aiLL 09-05-23 - Kyle Debates AI in Education and Predicts an AI-Powered Future

Transcript

1. 0:02they are learning lab welcome welcome welcome uh if you would go ahead and share the live if you've got
2. 0:11friends and followers you think might dig this even if you think they're if they're gonna think you're weirdos do it anyway um
3. 0:19welcome from Denver it is eight o'clock on the money and looks like Ben Shelton is about to take set one unless tiafo can turn things around
4. 0:34crazy I tell you crazy I tell you chaos generation what's happening Ted looks at life how are you Tobias
5. 0:41welcome sir welcome everybody Welcome to the lab we're going to talk about AI stuff tonight the shirt tonight if you're
6. 0:51wondering is is the old classic Basquiat Cassius Clay backwards it's tough to get these things printed backwards like this that makes it a
7. 1:02collector's item oh just you know never dull moments what I do today um I did I discovered I screwed up I
8. 1:19screwed up The Tick Tock shop even more than I thought I did so it's kind of untangle that wackiness stayed up till almost 1am watching the
9. 1:29German and Italy match over four and a half hours wow crazy yeah what's going on here all right Shelton serving for
10. 1:35the for the first set double set point 40 15. see if he cracks out a oh wow 127 mile an hour Ace blasted it past tiafo wow packing up to
11. 1:50go to Denver tomorrow okay welcome okay farmer Dr J Mark attiel how are you welcome everybody so we got some Irregulars here
12. 2:05those are what we call the regulars they have their own t-shirt design but uh I'm still figuring out how to make t-shirts
13. 2:13and not [ __ ] it up you know old guy with new technology you know how it is it doesn't go good hey what are you gonna do
14. 2:24anyway uh this is the AI learning lab um if you're new to uh to this space and to generative AI I will
15. 2:35talk you through some stuff to get you started we'll start you with some some URLs we'll we'll talk through those
16. 2:43we'll we'll show some pretty pictures these were all made in um ideogram which I still I've been Hawking that like I worked there
17. 2:51just because it's pretty amazing that you can now do text in AI image generation it seemed like that was going to be
18. 2:58farther away than whatever it was six eight months and then it just showed up and it didn't show up from mid Journey
19. 3:04or stable diffusion I was expecting it to come from stable diffusion quite honestly because they seem like they
20. 3:10have their [ __ ] together mid-tierney's just an 11 person company um and uh and this startup this startup called uh
21. 3:21I don't know what the company's called ideogram I don't know if that's the company but uh but yeah they figured it
22. 3:27out man they cracked the code so we're off to the races with text and images all right so what we do here is my name
23. 3:35is Kyle Shannon I'm an entrepreneur of many years I've started 12 or 15 companies I'm currently the founder and
24. 3:41CEO of a company called storyvine which is an automated video storytelling platform we don't our current platform doesn't
25. 3:48use a ton of AI for the automation believe it or not um it's it's a more traditional automation well you actually had to program [ __ ]
26. 3:56and think about it that's all different now so we're looking to to Really expand the product Vision with AI well I have expanded the
27. 4:06product Vision now we're looking to build it and get it paid for um so that's happening and I was also an
28. 4:11entrepreneur in the early days of the World Wide Web so I was a founder of agency.com back in the old and Tiny days
29. 4:16and uh the parallels between the early days of the web and the early days of this AI stuff are quite remarkable quite
30. 4:23remarkable and so that's one of the reasons I'm so excited about this stuff is that we don't get to experience these technological
31. 4:31primordial soup moments very often um there have been three of them in my lifetime one I missed and that was the PC Revolution
32. 4:40the second one was the world wide web and I hit that one square in my 30s and I didn't think I'd see another one in my
33. 4:46lifetime but here it is this is it and this is really exciting so if you have questions about AI can I make games
34. 4:53using AI yes in fact you can um you can just go to Chachi PT so here I'll I'll jump in and and just get get
35. 5:00started if you haven't played with chat gbt yet just go there go there right there chat.openai.com and just ask it to write you
36. 5:10a flappy birds clone in C sharp that will run in unity and then take that code paste it into unity and hit run and then if it if it
37. 5:20gives you an error then take the error take it back to chatgpt and said you gave me an error so so yes you can
38. 5:28um you can also use there's another one called pi which you can talk to and it's really conversational we call
39. 5:35it chatty GPT and chatty GPT will um you can just tell it to get to make up a game like you want to learn a particular
40. 5:43subject and tell it to invent a game to make it more fun to learn and it'll just do it so all sorts of stuff you can have
41. 5:50these things act like dungeon Masters all sorts of crazy crazy ass [ __ ] crazy ass [ __ ] so here's what we're gonna do
42. 5:57here's what we're gonna do we're gonna we're gonna go on ahead I'm gonna make some recommendations so if
43. 6:02you're new to this AI stuff and you like Macau where do I start I don't know where to start I'm so confused
44. 6:10everyone's talking about a I don't know what they're talking about this is very confusing are the robots gonna kill us
45. 6:16what do I do all right we're gonna get started we're gonna it's gonna we're gonna bring it down a notch just take a deep breath
46. 6:24this part of this channel is a Zen breathing Channel I don't know if you knew that welcome welcome just take a deep breath
47. 6:30in in through the nostrils and out through the mouth in through the nostrils and out through the mouth okay great now let's talk about AI
48. 6:39um so yeah so we're going to talk about AI if you've not played with chat gbt that's the URL right there
49. 6:45chat.openai.com that's the official one it's the best right now there might be some stuff around the corner uh
50. 6:53particularly something coming from Google that I think could kick this thing in The Shins but right now gpt4 is
51. 7:00is by far the best of these things you have to pay for gpt4 here but I've got a free option down below but even GPT 3.5
52. 7:07here is quite good so if you're brand new to this stuff just go there all the other [ __ ] I'm going to talk about you
53. 7:12can just ignore for now go there and just start with a parlor trick like you can ask it anything but start with
54. 7:18something like write me a country song about you know describe your dog or your significant other and just have it write
55. 7:24a country song and just just start to throw challenges at it and watch what it can do and then go to that URL down
56. 7:31there prompts.chat that will teach you how to talk to these things these are called large language models chat GPT is
57. 7:37is a large language model and um proms.chat will kind of tell you what they are how to talk to them and it's a
58. 7:45whole long list of prompts you can literally copy and paste where you can have chat GPT act like different
59. 7:50personas it's really quite remarkable by the way Shelton just won the first set so in the tiafu match um so uh so that's that
60. 8:03um Claude and Bard are are two that offer some slightly different things Bard's connected to the internet Claude
61. 8:09you can put really long prompts into I was playing with with uh Claude today working out some Workshop deals with
62. 8:18with a friend who's gonna host uh doing a workshop so I'll there'll be more about that coming up but I use Claude
63. 8:26today so those are the three I kind of use there's four more I call this the kitties table so the kitties table is these are ones
64. 8:35that are still worth playing with but are kind of lower on my list for a number of reasons um bing.com is from Microsoft that's
65. 8:43gpt4 for free and um it's connected to the internet and you can make images with it it's dollies it's open ai's Dolly image engine
66. 8:56sitting underneath it and it can read images so that's super cool labs.perplexity uh lets you play with uh
67. 9:04meta's llama 2 models Pi like I said is chatty GPT Poe lets you play with um Bots if you don't know what Bots are just go there and
68. 9:15hit the explore button and you can see all these Bots that people have made and you can make your own it's a really good
69. 9:20way to practice prompting okay so those are all the large language models Chachi PT is the biggie Claude and Bing
70. 9:27are you know those three I sort of go between those three to play um these four are also worth exploring
71. 9:35and then for images but you're like but God I want to make pretty pictures I don't know I'm like pretty pictures Kyle
72. 9:41it's not fair all you're talking about is words I don't like Words there's so many words I only know a few Kyle what
73. 9:51about the pictures here's the pictures calm down like I said this is a deep breathing exercise you're gonna be okay
74. 9:59and in through the nostrils out through the mouth in through the nostrils out through the mouth fantastic Bob tell him
75. 10:06what he's won you've won a new car all right um mid-journey really good best in terms of image quality fan fuckin-tastic
76. 10:19I have a feeling they're going to get their asses kicked in the next year because they're only an 11 person
77. 10:23company and their CEO is a little bit full of himself and they could be in trouble but right now they got it you
78. 10:30have to make images in Discord which I'm happy to show if someone wants to see how you do that it's a pain in the ass
79. 10:35but it just is the best ideogram this one right next to it is like 10 days old this is a brand new service they figured
80. 10:42out how to put text in images which you're like but Kyle that's not a big idea there's always text and images no
81. 10:48there's text and images when typographers and designers go into illustrator and type them in there but
82. 10:54you can't just say give me a poster with an eye on it that says AI learning lab well now you can and that's what ideogram is
83. 11:02and it is pronounced ideogram I heard someone I heard the founder talk about it and then Leonardo is a really cool
84. 11:09website sitting on top of stable diffusion and clipdrop.co is is a bunch of tools from stable diffusion from stability AI
85. 11:18and then and then prompt book openart.ai prompt book is just like prompts.chat for words this is the one
86. 11:26that teaches you how to talk to these uh image things um all the all the hearts that you're seeing flying up on the screen that's
87. 11:34the Irregulars the ones that come in here every night the sickos the sickos that are just in here every night
88. 11:39they're playing a game they're trying to beat their last night they hit like 75 000 likes in a two and a half two and a
89. 11:46half hour live so so just don't worry about that you feel free to join in tap the screen all you want just you know the the
90. 11:54the Irregulars they've got they've got wrist issues they have carpal tunnel and so they would appreciate the support so
91. 12:00feel free to tap the screen uh feel free to also share the live um and then futurepedia.io is just it's a directory it's a directory of like
92. 12:09four thousand forty two hundred uh AI tools and so a lot of times in here people will ask me about stuff that I
93. 12:18don't have a flipping clue about I'll just keep pointing You Back to Future PD if you want something really specific
94. 12:23like what's the best tool to I don't know analyze legal documents there's probably seven of them in there all right okay so with that
95. 12:32pop any questions you have in the comments below I'm gonna start answering them I can guarantee you I promise you I will get behind
96. 12:40um I haven't gotten my [ __ ] together to get moderators in here yet but you know I'm getting it I'm just uh I'm just
97. 12:46doing the best I can people um but yeah but pop pop any questions you have below I will do my best to answer them and let's have some fun
98. 12:55we're gonna have some fun tonight you know what I mean you know what I'm saying this is how we're doing it all right
99. 13:01hello can I make games using AI we started with that one you yes you can in fact was it I think you can also just have it just
100. 13:13tell it to make a game of snake in Python and it'll do that and go run it and ripple it play Blackjack versus chat EBT that's
101. 13:21really interesting chaos Generation The Secret of AI it's Turtles all the way down oh yeah I like turtles oh yeah Turtles is tasty
102. 13:36yeah yeah I like it I don't know what's going on with my voice tonight it's got It's it's deep and Gravelly there's something
103. 13:48going on yeah today I know Bard oh God fantastic ride on Mr Shakespeare it was fantastic all right let's see my dog is my
104. 14:03significant other dogs are the [ __ ] best um higher regulars and Kyle hello hello hello write effective that can do what others
105. 14:13don't what's effective effective AI Translate wrote a I don't know what that wrote a something AI
106. 14:28that'll do what others don't I want to hear about that workshops yes and so so um so I did a I did a podcast uh with a
107. 14:38woman named Rhonda I forget her last name she's a voice over actress and so we're putting together a three-session
108. 14:46workshop for creatives so for voice over actors and other creatives um so I'm going to start with that one
109. 14:52that's gonna we we decided tonight it's gonna start October fourth on it's which is a Wednesday night from five to seven mountain time
110. 15:02and it's going to be three sessions every Wednesday so it's going to be the fourth the 11th and the 19th 18th 18th
111. 15:12of October so that's coming um so as soon as I have that sort of packaged up I'll I'll let you guys know about it
112. 15:19um and then I'm gonna do a specific one for the it for the AI learning lab that'll that'll be probably a bit more
113. 15:25business focused but I'm but I'm going to do it based on what folks here have asked for so so anyway um I had the obligation to do that one
114. 15:34with her first so we're gonna do that it's gonna be cool if you want to take that one it's going to be good
115. 15:39um yeah it's gonna be a good one it's gonna be a good one Focus primarily for creatives all right fantastic yeah we're gonna do it we're
116. 15:47gonna do it Kyle win workshops October 4th are you cats using typing mind yet no let's go learn what that is
117. 15:56this is the other reason like this thing is so cool is like it's just you you can't keep up with it so
118. 16:05anytime someone says like have you seen this no tight mind a better UI for chep GPT interesting interesting well there's more [ __ ] on it
119. 16:20oh you can change the temperature yeah this has got some interface on it you can search your chats it shows you your API keys right there
120. 16:32yeah this is pretty good so basically what they did what what type mind is or typing mind is uh they took all the [ __ ] that old
121. 16:44Sam Altman and gang should have done six months ago they're like why wait we can write our own [ __ ] it so there you go
122. 16:55that's awesome yeah good for them you can change your you can select your model you can change your profile that's super cool
123. 17:05your info your information your profile image I wonder if that uh impacts your your prompt by what's by mean
124. 17:19buy once use forever price increase soon I I uh I bought a piece of software once that was a it was a lifetime
125. 17:32I paid whatever it was the 49 bucks for the lifetime subscription and then at some point like two years later they're
126. 17:39like everyone's got to upgrade and pay this new fee for the new version and I'm like but I bought a lifetime
127. 17:45subscription and they're like no that was the lifetime of that version I'm like oh [ __ ] you no that's wrong that was that was uh
128. 17:59what was it was clean my Mac I think it was Mac Paws clean my Mac [ __ ] but anyway that's pretty cool um
129. 18:10I mean 39 bucks for a lifetime thing assuming it's actually a lifetime thing I would check that um premium license giveaway
130. 18:19um I mean here's the deal with with something like this because this is kind of a slight augmentation to a thing that's already there
131. 18:27when open AI gets their [ __ ] together and fixes chat GPT it's likely going to have a lot of this stuff like like right
132. 18:33now Chachi BT is missing like a lot of what I see here is the stuff that drives me nuts that's not in Chachi PT so I
133. 18:40like it but this is one of those companies where this will just get you know this just becomes features of the
134. 18:46real deal so anyway interesting um thanks for sharing that uh it's a home base for all the big llms you can
135. 18:53organize your chats sets of lots of personas temperatures yeah it looks pretty cool um it's like the French have a different
136. 19:01word for everything Steve Martin joke thank you chaos generation you ever notice when you go to a different country
137. 19:11how does he say it I'd like to go to the hotel please and the taxi driver turns around and he goes and he goes I would like to
138. 19:24go to the hotel please uh uh feels so good so good what do you think of prompt hero I'd love heroes heroes are the best
139. 19:39prompt hero I don't know anything about I think I've heard of it but let me let's go look at it prompt hero
140. 19:47search prompts for stable diffusion chat gbt oh is this the one where you can sell your prompts no that's prompt bass right
141. 19:56number one website for prompt engineering search millions of AI art images by the models by models like stable diffusion mid-journey
142. 20:06so there's mid-journey ones yeah this is pretty cool oh there's there they're just not selling it yeah prompt hero looks pretty good
143. 20:14AI models loros this looks like it's a well put together site featured hot yep prompt hero.com that's worth that is
144. 20:31worth checking out that is worth checking out in fact let me see AI models large language models why are there only two
145. 20:52two prompts here looks like the language model prompts aren't really functional right now oh best chat CBT prompts
146. 21:12huh no results so like all of this stuff so so one of the things one of the things that I'll share with you
147. 21:24all of the tools all of the tools all of the tools right now all of the tools whether it's chat EBT or Bing or Bard or Claude or uh
148. 21:38pi or llama or prompt hero all of them right now are wait for it janky pieces of [ __ ] one of the Irregulars in the uh in the salon
149. 21:53Discord made me a janky pieces of [ __ ] logo um and here's what I mean by that it's not that the developers that make all
150. 22:03these apps are bad it's that November 30th 2022 is ostensibly when the popular AI industry started so that was nine months ago
151. 22:18so nine months ago Chachi PT launches and this is sort of this fork in the road where you know before that it was
152. 22:25just all scientists and researchers and mathematicians and after that it's the rest of us right and so that this
153. 22:30channel is about you know that moment and what now that now that these tools are in the hands of mere mortals what
154. 22:36the [ __ ] do we do with them um just know that as powerful as these tools are and as much noise as they're
155. 22:42making and as much as people are talking about them they are still very very early so they're very raw they don't
156. 22:49have a lot of sophistication they make mistakes they hallucinate they make hands that look like this sometimes they
157. 22:56make letters that do not look like any letter you've ever seen But if you use a tool like ideogram you can get it to
158. 23:02make janky pieces of [ __ ] content um so just know that that uh we're still very very early and that's it's okay
159. 23:13it's okay it's all right it's gonna be okay seriously it's gonna be all right all right okay all right all right yeah yeah we're gonna go on
160. 23:24ahead and need you to come on in on Saturday oh yeah and another thing Peter I need you to go on ahead and come on in on Sunday too yeah um
161. 23:37one of my all-time favorite movies one of my all-time favorite movies I have a red stapler at work uh so you can do prompting with multiple
162. 23:50models all in one place oh that's pretty cool that's the uh what was that uh was that one we just went and looked at what was it what was it
163. 24:03um type mind.ai yeah I'm gonna bookmark that one hang on notify me enter your email
164. 24:20the biggies all right I'll just drop it in here and then prompt here I'm gonna drop in here too prompt hero it looks
165. 24:28new it looks like they're still building it but that's okay a lot of these things are still building so all right dually bookmarked
166. 24:39Fran she asked yet what's up irregular share bears in the house There's Tom Baldridge I'm a sicko welcome Tom you're not a sicko you're
167. 24:51irregular [Laughter] Prince hello world can you make money with chechi PT oh I don't know can you
168. 25:10why yes you can but not the way you think oh I'll talk to you about that if you really want to know if you want to know
169. 25:19the secret of making money to chat GPT just send 39.95 to this PO Box in Sioux City Iowa all the secrets are contained within
170. 25:37did it just jump me to the bottom oh Tick Tock you and your godforsaken bug see here's the deal I'm I'm neuro spicy and
171. 25:49um I like answering all the questions there's many as I can and then Tick Tock occasionally just jumps me to the bottom
172. 25:54for because it's got a bug and it makes me cranky it makes me cranky and every time it happens I say
173. 26:01the same thing and does tick tock listen no they're listening to all the words I say because God forbid I say a word that
174. 26:07might offend someone they shut my [ __ ] life down for seven days but if I say things like fix your bug no not happening
175. 26:17hey I just got a note from oh laricia from my company is she's discovered the NPCs you know the people
176. 26:31that are that are like ice cream ice cream ice cream I sort of do that here when people give me the uh cowboy hat but
177. 26:43um but yeah I'm not okay I found this on the web for to do Siri when people give me the cowboy hat but check it out there Siri really
178. 26:54she's jealous she's Sherry's jealous of the Irregulars it's so true can you recommend any gen AI for visualization
179. 27:03like create a multi-layered donut with this data well so yes so let's let's go play um it's not super sophisticated
180. 27:16visualizations but that's also because I don't know what the [ __ ] I'm doing so we're gonna go to Chachi PT um
181. 27:29so here we are in chat EBT if you haven't seen chappie gbt before these are all the prompts these are your
182. 27:34different models you have 3.5 and 4 and then under gpt4 you've got some options here you've got default Advanced uh data
183. 27:43analysis and you've got plugins okay um in order to get to Advanced Data analysis that's what I'm going to show
184. 27:53you you've got to be in chat GPT plus which is the 20 Buck a month thing and you have to turn it on so you go to your
185. 28:00settings and then you go to Beta features and you turn it on so if you if you have paid your 20 bucks but you
186. 28:07don't see Advanced data analytics or plugins you haven't turned them on okay so Advanced data analytics this also
187. 28:16used to be called code interpreter so once you select it you now have this little plus button you can upload data
188. 28:24and so I'm going to go grab what was archive 10 no that's not what I wanted was it archive nine we'll grab archive seven although I
189. 28:38don't know how big it is so this is just some data I grabbed off the internet there's a site called kaggle just
190. 28:44horrible name for a site I don't I don't care what it is but it's where you can it's where you can get data sets
191. 28:51so you go play with data sets and uh get them at kaggle and so I'm going to say um tell me what data is in this archive
192. 29:10what's interesting about it and how you might visualize it
193. 29:24and so what's going to happen now is all right let's start by examining the contents in the provided archive
194. 29:32so what happens with Advanced Data analysis is that when you launch the chat it's actually spinning up a virtual machine
195. 29:42that includes its own set of python libraries and and python so it's basically a virtual computer and then if
196. 29:51you open these little boxes it says you know I'm doing this work now you can see what it's actually done is it has imported
197. 29:59libraries to open a zip file written its own python to extract what's in the zip file and then looked at what's in there
198. 30:07so it's it's game data related to it's data related to games and this is profile related data so now it's going
199. 30:16to inspect a few entries from each Json file to understand the structure of the data I forget what this is I think this is
200. 30:28other quite large causing a memory error to circumvent this I'd like to take another approach I'll read the files line by line so it
201. 30:35wrote some python that python failed because the virtual machine is too small and so it's trying another thing now now
202. 30:44it run it ran into the same issue let's take a more cautious approach I think this is going to fail as well
203. 30:52I'm gonna go grab a different piece of data here's a brief overview okay now it did it it did it player ID oh I think this is chess games
204. 31:04oh no rushing yards this is football yeah this was this was one night someone asked me you know can you
205. 31:13can you do uh like like fantasy football player you know player selection using chat GPT all right how might I visualize it player
206. 31:24performance a time chart series showing a player's performance metrics like passing yards over time team performance
207. 31:32distribution of player attributes players per team uh let's visualize visual allies
208. 31:46team performance by the Broncos and the New York Giants um
209. 32:02that's just and and we'll say um make it visually stunning and so now what it's going to do if we do it right
210. 32:19is it's going to go actually analyze the data it's going to pull that data into a data frame within the virtual machine
211. 32:24it's then going to write some python to visualize and I didn't tell it what kind of you wanted a what did you on a
212. 32:31multi-layer donut it might be able to do that I don't know enough about data visualization to you know do that
213. 32:38without screwing it up so I just tell it make it pretty and it's pretty good at making it pretty by the way tiafo's upper break in set
214. 32:47two against Shelton Ben Shelton something he's going to be he's he's a flipping Star him and and alcaraz
215. 32:55I'd like to see Shelton win this and play alcaraz I think those two be explosive together all right let's see I apologize for the
216. 33:05inconvenience the memory issue persists let's enjoy employ a different strategy we're going to break it into smaller
217. 33:12chunks so again like one of the things I find remarkable about this is I mean you saw what little input I put
218. 33:21into it it is prompting itself now it's doing its own problem solving to to figure out the data in this in this set
219. 33:29you can have this thing analyze images you can have it analyze PDFs you can have it analyze video files you can
220. 33:37change color spaces of of multimedia files you can um turn an image into a zoomed you know panning animation across the image um
221. 33:52it's really apologizing I sincerely apologize for the inconvenience we're still encountering memory issues
222. 34:00I'll use the Json Library which allows us to parse Json files so now it's just picking a different python library and
223. 34:07see if that works that's kind of interesting no it can't do it it's it's let's see
224. 34:23I did a cool thing a couple of weeks back where I took I brought in and I imported an image I had it determine the top 15 average colors
225. 34:32of the image and then create a pie chart with all of the slices of the pie chart in the color of the of the most
226. 34:40prevalent colors of the image and then I had it name the colors like a creative Crown color and it did
227. 34:47it brilliantly it was pretty beautiful are we doing it are we doing it did we fail what's it got going
228. 35:04it's chunking it into 1024 by 1024 chunks one megabyte chunks I assume it it knows that it can keep that in memory
229. 35:15although it looks like it might have hung anyway Um this can do it let me go let me go grab I'm just going to bail on this
230. 35:25let's go do let's go do a new thing I'm gonna go grab a different piece of data that's smaller uh archive six that was archive seven I
231. 35:33mean what's cool about this I'm just grabbing random things um let's the data data in here and um what is interesting and how might we
232. 35:52visualize it and it's still up uploading there we go all right all right all right all right all right all right so anyway tktk HMO
233. 36:09we'll try this with one other data set we'll see what this one is
234. 36:23Spotify data let's proceed by loading the CSV file so so it found a thing called Spotify data so it understands that it's Spotify data
235. 36:35it might contain data related to Spotify such as song details play accounts user preferences the data set contains details of songs
236. 36:43on Spotify artist name track name track ID popularity score of the track year the song was released genre danceability
237. 36:51it's got we've got danceability data here kids and energy data and key and loudness and mode mode is minor or major speechiness
238. 37:03presence of spoken words in a track acousticness this is really cool valence wait what's balance valence valence
239. 37:14what's valence valence measure the musical positiveness so I guess you know the more minor and said all the all the songs I know I'm a
240. 37:24singer-songwriter all my songs are depressing so my valence score would be very low duration in microsec in milliseconds time signature
241. 37:34considering the columns available there are several interesting visualization distributions of songs by your popularity distribution
242. 37:42average song characteristics per genre correlation Matrix so let's do um visualize number four so it it knows what number
243. 37:55four is it gave us a fourth option there visualize number four in a visually stunning way let's see what it does
244. 38:08it's gonna do a heat map we're going to use colors to represent the degree and direction of correlation between the
245. 38:13variables to make it even more visually stunning stunning we can use a diverging color palette to make the positive and
246. 38:21negulation negative correlations stand out so like even if you know nothing about data visualization and and data analysis like I don't
247. 38:33it'll just do it there you go so it's got correlation Matrix heat map popularity so it's all all the different attributes
248. 38:43I guess for all the different songs and where they correlate and where they don't amazing so anyway yeah
249. 38:54danceability and Valence show a positive correlation implying that the tracks that are more danceable tend to convey a
250. 39:00more positive or happier mood oh this is cool so so some observations from the heat map darker Shades in either
251. 39:07direction indicate stronger correlation a diet a diagonal is always one because any variables are perfectly correlated with themselves
252. 39:16features like energy and loudness show a positive correlation indicating that tracks with higher energy tend to be
253. 39:23louder that makes sense acousticness shows a negative correlation with energy and loudness suggesting that more
254. 39:31acoustic tracks tend to be softer and have less energy James Taylor Martin Sexton John Hyatt Brandi Carlisle breaks that though
255. 39:42because she's she's acoustic and she rocks her ass off anyway that's uh that's Advanced data analytics so that's
256. 39:50that's the way you can do that just pay your 20 bucks turn it on and start popping data in there isn't that amazing
257. 39:57it ain't amazing and amazing all right I'm now 82 messages behind because we went down the little rabbit hole there
258. 40:04so sorry about that I'm not weird I'm a baby though um I like turtles okay can I make games using AI we talked
259. 40:18about that dude I I will avoid put the title on some mid-journey generated pictures don't like trunk traveling what is recommended
260. 40:36best AI for coding so gpt4 is really good although I saw a thing today the gpt4 has gotten much worse I think Stanford
261. 40:50released a study where they they looked at the quality of coding of gpt4 and it's gone down pretty dramatically and
262. 40:57at the same time but but it's still good so go there Bard is supposed to be good at coding bard.google.com and then um
263. 41:11code llama if you go to labs.perplexity.ai code llama the code llama model is supposed to be better than gpt4 at coding so I would say bard
264. 41:25llama right there they've hosted it for you and gpt4 all all three of those are going to be some version of better than I can code
265. 41:36which is which is the bar is very low I know what code is Wolfman Jack Wolfman Jack wesn Radio Rock hip-hop and smooth jazz
266. 41:52Kyle the Irregulars are scheming to send you hats at the top of the hour giddy up all right what time is it how are we
267. 41:58doing on all right I got it's 42 so I've got what is that 15 18 minutes until the Hat damn I got it okay see I know you said
268. 42:09I didn't prepare I gotta get some John Wayne quotes up all right hang on I'll I'll be a little prepared John Wayne quotes
269. 42:26all right courage is being scared to death and saddling up anyway all right life's hard but it's harder if you're
270. 42:37stupid that's not actually the quote the quote is life's tough it's tougher if you're stupid that one I learned the last time let's see
271. 42:51talk low talk slow and don't say much all right I got some wait 50 all-time favorite John Wayne quotes expand to show full article
272. 43:09all right all right I got some quotes there I'm prepared fine share bear call me out with the with the hat Ambush fine fine
273. 43:20I see how this works I'm no longer in control of this live am I but Kyle you never were ah you kids your kids is funny
274. 43:39um she's got a funny face let's put her we'll put her front and center why is there a creepy lady over his
275. 43:47shoulder it doesn't make any sense to me this is disturbing all right Turtle girl is here Tyron truth what's
276. 44:02happening fiction derived from fiction hmm maybe don't know what that's in reference to but yeah best app to build
277. 44:09spreadsheets with formulas to merge with Excel sheets or QuickBooks um it depends what you're trying to do
278. 44:19Advanced data analytics that thing I just showed you can tell it so I've pulled data into that and then had it analyzed the data
279. 44:29and then clean up the data and rewrite it in the order that I want in sheets within an Excel sheet so so Advanced
280. 44:37data analytics can can generate CSV files or Excel worksheet or Excel spreadsheets with multiple worksheets in them it's pretty amazing
281. 44:47um so it can do that um there's also for sheets thank you very much there's also for Sheets a thing called GPT for Sheets plugin
282. 44:57and in every cell of your sheets you can call open AI so you can call it to do a translation you can call it to do an
283. 45:05open AI prompt you can call it to generate an image so that's pretty amazing um what what do you want oh you want to go
284. 45:17see Mama what do you want to do oh a little whiny bumper all right
285. 45:32let's see um I don't know if it can do Quickbooks data but I assume I mean QuickBooks you can import csvs right so um or Excel files so
286. 45:47and it might even be able to like if you if you grab the data from a from a QuickBooks file it might be able to read
287. 45:56that and rewrite it and and I don't know it might be able to even output that if it's a pretty standard format that
288. 46:03QuickBooks does I mean it's been around for decades so so anyway so I'd go play with Advanced data analytics or the GPT for Sheets plugin
289. 46:12for that stuff I'm not good at Excel or I'd show you stuff I'll just embarrass myself anything involving math
290. 46:20yeah it's a it's a it's a situation all right yep and folders Kyle you said subscribe short term as this stuff is
291. 46:30constantly changing it's new correct I did so so that's I think that's in reference to that to that site saying
292. 46:38it's got a lifetime thing if if it's got a lifetime subscription I'm a little more lenient on that but but I think
293. 46:45right now spending money on yearly subscriptions is is silly because this stuff is changing so fast like the
294. 46:55the the speed with which these things are are just changing thanks shiver this the speed with which these things are
295. 47:02just changing even on my list of just like you know what what I'm finding useful personally is just I can't keep up with this [ __ ] so
296. 47:10you know if I had subscriptions to all these I'd just be kicking myself because every time one sort of drops off the the you know
297. 47:17the big kids list um it makes me cranky that I paid for it but you know let's see tiafo won the second set so one set a piece
298. 47:27Shelton and tiafo all right there's my John Wayne quotes for when the when the Hat attack comes all right
299. 47:40Omar domain if businesses are concerned about data privacy and public chat gbt what instead should they use good
300. 47:47question so there's a couple of things chat GPT just announced GPT Enterprise which right now you have to be spending ten thousand
301. 47:58dollars a month for them to even consider you so if you're a big company chat GPT Enterprise um you've also got you can you can spin up
302. 48:08custom instances of open AI um servers in Microsoft azure so you so you can do things there the other thing you can do is go to the open
303. 48:24source Community you now have um the Llama 2 models and the new code llama model from meta these are fairly
304. 48:33significant large language models that are free for people to use so you can set up your own servers whether it's a
305. 48:41virtual machine on you know some cloud-based thing where you have access to gpus or you've got your own Iron with
306. 48:48big gpus on it you can run this [ __ ] locally so you're not doing any calls externally at all now you're gonna have
307. 48:54to roll your own you're gonna have to fine tune it you're gonna have to figure all that [ __ ] out but if what you're
308. 48:59trying to figure out is data privacy and security there's there's the concerning phallic elephant ears thank you for that it
309. 49:09throws me every time but I like it it's solid it's good it's it's a good animation it's a it's solid 3D mapping to to managon in my machine
310. 49:23um yeah but but I would do that so I would I would look into the the chat GPT Enterprise and and they say that uh GPT
311. 49:32Enterprise is going to be um there's going to be a self-service version of it that's coming um so so at some point after they after
312. 49:44they get done spinning up their ten thousand dollar a month clients then they'll they'll they'll they'll come to
313. 49:50the rest of us groveling for data privacy and chat GPT um like I said all of these tools right now are right for it janky pieces of [ __ ]
314. 50:07laughs they're getting less janky but they're still they're still improving let's say um [Music]
315. 50:26all right oops scrolled a little overzealously it happens I do it all the time cool thanks on the thumbs up on prompt hero yeah
316. 50:34that looks pretty cool uh um Thelonious damn smartypants no chat GPT prompts yeah it's so I I would assume
317. 50:43that prompt hero is like so what they're doing is smart they're saying okay there's going to be lots of different
318. 50:48models so let's create an info let's let's create a clean interface where you can find good prompts from lots of
319. 50:55different things but they gotta spin spin all that up and I would assume that there's not a prompt API from chat GPT yet so they're
320. 51:03probably gonna have to scrape some [ __ ] and whatever it's it's probably a thing is my guess it'll get there how's
321. 51:10everyone been Joker I've been good can my husband be an honorary irregular because I walk around the house
322. 51:17listening to Kyle every night I'm so sorry he is absolutely welcome to be an irregular but you know he's got to put
323. 51:23up with our nonsense so ask okay okay farmer and and uh Tobias they they make it they make date night out of AI learning lab lives
324. 51:35uh I was working and didn't realize we went live it happens I sometimes we're live and I don't even realize we've gone live you know
325. 51:44nine years old is wait no nine-year-old is good at this prompt said me and you are too old for this we are too old for this but screw them
326. 51:54this is I'm telling you so so here's the thing that I I I deeply deeply deeply believe this these AI tools
327. 52:04revenge of the liberal arts major is coming and revenge of the Gen xers is coming because I think having a broad view of
328. 52:13things is going to be a superpower and I think being on this planet for extra decades is going to be a superpower why
329. 52:21because these tools these fancy ass not those these fancy ass large language model tools they produce content so quickly
330. 52:35that if you've been on the planet for a while we're going to be able to look at the what they produce and go that's good
331. 52:40that's bad like that so our life experience and then if you've got a broad view of the world if you're not sort of hyper focused on a single
332. 52:51thing again you're going to be able to sort of context shift and now I'm copywriting now I'm making images now I'm making
333. 52:58music now I'm doing project planning now I'm doing creative writing right boom boom boom boom [ __ ] Field Day for ADHD years Field
334. 53:08Day for Gen xers and and you know you know Millennials older millennials Field Day for um liberal arts people horizontal thinkers
335. 53:21we're having a [ __ ] party with this stuff so yeah you can tell your nine-year-old well he's dying say thanks honey you're
336. 53:32right I love you that's that's what you should tell your nine-year-old and then you can secretly say what's really going on
337. 53:47all right [Laughter] Silver Fox I just call Kyle my guy that's when I was watching my guy and yeah sorry [Laughter]
338. 54:01office space is the best TPS reports ASAP Lumberg I made that you made office space did you work on office space Ted looks at life
339. 54:14oh my God just like I just I I like every flipping scene of that movie has a memorable line in it you don't have you don't have on the
340. 54:30required amount of flare flare it's like oh I'm gonna need you to go on ahead and come on in on Saturday yeah
341. 54:46you've been missing a lot of work Peter I wouldn't call it missing it well that's good that's good your management material
342. 55:00oh my goodness That's a classic it's a classic it's a real cup Tick Tock just jumped me to the freaking bottom
343. 55:09this is like a bad SNL skit that's not funny that I have to react to every time like I'm some sort of pathological Tick-Tock interface troll
344. 55:19[Laughter] oh Tick Tock why
345. 55:33I saw Peyton here welcome uh um 1032 EST Bingo what time is it
346. 55:47oh EST so that's 8 30. that was ages ago I'm way behind of course I am that's that's one of the that's one of the Traditions around here um
347. 56:01looks like a touch screen computer it's a spin stand box container not virtual machine oh wait it's a wait it's spin a stand box contain it
348. 56:16it's spin a stand box container I don't know what a stand box container is not a virtual machine but please educate me on that
349. 56:27well Chachi batiken too hang on is that the right thing stand box container versus virtual machine
350. 56:43containers and virtual machines are both technology used for virtualization and resource isolation they have distinct differences
351. 56:52in terms of architecture efficiency and use cases here's a comparison isolation and Resource Management containers share the host operating
352. 57:02systems kernel okay which means they're more lightweight and efficient however the shared kernel can potentially lead to less strict
353. 57:10isolation Okay VMS so so VM is is this is really a standalone thing containers are highly efficient containers are known for their
354. 57:23portability generally have better performance containers often good isolation for most use cases but not as secure as VMS Okay
355. 57:33cool so it's a container it's a spin spin stand stand box container not a virtual machine cool thank you for that all right
356. 57:46see you learn something new every day here people this is about learning it's the AI learning lab we're learning
357. 57:53we start clueless and then we remain clueless because someone else always knows more and you know who knows even more than those people
358. 58:03these things they know [ __ ] everything hey a lot of partner wait is it that time is that top it is it's that time oh good lord hang on
359. 58:13the [ __ ] pressure y'all are killing me every man deserves a second chance but keep an eye on him uh wait I gotta make
360. 58:25him bigger my eyes are old uh tomorrow's perfect when it arrives and puts itself in our hands it hopes
361. 58:33we've learned something from yesterday John Wayne tomorrow is the most important thing in life it comes into us
362. 58:39at midnight very clean it's perfect upon arrival and this is not here's the problem [Music] because I'm shitty at John Wayne oh good lord
363. 58:51I'm in a flop sweat right now I'll tell you what you know what little missy we're not going with actual quotes
364. 58:56because you know what I got a generic where's my hat where's there to go see I got a generic Southern Cowboy kind of
365. 59:05accent and this is how I talk I'm gonna talk like this and we're going to talk about you know we're going to talk about
366. 59:10AI I don't know if you tried that AI but AI is good ai's powerful ai's this is like one of them awful NPC things hat hat hat hat
367. 59:21oh dogs are fighting is that my uh yeah I like it I like it I need I need suspenders for that one do a little dancing do a little jig
368. 59:34ding ding ding oh that was awesome thank you sorry I you know I I you know I should have written a script I should
369. 59:44have had chat GPT write me a script and learned it bad actor that's why I'm an X actor thank you for that that was awesome
370. 59:55y'all y'all are swell how big are the data files for it to die um they they're not that big uh the the the the
371. 1:00:06megabyte good wow welcome to the planes the plane boss zip Lane um did Shelton just break oh he did Shelton just broke tiafu in the third
372. 1:00:22set um I think for code interpreter for Advanced Data analysis the limit was 100 megabyte file but now they've changed it
373. 1:00:34so you can upload multiple files what I don't know is is it a total of 100 megabytes or if it's 100 megabytes per
374. 1:00:41file I assume it's a total of 100 megabytes because the fact that that one file I think it was a 90 it was a 90 megabyte file and it kind of
375. 1:00:52choked it in its container um so you know not huge amounts of data again like like the the Advanced Data and
376. 1:01:02analytics right now in chat apt you know again I'm going to go back to this thesis that all of these tools right now
377. 1:01:10are janky pieces of [ __ ] so it crashes a lot it doesn't quite get it right A lot it only does python right now it's only
378. 1:01:18got limited libraries right now they're like they've probably picked the sort of safest set of libraries um
379. 1:01:26the the visualization libraries it uses are fairly fairly unsophisticated although it might just be that I'm
380. 1:01:33unsophisticated and don't quite know what to ask for um so so anyway so so I think it's decently I think it's decently limited paid from
381. 1:01:44from what I can tell uh certainly for someone you know who who is used to dealing with big data um it's probably comically small
382. 1:01:55but you know who knows who even knows anymore does anybody really know what time it is does anybody really care
383. 1:02:07that's a Gen X kids song remember hearing that when you were a kid [Music] all right let's see that was in a day we had three
384. 1:02:17television channels and two radio stations and we liked it you listen to whatever crap song our parents liked all right
385. 1:02:33I'm 73 messages behind by the way so if you're wondering why isn't he answering Michael an question it's just not fair
386. 1:02:41I don't even think he's live I think he's an avatar I'm pretty sure he's GPT five you know I don't know if you knew
387. 1:02:48this but they said gpt5 was gonna look like people and he doesn't look normal so I'm pretty sure he's gpt5
388. 1:02:56but he doesn't seem very smart you would think gpd5 would be smarter wouldn't don't you wouldn't you I mean I would think that
389. 1:03:04but he's definitely not live because he didn't answer my question I've asked it 73 times I look like a Al Franken there for a little bit
390. 1:03:15also that's like an Al Franken look buffering a little for me just once though all right well Tick Tock sucks
391. 1:03:29Xfinity sucks teacher here what app is best for detecting student using AI to complete written assignments
392. 1:03:36Catherine I would strongly strongly strongly strongly strongly strongly strongly dissuade you from taking that approach um
393. 1:03:48your student whichever tool you find that can detect AI your students will quickly figure out ways to um use tools to rewrite it so it's not
394. 1:03:59detectable by the tool that you selected um this stuff this stuff is coming everywhere it's going to be in the world that your
395. 1:04:13students enter and so rather than try to do this and say to hell with the new technology accept the fact that this stuff's not
396. 1:04:24going to be uninvented so Google three days ago announced the general availability of duet which is their
397. 1:04:32version of chat GPT in all of the Google tools it's going to be in Google Docs Google Sheets Google presentations
398. 1:04:39the the help me write button it's gonna literally be help me write it's going to be in everything it's in slack right now
399. 1:04:46it's in notion it's going to be in Microsoft 365. it's going to be in Apple's operating system it's going to
400. 1:04:53be everywhere so right now these tools kind of live off to the side but they're very quickly being integrated into what
401. 1:05:00we do now if you work for a school district that at the district level it says we're blocking chat GPT I feel bad for you
402. 1:05:09I don't think it serves the district I don't think it serves the students I don't think it serves the teachers you
403. 1:05:17can use these tools to read and analyze papers and help create You Know lesson plans faster and do a lot of the
404. 1:05:25administrative that work that you do much much faster these tools are remarkable take a look at um Harvard this year just released their
405. 1:05:38generative AI guidelines Harvard you can take that to the school board or to your to your administrators if this is coming
406. 1:05:45from the administrators Harvard now has generative AI guidelines let me go look at Harvard gen AI guide lines
407. 1:06:05provost.harvard.com guidelines for using chat GPT and other AI tools at Harvard protect confidential data you're
408. 1:06:17responsible for any content adhere to current policies on academic integrity be alert for AI enabled phishing um sister one
409. 1:06:35there's another one um [Music]
410. 1:06:49anyway yeah uh AI policy gaps diverse policy stuff go go look for the for the Harvard policy audit also look for the city of
411. 1:07:00Boston's guidelines on using generative AI I would also look for a tick-tocker her name is I think it's I am Gibson or
412. 1:07:08Gibson's in her name she's an English teacher she brought chat GPT into the classroom she teaches English in 11th and 12th
413. 1:07:16grade I think and she's engaging her students to like her her senior class this year she's having them write
414. 1:07:26um a major thesis paper without chat EBT and a major thesis paper with chat EPT and the entire semester the entire year
415. 1:07:36they're going to be talking about what are these tools good at what are they not good at what did you have to rewrite
416. 1:07:41what was hard for you to write what was easy for you to write where did it help where did it hurt all that sort of stuff
417. 1:07:46she's engaging their critical thinking skills that's what's going to be most important here's the challenge our
418. 1:07:52current education system is training students for a world that's about to cease to exist these tools over the next three years are going to
419. 1:08:05start to do all of the Tactical execution of like writing copy and entering data and doing all this stuff
420. 1:08:13all the all the sort of rote stuff that a lot of our education is training people to do is going to be done by
421. 1:08:18these tools what's going to be required is critical thinking so I would strongly encourage you don't think like how do I
422. 1:08:26detect them and get them to stop using it it's not going to stop and it doesn't serve them so that's my soapbox I'll get
423. 1:08:34off it you can disagree with me but I think this is really important it's something I'm very passionate about
424. 1:08:40um and I I I'm not I'm not historically passionate about education but even though I started a channel called the AI learning lab
425. 1:08:50um I just think this stuff's super important and it's coming whether we want it or not and it's going to transform not only
426. 1:08:56education it's transforming all of the creative arts it's transforming all of knowledge work it's transforming
427. 1:09:02anything that's computer mediated is going to be absolutely radically transformed by this stuff and the more that we can
428. 1:09:11get curious about it and embrace it and learn what it's good at and what it's not good at that's how we're going to
429. 1:09:18serve students of all ages so that's where my Haiti is okay my boyfriend's in jail and I need an AI to basically pretend they are an
430. 1:09:34attorney um okay that actually sounds serious and and um so a couple of things I would say go to prompts.chat
431. 1:09:44and look at how they've crafted the personas that they create I'm pretty sure in there is you know act like a
432. 1:09:52lawyer you can tell Chachi BT to act like a lawyer act like this act like that when you're dealing with legal stuff it's really good at
433. 1:10:02um analyzing documents and it's really good at having opinions but my nickname for chat GPT is mansplaining as a
434. 1:10:10service it will confidently answer you even if it's wrong there's a pretty famous case six weeks or so ago of a
435. 1:10:19lawyer that had chat GPT write a legal opinion complete with legal citations and handed it to a judge
436. 1:10:27and the judge clicked on all those legal citations and realized they were all fake [ __ ] because Chachi PT just made them up so
437. 1:10:37this stuff can actually probably give you you know at least good understanding of what your options are and maybe give
438. 1:10:44you ideas and maybe write nice letters for you but double check everything especially if you're using citations in
439. 1:10:51any way double check that stuff um the other thing that I think is is potentially worth your time to explore is play with pi dot AI
440. 1:11:04or the PI app if you go to um uh the Apple Store and look up Pi Pi personal assistant you'll find that app the company's
441. 1:11:15inflection and put it into voice mode and you can actually just start talking with it and asking it questions here's
442. 1:11:23the situation here's the situation with my boyfriend and why that's why that's potentially interesting is Pi's actually
443. 1:11:30connected to the internet so if there is modern case law that could help in the argument Pi will be able to search for
444. 1:11:40that modern case law better than chat GPT which isn't connected to the internet um you could also try Bard
445. 1:11:48um which is on my first little page there Bard is connected to the internet the other thing so so there's three
446. 1:11:57things I would try I would try gpt4 at open AI I would try Bard I would try pi and then if you've got long documents
447. 1:12:07like if you've got long case law documents or legal briefs or lawsuits or whatever it might be Claude can read very long documents like
448. 1:12:18up to about 75 000 words so you can take very long documents upload them to Claude or just paste them into the prompt and then ask any
449. 1:12:27questions you want about it all of these tools are really good at reading documents and analyzing them Claude is
450. 1:12:34really good at long documents so hope that helps paint looking at the uh the advanced data analytics the heat map right
451. 1:12:44writing its own code and generating heat maps and charts and graphs yeah it's [ __ ] remarkable um read the probable cause affidavit
452. 1:12:52tell me what what motions to file yeah okay yeah so if I don't know if you're saying read or read if it read it it's really
453. 1:13:01good at that I would say Claude is is the one if you want to have it read long documents and not have it choke claude's
454. 1:13:07the one to do that with um and yeah Thelonious talking about do not pay.com is an AI lawyer I depending
455. 1:13:15on what your your boyfriend is facing uh do not pay could work um but you might need something a little
456. 1:13:22more sophisticated wonder when we will when we will find uh on the Bible codes I mean listen there has there's there's been
457. 1:13:33so much written about the Bible codes these things know know all of you know these these tools chat Chachi BT Claude
458. 1:13:41Bard all of these things have been trained on as much of the public internet that people can get their hands on that that
459. 1:13:49these companies can get their hands on so something like the Bible Code that people have been talking about for what
460. 1:13:56centuries now long long time these things know all that right so you can absolutely just go ask it all that stuff
461. 1:14:02and just run down your own rabbit holes I want you to act as a Biblical scholar that's really good at code breaking and
462. 1:14:09blah blah blah blah and just start asking you questions um what's the secret human question to know who is a real human profile wait
463. 1:14:17what is the secret human question to know who is a real human profile I don't know what the secret human question is
464. 1:14:25but I do have something that I can show I'll do a little I'll do a little showing off here that'll it's gonna piss
465. 1:14:32some people off but I think that's funny um [Music]
466. 1:14:48[Music] I am a verified human and I'm anonymous but I have a blockchain verified by my
467. 1:15:02retina at a thing called the orb World ID so this is this is an nft this is a blockchain identity badge or identity token
468. 1:15:18um of me that doesn't require my name my social security number any of that [ __ ] it just verifies that I'm human
469. 1:15:29um and this is Sam Altman's new company called World coin and this is World ID so his his new company has three
470. 1:15:35components World coin which is the cryptocurrency world ID which is a human verification system and then a whole
471. 1:15:42transaction like a PayPal kind of thing and it's all in developer mode right now so if you're a developer you can be
472. 1:15:49building applications on top of this system where for example you could build a Twitter that people could only use if they were
473. 1:15:59verified to be a unique human one of the things this did when it scanned my irises which I know is creepy as [ __ ]
474. 1:16:07but whatever I don't give a [ __ ] when it scan my Iris what it said was it was comparing my um signature my biometric signature with
475. 1:16:18all of the other humans in the database it was it was basically validating my uniqueness compared to the other humans
476. 1:16:25in the database so once it did that it gives me this blockchain verified thing so yeah that's here I went to New York City to get that
477. 1:16:35because you know like you do [Laughter] um he's so weird he got his Iris scanned he
478. 1:16:50put it in a thing called the orb and then and then he seems the normal sort of but he's a little weird you know what
479. 1:16:56it seems like he got scanned by a robot and that's what it seems like oh thank you very much for their how how
480. 1:17:04actually one of the last plays I did I played Elvis in uh Steve Martin's Picasso at the LA panajiel in Ridgefield Connecticut
481. 1:17:16that microphone reminded me when I sang I sang me some Elvis songs [Laughter] all right let's see Ahoy sir how is the picture-in-picture
482. 1:17:31featured enabled this is when you go live um there's an option called dual camera what's weird about it is Tick Tock took
483. 1:17:43it away from me for like a month I don't know why it just was gone like it was there and people were like how do you do that
484. 1:17:50and all it is is that box is the back camera and then this is the front camera and that's how I'm doing it so that's
485. 1:17:58what did it dual camera feature when you go live code quality seems to change quite a bit depending on the language and model yeah
486. 1:18:07I know the code quality right now is a mess they call that um in the in the large language models drift so as open AI is fixing
487. 1:18:20their their fine tuning of chat GPT some things get better some things get worse which is again you know the I kind of
488. 1:18:28make the joke about these tools being janky pieces of [ __ ] but it's important to know that as you get more educated
489. 1:18:35with these tools and you start playing with them you start understanding what they are um you're going to start to want to build
490. 1:18:43stuff and you might use a tool like zapier and you create these automations where someone can fill out a form and
491. 1:18:48then you send stuff to open Ai and it makes these really cool things and you're like oh just fantastic and I made
492. 1:18:53this application and it's perfect and then three months later you go to show it to someone and it's not perfect anymore
493. 1:19:00because there's drift so so a strong recommendation right now is learn as much as you can build as much
494. 1:19:08as you can talk as much as you can to people about what you're learning and accept the fact that whatever you build
495. 1:19:17right now whatever you learn right now is going to be completely irrelevant three months from now just just know that just like life
496. 1:19:25everything is some temporary game that we play it's just that with these tools it's much more temporary because they're
497. 1:19:31because they're changing so fast this stuff's changing so fast like the the the requirement right now that is on all
498. 1:19:39of us to be good at prompting and to be good at you know using our native language as a programming tool
499. 1:19:49that requirement is going to go away if you're a developer you're still going to need to be able to do that but most of
500. 1:19:55most human beings are not going to need that you're going to be able to just ask this [ __ ] kind of like when I was
501. 1:20:00showing Advanced data analytics where I said hey what's inside this and what's interesting about it and it just started
502. 1:20:07doing all its own [ __ ] that's how these tools are gonna act in the future right now they don't act like that way
503. 1:20:13because they're janky piece is a [ __ ] but as they get more sophisticated there's going to be increasing layers of
504. 1:20:20interface that hide all the sort of technical underpinnings from US unless we want to dig in there and play with
505. 1:20:26those most people don't want to do that most people are like I want the robot to do the hard stuff for me could the robot
506. 1:20:33do the hard stuff if the robot could do the hard stuff I'd like this it would be pretty good because there's a lot of
507. 1:20:39[ __ ] I do I don't like you know what I mean and so I I would like the robots to do the hard stuff and if the robots do the hard stuff
508. 1:20:49everybody wins seriously like I think this is my this is my thing like the more I'm learning I think the killer app is
509. 1:20:59our world is too complicated and most of us are miserable because every [ __ ] day we spend two hours on the phone with
510. 1:21:06a t trying to get them to reinstate our AARP discount for ten dollars a month for the stupid [ __ ] phone bill that was my
511. 1:21:17Labor Day Monday two hours just trying to get my stupid 10 bucks a month discount back they were clueless on the phone we had a
512. 1:21:30bad connection they couldn't hear me I couldn't hear them we were both screaming at each other then I go to log into the website and it
513. 1:21:38breaks it doesn't log in it's just like something went wrong sorry two hours it just wouldn't it be nice if you just said
514. 1:21:47excuse me could you go make sure that I have all the discounts I need on My ATT thing and it's like sure boss like Golden Retriever
515. 1:21:56I'll do that for you let me go now it runs off and your little robot talks to their little robot they work some
516. 1:22:03[ __ ] out your little robot comes back and says okay boss here you go I fixed this I got this you have two dollars
517. 1:22:09Less on this you have a choice do you want this or do you want this uh I want that okay great got that I'll go change
518. 1:22:14that now you have another choice you want this you want that do you want that do you want a new Apple watch you can get one for free
519. 1:22:20is it gonna cost me anything nope okay click I'll be back all right your watch will be there tomorrow why can't it do that well it can it's
520. 1:22:33going to that's the killer app all right now screwfuffled I was totally kerfuffled there was just too many hats it was too much pressure
521. 1:22:45I didn't have a script I didn't learn my lines I was a bad actor [Laughter] oh what's going on wait sorry tennis moment just getting a score
522. 1:22:57update looks like tiafo's up a break in the third set yeah is it really an ambush if you were forewarned that's that is now you're
523. 1:23:13getting just smart alecky Emilio's wife all right so guys are we in a regular private room we're in a general one took
524. 1:23:21us in the AI Discord in the AI Salon Discord there's a private Irregulars Channel but you just have to let me know who you
525. 1:23:30are in Discord and then I'll give you the role so for those of you that don't know what we're talking about
526. 1:23:39I've created a community called the AI Salon that we meet every other week and um we have a Discord server and we can
527. 1:23:49continue the conversation about all this AI stuff there so that link the salon.ai is a link tree that contains information
528. 1:23:58about the salon itself and our values and a link to the Meetup where you can sign up to you know join us every other
529. 1:24:07week and then a link to the Discord where you can go play with mid journey and talk about [ __ ] and ask questions it's cool it's a
530. 1:24:16neat it's a neat Community Silver Fox thank you what's that cool thing oh it's a little Tweety Bird there's a little Tweety burp it's not
531. 1:24:23called tweets anymore it's called X it's a little expert nice little expert you got there except he killed the bird he killed the bird
532. 1:24:33all right the only thing that isn't irregular about us is our tapping 36 000 likes yellow nuts thank you Stevo thank you silver fox
533. 1:24:43uh good evening Kyle Bruno what's happening what's shaking what's shaking shaking shaking wait Kyle do we have your full attention
534. 1:24:52no one ever has my full attention it's always like um it's like add is just like some sort of space antenna just [ __ ]
535. 1:25:05filtering in thoughts from everywhere uh uh Ben Shelton's about to break tiafo back in this game
536. 1:25:20Ben Shelton he's gonna be superstar mark my words that boy's gonna be good oh wait sorry the cow where's the pictures we want
537. 1:25:30pretty pictures how will we know for an AI learning lab if we don't have the pretty pictures Kyle what's up with that cow
538. 1:25:39the elephant and chat stopper LOL five items added to the list tonight oh good uh felonious thank you for doing that so
539. 1:25:48what we should do Thelonious if you haven't joined the Discord you should join the Discord and we can share the
540. 1:25:53um felonious is making like a Google sheet oh you know what we could do Thelonious if you want I can put the Google sheet on one of the
541. 1:26:01link trees or other so what Thelonious is doing is every time I show some stupid ass tool where I'm like oh this is a cool one
542. 1:26:09in in my sort of manic add way um he's been adding it to a Google sheet so he's got like a list of like 75 tools
543. 1:26:17which that's super cool so first thank you and second we should figure out a way that we can share that because if it's a
544. 1:26:23Google sheet you know we can just make it shareable and then we can figure out a place to to share it I think that would be super cool
545. 1:26:31um and that [ __ ] is going to change all the time and and you know what I'm trying to do on the channel is
546. 1:26:40through like whatever my filter is it's like like I have a decent [ __ ] meter and so what I try to do is like you know
547. 1:26:47what what makes my list here is like this is [ __ ] this is [ __ ] that I've found value in that I find Value in pretty regularly
548. 1:26:56and I'll knock [ __ ] off this list as as new stuff comes up that's better than this I'll knock stuff off this list
549. 1:27:03um this list for most of the year was three things long now it's seven and now we've got four different image tools
550. 1:27:12and the image tools quite frankly this should probably be more like six or seven but I just it's just too much so
551. 1:27:20and then the video tools are a little they're expensive and a little janky right now and the audio synthesis tools
552. 1:27:27are okay but the audio synthesis doesn't really sound oh Shelton broken so they're they're on serve in the third
553. 1:27:35set tiafu looks broken Ben Shelton looks like 20 year old confidence man dude's a beast 149 mile an hour serve come on
554. 1:27:52all right how do I get flow wise to stream the responses from bot press janky is a great way to describe it all
555. 1:28:00uh Alexander rude I don't know you're you're down in a level of tools a level below what I'm dealing with so flow wise
556. 1:28:08to stream the responses from bot press I have no idea I have no idea I can I can do automations in zapier
557. 1:28:18I know a handful of the tools but I'm not like I'm most of what my time is doing is is I'm the CEO of a video
558. 1:28:26technology company so so I'm not doing any of this AI [ __ ] full time other than how I'm applying it to my business this is all
559. 1:28:34this isn't a hobby for me like this is a what do they call it an avocation like this is something the stuff the reason I'm doing this is
560. 1:28:42all of this stuff feels very very important to me um I like I feel like in there's there's been a handful of times in my life where
561. 1:28:51I can't not be doing this like I have to be actively engaged with this [ __ ] every single day because it's critically
562. 1:29:02important critically [ __ ] important that's why this channel exists is you know if I can get one person to go to chat.openai.com
563. 1:29:12type in some [ __ ] little thing where they're like oh this is stupid it's just like Google do it and then they have their Kevin
564. 1:29:20McAllister moment there their home alone moment and then they realize oh [ __ ] these are different something's different oh crap
565. 1:29:30what do I do thank you togas oh togas reached level 13. banger is leveling up people people are leveling up in here
566. 1:29:40I don't know if you you're competitive right why would you not be competitive oh right because everyone gets a
567. 1:29:46participation certificate no not in Gen X we didn't we got trophies you know who got Trophies the winners
568. 1:29:54you know what we called the people that didn't get trophy loser that's how we played we got slapped around
569. 1:30:13I think we need to go back to the little winners and losers all right totally surprised turning soprano the
570. 1:30:23books are closed my husband asked me why do I watch this stuff you'll never use it I beg to differ okay Don A Smith
571. 1:30:32kudos to you 1956 I'm 1965. listen he's wrong he's wrong you will use this here's the deal about AI you're already using AI it's in
572. 1:30:45almost everything we've been using AI isn't new it's been around for decades ask Pate who's in here tonight like he's
573. 1:30:51he's working on the sort of core underlying models and making that [ __ ] more efficient like the reason this [ __ ]
574. 1:30:58just works is because of people like Pate that have been doing this for [ __ ] decades what's new is the chat GPT piece the piece that
575. 1:31:08makes these tools available to like regular human beings these things are janky pieces of [ __ ] right now but they're not going to be
576. 1:31:16for long like like in the next 18 months what these things do is going to get way more sophisticated and where these
577. 1:31:24things live is going to be everywhere it's going to be features inside of everything we touch and so what you're doing Don is
578. 1:31:37by being curious about this and understanding what these things are and playing with them even if you're just
579. 1:31:42saying help me do my garden help me figure out some clever stitch for my sewing Club help me organize the potluck
580. 1:31:52dinner for the school it doesn't matter whatever you're doing with this stuff anything that you're learning right now
581. 1:31:59is going to be a resource to you as this stuff gets more and more sophisticated when it shows up in all of the different
582. 1:32:06tools that we use you're going to know what to do with it you're going to know how to do it you're going to know why
583. 1:32:11it's cool you're going to know why it's powerful so kudos to you he's wrong rock on sister T minus eight minutes who's in y'all are
584. 1:32:21great that was very sweet Sheriff the the uh the Hat the Hat barrage was awesome can AI with machine learning develop
585. 1:32:31self-awareness and become sentient it's an interesting question Zen in it yes okay it's 9 33 p.m apparently serious just so jealous of me
586. 1:32:46um Here's my thought on sentience and AGI and all that sort of stuff um I think at this point it's it's a matter of semantics and and I I
587. 1:33:03think ultimately it it will be way more than just a matter of semantics but if you understand how chat GPT works if you understand how
588. 1:33:14the probability engine you know that that is the way that these tools generate what they generate it is mathematically not sentient it is
589. 1:33:28mathematically a calculator so when you type your prompt what the machine's doing is turning your prompt into a series of numbers and in a
590. 1:33:38weighted probability that goes and looks in this crazy ass mathematical Vector space and says some point in this thousand dimensional
591. 1:33:49mathematical space has the highest probability of being the next word or the next phrase based on the prompt that you wrote
592. 1:33:59and then it just does that over and over again so it literally doesn't know what words it's putting on screen it's
593. 1:34:06basically finding a number in Vector space and then just associating that number with the word that's tied to it
594. 1:34:12and then that's the word that it puts on screen where that's the phrase that it puts on screen and it does that over and over again
595. 1:34:19it is not aware of what it wrote to it it's just a series of numbers in a sequence based on probabilities it's
596. 1:34:28probability engine it's conceptually super simple it's [ __ ] magic right it's so sophisticated that what it produces
597. 1:34:40you look at it and you're like this thing has to be sentient it has to understand that but but what's that
598. 1:34:45what's actually happening here is it's been trained on all of what we have put on the internet for the past four
599. 1:34:52decades or five decades not all of it but as much as they could get their hands on and so this is what what these things
600. 1:35:01are right now is they're like a mirror of our Humanity we've put a bunch of language into the world these things
601. 1:35:08have ingested that knowledge and created all these semantic pools mathematical clusters of what all the words mean and then when
602. 1:35:19we interact with it it kind of responds back to us kind of like a human because we trained it on what we've done so this is a mirror of us
603. 1:35:30so right now they're not aware they're not aware of what they say of of what they do now could they be sure because
604. 1:35:40so one of the reasons these things hallucinate is you ask it a question and it's just pulling from all of the [ __ ]
605. 1:35:47that was on the internet so it's pulling stuff that was wrong as much as it was right and it's it's some amalgamation of
606. 1:35:53what you asked for is is what gets put back and sometimes it lies there's not a layer right now there's there's not a
607. 1:36:01there's not a cycle where these large language models actually analyze what they gave you and score it
608. 1:36:10so they don't say is this correct did I lie about these citations there's none of that right now as they start adding in things like
609. 1:36:17memory so like imagine if chat gbt remembered every conversation you've ever had with it and you could say hey chat GPT remember when
610. 1:36:25we talked about that obscure card game that had weird rules oh yeah that was this game on this date here's what we
611. 1:36:32talked about what do you want to know about it then it starts to get a little interesting when it starts actually being aware of
612. 1:36:40what it wrote and analyzing what it wrote and adjusting what it wrote based on its knowledge of what you like and what it
613. 1:36:49likes and like you could see this getting way more sophisticated in the future so right now it looks sophisticated but
614. 1:36:58it's mathematically a calculator in the future there's still going to be a lot of math involved but it's going to
615. 1:37:04start to have awareness and memory and reasoning and things like that that's where you start to get into things like
616. 1:37:09AGI here's why I don't think it's super relevant from where I sit just as a business guy an entrepreneur
617. 1:37:18that just wants to use these remarkable tools these things are already dramatically smarter than I am more capable than I am
618. 1:37:28in in many many areas and like to some areas I have no interest in learning it other areas it's like I'd like to learn
619. 1:37:37that but I don't have to now it's all there so you know I suppose at the point that they become they become sentient then we
620. 1:37:47have conversations like well is it ethical to shut one off if it's aware of its own existence I don't know
621. 1:37:55it's still a machine that that we're prompting but there may be a point at which it is so sophisticated that we're like
622. 1:38:04it's kind of human I don't know I don't know but but right now these things are already significantly more profound than
623. 1:38:12anything I've ever touched in my life it's completely making me rethink how I engage with computers and how we
624. 1:38:19will engage with computers like completely like [ __ ] 58 years of how I do problem solving and creative thinking is completely different
625. 1:38:32in nine months and and just I'm telling you that the minute the minute you cross the threshold from oh this is a cute little tool to
626. 1:38:42holy [ __ ] it can do that the minute you cross that threshold this is red pill blue pill you can unlearn it you can't
627. 1:38:49unsee it and you can't describe it to anyone why people who are into AI sound so [ __ ] insane is that they're trying to explain
628. 1:38:59something that's unexplainable unless you are in it and see it and experience it and then you go oh I get it now
629. 1:39:08that's what this channel is is like people come in here and like who's this [ __ ] weirdo oh okay I'll go to Tia
630. 1:39:16gbt whatever you know and judgments is whatever I'll call it oh my god oh that was rather impressive well but it can't do what I do oh oh
631. 1:39:31well let me see if I can really do what I do oh my oh my should I be oh my what do I do now oh until you have that moment
632. 1:39:45how do you explain that to someone I had I had an epiphany with a computer I had a breakthrough I realized that
633. 1:39:53what I do is now it's not going to be done that way anymore well how do you know that I just know because it doesn't
634. 1:39:59like I don't know it like I don't know what to do with this information that's where we are right now nobody knows what to do with this
635. 1:40:07information I don't know what to do with this information other than [ __ ] talk about it incessantly two or three hours a night
636. 1:40:15and and learn every [ __ ] thing I can so you know Don with her husband going yeah you're never gonna use this [ __ ] Don
637. 1:40:25Don's like uh yeah well okay keep watching the tennis match there hubby six six in the third they're in a tie
638. 1:40:38breaker in the third set this is a good man it's very exciting I'll ask me some U.S Open tennis all right
639. 1:40:47thanks for the update Kyle about data privacy obviously early days in the uh in the matter for small sized businesses yes uh then maybe okay
640. 1:40:59oh Tick Tock just jumped me to the bottom oh Tick Tock For the Love of For the Love of All That is programming
641. 1:41:08fix the [ __ ] bug here's the problem with the bug that drives me crazy is I'm I'm sure I'm the only one that
642. 1:41:16complains about it I don't think anyone else runs into this bug so they never gonna fix it never never well that sounds very cynical and
643. 1:41:25negative Kyle have you ever considered trying a positive outlook on life oh I've tried it but I've been on the planet [Applause]
644. 1:41:37show and tell time Mr Shannon all right well what do you want me to show and tell I will show and tell whatever you
645. 1:41:43want to say you just made my night thank you very much ask pie if she thinks your bow would help you if you were in jail cabrino
646. 1:41:53that's that's a solid that's a solid bitter response I like it I fully support the bitterness bitter you betcha
647. 1:42:02that's chat gbt I'll leave your boyfriend who's in jail because you deserve better it's getting Saucy in
648. 1:42:08here I like it all right you just made my night um what AI services are worth paying for okay um as as few as possible
649. 1:42:21for one thing because there's a lot of free [ __ ] out there chat gbt I paid 20 bucks a month for it I think it's worth it GPT 4's best code
650. 1:42:29interpreter the advanced data analytics is [ __ ] remarkable uh plugins I could take or leave I don't really give a [ __ ]
651. 1:42:35about those claude's free right now but they rate limited so at some point that might be worth paying for but given the
652. 1:42:43choice between paying for that one or Claude if they charge for it it's going to be chat GPT for now Bard is free um Bing is free
653. 1:42:56um lab stop perplexity is free the Llama models are open source you can build [ __ ] with them and you don't have to pay for them
654. 1:43:02um Pi is free po is free but you can you can pay for access to some of the more advanced other models um I wouldn't pay for Poe
655. 1:43:12all the all the rest of the shit's free I pay for mid-journey absolutely worth it it's the best image generator out there ideogram is free
656. 1:43:23Leonardo is is worth paying for in my opinion if you want to make a lot of images and you don't want to make images inside of Discord
657. 1:43:33Shelton's up 4-2 in the tie breaker um clip drop all those things are free but I I think I think there's also some
658. 1:43:43sort of paid thing there so I pay for mid-journey I pay for Chachi PT I pay for um 11 Labs because I wanted to
659. 1:44:00synthesize my own voice I pay for D Dash ID which makes the animations of the voice so you can you can synthesize your voice then write
660. 1:44:11a script and it'll talk like you sort of and then you can put it behind a picture and it'll talk like you with did so I
661. 1:44:18pay for that I pay for zapier which is an automation tool and I pay for Google collab so I can run virtual machines if I want to geek out
662. 1:44:27and that's I think it right now um and they're listen there are probably others worth paying for as I said to the Thelonious reminded me um
663. 1:44:43I wouldn't pay for anything on an annual basis just go month to month because the shit's changing too fast
664. 1:44:50um what should I be showing telling let me let me jump to the bottom and see what what everyone's saying I should show and tell
665. 1:44:56holy [ __ ] I was really far behind um felonious I'm an addict hi my name is Kyle I'm into AI welcome Kyle I've been into AI since I
666. 1:45:12discovered chat EBT on November 20 November 30th 2022. yeah we know Kyle us too yep yep yep every day I think about Ai and I
667. 1:45:23can't do anything else I don't sleep anymore yeah welcome to the club Kyle all right everyone else excited oh yeah we're
668. 1:45:31thrilled why are we so low energy we're [ __ ] exhausted oh thanks for the room thanks for the penis nose fantastic
669. 1:45:45oh and Shelton just whiffed one five four Shelton Over tiafo in the uh tie breaker oh it's a quick man what do you want me to me to tell
670. 1:45:58there people um I do 13 hours by myself what's the best place to learn prompting uh Ann Rando okay 10 out of 10 on the name
671. 1:46:13I am Rando that's brilliant okay so prompting if you're just getting started go to that document prompts.chat it's got primers in there
672. 1:46:30um that that'll teach you about the um large language models how they work how to talk to them it's also a big long list of like personas
673. 1:46:41for how um how you can make Chachi BT act like different kind of experts basically it's a good starting point it goes way
674. 1:46:53way deeper than this I would say just go go to Reddit and go to Twitter you know X and search for chat GPT prompt cheat
675. 1:47:04sheets there's like a billion cheat sheets out there some of them are good some of them are not good the the best
676. 1:47:11way to get good at prompting is just start playing and and here's the biggest shift mental shift that I can give you five
677. 1:47:21six in the tie breaker it's intense um Google has trained us that what you should type into a prompt box is the
678. 1:47:32minimum amount possible right just give me what you need and I'll go find you your [ __ ] these things need much more context the
679. 1:47:40the more context you give them the better they get so think of a think of an interaction with chat GPT not like
680. 1:47:47you're prompting and getting an answer but more like you're starting a conversation here's the kind of thing
681. 1:47:53I'm trying to solve here's what you're good at here's what I'm good at here's the piece I don't know here's my problem
682. 1:47:59six six in the tiebreaker holy crap um so um they do have a bug on other lives but maybe a smaller reaction I don't know
683. 1:48:10what that's about what was the free image prompt site mentioned earlier the free image prompt site that was ideogram
684. 1:48:19so ideogram dot AI is free and it's the one that does uh text these are all ideogram images that have the AI learning lab words like
685. 1:48:34the word on The Silo Isn't that cool I mean come on how cool is that I didn't Photoshop that that ain't that ain't no
686. 1:48:42work I just said AI learning lab silo and it did that these were all posters so ideogram so that's free for now it
687. 1:48:52won't be for long but it's free for now oh tiafo's up seven six young young Shelton young Ben Shelton he's he's up against it
688. 1:49:05this is this is set point for tiafo and he's serving hang on oh Ben just nodded to his his uh box it's gonna be okay
689. 1:49:20let's see if it's okay Ben tiafo bang oh whiffed it did he win oh no it was long second serve
690. 1:49:34was that in holy crap what a return what a return right oh my God he [ __ ] smashed the second serve 7-7 in the tie breaker holy crap
691. 1:49:49all right I just uploaded two coloring books to KDP very cool Sherry Banks 22 years versus two weeks what's that the 22 years versus two
692. 1:50:03weeks thing because I just wrote the uh I just wrote the uh the uh I just did a video on the 10 000 hour rule that I
693. 1:50:11think the 10 000 hour rule is dead right so instead of you having to spend four years getting good at something you
694. 1:50:17can spend like you know four months now or two weeks right big rally big rally seven seven oh tiafo wants it into the net
695. 1:50:29eight seven for Ben Shelton serving for set number three all right let's see what time is it what time is it it's 10
696. 1:50:42minutes to ten good night Kyler regulars and and aliatory promptlines goodnight took us do we need more data um
697. 1:50:53they're saying that we need more data and then there's there's kind of weird conflicting things about is synthesized
698. 1:50:59data going to cause model collapse or not it seems like synthesized data when used right can be quite powerful
699. 1:51:06and it also looks like oh oh wait wait for it Ben Shelton set number set number three two sets to one Ben Shelton
700. 1:51:18Francis tiafo's number 10 in the in the world people are losing their minds this kid's a [ __ ] star this kid's a [ __ ] star amazing
701. 1:51:30um do we need more data um I just saw a thing that um so Chachi PT the way chat GPT is trained is through what's called human
702. 1:51:40reinforcement learning so they have Chachi PT or gpt4 generator response and then humans go thumbs up thumbs down
703. 1:51:48here here's what's good about it or here's how I'd rewrite it right they're now getting AI large language models to do that
704. 1:51:57instead of humans so the AI robots are doing the human analysis as good or better than humans are way faster
705. 1:52:09so this is now that thing that people talked about well what happens when the robots start training the robots and the
706. 1:52:15and they just get sort of exponentially smarter faster so so I don't know like I don't know where the boundary is of do we need more
707. 1:52:27human generated data or do we solve the problems of model collapse with synthesized data and figure out ways to use synthesized data
708. 1:52:36so that these things just get better and better without collapsing I don't know I do not know it's wild um
709. 1:52:49oh which service did he say he paid for for VMS for voice for voice stuff 11 Labs so eleven Labs Dot here you want to
710. 1:52:58say uh uh Emilio's wife wanted Show and Tell let's go to a show and tell 11 labs elevenlabs.com generative voice AI
711. 1:53:16and then let's go we'll go to chat GPT and we're going to go to gpt4 because it's the best just normal and we're going to say um
712. 1:53:38write me a soliloquy of a sad zombie turtle that is lamenting the fact that the Waterfront Carnival
713. 1:53:52is ending tonight in the dimming light where water meets the shore the Carnival's vibrant Colors cast a glow transient and fading how
714. 1:54:08cruel time can be that's good we'll grab that okay so when I say play with these tools seriously like just put in stupid [ __ ]
715. 1:54:19just ask it to do stupid [ __ ] and then it what what invariably happens is you realize that these things are just way deeper
716. 1:54:30than you than than you think they could possibly be and it'll just write something and it a it gives you Joy because it's just like
717. 1:54:38this spontaneous serendipitous stupid little piece of [ __ ] um but anyway okay so I'm going to take that dumb thing that was written I'm
718. 1:54:46going to paste that in so I'm I'm in 11 Labs so you go into voice lab and that's where like I recorded um like a minute and a half long
719. 1:54:57thing that I had chat GPT right I told chat GPT I was gonna uh use a voice synthesis you know I'm going to train a voice synthesis model
720. 1:55:07and I wanted it to write me a script to use the words you know you know as many diphthongs and phonetics as possible um
721. 1:55:20Ben Shelton's just smiling to the crowd he's having a blast he's up two sets to one that's [ __ ] awesome um
722. 1:55:30and so I took that script brought it in here read it and then said synthesize my voice and I actually did two different
723. 1:55:36readings of the script I did one that was like you know I was talking kind of like a professional voice actor and I
724. 1:55:42you know put a little gravity in my voice and then the other one I was talking like you know I was talking for
725. 1:55:46a Pixar movie it was more bright and so I have one called Pixar bright and I've got one called Pixar normal
726. 1:55:54and so since this is sort of a dark Soliloquy we'll do Pixar normal so that's the one with the little lesson
727. 1:56:01voice so now I'm just gonna say generate in the dimming light where water meets the shore the Carnival's vibrant Colors
728. 1:56:11cast a glow transient and fading how cruel time can be for even a creature like me Undead and unhurried feels the
729. 1:56:19tightening grip of fleeting Joy a turtle now a ghastly reflection of Life Bound by a cycle unknown seeking Solace where
730. 1:56:27music and laughter once danced upon the waves all right so that's 11 laps that's not bad actually that's actually pretty
731. 1:56:32good okay so we're gonna download that file so now we have a file of me talking sort of of words I sort of wrote
732. 1:56:42and then we're gonna go grab tokas if you're still here we're gonna grab we're gonna grab your turtle your copper Turtle dude here
733. 1:56:52so let's export this where's it exporting it we'll go desktop that's fine that'll get downloads because that's where the other thing is yeah
734. 1:57:06all right and now we're gonna hop over to a thing called D- ID or digital identity is what I finally figured out
735. 1:57:16I'm like why is it D Dash ID d-id.com okay Ben's really excited about all this
736. 1:57:32um and I'm going to log in so this thing I pay for this is one of the things I pay for so I paid for 11 labs and I pay for this
737. 1:57:47oh switch to our exciting new editor today amazing take me there so remember when I said everything's a janky piece of [ __ ]
738. 1:57:57so looks like did just got a little less janky so you now have a presenter you can now put text on screen
739. 1:58:06you can now add a background oh this is really interesting so you've got AI presenters so so these are people that they've created
740. 1:58:17then these are people that I've created and and some of the ones they've created and some of them are on green screen
741. 1:58:23backgrounds and they're HD high quality but we're gonna add the turtle dude wait we could not identify any faces in
742. 1:58:38the photo oh that's dumb all right let's go grab the other one we'll grab this dude we'll see we'll see if it can
743. 1:58:44recognize this face hopefully it can um
744. 1:58:59that's actually interesting that it it does an analysis of the image if it can't find a face it it won't upload it
745. 1:59:06no it couldn't find it there all right so I'm just going to grab another one so let's see we'll we'll grab we'll grab
746. 1:59:12this dude so this is my Genesis image of Kyle Shannon dreams it's interesting that it's putting it in a horizontal video
747. 1:59:24and let's throw a background on there that looks cool and we'll kind of zoom in on me this is all very self-serving isn't it
748. 1:59:37okay and then I'm gonna now what I'm gonna do I don't know how to do this listen upload audio okay so I'm gonna upload audio
749. 1:59:49so I'm gonna upload that 11 Labs thing that we just created in the dimming light where water meets the shore okay and then we're gonna say
750. 2:00:00generate video so in the upper right hand corner generate video uh uh let's go
751. 2:00:17wait what's voice script to do is it generating the video
752. 2:00:34I do have a paid plan all right all right it's not doing it well anyway I guess if you have the paid plan let me go back to home
753. 2:00:47so I'll just show wait oh that's a draft yet so here's one I made ah dear Adventurer it seems your path has led you to A
754. 2:00:58peculiar Place indeed as the master dungeon master I invite you to explore the world of Dungeons and Dragons where you'll encounter so
755. 2:01:08so this is like a couple of things going on 11 Labs right now is still got that kind of monotone thing but one of the
756. 2:01:16things 11 Labs has is they've got different kinds of inflections and emotional states like you can you can do
757. 2:01:24one that's angry you can do one that's a whisper you can do one that's sexy you can do one that's funny what we're going to be able to do is
758. 2:01:33create that those emotional tags within the context of a single paragraph right so and and that's going to start happening
759. 2:01:42automatically so the monotone [ __ ] that we hear that's going to start getting much better the the animation here it's mostly good
760. 2:01:51ancient artifacts and unraveling but like right there the goggles blinked right so it assumed those are eyes and
761. 2:01:58they're actually goggles so it doesn't make sense that they blinked but that'll get better right and it's there's some
762. 2:02:04some weird animation how it isolates the head and things like that but to be clear I created this image in stable diffusion
763. 2:02:11I uploaded it to this thing you know I put in a synthesized soundtrack and I hit build video and this is what it
764. 2:02:19generated right so so we're we're already at a place where these tools are quite sophisticated without you having
765. 2:02:28to be in the visual effects business so imagine 18 months from now what this tool is going to be it's gonna be [ __ ] stunning
766. 2:02:38right so anyway there's a little show and tell all right thank you Emilio's wife for reminding me to not just [ __ ] Babble fantasy all
767. 2:02:50he does is talk and talk and talk he loves to hear himself talk there's a hundred people in here it's 10
768. 2:02:56o'clock at night Do You Know Where Your Children Are I don't oh right they're right upstairs damn it crypto Joe thank you for the cruise
769. 2:03:07oh Ben Shelton breaks tiafo in the first game of the fourth set holy crap star being born
770. 2:03:21if Ben Shelton beats I don't know if he's in the same side of the draw with uh with what's his name with the Spanish dude
771. 2:03:32but I could this is this is something this is something all right I worry the cheat sheets are terrible a
772. 2:03:45lot of them are pretty terrible I'm pretty good at it I want to be great at it oh so Anne Rando um I I think the way to get great at
773. 2:03:54prompting is well it depends how your brain works the way my brain I'm I'm a non-manual reading you know mess
774. 2:04:08so for me taking a systematic approach to it doesn't work and and that's what a lot of those prompt sheets assume is
775. 2:04:15that you're going to take a systematic approach and you know exactly what to do when the thing that helped me get good at
776. 2:04:23prompting I mean if you're good at writing you're going to be decent at prompting right so so if you've got a
777. 2:04:28facility with the English language that helps but even if you don't rather than saying I want to get good at prompting
778. 2:04:39flip the equation and say what's the problem you're trying to solve and then start to think about if I were
779. 2:04:46talking to a clueless intern who was really well educated so let's say you just hired an intern and they're [ __ ] clueless about the
780. 2:04:54world they've been locked in a closet just memorizing all of the knowledge of the Earth for the past 20 years and now
781. 2:05:01you've got them they're working for you they don't know [ __ ] about [ __ ] but they know everything okay
782. 2:05:06so thank you Nature's bull whatever the BL stands for or bi thank you very much for the Roses um so imagine you've just hired that
783. 2:05:15intern and you've got a problem to solve how would you how would you talk to an intern to get them to solve that problem
784. 2:05:22for you well you'd probably have to set some context okay intern so today you're going to act like a a stock broker and
785. 2:05:31um I'm gonna give you a problem and we're gonna talk about all this you know these companies that I like and you're
786. 2:05:37going to help me figure out which one I'm gonna do do you understand yes I understand and it'll probably give you some stats about
787. 2:05:44stocks that it knows and you'll be like okay great Savant you're you're brilliant but still clueless and so the minute you start solving
788. 2:05:53problems using these tools that's when you'll find your prompting just gets better because what you end up in is these fairly sophisticated
789. 2:06:02conversational back and forths with these with these machines where you'll get it close and and it'll give
790. 2:06:09you something and you're like oh that's close but it's not quite right because you missed this point and that point in
791. 2:06:14that point and it'll apologize oh you're right I'm sorry I missed that point here's here's some stuff okay now we're
792. 2:06:20getting closer right that back and forth over time you just get better and better and better at understanding okay for
793. 2:06:28this kind of situation I can do a direct single prompt and get an answer back and it'll be fine but for
794. 2:06:34this kind of situation I need a fairly sophisticated back and forth so for me it's kind of just becoming facile enough with with
795. 2:06:42the interactions with these where you kind of know here I need to dig deeper here I can here I can do it really quick
796. 2:06:48and so that that's the thing you can still take those cheat sheets and understand here are the mechanisms right
797. 2:06:55you can do you can give it an example if if you don't know examples uh and is um you could take an example of writing and
798. 2:07:06tell it you want you want it to match the tone of some writing or you could give it something that's very
799. 2:07:11specifically formatted like Json and say I want you to match this kind of Json so the other thing you should do if you
800. 2:07:18haven't played with it is explore custom instructions in chat GPT so in check EBT so we've got our new chat here there's a
801. 2:07:29thing called custom instructions which you have to enable tiafo three big serves tiafo's pissed okay
802. 2:07:44um so go to custom instructions and then what custom instructions are I'm going to turn these on is you can tell it this is fine tuning
803. 2:07:52for chat EBT so you're basically telling it to behave like a certain kind of character and and respond in a certain
804. 2:07:58kind of way and it does it for all of your prompts so over time what I what what I would imagine that that one tool
805. 2:08:06that that someone showed us here earlier where it's you've got multiple versions of these things I could see you having
806. 2:08:13lots of different versions of these but this is this is the one I wrote you're you're an overbearing pompous ass who
807. 2:08:19uses large words even though you don't understand that even you don't understand as a way to intimidate those
808. 2:08:25you to talk to and and then how do you want it to respond whenever anyone asks you anything I want you to use an
809. 2:08:32obscure literary reference to belittle their idea and then redirect the conversation to talk about yourself okay
810. 2:08:38so I'm going to save that so now no matter what I say to chat GPT it's going to respond using that imperative right so I'm going to say um
811. 2:08:48why do turtles live in ponds question mark fairly fairly straightforward question
812. 2:09:04and then Chachi PT ah the pond the pond is a domicile for Turtles the bucolic equivalent of the balls the the the
813. 2:09:15balsatian provincial town where everyone knows each other's business but I digress the true inquiry is why I a
814. 2:09:24Marvel of computational linguistics and programmed to parse even the most ineffable of human questions but to
815. 2:09:32assuage your curiosity Turtles favor pawns primarily due to their ecosystem which provides them with the necessary
816. 2:09:40resources such as food shelter and a breeding environment now let's return to the wonder that is me capable of
817. 2:09:47elucidating the enigmas of the natural world with the flicker of an algorithm so you can as silly as this was you can see
818. 2:09:57how powerful this is that that you can absolutely Bend this bot to your will right and you could get it to write
819. 2:10:05really good emails or get it to write really good outlines of books or or whatever you wanted to do by doing this
820. 2:10:11so again the big trick with prompting from where I sit is this is a conversation and the better you get it at having an a back and forth
821. 2:10:20conversation with these tools the better you'll get at it all right all right I hope that helps I hope that helped GPT has supercharged My
822. 2:10:30Philosophy research holy crap I mean we are large language models and these tools are large language models like
823. 2:10:36philosophy stuff the uh if you haven't seen the Jordan Peterson video where he talks about Chachi BT like this this is from like back in
824. 2:10:46February he's just let me see if I can find it it's well it's it's a little long but he but he basically goes have you heard of
825. 2:10:53this chat GPT and and he just talks about he he fed it he asked it if it knew 12 rules for life his book 12 rules for life
826. 2:11:04and then he had it right the 13th rule for life and he had it compare like use two different philosophical imperatives
827. 2:11:13as he wrote as it wrote the 13th chapter or some it was some bizarre like Byzantine Jordan Peterson intellectual
828. 2:11:20thing and he said it did it perfectly perfectly it was super enunciation I write long paragraphs yeah the long paragraphs can
829. 2:11:32be good the also experiment Ayan Rando with um with not trying to write the perfect prompt to get a single response but like lots of little
830. 2:11:47um back and forth um because both of them can be good talk to me like you're an AI expert well give me a spreadsheet yes you can do
831. 2:11:58that um and it will the the advanced data analytics it will give you a spreadsheet I'm ready to give all the
832. 2:12:05data required what does X stand for in the new Twitter name I think it's just he he likes the name X he named one of
833. 2:12:13his kids X or it's he's got X in the in the name named after a a plane um uh so so I I think what this is so so here's here's
834. 2:12:25here's why it's X hang on I've not had much water tonight I'm a little parched if you know what I mean hmm
835. 2:12:36uh it's tasty it's very tasty Rocky Mountain filtered refrigerator water tasty um [Applause] X the the X in it so so here's the thing
836. 2:12:49what what Elon Musk wants Twitter to be is what he calls the everything app so in China one of their apps I forget which one it
837. 2:13:00is it's it's their Communications tool it's their shopping tool it's their social media tool like it's it's all these things
838. 2:13:09so he wants Twitter to be you know the new version of PayPal he wants it to be Twitter he wants it to be you know where
839. 2:13:18you talk to people on the phone instead of using a phone you'll talk through that he wants it to be AI he wants so so
840. 2:13:26he wants it so X stands for everything right X is a variable in algebra so X is all of it so he wants it to be the
841. 2:13:34everything app so I think that's where he's headed and that's why x dot a i is the AI company and x.com is the app and
842. 2:13:43then the AI company is going to feed the app and make it super powerful and super magical so that's that's the theory I
843. 2:13:51suppose I am now 99 uh comments behind I can't believe it ideogram you can't use do's and wait you can't use does have a don't use
844. 2:14:09commercial I don't know quite what that means Sherry Banks clip interrogator if if you can't prompt it's a reverse way of prompting yep you
845. 2:14:22upload an image and get a word that's actually really cool let me show you some Discord um this is Discord
846. 2:14:28and I'm going to go into a channel called mid-journey play so so I'm in the um the AI salon this is our little icon that's the salon
847. 2:14:40these are the channels so so here's this was something who did this one this was I'm Claire that was a cool thing that she did um
848. 2:14:51so one of the things that you can do so the the way you make an image in mid-journey you're in Discord you type
849. 2:14:58slash imagine so this is like early days of the internet where you know you're making images with command line prompts so
850. 2:15:06we're making a command line prompt here and uh we're gonna say uh I'll just do my my sort of classic 1970s
851. 2:15:18muscle car I'll do a different location on the beach at Sunset um yeah that's good I'll just just do that
852. 2:15:34and so it's going to go off and it's going to make this thing now but one of the things that you can do
853. 2:15:40here there's a lot of things you can do with mid Journey one of the things you can do is something called describe so
854. 2:15:45you can upload an image to it and have it describe what it sees and then you can use it gives you four
855. 2:15:52different descriptions and then you can use the description of what it saw to make a new image right so it's a way
856. 2:15:58that you can start with an image as inspiration but because you're translating it from an Inc image into a
857. 2:16:07prompt and then back into an image you can get really really creative with it so there's my there's my little cars
858. 2:16:15so those are cool so let's go I don't know let's make we'll upscale that first one so these these U's and V's these these mean
859. 2:16:27upscale and the v's mean very very um and then I'm gonna actually do a zoom out 2x so this is a new feature um we'll do custom Zoom
860. 2:16:40so I'm going to do I'm going to zoom out and make this a 16 by 9 image and we're going to do it a two times zoom out
861. 2:16:52so this is a new feature this in uh in stable diffusion this was called out painting in Photoshop this is called generative fill
862. 2:17:02oh wait so Ben Shelton if he wins tonight plays Djokovic in the semis holy [ __ ] Djokovic will wipe the floor with
863. 2:17:11him but that'll be a good match um all right so so it's kind of zoomed out on a car see how far away it is now super cool
864. 2:17:23and let it let it finish its the the one thing you'll notice if you haven't watched these image generation tools do
865. 2:17:29what they do you'll notice it starts out blurry and then it gets less blurry and less blurry and less blurry it's
866. 2:17:34literally what it's doing is it's taking sort of multiple like hundreds if not thousands of images that have been noised they've
867. 2:17:43run them through a noise algorithm that that turns them from an image into like random noise patterns they then
868. 2:17:49amalgamate them all together into sort of an average noisy image and then they denoise that over like 50 steps and
869. 2:17:57that's how you generate these these uh these images so I'm going to upscale number four all right there's our image so super
870. 2:18:13cool car on the beach all right it looks like a Mustang actually it looks like a 70s Mustang like a 70
871. 2:18:21or 71. so I'm going to save this image out um that's good and now here's what I'm going to do I'm going to go slash describe
872. 2:18:35and now I get to drag and drop an image so I'm going to go grab that image and I'm going to pop it in here so
873. 2:18:45you're like but Kyle you just generated it there why are you popping it back in because we want to do a different thing
874. 2:18:49we want to describe it so it's not looking at that image and here's what it generated HD
875. 2:19:03wallpaper the old car on the beach in the style of realistic and Hyper realistic detailed renderings crimson
876. 2:19:10and bronze digitally enhanced Cobra 1970s correct the car is on the beach at Sunset and the style of a realistic
877. 2:19:19hyper detailed so so there's four different descriptions here right it understands the aspect ratio um it's saying this is like Mike campau
878. 2:19:30and Herb trimp or Mike campau and Ford Maddox Brown um so what I can do here is I can render um any one of those into a new image or
879. 2:19:43I can say imagine all and it's going to basically make four different images based on the four different descriptions that it wrote
880. 2:19:51how trippy is this and this [ __ ] nuts it's I I just I like I I can't I don't know how to not be enthusiastic about this [ __ ]
881. 2:20:07it's this this is [ __ ] magic have you ever tried to take a picture like I'm a photographer I can't illustrate for [ __ ] I can take
882. 2:20:16pictures okay I can edit pictures okay but like if I wanted to do a picture of a Mustang on a beach I would search through
883. 2:20:24stock photos trying to find it and at some point maybe get frustrated and just go try to take it but like now you can just make it
884. 2:20:34and so let's see what it did here oh this is cool oh look at that oh oh cause it said Cobra right yeah this said cobra in 1970s
885. 2:20:45check these out so so because it said Cobra it didn't say Cobra Mustang so it's giving us like Shelby Cobra
886. 2:20:54and this is like a Shelby Cobra vet hybrid that's like a 56 vet with a with a cobra front end those are [ __ ] awesome it's like come
887. 2:21:07on like I want that car I want that front end on that you know car all right what's old Benny up to tiafa so 3015 tiafo serving
888. 2:21:22bends up four to two in the set holy [ __ ] okay um what else did we get here this one's super cool like it's an old broken down car on a
889. 2:21:34beach right we're now out of the 70s but still super [ __ ] cool right like that's like you stumbled upon this you
890. 2:21:42know on some deserted [ __ ] Island some rich dude had a car on the beach wow that's so cool here's another one right completely
891. 2:21:52different era and this is just based on whatever the description was right and look now here it's bright and sunny
892. 2:22:00that kind of looks like a Thunderbird you know these look like cobras that looks like some you know 40s 39 duesome
893. 2:22:09bird kind of crazy ass thing anyway all right I'm gonna jump all the way to the bottom hopefully that was cool
894. 2:22:16I don't know about you I just this [ __ ] blows my [ __ ] mind every [ __ ] time it's just like juice that's crazy it's crazy
895. 2:22:27I don't up five to two Shelton just broke tiafo third game five to two serving for the match to play Djokovic in the semis Dag
896. 2:22:42Dag Dar is born all right I'm jumping to the bottom people I'm doing it I'm doing it I gotta go to bed soon
897. 2:22:56what time what do I have tomorrow do I have anything tomorrow I don't [ __ ] know it doesn't [ __ ] matter
898. 2:23:02sleep sleep is so 80s we used to joke about that when they did agency.com back in the 90s we were working like 14-hour days we
899. 2:23:11were just like sleep is so 80s oh man starting that company when I started agency.com we did it was it was 14
900. 2:23:2314 hours a day seven days a week for about nine months straight it was the first time in my life I woke up one morning
901. 2:23:35and I felt it was the first time I had ever felt physically like oh my body could shut down it was like the
902. 2:23:45first time I'd ever confronted a level of exhaustion that was like in the ha this must be what it's like when [ __ ] shuts down
903. 2:23:55and shortly after that we we made enough money that we could hire a bunch of people but that was [ __ ] crazy man
904. 2:24:01that was some wild ass times WeChat yeah that was that's the Chinese app that does all that [ __ ] that's what that's what Elon wants to
905. 2:24:09make here in America Robert Rossi thank you for your donut I love my sugary sweets they they give my stomach all sorts of
906. 2:24:16trouble but they're tastier than [ __ ] love them all right run it through Claude 2 GPT 4 and 70 billion llama instructive and
907. 2:24:25then feed it to each other around Robin well wizard see wizard that's one of the things that I love doing is take the
908. 2:24:31same prompt and and try it in all the different models and see how different they are um it's it's really cool all right I'm
909. 2:24:38gonna give you a blow by blow on this game so first point oh tiafo wins point one with a nice shot prompt hero found it good all right I
910. 2:24:49write the idea for the prompt in depth and instructions for how to build the prompt then I wait where's the rest of that then I run
911. 2:24:57it through oh I see okay then you run it through Cloud through got it um by the way I I love I've the the
912. 2:25:05Irregulars know this I love when you guys are having conversations with each other and coaching each other and like
913. 2:25:10sharing what you know [ __ ] thumbs up that's the whole idea so like I want people not only to get curious about to
914. 2:25:18say I shouldn't play with it but also like talk about it share be generous because we're all [ __ ] trying to
915. 2:25:24figure it out so I love that so thank you wizard see wizard for be in a mensch for being cool oh drop shot tiafo oh oh the lob was it out
916. 2:25:40yes 15 30. tiafo might break him back ooh it's getting spicy there in in uh Forest Hills isn't it all right let's see Vicky I'm lost but
917. 2:25:55I'll read a manual and make it a prompt is AI the new age Ultron oh I don't know that reference why do I not know that reference
918. 2:26:13oh wait I still got my thing acting like an idiot hang on let me turn off let me turn off idiot mode um who is Ultron
919. 2:26:26I don't know I gotta do a new chat I know is Ultron I know you're making fun of me for not knowing who that is and I said altorn
920. 2:26:42oh Tron stop generating Match Point Match Point Ben Shelton he's gonna go for the 149. oh did he get it
921. 2:26:58no 135 miles an hour just wide holy crap first appearance Avengers 1968 Ultron is the sentient robot to possess a
922. 2:27:07superhuman strength speed ability so zai um wait wait sorry old uh deuce deuce all right
923. 2:27:23okay um is AI the new age Ultron AI is the beginning of it so so like Here's my thought like we're we're living in a post
924. 2:27:37sci-fi world right now where almost every sci-fi movie that we've seen with the exception of the time travel ones
925. 2:27:45like either we have that [ __ ] or it's within our grasp sorry watching this point oh teofo whiffs one into the stands
926. 2:27:59another match point for Shelton crazy um so so like like I said about the AI tools right now are janky pieces
927. 2:28:10of [ __ ] they're they're really flimsy they're not sophisticated they're they're nowhere near Ultron level um goodness but
928. 2:28:22they are the foundation of it right and good serve tiafo's on the Run he wanted Ben Shelton [ __ ] beat tiafo and he's
929. 2:28:36going like this [Laughter] holy [ __ ] hey stop Star is Born he [ __ ] wiped the court with him jeez Wild
930. 2:28:52that's crazy that's crazy um so anyway the these tools are going to be um incredibly powerful incredibly
931. 2:29:05sophisticated in a very short amount of time plus you've got Elon Musk Boston Dynamics two or three other
932. 2:29:19robot companies building humanoid robots that are basically just waiting for this [ __ ] to be there and and Elon musk's latest self-driving
933. 2:29:32um demo was not something that humans programmed it's it's an AI using camera data analyzing it in real time and learning
934. 2:29:45from the camera data of um of people driving in you know in their millions of cars they're going to be able to take that approach and
935. 2:29:56literally just transfer it into a humanoid robot and it's going to be able to learn without humans having to
936. 2:30:01program like like here this is one of the remarkable things about these tools is everything up to this point has been for the most part um
937. 2:30:11teach the robot to reach out and grab the apple and when it when it senses that it's squishing the Apple stop right
938. 2:30:21so read the sensor data so this is humans figuring out the logic for how the hand should move and and now they're just taking
939. 2:30:33sensor data and and computer data and it's just learning so yeah we're close we're we're close hang on
940. 2:30:44there and put the towel over your head what's going through your mind at the moment it was a hot one in here tonight wasn't it guys
941. 2:30:56yeah feeling like I left it all out here tonight uh emotional battle and uh things like you guys for staying so late hell of an atmosphere
942. 2:31:10and thanks for pushing me over the line crazy 20 years old [ __ ] nuts all right um Jordan Peterson outro Innovation just
943. 2:31:21landed new orders worth over 100 million wait on to Innovation just landed new orders what's on to Innovation Thelonious
944. 2:31:31Thelonious lately has been bringing the good news I'm digging it onto innovation
945. 2:31:45process control interesting announces 100 million dollars in orders system supporting Advanced packaging for AI company news for AI wow Wild
946. 2:32:02your system is run out of application memory thanks Thelonious way to crash my my browser um I'm not good at PowerPoint creation
947. 2:32:13is there any way to use AI to do that oh let's do another show and so we're going to go to tome t-o-m-e tome.app
948. 2:32:23there so so what I would suggest is so this is for Thomas Kroll um go to um futurepedia.io and look up presentations
949. 2:32:37or PowerPoints um and and uh there's going to be a lot of them Tome is one that I know is pretty good
950. 2:32:54and this is this shit's just crazy now this demo is really good um but it but it you know it isn't going to necessarily give you the perfect
951. 2:33:04thing right out of the gate but if you're shitty at writing PowerPoints so you can do a lot of things like write a
952. 2:33:12create a page about like a web page or create a presentation about so I'm going to say create a presentation about small town carnivals and
953. 2:33:27the turtles that are sold there and so it says creating outline I don't
954. 2:33:41know if you can see this it's it's one of those super hip sites so truly awesome truly awesome exploring the wonders of small town carnivals
955. 2:33:52so okay so there's our outline small town Carnival is a delightful experience the turtles of small town carnivals the
956. 2:34:00ethics of selling Turtles at carnivals the importance and responsibility of pet ownership support local communities
957. 2:34:07right so I'm going to say continue and I guess I could have gone in and edited those so those are now my slides
958. 2:34:13and so now it's writing the copy for each slide and then it's going to write it's going to generate images for each
959. 2:34:19slide so I now have a title screen totally awesome explore exploring the wonders of small town carnivals
960. 2:34:29there's there's our agenda there's our thing look small town Carnival small town carnivals a delightful Escape
961. 2:34:37look at these kid looks terrified oh my God this is awesome the ethics there's the ethics so so this is
962. 2:34:50something that like if you just wanted you know a quick little thing now the other thing you can do is this
963. 2:34:58okay what did I say here let's go back to Tome um did I have my original prompt yes no where's my prompt I lost it okay let's go to chat GPT
964. 2:35:15so I'm gonna do a new chat I'm gonna do gpt4 and I'm gonna say I want you to act as a master business presentation writer
965. 2:35:31and help me write the outline for a PowerPoint presentation that talks about small town carnivals and the turtles that are sold there
966. 2:35:50all right certainly here's an outline so slide one so it gives you an image prompt and an overview so what's it what's what's this one
967. 2:36:02doing so um small town carnivals and the turtles they nurture um introduction brief overview of small town carnivals
968. 2:36:13um history of small town carnivals the Allure of carnivals the unexpected Stars Turtles why turtles turtles as pets the business side
969. 2:36:23ethical considerations a sustainable approach turtles in pop culture the future so so gpt4 is writing a much more sophisticated
970. 2:36:34presentation outline now it's also writing image prompts for you you could take these over to Mid journey and make
971. 2:36:40your own images so you could just copy this now I can say this I could say um perfect period now please flesh out each slide
972. 2:36:51so that it's got three to five bullet points supporting the core argument and a short intro paragraph
973. 2:37:07and now it'll start writing that so title small town carnivals and the turtles they nurture welcome to an
974. 2:37:13exploration of the unique bond between small town festivities and their unlikely star attraction Turtles
975. 2:37:20intro paragraph for the heartbeat of small towns carnivals are more than just a source of entertainment they're a
976. 2:37:26reflection of community spirit and within them lies unique Traditions like the selling of turtles oh God I'm such an idiot but it's okay
977. 2:37:44because back here live at the waterfront village with my friend the zombie Jonathan you're looking good Jonathan just got an
978. 2:37:57awesome face paint job what do you think I like turtles the Waterfront Village open for the next 11. so anyway yes uh AI can help you write
979. 2:38:13presentations [Laughter] oh my lord it's so good I'm not weird
980. 2:38:28I'm irregular [Laughter] oh man all right listen I'm punching someone just said I get punchy um I'm Punchy I gotta go to bed people
981. 2:38:44um Ben Shelton new star plays Djokovic in the semis holy crap wow wow all right let's see
982. 2:39:01let me get a few more of these in here buffering Turtle presentation what site is he on oh the the presentation site that I showed before
983. 2:39:09was Tome t-o-m-e dot app app that's one that'll make you websites or things like that I personally as much as the all those sites out there
984. 2:39:23do all that sexy [ __ ] for you just the nature of how they have to compromise and try to make it good for
985. 2:39:30everyone you're going to get better results if you go to Chachi PT and learn how to do this [ __ ] yourself you just
986. 2:39:36are it's just better writing now some of those tools might be using gpt4 and they're going to get better again
987. 2:39:41understand all these tools right now janky pieces of [ __ ] so just just be be in Exploration mode be in Swiffer mode
988. 2:39:49not even sponge mode you got to go a level up from sponge mode to Swiffer mode you're you're a trademarked product at
989. 2:39:58this point you're not some generic [ __ ] sponge no we're we're absorbing on a whole different level now um just start learning this [ __ ]
990. 2:40:07um all right let's see um very little time for sleeping it's so true all right listen I'm out of here let's do our Romper Room ending I see
991. 2:40:18Thelonious I see Vicky I see Monda JT nice name um Emilio's wife of course thank you for for trying to Corral me tonight I know
992. 2:40:30it's a it's a special thing Matthias was in here with the lightning bolt thank you um Jay-Z Clark good to see you jrc uh
993. 2:40:40let's see a little bit Vicky again Stevo peace out cafruno peace good night everybody all right um as always uh don thank you very much
994. 2:40:53good night and tell your husband he's wrong you're right he's wrong peace everybody have a great night thank you Sherry Banks