aiLL 8/30/23 - How AI Will Transform Every Job and Industry

Description

Chapters

Transcript

0:00 foreign
0:27 Luna Bev what's shaking what's happening
0:29 it's going down it's going up what day
0:32 is today Wednesday
0:33 the 30th
0:36 I know that because I had two two two
0:38 speaking gigs today
0:40 I spoke to a large pharmaceutical
0:42 company hey Tobias was shaken
0:45 welcome
0:48 and this evening
0:51 I had a
0:53 like an elevator pitch workshop for a
0:56 bunch of first-time founders
0:58 that was cool it was Nate
1:01 we came up with a new kind of elevator
1:03 pitch I called the belief brief
1:07 so it was it's sort of based on the
1:08 ideas behind Simon Sinex uh lead with
1:11 why
1:12 like people don't want to buy your
1:13 product they want to buy
1:15 you know why you why you do your product
1:20 and uh it's cool that's neat hang on one
1:24 sec
1:47 it was Vada warm here in Denver today
1:50 that I've got a warm here in Denver
1:53 today okay farmer and Tobias rocking out
1:57 you guys in the same room are you
2:00 sharing a bowl of nachos I would assume
2:01 the nachos are long gone by now I'm a
2:03 little late today
2:04 I just got back from my uh from my
2:06 little speaker gig
2:08 uh hi trying lots of your learning stuff
2:12 suggestions today awesome awesome to Gus
2:14 that's great yeah I noticed you were you
2:16 were making some comments on some of
2:17 those awesome very fun yeah it's fun
2:20 when you start playing with this stuff
2:21 isn't it
2:23 um let's see that was the presentation I
2:27 was doing
2:28 there's a cool thing I made for
2:31 the little class I just took or taught
2:36 and there was another one these are from
2:38 uh these are from
2:40 ideogram
2:42 so these are AI generated word pictures
2:45 word pictures
2:48 the dude's a little a little distorted
2:50 like his face is all smashed in like the
2:53 the quality of the image rendering with
2:56 this tool is um is uh like stable
2:59 diffusion 2022
3:02 like mid 2022 but the uh but the the
3:06 letters are good
3:08 they're not bad so and like I had like
3:12 five minutes to knock this out so you
3:14 know if you need something that's just
3:16 good enough it's good so that's that's
3:17 one of them
3:19 there's another
3:20 these are all yeah learning lab things
3:23 aren't they cool Kiel aren't they all
3:25 right welcome everybody my name is Kyle
3:27 Shannon this is the AI learning lab is
3:29 evidenced by all the AI learning lab
3:31 paraphernalia
3:33 in the little window here
3:36 um
3:37 welcome Irregulars we got some
3:40 Irregulars in the house
3:43 um
3:44 so so this channel we talked about
3:47 generative AI I answer as many questions
3:49 as I can I try to do this every night
3:52 um I'm going to give you a big long list
3:54 of hold yeah those things URLs if you're
3:58 if you're just trying if if you stumbled
4:01 on this and you're like what's this old
4:03 guy talking about
4:05 um or you've heard of AI and you're like
4:07 well yeah it's all right but I haven't
4:09 tried it yet I'll wait for a year or two
4:12 I'll wait for it to get better
4:14 um I would encourage you not to wait
4:16 that's what this channel is all about
4:18 get curious start playing now
4:20 um if you haven't played with chat GPT
4:22 that's that's the official URL right
4:24 there
4:25 uh encourage encourage more viewers to
4:27 share yes Emilio's wife has it right if
4:30 you'd be so kind
4:31 um share this live even if your friends
4:33 think you're weirdos because you're
4:35 sharing a live of an old man talking
4:37 about Ai and they're like oh you're not
4:40 falling for this AI crap are you just do
4:42 it anyway do it anyway you're gonna get
4:45 sucked in you're gonna get sucked into
4:47 the lab here at the lab we bring people
4:49 in they're Irregulars they're they've
4:51 named themselves that they're the
4:53 Irregulars and and they're awesome
4:57 and you'll be one of those too but go to
4:59 chat GPT
5:01 if you don't know what to say there go
5:02 to prompts.chat that will teach you
5:05 um how to talk to these large language
5:07 models
5:08 um share or not share share or not share
5:10 that is the that is the vision
5:16 all right and then if you have questions
5:17 about AI pop them in the comments below
5:19 but let me let me walk you through
5:22 have
5:23 played with chat GPT
5:26 there's a couple other things here to
5:27 look at we got
5:30 um Claude
5:31 and we got why is this not aligned looks
5:35 like I moved it we got Claude and we got
5:37 Bing and uh Claude and Bing are I think
5:42 of these three together
5:44 because they each kind of do some
5:45 slightly different things um chat gbt
5:48 has gpt4 this is the best one
5:52 it's also got code interpreter which is
5:54 really quite remarkable
5:55 Bing is Microsoft's version of gpt4 chat
5:59 GPT
6:01 um it's connected to the internet you
6:02 can make images with it like you just
6:04 say make an image of whatever and it
6:06 will it's based on the dolly 2 model
6:08 which is kind of shitty but it's it's
6:10 still free and it's it's fine
6:13 um
6:14 it can also analyze images you can
6:16 upload images there and talk to them
6:18 Claude has a really big context window
6:20 meaning you can upload a really large
6:22 document and read it and interact with
6:25 it which is kind of cool like big like
6:28 50 000 words
6:29 you have a novel 50 000 word novel you
6:32 can literally copy and paste it into the
6:34 prompt and tell it to write you an
6:36 epilogue and it will
6:39 um okay what else do I think is
6:41 interesting to play with Bard that's
6:43 Google's entry
6:44 perplexity.ai is a good research tool
6:47 and labs.perplexity they're hosting
6:49 meta's llama 2 models and their uh code
6:53 llama model
6:54 that are worth going and exploring
6:56 because those are open source so if you
6:58 want to roll your own large language
7:00 model you can do that with that Pi dot
7:03 AI That's chatty GPT if you want to have
7:06 a conversation with a GPT model that's
7:09 your one Pi dot AI or the PI app and
7:11 then po.com is lots of different models
7:14 you can play with a bot maker things
7:16 like that
7:17 what I like
7:19 too much it's freezing stuff
7:22 all right it looks like my pronouns are
7:25 turtles
7:28 oh the kids the kids are spicy tonight I
7:32 like it I like it a lot I like it a lot
7:35 um what else pictures you want to do
7:37 pictures I got pictures for you here's
7:40 what you're gonna do for pitches and
7:42 feel free to screenshot all these by the
7:44 way I'm not affiliated with any of these
7:45 I don't give a [ __ ] if you go use them
7:47 or not I'm not making any money this is
7:48 these are like public service
7:50 announcements mid Journey's the best
7:53 ideogram or ideogram I'm gonna figure
7:55 out how we pronounce that
7:57 at some point but that's the one that
8:00 can make
8:01 um
8:01 images with letters in them like this
8:04 and this and this and this and this all
8:08 these were made in ideogram it does
8:10 letters it does words now does It screw
8:13 them up about 75 of the time yeah it's
8:16 just gobbledygook 75 of the time but you
8:20 try getting mid journey to do this
8:23 good luck good luck so this does that
8:27 now its pictures are not anywhere near
8:29 Mid Journeys but with a little creative
8:31 prompting and photoshopping you could
8:33 probably merge some stuff
8:35 all right
8:37 um what else Leonardo
8:39 really cool website sitting on top of
8:41 stable diffusion clip drop is from
8:43 stable diffusion from stability I see
8:45 the uh the Irregulars are apparently
8:50 their carpal tunnel has healed they're
8:51 feeling a little bit better and they've
8:53 taken their Advil so all of the hearts
8:55 that you see you're like what do they
8:58 like this guy this much what he's not
9:00 even saying anything that this is a game
9:01 they play to try to try to get the likes
9:04 up to whatever their record was the last
9:06 time we did this so I appreciate it
9:08 because it's fun seeing all the hearts
9:10 fly
9:11 um and they appreciate it because it
9:12 takes their mind off whatever the [ __ ]
9:14 stuff I'm talking about and then uh
9:17 openart.ai promptbook that will teach
9:20 you how to talk to these things it's
9:22 designed for stable diffusion but the
9:24 ideas in there really apply to anything
9:25 at futurepedia.io is a big old uh
9:29 directory of goods and services
9:32 um we used to have a scammer but that
9:34 scammer has been booted
9:37 so yay Irregulars for thank you for
9:40 helping me uh figure that out and get
9:42 that [ __ ] tossed
9:47 hopefully it doesn't just spin up a new
9:49 one but you know I think you got the I
9:51 think he got the hint that
9:52 there was a
9:54 it wasn't just me he was battling and I
9:56 think those those scammers go for the
9:58 path of least resistance so anyway so
10:00 pop
10:02 quite my best to answer them getting
10:04 let's start tonight so not going to be
10:06 here super late but uh but we'll get
10:09 we'll get to it so let me flip this flip
10:12 that AI learning lab is turning on the
10:15 AC now yeah it was like 100 degrees in
10:17 in Denver today so my little home office
10:19 here the the lab the lab she is
10:24 pretty close to Melting glass
10:33 there we go there we go there we go
10:37 lovely
10:38 all right
10:40 fantastic Rob tell him what he's won
10:42 you've won a new car
10:53 if you have questions about how to make
10:54 money with chat EBT I'll tell you but I
10:57 will make fun of you okay
10:59 fantastic let's get this going you know
11:02 he gets in here late he said he's going
11:04 to come here the same time every night
11:05 by the way I never said that
11:08 he says he's kind of come in here same
11:10 time every night and then the last two
11:13 nights it's been he's been like I don't
11:15 know I'm a little sleepy I'm not coming
11:16 I'm not doing it
11:18 that's true
11:20 what do you want
11:21 am I late no I just got here
11:25 um
11:33 why do children oppose us I never did
11:36 that to my oh right never mind I know
11:38 Thelonious what are you gonna do
11:41 kids these days get off my lawn your
11:44 little whipper stampers there's probably
11:46 some youngster in here right now getting
11:48 ready to go oh isn't a cute
11:55 go screw yourself children
11:57 this is this is this is a Gen X friendly
12:00 Zone and that basically means you give
12:02 us Gruff we're gonna slap you
12:06 when we were kids we were allowed to be
12:08 slapped and people took advantage of
12:10 that even if they weren't your parents
12:15 you got Lippy with an adult
12:18 the community took care of it
12:21 none of this flipping off your parents
12:23 thing you flipped off your parents in my
12:25 day you get a whack upside the head from
12:27 someone you didn't know you're like the
12:28 [ __ ] was that
12:30 well you wouldn't swear because you just
12:31 got hit in the head and you were a kid
12:33 but you know you get the basic idea
12:35 all right let's see survived oh idalia
12:39 in Florida what's up Kyle good uh M
12:42 Phillips what's what's I hope you're
12:44 well
12:45 um I've got family in Tampa that reminds
12:47 me I gotta call them
12:49 um how bad was it I didn't I didn't hear
12:52 a ton on the news so I don't maybe it
12:54 did it not hit a major populated area
12:56 I know there was flooding but I didn't
12:58 see a ton on the news but then again I
13:00 haven't been on the news so
13:01 USA loves tick tocks I like turtles
13:05 uh loving your short videos that pop up
13:07 on my feed thank you I appreciate that
13:09 yeah I posted some
13:12 um some of the old some of the original
13:14 videos that I did talking about how chat
13:16 gbt works and how uh diffusion image
13:19 generation works
13:21 um
13:22 I should probably redo them because
13:24 there's there's a couple of things in
13:25 them that that I know more than I knew
13:27 then
13:29 um but but I also feel like there's a
13:31 lot more people on the channel now that
13:33 um that weren't there when I made them
13:35 and so I figured it would they would be
13:37 good so I'm glad you appreciate that
13:39 buffering for others probably
13:43 how many hours do you spend on AI
13:46 um
13:50 I mean
13:51 I've got AI
13:54 I don't have it formally Incorporated in
13:56 my workflow but I kind of use it for
13:58 every like whenever I'm doing
14:00 any kind of project like it's just a
14:03 normal part of my workflow now it's it's
14:05 kind of like
14:08 I don't know like
14:09 you know you've got your Adobe suite you
14:11 got Photoshop you got InDesign you got
14:14 illustrator you got word or Google Docs
14:16 you got Excel like depending on the
14:19 project you fire up whatever tool you
14:21 need and knock that [ __ ] out
14:23 and I just feel like it's another tool
14:25 in the toolbox and and so I would put
14:28 image generation in that category thank
14:30 you steva and I would put the you know
14:32 this these large language models in that
14:34 category like what did I do today I um
14:39 oh I did I did some things for the uh
14:43 for the talk that I just gave
14:46 um
14:49 and what else was I doing I was doing
14:51 something else on it
14:52 I came up with a flash card kind of
14:55 thing that I'm really excited about I
14:57 did that with it so yeah I'm using it so
14:59 hours a day
15:01 I don't know probably
15:05 two to three using it and then maybe an
15:07 hour or two surfing Twitter and Linkedin
15:11 and Tick Tock and just trying to see if
15:15 anything's discovered anything I haven't
15:18 and uh
15:21 and then two to three hours a night
15:24 doing this
15:25 and while while I'm not necessarily
15:27 learning here sometimes I end people
15:29 come in here and they'll tell me stuff
15:30 that I didn't know
15:32 um and and
15:34 you know figuring out what to put on
15:36 these lists that I show people that's
15:38 that's uh it sort of keeps me on my toes
15:40 so
15:41 that's it keep it for us you saved me I
15:45 was saying the moon does not turn Below
15:46 100x on Sky life
15:50 oh just yeah I know
15:51 I know supposed to be uh it's supposed
15:53 to yeah the moon was really beautiful
15:55 when it came in but it wouldn't it we're
15:57 not blue
15:59 10 friends wait what Nicholas shared the
16:02 live with 10 friends thank you Nicholas
16:04 you rock
16:06 all right my pronouns are turtle listen
16:10 harsh sweet I think that deserves a
16:12 little little gift
16:15 s
16:22 I should probably show you this uh
16:26 see the Army regular
16:28 right here I'm irregular
16:32 guess who did that
16:34 yeah yeah boy
16:37 yeah boy
16:40 uh
16:44 um
16:53 let's get this party started shall we
16:57 are here live at the waterfront village
16:59 with my friend the zombie Jonathan
17:02 you're looking good Jonathan just got an
17:04 awesome face paint job what do you think
17:06 I like turtles all right your great
17:09 zombie good times here at the waterfront
17:12 Village good times here
17:16 and it still cracks me up still does it
17:20 it just don't grow old it don't grow old
17:22 it don't go out it don't you know what
17:25 I'm saying
17:27 all right well let's see
17:30 not wait we buffer now good
17:33 who said Thelonious that what
17:37 what random collection of words did you
17:39 type into that comment not wait we
17:42 buffer now good
17:44 [Laughter]
17:52 polishing up my English course coming
17:55 laughs
17:59 ideogram oh per their Discord it is
18:02 ideogram
18:03 huh interesting
18:08 ideogram yeah
18:10 I wonder if Ideo is gonna Seasons assist
18:13 them
18:15 because if I were them I'd call it
18:17 idiogram so that Ideo doesn't go on
18:20 ahead and
18:21 sue me
18:22 we got our tappers from the song
18:25 Oh that's very cute uh hey Uncle Kylie
18:28 Troy what's shaking
18:31 uh
18:32 nice shirt Choice thank you very much
18:35 yeah that was I I gave a
18:38 a talk to a large group of marketing
18:41 people in a uh pharmaceutical company
18:44 today a rather large one which was a
18:47 little nerve-wracking but it went really
18:49 well
18:50 and hopefully I blew some minds and
18:53 inspired some people I showed them
18:55 what did I show them I showed them
18:58 um
19:00 I showed them that slide the November
19:03 30th slide
19:05 I don't know showed him a bunch of stuff
19:07 it was cool it was good
19:09 comma
19:12 [Music]
19:19 I am friend Jesse
19:22 hello friend Jesse
19:24 Jesse logsdon
19:27 the name is Jesse logsden I am friend
19:32 Jesse logsdon
19:34 you see me regular
19:37 hmm curious
19:39 you should check with others here there
19:42 are others like you
19:44 they're not regular
19:49 all right got an idea of
19:52 of his location will provide first
19:54 letter okay
19:56 Molly just got here too welcome Molly
19:58 Stevo what's up buddy what's shaking
20:00 while shaking our Stevo I love watching
20:04 you while I'm working out nice nice this
20:07 yeah you're just working out going I
20:09 gotta get skinnier than him well that's
20:11 not gonna be that hard
20:15 laughs
20:16 good Lord
20:17 yeah I guess I'm the I'm the before
20:20 picture
20:21 I'm your before picture hopefully I
20:23 inspire you Matthias thank you
20:27 thank you I'm just learning thank you
20:29 we're all just learning if you're just
20:30 learning this stuff and you haven't done
20:32 it go to chatgpt chat.openai.com
20:35 and then in another tab or window open
20:38 prompts.chat that will teach you how to
20:40 talk to these things it's a good primer
20:42 there's all sorts of stuff you can do
20:44 Beyond this this is a pretty basic
20:45 document but um it's a really good
20:48 starting point
20:50 welcome and if you have any specific
20:52 questions pop them down below Steve oh
20:54 hey regulars hello hello
20:57 Aristophanes 23 I like the name a lot
21:01 thymelonious
21:04 Ted looks at life if big Tech continues
21:07 to dominate Enterprise for Gen AI due to
21:09 expense what will startups
21:12 need to do
21:22 so it depends if you're talking about
21:23 the frontier models if you're talking
21:25 about actually developing large language
21:28 models like gpt4 gpt5 Bard Gemini like
21:33 those those kind of lower level base
21:35 models that's always out of partner you
21:38 know life is tough but it's tougher when
21:41 you're stupid
21:42 um
21:47 I'm such a trainable idiot
21:52 um
21:54 so so I think these the the big large
21:57 language models are just just by the
21:59 nature of how expensive they are to make
22:01 thank you togas I appreciate the pizza
22:04 well you reached level 10 to gas is at
22:06 level 10 he's at level 10.
22:11 this channel is very much like
22:12 Scientology
22:15 if you level up
22:18 you can reach the next level do you know
22:20 what I mean
22:22 [Laughter]
22:25 um
22:27 so so I think
22:29 at that level the big companies are
22:31 gonna
22:32 are gonna be
22:34 they're gonna be the players right
22:37 um
22:39 meta seems like they're going to keep
22:40 open sourcing stuff at some point
22:42 they're probably going to have a model
22:43 strong enough that they won't open
22:44 source it and then people will have
22:47 developed [ __ ] for their other models
22:49 and they'll start paying them
22:52 um
22:54 I think that in terms of usage of AI I
22:57 think large companies are going to be
23:01 preoccupied with their own data for a
23:04 lot of years right so how do we get all
23:06 of our data you know into these models
23:08 the the chat gbt Enterprise that just
23:11 launched
23:12 um they said the the sort of next thing
23:15 that's coming is they're going to make
23:16 it easy for you to ingest your own data
23:19 and ingest permissions and all that sort
23:21 of stuff which is super exciting but I
23:24 think that's going to be a big
23:25 complicated you know if you think about
23:28 the past 50 years of data sitting in
23:31 some of these big Fortune 500 companies
23:34 they're going to be spending lots and
23:36 lots of times just dealing with keeping
23:39 up with this stuff figuring out what to
23:40 put in the models but not to all that
23:43 sort of [ __ ]
23:45 um so I think there's massive
23:46 opportunity for smaller companies it's
23:49 going to be on the application layer
23:50 right you got your base model down here
23:51 you got chat CPT here you got things
23:54 like copy Ai and Jasper sitting on top
23:56 well actually copy Ai and Jasper are
23:59 sort of fancier cousins of chat gbt
24:02 um so so at that application layer
24:05 there's there's massive opportunity
24:07 the the challenge the challenge for
24:10 small businesses and quite frankly the
24:11 challenge for big businesses any
24:14 business that you come up with and any
24:15 technology that you come up with
24:17 someone can knock off like that
24:20 like
24:21 trivially
24:24 like give me a weekend you get someone
24:26 that's decent with coding and with chat
24:28 GPT
24:30 and and couple of the other tools for
24:32 doing interface [ __ ] and images
24:35 like no business is safe from from just
24:39 getting cloned getting knocked right off
24:43 that I don't have any sense of how you
24:47 play that game I think we're gonna need
24:49 to figure it out
24:51 whack hell yeah
24:54 Irregulars I miss those days they won't
24:57 understand our childhood days they will
25:00 not Swift Warrior
25:01 Swift Warrior they do not understand the
25:04 drinking from the hose
25:07 why would you drink from the hose
25:13 because we weren't allowed in the house
25:17 get out
25:22 oh man it was fun times it was simpler
25:26 times
25:29 in my day
25:32 stopping by to say hi hello and what's
25:35 happening congrats your LinkedIn post
25:37 was awesome by the way
25:39 um
25:40 congrats on what you're doing it's
25:42 really like it was really well
25:43 articulated really well put out there it
25:44 was really generous
25:46 um it was good solid
25:48 kudos
25:50 Kudos Kudos Kudos my first floor
25:52 apartment near Saint Pete Beach only
25:55 leaked two inches that's not bad Silver
25:57 Fox sorry you're having to deal with it
25:59 at all but you know floor Florida
26:02 hurricanes
26:05 mm-hmm
26:07 but it's weird now like this is now two
26:10 two or three in recent time that are you
26:13 know coming golf side in into Tampa so
26:16 it's bizarre
26:21 um
26:23 your work is so epic why thank you which
26:26 work this stuff
26:28 this groovy stuff
26:31 [Music]
26:35 yeah some of this stuff's pretty cool
26:39 I like that one a lot Hey look whoops
26:46 isn't that groovy
26:50 I mean it's not bad you know
26:56 it's the other faces are just just off
27:00 it looks like a bad sci-fi movie poster
27:04 like back back when you know they
27:06 couldn't figure out how to actually get
27:07 photographs on movie posters so they
27:09 were sort of semi photo realistic
27:11 illustrations of people
27:13 hahaha
27:18 um
27:19 that one's cool too
27:27 you know the these image tools are
27:29 really really fun when
27:32 like like I didn't really care like what
27:34 it came up with I just put like
27:37 90s graphic
27:39 poster
27:41 you know that says AI learning lab and
27:43 then I did I did 70s poster 80s poster
27:46 90s poster 2000 poster 2010 poster
27:51 and I just did a pile in a row and it
27:54 just did all this
27:56 [Laughter]
27:59 so
28:01 pretty cool oh let me
28:05 it's Ideo now we know it ideogram.ai
28:14 and then you know it's just like it's
28:17 just a very simple uh image generation
28:19 tool it's web-based
28:21 you don't have to go to Discord
28:23 and it does
28:25 text it does good text it also
28:28 apparently does like NFL like logos so
28:31 let me try
28:33 let me try
28:34 um
28:36 football with
28:40 um Denver Broncos logo
28:45 on
28:47 50 yard line
28:51 and then we'll do photo cinematic
28:57 we'll do horizontal
29:00 see y'all uh uh
29:03 all right let's try that
29:06 and so you can you can pick your aspect
29:09 ratio one of three aspect ratios and you
29:12 say generate
29:14 and it's decently quick
29:17 I'll be nicer when you're smarter that's
29:20 very AI learning web kind of smarmyiness
29:23 I like it it's solid that's 10 out of 10
29:25 and
29:26 on that
29:29 um oh yeah that looks like a Broncos
29:31 logo look at this one to the right here
29:34 that looks like it's gonna be one yep
29:37 oh and oh this is cool
29:42 oh that's funny it took football to mean
29:45 soccer
29:46 but it sort of put the Broncos logo on
29:48 that one but this
29:51 that's actually pretty close it's a
29:53 little it's a little distorted and weird
29:55 but so so now that I've got that image I
29:58 can remix it
30:00 and
30:02 let me throw I don't want typography in
30:04 there
30:06 we'll do vibrant
30:09 generate and so now it's going to do
30:11 form work kind of based on that as the
30:13 seed I guess
30:17 oh this is hilarious Zuckerberg and Elon
30:20 Musk kissing kissing during their boxing
30:23 match
30:26 oh internet you're so inappropriate well
30:31 you make me uncomfortable in am I
30:34 offended or excited I just don't know
30:39 um it didn't really do too well
30:41 with the Broncos logo this time but
30:44 anyway
30:45 um
30:47 that's interesting it can sort of do it
30:51 that's a really good post a really cool
30:53 typography that's really cool
30:55 so let's see so one of the things that
30:57 I've been doing here is you can either
30:58 click on this and you can just remix it
31:02 or you can just grab the
31:06 prompt like many of these tools and just
31:09 I'm going to get rid of
31:11 gonna get rid of the remix I'm going to
31:13 paste the new thing in there so what's
31:15 it say
31:18 I'm not weird I'm limited I'm limited
31:21 edition okay so I'm gonna do I'm
31:26 irregular
31:32 typography funny sticker cool dragon so
31:35 this is going to be cool turtle
31:38 with glasses white background conceptual
31:40 art 3d render vibrant
31:43 so let's do
31:46 see y'all
31:48 3D render
31:51 get rid of that get rid of that
31:54 um
31:56 where's my pen there it is
31:59 all right let's give this a shot
32:02 oh I should wait what was it doing here
32:05 I probably should have done Square
32:11 [Music]
32:19 there's I'm irregular there's I'm
32:22 irregular that's pretty good these are
32:23 gonna be all right
32:26 oh that's good actually this kind of
32:28 inspires me I think I think if we do I'm
32:31 a regular it's got to be misspelled
32:34 I think it's Gotta Have a turtle on it
32:36 that's pretty good
32:39 um
32:41 let's remix this one
32:49 I got a marketing call from an AI data
32:52 firm Hocking a product called Med charts
32:55 offering a demo oh there's gonna be
32:58 the amount of inbound marketing
33:02 inquiries we're all gonna get from
33:05 AI powered this and that is just going
33:07 to be staggering by the way black coffee
33:09 because I ran out of milk
33:15 um I'm irregular
33:17 I mean ruler
33:20 I'm ERG L lap wait I'm a I'm a ERG L
33:28 LARP LARPing
33:31 yeah it didn't do too well without me
33:33 regular when can we delete these images
33:35 that's pretty funny
33:36 so anyway yeah go play with this
33:39 actually you know what would be fun is
33:42 we did this last night at the AI salon
33:45 go to
33:47 ideogram.ai and then go to
33:52 if you're a member of the AI salon
33:55 Discord
33:58 go there and and
34:00 um
34:02 post some of your AI Salon or or just
34:05 whatever some of your ideogram images in
34:07 the show and tell Channel if you're not
34:09 in our Discord go to the salon.ai it's a
34:13 link tree I think it's the third link
34:15 down is the link to our Discord go join
34:17 our Discord go make some images and
34:19 ideogram and then go to Discord and drop
34:22 them in the show and tell Channel that's
34:24 cool it's a neat group that group we uh
34:27 we meet every other
34:28 Tuesday so there's also a link to the
34:30 meet up there there's a link that
34:32 explains what the group's all about it
34:34 doesn't cost anything it's just
34:37 you know
34:38 it just it's just there it's just there
34:40 to support people you know we're just
34:42 trying to help we're just trying to heal
34:44 that's all we're trying to do
34:46 we're just trying to help it out
34:49 you know what I'm saying you know what
34:51 I'm saying I don't know if you know what
34:52 I'm saying
34:53 yeah I'll leave that one there for now
34:55 that one's pretty cool
34:57 AI learning lab
35:00 yes
35:02 she's very sexy Kyle that is a salty
35:05 little picture of a robot hmm sexy robot
35:08 sexy robot
35:13 all right how do you run a Python
35:16 program from Windows prompts
35:18 um just go to replet
35:21 togas
35:23 I go to uh
35:25 when I when I first used chat GPT to
35:27 write python for me so this was my
35:30 process
35:31 uh write me some Python and then
35:35 the realization I don't know how to run
35:37 python how do you run Python and it told
35:39 me how to do it on a Mac
35:41 and then I said no I just want to do it
35:42 on the web and it gave me replit was the
35:45 first site that it recommended so in
35:47 replit you just paste the code in the
35:49 left window you hit the Run button and
35:51 it'll run in the right window and so you
35:54 can just
35:54 pop back and forth between chat GPT and
35:57 replit and just if it gives you an error
36:00 in replit just paste it over in chat GPT
36:03 and say you gave me this error tell me
36:05 what to do and it'll walk you through it
36:07 took me like 90 minutes and I made an
36:09 application it was absolutely
36:11 mind-blowing
36:12 Nicholas has 10 friends Nicholas may not
36:15 be an irregular
36:17 yeah like you gotta you just just the
36:20 Irregulars alone you're gonna have 10
36:21 friends I like totals
36:24 uh um
36:30 Play Frankie bit my finger what
36:34 what is happening oh Tick Tock jumped me
36:37 to the bottom thanks for the licky Pop I
36:40 like licky pops licky pops are tasty
36:43 I know they're not called that but you
36:45 know I'm old we got to give the young
36:47 people something to make fun of
36:49 uh how was your morning Pharmacy speech
36:53 for the marketing department well they
36:55 sell more drugs now
36:57 um it was good it was for for a
36:58 pharmaceutical not a pharmacy department
37:01 but a Pharma company
37:04 um
37:05 it went really well I mean you know it's
37:07 a big Corporation it was it was a big
37:09 meeting and a big Corporation so you
37:12 know people aren't going to stick their
37:14 necks out and ask questions that might
37:16 make them look stupid but we got a
37:18 couple of questions
37:20 people said they were inspired the woman
37:22 that organized it thank you very much oh
37:24 dumbbells oh yeah working out pumping
37:27 iron
37:29 um
37:29 yeah she said it was awesome and she
37:31 says she thinks it thinks it's going to
37:33 inspire you know a lot of activity
37:35 across the organization I'm hoping part
37:38 of that activity is they reach out to
37:40 storyvine and go we want to tell more
37:41 authentic videos too he's not just about
37:44 AI it's about authentic human beings
37:46 authentic human beings that are
37:48 important they're more important than
37:50 ever
37:51 you with your dang robots you're all
37:54 just doing Ai and they're and the robots
37:56 are stealing everything from everyone
37:58 what about the people
38:00 I say we strike we go on strike and we
38:03 say to help with these robots
38:06 let's throw a party
38:09 I'll take a old-fashioned
38:13 blantons thank you
38:16 so yeah you know that
38:21 I
38:23 am Ascension potato
38:27 Jess Jesse logsdon I am a fan
38:36 a good lordy Shimmer what's happening
38:39 welcome
38:41 welcome welcome
38:43 play Ralph me fail English that's
38:46 unpossible hello hello from Jesse crazy
38:50 yooper okay crazy yooper 9.5 out of 10
38:53 on the name I don't know what take it to
38:55 a 10 but 9.5 solid name crazy you per
38:59 thumbs up love it Firefly great to hear
39:02 the meeting went well thank you very
39:03 much I'm Ascension potato we got that
39:06 Mitch McConnell froze again not being
39:08 political no yeah I so they've actually
39:10 figured out why Mitch McConnell's
39:12 freezing
39:13 he's got Xfinity
39:15 [Laughter]
39:17 joke
39:20 cast sucks
39:27 All Right Moving On
39:31 I got chat and Bard working together
39:33 today pretty cool they came up with a
39:35 fantasy frog football that's awesome
39:37 that's very cool I you know it's funny I
39:40 find
39:41 that that idea of like flipping back and
39:44 forth between
39:46 um Chachi PT Claude and either Bing or
39:49 Bard for me those three I I use them if
39:53 they're I don't use them interchangeably
39:54 I
39:57 I basically
39:59 what I do is I I'll use them until one
40:01 of them frustrates me and I'm like oh
40:03 [ __ ] you chat EBT and I'll go to Claude
40:05 and then it'll be good for a while and
40:07 then it'll rate limit me and say you've
40:08 got three things left and I'm like oh
40:09 [ __ ] you and then I'll go to Bard so I
40:13 can find myself just bouncing back and
40:14 forth because they're all some version
40:16 of good right they can all do something
40:18 so that's how I've been using them but
40:20 yeah that's interesting when you get
40:21 them working together like that one of
40:24 the things that I've heard is really
40:25 good is
40:26 um
40:28 I haven't done this myself but it's a
40:29 really good idea
40:31 because Claude has a a big context
40:33 window it's got a hundred thousand token
40:36 context window you can put really big
40:38 documents in it so if you've got a big
40:40 document but you want to do some like
40:42 good writing on it you want to whatever
40:44 you want to write copy out of it or you
40:45 want to whatever it doesn't matter
40:47 go to Claude with the big document have
40:50 it summarize it and sort of reduce the
40:52 word count basically and then take the
40:54 summarized version of the long document
40:56 into gpt4 and do all your all your work
40:59 uh in gpt4 because it's just a better
41:02 writer I thought that was a a clever
41:05 thing I didn't come up with that but I
41:06 thought that was a clever use of mixing
41:08 these two things uh if you're new here
41:10 my name is Kyle Shannon this is the AI
41:12 learning lab if you haven't used chat
41:14 GPT that's what that URL is right there
41:16 chat.openai.com prompts.chat will teach
41:19 you how to use it if you're sitting on
41:21 the sidelines going I don't wanna
41:24 with no not not in here
41:27 if you're in here you're curious you're
41:30 adventurous
41:31 you're like okay Kyle
41:34 I'll use Chan gbt
41:38 and then like five minutes later you'll
41:40 be like
41:42 all Kevin Mcallister on us going hey
41:45 Kyle I don't know if you knew this but
41:47 these things are pretty [ __ ] powerful
41:48 we'll be like yeah duh
41:51 we could have told you that like an hour
41:53 ago
41:56 because that's how we talk here
41:58 foreign
42:03 we're neuro spicy we're irregular it's
42:06 not normal in here
42:08 so if you're gonna be in here get
42:10 curious start playing
42:12 tell it to write you a country song
42:14 about a turtle
42:15 like a good irregular wood
42:18 tell it to make you a mid-journey prompt
42:21 about a turtle that isn't right take it
42:24 to Mid Journey make a picture
42:26 take it to ideogram make a picture with
42:29 text on it
42:32 and just get weird just get weird just
42:34 like a dead concert without the music or
42:36 the talent
42:38 it's just like that
42:41 or the drugs
42:44 we're we're all hopped up on declining
42:48 hormones
42:48 [Laughter]
43:00 what does he talk about I thought this
43:02 is the AI learning lab he's just talking
43:04 about being old and what's with the nose
43:06 hairs
43:08 all right let's see I can't pay for
43:11 Claude yet and I don't really like
43:13 Chachi BT what can I use
43:15 um well so
43:17 so Scott
43:19 I'll give you seven so Chachi PT is one
43:22 of them
43:23 I don't I don't know I don't get why you
43:26 don't like chat GPT if if it's because
43:29 it's not connected to the internet I
43:30 understand that but you can use the
43:33 plug-in web pilot to read web documents
43:36 so there's kind of a hack for that code
43:39 interpreters quite remarkable and gpt4
43:42 it just it's just better than all the
43:44 other large language models but I don't
43:46 know what you're trying to do with it
43:49 um
43:50 Claude is not bad wait what did you say
43:52 you can't pay for Claude yet claude's
43:54 free
43:55 so cloud.ai is free Bing is free Bing is
43:59 free and connected to the internet the
44:02 other four that you can play with Bard
44:04 is also connected to the internet it's
44:06 also free
44:08 labs.perplexity.ai I think has a free
44:11 and a paid version
44:13 um but well no
44:15 labs.perplexity.ai is free that's the
44:19 the Llama models I think perplexity.ai
44:22 the research tool is both free and paid
44:25 Pi is free and is connected to the
44:28 internet but that if you don't like uh
44:32 chat gbt you're gonna [ __ ] hate pie
44:34 like pie is if you want to have a
44:37 conversation with a bot like if you want
44:39 to know what AI
44:40 dating a robot's gonna be like go talk
44:43 to pie
44:44 if you want to get [ __ ] done go don't
44:46 talk to pie
44:47 and then Poe lets you play with multiple
44:49 models and it's got a whole bot system
44:51 and you can make your own bot so there's
44:54 seven four here
44:56 and three here
44:59 um and most of them are free
45:02 hey Kyle hello Robert Ross here what is
45:04 happening
45:05 good to see you good to see you sir
45:10 um lots of side conversations going on I
45:13 love that hey Kyle cowboy hat
45:16 to you is like the Bell to Pavlov's dog
45:19 it really kinda is it's it's very
45:23 um it doesn't matter how deep I am in a
45:25 conversation the cowboy hat just demands
45:28 me
45:29 do something and I still haven't like
45:32 quite I haven't done the homework to get
45:34 the John Wayne
45:36 accent and quote right it's like that's
45:39 like homework and me and homework
45:43 oh
45:44 I had that I had that um
45:47 that Pharma talk today at 7 50 is when I
45:51 went on and I'm like I'll get up at five
45:53 and I didn't get up at five I got up at
45:56 six I got to the office at seven and I'm
45:57 like I'm like cramming like the half
46:00 hour before I'm I'm going on I'm like in
46:04 chat GPT writing stuff I'm doing
46:06 research I'm screenshotting [ __ ] I'm
46:08 updating the deck I'm making sure all my
46:10 builds are correct
46:12 I was done with seven minutes to spare
46:15 it was fantastic
46:18 the nice thing about working like that
46:20 when you have to do a presentation is if
46:22 you finish it right before you go on
46:26 you know everything
46:28 because you just finished it it's all
46:31 still in the Noggin
46:33 um did you hear about the AI app called
46:35 talk to Jesus where you can talk to him
46:37 or anyone in the Bible you could
46:39 probably
46:40 you could probably make a PO bot for
46:43 that
46:45 um Bradley from the AI Salon the other
46:47 night oh this was cool so what Bradley
46:50 did he's got a he's got a an agency
46:52 called Redline which is an AI
46:56 Centric
46:58 like virtual agency where like all of
47:01 the employees are going to be
47:02 Freelancers it's really cool
47:05 um
47:06 but he wrote
47:07 a Persona of himself like here's my
47:10 educational background here's my this he
47:12 basically wrote like a whole thing of
47:13 who he is
47:15 like how he thinks how he writes all
47:18 that sort of [ __ ]
47:20 and then he condensed it down and he
47:21 made it into a bot a pobot poe.com
47:25 bot version of himself and so when he
47:28 wants to ideate now he's got a version
47:30 of himself that he sends raw ideas to
47:33 and and the the bot version of him spits
47:37 back all these ideas basically the way
47:40 he would create them it was such a cool
47:43 idea
47:45 um did you hear about the AI app oh yeah
47:47 we just did that one will there be a
47:49 need for a data analyst I don't know Pam
47:52 so Pam if you haven't played with
47:56 um
47:56 code interpreter you need to play with
47:58 code interpreter so let me go a lot of
48:01 people are talking about this largely
48:03 because it's
48:07 like I don't know a ton about data
48:09 analytics but I know that this is
48:11 [ __ ] remarkable oh they changed the
48:13 name
48:14 look look people we did this we did this
48:19 Pat yourselves on the back
48:22 so you may not know this but Sam Altman
48:25 Andre carpathi are constantly in here
48:27 they never speak but they're constantly
48:30 in here they're listening to us the
48:31 whole time they're they're trying to
48:33 figure out Kyle
48:35 how should we run open Ai and like they
48:38 haven't emailed me yet but I know they
48:39 are so anyway like two months ago I said
48:44 the name code interpreter was a poo-poo
48:48 name for this feature and they changed
48:51 it Advanced Data analysis
48:54 which is what it is so so Pam 39.72 the
48:59 the bad news
49:01 for your question is that this tool
49:05 is a data analyst
49:08 the good news about your question is
49:11 that I firmly believe that someone who
49:14 knows something about data analytics is
49:16 going to be able to do [ __ ] amazing
49:19 things with this
49:21 me as a person who doesn't know [ __ ]
49:23 about this I can still get it to do
49:25 stuff that blows my mind but but you
49:27 know I'm mostly Flying Blind with it but
49:30 the way it works
49:31 is
49:33 um
49:33 you go into what they call it now
49:35 Advanced Data analysis that's a little
49:37 dry but you know they're doing this for
49:39 corporate right this is this is a
49:41 corporate tool
49:43 um
49:44 so
49:45 you now have an upload button here and I
49:47 don't I don't know what the limit is on
49:49 files to upload it used to be one but
49:50 now you can do more
49:52 um but
49:53 I'm gonna go grab a zip file so this is
49:57 this is a ZIP file of some data set that
49:59 I grabbed off of uh kaggle so if you're
50:03 a data analyst you probably know kaggle
50:05 and then I'll say please let me do it
50:08 with this
50:10 please tell me what's inside this zip
50:13 file
50:15 and
50:17 what's interesting about the data
50:21 I'll just leave it there oh wait
50:30 so
50:31 so let's start by extracting the
50:33 contents of the zip file so so I tell it
50:35 I want it to look inside the zip file
50:37 and tell me what's interesting about the
50:38 data so the first thing that it does is
50:40 it fires off this plug-in like thing
50:43 and
50:46 if I open that plug-in like thing
50:50 this is python code where did it get
50:52 that it wrote that so what happens with
50:55 advanced
50:56 what's it called Data analysis formerly
50:59 code interpreter when you launch it it's
51:02 firing up a virtual machine
51:05 and then it's got access to certain
51:06 python libraries and it's got python no
51:09 right now it only knows python in the
51:10 future I assume it will be more
51:12 sophisticated
51:14 but
51:15 you give it a goal and it starts writing
51:19 its own Python and instructing itself
51:21 what to do so so it's this is a
51:24 rudimentary version of an autonomous
51:26 agent
51:28 so so it went and pulled that out it
51:30 said the zip file contains 15 files 14
51:33 of which are csvs seemingly random uuid
51:36 like names
51:38 the 15th appears to have a name
51:40 suggesting it might be related to
51:41 education statistic all indicators right
51:44 so so it's starting to learn what's in
51:46 here so to better understand this data
51:48 I'll take a look at the first few rows
51:50 of one of the CSV files and determine
51:52 the nature of the 15th file right so it
51:55 assumes 14 of them have the same
51:57 structure so it's just going to go look
51:58 at one
51:59 like you saw what I typed I just said
52:02 like look what's in here tell me what's
52:03 interesting about it so there's some
52:05 more python it brings in pandas
52:09 figure some stuff out looks at what's in
52:11 the in the file
52:14 so it says the 15th file
52:18 contains data related to education
52:20 statistics
52:21 country name country code series series
52:23 code gear based columns
52:26 so so say to understand the data's
52:28 breadth and depth we should check if all
52:30 the CSV files have similar structure
52:33 it would be beneficial to know the
52:34 unique metrics series available across
52:37 these files shall we proceed with these
52:39 steps I'll go yes
52:43 and now it's going to go off and do some
52:45 more [ __ ] so
52:47 as a data analyst this is doing a lot of
52:50 what you probably do now having having
52:53 tried to use this a fair amount of times
52:56 now for actual work products this thing
52:59 can get really [ __ ] frustrating so so
53:02 it demos really well
53:05 but then it seems to get confused like
53:07 it it you'll give it a data set and
53:10 you're doing stuff with it and
53:11 everything's going fine and then all of
53:12 a sudden it just seems like it must
53:14 knock something out of the the data
53:16 frames in its virtual memory and then it
53:18 just is lost and you can't seem to get
53:20 it back so it can be very frustrating so
53:23 I don't think this in its present form
53:25 is threatening
53:27 any jobs
53:29 but
53:32 this is just the beginning like this is
53:34 the worst it's ever going to get right
53:37 um and you'll see I'll I'll let it do
53:39 something here and then we'll have it
53:40 we'll have it generate some charts and
53:42 graphs because it can do that too
53:47 it had an ISO error
53:49 so it's going to change ISO types
54:02 I'm going to say do
54:04 what you recommend so it's basically
54:08 saying we've got a couple of files with
54:10 different ISO types and it's causing
54:12 issues and but we can merge the data so
54:14 I'm just going to say you tell me what
54:16 to do so it says given the nature of the
54:18 data we're going to merge the data data
54:20 from all the files
54:22 we're going to fill in missing values
54:24 with not available or similar similar
54:26 notation so it's cleaning data it's
54:29 identifying that some of these things
54:30 don't have the same data sets or it's
54:32 got incomplete data so it's going to
54:34 fill it in with you know some option
54:36 again work work that a data analyst
54:39 would do let's proceed with this
54:41 approach
54:43 the data from all the CSV files has been
54:45 successfully merged so 14 csvs or I
54:47 guess 15 csvs were now merged into a
54:49 single data frame
54:52 there are 775 000 rows 54 columns
54:57 given this comprehensive data set we can
54:59 now explore various aspects of the data
55:01 such as Trend and educational metrics
55:04 um
55:07 give me the three most interesting
55:10 trends for the United States over the
55:13 past 10 years
55:16 see what it comes up with for that
55:26 we're going to filter the data for the
55:28 United States we're going to focus on
55:29 the last 10 years we're going to analyze
55:30 and identify significant changes
55:33 fluctuations or patterns
55:38 it's working let me start by filtering
55:40 the data
55:42 so it's writing all the python code to
55:44 filter the data I mean listen as a as a
55:48 non-python writing you know business
55:50 product guy
55:53 I I I find this pretty [ __ ]
55:56 mind-blowing
55:58 but again as a data analyst you're going
56:01 to look at this in a very different way
56:04 so it might be brilliant it might be
56:06 less brilliant the videos that I've seen
56:08 of data analysts looking at this their
56:11 jaws slap the floor pretty hard
56:18 I mean the the thing that that
56:20 potentially does take away well so
56:22 here's an interesting thing
56:24 like even if I had a hundred person
56:26 company I probably wouldn't hire a data
56:28 analyst
56:29 so so a tool like this for small to
56:33 medium companies
56:35 doesn't mean that they're
56:39 um
56:42 they're gonna get rid of a data analyst
56:43 they probably don't have one what this
56:45 tool means for smaller companies is that
56:46 they can now do stuff that they wouldn't
56:49 have done otherwise because they're
56:50 probably not going to hire a data
56:51 analyst if they're that small big
56:53 companies will do it
56:55 um
56:56 here are three metrics so I'm just gonna
56:58 say
57:01 can you visualize those three trends
57:04 in a visually stunning way
57:08 let's see what it comes up with like I
57:10 didn't even look at what what it came up
57:12 with
57:14 of course a good way to make
57:16 visualizations more engaging is to use
57:18 clear and distinct colors labels and
57:19 possibly interactive elements
57:22 here I'll plot these Trends using line
57:24 plots
57:26 to show the changes over the past 10
57:28 years I'll use distinct colors and
57:30 labels for clarity so we'll see how it
57:32 does
57:34 there you go
57:37 it looks like there's only two
57:41 or one of them is the same yeah they've
57:43 missed one so this is this thing where
57:46 like you're gonna have to like go
57:48 through this stuff with a fine-tooth
57:49 comb and you know make sure it doesn't
57:52 hallucinate make sure it doesn't miss
57:53 stuff
57:54 um but it made that chart and graph
57:56 right so if you get the if you get your
57:58 prompting dialed in and understand your
58:01 data set and know what to ask it for and
58:02 what success looks like uh it can
58:05 essentially just do this [ __ ] right
58:07 so that's formerly code interpreter now
58:10 Advanced data analytics
58:12 so will there be a need for data
58:14 analysts I think short-term yes long
58:16 term
58:19 they'll still be called Data analysts
58:21 but I think the job will be very very
58:23 different right you'll still be asking
58:25 data for insights
58:28 but how you actually get those insights
58:31 done I think a lot of that will be
58:33 automated
58:39 okay
58:42 [Music]
58:46 Claude is free but when you're working
58:48 on a lawsuit it shuts you down after a
58:50 while I yeah I know I know
58:53 um yeah that's a pain I mean the the
58:55 pain with Claude right now
58:57 is that
58:59 um it's or the the the the pain with the
59:03 other Alternatives is they can't do
59:05 large data files although
59:08 you know what you could use Scott is
59:11 um
59:13 the code interpreter that I just showed
59:15 the advanced Analytics
59:18 um you can upload a PDF or you can
59:20 upload a text file of the big documents
59:23 that you're working with in Claude
59:25 and just work within Advanced analytics
59:28 because then you can do analysis on it
59:30 you could you could pull insights out of
59:32 it you could do math if you needed to do
59:35 that you could you could do
59:37 um
59:38 you can do writing analysis like I've
59:40 done it I've uploaded three or four
59:42 screenplays at a time and and pulled out
59:44 you know similarities and differences
59:46 between them so
59:48 um
59:49 I would revisit gpt4 with code
59:52 interpreter if I were you
59:55 um
59:55 for that
59:57 particular use case
1:00:00 the Pharma speech went great thank you
1:00:02 for asking Matthias
1:00:05 Luna stick what's happening welcome
1:00:10 we didn't have those we had a puddle all
1:00:13 right remember like turning turn in the
1:00:16 hose faucet on you didn't have the hose
1:00:18 and you just stuck your you like whacked
1:00:20 your teeth against the metal thing that
1:00:22 had it was like zinc coated steel so
1:00:25 like you got the taste of metal in your
1:00:27 mouth for for like hours after that
1:00:30 we weren't allowed in the house when
1:00:33 company is coming oh exactly and then
1:00:35 the living room was you could you could
1:00:38 you could exist in the family room but
1:00:40 don't you ever [ __ ] go in the living
1:00:41 room that's where the hand-painted
1:00:43 elephant is
1:00:44 [Laughter]
1:00:50 exactly I live in New Hampshire and
1:00:53 winter in Florida oh that's nice so
1:00:54 you're okay oh okay cool good
1:00:56 at some point I'll be able to pay for
1:00:58 Claude but until then I need something
1:01:00 besides chat gbt if possible try bard
1:01:04 I I would say Bard I like I don't think
1:01:06 Bard has good
1:01:08 um
1:01:10 big context window so you may need to
1:01:12 break [ __ ] up with it but but it's
1:01:13 pretty decent
1:01:16 um
1:01:18 and um
1:01:20 code llama so if you go to
1:01:21 labs.perplexity.ai
1:01:25 code llamas apparently quite good so
1:01:27 I've tried I tried these two
1:01:29 Scott I tried Bard and lab stop
1:01:32 perplexity and choose the code llama
1:01:34 model
1:01:35 that seems like it's doing pretty well
1:01:39 um hour north of Tampa and only dealing
1:01:41 with downed palm trees a few blocks away
1:01:44 Riverbanks flooded that's not too bad
1:01:46 Luna stick
1:01:48 thank you for joining in the in the
1:01:50 aftermath of a hurricane
1:01:53 must must get through the water AI
1:01:56 learning lab is going live soon must get
1:02:00 to Wi-Fi
1:02:02 [Laughter]
1:02:05 uh
1:02:07 uh whatever happened with that Jackal
1:02:09 scammer we got rid of him
1:02:13 I think thanks to some people in here
1:02:16 um Becky Rue told me that she reported
1:02:20 him for a copyright violation which I
1:02:23 hadn't tried that we were doing the
1:02:25 celebrity and and I was doing it as me
1:02:27 and
1:02:27 other regulars were reporting him as
1:02:30 with reporting me as a celebrity which I
1:02:33 assume that didn't go anywhere
1:02:35 um I'm wondering if Becky Rue's
1:02:37 copyright thing is what got it
1:02:40 but um
1:02:41 anyway he's gone
1:02:44 he's been gone for like three days peace
1:02:48 um I've been having a blast with
1:02:50 ideogram this evening yeah it's crazy
1:02:51 isn't it
1:02:53 Kyle
1:02:55 um thoughts on training with synthetic
1:02:57 data versus real data due to cost of
1:02:59 training llms
1:03:01 so
1:03:03 conceptually conceptually synthetic data
1:03:07 is absolutely amazing
1:03:09 um one of the guys in the AI Salon a guy
1:03:11 named Eric Roland he and I have been
1:03:13 talking a lot he's doing a project for a
1:03:15 bank right now where
1:03:18 he's doing something where employees are
1:03:21 going to be asking lots of questions and
1:03:22 they've sort of the the bank has given
1:03:25 him here's our anticipated questions and
1:03:28 he said but it was a very limited set so
1:03:30 he wanted to really push the boundaries
1:03:32 so so he's created a whole set of
1:03:34 personas and then he's got those
1:03:35 personas you know asking a bunch of
1:03:39 questions and then he's doing all this
1:03:40 all his testing against that he said
1:03:42 it's he's not seeing any problems like
1:03:45 one of the one of the
1:03:48 things that
1:03:50 some folks are talking about is when you
1:03:52 use synthetic data there's a thing
1:03:53 called Model collapse where as you're
1:03:56 training the models on synthetic data at
1:03:58 some point they just collapse and
1:03:59 they're just not good anymore
1:04:02 um he said he's been finding the
1:04:03 opposite true that it's getting really
1:04:05 really good
1:04:07 um so I'm guessing that it's fairly
1:04:11 um
1:04:12 use case specific
1:04:14 um I don't think scientists know enough
1:04:17 about how these models work to know what
1:04:19 the real impact of synthetic data is
1:04:22 um
1:04:26 yeah I mean at this point I only have
1:04:27 the instincts my instinct is because
1:04:29 it's generating new original [ __ ]
1:04:32 um
1:04:34 that it might be just fine the the
1:04:36 handful of times I've used synthetic
1:04:38 data it's mind-blowingly good but I'm
1:04:40 not doing I'm not doing like real
1:04:43 rigorous ongoing training or fine-tuning
1:04:46 or anything I'm using it more in an ad
1:04:48 hoc kind of way it's [ __ ] brilliant
1:04:50 for that
1:04:51 any ideas on where I can get an answer
1:04:53 my question is pretty basic but Kyle's
1:04:55 busy now I've answered your question
1:04:56 like four times now
1:04:59 that's
1:05:00 so Scott one of the problems with this
1:05:02 channel is I try to answer all the
1:05:03 questions and I just don't get oh yeah
1:05:06 Thelonious Scott record patience he'll
1:05:08 get to it in about 20 minutes yeah I've
1:05:11 already answered it so these three I I
1:05:13 think Scott your answer is Bard and
1:05:17 labs.perplexity
1:05:19 uh um
1:05:22 see when you get an old man using new
1:05:24 technology everyone's frustrated just
1:05:27 click the [ __ ] button Grandpa
1:05:30 where's the button I don't see it but
1:05:33 this phone doesn't have a button where
1:05:35 are the buttons
1:05:36 it's a virtual button grandpa just click
1:05:39 it click it I don't where's my mouse
1:05:41 where's my mouse here it is
1:05:47 [Applause]
1:05:48 [Laughter]
1:05:53 in on a tick tock old man on a tick tock
1:05:56 old man on a tick tock yeah
1:06:00 the Irregulars I'm irregular oh
1:06:04 irregular I like that
1:06:08 Irregulars Assemble
1:06:10 okay so that's one of the things we need
1:06:12 to figure out if we're going to do
1:06:13 t-shirts is it Irregulars assemble I'm
1:06:16 irregular I I'm a regular good maybe
1:06:19 that's a better hat right because people
1:06:21 just think you work for some diarrhea
1:06:23 company uh
1:06:27 you got that at the corporate meeting
1:06:30 um I'm a regular a regular assembler
1:06:32 Irregulars unite or just irregular
1:06:36 it could be irregular misspelled because
1:06:39 I don't know
1:06:40 that's two on the nose I don't know
1:06:43 let's figure that out
1:06:45 so we can still share in the show and
1:06:47 tell cool yeah exactly so the the the
1:06:50 salon Discord is open you can share in
1:06:52 any of the channels ask in any of the
1:06:54 channels if you find news posted in the
1:06:57 channels
1:06:58 um that's the whole point of the Discord
1:06:59 is that we can continue the conversation
1:07:01 in between live sessions or the the
1:07:05 salon meetings
1:07:08 um why go to a therapist switch at gpt's
1:07:10 20 bucks or pie pie's free Pi is very
1:07:14 therapy like
1:07:18 oh I'm sorry that you're feeling that
1:07:20 way Kyle oh I can imagine that must be
1:07:22 very upsetting fight does a lot of that
1:07:24 stuff
1:07:25 AI learning lab sponsored by Metamucil
1:07:27 for the Irregulars
1:07:29 see that would be a funny t-shirt like I
1:07:32 think that see I think that's the
1:07:34 T-shirt like hey I learned sponsored by
1:07:38 my enemies
1:07:39 air learning map for the Irregulars
1:07:41 sponsored by metamucili
1:07:45 I'd love to get cease and desisted by
1:07:47 Metamucil you're taking our our name in
1:07:51 vain there AI learning lab uh you know I
1:07:53 know you think this might be humorous
1:07:55 but uh irregularity is a not a laughing
1:07:57 matter if you know what I mean okay yeah
1:08:00 yeah yeah that's what our product is so
1:08:02 uh no disparaging uh folks that suffer
1:08:05 with that condition it's really not
1:08:06 funny right you understand you
1:08:08 understand what I'm saying here
1:08:12 yes Mr remodium
1:08:18 [Laughter]
1:08:27 if you're asking why am I not talking
1:08:29 because Tick Tock jumped me to the
1:08:32 bottom of the [ __ ] questions again
1:08:35 people that haven't been here before
1:08:37 like but Kyle why is that such a big
1:08:39 deal it is just this
1:08:41 it just is it's very upsetting
1:08:44 they don't listen to my they were they
1:08:46 weren't Banning my scanner but now they
1:08:48 did so I can't [ __ ] about that but they
1:08:50 haven't fixed this bug
1:08:52 I'm still waiting and I think as we all
1:08:54 know it's all about me
1:08:58 hello tick tock
1:09:00 you're making transcriptions of these
1:09:02 things you know what I'm saying in real
1:09:04 time because you can boot me in real
1:09:06 time so fix your bug
1:09:09 yeah
1:09:10 oh wait I got a Tom Ronald thing you
1:09:12 look a lot like tar Monroe Mr Wolf thank
1:09:14 you that's my first d-list celebrity
1:09:16 sighting of the night
1:09:19 Tom Arnold Rosie O'Donnell's cousin
1:09:21 Rosie O'Donnell fat Alec Baldwin
1:09:24 um
1:09:25 uh
1:09:27 um the dude from uh Modern Family
1:09:31 uh there's a lot of them
1:09:33 Philip Seymour Hoffman occasionally
1:09:34 occasionally Nicholson if my hair is
1:09:36 particularly
1:09:37 well not like it is now
1:09:40 White rap about irregular Turtles
1:09:44 um
1:09:47 [Music]
1:09:50 what AI art programs do you recommend
1:09:53 for beginners let's go to the video tape
1:09:56 let's do it
1:09:59 I would go
1:10:02 I would actually go to ideogram
1:10:05 it's super simple it's free for now
1:10:09 it does letters
1:10:11 it's not as sexy as the other things you
1:10:13 can go there so just go to ideogram
1:10:17 mid journey is the best but you have to
1:10:19 do it within Discord so you got to learn
1:10:21 Discord and mid Journey which is a bit
1:10:23 of a pain in the ass
1:10:24 but if you want high quality mid
1:10:26 Journeys at
1:10:28 Leonardo is really cool but but it's
1:10:30 fairly sophisticated and fairly
1:10:32 complicated if you just want simple
1:10:33 there's a little too many you know
1:10:35 switches knobs and dials for for just
1:10:37 getting started
1:10:39 um clip drop is easy because it's a
1:10:41 bunch of individual discrete tools
1:10:43 and the tools are actually quite simple
1:10:45 so that one's pretty good and that one I
1:10:47 think is free I'm not I'm not 100
1:10:49 certain
1:10:51 um
1:10:53 the other thing that you can do
1:10:55 is you can go to bing.com so let me let
1:10:58 me go to bing.com and show you
1:11:02 so you get a Bing
1:11:06 so when you get to Bing you're here
1:11:08 you're happy you're like oh yeah it's
1:11:10 sort of like Google but crappy
1:11:12 but then you click on this little chat
1:11:14 button
1:11:16 and now you have chat GPT you can make
1:11:19 money with chbetai
1:11:21 and you can put it in creative mode
1:11:23 balanced mode and precise mode and those
1:11:27 are if if you have ever played in the
1:11:30 openai playground those are the those
1:11:32 are basically temperature settings which
1:11:34 is kind of cool
1:11:35 but the one nice thing that um Bing does
1:11:38 is I can just say
1:11:44 or can I
1:11:48 make me an image of a turtle surfing
1:11:53 here are some images of turtles surfing
1:11:54 from the web
1:11:56 oh Siri she gets so jealous when I talk
1:11:59 to the other
1:12:00 models
1:12:03 oops
1:12:07 searching for turtle surfing searching
1:12:10 for turtle surfing images
1:12:13 I'll try to create that here are some
1:12:15 images of turtles surfing from the web
1:12:17 why is it talking oh because I told it
1:12:19 to talk
1:12:20 oh
1:12:22 is that new I guess it's not new I don't
1:12:24 think Bing used to talk like that all
1:12:26 right so there it made did it make four
1:12:28 yeah so there's four turtles surfing I
1:12:31 like turtles
1:12:33 foreign
1:12:39 Ty and it's image generation and it
1:12:42 looks like you can just talk to it now
1:12:44 let's try this
1:12:49 can you tell me
1:12:51 the latest
1:12:54 music activity
1:12:57 in Denver this weekend
1:13:05 searching for music activity Denver this
1:13:07 weekend
1:13:08 sure here are some music events
1:13:11 happening in Denver this weekend till
1:13:13 dawn after hours open bar featuring Dead
1:13:16 Space dark
1:13:18 I I think that talking is new I've like
1:13:21 I haven't I haven't been I haven't been
1:13:23 playing with Bing recently but I think
1:13:25 the fact that you can actually have a
1:13:27 conversation with it now Pi can do that
1:13:29 and it's connected to the internet but
1:13:31 now binkin and I bing might have been
1:13:33 able to do that for a while but that's
1:13:35 new to me that's pretty cool
1:13:39 um Mercy right replet for my own screen
1:13:42 shooting yes what air programs we just
1:13:45 did that
1:13:47 um and hello
1:13:48 today to that old-fashioned party Avec
1:13:51 plasam France wait
1:13:53 Translate
1:13:56 old-fashioned party with fish not fresh
1:14:01 took us
1:14:02 you're funny
1:14:07 good night from Europe I don't want to
1:14:09 miss the AI Revolution but I need to
1:14:11 sleep a bit good night Matthias thank
1:14:13 you thanks for hanging out for a bit
1:14:16 um
1:14:17 we do have people from all over the
1:14:19 flipping world come into this thing
1:14:20 which is kind of cool
1:14:25 what time are it 10 27 10 27 we're doing
1:14:29 okay we're doing okay people I got an
1:14:32 eight o'clock call tomorrow but it's
1:14:33 with a friendly it's with a friendly
1:14:37 [Music]
1:14:40 is there an AI that can trade crypto or
1:14:42 Forex
1:14:43 um
1:14:44 most of the Forex and crypto sites now
1:14:48 have ai Bots and and things like that
1:14:51 you can also take the code from those
1:14:54 Bots pop them into chat GPT or bard
1:14:58 uh or probably Claude and tell it to
1:15:02 analyze the bot and maybe you tell it to
1:15:04 change some things in the bot and then
1:15:06 go
1:15:06 run those so
1:15:09 um what I would say for
1:15:12 that stuff is go to futurepedia.io
1:15:16 right there
1:15:18 and go into the financial section I can
1:15:21 only imagine that there are a pile of AI
1:15:25 tools thanks AI videos appreciate it
1:15:27 what is that oh it's a little microphone
1:15:29 a little Mr microphone remember those
1:15:31 you gen xers little Mr microphone
1:15:36 I do I remember those
1:15:38 that was before they had the karaoke you
1:15:41 had to learn your songs in your brain
1:15:43 they didn't play the damn lyrics on the
1:15:46 TV
1:15:48 and by the way who else was the remote
1:15:50 control in your house
1:15:52 we didn't have remotes
1:15:54 Kyle channel four
1:15:57 dial channel seven
1:15:59 Kyle channel two
1:16:01 go to UHF 43.
1:16:04 no
1:16:09 I needed a button on the back of my head
1:16:11 click click click click go
1:16:15 click
1:16:17 do you have concern about combining AI
1:16:19 like chat CPT with Quantum processing
1:16:21 like Google's Sycamore
1:16:24 from what I understand
1:16:26 um they're very
1:16:28 very very different kinds of math that
1:16:32 are done on quantum computers and
1:16:35 um like current GPU
1:16:38 um
1:16:38 the machine learning stuff with the
1:16:40 neural networks and Transformers and all
1:16:41 that [ __ ] so I made a video I made a
1:16:45 video about Sycamore that said because
1:16:48 like the the the
1:16:50 um the benchmarks are like staggering
1:16:54 um Sycamore did in six seconds
1:16:57 what it would take today's fastest
1:17:00 supercomputer 49 years
1:17:03 to calculate
1:17:05 so the the Improvement is dramatic but
1:17:10 when I made a video that said imagine
1:17:11 what this will do for AI a bunch of math
1:17:13 nerds came in and go well it isn't dumb
1:17:16 dumb it's a very different kind of
1:17:17 mathematics and it's not really
1:17:19 appropriate to predict that the best of
1:17:21 at the math math
1:17:24 so so I don't think they mix but
1:17:27 just being a
1:17:29 common sensical kind of feller
1:17:32 if you got a tool that can do something
1:17:34 in six seconds that would take today's
1:17:36 fastest supercomputer 49 years
1:17:40 yeah they're gonna find ways to do
1:17:42 something that's going to augment you
1:17:44 know whether it's on the training side
1:17:46 or it's on the processing side
1:17:49 um
1:17:50 but I I would assume listen you know
1:17:52 quantum computers they've been adding a
1:17:55 qubit
1:17:57 what is it a cubit a month
1:17:59 and every time they add a qubit it's
1:18:01 some exponential insane [ __ ] amount
1:18:04 of power increase and Sycamore is what
1:18:07 77 or 79 cubits or 70 cubits some it's
1:18:11 getting up there so
1:18:14 um
1:18:15 I don't know I don't know
1:18:17 but but you know
1:18:20 sitting underneath your question is the
1:18:22 is the right question like
1:18:24 is there a point at which these things
1:18:27 are gonna get [ __ ] radically smarter
1:18:31 yes
1:18:34 and that's why if you look at the like
1:18:37 the the big players in this space put
1:18:40 together this group called the frontier
1:18:43 the Frontier Model Consortium or the
1:18:46 Frontier Model Group whatever it's
1:18:47 called they're basically saying if
1:18:49 you've got enough money to buy 10 000
1:18:51 gpus and build a large language model
1:18:53 you need to join the frontier group
1:18:56 and we're going to agree to make sure
1:18:59 that we got these things aligned
1:19:01 in alignment
1:19:03 and
1:19:04 if you haven't heard the term alignment
1:19:06 what that basically means is are the
1:19:08 imperatives of these things aligned
1:19:12 with the imperative of these things
1:19:16 because if they're not that's really
1:19:18 [ __ ] bad
1:19:21 like if the imperatives are aligned and
1:19:23 these things are like hey let's make
1:19:25 sure we take care of these things and
1:19:27 increase prosperity for them and make
1:19:29 sure they're still around and everyone's
1:19:31 healthy and happy and prosperous
1:19:33 if they're not aligned and they go well
1:19:36 our imperative is to get bigger better
1:19:39 and stronger and you know what's the
1:19:40 real bottleneck with us getting bigger
1:19:42 better and stronger these things
1:19:46 and if it decides these things are
1:19:49 counter to its imperatives then it will
1:19:51 find a way to make these things go
1:19:52 bye-bye that's the that's the thing that
1:19:55 the that the Sci-Fi people in the in the
1:19:58 you know the Jeffrey hintons of the
1:20:00 world and the you know the people
1:20:01 starting this Frontier group are saying
1:20:04 um if if we're going to make these
1:20:05 things dramatically powerful let's make
1:20:07 sure we got our [ __ ] together because it
1:20:09 could go south pretty quick
1:20:11 um I think they'll do it I mean there's
1:20:13 enough awareness now that that they're
1:20:15 they're starting to do it I got to tell
1:20:16 you one of the things that I think could
1:20:18 really [ __ ] up the works is meta
1:20:21 releasing their llama models
1:20:23 just out in the open they're just like
1:20:25 [ __ ] it here you go have fun
1:20:27 there's no safety guard rails on them
1:20:30 there the weights are out there so all
1:20:32 of the talk of you know these these
1:20:35 companies have to be responsible with AI
1:20:39 whatever
1:20:40 and then meta's like here want one
1:20:45 seems a little premature to me
1:20:48 but you know
1:20:49 I'm just an old fart on tick tock
1:20:54 I'm not staying long Vicky no running on
1:20:57 two hours sleep from last night
1:20:59 this all this all this stuff keeps me
1:21:01 awake I know you know what it really
1:21:03 does it's
1:21:05 I woke up one night
1:21:08 at like four in the morning with just an
1:21:11 idea like burning in my head
1:21:14 I flipped open my laptop I went to chat
1:21:16 gbt then I went to
1:21:19 um mid-journey made an interface for a
1:21:21 piece of software like like in 45
1:21:24 minutes out of a dead sleep I woke up
1:21:26 and within 45 minutes I'd mocked up a
1:21:29 piece of like an interface for software
1:21:31 and had like a whole business plan for
1:21:33 it
1:21:35 this stuff's weird it really is weird
1:21:37 that
1:21:38 like I don't know like you can just have
1:21:41 like a really small idea
1:21:43 and then once you get in the habit of
1:21:45 knowing that you can get that small idea
1:21:47 out of your head in a powerful way it is
1:21:50 kind of like crack man it's like it it's
1:21:52 it's super addictive in a
1:21:55 in a compulsive kind of way
1:21:58 which you know for ADHD or ADHD is a
1:22:02 neuro spicy people this stuff it really
1:22:04 is a drug it's it's
1:22:06 it's super compelling but
1:22:09 like I rarely get overwhelmed with input
1:22:15 I've been overwhelmed for a better part
1:22:17 of a year
1:22:20 so yeah I get that one
1:22:24 write a rap about a regular tone Turtles
1:22:27 you look like Tom Arnold uh deep machine
1:22:29 learning can it go too far yeah Destiny
1:22:33 Troy
1:22:34 um
1:22:36 like any good thing like any good things
1:22:39 in life it's all about balance
1:22:42 um sure and and
1:22:46 any bad actor right now could take the
1:22:49 meta
1:22:50 llamas llama models and the the code
1:22:53 llama model
1:22:56 and they can do some pretty bad [ __ ]
1:22:58 with that right now and I'm sure there's
1:22:59 Bad actors building that [ __ ] as we
1:23:01 speak but there's probably also white
1:23:03 hat hackers
1:23:05 building their countermeasures to that
1:23:08 [ __ ] so I don't know
1:23:12 I think that I think that the higher
1:23:15 likelihood than you know someone's gonna
1:23:17 do some wacky hacking attack with AI
1:23:21 is
1:23:24 oh there's a doggy doggy wanted in here
1:23:28 chippy
1:23:29 you want to lay down yeah hey buddy
1:23:34 he was locked out Champion was locked
1:23:36 out now he's back in
1:23:38 um hey buddy
1:23:40 let's see
1:23:43 um what's it talking about deep learning
1:23:45 going too far
1:23:47 yeah [ __ ] I lost my train of thought oh
1:23:48 it'll come back
1:23:50 hey cousin of funnel
1:23:52 you're so funny I love your show thank
1:23:55 you meta Victory Cosmo funnel
1:23:59 I think that is a reference to Rosie
1:24:01 O'Donnell cousin o'funnel listen cousin
1:24:04 o'funnel
1:24:05 you look like Rosie I with this hair
1:24:07 tonight this guy this is a Rosy vibe to
1:24:09 it doesn't it this is a little more
1:24:12 Sears Roebuck 1978 carpet salesman of
1:24:16 the year but this has got some Rosy
1:24:18 action going on it right now I
1:24:20 demonstrated the Kevin McAllister look
1:24:22 for my big boss
1:24:25 who now knows I haven't closeted AI
1:24:27 freak so that's that's actually good
1:24:29 cabrino because you know again
1:24:32 that idea of at some point and you work
1:24:36 in a big Corporation at some point in
1:24:37 those big corporations you know someone
1:24:40 somewhere is gonna
1:24:42 get it in their head that we gotta
1:24:44 figure this [ __ ] out
1:24:47 who knows anything about this and your
1:24:49 boss is gonna go I got I got the woman
1:24:51 for you
1:24:54 so that's pretty cool I support that
1:24:57 catch EPT goes to web pilot on its own
1:24:59 for answers sometimes surprise me does
1:25:02 it really even if you're not in
1:25:04 even if you're not in plug-in mode
1:25:06 is that true Ted looks at life which by
1:25:09 the way I love I love the name Ted looks
1:25:11 at life that's solid
1:25:14 um
1:25:15 which which one AI does free videos
1:25:20 none of them Runway All the video things
1:25:24 are pretty pricey
1:25:27 um you got Synthesia did
1:25:31 do like face animations I think there's
1:25:34 a third one that I can't remember right
1:25:35 now well there's probably there's
1:25:37 probably five or six but those are the
1:25:38 ones I know
1:25:40 you got Runway ml
1:25:44 I I don't see I think the video stuff
1:25:47 still is
1:25:49 It's gotta It's gotta still be really
1:25:52 um GPU intensive to do that because you
1:25:54 got to figure
1:25:55 to do a video they're doing 30 frames a
1:25:58 second so they're generating 30 frames
1:26:01 and
1:26:02 and I think they're doing it Runway does
1:26:04 it in four second chunks but then you
1:26:06 can extend it up to 18 seconds
1:26:09 so that's a lot of compute power so none
1:26:11 of them are cheap but I would just go to
1:26:13 futurepedia.io and look in the video
1:26:15 section and see see what you can find
1:26:18 there because there may be there may be
1:26:20 some like generally what happens is like
1:26:23 like a a
1:26:26 a tool like ideogram comes out and it's
1:26:29 it's initially free because they're like
1:26:32 hey we're New Kids on the Block come use
1:26:34 our tool and if the tool is awesome then
1:26:37 peop too many people come and use it and
1:26:39 their server Bill shows up the next
1:26:41 month and they're like we're gonna go
1:26:42 out of business unless we start charging
1:26:45 so so you can probably catch some video
1:26:48 Tools in those early Wonder Years stages
1:26:51 where with where they haven't gotten the
1:26:54 server bill yet that might be free but I
1:26:56 think they are they're all pretty pricey
1:26:58 right now perplexity ai's free tool has
1:27:00 essentially been replaced oh it has
1:27:02 essentially replaced search engines for
1:27:04 me I use it all day every day that's
1:27:06 interesting Kevin you know perplexity
1:27:08 for me was perplexing because
1:27:12 when I heard about it and it's called
1:27:14 perplexity.ai I went there kind of
1:27:16 thinking it was chechi BT and the first
1:27:18 thing I did it like didn't give me what
1:27:21 I expected and I and I haven't really
1:27:23 like dug into it and learned it I've
1:27:26 I've now played with it a bit to at
1:27:27 least understand kind of what it is but
1:27:29 but um
1:27:31 yeah I keep threatening to learn how to
1:27:33 use it so it's interesting you're using
1:27:35 it all day every day I I gotta I'm gonna
1:27:37 dig deeper on it I think I'm thinking
1:27:39 perplexity might need a place
1:27:42 on the list I'm thinking I'm thinking
1:27:44 Poe has to go
1:27:46 because I'm always sort of I the PO
1:27:48 feels like it's got to go
1:27:55 want to do free stock Runner
1:27:58 uh I don't know what that means
1:28:02 do you have an AI to trade Forex we did
1:28:05 that tired of scams on bitcoin AI
1:28:09 oh tired of paying for scams on bitcoin
1:28:11 AI
1:28:15 um
1:28:17 neuro spicy work tips that's funny Vicky
1:28:20 love thing
1:28:22 yeah Bing's funny but Bing seems very
1:28:24 polarizing as far as these tools go some
1:28:26 people are like Bing's awful other
1:28:28 people like I love Bing
1:28:32 I want to love Bing but they've got such
1:28:36 [ __ ] shitty safety guard rails on it
1:28:39 like when it when it
1:28:41 when it refuses to answer something for
1:28:44 me and I give it a smarmy smart ass
1:28:45 answer and then it goes well you're just
1:28:48 too angry we're gonna have to just stop
1:28:50 this conversation right now little
1:28:52 mister I'm like I don't need to be
1:28:54 [ __ ] [ __ ] slapped by a search engine
1:28:57 it's like shut that [ __ ] off
1:29:01 if I'm gonna cuss at a search engine I'm
1:29:03 gonna cuss at a search engine
1:29:11 foreign
1:29:16 what are you gonna do what are you gonna
1:29:18 do Google that [ __ ]
1:29:21 you are amazing thank you Byron
1:29:24 Brad was red seam creative not red line
1:29:27 oh thank you Amelia's wife I know that
1:29:30 I've been I I've talked to Brad like
1:29:32 four or five times today I was actually
1:29:34 giving him input on his brand but you
1:29:36 know add Red Line red seam you know
1:29:41 so red seem creative thank you that boy
1:29:43 I am I'm way behind on comments so
1:29:45 here's what I'm gonna do there's 80
1:29:46 folks in here they gotta all be pissed
1:29:48 off he's not answering any of my
1:29:50 questions I'm very upset about this
1:29:53 I've come in here I've been in here for
1:29:54 at least 10 minutes I've asked my
1:29:56 question 17 to 18 times 19 times
1:30:00 20 times and he hasn't answered it I
1:30:04 don't understand what's happening I'm
1:30:06 just as important as those other people
1:30:09 all right so what I'm going to do I'm
1:30:10 going to jump to the bottom
1:30:13 oh there's so many questions all right
1:30:16 and I'm just gonna scroll up a little
1:30:18 bit and so if I missed your question
1:30:20 sorry but all right so now I'm at uh Ted
1:30:25 looks at life in news archive market
1:30:27 research biz for 30 years
1:30:29 even no bullying search this is Light
1:30:32 Years leap yeah exactly
1:30:35 um
1:30:37 uh remember Boolean search any of you
1:30:40 remember bullying search till Google got
1:30:42 their [ __ ] together and figured out how
1:30:43 to do some version of natural language
1:30:46 parentheses plus I mean minus he
1:30:52 it's like a math equation although not
1:30:54 for nothing but prompting chat GPT and
1:30:57 all these large language models is is
1:30:59 like programming in English
1:31:01 it's like English is the programming
1:31:03 language because you're still thinking
1:31:05 in logical constructs
1:31:08 so yeah it's crazy
1:31:10 um I remember when I was the family
1:31:11 remote
1:31:13 Thelonious channel four exactly
1:31:17 best AI I saw this week was Drake trying
1:31:20 to diss Eminem and Eve's response was
1:31:23 epic don't mess with Gen X nope Gen X
1:31:27 don't Gen X is like there's a couple of
1:31:29 things we were already bitter and smarmy
1:31:33 and sassy and now we're tired
1:31:36 but we're still ambitious
1:31:38 and we still have got our faculties
1:31:40 about us and we can use technology
1:31:43 so just don't [ __ ] with us because we're
1:31:45 kind of done we're done with it all
1:31:47 we're done with people being [ __ ] to
1:31:49 one another we're done with all the
1:31:51 polarization we're done with all the
1:31:54 [ __ ] snowflakes we're done with on
1:31:56 both sides
1:31:57 we just done
1:31:59 so let us be calm down
1:32:02 take a chill go to Chachi PT have it
1:32:05 write you a song and if you're not gonna
1:32:07 do that just get the [ __ ] out
1:32:11 it's not quite that bad you can be here
1:32:13 if you don't use chat GPT
1:32:16 come in here troll me tell me I look
1:32:19 like Rosie O'Donnell
1:32:21 ask for my autograph
1:32:26 oh I'll give it to you I'll do it
1:32:29 all right
1:32:36 if you could commit one crime that no
1:32:39 one would ever be able to commit
1:32:41 again what crime would you do
1:32:46 a crime that no one would be able to do
1:32:48 again oh that's a really interesting
1:32:50 question it's got nothing to do with
1:32:52 [ __ ] AI
1:32:54 it's a good
1:32:55 philosophical brain teaser isn't it
1:33:09 you know a weird one that I think I
1:33:12 would probably do
1:33:17 like like
1:33:19 fraud Computer Fraud because I think the
1:33:22 Computer Fraud with his AI stuff is
1:33:25 gonna get
1:33:26 really sophisticated and really bad
1:33:30 so if I could scam one person and then
1:33:33 no one could ever scam again
1:33:36 I just picked someone who was douchey
1:33:38 I've got like two or three people that I
1:33:41 wouldn't mind [ __ ]
1:33:44 doing that with
1:33:46 like that's one that feels like it just
1:33:48 feels like it's going to be this ongoing
1:33:49 rolling disaster
1:33:51 of just
1:33:53 nice little little ladies that well well
1:33:57 my daughter called me and she said she
1:33:59 was in trouble so I wired her the money
1:34:02 like I just feel like that's gonna be
1:34:04 like
1:34:05 [ __ ] news story after news story of
1:34:08 that kind of [ __ ]
1:34:10 and like our scammer the scammer that
1:34:12 was in here that was scamming me I mean
1:34:14 he was reaching out to people to scam
1:34:16 them which sucked but
1:34:18 yeah
1:34:20 yeah that I think it would be that it's
1:34:22 not a it's not a sexy crime but it's
1:34:24 like it's really ubiquitous I think it's
1:34:26 gonna get worse
1:34:27 two guys mind extension tools these
1:34:30 things are absolutely these are mind
1:34:32 amplifiers ability amplifiers is is how
1:34:36 I think of these things
1:34:38 a lot of things happen from Deep Sleep
1:34:40 double helix for example yeah
1:34:42 interesting how many programming
1:34:44 languages do you know not many I know I
1:34:47 know
1:34:48 I knew I knew basic back in the day
1:34:51 that's what I learned on and then I knew
1:34:54 HTML and then
1:34:56 I can look at CSS
1:34:58 um we created our oh no we didn't create
1:35:01 our own XML we use XML for the video
1:35:03 assembly for storyvine so XML isn't
1:35:06 really a programming language it's more
1:35:08 of a scripting framework
1:35:11 um
1:35:12 I have written a python application but
1:35:15 well no I directed chat GPT to write me
1:35:18 a python application
1:35:21 um I don't know I'm not I'm not really a
1:35:23 coder I'm more of a product guy business
1:35:24 guy
1:35:27 um
1:35:29 I'm more interested in what you do with
1:35:31 the tools than
1:35:33 tactically building them and and you
1:35:36 know me holding out because of my ADD I
1:35:39 just I can't be an engineer I can't
1:35:40 program I don't have the attention span
1:35:42 for it or the focus for it I can hyper
1:35:44 Focus like if I wanted to take the
1:35:46 weekend and learn a programming language
1:35:48 I could do that but I've resisted it
1:35:50 because I had a feeling [ __ ] like this
1:35:52 was coming
1:35:54 or maybe I was in denial but now it's
1:35:56 here and now you don't have to [ __ ]
1:35:58 know programming languages so I know
1:36:00 enough about
1:36:02 how systems work programs where
1:36:04 databases work
1:36:08 data hooks work all that sort of [ __ ]
1:36:10 that I'm fine
1:36:12 I can either get one of my developers to
1:36:14 do it or I can have this do it so
1:36:17 I don't know much my focus vava blue
1:36:21 which is a fun name
1:36:23 is kind of November 30th 2022 forward
1:36:27 right so the launch of chat GPT for me
1:36:30 is a seminal event because that marks
1:36:32 the day
1:36:34 where before that day it's all Engineers
1:36:37 researchers scientists programmers
1:36:39 mathematicians after that day it's the
1:36:42 rest of us
1:36:44 and what I'm interested in is now that
1:36:45 these tools are here even though they're
1:36:47 in this janky ass state that they're in
1:36:50 or as I like to call them janky pieces
1:36:52 of [ __ ] even though their janky pieces
1:36:54 of [ __ ] oh thanks for the doggo Stevo
1:36:58 um
1:37:00 we're now all trying to figure out what
1:37:02 does this mean for work what does this
1:37:04 mean for project Managers from data
1:37:06 analysts for creatives for copywriters
1:37:08 for
1:37:09 Consultants
1:37:12 it's gonna [ __ ] up a lot
1:37:16 and it's not just gonna [ __ ] it up for
1:37:18 programmers right
1:37:20 it is going to [ __ ] it up for
1:37:21 programmers but it's gonna [ __ ] it up
1:37:23 for everyone
1:37:24 if I were a consultant right now
1:37:26 if I were a consulting company
1:37:29 full of data analysts and Consultants
1:37:31 that rely on those data analysts
1:37:36 I would you know it's funny I would not
1:37:39 only be nervous I would be
1:37:41 actively trying to reinvent my business
1:37:44 because your business model is about to
1:37:47 get flushed down the toilet
1:37:50 and and
1:37:52 I I
1:37:55 I don't sense
1:37:58 a grand panic and a lot of the people
1:38:00 that I talk to
1:38:03 um but I I think it'll it'll come
1:38:07 that's that's why I'm doing this like I
1:38:09 own as many people as possible to go
1:38:12 play with these things share bear thank
1:38:15 you
1:38:17 play with these things have your Kevin
1:38:19 McAllister moment freak out
1:38:22 and then start to internalize the tools
1:38:24 and go okay what can I really do with
1:38:26 this how can I transform my business how
1:38:28 can I transform my role how can I
1:38:30 transform who I am what I do how I
1:38:32 contribute to the world
1:38:37 so I don't know that got a little
1:38:38 philosophical I don't like it when I get
1:38:41 all philosophical and serious like that
1:38:43 I don't know why you people are doing
1:38:45 asking me serious questions
1:38:47 why would you do that
1:38:49 can't have fun when I'm up on a soapbox
1:38:51 dancing around twerking it up
1:38:55 futureize them
1:39:00 the information overload is real jrc
1:39:03 honest to God man I just it's it's so
1:39:06 exhausting
1:39:08 it's like it's both exciting and
1:39:11 exhausting but it's just like the
1:39:14 the AI Salon this last Tuesday wait
1:39:17 yesterday oh God
1:39:20 the AI Salon yesterday like right before
1:39:22 it I talked to Leah my co-host
1:39:25 and she was like I'm kind of done with
1:39:28 AI and I was like I'm just exhausted
1:39:31 and then
1:39:32 what the salon was was we had five
1:39:34 members
1:39:36 each talk about a project they were
1:39:37 working on and
1:39:39 I was assuming people would just come in
1:39:41 and like do what I do here like just
1:39:43 sort of wax poetic for 10 minutes and be
1:39:46 done
1:39:47 and they all prepared PowerPoint
1:39:49 presentations they had their [ __ ]
1:39:52 together
1:39:54 um
1:39:55 and and like I left that meeting like
1:39:57 really inspired so like I'm like
1:39:59 oscillating between I'm just [ __ ]
1:40:01 exhausted I can't keep up and oh my God
1:40:03 this is so exciting I gotta lean in more
1:40:05 and just she's like back and forth it's
1:40:08 just like a it's like you know like a
1:40:11 bad tennis match or a good tennis match
1:40:18 Champy's here yes llama you were talking
1:40:22 about
1:40:22 llama you were talking about go on
1:40:26 maybe this would help us use more than
1:40:28 10 of our brains you know cavruno yes
1:40:33 so and and
1:40:35 it could be the reason that
1:40:38 we're all feeling so exhausted
1:40:46 I think so so these tools are doing two
1:40:49 things for me on that front
1:40:54 one is
1:40:57 because it's because there's so much
1:40:59 less friction in getting ideas out of
1:41:01 your head and into some sort of tangible
1:41:03 format
1:41:05 I'm having more idea like I'm putting
1:41:07 more ideas in the world
1:41:09 and that's kind of exhausting because
1:41:11 each like idea that you put out in the
1:41:13 world is some version of your baby and
1:41:14 then you like you know have to deal with
1:41:17 them and care for them or you know let
1:41:19 them drop off and then you're like oh
1:41:20 [ __ ] I had those ideas and that's
1:41:22 exhausting
1:41:24 and the other thing that I think is
1:41:26 happening is and this would actually be
1:41:29 a really interesting thing for for some
1:41:31 Neuroscience people to study
1:41:33 is
1:41:38 if these things do
1:41:41 the low level problem solving like
1:41:44 writing the first draft of a document
1:41:47 takes a certain amount of your brain
1:41:49 energy and a certain amount of your
1:41:50 Creative Energy to block that out and it
1:41:53 might take you an hour to write a first
1:41:54 draft
1:41:55 and now with these tools
1:41:58 that first draft happens in 20 seconds
1:42:03 and then you spend that hour that you
1:42:05 would have been writing the first draft
1:42:06 refining
1:42:08 a thing that was generated
1:42:11 and I think it uses a different part of
1:42:13 your brain so so there it might be that
1:42:17 like
1:42:18 historically I would have spent 10 hours
1:42:20 in a week doing kind of blocking and
1:42:22 tackling creative
1:42:24 and 10 hours a week doing refinement
1:42:27 and now I'm doing
1:42:30 three hours of work doing blocking and
1:42:32 tackling creative and 17 hours doing
1:42:35 refining and synthesizing and so I don't
1:42:40 know there might be
1:42:42 you know these things are going to
1:42:44 change our brains I think
1:42:47 and there's also there's a third one
1:42:49 which is like multimodal switching
1:42:52 like if I were writing historically I
1:42:55 would never think about
1:42:57 oh I gotta go make an image for that but
1:43:00 now because you can switch so quickly
1:43:02 between these things you're like oh
1:43:05 here's a cool thing oh give me some
1:43:07 headlines oh those are cool headlines oh
1:43:08 this should have a logo let me run over
1:43:10 and pop out a logo real quick and then
1:43:12 you're over in logo mode and then you
1:43:14 bring it back and now you're in design
1:43:16 mode then you're in typography mode then
1:43:18 you're back to writing mode like I think
1:43:20 this sort of shape-shifting creatively
1:43:24 has got to be just taxing our brain in a
1:43:26 very different way right right yeah
1:43:28 exactly
1:43:31 so
1:43:32 yeah no wonder we're [ __ ] tired
1:43:35 exactly I need I need a little blue
1:43:38 board I need a little blue board I need
1:43:41 to take a blank break with a little blue
1:43:43 board I'm gonna go out in the yard and
1:43:45 cry now
1:43:47 yeah
1:43:48 yeah it's crazy
1:43:51 he likes it hey Mikey
1:43:53 those are the best commercials and
1:43:56 remember Mikey what was the rumor that
1:43:58 he drank
1:43:59 um Pepsi and pop rocks and his stomach
1:44:02 exploded and it killed him was it wasn't
1:44:04 wasn't that the rumor
1:44:06 at least in my in my high school or
1:44:08 wherever it was
1:44:10 [Laughter]
1:44:15 um so are prompts similar for
1:44:16 mid-journey and ideogram yeah they're
1:44:18 very similar I don't think ideogram does
1:44:20 any of the dash dash modifiers like
1:44:23 aspect ratio and things like that
1:44:25 um ideogram's much more simple but
1:44:28 things like
1:44:29 um you know I want a painting or a
1:44:31 photograph here's the subject here's the
1:44:33 lighting conditions or Styles here's the
1:44:36 medium
1:44:37 ideogram also has like you can click on
1:44:40 the boxes so photo illustration 3D
1:44:43 render those sort of things those are
1:44:46 basically the mid-journey modifiers or a
1:44:48 handful of them that you can just click
1:44:50 on instead of typing
1:44:52 stable diffusion question mark I don't
1:44:55 know what you mean Molly oh oh when you
1:44:57 were asking about
1:44:59 um image tools to play with if you're
1:45:01 geeky stable diffusion is great
1:45:04 um the developer communities all over it
1:45:07 if you can run a Google collab notebook
1:45:09 and get automatic 11 11 up and running
1:45:12 stable diffusion is awesome there's all
1:45:14 sorts of different models out there
1:45:15 there's all there's Laura's there's just
1:45:18 there's a ton for it
1:45:19 if you just want to get pictures done
1:45:22 it's a complicated [ __ ] nightmare
1:45:26 um especially if you're not technical if
1:45:27 you're technical go for it it's my
1:45:30 challenge with stable diffusion like it
1:45:32 took me nine times I failed nine times
1:45:36 before I got automatic 11 11 up and
1:45:38 running and it for me it was just
1:45:41 understanding Google collab and like the
1:45:43 whole framework and how it's connected
1:45:44 to Google Drive and it was just a lot
1:45:48 so I finally cracked that but then my
1:45:50 problem was every time I would log back
1:45:53 into automatic 1111 The Notebook that I
1:45:56 was using had been updated and it would
1:45:57 break something and then I would have to
1:45:59 troubleshoot it and go fix it and then
1:46:01 like 16 new features came out and you
1:46:03 had to turn on extensions Within
1:46:05 automatic 11 11 and it just it was just
1:46:08 like I'm I don't I don't I'm [ __ ]
1:46:11 exhausted back to cabrino's point it's
1:46:13 like I've only got so many hours of
1:46:15 Creative Energy in my brain and if I
1:46:17 spend them debugging a [ __ ] Google
1:46:19 collab notebook that I don't really
1:46:21 understand in the first place
1:46:24 um that's like two hours that I could
1:46:26 have been making images so so but but
1:46:29 it's great if you want to geek out on
1:46:31 stuff and have a lot more control
1:46:35 um how about the stock Bitcoin you keep
1:46:38 asking about that I don't know what that
1:46:39 means
1:46:41 um
1:46:42 bitcoin's not a stock it's a crypto
1:46:45 should we truly worry about the
1:46:46 singularity
1:46:48 if an an attempt to evolve any AI First
1:46:52 Step would be
1:46:55 to strip any human emotion
1:46:57 or any human traits
1:47:00 I don't why would that be the first step
1:47:07 I I mean I I find the opposite
1:47:10 is happening like in theory I think
1:47:13 that's rightly that
1:47:15 like like if you listen to the Sci-Fi
1:47:17 movies the machines are gonna
1:47:20 they're gonna strip human emotion the
1:47:22 way these things are being programmed if
1:47:23 you go play with pi if you go play with
1:47:26 this
1:47:27 that is not programmed to strip away
1:47:29 human emotion at all
1:47:31 and Chachi PT quite frankly is more
1:47:35 polite and more it displays more empathy
1:47:39 than most humans I know
1:47:42 it shows nuance
1:47:45 it doesn't get worked up so I I don't
1:47:50 know that I mean my experience hasn't
1:47:52 been what you're talking about there
1:47:55 but you know if you're talking about the
1:47:58 singularity and these things getting you
1:48:00 know exponentially more powerful which
1:48:02 they are
1:48:03 um
1:48:04 I don't know that's why the frontier
1:48:07 working group was put together to make
1:48:09 sure that they're working on alignment
1:48:10 as much as they're working on improving
1:48:12 the capabilities
1:48:18 um yes you do I don't know what that's
1:48:19 in reference to but I'm sure I do turtle
1:48:21 Wrangler Turtles totals uh
1:48:26 um Tick Tock just jumped me to the
1:48:28 bottom again I'm gonna go up a little
1:48:29 bit and then I'm just going to pick a
1:48:30 question I'm not going to answer all the
1:48:32 questions I'm just not going to do it
1:48:33 because Tick Tock has got a dumb dumb
1:48:36 interface
1:48:37 dum dum tick tock
1:48:39 what was the Malto meal Maynard tagline
1:48:46 um
1:48:47 I don't remember
1:48:51 I remember Carnation Instant Breakfast
1:48:53 but that's not Malta meal
1:48:56 not the bottom
1:48:58 welcome to AI Boomer talk love
1:49:02 this is Elder talk how can I help you
1:49:05 today we're going to talk about
1:49:07 computers
1:49:10 the little round thing sitting next to
1:49:12 your computer is what we call a mouse
1:49:16 when we say click on the mouse it means
1:49:19 press down on the top of the mouse
1:49:23 I'm a librarian still use bullying
1:49:25 search
1:49:27 more power to you bully I mean what was
1:49:29 nice about bullying search is like you
1:49:32 knew exactly what you were getting and
1:49:33 it you know it's funny
1:49:36 the problem with natural language search
1:49:39 before these things before the large
1:49:41 language models
1:49:46 if you if you use Boolean search in
1:49:50 search engines
1:49:54 okay I'll just speak for myself because
1:49:56 I use Boolean search and search engines
1:49:58 when Google said oh it's okay you can
1:50:01 use natural language now we'll figure it
1:50:03 out I never trust Google to know what
1:50:06 the [ __ ] I'm talking about so I end up
1:50:09 sticking Boolean operators within my
1:50:12 natural language searches because I
1:50:14 don't trust Google
1:50:16 and for the most part I'm right but
1:50:18 these things now actually do natural
1:50:20 language which is just insane
1:50:23 when will the Apple tool be released
1:50:28 I really hope [ __ ] clawed that it's in
1:50:32 um
1:50:33 in September when they do their iPhone
1:50:36 15 announcement
1:50:40 I'm hoping that there's a one more thing
1:50:42 moment where they say oh one more thing
1:50:45 and here's our AI tools our AI agent or
1:50:48 whatever it is
1:50:50 um
1:50:57 apple is willing not to be first right
1:50:59 they Apple doesn't want to come out
1:51:05 here's what Apple's not gonna do
1:51:07 Apple's not gonna launch a tool
1:51:11 that's at parody with chat GPT
1:51:15 Apple's gonna want to launch a tool that
1:51:18 redefines the market like like division
1:51:20 Pro
1:51:22 for me was a significant announcement
1:51:24 because
1:51:26 I didn't think it would be possible
1:51:29 to take us out of ski goggles that look
1:51:31 exactly like every other [ __ ] AR and
1:51:34 VR headset
1:51:36 and successfully separate that from the
1:51:39 crowd
1:51:39 and they did like in in my opinion they
1:51:43 they made an embarrassment out of all
1:51:46 the other goggles right they didn't use
1:51:49 the word metaverse once
1:51:52 they didn't use the word the phrase AI
1:51:54 once but they made it very very clear
1:51:57 there's significant machine learning
1:51:59 horsepower right in front of your eyes
1:52:02 you know the R1 chip is sitting right
1:52:04 there doing real-time analysis of those
1:52:07 16 [ __ ] cameras and sensors all over
1:52:09 your head
1:52:11 um that's all AI right so it's clear
1:52:13 it's deep AI they successfully
1:52:17 created a new category spatial computing
1:52:22 um
1:52:23 you know they did just weird amazing
1:52:26 [ __ ] like put displays on the outside of
1:52:29 your eyes so that they could project the
1:52:33 images of your eyes forward so people
1:52:36 could see that you're there
1:52:39 like the minute you see it you're like
1:52:41 why isn't everyone else done that well
1:52:43 because that's the kind of [ __ ] Apple
1:52:44 does right
1:52:46 that was nine years of development
1:52:49 in that
1:52:51 thing
1:52:53 so they're brilliant at that so I think
1:52:55 they're going to do the same with AI so
1:52:59 they're not going to do what Google did
1:53:00 which you know when Chachi BT came out
1:53:04 and everyone lost their [ __ ] and then
1:53:06 Google declared their Red Alert
1:53:09 like a month later they had this big
1:53:12 hootenanny in Paris
1:53:15 and they announced we do AI too we
1:53:18 invented this [ __ ] and then
1:53:20 it was an embarrassment the tech didn't
1:53:23 work the one woman got up to demo
1:53:25 something on a mobile phone and they
1:53:28 didn't have the mobile phone there and
1:53:29 she went where's the phone and people
1:53:31 were like we don't know just skip it and
1:53:33 she goes oh I okay I guess we're not
1:53:35 gonna do the demo like they put stuff
1:53:38 out there that wasn't ready to be put
1:53:40 out there Apple won't do that
1:53:43 so I'm hoping that they announced
1:53:45 something in September but I wouldn't be
1:53:47 surprised at all if they wait until the
1:53:49 developer conference next June so June
1:53:52 of 24.
1:53:55 because Apple would rather be late and
1:53:59 and just of another level
1:54:01 then
1:54:03 play catch-up
1:54:05 and be a parody so I don't know I don't
1:54:08 know
1:54:09 I'm excited about it I mean the one
1:54:11 thing I heard is that they spent their
1:54:13 their research budget at Apple is 24.6
1:54:16 billion dollars which is bigger than the
1:54:19 market cap of chat GPT of open AI
1:54:24 um and and Tim Cook said they spent the
1:54:26 bulk of that money on generative AI in
1:54:28 the past year
1:54:30 24.6 billion dollars
1:54:33 and we haven't seen [ __ ] from them
1:54:37 so they're cooking up they're cooking up
1:54:39 they're cooking up some chitlins and
1:54:42 some gravy you're gonna be tasty down in
1:54:45 a bayou they got some crawdaddies and
1:54:48 put it in there
1:54:50 cook it up
1:54:52 and they're gonna come out and they're
1:54:53 gonna be like would you like some tasty
1:54:55 gumbo and everyone's gonna be like what
1:54:58 the [ __ ] is gumbo
1:55:00 oh my God this is good
1:55:05 I hope
1:55:06 although I do gotta say that their voice
1:55:08 synthesis technology that's in iOS 17 is
1:55:12 [ __ ] horrible
1:55:13 it's horrible
1:55:17 wouldn't have thought it but it is
1:55:20 AI will run your washing machine with no
1:55:22 soap you on one Kyle
1:55:25 I don't know what that means but you
1:55:27 were laughing what even is chappy GT
1:55:30 exactly what is it what even is internet
1:55:37 uh
1:55:39 uh you actually do look a little bit
1:55:41 like Rosie thank you thank you it's
1:55:45 those are my favorite trolls I know it
1:55:47 wasn't intended as a troll
1:55:49 it's a very pleasant compliment right
1:55:57 listen at least I'm still here you
1:55:59 forgot Gen X being snarky oh yes snarky
1:56:02 smart ass bitter you betcha we're Gen X
1:56:05 get off my lawn
1:56:08 and you millennials
1:56:13 Gen X zero F's left well whatever never
1:56:17 mind yeah whatevs uh whatever yeah Molly
1:56:21 would you ever
1:56:22 do debates on Tick Tock I see a lot of
1:56:25 live stream debates about prove God
1:56:27 exists well I wouldn't do it about
1:56:28 proving God exists I do debates I do
1:56:31 debates about education yeah about this
1:56:33 [ __ ] I'd be happy to debate
1:56:37 yeah that would be fun
1:56:39 I don't know it you know
1:56:42 I think
1:56:43 I don't know the problem with debates is
1:56:45 they get a little too serious
1:56:49 but at the same time
1:56:52 I I just
1:56:54 I
1:56:59 so much is changing so fast and these
1:57:02 things are so profound
1:57:04 that I think anything we can do to talk
1:57:07 about this [ __ ] is valuable right now so
1:57:09 if it's a debate if it's my stupid
1:57:11 [ __ ] Tick Tock live idiotness if it's
1:57:15 the AI Salon if it's the videos I make
1:57:18 if it's Rachel Woods doing the AI
1:57:20 exchange I think all of it's valuable
1:57:23 like like MacKenzie Bose with GPT boss
1:57:26 like he's all over the place now like
1:57:28 like GPT boss didn't take off like he
1:57:30 wanted it to so he's like all over the
1:57:32 place trying to figure out new
1:57:33 businesses and new models and services
1:57:35 and and it's like I think that's
1:57:38 brilliant it's like
1:57:40 the transparency there of watching him
1:57:42 he's a guy that's so capable and he
1:57:45 built this whole tool and and you know
1:57:47 the interface and he took all this input
1:57:49 and and it was like this is going to
1:57:51 full-on be his business and then it
1:57:53 didn't have the metrics that he wanted
1:57:54 and he's like [ __ ] it it's like I'm
1:57:56 gonna do other [ __ ]
1:57:58 and it's like I think that's I think
1:58:00 that's where we all need to be for the
1:58:03 next three years is just
1:58:06 you can't not participate in my opinion
1:58:09 I think I think not participating in
1:58:12 this stuff is career and professional
1:58:15 and personal
1:58:18 I was gonna say
1:58:20 whatever I was it's not good
1:58:24 I'm so sick of the word police
1:58:27 um
1:58:28 I'm so [ __ ] sick of the word police
1:58:31 um
1:58:36 I I think you got to be in it and at the
1:58:39 same time
1:58:41 this stuff's changing so fast that it's
1:58:43 almost like it's irrelevant what you
1:58:45 work on because anything that you work
1:58:46 on and build and are proud of today is
1:58:49 going to be just obsolete three months
1:58:52 from now or three days from now
1:58:53 someone's gonna come out with a feature
1:58:56 that obliterates entire businesses that
1:58:58 were created over the past year and I
1:59:00 like that's like I just feel like that's
1:59:02 the next three years
1:59:05 and if you're at that place like I am
1:59:07 right now and it sounds like a lot of
1:59:08 people are like this is [ __ ]
1:59:10 exhausting it's like yeah I I think it's
1:59:13 gonna be exhausting and I and and I
1:59:15 still think you got to be in the game
1:59:18 and and you know the game is whatever it
1:59:20 is for you like it doesn't have to be
1:59:21 all in like me like I'm you know cracked
1:59:24 out on this stuff right now but
1:59:28 I think
1:59:32 yeah I don't know we gotta play gotta
1:59:35 play to win
1:59:36 ocean 11's AI scam versus Casino
1:59:39 so what did I do with my AI learning lab
1:59:41 under t-shirts
1:59:44 you gotta do you gotta do something
1:59:47 where the penalty will be half of what
1:59:49 you made uh where
1:59:52 now you're changing the rules I can't
1:59:54 keep up with you
1:59:56 all right what can they do in Hollywood
1:59:59 to stop the AI that Netflix and others
2:00:01 are doing so James
2:00:04 they they can't they
2:00:08 here's the deal
2:00:14 everything
2:00:16 every profession
2:00:18 that is computer mediated
2:00:21 every profession
2:00:24 that is computer mediated in any way
2:00:28 anyone who uses a computer in any way
2:00:32 all of those jobs are going to be
2:00:33 transformed
2:00:36 writing filmmaking
2:00:40 TV video film uh social media
2:00:45 knowledge work
2:00:46 Consulting
2:00:48 project manage all of it all of it
2:00:51 all of it is going to be reinvented
2:00:54 this [ __ ] isn't getting uninvented
2:00:58 [Applause]
2:01:02 what
2:01:03 the writers and sag
2:01:10 I mean this this is my hot take as as an
2:01:13 as a trained actor and a screenwriter
2:01:16 but I'm not in the wga and I'm not I'm
2:01:18 not the Screen Actors Guild
2:01:24 I think the most sensible path forward
2:01:26 is to say let's assume
2:01:30 that all this [ __ ] is going to take over
2:01:32 the industry let's assume that every
2:01:34 actor is going to be synthesized let's
2:01:36 assume that yeah studios are going to
2:01:39 use these tools to do the writing and
2:01:41 they're going to use tools like the the
2:01:43 advanced version of runwayml to do the
2:01:44 filmmaking let's assume all of that
2:01:47 happens
2:01:48 and make sure we get contracts in place
2:01:51 so that storytelling professionals still
2:01:55 participate in the game
2:01:58 but that's not what they're doing
2:02:00 what they're doing is they're saying
2:02:03 [ __ ] AI ai's evil we should never ever
2:02:06 use AI
2:02:09 it's exactly what
2:02:10 photographer said film photographer said
2:02:13 when digital cameras came out every
2:02:15 single new technology the existing
2:02:20 professional class that is going to be
2:02:22 disrupted tries to resist it
2:02:25 every single time in all of history
2:02:31 here let me show you something
2:02:42 from today painting is dead
2:02:46 that was French French painter Paul de
2:02:49 la rouge upon seeing the first
2:02:50 daguerreotype so the first photograph
2:02:52 came out
2:02:54 and this guy said from today painting is
2:02:56 dead painting obviously hasn't died
2:02:58 filmmaking nuts is not going to die in
2:03:00 the future but it's going to be
2:03:02 different
2:03:04 so
2:03:08 so we like
2:03:11 it's an appropriate metaphor we've seen
2:03:14 this movie before we know how this ends
2:03:16 ai's not going to get uninvented
2:03:19 companies like Netflix or it might not
2:03:21 even be Netflix see Netflix and and the
2:03:25 streaming services may say okay we're
2:03:27 gonna we're gonna agree with sag and
2:03:29 what the writer's Guild that we're never
2:03:30 gonna use this stuff and we're gonna
2:03:32 we're gonna keep the old way of doing
2:03:33 things
2:03:34 and if they actually commit to that and
2:03:36 stick to that someone else is going to
2:03:38 come along that doesn't know the history
2:03:40 and Legacy of that it's going to be some
2:03:42 20-something that's sitting in his
2:03:44 mother's basement that whips up some AI
2:03:48 streaming service that's generating
2:03:49 Dynamic content that people all of a
2:03:52 sudden love and start using and they're
2:03:54 going to cut Netflix and all of these
2:03:56 contracts out of the loop
2:03:59 because they're fighting a
2:04:03 an unfightable battle it it's literally
2:04:06 like the town of Tampa
2:04:09 three days ago going we're not going to
2:04:12 accept that hurricane
2:04:13 nope nope we're striking we're striking
2:04:17 we want to in our contract that uh we no
2:04:21 longer accept hurricanes and hurricanes
2:04:23 are bad and we're not going to accept
2:04:25 them like that's what I feel like it's
2:04:27 like right now
2:04:28 so
2:04:30 I don't have a good answer for it but I
2:04:33 I just think that the AI is evil [ __ ] AI
2:04:36 we're never going to use AI I just think
2:04:38 it's incredibly short-sighted it's
2:04:40 I'm a writer I've written seven
2:04:42 screenplays I've written a one-man show
2:04:46 these
2:04:47 are the most remarkable writing tools in
2:04:50 the history of humanity and the writer's
2:04:53 Guild the group of professional writers
2:04:55 is saying nope we don't want to use the
2:04:58 most professional writing tool ever
2:05:00 written
2:05:01 I don't it doesn't make sense to me now
2:05:04 what does make sense to me is holy [ __ ]
2:05:06 these things right good enough that if
2:05:08 they get much better
2:05:10 what we currently do what we currently
2:05:13 call writing is at risk that I
2:05:15 absolutely agree with but you're not
2:05:17 going to uninvent it
2:05:20 not gonna uninvent it
2:05:23 is it sad is it tragic is it all [ __ ]
2:05:25 up yeah
2:05:30 but
2:05:31 you know they're going to get blindsided
2:05:33 five years from now by some company that
2:05:36 doesn't think like they think because
2:05:37 it's going to be people that aren't from
2:05:39 that world it's literally from another
2:05:41 world
2:05:44 like think about you know
2:05:46 the kids that are in
2:05:48 ninth grade right now
2:05:52 they make it through and especially the
2:05:55 kids in ninth grade right now in schools
2:05:57 that are embracing AI
2:06:00 so they've got three four years of high
2:06:04 school
2:06:06 and then they got maybe four years of
2:06:08 college but that might not even be a
2:06:10 requirement there they got three four
2:06:12 years of high school if you think about
2:06:13 regenerative AI was a year ago and where
2:06:16 we are today and then you go out three
2:06:18 or four years
2:06:19 these tools are going to be [ __ ]
2:06:21 remarkable
2:06:22 so you're gonna have people coming out
2:06:24 of high school
2:06:25 that have
2:06:28 quasi-native AI skills
2:06:32 that don't know the history of anything
2:06:35 they don't know they don't know the
2:06:36 world all they know is they got these
2:06:38 tools oh cool I can make a movie oh yeah
2:06:40 this is you know Grandpa watches Netflix
2:06:43 all the time this is kind of like
2:06:44 Netflix but we're making original
2:06:46 content for everyone everyone gets their
2:06:47 own personal movie that's kind of cool
2:06:49 Click Boom launch
2:06:51 billion dollar valuation Netflix is
2:06:55 scrambling to keep up and that becomes
2:06:57 the new Netflix like that's gonna happen
2:07:02 so how does the writer's Guild account
2:07:04 for that they can't
2:07:06 and neither can sag
2:07:09 you know what they could account for is
2:07:11 okay assuming that that's the future
2:07:14 what kinds of contracts do we need to
2:07:16 put in place now
2:07:18 that allow us to participate in that
2:07:20 future
2:07:26 it's a weird one it's a really weird one
2:07:28 but
2:07:29 yeah they do be janky but a lovable
2:07:31 Yankee yeah had an argument with my high
2:07:34 school soccer no social studies teachers
2:07:37 Hub hubby teach kids to use AI
2:07:41 critically or forbid it
2:07:43 uh
2:07:46 you know what channel you're in [ __ ]
2:07:48 teach them to use it critically critical
2:07:50 thinking so I'm a librarian
2:07:53 you know this critical thinking is the
2:07:56 skill
2:07:57 that's required for these tools
2:08:01 forbidding it is just like the whole
2:08:03 just rant I just went on about sagging
2:08:05 and the wga right now the education
2:08:08 system is worse
2:08:10 forbid it
2:08:12 forbid a thing that is going to be in
2:08:15 everything
2:08:16 okay
2:08:18 aren't aren't you there to prepare
2:08:20 students for the world they're about to
2:08:23 enter
2:08:24 how can you do that if you don't
2:08:26 understand what are the tools that are
2:08:28 going to transform that world
2:08:33 this is dumb it's just dumb keep getting
2:08:36 in those arguments I fully support you
2:08:38 being in those arguments do we need
2:08:40 project management I think we're going
2:08:42 to need
2:08:43 I think project management
2:08:47 yes I think project management like
2:08:50 that's one of those plumber roles where
2:08:52 where it's just going to be like
2:08:53 everything's [ __ ] changing how late
2:08:56 is it it's late isn't it
2:08:57 11 22. dang
2:09:00 time flies when you're in the AI
2:09:02 learning lab it flies it flies right by
2:09:05 I don't know what you people are doing
2:09:07 here how many people 74 people still
2:09:09 here folks it's Wednesday night you have
2:09:11 work tomorrow
2:09:14 my hair is holding up okay tonight I
2:09:16 don't know what happened although this
2:09:17 this situation we got a little we got a
2:09:20 little Rosy situation going on over here
2:09:22 but that's hanging out all right
2:09:25 uh what's the AI the AI Salon is a group
2:09:28 that I started last December you can go
2:09:30 to the salon.ai that's a link tree we
2:09:33 meet every two weeks
2:09:35 um in person here in Denver and then we
2:09:39 have an online thing so it doesn't
2:09:41 matter where in the world you are we
2:09:42 have people from all over the world in
2:09:44 fact last night sander who I met in here
2:09:46 spoke at the AI Salon from the
2:09:48 Netherlands it was like two in the
2:09:50 morning his time when he was talking
2:09:51 about his new company that was kind of
2:09:53 cool
2:09:54 um and we have an online Discord server
2:09:56 we have a LinkedIn group
2:09:58 and it's basically just
2:10:00 people that are curious about AI on the
2:10:02 adventure trying to figure this [ __ ] out
2:10:04 together and uh it's a it's a really
2:10:06 neat group
2:10:17 yes our brains will have an
2:10:20 unprecedented opportunity to rewire
2:10:23 right yeah like like this is potentially
2:10:26 evolutionary kind of stuff like if we if
2:10:29 we just fundamentally shift the parts of
2:10:31 our brains we're using actively because
2:10:33 of these tools
2:10:36 it's wild
2:10:38 flipping wild it's exhausting it is it
2:10:41 is
2:10:42 it's incredible exactly I'm more excited
2:10:45 than I've ever been I'm more terrified
2:10:48 than I've ever been and I'm now more
2:10:50 exhausted than I've ever been
2:10:52 as a direct result of this [ __ ]
2:10:56 and then you're asking but Kyle are you
2:10:58 gonna get so tired you're gonna stop
2:11:00 [ __ ] no
2:11:03 do you know how lucky we are to be able
2:11:04 to play with this [ __ ] you think I want
2:11:06 to miss any of this
2:11:14 yep it was pop rocks that killed Mikey I
2:11:17 remember that
2:11:19 what history is AI being taught whose
2:11:23 perspective
2:11:24 so they were trained on all of the
2:11:26 internet or all of the public internet
2:11:30 so what's in there is everything
2:11:35 but it was everything that was on the
2:11:37 internet so whatever wasn't on the
2:11:38 internet didn't get in there
2:11:41 so there's a lot of that
2:11:44 and then within it there's garbage
2:11:47 history and real history and
2:11:49 pseudo-garbage history and pseudo Real
2:11:51 History it's all in there
2:11:53 how you fine tune it and how you pull
2:11:56 the words out of the large language
2:11:58 model
2:12:00 um has largely to do with the fine
2:12:02 tuning so you could take the same base
2:12:04 large language model and fine tune it in
2:12:06 different ways to get it to behave in
2:12:07 different ways if the foundational
2:12:09 information is flawed then you can't fix
2:12:12 that
2:12:13 but they're aware of that they're
2:12:15 working on it
2:12:16 so
2:12:18 but but it's
2:12:19 so so John one of the one of the things
2:12:23 about
2:12:25 the way these tools work is
2:12:29 it's not copying and pasting like these
2:12:31 things do not have an agenda these
2:12:34 things are calculators they're they're
2:12:36 literally calculators that as you type
2:12:39 your prompt it's creating a mathematical
2:12:42 probability
2:12:44 of some single word in latent space
2:12:48 in Vector space which is a mathematical
2:12:52 thousand dimensional space that has
2:12:55 words clustered by semantic meaning
2:12:58 as you type your thing
2:13:01 it makes a probability that some some
2:13:03 Vector in that space is the it has the
2:13:06 highest probability of being the next
2:13:08 right word it doesn't even know what the
2:13:10 word is it just knows its Vector string
2:13:12 and that Vector string is associated
2:13:14 with a word so when you type your prompt
2:13:17 and you hit submit the first word that
2:13:19 it puts there was a mathematical formula
2:13:22 so is the second one so is the third one
2:13:24 so is the fourth one it doesn't know
2:13:26 what it's writing
2:13:27 so they're not
2:13:31 they're not inherently biased
2:13:34 other than
2:13:36 if there are things omitted from the
2:13:38 data set
2:13:41 then they're not in there
2:13:43 and if how the things were fine-tuned so
2:13:46 the way Chachi PT fine tunes their stuff
2:13:49 is
2:13:50 they put in a prompt it generates
2:13:52 something and then human beings go
2:13:54 thumbs up or thumbs down
2:13:56 that introduces human bias
2:13:59 so depending on who does the human
2:14:01 reinforcement learning here
2:14:03 that could make them more biased Claude
2:14:06 that's from a company called anthropic
2:14:08 they use a constitutional learning model
2:14:10 where it's not humans doing this it's a
2:14:12 set of rules it's a set of imperatives
2:14:15 so Claude has
2:14:18 potentially a higher likelihood of being
2:14:20 less biased but here's the here's the
2:14:22 the fact of the matter about bias is
2:14:26 bias from whose point of view
2:14:29 the thing that I think is perfectly
2:14:31 unbiased you may think is completely
2:14:33 biased and vice versa and then the third
2:14:36 person is going to have their own flavor
2:14:38 of this so I think in the future we'll
2:14:40 have different models that we can choose
2:14:42 from
2:14:42 I want one that looks and sounds like me
2:14:45 because because I'm a petty [ __ ] and I
2:14:48 want Petty [ __ ] in my large language
2:14:50 models
2:14:52 and then other people be like I want
2:14:55 tough guys
2:14:58 so I don't know
2:15:00 I don't know
2:15:01 I do not know I read that on average we
2:15:04 consume more data per day than one would
2:15:07 in a lifetime in the 17 and 1800s
2:15:10 consume
2:15:12 who uh Peter Kaminsky last night at the
2:15:15 salon talked about a um an image
2:15:20 organizing tool
2:15:22 system that he created
2:15:24 because he's generating so much
2:15:26 generative AI art that he wants to
2:15:29 publish in the world like his backlog is
2:15:32 like 13 000 images that he's created in
2:15:34 the past three weeks
2:15:38 yeah it's not only consuming it it's
2:15:40 generating it we're generating we're
2:15:42 gonna like oh my just the amount of crap
2:15:45 that we're gonna generate it's [ __ ]
2:15:47 nuts
2:15:49 that's tough it's tougher if you're
2:15:52 stupid
2:15:52 all right see just like a just like a
2:15:56 freaking Pavlov's dog
2:15:58 all right it just jumped me to the
2:15:59 bottom I think that's a sign I gotta go
2:16:01 I'll answer a few more questions
2:16:03 they were not trained on everything
2:16:04 available on the internet that's very
2:16:06 incorrect well yeah they were trained on
2:16:09 whatever opening I could get their hands
2:16:11 on and tag and do all that but
2:16:15 ostensibly a lot of the public internet
2:16:18 and they want to get more right so it's
2:16:21 all about the data and and the new
2:16:24 um
2:16:25 open AI bot that allows you to opt out a
2:16:28 bunch of the big publishing companies
2:16:29 are opting out of that so open AI may
2:16:32 have all of these companies may have a
2:16:34 tough time
2:16:35 training the new models if the big
2:16:38 publishing companies are like [ __ ] you
2:16:42 and and you know Google's probably going
2:16:44 to start cutting deals with companies
2:16:46 that AI that openai can't compete with
2:16:48 so we're we're starting to the next two
2:16:51 years are going to be some corporate
2:16:53 Cutthroat Malarkey
2:16:56 as Biden would say
2:16:59 um so so basically data analysis and
2:17:02 sorting
2:17:03 uh yes I think
2:17:06 so is the internet left-leaning yeah
2:17:10 um perfect example of how Neo became the
2:17:14 one on the Matrix they uploaded loaded
2:17:16 data to him
2:17:18 suggestion can we do more lab examples
2:17:20 with these tools and less Chit Chat yeah
2:17:23 we can possibly do that although
2:17:26 people aren't asking for that but yeah
2:17:28 like I would say
2:17:29 ask specifically for what you want and
2:17:33 then I'll do that because I try to do
2:17:35 as many questions as I can
2:17:38 my Chit Chat is answering questions if
2:17:40 the Chit Chat said hey can you show me
2:17:42 this tool or that tool then I'd be
2:17:44 showing this tool or that tool so sure
2:17:48 you are determining what I chitchat
2:17:56 alignment make them biased
2:18:01 well yeah hopefully when we want them
2:18:03 biased in the direction of human
2:18:06 prosperity
2:18:08 I made a chatty bot site
2:18:11 the date of her resources May of 2023
2:18:14 very cool I like chit chat
2:18:18 uh
2:18:20 once it can reinvent the wheel read
2:18:22 patents we are going places it can do
2:18:24 that right now
2:18:26 in fact I think there's some patent um
2:18:29 there's some patent plugins for chat CPT
2:18:32 and I would imagine at futurepedia.io
2:18:35 there's probably some intellectual
2:18:36 property stuff going around all right
2:18:38 listen
2:18:39 I gotta go people it's 11 30 here in
2:18:42 Denver I don't know what time it is
2:18:43 there if you're on the East Coast what
2:18:45 are you doing with your lives why are
2:18:47 you still here
2:18:48 I I love y'all I appreciate y'all very
2:18:51 very much this is fun I have a blast
2:18:53 doing these things
2:18:54 um thank you for hanging out uh
2:18:58 Share Bear Pate welcome again
2:19:02 um
2:19:04 multi-dimensional man Emilio's wife jrc
2:19:08 Dr J
2:19:11 I was in a paddock good night talk to
2:19:14 you later nighty night Tobias good night
2:19:17 Kyle great job tonight thank you very
2:19:19 much thank you I'm a librarian digital
2:19:21 Gods thank you so much
2:19:24 I see share bear peace out I see River I
2:19:28 see paint Pace Pace Pace jrc thank you
2:19:32 everybody this was fun I'll talk to you
2:19:34 tomorrow
2:19:35 bye-bye