aiLL 9/6/23 - How Today's Janky AI Will Reshape Tomorrow's Work

Description

Chapters

Transcript

0:05 thank you
0:21 all right
0:23 let's get this show on rude
0:27 let's get this show on the road
0:29 Wednesday Siri stop bugging me
0:34 stop bugging me
0:37 [Laughter]
0:39 welcome everybody welcome welcome
0:41 welcome Happy Wednesday US Open update
0:44 Carlos alcaraz just won set number one
0:51 in the quarters against zavera zaverev I
0:56 think it is
0:57 Ted looks at life what's happening
1:00 Silver Fox welcome welcome welcome
1:02 welcome welcome
1:05 oh
1:07 a little stiff little tight
1:10 there's a I had a crappy day at a crappy
1:14 day kind of whiskey after dinner I've
1:17 pre-whiskied now I'm back to coffee
1:19 welcome welcome everyone we'll get
1:22 started here in a minute
1:27 Silver Fox shared the live and Tick Tock
1:30 is telling me
1:32 hey I was just flying by the page and
1:34 you were like yeah I just I just went
1:36 live silver fox just happened
1:40 I waited for uh for alcaraz to win the
1:42 first set of the uh of the open tonight
1:46 all right let's see 803 getting rolling
1:50 it's cool here
1:53 um
1:55 welcome to the AI learning lab as
1:58 evidenced by The Groovy sign these
2:00 groovy signs were made courtesy of a new
2:05 sweet sweet image generation tool called
2:07 ideogram if you haven't seen it you
2:10 should go see it it does letters
2:14 it does letters like janky if you want
2:17 to sign it says janky piece of [ __ ]
2:20 you can have it
2:21 [Laughter]
2:24 oh yeah Nancy welcome
2:28 Ted's here travel lovers Escape
2:31 I like it
2:34 Tobias what's happening man
2:37 you got into the Irregulars channel on
2:40 Discord right you have the role
2:44 so hopefully you're in there
2:47 all right
2:48 I'm a little stuffy I had a crappy day I
2:51 had a whiskey
2:53 today was weird man everything was weird
2:57 started out the morning on a meeting
2:58 where we showed up and there were a
3:00 bunch of people on the meeting and we
3:01 weren't supposed to be there it's a
3:03 rescheduled meeting it somehow ended up
3:05 on my calendar as if I made the meeting
3:07 someone screwed something up but I mean
3:10 not any of the people some technology
3:12 screwed up it was just bizarre and then
3:15 the whole day was just off like that
3:17 just crap was weird anyone else have
3:19 janky weird ass crap going on we've got
3:21 like seven planets in retrograde or some
3:24 crap like that so the universe is is not
3:27 really happy right now so
3:30 Farmers should be rolling into Denver in
3:32 an hour or so oh cool Tobias
3:35 very very cool
3:39 um hey everyone hey Winston what's
3:41 happening
3:42 got into the group awesome terrific
3:45 welcome everybody my name is Kyle
3:47 Shannon this is the AI learning lab
3:50 there we go that's a good one we'll go
3:52 with that one it's good and clear
3:55 um so what we do here is uh my
3:58 background I'm an entrepreneur and a
4:00 founder and CEO of a video wait
4:05 Kyle did you get a diatopic problem too
4:09 I don't know what a diatopic problem is
4:11 possibly
4:14 I have many I have many problems dytopic
4:17 is probably one of them Dr J hey
4:20 Irregulars what's happening
4:22 um
4:23 okay so here's what we do here uh I'll
4:26 answer questions about AI we talked
4:28 about AI I make voices we have fun I
4:31 cuss occasionally I'm a little salty I'm
4:33 a little bitter I'm a gen xer we we just
4:36 we don't rules in us don't really get
4:38 along so I do what I do if you don't
4:40 like that kind of thing I'd move on uh
4:43 some call it tacky others call it
4:45 refreshingly honest thank you uh
4:50 but if you've never tried chat gbt
4:53 before first thing we do here is kind of
4:54 go over some uh
4:56 some like where do you start if you're
4:58 trying to figure this stuff out where do
5:00 you start well you start right there
5:01 wait let me flip these
5:03 put myself in a Box nobody puts baby in
5:07 the corner
5:08 all right so these things
5:12 this is like my little list of you know
5:14 everyone's got their let's bass what's
5:16 best this isn't necessarily what's best
5:18 I suppose it is what's best but it's
5:19 like where do you get started with this
5:21 AI [ __ ] everyone's talking about AI what
5:24 are you supposed to do
5:25 well if you're brand new to it if you
5:27 haven't played with chat EBT just go
5:29 right there go there right now
5:31 chat.openai.com I think you have to sign
5:34 up you know get an account or whatever
5:35 but it's free
5:37 and you can you can pay 20 bucks a month
5:39 to get the fancy ship but the the base
5:41 one the uh chat GPT 3.5 is really good
5:44 so just go there and just start playing
5:47 just start putting stuff in while you
5:48 listen to me ramble
5:50 and then you can also play with Claude
5:52 and Bard if you have played with this
5:54 stuff before this top three list these
5:56 are kind of the three that I bounce
5:59 between I actually today what was I
6:01 trying to do I was trying to get oh I'm
6:03 giving I'm giving a like a an hour-long
6:07 talk at a policy Center at the
6:10 University of Colorado
6:12 and I don't know crap about policy so
6:15 I'm trying to get these large language
6:16 models to prepare me and like what do I
6:19 need to know about policy and what are
6:20 people doing and what are some uh
6:23 articles that people have written and so
6:26 um chat sheep is not connected to the
6:28 internet Cloud's not connected to the
6:30 internet Bard is connected to the
6:32 internet
6:32 but it would give me the names of
6:35 articles but it wouldn't link to them it
6:37 said well I'm sorry I can't do that so
6:39 then I went to Bing which is also
6:42 connected to the internet and it gave me
6:44 links to articles but like the name of
6:46 the Articles were wrong and it linked me
6:48 to different articles it was just
6:49 bizarre it was just nothing was helping
6:51 today so I just gave up I just I just
6:53 said screw it so let's make sure that I
6:57 don't need actual research
6:59 uh Bing is a chat gbt as well that's
7:02 Microsoft's version of it it can also
7:04 make images for free it's free uh the
7:07 safety guard rails on Bing are a little
7:08 annoying as our Bard I found out today
7:11 um perplexity.ai is actually a good site
7:13 that's connected to the internet as well
7:15 and labs.perplexity lets you play with
7:18 meta's new llama two models and their
7:20 code llama model and then Pi is chatty
7:24 GPT if you want to talk to one of these
7:26 uh these AI tools like like a robot in a
7:29 science fiction movie Go download the PI
7:32 app and put it in voice mode it's insane
7:35 um if you want to get [ __ ] done don't
7:37 use pie because it's too chatty but um
7:39 but if you want to talk to something
7:40 it's actually really interesting and
7:42 then pose worth exploring as well the
7:44 other thing that people always ask about
7:47 is what about pictures how do I make
7:48 pictures cut I don't know how many
7:49 pictures got I want to make pretty
7:51 pictures Kyle how do you do it how do
7:52 you do what do you do how do you do it
7:54 you just put in text and then it makes
7:55 picture what do you do how do you do it
7:57 okay these are them
7:59 these are the ones
8:00 um ideogram I I'd start with ideogram
8:02 right now because it can do text so it
8:05 can do all of these images that I show
8:07 here
8:08 with images and words these were all
8:11 done in ideogram you can't do this in
8:13 mid-journey
8:14 mid journey is the best from a quality
8:17 perspective but it can't do text
8:19 clip drop in Leonardo are also really
8:22 good they're based on stable diffusion
8:24 so mid journey is its own thing
8:26 ideograms its own thing stable diffusion
8:28 as a whole ecosystem of crap here's two
8:31 places to start Leonardo's a really cool
8:33 website on top of stable diffusion and
8:36 did he just break him
8:38 yes he uh no he didn't
8:41 um
8:43 Leonardo is just a cool website in front
8:45 of stable diffusion and clip drop is
8:47 from stability AI the people that make
8:49 stable diffusion so it's a bunch of Cool
8:51 Tools from them all those are worth
8:53 playing with and just start start here
8:55 here's my advice if you're just getting
8:57 started with this stuff just start
8:59 playing just get curious just start
9:02 playing you can't break it
9:04 you'll find Happy accidents at some
9:07 point you'll stumble into something
9:08 where it will blow your mind and you'll
9:10 have a your Kevin McAllister moment your
9:14 your home alone moment where you're like
9:16 holy [ __ ] I didn't know they could do
9:17 that
9:19 um if you're not quite what to do in
9:21 these image things
9:22 openart.ai promptbook will teach you how
9:25 to talk to the image generation tools so
9:28 just like prompts.chat does for these
9:31 things I don't think I talked about that
9:33 before prompts.chat will teach you how
9:35 to talk to the large language models and
9:39 prompt book at openart.ai we'll teach
9:42 you how to talk to the image models and
9:44 then futurepedia.io is
9:47 um a big directory of sites of AI tools
9:50 okay now
9:52 AI learning lab very popular place I a
9:56 lot of people don't know this but I'm
9:58 very very famous and I'm so famous that
10:00 we have a scammer
10:02 we have a scammer that copied
10:05 the channel name and put an extra l in
10:08 it because it's really hard to see if
10:10 there's one or two L's in it so AI
10:12 learning lab with 1l that's me AI
10:14 learning lab with two L's that's the
10:16 [ __ ] that's probably on this live
10:17 right now waiting it for it to be over
10:19 so he can reach out to you and ask you
10:21 to buy some of his shitty crypto stuff
10:24 so don't buy his shitty crypto stuff
10:26 it's shitty it's shitty crypto and it's
10:30 a scam and [ __ ] him so if you want to
10:33 report him I would appreciate that you
10:35 can report him for copyright violation
10:37 you can report him from impersonating a
10:39 celebrity like me I'm famous I look like
10:42 Rosie O'Donnell I look like fat Alec
10:44 Baldwin
10:47 I I look like Tom Arnold I've been told
10:49 and and and some other guy related to
10:53 Roseanne and a lot of people I'm nearly
10:57 famous I'm so close I'm this close to
10:59 being famous
11:01 so you can report him for that you can
11:03 just report him I'm pretty sure Tick
11:04 Tock has uh he's an [ __ ] so you just
11:07 report him for being an [ __ ] all
11:08 right fantastic go [ __ ] yourself scammer
11:11 okay moving on
11:14 so if you have questions about AI pop
11:16 them in the comments below I will do my
11:18 best to answer them I promise you
11:21 I will get behind on the comments so
11:24 much so that uh you're just gonna be
11:26 like he's not even real he hasn't
11:28 answered any of my questions I've been
11:30 in here for 45 minutes it's not possible
11:32 oh it's possible
11:34 sometimes someone will ask me a question
11:35 I go down a rabbit hole we're down the
11:37 rabbit hole for a while and then I
11:38 answer the next question and then I'm 99
11:40 questions behind so
11:43 um yeah all right so let's do this let's
11:45 get rolling let's get rolling let's get
11:48 singing let's get Dancing Yeah Yeah Yeah
11:52 I feel like I think like Bob Dylan
11:56 in the 60s yeah yeah
11:59 um
12:03 you're like
12:04 is he touched oh I'm touched
12:09 okay
12:12 let's see
12:14 got into groups
12:16 oh I'm the creator of janky Ted looks at
12:19 life took me a while to get ideogram to
12:22 recognize the word yeah thank you Ted I
12:24 I am a big fan of the janky graphic and
12:28 uh I used it last night for the first
12:30 time and and well you know what let's
12:32 use it again now so so here's the thing
12:34 I just spent 10 minutes telling you
12:36 about all these Cool Tools right like
12:38 chat GPT is amazing and awesome and
12:41 clawed and barred and all of them and
12:44 you can go learn how to talk to them at
12:45 prompts.chat all of these things
12:49 our janky pieces of [ __ ] so you need to
12:52 know that
12:53 what's that mean Kyle well they
12:55 hallucinate which is
12:57 um the engineers thank you for that they
13:00 they uh the engineers rather than saying
13:03 they lie came up with this term they
13:06 hallucinate wow wow wow uh so so when
13:10 you get answers from chat GPT check them
13:15 if it gives you a link to a citation for
13:18 an article
13:19 click on it see if it goes anywhere this
13:21 good chance it won't now there's there's
13:24 ways that you can make it better and do
13:27 that less but just know that that
13:29 happens
13:31 um sometimes these things are brilliant
13:33 to the point that they melt your face
13:35 off sometimes these things are
13:37 frustrating to the point that you're
13:38 like really so they're completely
13:40 useless
13:41 um
13:42 more the former than the latter but just
13:44 know that they're all kind of and here's
13:46 the deal it's not that any of the
13:49 development teams Behind These are are
13:51 bad like the the underlying technology
13:53 under these things is [ __ ] magic it's
13:55 just [ __ ] magic
13:57 um but we're very very new chat GPT is
14:01 as much as everyone's been talking about
14:02 AI Chachi PT came out November 30th of
14:05 last year so it hasn't even been a year
14:08 since chat EPT has been out so all of
14:11 these tools are very very new they're
14:13 very raw and that's why they're a little
14:15 janky and that's why I have the Yankee
14:17 graphic they're just Yankee pieces of
14:19 [ __ ] so just know that uh and still
14:21 learn them still learn to play with them
14:23 and and do all that happy stuff so all
14:26 right
14:27 let's get rolling
14:30 um
14:32 the Irregulars are coming in by the way
14:35 the the the the there's a gang in here
14:38 this this is a little gang gang in the
14:41 AI learning lab where uh people come in
14:44 here every night and they they call
14:45 themselves The Irregulars don't mess
14:47 with them they're not normal they are
14:49 irregular so you might see lots of
14:51 Hearts going up that's that's not
14:54 because they love me that's because
14:55 they've got a side game going they got
14:58 some sort of bet with themselves and uh
15:00 and that's what's happening there but
15:02 they're awesome and actually they they
15:03 do a lot of help so as I get behind
15:05 answering these questions if you have
15:07 questions a lot of times the Irregulars
15:09 will answer questions and say things
15:10 like he'll get to your question
15:12 eventually there Betty
15:14 he tries to answer them all but he's old
15:17 you know he's old
15:21 so they support me like that
15:26 being the being the king of the kids
15:29 table it's so true
15:31 big ping ping is at the kids table and
15:35 and you know I thought I might because I
15:38 got so frustrated with Bard today I
15:39 almost moved barred
15:42 um back to the kids table and Bing back
15:45 to the grown-ups table but no Bing
15:48 stayed here it frustrated the [ __ ] out
15:50 of me again today
15:51 uh what are you gonna do
15:54 what are you gonna do these tools what
15:56 are they come on people Yankee pieces of
15:58 [ __ ] exactly exactly
16:02 all right let's get let's get some
16:04 graphics going here let's see which
16:06 graphic are we gonna use I think I'm
16:08 gonna use that one because it just looks
16:11 it looks I don't know it looks like the
16:13 future
16:15 looks at the future right next to my
16:17 head there's a little glow
16:20 all right see your YouTube live
16:22 conferences
16:24 for students and others you give and
16:27 awesome oh you you saw my YouTube live
16:29 conference for students and others you
16:31 give a great presentation thank you very
16:33 much I assume you mean the uh I did a I
16:36 did a talk for Creative mornings was it
16:38 that one
16:40 um where I talk about all this stuff
16:41 yeah that was a fun talk and it was very
16:43 well received so thank you I'm glad you
16:44 enjoyed that Kyle which browser do you
16:47 use I'm using Edge so so here's the deal
16:51 I've been I've been an apple Fanboy
16:53 since
16:55 the the Apple II like before the Mac was
16:58 around
16:59 I have hated Microsoft with a passion
17:02 for decades I've been a Microsoft basher
17:06 if you had bet me a thousand dollars a
17:09 year ago that my default search engine
17:11 would be Bing and my default browser
17:15 would be Edge
17:16 I would have laughed in your face and
17:19 bet you the thousand dollars and yet
17:22 here we are my default search engine is
17:25 Bing and my default browser is Edge now
17:29 is it great it's good except it's got
17:32 such memory problems that I have to like
17:35 if it [ __ ] tanks my computer about
17:37 every third day and and chrome only did
17:40 did that like once every three weeks so
17:43 it's got it it it leaks like a sieve
17:47 um but it's got some slick things to it
17:49 and it's just the the integration of all
17:51 the the chat GPT [ __ ] is actually just
17:53 kind of nice and you know I wanted to
17:57 experience what it was like to to play
17:59 with something that that was that
18:01 integrated and and you know it's it's
18:03 interesting it's good
18:05 um I actually kind of like being as a
18:06 search engine it's fine
18:08 um I like being able to just flip into
18:10 chat GPT mode I mean most of the times
18:13 when I go to Bing
18:15 I go immediately into chat I don't I
18:17 don't really use search engines anymore
18:20 I don't like them I can't stand them
18:22 they're just also SEO injected like like
18:26 I I just
18:27 they're just awful
18:30 I just want the answer I would I would
18:32 rather wait for gpt4 to slowly give me
18:35 the wrong answer four times in a row
18:37 then Google to give me a really quick
18:40 response of shitty links that I have to
18:42 go sift through
18:43 right and the first 10 are just you know
18:46 SEO garbage sites
18:51 or or you know they have so many ads in
18:53 them that you literally can't find the
18:55 content so I'm like [ __ ] it I'll I'll
18:57 wait
18:59 I'll wait at this point I'm so pissed
19:01 off at the state of search and the state
19:03 of the art that that uh sometimes now I
19:07 just I just go to the the large language
19:09 models just out of spite yeah [ __ ]
19:11 Google
19:13 but uh it's it's interesting times
19:16 interesting times
19:20 um were you one of Jerry's Kids yes I am
19:22 so RJ I uh I met Jerry mikulski so
19:26 Jerry's Kids Jerry mikulski is a
19:29 remarkable human being he's one of the
19:31 most networked human beings I know
19:34 really smart really passionate
19:36 and I met him when he was the editor or
19:39 whatever his role was managing editor of
19:42 uh Esther Dyson's release 1.0
19:45 and then he started this thing called
19:47 Jerry's Retreat where he invited people
19:49 that he had interest in conversations
19:50 with to this three-day Retreat and
19:53 that's where I met my co-founder Monique
19:54 so RJ my
19:57 um my company storyvine I founded it
19:59 with Monique who I met at Jerry's
20:01 Retreat and our senior advisor is a guy
20:04 named Kevin Clark who I met at Jerry's
20:07 Retreat so yeah so I am one of Jerry's
20:10 kids and
20:11 uh it's a that's a remarkable community
20:13 and I um
20:16 I feel bad about two years ago I got all
20:20 hype I got all like crack addicted to
20:22 blockchain and nfts and I told all of
20:25 Jerry's Kids oh nfts are going to be the
20:26 big thing blockchain's gonna be a big
20:28 thing
20:29 and then it kind of [ __ ] the bed
20:30 appropriately so and then uh and then I
20:33 did the same thing with AI but AI is
20:35 actually going to change the world and
20:36 and it's going to do it in a profound
20:38 way so uh so that's why this channel
20:40 exists but anyway
20:42 um yeah what's what's your name pop it
20:44 in the comments that's very cool
20:47 um
20:48 do Kyle or anyone know the best image
20:50 API to work with programmatically well
20:52 you can use
20:54 um stable diffusion's got one but it's
20:56 open source so you got to find a place
20:58 where it's hosted or host your own
21:00 um
21:01 probably the easiest one is the Dali API
21:04 because you can just use it like like if
21:07 you use zapier automations
21:10 um and you use the
21:12 um the open AI
21:14 uh widget within zapier you can just
21:17 make an image call or a chat GPT call
21:20 like a GPT call and
21:23 it'll make images for you and then
21:25 um Mid Journey does not have an API yet
21:30 and
21:32 I don't think any of the Adobe stuff
21:34 does
21:36 there's probably something else out
21:37 there but I forget
21:39 but I think the easiest is probably just
21:41 the the Dali one within openai that said
21:44 Dali sucks like it's just not as good
21:47 like like mid journey is really good I
21:49 know that there's a um
21:52 I'm working with a group that's that's
21:53 working on doing some stuff with images
21:55 and
21:57 I know that someone just came out with a
22:02 it's a tool that's got an API where you
22:04 can generate mid-journey images but I
22:06 think what they're doing is actually
22:08 scraping a Discord server with a
22:10 mid-journey channel in it
22:12 and then sort of passing the [ __ ] to it
22:15 to their API so so it's kind of a fake
22:18 API but the problem with that is they're
22:20 going to charge you 30 bucks or whatever
22:22 you're going to build an application on
22:23 it and then mid journey is going to go
22:26 uh [ __ ] anyone who's scraping our thing
22:27 they're gonna you know they're gonna
22:28 change their service and then and then
22:30 that thing's gonna break so
22:33 um so I I would love for mid-journey to
22:35 have an API but I don't it doesn't look
22:38 like they really give too much of a [ __ ]
22:39 right there
22:41 um I can locate Google teapots for
22:44 mid-journey uh oh do you want you want
22:46 me to find those they're hard to find
22:48 where did I
22:50 what did I just do
22:52 oh I see
22:54 um hang on so what what uh toe Dance by
22:58 the way toe dance good name I like the
23:01 name
23:02 9.5 out of 10 on the name
23:05 um dancer you a ballet dancer that kind
23:08 of toe dance
23:09 um
23:10 the teapots for mid-journey and stable
23:12 diffusion let me see if I can go find
23:13 that so so what toe dance is referring
23:17 to is this really amazing resource that
23:20 shows you the power of these tools
23:23 um what was it called
23:25 something teapot um let's just say
23:28 teapot
23:32 mid
23:35 Jersey let's see
23:41 teapot sets by mid-journey
23:45 that's it
23:59 [Music]
23:59 um
24:02 teapot mid-journey reference
24:13 [Music]
24:28 [Music]
24:36 I know I have it linked somewhere I
24:38 can't remember what I did with it though
24:42 in one of my videos in my Tick Tock is
24:44 is a thing that talks about the
24:48 the teapots
24:55 well I'll keep looking I don't wanna I
24:56 don't want to waste a ton of time I'm a
24:58 Little Teapot mid-journey assistant
25:01 I spent 40 hours of my life on this
25:05 no
25:06 uh I don't know what it is is someone
25:09 took the a prompt of just a specific
25:12 kind of teapot I forget what it's called
25:13 or I could find it and then they they
25:16 applied all these different styles to it
25:18 like in the style of rembrand and the
25:21 style of of Dali in the style of
25:23 whatever
25:24 um you know uh uh watercolor oil
25:28 painting you know acrylic things like
25:30 that and it and it's so this single
25:32 prompt and it's it's just a spreadsheet
25:34 full of you know hundreds and hundreds
25:37 of different renderings of the teapot
25:39 and you can see how how varied results
25:42 you can get with something like mid
25:44 Journey it's it's a pretty cool resource
25:45 but I can't find it right now
25:48 um
25:53 the documentation is moved the ultimate
25:56 mid-journey cheat sheet is this
26:04 [Music]
26:15 I mean this looks kind of cool but it's
26:17 not the it's not the teapot thing anyway
26:19 sorry I can't find it I'll see I'll see
26:21 if I can find it is it stable diffusion
26:26 or is it maybe the open AI API for which
26:30 for for
26:32 um for ideogram I don't I don't know I I
26:36 I
26:37 think when I did my my video about
26:40 ideogram I said I thought it was based
26:42 on stable diffusion and someone in the
26:43 comments said that it wasn't
26:45 that it's their own thing they raised
26:47 like 160 million dollars or something
26:49 like that or no 65 million dollars
26:53 a lot of money
26:55 um
26:56 so yes you're Steve Jobs you admire and
26:59 want to uh to join the team 10 years too
27:02 late yeah exactly that's right togas
27:06 I was too young to be part of the PC
27:09 Revolution but then I was just right to
27:11 be part of the web Revolution and now
27:14 um old and crusty for this AI Revolution
27:17 but I'm digging it this is a cool time
27:21 to be alive so it is very rare as humans
27:24 that we get to uh we get to play with uh
27:27 Technologies when they're when they're
27:29 being born they're that are going to
27:31 change everything
27:33 um yeah the PCU is one that I missed lab
27:36 I got and then this one I'm getting two
27:38 in my lifetime that's pretty [ __ ]
27:41 exciting you're pretty exciting it's
27:43 pretty good experts say Chrome is the
27:46 best browser for chat gbt because the
27:47 extensions Aid user experience although
27:50 cabruno Edge is a Chromium browser so
27:53 Edge can use all of the the Chrome
27:55 extensions
27:58 um so you can do all that anyway I
27:59 haven't gotten into extensions because
28:00 they tend to be buggy and they tend to
28:02 go out of uh they get not supported
28:05 really quickly so I just I can't
28:07 anything where I've got to manage
28:09 actively is not good for my brain type
28:13 [Laughter]
28:16 I got addicted to pop-ups and Robo calls
28:19 I saw the future then but they still
28:22 just bug me
28:24 oh that's funny uh it's something you
28:28 are regretting uh I don't know what it
28:31 is but there's there's not many things
28:33 I'm regretting but there's a handful
28:36 there's a handful oh yeah toe dance is
28:39 just living life like murderedly nice
28:43 oh I see toe dance is just like yeah
28:46 just dancing around nice I like it and
28:49 it's solid
28:51 um are you late Emilio's wife you are
28:53 not late so by the way welcome everybody
28:55 if you're new here
28:57 um if you haven't played with chat GPT
28:58 go there that's the chat GPT website if
29:02 you don't know what to say to it you
29:03 don't know what to do go there
29:04 prompts.chat that will teach you how to
29:06 talk to that and all these other things
29:08 if you have questions about AI pop them
29:10 in the comments below
29:12 um I'll do my best to answer them
29:15 um happy to talk about whatever just pop
29:17 it in there thanks Kyle we'll figure
29:19 something out but don't want to screen
29:21 scrape from mid-journey yeah I wouldn't
29:23 do that I
29:24 I would say stable diffusion is your
29:26 best is your best bet Winston just
29:28 because the developer Community for Mid
29:30 journey is insane there's all these
29:32 loras out there there's all these models
29:34 out there so if you want to get
29:38 um if you want to be able to dial in a
29:40 particular style in in the API calls you
29:43 want to make I would say stable
29:46 diffusion is the way to go
29:48 um
29:49 because Dolly just sucks right now now
29:53 is Dolly gonna get better probably but
29:56 have they had plenty of time to launch
29:59 something better and not yes so it still
30:03 sucks and they haven't even Incorporated
30:07 ah Madrid reached level 11. wow awesome
30:10 congratulations welcome to level 11. I
30:14 have no idea what that [ __ ] means other
30:16 than if you hit Level 33 some [ __ ]
30:19 happens
30:20 I think that's I think that's
30:22 disorganization I think if you hit Level
30:24 three you you have attained uh sainthood
30:26 in the AI learning lab
30:30 um
30:31 oh tick tock
30:34 and Tick Tock just jumped me to the
30:35 bottom so those of you that have been
30:37 here before you know it's the bane of my
30:38 existence those of you who don't don't
30:40 know what I'm talking about and it
30:41 doesn't [ __ ] matter okay
30:45 wait can I turn off screen text blocking
30:48 your face on the screen no the text
30:51 blocking my face on the screen I think
30:53 is an advantage I just want to cover
30:55 this up as as much as possible I
30:58 honestly don't know I think if if you
31:01 try sliding your screen left and right
31:04 um I think there's one that like like
31:06 magnifies the text all over the screen
31:08 and then so it might be a swipe writer
31:11 swipe left but I don't know I don't know
31:13 Tick Tock I'm just just an old gen xer
31:16 just trying to [ __ ] deal
31:19 um I love the janky image I know the
31:20 janky image is great right
31:22 [Laughter]
31:24 can I turn off the test thank you very
31:26 much for the donut
31:28 AI merch 99 questions behind 99 plus
31:31 questions behind It's Always 99 plus
31:36 oh man it's so true
31:41 um reporting live
31:44 wait reporting oh reporting live from
31:46 back surgery rehab hot tub in Midwest
31:49 Oklahoma
31:50 fabulous unknowns in the sky
31:52 oh good yeah hopefully you'll see Tobias
31:55 so if you see some UFOs tonight or if
31:57 like you know some crafts land in your
31:59 yard and the dudes come out like just
32:01 give us the play-by-play I'll jump down
32:03 to real time commenting if you're in the
32:05 hot tub you know meeting aliens because
32:07 we want to know that that's awesome hope
32:09 your back is feeling better back
32:11 surgeries no [ __ ] joke I've got my
32:13 back almost went out twice yesterday and
32:15 once today so I'm like
32:19 there's [ __ ] nothing worse my back
32:21 went out the other well not the other
32:23 day it was like a year ago I was sitting
32:25 on my son's bed and he goes hey can you
32:27 turn on the heater I literally I
32:29 literally went like this my back went
32:32 out
32:33 I [ __ ] went right to the floor
32:36 looking awful it's just like oh so this
32:40 for those those of you that are young in
32:42 here big piece of advice this is AI
32:45 advice don't get old it [ __ ] it's
32:48 just awful so just don't do it just stay
32:50 Wherever You Are
32:52 uh
32:54 okay
32:57 won't be able to Turbo tap till I'm out
32:59 Tobias that's just so sad
33:03 all right so if these models are
33:06 becoming more persuasive and they
33:08 hallucinate how do we protect ourselves
33:10 from BS whose thing so Jeff
33:14 you know
33:16 the the
33:18 way you protect yourself from BS with
33:20 humans right like humans hallucinate
33:22 humans lie humans write shitty stuff too
33:24 and do shitty research too I think one
33:27 of the
33:28 one of the
33:29 one of the traps of these AI tools is
33:32 that because they're so fast and because
33:34 they're so capable it's very easy to get
33:36 seduced into thinking that they're they
33:40 they know what they're saying
33:43 the reality of how the technology works
33:45 is that they don't they they actually
33:47 have no they have zero comprehension
33:50 of what they just wrote zero which is
33:52 why they hallucinate
33:54 it's actually just a mathematical
33:56 probability engine when you type your
33:58 prompt it turns it into a mathematical
34:01 weighting and it basically goes and
34:03 looks at all of the words in this
34:05 thousand-dimensional mathematical word
34:08 space or it's actually just Vector space
34:09 it's actually just numbers
34:11 and it says which of all of those
34:14 numbers those points in Vector space is
34:17 the most likely to be the next probable
34:20 word right or phrase sometimes they use
34:22 multiple words
34:24 and then it just puts those on screen
34:26 and then it does it again and it puts
34:28 those on screen so when it writes you a
34:30 whole thing it's just doing that as a
34:32 mathematical probability engine so it
34:34 doesn't know what it's doing so a couple
34:35 of things there's two ways that you can
34:38 um that you can deal with this one is
34:41 you're in control the the robots are not
34:45 doing this [ __ ] to us this is a tool
34:48 that we get to use right so you're
34:52 saying Hey I want you to give me this
34:54 information
34:55 if you're like the idiot lawyer a few
34:58 weeks back six weeks back or so that had
35:00 chat EBT write a legal brief
35:03 and then he handed it to a [ __ ] judge
35:07 and the judge checked the citations in
35:10 the legal brief and they were fake
35:13 and the judge said uh Mr lawyer what the
35:17 [ __ ]
35:19 that's on the lawyer that's not on these
35:21 things right so so so one is just have
35:24 situational awareness of what you're
35:26 creating and if you're creating
35:29 something you're going to hand to a
35:30 judge I don't know check the references
35:32 go nuts
35:35 the other thing that you can do is you
35:36 can get better at prompting so the so
35:38 one of the reasons I have prompts.chat
35:41 on screen from for most of my lives is
35:43 it if you're just playing with these
35:46 things this thing will give you a primer
35:48 and how these things work and and what
35:50 what prompts.chat is is it's a big long
35:52 list of prompts where you tell chat gbt
35:55 to act like a certain kind of character
35:57 and and basically the the way large
36:00 language models work the more context
36:02 you give them the better they get at
36:04 answering things it doesn't mean they're
36:07 not going to hallucinate they it's weird
36:10 sometimes you can get them to be really
36:11 really good and they don't hallucinate
36:13 much at all other times they just
36:15 [ __ ] go off the rails but if you give
36:17 them a lot of context and say Hey I want
36:19 you to be a history profession Professor
36:21 that specializes in Egyptian history and
36:24 blah blah blah blah and then I want you
36:26 to tell me about the history of this
36:27 particular thing in Egyptian history
36:29 you'll tend to get it very very accurate
36:31 because you've kind of narrowed the
36:34 knowledge that it's drawing from to a
36:37 smaller space if you just say tell me
36:39 about the pyramids it'll tell you about
36:42 everything from the Luxor pyramid in
36:44 Vegas to pyramid schemes to just
36:47 everything else and it's sort of
36:48 amalgamating you know words from all
36:51 these different kind of semantic
36:53 clusters
36:54 but if you if you kind of narrow it into
36:56 a space then then you're pulling from a
36:59 much more focused knowledge set and
37:01 that's that's why it makes it better
37:04 um
37:05 but but the other thing Jeff is
37:08 recognize that these things are
37:12 wait wait for it
37:16 janky pieces of [ __ ]
37:19 the these tools are very very new
37:21 they're incredibly powerful the fact
37:24 that we as mere mortals and not you know
37:27 research scientists get to play with
37:28 them is remarkable
37:30 um but they're new they're new and
37:32 they're rough around the edges it
37:33 doesn't mean you shouldn't Play Just
37:35 Because
37:37 they're not perfect get curious about
37:40 them figure out what they're good at
37:41 what they're not good at they are
37:44 remarkable at some point you will have
37:46 what I call your Kevin McAllister moment
37:48 which is the Home Alone moment of just
37:50 like I had no idea
37:53 but you kind of have to push through
37:55 playing with them and being curious
37:57 about them to get past like oh it's just
37:59 like a fancy Google search I get it I
38:02 get it it's you know I heard that they
38:04 hallucinate you know what they're
38:05 stealing from copyright and works
38:08 there's a lot of people sitting on the
38:10 sidelines soap boxing without ever
38:13 having tried them that's the thing that
38:15 I'm trying to this channel is the anti
38:17 that like I don't mind if you hate AI
38:19 but Heyday I haven't used it
38:23 not hate AI because conceptually it's
38:26 the evil Bots are gonna kill us or it
38:28 hallucinates
38:30 go play go play
38:32 go have your mind blown go figure out oh
38:35 [ __ ] it really can do something
38:36 remarkable and then figure out what the
38:38 boundaries are and you might you might
38:39 realize oh I actually don't like how
38:41 this does what it does I'm going to keep
38:43 doing things the way I do them
38:45 totally respect that love that
38:48 but most of the loud haters of this
38:52 stuff haven't tried it
38:54 because the minute you try it the minute
38:56 you have your Kevin McAllister moment
38:58 you kind of go ah I should probably I
39:02 should probably learn what these things
39:03 do do they do more than that
39:07 but surely they can't do what I do
39:11 what I've trained for 30 years to do
39:13 they surely they can't oh they can do
39:15 that too
39:18 what do I do what do I do Bob Bob can we
39:24 turn off the internet
39:26 yeah
39:28 I think Google had some kind of reset
39:31 today did they
39:34 I know I wonder if they're getting ready
39:36 for
39:37 um
39:37 Gemini
39:39 I wonder if they rolled Google
39:42 reset today
39:44 let's reset mean though
39:47 reset Google Chrome
39:50 let's see Google
39:53 news
39:58 Google will soon require
40:02 disclaimers for AI generated political
40:04 ads
40:10 I don't see anything in the news but
40:12 it's possible
40:14 it wasn't big enough Google breaking
40:17 news today
40:19 I don't see anything
40:20 but if something happened let me know
40:23 wait what wide Bing or Edge
40:27 um so I Galactic okay so let me show you
40:30 why I'm using Bing and Edge so so here's
40:33 Bing
40:34 so I'm just going to go to bing.com okay
40:38 and then hang on
40:42 okay question from igalactic which by
40:45 the way I'm gonna go nine out of 10 on
40:46 the name I love it
40:48 is it is it I galactic
40:53 alcaraz is mopping the floor was
40:56 zarev how do you say it
40:59 zverev
41:02 all right he's also 20. good Lord these
41:06 20 year olds
41:07 um okay so you come to Microsoft Bing
41:11 and
41:12 a couple of things going on here you got
41:14 a little microphone you got a little
41:15 camera you can upload [ __ ] you can do
41:17 like a normal search but
41:18 click on the chat button
41:21 and this is chat GPT
41:24 so this is pretty cool it's also got
41:27 this thing like
41:28 um more creative more balanced more
41:31 precise if you've ever played around in
41:33 gpt's playground there's a thing called
41:35 temperature it's a slider that you can
41:37 basically slide from less creative to
41:39 more creative and so Microsoft has
41:43 basically revealed that capability you
41:46 know as an end user product which is
41:48 it's pretty slick
41:50 Bing is also connected to the Internet
41:52 so I can say
41:54 um
41:56 tell me what the weather in Denver is
41:58 going to be like tomorrow period
42:01 is it a nice day to take a hike
42:06 so it's connected to the internet
42:11 according to the weather forecast it's
42:13 going to be 90 degrees
42:15 and a low of 55 seems like it would be a
42:18 nice day for a hike but it's always good
42:20 to check the weather again before
42:21 heading out
42:22 and then it's got links to the sources
42:25 that it used right so it's doing
42:27 citations right so pretty slick okay
42:30 then I could say
42:34 can you write me
42:37 write me a poem
42:39 about the weather tomorrow
42:43 that
42:45 is in iambic pentameter
42:49 so this is this is uh what I call a
42:51 parlor trick let's let's have chat GPT
42:54 do a parlor trick wait where did it
42:58 it ain't my it ain't my prompt
43:01 um
43:02 write me a poem about tomorrow's weather
43:05 in iambic pentameter
43:08 do a little Shakespeare action
43:11 all right sure
43:13 in Denver where the mountains touched
43:16 the sky the sun will shine and Rise
43:18 attempts rate high with 90 as the peak
43:21 and 55 the low a hike would make one
43:24 feel alive so it remembered that we were
43:26 going to go on a hike
43:27 so lace your boots and fill your flask
43:30 with care and don't forget the sunscreen
43:32 for your hair now you don't put
43:34 sunscreen on your hair but anyway it can
43:36 do [ __ ] like that
43:38 um it can also do this
43:40 um make me an image of the flat irons
43:45 in Boulder
43:48 and shows some
43:50 nice dogs walking on the trails with the
43:54 mountains in the distance at sunset
43:59 it can make images it can read images it
44:02 can so so
44:04 it's it's good interesting stuff I'll
44:07 try to create that
44:10 so it's got this really horrible
44:11 Microsoft looking 90s the computer is
44:15 working on your bits now graphic I I
44:18 think I can get rid of that but anyway
44:20 there's the flat irons that's actually
44:22 pretty close to what they look like
44:24 they're not quite right but there's your
44:26 doggos right so you got dogs there
44:31 so it does that
44:34 um so that's that now here's the other
44:36 thing so I'm going to go to
44:37 my company's website storybine
44:42 so here's storyline now I'm in I'm in
44:44 Edge and so in Edge you've got all these
44:47 functions down the side one of the
44:49 things it's got is it's got a a bing
44:52 sidebar
44:53 like a Chachi PT sidebar that is aware
44:57 of the sight to the left of it so first
45:00 thing it did it's in creative mode it
45:01 grabbed a summary of storyline I'm going
45:03 to put it in precise mode
45:06 but it says what is storyvine
45:07 storyvine's a video platform what are
45:09 the benefits of storyvine how does
45:11 storyvine work
45:12 storyvine uses a video guide that asks
45:15 specific questions one at a time to
45:17 structure the storytelling that's
45:18 correct and that's not
45:20 um text from our site the app records
45:23 answers and uploads them to the cloud
45:25 where they're stitched together with
45:27 graphics music and transitions you can
45:29 review the videos in the back end
45:31 dashboard and share them on social media
45:34 right and then learn more here's here's
45:36 your links let's see where it takes us
45:39 yeah so so now it's showing us so it's
45:41 got a link here and it's saying here's
45:44 the here's I wrote this from this
45:46 information right
45:49 so so I can start asking now like
45:52 um
45:54 what kind of companies does storyline
45:56 work for
46:00 um
46:01 [Music]
46:06 story finds a Denver based company
46:12 variety of companies including Alcon the
46:15 hospital Corporation of America and
46:17 America succeeds
46:20 platforms designed to fill the gap
46:22 between expensive user gen or
46:24 professional video and hard to control
46:26 user generated video so I got it mostly
46:28 there in essence storyline works for
46:31 companies that are looking to create
46:32 authentic branded video content in a
46:34 structured way without the high costs
46:35 typically Associated so so anyway so
46:39 it's just like just that ability to
46:42 no matter where I am and what I'm doing
46:44 to just quickly flip into chat mode
46:46 without having to you know go over to
46:48 another site is is pretty slick
46:52 so that's that's why I shifted and and
46:54 quite frankly like I don't know I was
46:57 kind of done with Google I think they
46:59 are so over SEO optimized and they
47:01 effectively had a monopoly so part of
47:04 this is Google Microsoft is acting like
47:07 a startup for the first time in decades
47:09 and they're [ __ ] with Google and I
47:11 think it's fun so
47:14 so I'm gonna I'm gonna reward their bad
47:16 attitude
47:20 oh man yeah can you explain the layers
47:24 in a neural network
47:26 not really not not really I can explain
47:30 how chat GPT works and I can explain how
47:33 diffusion models work
47:36 um but once you get into the sort of
47:37 data science of it below at the neural
47:39 network level that's that's below my my
47:42 uh my area of expertise uh you might
47:45 also ask what are your qualifications
47:47 sir so here's the deal this channel is
47:50 about the inflection point that that
47:53 we entered on November 30th 2022 so
47:57 prior to that inflection point you had
47:59 to be a mathematician a scientist a
48:01 researcher an engineer
48:03 if you wanted to play in machine
48:05 learning
48:06 I've been fascinated with machine
48:08 learning for a long time like decades
48:12 and in the past seven years or so
48:16 Google would put out these announcements
48:18 oh there's a new version of tensorflow
48:20 that's really easy to use and I'd go oh
48:22 that goody goody let me go play with
48:23 tensorflow I would go look at the
48:26 documentation and I would go oh that
48:29 makes me sad
48:30 I still haven't gone for Stanford and
48:32 got my mathematics degree yet I'll get
48:34 to it I will I promise
48:37 and then
48:38 um and then in early 2022 I started
48:41 playing with these image generation
48:42 tools I started on a an nft project
48:45 called pixelmind which is a really cool
48:47 project built on top of stable diffusion
48:50 and then I [ __ ] slammed my face
48:52 against the brick wall and failed nine
48:55 times to get stable diffusion up and
48:57 running on a Google collab server and I
48:59 finally got it and then I got dream
49:01 Booth working and then I started making
49:03 all these crazy images
49:06 um
49:07 and I created an art project called Kyle
49:09 Shannon dreams and
49:12 created all these crazy sort of weird
49:15 self-portraits of myself but not myself
49:19 and I'm like this this generative AI
49:21 stuff like so so like now I'm starting
49:23 to play with this stuff and then
49:25 November 30th chat GPT comes out and
49:28 it's what's funny is I had [ __ ] around
49:30 with gpt3 and the GPT playground you
49:34 know while I was learning about this
49:35 stuff you know you kind of bounce around
49:36 you learn this stuff I had played with
49:39 that
49:39 I don't know for six months and I sort
49:42 of got it and I sort of got that you
49:44 could do cool stuff and it did cool
49:45 stuff and then Chachi PT came out
49:48 and within about two weeks I had my my
49:52 first Kevin McAllister moment I had this
49:55 moment where
49:57 um within 90 minutes never having
50:00 written the line of python
50:03 I created a fully functional python
50:05 application that took some inputs from a
50:08 user and generated five pieces of social
50:10 media content
50:12 and I was like
50:14 holy [ __ ] the world has changed and it's
50:17 the same Epiphany I had in 1994 when I
50:21 published an online magazine using this
50:24 crazy technology called the world wide
50:26 web and three weeks later it was written
50:28 about in Paris
50:30 and I was like holy [ __ ] the world just
50:32 changed like those moments don't come
50:34 along that often so the minute that
50:37 happened with chat GPT I just I just
50:40 kind of went ballistic I started the AI
50:42 Salon I started this Tick Tock Channel
50:45 um I started a newsletter I completely
50:47 started rethinking my company's value
50:50 proposition and what our product is and
50:52 how it works and I haven't looked back
50:55 and it's just like every day I'm doing
50:57 everything I can to learn as much as I
50:59 can about it but
51:00 I'm sort of post inflection point so
51:03 when if you have questions about stuff
51:06 down in the you know engineering side of
51:09 things I'm not the one to ask but the
51:11 good news is there's lots of resources
51:12 for that right that shit's been around
51:14 for decades
51:16 so but if you want to know how this [ __ ]
51:18 works and what the implications are and
51:20 how I think it's going to impact work
51:21 and how I think it's going to impact
51:22 people and all that sort of [ __ ] that's
51:24 all I've been thinking about and all
51:25 I've been doing for the past year and a
51:27 half
51:32 [Music]
51:36 I actually kind of like Bing he said
51:39 and then he knew he would never be young
51:42 again I know I Galactic listen I have I
51:45 like it hurts me inside to say good
51:48 things about Microsoft products it
51:51 really does
51:53 but but you know I
51:56 app like they just sucks ever ever since
51:58 Steve Jobs died like there's there's no
52:01 one there that's [ __ ] maniacal about
52:03 quality and Innovation anymore
52:05 there are people there that are
52:07 passionate about it but there are no
52:08 [ __ ] maniacs
52:10 so and their products have suffered as a
52:12 result
52:13 um I think the uh
52:15 The Vision Pro things are actually
52:17 promising but you know it's a little
52:19 disappointing like their [ __ ] doesn't
52:20 just work like it used to
52:24 Apple has gotten jankier
52:27 it's not janky but it's gotten jankier
52:29 so anyway so I'm with you I know
52:32 I know Microsoft old people I get it I
52:35 don't disagree but they've they've
52:37 Microsoft has acted more like a startup
52:39 than Google has for a year now
52:44 and and like gloating in it
52:46 and there's something really exciting
52:49 about that and they've they've [ __ ]
52:51 knocked Google back on their heels and
52:53 Google has been spending the last year
52:55 scrambling to try to catch up
52:57 with the technology they [ __ ]
52:59 invented that's the thing that blows my
53:00 mind they invented the transformer in
53:03 2017 and then they [ __ ] mothballed it
53:06 why because it threatened their business
53:07 model
53:08 they didn't expect little old snot-nosed
53:11 Sam Altman to like whip up a little chat
53:13 GPT that would get 100 million users in
53:16 six weeks no one expected that including
53:18 Sam Altman
53:20 but when that happened and Microsoft
53:22 went from We're investing a billion
53:24 dollars to like a month later oh make it
53:26 10 billion dollars and we're rolling it
53:29 into everything we do
53:30 because they knew they had they had not
53:33 Google for a loop
53:35 I find that really exciting some actual
53:38 competition in a space that it that had
53:40 an effective Monopoly for 15 years
53:43 I mean November of 2022 Google's market
53:48 share was 92 or 94 of the search Market
53:56 I wonder if they've lost Danny
53:58 I heard that
54:00 what was it every for every percentage
54:03 point in search
54:07 it was like two billion dollars for
54:09 Microsoft
54:11 some some crazy amount of money but but
54:13 the CEO of Microsoft said he's just he's
54:17 he wants to [ __ ] with Google he'll he
54:19 will take a loss on chat GPT just to
54:22 mess with Google
54:24 um let's see uh what is Google's
54:30 market share in search
54:36 92 is that when is this from
54:44 uh
54:47 [Music]
54:50 it's still steady
54:52 they're still steady nothing's moved
54:54 they still effectively have a monopoly
54:55 here's the other thing about Google
54:58 um
54:59 they um
55:02 they're pissed off
55:05 and they're they're apparently working
55:07 on this thing Gemini which is is
55:09 supposedly a chat GPT 4 killer I'll
55:12 believe it when I see it but but they're
55:14 gonna they gonna come out swinging it's
55:17 gonna be fun I mean here's here's the
55:20 deal when these big companies start
55:21 throwing billions of dollars around like
55:24 apple apple said a month ago they spent
55:28 the bulk of their research budget in
55:31 2023
55:33 on generative AI
55:35 Apple's research budget is
55:39 24.6 billion dollars
55:43 so who benefits when these big companies
55:46 start swinging their
55:49 their private parts around at each other
55:51 we do we get to play with all the toys
55:53 because they're trying to get us to
55:55 click on their [ __ ] so
55:59 um so that's that's why this stuff's
56:01 really exciting because
56:03 um whatever you think of chat GPT and
56:05 whatever the state of the the art is now
56:06 the fact that it took off so quickly
56:10 um just forced everyone to like reveal
56:14 their hands and and and start pushing
56:16 hard on this stuff so that's that's why
56:18 this is really exciting right now
56:25 I got addicted to okay let's see togas
56:28 thank you for the cruises the Roses on
56:31 who's buying out of the name
56:37 thanks Kyle we'll figure something out
56:39 oh it so tick tocks just put me in I I
56:43 don't know how Tick Tock does what they
56:44 do with these [ __ ] comments I really
56:46 don't
56:49 um actually those Epiphany moments will
56:51 keep on doubling quicker and quicker as
56:53 time goes on yeah I agree with that Kurt
56:55 where am I I think I'm fairly far down
56:57 in the comments I think I just skipped a
56:59 bunch of questions but that's okay I was
57:00 pretty far behind
57:02 so I'm just gonna go grab one here that
57:05 looks interesting may I inquire if you
57:08 possess a concise reference guide
57:11 I do not the closest thing I have so
57:14 Thelonious who's one of the Irregulars
57:16 here
57:18 he's he's been as I mentioned new tools
57:21 and played with new tools he's been
57:23 adding them to a Google sheet so
57:26 um Pony on up the Thelonius if you see
57:28 them in the comments and then at some
57:30 point we're gonna we're gonna
57:32 um
57:33 we're gonna put make that that uh
57:36 spreadsheet available somewhere
57:38 but that's that's why I have these
57:40 things on screen like so what I would do
57:41 is what I'll do is this
57:43 I'm gonna go through and I'll show you
57:46 all the all these uh
57:48 resources that I think are worth playing
57:51 with especially if you're just getting
57:52 started right now you don't need a ton
57:54 you could spend I seriously
57:57 I feel like I could personally spend the
58:00 next decade playing with chat GPT 3.5
58:03 and nothing else and still not uncover
58:06 all of what it can do these things are
58:08 profoundly powerful they're profoundly
58:11 deep
58:12 um so you don't need a ton and that's I
58:15 try to keep
58:16 what I recommend to a minimum because
58:18 you don't need more rabbit holes to run
58:21 down there's plenty even within these
58:24 things so what I'm going to do is
58:26 um
58:27 so so screenshot this so here's these
58:31 are these are my top three this is the
58:33 adult table
58:35 um chat EBT is the best gpt4 is worth
58:39 the 20 bucks to pay for in my opinion at
58:42 this site chat.openai.com you also get
58:45 this thing called Advanced data
58:46 analytics here which is [ __ ] mind
58:49 blowing
58:50 so that's that one Claude is free and
58:53 it's a competitor to chat GPT from a
58:56 bunch of EX open AI people so a bunch of
58:59 people left here and went to start
59:01 Claude because they had they had a
59:03 better way to make these things safer
59:05 and that's what that company is it's
59:07 it's called anthropic Claude is really
59:10 interesting and Claude has a hundred
59:13 thousand token context model which
59:15 basically means you could paste a 50 000
59:18 word novel into the chat window and it
59:22 will read it
59:23 you can literally paste in your great
59:26 American novel into the chat window and
59:28 tell it to write an epilogue or a
59:30 prologue or rewrite chapter 13 and it
59:33 will
59:33 so it's pretty amazing Bard is Google's
59:36 entry into here I have it here because
59:38 it used to suck so bad it wasn't even on
59:40 the list at all and now it's gotten
59:42 decent although today when it wouldn't
59:44 give me hyperlinks it about lost its
59:46 spot at the adult table
59:48 um Bing was at the adult table and got
59:50 kicked onto the Kids Table this is the
59:52 Kids Table
59:54 um and not because these things are bad
59:56 it's just I feel like these three are
59:59 things I use for my everyday work these
1:00:01 three are things that that I experiment
1:00:04 with more and and play with more just
1:00:06 just kind of trying to figure out what's
1:00:08 good and what's not so Bing I showed
1:00:10 earlier that's the one that is connected
1:00:12 to the internet can make images it's
1:00:14 tied to Microsoft
1:00:16 perplexity.ai in and of itself is an
1:00:19 interesting site you can do research it
1:00:21 does citations really well but
1:00:23 labs.perplexity.ai they're hosting metas
1:00:26 llama 2 models and the code llama models
1:00:29 so if you want to go play with those
1:00:30 open source models and see how good they
1:00:32 are or bad they are you can do that
1:00:34 there before having to take the time to
1:00:36 install them on your own servers Pi is a
1:00:40 kind of Chatty version of chat GPT
1:00:42 chatty GPT
1:00:45 um you can talk to it you can put it
1:00:47 into voice mode and actually just have a
1:00:48 conversation with it fascinating and
1:00:51 then Poe has a whole bunch of bots that
1:00:54 people have built and you can build your
1:00:55 own bot there it's really interesting
1:00:57 okay so take a screenshot of this
1:01:02 um the the thing at the top that's our
1:01:04 YouTube channel so I'm taking the
1:01:08 um archives of these lives and I'm
1:01:10 putting them on here and I'm
1:01:12 chapterizing them using Claude
1:01:15 it's a slow cumbersome process so I'm
1:01:19 very far behind but I'm doing it
1:01:22 um what else okay images
1:01:25 so these are images here's another one
1:01:27 to take a screenshot of okay hang on
1:01:29 hey there you go take a screenshot of
1:01:32 that so these are the image tools that
1:01:34 where I would start mid journey is by
1:01:36 far the best just it is it's the best
1:01:40 stable diffusion Geeks will go Stamford
1:01:42 Fusion is as good as mid Journey it's
1:01:44 not but what stable diffusion has that
1:01:47 mid-journey doesn't is that stable
1:01:49 stability AI has open sourced stable
1:01:52 diffusion in the models and so the
1:01:55 developer Community for stable diffusion
1:01:57 is massive these are just two projects
1:02:01 within the stable diffusion Community
1:02:03 there's one called automatic 1111 which
1:02:05 is if you're geeky you can install and
1:02:08 it's got all sorts of extensions and
1:02:10 wild stuff thanks for the beaver I like
1:02:13 a good Beaver okay
1:02:16 um
1:02:17 Leonardo is just a really good website
1:02:20 sitting on top of stable diffusion clip
1:02:22 drop is a bunch of tools from stability
1:02:24 AI
1:02:25 um ideograms free mid journey is not
1:02:28 Leonardo's not but you can do I don't
1:02:31 know seven or eight images a day for
1:02:32 free there you get credits every day in
1:02:35 the free tier and then clip drop most of
1:02:38 those tools are free I think you might
1:02:40 have to pay for those I can't remember
1:02:41 and then prompt book will teach you how
1:02:44 to talk to these tools all right and
1:02:45 then the other thing you can do about
1:02:47 almost a year ago I started this thing
1:02:50 called the AI salon so if you go to the
1:02:52 salon.ai there's a link tree there where
1:02:55 you can go to our Meetup we meet every
1:02:57 other week and we're our next meeting is
1:03:00 next Tuesday and then we also have a
1:03:02 Discord where you can hang out and geek
1:03:04 out and we've got a if you're if you're
1:03:06 a regular in the chat here we've got a
1:03:08 channel a private channel for the
1:03:10 Irregulars to hang out and you know ask
1:03:12 questions and stuff like that it's a
1:03:14 cool group so all right
1:03:17 that was a rabbit hole I'm exhausted
1:03:20 good Lord good Lord people what are we
1:03:23 gonna do all right
1:03:25 um but that's as close as I have to a
1:03:27 concise reference guide
1:03:29 uh I am gonna do so I'm putting together
1:03:32 a workshop right now thank you very much
1:03:34 Sheriff anyway share bear what's
1:03:35 happening
1:03:37 um I'm putting together a workshop right
1:03:38 now with a woman that I did a uh I spoke
1:03:43 on her podcast
1:03:45 um and we're doing a workshop for
1:03:47 creatives and then I'm gonna once I do
1:03:49 that I'm gonna follow that up I'm gonna
1:03:50 create a an AI learning lab specific
1:03:53 Workshop I just haven't done it yet
1:03:55 so that will be coming to
1:04:03 so much people I can only do so much
1:04:09 all right
1:04:10 I saw someone using the beta for
1:04:12 improved chat GPT today on Tech tick
1:04:15 tock
1:04:18 which cafruno was that the one that that
1:04:21 we showed last night that was the
1:04:26 the uh what was that thing called wait I
1:04:28 bookmarked it hold
1:04:33 typing mind
1:04:38 yeah was it typing mine that you saw
1:04:40 someone using or was it a new version of
1:04:43 chechi BT I'm curious which one can Bing
1:04:47 settle the debate as to who has the
1:04:48 better skiing of Utah or Colorado no
1:04:51 like no no of course not I put sunscreen
1:04:55 in my hair I know so for Fox the writing
1:04:58 of Bing is
1:04:59 here's the deal when when you have
1:05:03 Bing GPT right for you
1:05:06 the result of the writing for me with
1:05:10 Bing is always it's as if someone at
1:05:12 Microsoft wrote it
1:05:16 I mean it's professional-ish
1:05:21 so yeah I'm right there with you
1:05:23 um so no no one can ever debates like is
1:05:26 Utah or Colorado better you never win
1:05:28 those it's like it's like politics right
1:05:31 now in the United States or it's like
1:05:32 football teams you know that's just a
1:05:35 lifelong battle that's that'll just go
1:05:38 on forever did you create that image or
1:05:40 find it I don't know which image you're
1:05:41 talking about
1:05:43 um
1:05:44 these images I created on
1:05:48 um ideogram so here let me show you
1:05:51 ideogram if you haven't seen it yet it's
1:05:53 pretty flipping sick it's thick man it's
1:05:56 so sick it's so cool
1:06:02 oh that's a cool thing I'll show you too
1:06:04 all right let's see id-o so it's i-d-o
1:06:08 wait i-d-e-o Graham dot a i
1:06:13 and it's just like any one of these
1:06:16 other sort of image generation sites
1:06:17 it's very very simple
1:06:19 it's got some buttons for different
1:06:21 styles it's got
1:06:24 um three different aspect ratios tall
1:06:26 square flat or you know landscape
1:06:30 and but look what it can do
1:06:33 most likely to pick peek under the
1:06:35 Christmas tree it can do text it can do
1:06:37 logos it can do stuff that mid Journey
1:06:40 can't
1:06:42 and so how I made stuff like here's a
1:06:45 cool one so so we're just gonna grab
1:06:47 this one we're gonna go copy this
1:06:49 person's prompt because you can
1:06:51 this is a cool way to learn this stuff
1:06:53 as well a good way to learn prompting
1:06:55 and I'm just going to paste it in here
1:06:57 and we'll do vertical because it's kind
1:07:00 of a poster
1:07:01 so we'll say photo poster 3D render
1:07:07 I say cool futuristic pop art realistic
1:07:11 song poster in which the text is
1:07:14 AI learning
1:07:17 lab
1:07:19 is written on a hot air balloon
1:07:22 of Earth like shape in night sky photo
1:07:26 fine whatever click
1:07:28 go experiment go spearmint just go copy
1:07:32 and paste other people's prompts just
1:07:34 learn this stuff any way you can you
1:07:36 don't need to you one of the one of the
1:07:40 misconceptions about these tools right
1:07:42 now is oh I've got to be a scientist or
1:07:44 I got to be a prompt engineer I got to
1:07:45 be this I got to be that no you don't
1:07:47 actually you can just go play these
1:07:50 tools are even though they're janky
1:07:52 pieces of [ __ ] they are at the stage now
1:07:54 you can just go play with this stuff and
1:07:56 you can see oh that's actually useful or
1:07:58 not or you know
1:08:01 you know here's a rabbit hole to go run
1:08:02 down it looks like we got close with
1:08:04 this one and close with this one we
1:08:06 didn't get the AI but we got learning
1:08:08 lab
1:08:09 and these aren't quite done yet
1:08:11 generation progress 99 still doesn't
1:08:15 quite look darn
1:08:17 but this is kind of a cool image yeah
1:08:20 there it goes there it sharpened up so
1:08:22 now what you can do with this is you can
1:08:24 actually go remix it
1:08:25 so one of the things that I've noticed
1:08:28 is if you can get one that's close
1:08:29 that's got some letters in it if I just
1:08:31 do a remix and then just generate it
1:08:34 it will sometimes give me the AI
1:08:37 learning lab thing
1:08:39 because it's still got the prompt in it
1:08:43 and it's using that image as kind of a
1:08:45 seed image I suppose look there's AI
1:08:47 learning lab right there
1:08:48 there's another one we got it yay you
1:08:51 did it yay
1:08:53 you did it yay
1:08:56 you're a winner yay
1:08:58 [Laughter]
1:09:01 so this is ideogram.ai
1:09:06 audiogram.ai
1:09:08 welcome to the AI learning lab
1:09:11 we'll take you to new heights of
1:09:13 understanding huh
1:09:16 how's that for marketing with the little
1:09:18 balloon image
1:09:20 kind of cool and it actually has the the
1:09:23 world on it this one has the world on it
1:09:24 pretty groovy
1:09:27 I think that's the best one
1:09:29 so you pop that bad boy up and actually
1:09:31 you know what I'm gonna do I'm gonna
1:09:34 save images
1:09:37 balloon
1:09:42 I'm gonna save it
1:09:44 and then I'm gonna go here and then I'm
1:09:47 gonna go here
1:09:50 and then I'm gonna go wait where do I
1:09:53 have that no that's that downloads
1:10:04 a i l l name contains
1:10:13 this is it no
1:10:17 where did I
1:10:19 what I say balloon right balloon
1:10:22 balloon
1:10:24 there it is all right so let's go over
1:10:28 here to library
1:10:30 we'll go over here to this I'll drag our
1:10:33 balloon in here
1:10:35 there's our balloon and then we'll drag
1:10:37 that into our little
1:10:39 library and now we got
1:10:41 Bang
1:10:43 look
1:10:44 you did it yay
1:10:47 [Laughter]
1:10:50 you did it you're a winner
1:10:55 so you know listen
1:10:57 I know I'm kind of making fun of this
1:10:59 and this is silly but
1:11:01 how [ __ ] remarkable is that
1:11:05 I mean we get to do this now
1:11:09 this works
1:11:11 you just watched me do it this works
1:11:13 this works now
1:11:15 and you're like but Kyle it's the
1:11:17 kerning isn't quite correct on those
1:11:19 fonts I don't know if you know
1:11:21 typography like I know typography but
1:11:23 then yeah sorry
1:11:26 30 seconds
1:11:29 30 seconds
1:11:31 this is the worst this technology will
1:11:33 ever be
1:11:34 imagine this in three years 18 months
1:11:38 ago stable diffusion and mid-journey
1:11:41 when it generated images they were
1:11:43 laughably bad laughably bad it didn't
1:11:46 look anything like the prompt mid
1:11:49 Journey right now is doing full-on
1:11:51 photorealistic insane
1:11:55 stuff and this stuff is doing this level
1:11:57 of typography
1:12:00 three years this stuff will be Insanity
1:12:04 in fact the I saw an interview with the
1:12:07 uh
1:12:09 with the CEO of stable diffusion and
1:12:13 their stability Ai and he's saying the
1:12:16 the internal tool that they're playing
1:12:18 with it's stability AI
1:12:20 it's generating 60 images a second
1:12:26 why is that relevant Kyle well
1:12:29 that's the speed that games run at and
1:12:32 sports are you know broadcast at
1:12:36 you know movies are 24 frames a second
1:12:38 videos 30 frames a second games and
1:12:41 Sports and things like that are 60 or
1:12:44 100 or 120 or right
1:12:47 so it's generating
1:12:50 like stable diffusion quality images 60
1:12:54 of them in a second well what's that
1:12:56 mean Kyle well
1:12:59 make me a movie right if you have you
1:13:01 seen oh you want to go do something fun
1:13:03 let's go take this image we'll go over
1:13:05 to runwayml
1:13:06 we'll turn you want to turn this into a
1:13:08 video wait can we just marvel at the
1:13:10 hair what the [ __ ] just happened to my
1:13:12 hair
1:13:13 good Lord I think the raccoons came out
1:13:16 of the storm drains again and did
1:13:17 something up there all right we're gonna
1:13:19 go animate this thing
1:13:22 with Runway ml
1:13:25 [Laughter]
1:13:29 so you go to runwayml.com
1:13:34 and of course you have to log in these
1:13:36 image these video generation tools are
1:13:39 expensive so they're not cheap but
1:13:42 they're pretty remarkable
1:13:44 so you go text to video and you can just
1:13:47 type in a prompt and it'll make you a
1:13:48 video but look at this
1:13:51 look at this magic little button here
1:13:52 let me make it bigger
1:13:56 upload an image
1:13:59 Bang
1:14:02 where's my
1:14:08 oh sorry no that's not right you gotta
1:14:10 do it here it'll show it there
1:14:13 all right so I'm gonna go find
1:14:18 there's my balloon image right there's a
1:14:20 little balloon image so we're going to
1:14:22 open that
1:14:23 and so now that becomes
1:14:27 I think it shows up here yeah so that
1:14:29 becomes our or that so this is what
1:14:30 we're going to animate
1:14:32 and then I can change my motion from
1:14:36 you know zero to ten so I'll put it at
1:14:38 like seven I guess I don't know what's
1:14:40 this
1:14:41 we're going to interpolate we're going
1:14:42 to smooth out the frames we're going to
1:14:43 upscale we're going to remove the
1:14:45 watermark okay whatever
1:14:48 and so I have a 197 seconds left and I'm
1:14:51 going to say generate
1:14:52 Bang
1:14:54 wait why can I not do this oh I think I
1:14:56 have to put in a prompt
1:14:57 um
1:14:58 balloon
1:15:03 drifting through
1:15:05 the clouds on uh at
1:15:11 dusk
1:15:13 all right and now I said generate bang
1:15:16 and now it's going to go off and it's
1:15:18 going to do its thing and it gonna make
1:15:20 us a video it gonna make us a pretty
1:15:22 pretty video
1:15:23 and you know this is this is very
1:15:26 similar
1:15:28 um to a lot of the other stuff where
1:15:31 where
1:15:33 um
1:15:36 the results of this are like so so when
1:15:39 I made this image that like
1:15:41 two of the images had no letters in them
1:15:44 one of them had some letters and one of
1:15:47 them had learning lab in it right and so
1:15:48 it was like 20 sort of there 25 sort of
1:15:52 there
1:15:53 um same with these video generating
1:15:55 things like it may or may not do this
1:15:58 well because for one thing I had a
1:16:00 vertical image and it's just sort of
1:16:01 cropping it so it might not even
1:16:03 recognize this as a as a balloon but
1:16:04 we'll see
1:16:06 um
1:16:07 but again you know you look at what
1:16:10 we're doing here you know
1:16:12 I I've put essentially no time or
1:16:14 thought into this right so we went over
1:16:17 to ideogram I I copied and pasted
1:16:19 someone else's prompt we turned it into
1:16:21 this which sort of looks like what they
1:16:23 did but different enough that not really
1:16:26 we made that cool image that's an image
1:16:28 we can use now for the for the thing and
1:16:30 now we're popping into Runway ml this is
1:16:33 going to turn into a video imagine
1:16:36 if you actually put some time and effort
1:16:38 and critical thinking and thought and
1:16:41 creativity and curation and editing into
1:16:45 this stuff
1:16:47 that's how people are making these
1:16:48 remarkable you know pieces of creative
1:16:51 these videos like if you haven't seen
1:16:53 barberheimer
1:16:55 um go see barberheimer oh it kept the
1:16:57 vertical aspect ratio that's pretty cool
1:17:07 so that's actually kind of cool I mean
1:17:10 like like
1:17:12 it didn't animate the clouds which it
1:17:14 wouldn't right if if this were If This
1:17:16 Were a balloon floating through the air
1:17:17 this is kind of what it would do right
1:17:19 balloons aren't the most exciting thing
1:17:21 and the light kind of you know
1:17:23 Shines on it differently so anyway and
1:17:26 then the other thing that you can do is
1:17:28 you can extend these for four seconds
1:17:29 you can do that four times so it allows
1:17:32 you to take a four second video and
1:17:34 extend it up to 18 seconds
1:17:37 so
1:17:38 um yeah just crazy crazy stuff
1:17:41 um let me see if I've got other other
1:17:43 videos that I did here
1:17:49 where
1:17:52 AI magic tools texted video
1:18:00 I don't know these these aren't ones
1:18:02 that I did but look here's here's ones
1:18:04 that other folks have done
1:18:08 you know good remarkable stuff
1:18:12 so you just generate an image
1:18:16 you come in here and it will make it
1:18:18 into video
1:18:22 look at that I mean come on man
1:18:25 you kidding me
1:18:27 let's see this one this looks cool
1:18:32 yeah not bad
1:18:36 that's slick
1:18:40 right
1:18:44 you need stuck you need stock video of a
1:18:47 certain kind of fabric fluttering in the
1:18:49 wind right you need a little
1:18:51 Hobbit Shire there you go done Bang
1:18:55 [Laughter]
1:18:58 trinkets blowing in the breeze I mean if
1:19:01 you did this carefully if you if you
1:19:03 really you know figured out what again
1:19:05 figure out what these things are good at
1:19:07 what they're not good at and then get
1:19:09 creative within the constraints it's
1:19:12 actually a really good creative exercise
1:19:13 is to try to be creative
1:19:16 inside of the imperfection of these
1:19:19 janky ass tools
1:19:21 really good creative exercise
1:19:25 all right 99 Plus messages behind see
1:19:28 what happens when you go down a little
1:19:29 rabbit hole
1:19:31 anyway I hope that helped
1:19:35 um I'm gonna jump all the way to the
1:19:37 bottom and kind of scroll up a little
1:19:39 bit
1:19:44 talk about your October series of
1:19:47 workshops okay
1:19:49 um
1:19:50 so
1:19:51 Rhonda just sent me the a write-up she
1:19:54 did for it I haven't looked at it today
1:19:55 but the evening of October 4th which is
1:20:00 a Wednesday and then every Wednesday for
1:20:02 three Wednesdays from five to seven
1:20:05 Mountain Time
1:20:08 I'm doing a workshop for creatives so so
1:20:11 Rhonda's group she's a professional
1:20:12 voice over actress so it's her group is
1:20:15 mostly voice over actresses her voice
1:20:17 over actors
1:20:19 um but she's also got just creatives in
1:20:21 general so we're gonna do a workshop on
1:20:24 you know practical use of AI for
1:20:27 creatives
1:20:28 um so October 4th 11th
1:20:32 18th I think are the three dates but I
1:20:35 know the fourth is the first one that's
1:20:37 right seven days right fourth eleventh
1:20:39 18th
1:20:41 five to seven Mountain
1:20:42 and I'll uh
1:20:44 just I'll I'll when I have materials
1:20:48 I'll talk about it in here but if you
1:20:49 want to do a save the date thing and
1:20:51 start hanging out in the lives you know
1:20:53 this is what we do here night after
1:20:55 night not after night after night after
1:20:58 night at tonight
1:21:02 all right you need another whiskey I do
1:21:04 need another whiskey you know you know
1:21:06 what the problem was I had the was I had
1:21:08 dinner and then I had a whiskey
1:21:10 and then I was watching the US Open I
1:21:12 got really sleepy I'm like I gotta go do
1:21:14 my live so I'm a little low energy
1:21:15 tonight but you know
1:21:18 another whiskey would make me higher
1:21:20 energy but I have an important call at 8
1:21:22 30 tomorrow morning so I'm I'm done with
1:21:24 whiskey for the night all right is there
1:21:26 a single AI that can make an episode of
1:21:29 a TV show because I know there was that
1:21:31 South Park one yeah that one was called
1:21:34 um showrunner
1:21:41 um
1:21:45 showrunner demos South Park episode
1:21:49 creator
1:21:52 yeah so
1:21:57 so it's so this is basically first first
1:22:01 uh this is an academic paper
1:22:04 um that but but basically what they did
1:22:06 was you you type in a prompt of what you
1:22:10 want the South Park episode to be and
1:22:12 then it creates you a full South Park
1:22:15 episode with music voice over actors
1:22:18 script animation all of it now is it is
1:22:23 it great no but again
1:22:27 these things are janky pieces of [ __ ]
1:22:29 they're the worst they're ever going to
1:22:31 be
1:22:32 so oh sorry we couldn't find the page oh
1:22:35 they took it down oh they might have
1:22:36 taken it down because
1:22:38 South Park kids Trey Trey probably told
1:22:41 him you can't you can't do that man
1:22:45 um
1:22:47 [Music]
1:22:49 of note I reached out to the actual
1:22:51 Parker and stone through their
1:22:54 production company
1:22:58 no comment provided wait
1:23:04 not create a
1:23:11 yeah I don't know I don't know so
1:23:14 um so this was kind of a proof of
1:23:16 concept I haven't really followed it so
1:23:18 I don't know
1:23:19 um if if there's other things behind
1:23:21 this but what what what this what this
1:23:26 makes patently clear
1:23:29 is that the reason that the writer's
1:23:31 Guild is striking at the the reason that
1:23:34 the SAG is striking are really valid
1:23:37 reasons
1:23:38 um as Justine Bateman said and she wrote
1:23:41 a piece for The Hollywood Reporter that
1:23:43 was really quite powerful she's she's an
1:23:46 AI hater but she's very very smart about
1:23:48 it and basically what she said is these
1:23:51 are the last two strikes where either of
1:23:54 these unions have any leverage
1:23:55 whatsoever
1:23:57 because this shit's gonna get that good
1:23:59 and it's it's not going to be long
1:24:03 um so they're gonna they should what
1:24:05 what the argument she basically makes is
1:24:07 you should get as much in those
1:24:08 contracts as you can
1:24:10 because this is this is basically it
1:24:14 um so yeah so I don't I I haven't paid
1:24:16 paid attention to what's going on with
1:24:18 this particular project but but if you
1:24:20 look up
1:24:21 um showrunner AI
1:24:26 Fable is the name of the company
1:24:29 and this was from July 18th
1:24:31 so this was the Venture beat article
1:24:36 I wonder if they have sorry the video
1:24:38 does not exist yeah they took it down
1:24:44 but you can go make your own because so
1:24:46 this is it's like an open source project
1:24:47 where you can go make your own I'm
1:24:49 pretty sure
1:24:52 interesting
1:24:53 but you know that's hey when you're like
1:24:56 like when you're in early days of an
1:24:58 industry like this like this like
1:25:00 everyone's trying to figure out where
1:25:01 the boundaries are right they're like
1:25:03 here's these new tools look we can do
1:25:05 this we can do that we can make Drake in
1:25:07 the weekend sing this popular song Get
1:25:10 600 000 views in a week and then Spotify
1:25:12 takes it down
1:25:14 you know showrunner AI makes it complete
1:25:18 episodes of South Park with a single
1:25:20 prompt
1:25:23 Trey Trey and and uh Stone take it down
1:25:30 all right
1:25:33 um
1:25:34 it was a beta of Chachi PT is I wonder
1:25:37 if there's a video of it where did you
1:25:38 saw it on tick tock
1:25:40 coverino
1:25:42 um YouTube
1:25:46 chat GPT beta
1:25:58 um I don't see anything yeah
1:26:01 uh
1:26:02 uh preview
1:26:09 uh
1:26:10 [Music]
1:26:13 well if you can if you find the uh if
1:26:16 you find that link Bruno to the to the
1:26:18 tick tock let me know just you can DM it
1:26:20 to me or something like that I'd love to
1:26:22 see it that's pretty cool
1:26:26 yeah it's so so it sounds like it sounds
1:26:28 like Google's gonna be releasing Gemini
1:26:31 in the fall which is their new
1:26:33 multimodal
1:26:35 um
1:26:37 chat GPT for competitor
1:26:42 open AI has been suspiciously
1:26:45 quiet
1:26:46 in terms of innovation I think they've
1:26:49 been a little busy with lawsuits and you
1:26:51 know the New York Times threatening to
1:26:52 sue them and things like that Sarah
1:26:54 Silverman suing them
1:26:56 um
1:26:58 but I assume they're cooking something
1:27:00 up I assume they know that something
1:27:03 something's coming from Google
1:27:06 so there's things like chat gpt4 can
1:27:11 generate images and can read images well
1:27:15 well uh Dali can generate images but you
1:27:17 could roll that into chat GPT they
1:27:20 haven't done that yet
1:27:22 um Bing has incorporated the fact that
1:27:25 gpt4 can read images chat GPT hasn't
1:27:28 done that yet like there's a lot of
1:27:29 stuff that should be coming from them
1:27:32 um and it's just not there I don't
1:27:36 you know it's weird
1:27:38 Irregulars is I can't see any totals of
1:27:41 the hearts so I have no idea how many we
1:27:43 have how many likes we have that's
1:27:45 pretty funny Utah doesn't have Casa
1:27:47 Bonita though Case Closed that's awesome
1:27:51 buffering it happens
1:27:53 good signal here
1:27:55 it'll come back maybe it's me
1:27:59 um yeah it was me okay okay yeah just
1:28:01 notice no lagging tonight yay yay that's
1:28:03 great
1:28:05 um okay dude
1:28:11 which application was that
1:28:14 um I do not know what was that feeling
1:28:17 so I was doing
1:28:20 I did ideogram
1:28:27 so ideogram is the image one I did
1:28:30 typing mind
1:28:33 no I don't want to do that
1:28:37 typingmind.com is basically
1:28:41 um
1:28:42 a bunch of additional features on top of
1:28:45 gpt4
1:28:46 like you can do multiple profiles and it
1:28:51 does plug-ins you can pick different
1:28:53 models it's got a prompt Library it's
1:28:55 it's got all sorts of [ __ ]
1:28:57 so that I was looking at
1:29:00 and then I was looking at ideogram and
1:29:04 then I was looking at Runway ml so
1:29:06 runwayml did the video
1:29:08 runwayml.com
1:29:11 ideogram did the images with letters on
1:29:14 them like our happy little AI learning
1:29:17 lab balloon
1:29:19 and then that type
1:29:21 one I can't remember what else
1:29:24 called these typing mind
1:29:27 is the sort of fancy
1:29:30 interface in front of chat EBT I don't
1:29:35 I'm not a particular fan of fan of apps
1:29:39 like this that are basically charging
1:29:42 you money
1:29:44 to to basically add features that openai
1:29:47 is likely going to add to chat EPT at
1:29:49 some point in the future
1:29:51 um
1:29:53 it's possible it's worth it but I just
1:29:55 I'm not a big fan of that approach
1:29:59 so yeah
1:30:02 uh what was the name of that program
1:30:04 again ideogram where else can people see
1:30:07 here like Kyle's really
1:30:09 that's right nowhere this
1:30:14 is what it is let's see alcaraz up two
1:30:19 sets to one
1:30:20 it's very zverev
1:30:23 up four games to three without caress
1:30:26 serving I think he's gonna take him in
1:30:28 straight sets
1:30:30 and it could be it's gonna be
1:30:32 interesting to see if young Ben Shelton
1:30:34 beats uh Djokovic I mean I'd like to see
1:30:37 Djokovic win just because you know he's
1:30:39 he's like having this amazing run as as
1:30:42 an older dude
1:30:44 um but these these young kids like the a
1:30:47 20 year old versus 20 year old US Open
1:30:50 final with uh with alcaraz and Ben
1:30:53 Shelton that would be electric so
1:30:56 all right enough geeky Sports Talk I
1:30:58 thought we were here to talk about AI
1:31:00 Kyle you know I don't know if you knew
1:31:02 this but uh I'm here to learn about
1:31:03 artificial intelligence okay uh you know
1:31:06 I'm trying to figure things out here uh
1:31:08 there's a lot of disruptions going on in
1:31:10 my life and if all you're going to do is
1:31:12 talk about uh Sports uh memorabilia or
1:31:15 whatever you're whatever you're doing
1:31:16 with your your uh your tennis Hobby
1:31:19 it's not really appropriate for an
1:31:21 artificial intelligence Channel if you
1:31:23 know what I mean
1:31:25 you understand what I'm saying
1:31:27 no
1:31:28 no all right well carry on then okay
1:31:34 very rude he was very rude to me
1:31:36 [Music]
1:31:48 [Music]
1:31:55 a lot of joking going on between the
1:31:57 Irregulars I like this I fully support
1:32:00 when I get boring the Irregulars kind of
1:32:03 just take over and have their own fun
1:32:05 which is which is how it should be it's
1:32:07 like Kyle if you're not entertaining us
1:32:10 we're just gonna entertain ourselves I
1:32:11 hope you know that
1:32:13 you know this is not just your channel
1:32:15 anymore you know there's a lot of us
1:32:17 here
1:32:18 we show up very regularly we expect
1:32:20 attention and if you're not giving it to
1:32:22 us then we're going to entertain
1:32:23 ourselves I hope you know that
1:32:26 the Irregulars are all from Minnesota I
1:32:28 don't know what's going on with that but
1:32:30 they all talk like Al Franken
1:32:33 all right morning in Oz time for the
1:32:36 voice I don't know what divorce is
1:32:40 October 4th 11th 18th plus 25 six to
1:32:45 seven Mountain creative lives Silver Fox
1:32:48 it's five to seven they're two-hour
1:32:50 workshops
1:32:52 five to seven Mountain Time
1:32:54 if I said six to seven that's because oh
1:32:57 MST is Colorado time yes Mountain
1:33:00 Standard Time
1:33:03 so what we decided is five to seven
1:33:07 in Denver will be four to six
1:33:10 on the west coast and seven to nine in
1:33:14 New York right so it's it's not bad so
1:33:17 Wednesday evenings the fourth the 11th
1:33:20 the 18th
1:33:21 at those times
1:33:23 for the for the workshop have those
1:33:26 strikes stopped production of shows Etc
1:33:28 how long could that last it can last for
1:33:30 a while they've got there's a weird
1:33:33 there's a weird
1:33:36 um
1:33:38 kind of production back door
1:33:40 for the strike where if the Productions
1:33:44 are independent which basically just
1:33:45 means it's not one of the big studios
1:33:48 that are producing them they can still
1:33:51 be filming so I think writers are still
1:33:53 writing I think actors are still acting
1:33:55 in this sort of weird loophole the big
1:33:57 studios are kind of screwed they they
1:34:00 can't like all the big Blockbuster
1:34:01 shit's not not going but I think all of
1:34:03 the more what they call Independent
1:34:06 production which a lot of production
1:34:08 these days is independent so like
1:34:09 there's a lot of [ __ ] still being
1:34:11 produced
1:34:12 um even though it's in the context of
1:34:14 the strike so I don't I haven't been
1:34:16 following it super closely but I know
1:34:18 that's a that's kind of a weird a weird
1:34:20 thing about it
1:34:21 uh they can go on for a while like the
1:34:24 the yeah it's just
1:34:27 striking in that business is rough
1:34:29 because like I have a degree in acting
1:34:32 I've moved to New York City out of
1:34:33 school to do this I've written seven
1:34:34 screenplays like I know that world well
1:34:39 if you have a crappy three-line
1:34:44 part in a movie
1:34:46 there are 10 000 people that would drag
1:34:49 you behind their car to take you out to
1:34:52 get your role
1:34:54 so it's a it's a rough business to
1:34:56 strike because like the the amount of
1:34:59 people that want to be in that business
1:35:02 dramatically outweigh the amount of jobs
1:35:05 available so so they don't have a lot of
1:35:07 Leverage so basically what it comes down
1:35:09 to is do the big stars can can the big
1:35:12 stars wield their Leverage
1:35:15 for the rest of
1:35:17 the the the Union
1:35:20 because I think the average salary of a
1:35:21 sag actor is 26 Grand a year
1:35:24 like Actors Equity
1:35:26 you know average unemployment is like 96
1:35:30 of union members right so if you're an
1:35:32 equity actor that's that's stage acting
1:35:35 but it's it's like some ridiculous
1:35:38 amount of unemployment
1:35:40 um so so none of the lower level like
1:35:43 even beat beat level stars and and down
1:35:47 they've essentially no leverage because
1:35:50 there's just thousands of people that
1:35:52 want to do that job so this is really
1:35:54 about the the
1:35:55 big stars you know
1:35:57 costing the studios hundreds of millions
1:36:00 of dollars and if they do then the
1:36:02 studios are going to bend and you know
1:36:04 quite frankly like all the streaming
1:36:06 services are not paying any royalties so
1:36:09 like I've got a friend a friend of the
1:36:11 family was on The Sopranos and you know
1:36:15 all that's all the shows on Netflix and
1:36:17 all that they get paid their their fee
1:36:19 for the show but they have no royalty so
1:36:21 they're trying to get royalties and then
1:36:23 they're trying to get with this AI [ __ ]
1:36:26 um if you synthesize my likeness and use
1:36:28 it again uh you should credit me and you
1:36:31 should [ __ ] pay me
1:36:33 and the producers are like well how
1:36:35 about we pay you 200 for a day rate and
1:36:38 then we can use your likeness in
1:36:39 perpetuity
1:36:43 and sag was like okay [ __ ] you we're out
1:36:46 so it could go on for a while
1:36:49 Tick Tock live with another Creator oh
1:36:52 interesting okay cool
1:36:55 um
1:36:57 27 000 very nice who should invest to
1:37:01 make some Investments oh who should we
1:37:03 invest in to make some Investments I
1:37:06 don't know listen I have a degree in
1:37:08 acting
1:37:09 do you want me to give you Financial
1:37:11 advice fine I'll do it
1:37:14 should you listen to it no
1:37:18 um
1:37:19 you know I'll tell you I'll tell you the
1:37:20 like so
1:37:22 in terms of who to invest in I don't
1:37:24 [ __ ] know and like a lot of these
1:37:26 companies are private right now I'll
1:37:27 tell you the companies that I think are
1:37:28 interesting
1:37:30 um I think Nvidia is interesting because
1:37:33 they somehow miraculously figured out
1:37:36 how to create a monopoly of GPU chips
1:37:42 that no one else seems to be able to
1:37:44 catch up with like like the entire chip
1:37:46 manufacturing industry seated
1:37:50 cede
1:37:52 gave
1:37:55 Nvidia the the GPU market right because
1:38:00 they were doing video games for a lot of
1:38:01 years
1:38:03 and then they were doing crypto [ __ ] and
1:38:05 blockchain [ __ ] and then the AI stuff
1:38:08 have you seen their earnings report that
1:38:11 their earnings report it was going like
1:38:13 this and like the last quarter it's like
1:38:15 it like doubled or tripled like it's
1:38:17 just some [ __ ] stupid amount of of
1:38:19 Revenue that they've got right now
1:38:22 um so they hit a trillion dollar market
1:38:23 cap and their revenues through the
1:38:25 [ __ ] roof open AI is clearly
1:38:27 interesting they just announced they had
1:38:29 a billion dollar run rate there there
1:38:31 was a there was an article out about
1:38:32 three weeks ago that was like open AI
1:38:35 spending 700 000 today and they're going
1:38:37 to be bankrupt by the end of 2024 and my
1:38:40 response to that was yeah no they're not
1:38:43 Microsoft invested 10 billion dollars in
1:38:45 them and
1:38:47 they're wrapping open ai's technology
1:38:50 into their entire product line from
1:38:53 Azure cloud services to uh to GitHub to
1:38:57 Office 365 to Dynamics 365 all of it to
1:39:01 Windows 11 all of it all of it's going
1:39:04 to have GPT in it so they're not going
1:39:06 to let them die so and and then you know
1:39:08 what last week open AI just hit a
1:39:12 billion dollar run rate 700 000 a day in
1:39:15 cost is about 250 million dollars 255
1:39:17 million or something like that
1:39:19 and and they just did a billion dollar
1:39:21 run rate so they're gonna be okay I
1:39:23 think I think open a is gonna be okay so
1:39:25 they're interesting
1:39:26 I think anthropic's interesting the
1:39:28 people behind Claude
1:39:30 um they've got a thing called the
1:39:31 Constitutional learning model which
1:39:33 feels to me like the future of how these
1:39:35 things how we're going to get these
1:39:36 things safe
1:39:38 um it might not be that but it's it's
1:39:40 the most interesting of the things that
1:39:42 are out there
1:39:43 um inflection the company behind pie the
1:39:47 the the the chatty GPT that we make fun
1:39:49 of they just raised 1.3 billion dollars
1:39:52 they're still private but that's they're
1:39:54 really [ __ ] interesting x dot a i
1:39:56 Elon musk's company I wouldn't bet
1:39:58 against him
1:40:00 um even though he's kind of making a you
1:40:02 know a mess of of X of Twitter of X
1:40:05 Twitter
1:40:06 um
1:40:07 you know they just put in a 10 000 GP pu
1:40:10 server farm and they've completely
1:40:12 shifted how
1:40:14 um Teslas are doing full full
1:40:16 self-driving
1:40:17 so the the
1:40:19 he he's gonna start melding all of his
1:40:22 company's Technologies together so
1:40:25 they're going to get efficiencies of
1:40:26 scale so I think interesting shit's
1:40:28 gonna come out of there
1:40:30 um apple is gonna they they're spending
1:40:33 the bulk of their 24 billion dollar
1:40:35 research Budget on generative AI They're
1:40:37 Gonna Come Out Swinging
1:40:39 um
1:40:42 mid-journey I think is going to get
1:40:44 their ass handed to them I think so
1:40:46 right now they're an 11 person company
1:40:48 they're really small the CEO is a bit of
1:40:51 a Cowboy and he's just kind of like I
1:40:54 don't want to do in painting so we're
1:40:55 not going to do in painting I don't like
1:40:57 apis I'm we're not like he's just he's
1:40:59 just a bit of a cowboy and
1:41:02 alcare has just won
1:41:04 so I don't know who alcaraz is playing
1:41:06 but you just want
1:41:08 um
1:41:11 so I think I think mid-journey is going
1:41:13 to get get slapped around a bit now
1:41:16 someone could buy them and because it's
1:41:18 really interesting technology it's
1:41:19 really good technology
1:41:21 um ideogram the fact that they came out
1:41:24 of the out of nowhere their 10 day old
1:41:26 company they raised I don't know 60
1:41:28 billion dollars with Andresen Horowitz
1:41:31 in the lead
1:41:32 um and they're doing text like this out
1:41:35 of nowhere all of a sudden they look
1:41:37 really interesting
1:41:39 um stability AI
1:41:41 um
1:41:43 super interesting company super
1:41:45 interesting CEO
1:41:47 uh
1:41:49 uh run my ml for video
1:41:51 descript for video
1:41:54 like again a lot of these are private
1:41:56 companies so I don't know that you can
1:41:57 really invest in them but but um
1:41:59 and and like I don't know anything I
1:42:01 like I'm I'm not looking at them through
1:42:03 a financial lens I'm looking at them
1:42:05 through a
1:42:07 what is the least janky
1:42:11 what is what is the least janky piece of
1:42:13 [ __ ] technology out there that's what
1:42:15 I'm I'm looking at like what's actually
1:42:17 interesting that's happening
1:42:19 okay
1:42:21 you know Google's been on their heels
1:42:23 but they're Google they have a lot of
1:42:25 money and if if Google is willing to
1:42:28 risk their current business model
1:42:32 to really go for this then they could be
1:42:35 really interesting meta is interesting
1:42:37 because they gave away
1:42:39 their llama models
1:42:42 um
1:42:43 I can only imagine there's some sexier
1:42:47 model that they're not going to give
1:42:48 away that they're going to charge for
1:42:49 like I think they're giving it a way to
1:42:51 stay relevant while they build up a
1:42:53 kick-ass model so there's there's some
1:42:55 [ __ ] coming that's going to be really
1:42:57 good I I don't think that tells you
1:42:59 anything about what to invest in but
1:43:04 that's the stuff I think is interesting
1:43:06 all right finally made it to dinner
1:43:09 Denver bugs are horrible my windshield
1:43:11 is covered oh if yeah if you're driving
1:43:13 up in the mountains or near a river yeah
1:43:15 you're gonna
1:43:17 you're gonna need to get extra washer
1:43:19 fluid make sure you get the bug
1:43:21 dissolving washer fluid
1:43:23 oh the other thing okay farmer if you're
1:43:26 listening to this if you're listening
1:43:27 while driving that's kind of cool
1:43:29 um if you're listening to this
1:43:31 um
1:43:31 when you're in Denver
1:43:33 well when you're in Colorado
1:43:36 this is a 32 ounce tumbler drink water
1:43:39 like it's your [ __ ] job if you wake
1:43:42 up and have a headache it's because
1:43:44 you're dehydrated you you cannot drink
1:43:47 enough water you're like but I'll pee
1:43:49 all the time no it's just like it it
1:43:51 it's a high desert right it's it's
1:43:53 incredibly dry in the altitude
1:43:56 um all the headaches and [ __ ] I it took
1:43:58 me two years to crack this I would wake
1:44:00 up and just have these ripping headaches
1:44:02 and I don't have headaches water if I
1:44:04 didn't drink enough water if I skip it
1:44:06 at night I'm in I'm in trouble
1:44:14 all right Kyle a-i-l-l balloon boy I'm
1:44:17 the balloon boy
1:44:19 I'm rooting for Djokovic I know I like
1:44:21 Djokovic too but I gotta say this
1:44:23 alcaraz is just a machine and actually
1:44:25 did you see there was an interview with
1:44:27 um
1:44:29 with Djokovic after the was it after the
1:44:32 the French it was after the French I
1:44:34 think
1:44:36 and he was just like the dude's
1:44:38 Relentless like
1:44:40 like [ __ ] calm down alcaraz Djokovic
1:44:44 was saying that he was like he was like
1:44:46 he just keeps coming he just keeps
1:44:48 coming and this new kid Ben Shelton is
1:44:50 just like that right so two 20 year olds
1:44:52 that are like they're both like you know
1:44:54 six foot plus beasts that you know have
1:44:58 crazy serves and crazy game
1:45:01 and then uh Djokovic man
1:45:03 I don't know
1:45:05 he's like the fact that he is just on
1:45:07 this run is just amazing what hope is
1:45:10 there for writers actors Etc if they
1:45:12 don't get a ban on AI in their fields so
1:45:15 a shoe it's a great point they're not
1:45:18 going to get a ban
1:45:20 they're not going to get a band here the
1:45:22 the problem so here's the deal
1:45:26 these tools not that tool not the
1:45:29 balloon but these
1:45:32 these tools these AI tools all of them
1:45:35 are not going to be uninvented and they
1:45:38 are going to be incorporated in
1:45:39 everything we do
1:45:41 they're going to be incorporated in all
1:45:43 the writing programs in all the
1:45:44 spreadsheet programs and all the
1:45:46 note-taking programs in all the film
1:45:48 editing programs and all the photo
1:45:49 editing programs it's it's going to be
1:45:51 ubiquitous it's going to be everywhere
1:45:55 so you can't escape it
1:45:57 and
1:45:59 it it would be the equivalent of when
1:46:03 films were shot on film why they're
1:46:06 called films
1:46:08 when digital cameras came out
1:46:12 the cinematographers likely said that's
1:46:15 not real film right that's not real
1:46:17 Cinema
1:46:19 so so you know cinematographers only use
1:46:22 film cameras we're never going to use
1:46:23 digital cameras well now we do right
1:46:27 every single time a new technology comes
1:46:30 out the existing creative class says
1:46:33 [ __ ] that that's not real art
1:46:37 when the daguerreotype came out in the
1:46:39 mid-1800s painters were offended and
1:46:43 said painting is over
1:46:45 that's not real art but everyone's just
1:46:47 going to click the button and take the
1:46:48 picture and that's not real art and it
1:46:51 took photography took a long time to be
1:46:53 assimilated into the art world and
1:46:54 considered an actual art form
1:46:58 every single time a new technology comes
1:47:00 along it's the same argument
1:47:04 it there is no stopping it you you will
1:47:07 not be able to block it because here's
1:47:09 what's going to happen
1:47:11 let's say
1:47:13 that sag and the wga get what they want
1:47:17 and they get the big studios you know
1:47:20 Sony and Universal and
1:47:22 you know
1:47:24 whoever the [ __ ] the conglomerates are
1:47:26 now they're all owned by BlackRock
1:47:27 anyway I assume
1:47:30 and they all agree we're never going to
1:47:32 use AI to make a film I don't think
1:47:34 that's going to happen but let's say
1:47:35 that happens
1:47:37 what's gonna happen is
1:47:41 a bunch of
1:47:43 idiots and Irregulars from this channel
1:47:46 are going to be like oh we were playing
1:47:48 with this with Runway ML and then like
1:47:50 five years later they're going to be the
1:47:52 best at whatever Runway ml becomes and
1:47:55 some filmmaker makers are going to
1:47:57 emerge
1:47:58 that don't know how to make films the
1:48:00 the old way like with Final Cut Pro and
1:48:03 with cameras and [ __ ] like that there's
1:48:05 going to be filmmakers that emerge that
1:48:07 just make things right
1:48:09 and they're going to get so good that
1:48:11 some of those films are going to get
1:48:12 popular and then they're going to get
1:48:15 together with their friends and they're
1:48:17 going to put together what's the
1:48:18 equivalent of a studio so five years
1:48:20 from now there's going to be some like
1:48:22 like Netflix started out sending DVDs in
1:48:26 the mail and they're now a major
1:48:28 production entity right how did that
1:48:30 happen something similar is going to
1:48:32 happen with this AI [ __ ] that even if
1:48:34 the big studios say we'll never use AI
1:48:36 there's going to be a whole creative
1:48:38 class that emerges that only uses Ai and
1:48:41 they're going to create competitive
1:48:43 entities
1:48:45 that will just obliterate
1:48:48 because they're going to be able to make
1:48:49 content faster better more personalized
1:48:53 now is that depressing [ __ ] yes it's
1:48:55 depressing
1:48:57 so what do you do well
1:49:00 that's why this channel exists you go
1:49:03 start it rather than saying AI is evil
1:49:06 I'm never going to use AI which is what
1:49:09 a lot of the creative class is doing
1:49:10 right now the ones that I know that are
1:49:13 not doing that they go in here
1:49:16 they start playing with it they go oh
1:49:20 [ __ ]
1:49:21 what do I do now
1:49:23 and then they keep playing and they keep
1:49:25 figuring it out and and what happens is
1:49:27 this
1:49:28 really good actors will figure out ways
1:49:31 to leverage this stuff that are better
1:49:34 than people that were non-actors same
1:49:36 with writers writers that amplify their
1:49:40 talents with these tools are going to be
1:49:42 better than non-writers who write
1:49:44 screenplays with these tools
1:49:47 creative professionals or creative
1:49:49 professionals right and and how you
1:49:51 how you use the tools creatively is what
1:49:55 makes you a creative
1:49:56 now what's filmmaking gonna look like in
1:49:59 five years I don't [ __ ] know but it
1:50:01 isn't gonna look like what it looks like
1:50:03 now there will be some Productions yeah
1:50:05 howdy partner life's tough it's tougher
1:50:08 if you're stupid all right
1:50:10 there will be some production companies
1:50:12 that keep doing things the way they are
1:50:13 and there's going to be a hunger for
1:50:15 that kind of filmmaking but there's
1:50:17 going to be this whole other kind of
1:50:18 filmmaking
1:50:20 if I'm a creative professional why would
1:50:23 I not want to use the tools that are the
1:50:26 future of the industry why would I not
1:50:28 start learning about them now because if
1:50:31 I become the best of the new kind of
1:50:34 filmmaking I'm gonna have a career for
1:50:36 myself
1:50:37 because we still got to make [ __ ]
1:50:40 right you still have to have taste
1:50:43 you still have to have refinement and
1:50:45 curation and critical thinking and
1:50:47 creative thinking these are just [ __ ]
1:50:49 tools
1:50:52 so I don't know like there's not a good
1:50:54 answer to it and and here's the deal
1:50:56 it's not just relegated to actors and
1:50:59 writers it's special effects people it's
1:51:03 knowledge workers of all kinds it's
1:51:05 Consultants it's data analysts like
1:51:08 every [ __ ] industry that's computer
1:51:11 mediated
1:51:12 is going to be disrupted by this [ __ ]
1:51:14 every single one
1:51:17 I like I can't there is no possible
1:51:20 future I see
1:51:22 where this stuff gets uninvented
1:51:25 and it doesn't
1:51:27 you know get absorbed
1:51:30 into the system and completely transform
1:51:32 what work is and how people work and
1:51:35 what they do and what's valued
1:51:37 so everyone is going to have to reinvent
1:51:40 themselves it's we're we're in for we're
1:51:42 in for some [ __ ]
1:51:44 and I think five years out it's gonna be
1:51:47 okay because there's going to be
1:51:49 entirely new categories of jobs we can't
1:51:51 even imagine today that that will be
1:51:54 created because of this technology but
1:51:56 between now and that five-year time that
1:52:01 gap of two to three years
1:52:04 kind of starting I don't know let's go
1:52:07 this time next summer for the next three
1:52:10 years after that it's gonna be a [ __ ]
1:52:12 [ __ ] show
1:52:14 and and you know I don't know how you
1:52:17 get on I I don't I don't think you can
1:52:19 avoid it but it ain't just it ain't just
1:52:21 actors and writers it's everyone
1:52:25 so sleep tight
1:52:30 what'd you do tonight I got really
1:52:32 depressed
1:52:34 this guy talks about AI it's all [ __ ]
1:52:38 it's like the robots are not gonna kill
1:52:39 us the robots are gonna
1:52:42 we're gonna use the robots to find ever
1:52:45 increasing ever ever efficient ways
1:52:50 it's like every every person that
1:52:52 leverages these tools is going to be
1:52:54 able to do more better work
1:52:58 more and better right there's this adage
1:53:02 in the in the ad world where
1:53:04 um you know you can have something fast
1:53:07 good or cheap choose any two
1:53:10 right you can have it fast and good but
1:53:13 it's not cheap you can have it fast and
1:53:14 cheap but it's not good right
1:53:16 with these tools
1:53:18 you can have all three and then you can
1:53:21 even probably increase the quality of
1:53:23 where you thought you could or decrease
1:53:25 the cost right
1:53:28 like what are the implications of that I
1:53:30 don't know
1:53:32 like the the the piece that I've got not
1:53:35 a [ __ ] clue about is what does it
1:53:37 look like
1:53:39 all I know is it's going to be different
1:53:41 and it will be the people who [ __ ]
1:53:44 hang out here and geek out or not just
1:53:46 here but just sort of hang out in this
1:53:48 mindset of okay [ __ ] it this is an
1:53:52 adventure
1:53:53 and I'm gonna [ __ ] go on the
1:53:55 adventure there's there's some large
1:53:57 fire breathing animal that seems to be
1:53:59 killing the villagers and oh it scares
1:54:02 the [ __ ] out of me but you know what I'm
1:54:03 gonna I'm gonna go figure out what this
1:54:05 thing is Right called to Adventure right
1:54:08 what is the dragon I don't know
1:54:12 is can we slay the dragon I don't know
1:54:15 but let's get together like five of us
1:54:18 with our little torches and you know
1:54:20 you're good with a bow I'm good with a
1:54:22 sword you know
1:54:24 Jesse over there is great with magic far
1:54:26 out let's go
1:54:28 and you get into the woods and the first
1:54:31 thing you come up against is a is a
1:54:33 torrential River and you have to cross
1:54:35 the river and you can't
1:54:37 how do you cross the river I don't
1:54:39 [ __ ] know I'm scared this is
1:54:41 miserable
1:54:42 and you figure out a way to cross the
1:54:43 river that's what we're in the middle of
1:54:45 and so the people that go on that
1:54:47 Adventure
1:54:49 are going to be rewarded now will some
1:54:52 of them maybe get killed by the big bad
1:54:55 you know
1:54:57 Heath and elf possibly
1:55:02 but I think it's the only way through
1:55:05 I think it's the only way through
1:55:10 it's just [ __ ] nuts and again like
1:55:13 like where I sit I'm
1:55:18 it is
1:55:19 it is incredibly rare in human history
1:55:23 to be alive at the beginning of
1:55:27 a transformative technological era it's
1:55:30 incredibly rare
1:55:32 and we've had multiple of them in the in
1:55:35 the 20th and 21st century I've had three
1:55:37 in my life personal computers World Wide
1:55:40 Web
1:55:40 generative AI
1:55:42 those are the three I think are
1:55:46 big
1:55:48 and I think this one's the biggest of
1:55:49 all of them by a lot
1:55:52 I'm in Minnesota and know the frankins
1:55:55 gosh darn it People Like Us that's
1:55:57 awesome I can taper type but not both
1:56:00 not simultaneously awesome I'm 37
1:56:04 messages behind so it's not too bad yep
1:56:06 not even Aaron Paul from Breaking Bad
1:56:08 yeah like Aaron Paul like no one's
1:56:10 getting
1:56:11 um
1:56:12 royalties like yeah I saw him
1:56:14 interviewed Shane uh Aaron Paul got
1:56:17 interviewed you probably we probably saw
1:56:18 the same interview where he's like he's
1:56:20 not getting any royalties from Breaking
1:56:22 Bad are you [ __ ] kidding me do you
1:56:24 know how much money Breaking Bad made
1:56:25 Netflix
1:56:27 and he's not getting any royalty so so
1:56:29 they've got plenty that they can get in
1:56:31 this in this contract
1:56:33 and they could probably get something
1:56:35 around AI right they can get something
1:56:37 around
1:56:39 you know
1:56:40 creating a mechanism for
1:56:43 um licensing their own likeness right
1:56:46 that's where I think the the if if you
1:56:48 paid attention
1:56:50 when the the Drake and weekend song came
1:56:54 out and it it got popular and people
1:56:56 talked about oh this is AI generated and
1:56:58 blah blah blah and I got all these views
1:56:59 and then Spotify took it down about a
1:57:02 week after that Grimes Elon musk's ex
1:57:05 the musician
1:57:08 put out a press release and said
1:57:10 do anything you want with my likeness
1:57:12 make all the music you want using my
1:57:15 voice whatever just give me 50 percent
1:57:17 I think it's a brilliant move because
1:57:19 what she recognized is that they're
1:57:21 gonna do it anyway right and
1:57:24 and if you fight that who are you
1:57:27 fighting are you fighting big corporate
1:57:29 machines well that's going to be one
1:57:32 group you're fighting you know how else
1:57:34 you're gonna be fighting your fan base
1:57:36 if you're a musician that's popular and
1:57:39 you've got fans that have embraced this
1:57:41 [ __ ]
1:57:43 and they start making fan generated
1:57:45 music that turns out to be a hit
1:57:48 what are you gonna do sue your fans some
1:57:50 are
1:57:51 but Grimes is like [ __ ] it
1:57:53 go celebrate my likeness do whatever the
1:57:57 [ __ ] you want with it give me 50 and you
1:57:59 know what people will
1:58:01 they're gonna generate good stuff and
1:58:03 like that to me feels like the future
1:58:05 like imagine creating a talent agency
1:58:07 centered around that idea
1:58:09 let's get artists everything we can so
1:58:12 that when all this stuff let's make it
1:58:14 really easy for fans to generate that
1:58:17 kind of content and get the artists paid
1:58:19 for it
1:58:21 and get the fan paid for it why not
1:58:24 because it's gonna happen anyway like
1:58:26 again
1:58:28 a lot of times people come in here and
1:58:30 they'll sort of pass judgment on it well
1:58:32 what do you think about the ethics of
1:58:33 this the ethics of that I'm like I I I
1:58:36 I'm in no position to sort of pass
1:58:39 judgment in any way all I'm looking at
1:58:41 is what's going to happen
1:58:44 and like are you slowing this stuff down
1:58:46 no are you stopping it no
1:58:48 if the tools are super easy
1:58:52 to be able to make a song that sounds
1:58:53 just like James Taylor when he was in
1:58:56 his 20s and I can write an infinite
1:58:59 amount of James Taylor songs that just
1:59:01 for me personally that I [ __ ] love
1:59:03 and I put them on some sharing service
1:59:05 and one of those turns into a new hit
1:59:08 why would I not do that
1:59:11 everyone's going to do that we're
1:59:12 entering a world of hyper
1:59:14 personalization and infinite content
1:59:17 generation
1:59:20 I want a song that that has the the
1:59:24 depth of Rage Against the Machine
1:59:27 and the lyrical sweetness of James
1:59:29 Taylor and I want it now and I want them
1:59:31 talking about my chubby dog champ
1:59:34 play
1:59:35 and it'll just start being James Taylor
1:59:37 singing about my dog champ with really
1:59:40 deep Rage Against the Machine lyrics
1:59:44 I love this song and I listen to it
1:59:46 every day for two weeks and I'm like you
1:59:47 know what I'm gonna put it on that site
1:59:49 where you can share
1:59:51 shitty cloned music
1:59:53 and you put it on the site and then all
1:59:55 of a sudden it becomes a hit
2:00:00 that's coming
2:00:02 what do you do
2:00:04 you fight it do you tell Sony Music they
2:00:07 can't ever synthesize a voice do you
2:00:10 think Sony music is going to give a [ __ ]
2:00:11 they're not
2:00:12 of course they're going to synthesize it
2:00:15 you know what else Sony music's gonna do
2:00:17 they're going to sign artists that are
2:00:19 AI
2:00:21 artists
2:00:25 and they'll find some [ __ ] loophole
2:00:27 where they'll put it off in a different
2:00:29 subsidiary
2:00:32 so anyway I know it's [ __ ] depressing
2:00:34 but it's also think about it like the
2:00:36 [ __ ] opportunities here insane
2:00:38 they're absolutely [ __ ] insane
2:00:41 in a gold rush invest invest in the
2:00:43 picks and shovels indeed Kyle that's a
2:00:45 really well thought response to
2:00:47 investing in AI for an actor exactly
2:00:50 what are your qualifications sir I would
2:00:53 really like to know your qualifications
2:00:54 what are your qualifications
2:00:56 [Laughter]
2:01:03 oh alcaraz Place Medvedev cool
2:01:07 so so Ben Shelton against
2:01:10 so alcarez is beating Medvedev there's
2:01:13 just not a question there
2:01:16 and yeah I just I think I think Djokovic
2:01:19 I think Djokovic is gonna
2:01:21 maybe Ben Shelton takes him to five sets
2:01:24 but I just Djokovic is is he's just too
2:01:27 mentally
2:01:29 he's a [ __ ] machine
2:01:33 he's just a machine
2:01:35 and it's at that like at this at this
2:01:38 level of the sport like when you're in
2:01:39 the when when you're in the semis this
2:01:42 is
2:01:42 this this is almost got nothing to do
2:01:45 with physical ability right this is
2:01:47 almost all mental
2:01:49 because why are all the same people
2:01:51 always in the finals because
2:01:54 it ain't about the racket it's about the
2:01:57 brain
2:01:59 it's just like these
2:02:02 it ain't about the racket it's about
2:02:04 what we decide to do with it
2:02:09 you know I haven't done this tonight
2:02:15 it ain't about making money with Chachi
2:02:18 PT
2:02:19 it's about you figuring out some problem
2:02:22 to solve in the world where you use chat
2:02:25 GPT thank you Winston oh level 11 for
2:02:28 Winston boom
2:02:32 you figure out something you want to
2:02:34 make better in the world using chat GPT
2:02:37 and get people to pay you money to do
2:02:39 that that's how you make money with chai
2:02:40 gbt you don't leave with technology
2:02:43 it ain't about the racket it's about the
2:02:45 game
2:02:46 what's the game you're playing
2:02:48 where's your head are you thinking more
2:02:51 cleverly than the next guy then you'll
2:02:53 do fine
2:02:55 if you're not you're not
2:02:57 all right infrastructure storage compute
2:02:59 data centers will be the winner winners
2:03:01 yeah that's you know that's why I'm just
2:03:04 I'm I'm flabbergasted that Nvidia has no
2:03:08 competition amd's the only one really
2:03:10 trying to break in but like none of
2:03:13 their software doesn't work according to
2:03:15 George hotz I was watching an interview
2:03:18 with him he's trying to incorporate AMD
2:03:21 GPU chips and just the [ __ ] software
2:03:24 drivers don't work it's like how did
2:03:26 someone let Nvidia get this far ahead
2:03:28 like how is it even [ __ ] possible it
2:03:30 is
2:03:33 oh they can have the gamers oh they just
2:03:35 got the crypto Bros yeah whatever
2:03:37 oh [ __ ] ai's really cool too oh man we
2:03:40 want to be in there
2:03:44 I don't know [ __ ] crazy to me what
2:03:47 three AI Services would you recommend
2:03:49 for online marketing
2:03:52 that's one
2:03:55 um
2:03:56 canva
2:03:58 for because they you can do that uh
2:04:00 automated
2:04:02 image generation [ __ ] so you just go
2:04:04 write all your copy here
2:04:07 you get good at prompting you could use
2:04:09 you could use things like Jasper and
2:04:11 copy AI but but like right now like I
2:04:14 feel like
2:04:16 all you're doing there is kind of
2:04:17 offloading your understanding of what's
2:04:19 possible to someone else
2:04:22 like I don't know it just it feels a
2:04:24 little early to me to be leaning on sort
2:04:27 of pre-written prompt libraries which is
2:04:29 essentially what they are so I would I
2:04:31 would do chat GPT
2:04:35 I would do canva and I would do
2:04:37 mid-journey or or now I no actually I
2:04:40 would probably do
2:04:42 um
2:04:43 but it would be four
2:04:45 before
2:04:49 mid Journey
2:04:51 um ideogram because it does the text and
2:04:54 then
2:04:55 um
2:04:56 canva because you can just [ __ ]
2:04:58 produce a bunch of [ __ ]
2:05:02 yeah that's what I do right now but I
2:05:05 guarantee you three months from now that
2:05:06 list is going to be different like
2:05:08 don't think that anything that you do
2:05:12 right now anything that you learn right
2:05:14 now is going to be relevant six months
2:05:16 from now it's not all the tools are
2:05:18 going to be different
2:05:20 I sell Tech market research to Wall
2:05:22 Street and other Tech investors
2:05:24 interesting times no [ __ ] isn't it
2:05:27 like it's fascinating times tech tech
2:05:30 market research like how do you how do
2:05:34 you research
2:05:36 I mean these tools are fundamentally
2:05:39 changing
2:05:40 how everyone including engineers
2:05:45 and I T people
2:05:47 all of them these tools are changing how
2:05:49 we interact with computers how we even
2:05:52 think about programming and what
2:05:53 programming is
2:05:56 right now today it still looks a lot
2:05:58 like programming you're you know they're
2:06:00 they're pair programming with GitHub
2:06:02 copilot so it's making them more
2:06:04 efficient but they're still kind of
2:06:05 approaching it the way they normally
2:06:06 approach it but when these things start
2:06:08 to get sophisticated enough that you
2:06:10 actually don't need to code
2:06:13 and we're close
2:06:16 go play with chat gpt's code interpreter
2:06:19 the advanced data analytics and just
2:06:21 watch it write its own code and execute
2:06:23 it and then realize it [ __ ] it up and
2:06:25 rewrite it and do that until it gets it
2:06:28 right just go watch that and now imagine
2:06:30 that five years from now
2:06:33 so what is even programming going to
2:06:35 look like I don't know
2:06:37 I think programming looks like I want
2:06:39 the computer to do this
2:06:44 and then the computer goes okay
2:06:47 and it goes and does it
2:06:49 so how do you like
2:06:53 I don't know how you predict what do you
2:06:55 what what do you predicted in in that
2:06:57 future
2:07:01 other than
2:07:03 you got to be in the game like you gotta
2:07:05 play to win gotta be in the game gotta
2:07:07 be in the game gotta be in the game
2:07:08 what's that mean I don't [ __ ] know it
2:07:10 means be curious be adventurous you're
2:07:11 on a [ __ ] Adventure
2:07:13 it's dark in the woods it's dark in the
2:07:15 forest
2:07:16 there are scary things there are
2:07:18 poisonous animals
2:07:19 I don't want to be in the forest okay
2:07:21 well do you want the whole village to
2:07:22 get killed no okay then be in the forest
2:07:25 go figure out how to kill the dragon
2:07:30 it's the only way it's the only way
2:07:32 through I think
2:07:35 I
2:07:37 I wish there were another way
2:07:41 but
2:07:43 holy [ __ ]
2:07:48 you know amazing times
2:07:51 really amazing times
2:07:53 like profound times
2:07:56 and here's here's what's
2:08:00 you know I talk about you know we're
2:08:02 lucky to be a part of it this early
2:08:05 many many many people have never even
2:08:10 heard of chat GPT
2:08:13 they don't know what this [ __ ] is
2:08:14 they're not interested in it they don't
2:08:16 give a [ __ ]
2:08:17 Mal what's wrong Champion you wanna go
2:08:20 out
2:08:36 yeah I mean like like
2:08:43 this is like like we're gonna we're in
2:08:46 an amazingly privileged position if if
2:08:49 you're if you're willing to be open to
2:08:51 it
2:08:54 to be actively participating at whatever
2:08:57 level you can like it doesn't matter you
2:08:59 don't have to be a
2:09:00 a data scientist you don't have to be an
2:09:03 engineer
2:09:05 to be engaged with this stuff you could
2:09:07 be using chatbt to help your church
2:09:09 group do the bake sale better
2:09:13 and you're still in a remarkably
2:09:15 privileged
2:09:18 situation
2:09:20 to have an awareness that these tools
2:09:22 exist
2:09:24 before
2:09:25 the tsunami wave actually hits
2:09:30 because anyone who's
2:09:34 actively engaged has just massive
2:09:36 opportunity in front of them I know it
2:09:38 doesn't feel like that right now but
2:09:41 you're like but Kyle they lie and
2:09:43 they're not perfect and they're length
2:09:44 and then and they're janky pieces of
2:09:46 [ __ ]
2:09:47 yeah
2:09:48 today
2:09:51 three years
2:09:53 where were these things three years ago
2:09:56 you had to be an engineer there was an
2:09:58 API for gpt3 well three years ago it was
2:10:00 gpt2 which kind of sucked
2:10:04 gpt3 was the first one that was good
2:10:06 that was two years ago
2:10:09 nine months ago was chat GPT
2:10:13 seven months ago was gpt4
2:10:17 this fall is going to be Gemini and
2:10:20 probably
2:10:23 May of 24 is going to be GPT five
2:10:28 Apple's gonna do something by that like
2:10:30 what's what's this [ __ ] going to look
2:10:32 like a year from now
2:10:34 holy [ __ ]
2:10:39 so if you think you're behind you ain't
2:10:41 behind you ain't behind at all
2:10:44 the fact that you're
2:10:46 spending a Wednesday night 50 people are
2:10:48 spending a Wednesday night
2:10:50 listen to some old
2:10:52 fat [ __ ] on Tick Tock ramble about this
2:10:56 [ __ ]
2:11:02 you're gonna have opportunity so I don't
2:11:05 know I don't know
2:11:07 an irregular creative it should be
2:11:10 lawyers judges doctors I agree
2:11:13 really enjoy your content learning AI
2:11:15 thank you Jesse from Austin I appreciate
2:11:16 that I'm a [ __ ] copywriter chat GPT
2:11:18 makes me look really good just a tool
2:11:20 exactly Shane
2:11:22 if you're a great copywriter this will
2:11:24 amplify that if you're a shitty
2:11:26 copywriter this will make it better
2:11:27 right
2:11:29 if you don't know a copy at all like
2:11:31 I'm a good photographer I'm a good
2:11:33 writer I'm a [ __ ] illustrator and and
2:11:37 designer
2:11:38 not anymore
2:11:40 not anymore
2:11:49 life has always been can't beat him or
2:11:51 join them or get left behind yeah
2:11:53 exactly
2:11:54 yep
2:11:56 we can cross the by I don't know what
2:11:58 that means oh cross the river with lots
2:12:00 of turtles exactly exactly Winston
2:12:06 that's [ __ ] awesome
2:12:08 uh we'll be we'll be buying and selling
2:12:11 self-insurance that's actually
2:12:12 interesting Shane so I was thinking
2:12:14 about this the other day every employee
2:12:17 think about this
2:12:19 so one of the things that these things
2:12:21 do is you can take any data from
2:12:22 business
2:12:24 you can embed it and talk and gptize it
2:12:27 right so so we're going to be able to
2:12:28 take any data that exists
2:12:31 and toss it into a large language model
2:12:33 and then be able to do remarkable [ __ ]
2:12:35 with it as these tools get more
2:12:36 sophisticated you're going to be able to
2:12:38 take like all the emails you've ever
2:12:40 written and just be able to interact
2:12:43 with all that [ __ ]
2:12:45 every employer is going to be able to
2:12:48 take the collective output of all of the
2:12:51 employees in you know in their history
2:12:55 take all their Word documents their
2:12:57 Excel documents the programs they wrote
2:12:59 the [ __ ] they wrote the the marketing
2:13:02 copy that they did
2:13:03 and within there there's going to be
2:13:05 some gems that they create
2:13:07 applications out of your work
2:13:12 so so just like the SAG actors are
2:13:16 fighting for hey they're going to take
2:13:17 my likeness and turn that into acting
2:13:19 well in a business context you know
2:13:22 IBM's gonna take employee 6257's
2:13:26 lifetime of work and turn that into a
2:13:29 cool application that they're going to
2:13:30 monetize
2:13:31 does employee 6527 get any of that money
2:13:34 do we all need agents
2:13:37 Maybe
2:13:39 because you can exploit any data easily
2:13:43 with these tools
2:13:45 what's any data I don't know anything
2:13:47 you've ever done
2:13:50 for a company for this company for that
2:13:52 company
2:13:54 depending on your employment agreement
2:13:55 you may or may not own the rights to
2:13:57 [ __ ] you do if you do own the rights and
2:14:00 they monetize stuff
2:14:03 based on your work
2:14:05 is there profit sharing I don't know
2:14:08 it's like it's fascinating
2:14:10 crazy I want a parody song of celebrity
2:14:13 by Brad Paisley
2:14:17 um it's my bedtime in Michigan oh and
2:14:19 it's 64 degrees what is it here it is
2:14:22 73. all right listen I gotta go
2:14:25 everybody I'm gonna jump to the bottom
2:14:27 here
2:14:28 um
2:14:29 I'm tired
2:14:32 I I I know this one got philosophical
2:14:35 tonight but you know that's the deal
2:14:36 like I just try to respond to what the
2:14:39 what the comments are and kind of what's
2:14:41 going on with me
2:14:42 um
2:14:43 this is this is deep deep [ __ ] this is
2:14:45 deep stuff and I know that the tools are
2:14:47 exciting and it's it's all fun but this
2:14:49 is uh
2:14:53 I don't know this is why I'm spending so
2:14:55 much time doing this because I want to
2:14:57 keep talking about it because I think
2:14:58 it's important to talk about it and
2:14:59 understand what's coming so anyway Vicki
2:15:02 loves the philosophical ones thank you
2:15:04 so let's do a little Romper Room tonight
2:15:06 all right who's in here
2:15:08 we got some good ones we got some other
2:15:09 Irregulars okay farmer
2:15:11 um welcome to Denver congratulations on
2:15:14 a safe trip glad you're here Share Bear
2:15:17 uh Emilio's wife Tobias Shane thanks for
2:15:21 commenting tonight Winston good to see
2:15:23 you uh Stevo Stevo's in the house cover
2:15:26 Uno
2:15:28 getting AI everything's gonna get a okay
2:15:31 every it's going to be everywhere it's
2:15:34 gonna be everywhere Ted looks at life
2:15:36 thank you for tonight thank you for
2:15:38 participating
2:15:39 Vicky all right Rebecca Johnson good to
2:15:42 see you all right everybody
2:15:45 peace peace peace peace out uh have a
2:15:48 good one I'll see you tomorrow oh look
2:15:50 the machine just went to sleep perfect
2:15:51 timing peace