:09:21 wondering are we on track for are you
2:09:23 ready for AI if you go to are you
2:09:25 readyfor ai.com this is a course that
2:09:28 we're putting together based on AI
2:09:30 festiv us which we did a few weeks back
2:09:32 um our goal was February 1 I'm
2:09:35 traveling um and and Murphy just got a
2:09:39 job and Vicky's got beta testing to do
2:09:42 so we're we're pushing it for two weeks
2:09:44 so think mid-February is when that's
2:09:46 going to happen but um this is an AI
2:09:49 Readiness training that we're together
2:09:52 that we're really excited about um we're
2:09:54 going to push it for two weeks so
2:09:55 mid-February it'll be out but you can go
2:09:57 sign up uh at at are you ready for
2:10:00 ai.com and we'll let you know when
2:10:02 things are coming so that's
2:10:04 that um we we love AI for Valentine's
2:10:07 Day yeah we should do a little we'll do
2:10:09 a little campaign um join the AI Salon
2:10:12 if you haven't joined the AI Salon go
2:10:15 there click on the button that says join
2:10:18 our community and then go join our
2:10:21 community step step two you know step
2:10:23 one is just watch a little welcome video
2:10:25 step two is introduce yourself go
2:10:28 introduce yourself in the AI Salon I
2:10:31 cannot tell you how important it is and
2:10:35 is going to be moving forward for all of
2:10:37 us to be in community around this AI
2:10:41 stuff it's going to be critical to your
2:10:43 employment I promise you promise you
2:10:47 shit's about to get weird 2025 is going
2:10:50 to get weird
2:10:52 you want to be in this game so go join
2:10:55 the salon it doesn't cost anything
2:10:58 okay that's it I'm getting out of here I
2:11:00 hope you all had fun tonight tomorrow we
2:11:03 will start with cing and see how our
2:11:05 video was and um and then we'll go play
2:11:08 with some other tools and we'll keep
2:11:10 tomorrow we're going to keep digging
2:11:12 into this idea of control in video all
2:11:15 right so I hope you had fun tonight and
2:11:17 peace out oh wait there's some comments
2:11:19 let me just make sure I didn't miss
2:11:21 anything saying Heather I'm a nurse
2:11:24 practitioner so I can manage everyone's
2:11:25 medication
2:11:28 perfect Heather's gonna put put uh the
2:11:31 AI learning lab content on replay all
2:11:33 night I love that thank you that really
2:11:36 helps we're we're just about to cross
2:11:37 the threshold where I can start to do
2:11:39 some business [ __ ] on the YouTube
2:11:41 channel so thank you for for that it it
2:11:43 means a lot so okay peace out everybody
2:11:46 have a fantastic evening