aiLL 10/18/23 - Unleashing AI's Creative Potential with ChatGPT, DALL-E & More

Transcript

1. 0:14hello hello hello got the got the Buster po Dexter hair going tonight what's shaking what's happening what's going
2. 0:28down welcome welcome welcome welcome Matias what's happening good to see you sir Who's Live Is It Anyway welcome to Tik Tock live it's a
3. 0:43game show where you twoo can play and if you play it right you can make money with Clint Copperhead what's shaking
4. 0:56buddy everyone in regulars and mods PR in do oh man my eyes is a little crispy Hello Apple user what shaking lunar stick good
5. 1:11to see you Dr J I taught a uh I taught the third um the Third Leg of the AI
6. 1:28LaunchPad Workshop so tonight was our the final the final leg it was Advanced Ai and holy crap it was
7. 1:44like it was just insane like how much how much stuff I ran them through in two hours I basically told them just sit
8. 1:53back don't listen to anything just go watch watch the recording when it's done it was nuts but I'm a little I'm a little
9. 2:04crispy Amelio's wife was there wait is is Tobias still banned no you're not banned are you Tobias cuz you just posted
10. 2:18Turtles I think you're good welcome everyone oh wow there's a lot of folks in here all right so let's get this party started so my name is
11. 2:30Kyle Shannon this is the AI learning lab and uh what we do here is we talk about this generative AI stuff and uh if you
12. 2:38have any questions or comments about what you're doing with AI pop them in the comments below I'll do my best to
13. 2:44answer them we'll walk through a bunch of stuff here so I'm going to I'm going to quickly get you up to speed on if
14. 2:51you're brand new to this AI stuff and you're like but I don't know what to do I'm not very technical I'm not a
15. 2:57mathematician okay all right I'm not an engineer I I don't do programming okay so I can't do this AI stuff no wrong
16. 3:05wrong wrong you can do it we're all doing it we're all figuring it out I have a degree in acting all right so take that as your
17. 3:12Baseline um I'm also also an entrepreneur but where we're going to start if you have not played with chat
18. 3:19GPT if you haven't tried it yet you're like but I don't know what it is Kyle I don't know what to do I think this is not bright enough yes
19. 3:28correct um if that's you if you're like I don't know where to start start there that one chat. open.com that's the official chat GPT
20. 3:38website it's it's the first and it's the best and you can make money with uh we make fun of those people um but
21. 3:50anyway um go there and just start talking to it it's it it looks like a Google search box and but it's not tell
22. 3:58it ask it things ask it to write your poems ask it to write you a business plan ask it to write you come up with
23. 4:04marketing copy for your pizza shop doesn't matter just go go there and start asking it stuff and watch what it
24. 4:10does um if you've been playing with this stuff for a while um what I've got here is a list of the things I think are
25. 4:15worth paying attention to these are the big four that are right here we call it the adult table and um these are all either very very good
26. 4:25now or about to get very very good Bard right now is okay not great um chat GPT now that it's gone multimodal and has
27. 4:34Dolly built into it is insane it's it's amazing it's the it's the it's the best 20 bucks you you could spend on a for a
28. 4:44monthly subscription for software um claud's really good Bing has potential um but it's not great uh cuz
29. 4:53it's they've got the safety guard rails too high then we've got the kids table piie perplexity AI where you can play
30. 5:01with the meta meta's llama models llamas is tasty and po um so go check those out now for images do a screenshot of this
31. 5:11one so Dolly 3 um you can play with inside chat GPT if they've given you access to it if not it's coming if if
32. 5:20you pay for chat GPT plus um but you can also do it for free at bing.com create or in Bing chat if you ask it to make an
33. 5:29image that's Dolly 3 mid Journey really good best in terms of art although uh you have to do things in Discord
34. 5:38ideogram okay but fading um Adobe Firefly from Adobe uh just went into version two and it looks like it's
35. 5:47getting pretty good here are some fun ones to play with I'm going to switch out Runway ml for either Pika or Moon
36. 5:54what's it called Moon Valley I think um for text to video hey Jen does video translation into a bunch of languages that are amazing um
37. 6:05those other things are worth playing with too so screenshot that and then the final little pile of these things future
38. 6:28pedia.com Siri go away Siri oh I said Sher D she thought I said Siri D Siri go away go
39. 6:43away okay Kyle melted my face off tonight with our last AI Workshop yeah it was intense Ronda it was it was a lot I told
40. 6:53you it was a lot going into that right so we did um so the the workshop that I did in in Rhonda's community
41. 7:00was called the AI Launchpad and actually I have a little graphic of it here so tonight was the final night and um and it was the first night
42. 7:16yeah the first night on October 4th was overview of AI the second night was um applying AI to your business so it was
43. 7:24mostly chat jpt but you know doing doing different use cases and things like that and then tonight was Advanced Ai and we
44. 7:31went through advanced chat GPT so like all of the new features in chat GPT we went through um music production tools
45. 7:41voice and animation tools video tools and then and then I even managed to squeeze in about 20 minutes of
46. 7:48automation using zapier it was just a machine gun of like here's here's 27 rabbit holes you can run down um so it
47. 7:58it was good but anyway thank you for having me Rhonda it was that was really great really by the way really nice
48. 8:05Community sweet community and Rhonda's great too um also the salon. a if you go right there that's a community that I
49. 8:12started last December um the AI salon and it's basically a group of AI adventurers so people just trying to
50. 8:19figure this stuff out helping each other Sher Dean nice I I got my first license plate look yay Kyle Shannon that's sweet thank
51. 8:33you um but but go join the salon if you go there there's a link to to the Discord where you can hang out with us
52. 8:40and ask questions and geek out um we've got guilds there now we got a beginners Guild and a a Art Guild and a writing
53. 8:49Guild and just it's it's cool it's cool group of people so go hang out there um you can put it on your bike
54. 8:57now um so yeah go do that uh story vine.com that's my company and yeah so that's all that stuff so let's let's
55. 9:06let's uh let's get started here so what I'm going to do is I'm going to go on ahead and I'll start answering questions
56. 9:14cuz I'll take my massive intellect you see how large my forehead is that indicates that I've got a very large
57. 9:20cranial uh crater as they call it and in my cranial crater there's there is some uh some activity and so I'm going to use
58. 9:30that activity here tonight to answer some of your questions you might want to ask there uh what are my qualifications all right mid Journey
59. 9:41just released a really big feature get it bigger pictures nice I didn't know that you don't pay attention for an hour
60. 9:51in this space and it's like good Lord really people let's see the announcements what are the announcements in in mid Journey all
61. 10:01right hey everyone we're releasing two new UPS scalers today a 2X upscaler and a 4X upscaler that's a big deal you can
62. 10:10see two and 4x upscale buttons under any normal V5 or nii job you can use it on old jobs using slow job ID nice the upscaler
63. 10:25is subtle and tries to keep the details as close as possible to the original image may not fix glitches or issues
64. 10:31with the old image the 4X up scaler costs roughly three times more GPU minutes than the 2x upscaler which makes
65. 10:44sense both upscaler work with fast GP fast GPU time right now we do not know if we have enough gpus yet to enable it in relax
66. 10:57mode known issue isues under very rare occasions it will drain your bank account no it doesn't say that I just I
67. 11:07just do you code machine learning no I don't um I understand it but I don't code it I am I am so what this Channel's
68. 11:19about is um so there's an inflection point that happened last November 30th right so we're coming up on the oneye
69. 11:28anniversary of the launch of chat GPT so in my opinion chat GPT marks an inflection point where before that time
70. 11:36you had to be an engineer a scientist a researcher um if you wanted to be involved in machine learning at all chat
71. 11:44GPT kind of opens it up for the rest of us very similar to how the worldwide web did with the internet back in the mid
72. 11:5190s and I was a co-founder of one of the early agencies in the mid90s company called agency.com so I've been through
73. 11:58one of these Cy before with this kind of a major technological and the parallels are kind of insane and this AI stuff is looking
74. 12:11oh I don't know significantly more profound and happening a hell of a lot faster so while I can't answer all the
75. 12:19technical stuff what I can talk to is you know um how do you use this stuff in your work how do you learn it why should
76. 12:26you learn it is it just hype um what are the ethical implications I mean I can talk to some of that I'm not an
77. 12:34ethicist but I'm an entrepreneur so my interest is in okay now how do we use these things now that now that we got
78. 12:41the people that are coding the machine learning and they're putting it out for the rest of us to use what do we do what
79. 12:48do we do what do we do what do we do and that's the that's the million dooll question and that's how we're all going
80. 12:55to figure out you can make money with J time how we do that um so yeah so let's go uh let's go look at some questions some questions who's
81. 13:06got some questions let answer the questions I so wish I could have been there I invited a friend of mine that is
82. 13:17awesome voice holy [ __ ] I don't know what R is Holy [ __ ] but that's cool uh did you just start I did just start
83. 13:26just wanted you to know how well you're educated us attended seminar today and I knew everything yes I listen I
84. 13:42mean this channel why I do these every flipping night is not for my health not for my sanity both of those
85. 13:53like um is to just keep in the convers ation if you just keep in the conversation with this stuff we're so early that you don't have to know
86. 14:05everything you just kind of need to know what are the major mile posts like what's possible today like chat jpt for
87. 14:12example today is a significantly different product than it was three weeks ago like a significantly different
88. 14:21product like I don't even know what it makes possible right now it can see it can hear it can talk it can make images it can program it can
89. 14:33do data analysis and it can do all the chat GPT [ __ ] it's [ __ ] amazing and it's really fascinating because right now all of those modes
90. 14:46they've got it programmed so they're all these distinct modes right I I can only assume that that's for them to do the
91. 14:54testing and fine-tuning that at some point all these modes are going to do this and chat jpt is just going to be like
92. 15:01this little magic box that you ask at anything and it does anything that's where it's headed I think I asked is mid
93. 15:07Journey free mid journey is not free they used to do like 25 free images and then they stopped that I don't even
94. 15:14think they do any now I think the minimum for Mid journey is 10 bucks a month that's what I pay I'm about to cancel that because of
95. 15:21Del um so yeah all right a large cranial crater with activity two and 4x that's what I've been Zippy to pin head hey it's me the
96. 15:38old Alice participant just just passing through to say hi and thanks for the new hobby absolutely welcome and good luck keep
97. 15:48having fun keep keep being curious came up with a new term last night that you you gota you got to work your curiosity knuckle
98. 16:00work that Curiosity knuckle I'm new and would like to start dabbling what is good to play with and free um perfect perfect
99. 16:18perfect so I am going to chat GPT so just little old Chad GPT what's going on oh I are we back are we alive yeah I think we're alive okay
100. 16:36so if you pay the 20 bucks to chat gbt which I think right now is the best deal in software um but if you can't pay that
101. 16:44that's fine but if you do pay it you'll have that little plus sign and you'll have access to chat gp4 and all these
102. 16:49goodies um even without that though just start with chat GPT and if if you have no ideas you can just like click on
103. 16:59these little things these are just ideas make a Content strategy make a plan an itinerary give me ideas um come up with concepts for retro
104. 17:07style arcade game whatever you can click on those things um yeah let's say let's just click on one of these see what it does give me
105. 17:16five concepts for a retro style arcade game this is odd this is normally very very fast so it might be my internet am I being all janky and
106. 17:37stuttery this does not look normal at all oh there it's going I have a feeling this is my browser I just woke my
107. 17:44machine up and and Microsoft Edge which is my default browser has got really bad memory problems here I'm going to I'm
108. 17:51going to stop this I'm gonna I'm gonna quit Microsoft Edge we're coming we're coming back to you Microsoft all right
109. 17:57listen listen Microsoft I don't know if you know this I know you're kind of still in startup mode and you haven't
110. 18:04you haven't really released any software before but if you could go on ahead and fix your Microsoft edge browser so that
111. 18:11it doesn't hose my [ __ ] Macintosh every time I wake it up from sleep that would be swell that' be super swell there Microsoft with
112. 18:22your your band of buddies up there in Redmond Washington all right okay back to this uh let's see show me a code
113. 18:32s show me a code snippet of a website sticky header I don't even know what that is but let's do that boom there
114. 18:40it'll write you code this is free so chat GPT will write you code uh we could also say
115. 18:56um oh there we go okay okay that's great could you also turn that into a country song well I'm not a songwriter but I'll
116. 19:07give it a try so now it's writing a country song about HTML in HTML we got a header so fine with navigation links
117. 19:16like a neon sign oh yeah um it can do that it can let me create a new chat here let me go um uh let's see
118. 19:30uh write me a 10 tweet tweet thread with Emojis and sequential numbering talking about chubby black and white
119. 19:46dogs 1 2 3 4 five 6 7 8 n 10 there's your Tweet thread you want social media content you got social media content so so
120. 19:58where should you start just start with chat GPT chat. open.com and just start playing run down as many rabbit holes as
121. 20:05you can and just start trying to try to stump it try to try to figure out where the boundaries are and what you quickly
122. 20:13figure out is that the boundaries are very deep and what it can do is is is quite profound um and then you can start
123. 20:23exploring other things I've got if you want to take some quick screenshots there's four you can play with there there's three more
124. 20:36there there's the image tools there's the video tools and these are just the ones that I think are worth
125. 20:44spending your time on these are just like personal editorial so there's lots and lots to do why do you use Microsoft Edge on a Mac
126. 20:54cuz I'm a dumb dumb so here's why Cory I cause I'm an OT CA I'm an idot um so
127. 21:09when when chat jpt first hit the market and it became clear that they were kind of in startup mode and they were [ __ ] with Google and it
128. 21:20seemed like Google for the first time in like 15 years had some competition um it just it became clear to me that um
129. 21:32they were going to get really aggressive with how they rolled out cat GPT and how they were going to incorporate it into
130. 21:37the Microsoft products so I switched my default search engine to Bing I searched I switched my default browser to
131. 21:46Edge and then I just and I just I haven't changed it back um but I probably should cuz it's at this point
132. 21:53it's it's unstable enough that I'm a I'm a tab hoarder and and if you're a tab hoarder with Microsoft Edge like it
133. 22:00doesn't matter if you're sleeping your your non-active tabs it just it [ __ ] crashes my machine every time I wake it
134. 22:06up so it's got to go oh wow look at Luna stick that's awesome so in the in the salon Discord we've got an
135. 22:18Irregulars channel so we got some inside jokes with our little characters and stuff like that but lunar sticks just
136. 22:23make it look like the salon the salon superhero game this is [Laughter] awesome oh that's awesome playing on
137. 22:37instruct GPT now chat GPT definitely more intuitive yeah exactly The [Laughter] Irregular it cannot do the word
138. 22:50irregular is not one it's obviously seen before because it cannot spell that word right that's really funny oh I love that one
139. 23:01that's cool those are awesome then private Peppa doing some Turtle action I like [Laughter] turtles oh man you guys are awesome
140. 23:17that's great okay so if you uh if you join the uh Salon Discord you can play along with what we do and
141. 23:29uh and if you make any cool images you can go share them in the show show and tell Channel bar just told me it could
142. 23:35be my CRM and keep track of real estate properties chat GPT told me no well so here's the thing log even if Bard told
143. 23:46you it can do that it may not so all of these large language models right now they do what's called hallucinating
144. 23:54which is basically where they lie like a Civ um I I like to call it mansplaining as a service um which is it will
145. 24:02confidently answer you even if it's wrong um if Google added something to Bard that allows it to act as a CRM I
146. 24:11would be very surprised but that would be cool um just be careful if these things tell you here's the an answer it
147. 24:20may not be the answer if they tell you they can do something they may not do it I was interacting with pi at one point
148. 24:27and and and said um hey can you email me that and it said sure I can email you that and it can't email it can't take
149. 24:35action so anyway um they will also Gaslight you yeah Joker exactly um so just be careful with that but cool all
150. 24:43right let's get some let's get some other questions going here uh did I broke Microsoft too could we make a Kyle bot impo a
151. 24:58super sized with shades of Steve Jobs bernee Brown and Einstein Einstein Einstein's a little much I think or do you just mean the eyebrows
152. 25:08yeah I've got Einstein's [Laughter] eyebrows but thank you for the compliment I like the the bernee brown
153. 25:17thing in there CU I think I think right now we we in so many ways need more empathy um just in general in society
154. 25:27but I think with this AI stuff this is it's it's profound stuff and I think a lot of people are very scared and you know
155. 25:38it's there's an interesting thing that happens when you sit outside of AI it seems really scary right because you're
156. 25:45like you're like oh the robots are going to kill us and it's taking my job and then when you get into it you start
157. 25:55playing with it and you do like little parlor tricks like write me a song write me a poem do this do that and then at
158. 26:02some point you do something with chat GPT or one of them that's close to what you do and and you will likely have what
159. 26:10we call your Kevin McAllister moment you know that one and at that point you get really scared and then you realize well it's
160. 26:19just a tool but it's actually kind of cool and I can do other things and then you get really excited and then you dig deeper and
161. 26:26you're like oh but that's really intense and that's going to screww up my buddy Jim's job and then you get scared then
162. 26:33you get excited then you get scared then you get excited and then those things just kind of meld and it's like this stuff's intense
163. 26:41it's exciting and it's scary and it's profound and it's awesome and it's terrifying and so I think that we all
164. 26:50need to just give each other a little Grace as we're trying to figure out how these things are going to impact our lives and how how
165. 26:58how we may have the opportunity to reinvent ourselves but how we might also be forced to reinvent ourselves and
166. 27:06that's [ __ ] terrifying so I think the brne brown um layer in this channel is is an important one so thank
167. 27:17you for for acknowledging that that's that was very sweet start by replacing Google with chat chibi yeah I don't really Google
168. 27:28anymore I'm a digital nightmare too many AI told me to put them on a laptop just for storing photos I'm the same way I've got hard
169. 27:40drives just laying around that I'm never going to throw out cuz like someday I might hook that up and see what's on it
170. 27:49like I'm a tab hoarder a a digital hoarder I guess I suppose I'm also a hoarder hoarder I a lot of crap um hey now don't be
171. 28:06browser I think my promps gave a AI that idea bar just told me oh yeah that we saw that one um hey wait he is bot
172. 28:17blocking me from Sydney accuses me of lying in gaslights me oh my God instruct GPT yeah we saw that one
173. 28:30I'll share a new video and show and ton oh right now oh cool oh it's Claire's birthday happy birthday Claire so let's see we got we got a
174. 28:45happy birthday Claire nice lovely that's actually really beautiful nice job Sherry D Claire I'm telling you Claire is like
175. 28:55a machine making making the art with mid journey in the AI Salon I have seven laptops I call photo albums that's
176. 29:03[ __ ] hilarious I've got I've got hard drives that are like this is like certain certain uh you know piles of years um that's really
177. 29:15funny all right here's a new video from Herod I made my first animation with sdl sdxl last night remade one of my
178. 29:23previous renders it came out really cool all right let's let's make this full screen this from
179. 29:38[Music] Herod wow that's that's super cool dude you know you you've
180. 29:56got it's cool the technique that you're using it it kind of hides the um the sort of um the weird animation
181. 30:08you know the sort of non-coherent Animation that a lot of that a lot of gen AI things have where you know you see someone's face and it
182. 30:15goes like the the technique you're using hides that it's it's really slick really good thanks for sharing that that's super
183. 30:24cool the other thing herod's doing in the salon so not only is he doing those cool videos but he's been teaching
184. 30:31people like how to make them um which is super [ __ ] cool so Kyle maybe AI will organize it for us at some point I think
185. 30:38it will actually uh Amelio's wife I think that um I think what we're all going to discover in the next three to 3 to 5
186. 30:51years maybe like two to two to four years I don't know I'm just guessing at this point I think these AI tools are going to help
187. 31:01businesses and people ReDiscover their digital histories right so so I think my fantasy with having all these hard drives is
188. 31:11that at some point I'll have the time to do digital anthropology and I'll dig and I'll find all those cool ideas I had and
189. 31:18I'll resurrect them somehow and there's a lot of problems with that one is a lot of the file formats of the things that are on those
190. 31:26hard drives the software doesn't exist anymore right so so so like there's that and then there's just there's massive
191. 31:34amounts of it and then there's all these duplications where you've got you know just it's impossible to organize well so
192. 31:40then imagine when generative AI is rolled into the operating system and it can just analyze all your data and
193. 31:50cluster it and and figure out what's in there and go find obscure file formats and update them from the Obscure old
194. 31:57file format to some new readable format so and I think that's going to happen for businesses I think it's going to happen for
195. 32:04individuals wow that's cool wow that's super cool it's another Herod look at that subtle animation that's
196. 32:21slick nice work Dag um yes so I think that's very much a possibility Jim Ross was at the Las Vegas sphere last night missed my AI
197. 32:34learning lab time I have a problem oh my God you're you're at the world's literally greatest spectacle in the
198. 32:42history of mankind you're like I want to get on the AI learning lab I'll I'll do my best to get a sphere going here I've
199. 32:50got I got a little light up moon this this he I got to I got to charge it though it doesn't turn on anymore I don't think does it let's see
200. 33:08on no batter's dead the battery [ __ ] dead it's Dick's 84th birthday wow Becky
201. 33:24amazing we're all ticking up there in aren't we jeez D Bears uh not totally it thanks Jim Ross cool though how was it I need a laptop HD I
202. 33:39think I tossed all my old ones out um yeah the the sphere looks just amazing like the scale of it what's it
203. 33:48hold 25,000 people or some crazy ass number like that apparently the food's good the hospitality is good the the you
204. 33:57know line Management's good being said I don't think you talk to Sydney only Microsoft employees can maybe it was an impostor that's really
205. 34:11funny um I posted the link I used to make that effect oh cool that's great Bing said I don't think you talked
206. 34:21to okay we said that um Kyle I made you where in the Irregulars Channel [Music]
207. 34:45look that's awesome thank you oh [ __ ] that's just that's a it captures me a little too [Laughter]
208. 35:04close um have you shown us how it can analyze data can it forecast future Trends oh you have shown us how it can
209. 35:12analyze data can it forecast future Trends maybe um let's go experiment uh
210. 35:27um let's go to um kagle K ggle so the site I'm on is called kaggle this is where you can go get data sets
211. 35:41so we're going to we're going to go do a little Advanced Data analysis and see if we can get chat GPT to help us so we're going
212. 35:51to search data sets for um stock data stock exchange data 5 megabytes two years ago all right this is a lot of
213. 36:06data so we're going to download it it's downloaded we're going to go to chat ji Tai you can make money with the CH J TI
214. 36:16if you do future forecasting on the stock market and invest all your money and then leverage it and then everything
215. 36:23wins and you can you can only make money ever all right so we're going to go to Advanced Data analysis we're going to click
216. 36:32plus we're going to pull in archive 15 zip so if you haven't seen Advanced Data analysis this is it and so I'm going to say in
217. 36:44this file is historical stock data can you please analyze it and tell me what trends you can identify oh wait what's going
218. 37:00on hello in this file is historical stock data I'd like you to analyze the CSV and tell me if you can identify
219. 37:16Trends within it so if you haven't seen this before future Trends is tough because there's not a solid way to detect hallucinations
220. 37:27yeah I know Herod although within within code interpreter because it's writing and executing python code what I'm going
221. 37:34to ask it to do is give us models for for for Trend analysis you know historical and carrying it forward and
222. 37:43maybe it'll do the math in the in the python code so um let's just see what it does the so like I did I did a a an exercise today
223. 37:56for work project where we had um we did a survey for for a project we're working on we had code interpreter analyze all
224. 38:04the survey data and write us a report which it did really well but we had a we kind of did three different versions of
225. 38:11the report and then we put them all in a Google doc and then I tried to have code interpreter read all of the reports that
226. 38:20it wrote and consolidate them into a single one that that sort of captured the best of everything and I kind of failed miserably at that so
227. 38:30um I'd like you to educate me to help me understand how an expert stock Trader might look at historical stock
228. 38:43analysis and project that into future Trends or even models that might indicate when to buy
229. 39:01it involves multifaceted Financial knowledge not anymore technical analysis chart patterns Trend analysis moving averages
230. 39:11technical indicators okay we've heard of all this stuff fundamental analysis quantitative analysis sentiment
231. 39:20analysis all right so these are all terms I've heard I know nothing about them risk management economic indicator Market
232. 39:28Cycles all right so there's obviously not all this data in this data set what I would like to understand is
233. 39:37that based on the data in the data set you have available to you now which of these might we use to create some future trending models as an
234. 39:49experiment the reason I put as an experiment is if if I'm asking it for financial advice like this safety stuff
235. 39:57might just say I can't do that for you so I'm just trying to get it to do [ __ ] um we can calculate short-term and
236. 40:06long-term moving averages analyze Crossovers and identify potential buy and sell signals uh uh uh uh uh would you like to
237. 40:16proceed with one or more of these analyses on your data if yes specify which ones you're interested and
238. 40:21also let me know if you have a specific index in mind you'd like to an I want you to pick the model I want you to pick the model and
239. 40:35the index to analyze and I'd like you to create a little report that discusses significant historical
240. 40:54Trends and how those might carry forward into the future certainly I'll select an index and a technical analysis model to create
241. 41:05a report for this exercise I'll choose the nyaa index which represents the New York Stock Exchange given its broad
242. 41:14representation of the market as a model I'll use moving averages and moving average convergent convergence
243. 41:21Divergence for Trend analysis the macd of course everybody knows that this is all so obvious I knew that I learned all
244. 41:30this in my first year of acting um all right introduction let's see data overview the data set comprises 13,000 Daily
245. 41:42records closing price at approximately 44.52 volatility Trend analysis through moving averages okay so now it's just a if you haven't
246. 41:56seen this work what's cool about this is if I open up this little box the way Advanced Data analysis Works in chat GPT
247. 42:05is it can write and execute its own python code so it's actually writing and and you know doing some sort of mathy
248. 42:14maath [ __ ] it appears there was an error due to the nump N Library not being explicitly imported right so there's the
249. 42:21the error stuff and so it recognized it errored out it says let's Rectify this and continue so it's it's figuring out
250. 42:29and problem solving on its own it appears there was an okay what's this oh look look data it also made a
251. 42:42chart and a graph did you know that it could do this stuff so it okay Trend analysis in the chart above we see the New York Stock
252. 42:54Exchange index historical prices along with its 50 and 200 day moving averages the 50-day moving average is shortterm 200 is long-term
253. 43:04Buy sell signals um next we'll calculate look for any points in the historical data where the macd crosses
254. 43:14the signal indicating buy and sell signals there's some of that [ __ ] I can't read that I don't know what it's
255. 43:23doing I don't know if that's valuable but it doesn't really matter momentum analysis by there are 551 instances in
256. 43:30the historical data where the macd crosses the above signal line which is commonly interpreted as a bullish Buy Signal conversely there are
257. 43:39550 in instances where it crosses below the signal line a sell signal it's amazing that they're 551 and 550 you know they say that the stock
258. 43:50markets a coin toss there you go historical volatility analysis so there's volatility discussion of historical
259. 44:01Trends moving average I mean this thing is just going and going and going conclusion historical analysis provides
260. 44:10valuable insights into its past Behavior oh it just it just crapped out so anyway I don't know if that's doing future
261. 44:17projection stuff but you could you can certainly be doing some some analysis good Lord good lordy lordy lordy
262. 44:26oh what do I no I only worked on Wall Street for two years oh I forgot that Herod yeah I don't know listen I have a
263. 44:33degree in acting you know that oh man I forgot about that that's hilarious all right he made money with chat [Music] gbt what was
264. 44:53that Kyle had $2 million in stock at one now I had more I had more than 2 million in stocks I had a lot in
265. 45:05stocks at some point in my history but it was all paper money there was none of that real money there ah Herod I'm just kidding I did work
266. 45:20there but I wasn't a stock Trader I was a code [Laughter] monkey was a wild experience turns out I hate Finance though yeah exactly yes 50
267. 45:29moving average crossover 200 moving average also rent rent in Manhattan is $150 million a month it is if you're not
268. 45:38taking companies public on a weekly basis you can't afford to live in New York oh man let's make a whole other
269. 45:47branch of math for the unknown chaos and chance and col it calculus um I spent the day researching demographics of an influencer's audience
270. 45:56without much luck yeah Source Camp I I mean my experience with code interpreter is Tainted by the fact that I don't know [ __ ] about data
271. 46:06analysis so I kind of get lost in like like because it's kind of prompting itself and and doing its own
272. 46:17work and coming to its own conclusions I think what sometimes happens is it grabs a subset of the data and puts that in a
273. 46:25data table that that it's doing analysis on and then if you ask it about data from the larger data set it just looks
274. 46:33in what it has in the data table and it's incomplete and so it just starts to fall apart so I don't quite know how to
275. 46:39use it yet but um I think it will get smarter and then I'll get smarter using it but it's it's it's I don't know it's
276. 46:49wild it's wild that it does what it does it it just every time I see it do that I'm just like what are you doing what are you doing no
277. 47:00literally what are you [Laughter] doing uh I see you ask code wait I've seen you ask it to code in Python before can it
278. 47:12do any other languages beside python oh yeah so within code interpreter it's only been given access to execute python but chat GPT right um
279. 47:30can you tell me two modern and two out ofd computer programming languages uh uh it's subjective modern
280. 47:46python JavaScript outdated cobal Fortran I'd like you to write a rocket
281. 48:03trajectory application in Fortran that can calculate getting a rocket into
282. 48:18the moon's orbit so so basically said this is really complicated I don't think I can do that
283. 48:33however I can provide a simplified example in Fortran all right so there's Fortran so I don't know if that works
284. 48:43but what's it doing changing effect of gravity as the rocket ascends the loss of mass as the rocket consumes fuel air
285. 48:49resistance thrust produced by rocket engines orbital mechanics the rotation of the earth other forces acting on the
286. 48:55rocket for real trajectory calculation one would typically use a numerical method to solve the okay whatever so
287. 49:04we've got there we've got some some Fortran um can you refactor that into cobal Koba is not is not is not a
288. 49:23recognize I meant Cobalt not Cobalt nevertheless here's a simplified
289. 49:38cobal version what was the other thing what was the other language it did here
290. 49:54JavaScript great can you now put this in a JavaScript application with a visualization of the
291. 50:14data uh uh uh uh so what did it say it said creating a JavaScript application with the data visualization involves both the
292. 50:24computational logic and the front end visualization we'll use Simple HTML and JavaScript for the front end and a
293. 50:30library like chartjs for the data visualization please note that trajectory calculation is highly simplified and not accurate for real
294. 50:40world rocket launches okay all right chat GPT I promise we won't launch a rocket tonight with your little math um so there's that
295. 50:51um yeah so there's the there's the trajectory math [ __ ] and there's that and then I'm going to say um can you now turn this into a
296. 51:04rocket launching game written in C intended to be run inside Unity uh uh uh uh uh uh uh it's a complex
297. 51:16task firstly you'd need to set up your Unity project here's a simple C script that might so so the answer is yes it
298. 51:26can do languages other than python oh man one of the one of the interesting things about about the coding in gp4 is
299. 51:42that it seems to there's a thing called Model drift where where the um as they're tweaking and fine-tuning the models
300. 51:54like maybe it was really good at programming two months ago then it gets shittier at programming and then it gets
301. 52:00good again and then it gets shitty again there there's this this oscillation that seems to happen with the quality of
302. 52:06these things um and I think it's just because it's so early they're they're they're fine-tuning them and you know
303. 52:12trying to deal with safety and probably bugs and exceptions and all sorts of weird [ __ ] you know hackers hacker going
304. 52:19to hack trying to break it so as they fix all that [ __ ] it probably you know affects the quality the code but anyway
305. 52:27there you go hope that helps Mitch Mitch there you go Mitch uh what platform should I be using for what source
306. 52:37Camp yes it can code in any language honestly yep right now your agent right now I'm your agent for your Tik Tock
307. 52:45live soundtrack featuring your non- AI voice [Laughter] awesome uh uh uh uh uh uh
308. 53:00all right how many folks are in here got some got some folks in here do me a favor if you would if you'd be so kind
309. 53:09um share the live share it with your friends let them know you're a weirdo and you watch other weirdos talking
310. 53:16about this AI stuff even if they're like Angry against it just just share it and then if you like if if you like
311. 53:26this idea of keeping up with this stuff staying curious about it um follow the channel and then uh yeah do do other you
312. 53:35know whatever else you want to do um hey Kyle where is it that you can test the output code from GPT well where wherever you would test
313. 53:43code so like if I'm if I'm doing like I'm not a I'm not a coder so I don't know a lot of the stuff um if I'm doing
314. 53:51things in Python you can go to a there's a site called repet but the other other thing you can do madcap and this is how
315. 53:57I learned to do it is I wrote some code and then I said to chat GPT I don't know how to run this code can you help and
316. 54:05then what chat GPT says is oh sure you on a Mac you open your terminal program and you run these command lines and and
317. 54:12then I went no I don't want to run command line Unix can is is there a website that can just do this [ __ ] for
318. 54:18me so I can see if the code works and they'll go oh sure here's five of them and then you just go find one just keep
319. 54:25every time you run into a roadblock with the programming just have chat GPT educate you you know the the first time I did this with
320. 54:35python it it gave me repet and then it said you copy the code and you paste it in the left Pane and then you hit the
321. 54:41Run button and then it'll run in the right pane I'm like okay I can figure that out paste run error and I said hey chat gbt it gave me
322. 54:50an error and it said well you need to install this Library I'm like I don't know how to install that libr Library it
323. 54:55said in the lower left hand corner click on this button and then in the upper right hand corner search for the library
324. 55:01and then when you see it install it and then hit the Run button again and I did that and it worked and so it's just you
325. 55:10know the you know that you know the old the old uh the old cliche let me Google that for you I kind of feel like chat
326. 55:16GPT is that on steroids it's like but I don't know anything about quantum mechanics well now you do but but I don't know anything about
327. 55:26interior design and color theory well now you [Laughter] do welcome to the new world oh you know what I haven't one of the things that
328. 55:37always blew me away about mid Journey was designers that were doing like interior design so so now I'm in chat GPT and I'm going
329. 55:49to I'm going to play with the dolly 3 so I've got Dolly 3 inside chat gbt if you pay the 20 bucks for chat GPT plus which
330. 55:58I think is the deal of the century right now um if you don't have Dolly 3 it will be coming soon like they're rolling it
331. 56:05out pretty widely across anyone who's got chat GPT plus so this has now got image uh generation within it so what I'm going to do is
332. 56:15this you are a world famous New York City interior designer and I need your help design ing my 55th floor Penthouse
333. 56:32apartment in lower Manhattan period I would like you to give me five different concepts for how you might
334. 56:48design my apartment I'm particularly interested in the big living room that has floor to ceiling windows and a view
335. 57:02of the Statue of Liberty period before you make any images can you describe your F five Concepts to me all right so our little interior
336. 57:17designer here elegant minimalism that seems like New York cliche to me so we're not going to do that classic Manhattan glamour
337. 57:25drawing on inspiration from the 1920s Art Deco area Urban Jungle Oasis no that's not right not for a penthouse Bohemian Loft again that's
338. 57:37cliche and if we were lower in the building that might be good modern industrial okay these are all very expected and
339. 57:52cliche for New York City I was hoping because you're an award-winning designer you would have some more innovative
340. 57:59ideas for me why don't you try [Music] again so this oh God even hearing for TR gives me PTSD that's really funny Herod
341. 58:14futuristic fluidity Envision a space where the boundaries between the walls ceilings and Floors blur seamlessly
342. 58:22curved structures would create a fluid continuous space think of undulating walls or furniture that emerges
343. 58:28organically from the floor I love that actually I mean cuz that'll be visually interesting cultural Fusion lab okay we
344. 58:36like that interactive art Lounge Echo futurism Retreat sensory exploration chamber okay so I like one two and five okay
345. 58:45great perfect I like one two and five can you now write up a hybrid between those
346. 59:02[Laughter] [Music] three fusion sensory Odyssey architectural fluidity the foundation of this design is seamless fluid architecture from
347. 59:18futuristic fluidity concept spaces flow organically from one to the next cultural integration the C cultural
348. 59:25Fusion lab room would incorporate design elements from Global Traditions Scandinavian fluid sofa vibrant African
349. 59:33textiles Zen meditation space beautiful sensory exploration room becomes an immersive experience of sensy
350. 59:40exploration center the view of the stat Statue of Liberty is framed with augmented reality great great awesome perfect
351. 59:50okay this sounds exactly like what I'm looking for for can you please give me four different renderings of design ideas for the
352. 1:00:04living room with the view of the statue period the Staten Island Ferry should also be seen in the
353. 1:00:18distance boom Dolly 3 now it's creating images Robert Stone hello what's happening so what I wanted to see was
354. 1:00:29could we get could we get Dolly 3 to create interesting um Innovative interior designs um as good as some of
355. 1:00:38the early mid Journey stuff where interior designers and I guess just regular people asking for wacky interior
356. 1:00:44designs there was some really amazing stuff so it will be curious to see what this thing can do and and you know the like so what I
357. 1:00:54just did there was oh this is beautiful so what I just did there was um I didn't I I don't have to do image prompting
358. 1:01:02right I can just I can just have a creative uh glad my mental trauma is so entertaining for you Kyle you're welcome Herod I'm here for you
359. 1:01:13buddy um [ __ ] that's cool look at that can you imagine walking up walking up into a a 55th floor Penthouse I mean
360. 1:01:25the Statue of Liberty is a little close but but that's okay there's the Staten Island Ferry as requested Staten Island Ferry Statue of
361. 1:01:36Liberty this is obviously not the 55th floor and by the way that's Manhattan in the background so this is Jersey he just
362. 1:01:42they just moved my apartment to Jersey [ __ ] I'm gonna I'm gonna tell it that I'm gonna say uh yeah it looks like three of those
363. 1:01:58apartments are in Jersey with a view of lower Manhattan that's kind of disappointing period could you put me on the 55th floor in
364. 1:02:07Manhattan looking toward the Statue of Liberty with jersey in the background [Laughter] period And that's pretty [Laughter] funny oh
365. 1:02:23man did chat GPT doesn't do good cobal yeah it probably hasn't been trained on it upload a photo of my family room and ask
366. 1:02:33that it be redecorated with Turtle Decor just saying you can do that there's a um there's a couple of apps if
367. 1:02:42you go to futurepedia doio there's a couple of apps that let you upload your any room and it'll it'll redecorate
368. 1:02:49it which is pretty pretty like all this stuff is just [ __ ] wild all right now now I'm down south of
369. 1:02:59the Statue of Liberty I think this is an apartment from the verizo bridge and and the uh Staten Island Ferry is it looks
370. 1:03:10like it's going about 8,000 miles an hour steaming toward the tip of Manhattan yeah it it just does not understand it doesn't understand the
371. 1:03:22geography but still pretty cool like that design right there that's pretty slick Nancy reached member level 13
372. 1:03:32awesome Nancy thank you so much all right um so this is Dolly within chat GPT um oh you know something I wanted to
373. 1:03:42try also that we'll we'll we'll do here as an experiment together no KY Kyle do tell me do tell me do tell me Kyle go on
374. 1:03:51do tell do tell all right I'm creating a children's book about a rag tag band of Nero well animals that
375. 1:04:08include a llama a hedgehog a turtle and a gummy bear period I would like
376. 1:04:23to create an image for each of them in a modern cartoon animation style to see if I like what they look
377. 1:04:39like all right so one of the things that they kind of promised one of the reasons we have Hedgehog is one of our animals that we
378. 1:04:49kind of make fun of is in the dolly 3 video when they did the announcement there was Larry the Hedgehog and they kind of maintained Larry
379. 1:05:00throughout throughout a series of prompts like once they declared that's Larry and then they said let me see
380. 1:05:06where Larry lives and let me see Larry dancing through fields and it me maintain that character across prompts
381. 1:05:13so I'm going to see if we can do that [Music] here all right
382. 1:05:28[Laughter] lunar [Laughter] stick this is what lunatick thinks of me like 900
383. 1:05:44lb and I have and I have a price tag of $7.99 oh my God that's hilarious man that that second one does
384. 1:06:00look like fat Alec Baldwin it's pretty funny all right let's go see our little images here okay you you apparently can't do a
385. 1:06:15cartoon hedgehog without making it Sonic the Hedgehog and that's pretty much Ninja Turtles and that's kind of uh
386. 1:06:24hey koola all right so we'll go with the Llama the llama's pretty cool all right so let's try this let's go okay I like the
387. 1:06:41Llama let's call him Lamar period I'd like to see Lamar sitting on a tricycle
388. 1:06:56in a Suburban culdesac so Lamar should be pretty cool
389. 1:07:10with sunglasses what's he got sunglasses he's got sunglasses a leather jacket oh
390. 1:07:33oh I had to grab my phone and put on my glasses didn't want to miss these results I know isn't it terrible it's
391. 1:07:40just like you hit a certain age it's like if these are off nothing I got nothing um here are different images of Lamar
392. 1:07:55sort of I mean it kept his it kept his leather jacket yeah it kept his leather jacket it kept his color that's pretty
393. 1:08:08close that's not that's really not okay of all of those number one is the one that's
394. 1:08:22closest so let's try to keep that style moving forward and let's keep that look and feel for Lamar period now I'd like to see
395. 1:08:34Lamar walking through a field of wild flowers on a fine spring day in addition to his leather jacket
396. 1:08:50he's also got on a swell baseball cap cocked a little bit to the side because if if if Del indeed does
397. 1:09:08like even semic coherence across prompts then you could start to really think about doing things like you know
398. 1:09:16characters in a graphic novel or characters in a in a children's book where where you're maintaining
399. 1:09:24um you're maintaining a character across [Laughter]
400. 1:09:40frames oh I'm sorry but I was unable to generate all the images due to our content policy well your dumb policy and
401. 1:09:49you write the wrote the damn prompts that's sort of looks like Lamar that is totally a different style that looks a lot like
402. 1:09:59him that's pretty cute so so it seems like with with enough kind of messing around you could you could probably get
403. 1:10:12Deli to be decently coherent a decent Storyteller well that's fascinating all right not perfect but usable look it looks just like me
404. 1:10:29see here's the here's the monster that I've created so when I first started doing these lives trolls would come in
405. 1:10:36here and they would stay say stupid [ __ ] and I would make fun of them and then I said listen if you're going to troll me
406. 1:10:42at least be clever about it so now the trolling is taking the form of hilarious parody images of me
407. 1:10:58Dolly did [Laughter] ITB man man man man futuristic just kidding Jersey residence maybe I
408. 1:11:17know what an imagination haha Kyle how do y'all get the image of you to start with in AI Salon uh which one the ones that people
409. 1:11:31are making of me of these I don't think they're taking pictures of me and uploading them I think they're just kind of describing like you
410. 1:11:40know if if Alec Baldwin had too many Quarter Pounders and had a had jol Santa face what might he look like
411. 1:11:54it's pretty funny I just realized you're wearing a catalyst shirt that's great I am that's the building I work in
412. 1:12:07exactly um I have really bad nightmares about teenage mut and Ninja Tur Turtles I can't wear green clothes that's
413. 1:12:15funny um number three is funny yep he's sitting outside the house on his motorcycle waiting on a gummy bear girlfriend and turtle dad spies
414. 1:12:26[Laughter] nice all llamas look the same to me can I get a copy of your YouTube Claude AI prompt for Tik Tock
415. 1:12:39videos um yeah I'll share that I guess I can share that in the in the salon somewhere I mean it's super specific
416. 1:12:53spefic I guess I mean just take it and you're going to have to you'll have to modify it cuz it's it's really super
417. 1:12:58specific um if you want to see what what uh go one of the challenges you have with chat GPT is it's got a really short
418. 1:13:06context window basically you can't your prompt can't be too long um it kind of loses its memory after a while but you
419. 1:13:15can use Claude so here's [Music] Claude there's cloud cloud hello my name is cloud and then we're going to go to the
420. 1:13:32YouTube and we're going to go to the old AI learning lab your channel and then I'm going to go to videos and oh I I did these all already
421. 1:13:44dang it oh well I'll show I'll show you how it works anyway so so these Tik Tok lives I do are generally between 2 and 3 hours long one
422. 1:13:53and a half sometimes but mostly between 2 and 3 hours and so every one of these videos has a little transcript but it's not a little
423. 1:14:03transcript it's two and a half hour transcript so so we're going to go in here where is my transcript all right so
424. 1:14:12there's a little button that says show transcript and so there's my transcript right so this is YouTube putting time code and all of the
425. 1:14:22[ __ ] I say as I ramble for 3 hours it's all been documented what do they they just need three seconds to clone your voice right I'm
426. 1:14:35[ __ ] all right so now we're going to go back to Claude and I'm going to just paste that whole transcript in and
427. 1:14:41notice what it does it turns it into like a text upload effectively so it's 13 3.46 kilobytes and then I'm going to go over to
428. 1:14:53a Trello board that I have where I've got a prompt that I wrote just for this situation and this is a long ass prompt
429. 1:15:08that it took this is this is about a 50 attempt it's probably more than that it's probably 75 or 100 attempts to get
430. 1:15:16this prompt to work basically there's three components to it act as an AI assistant who specializes in effective
431. 1:15:21copyrighting in YouTube metadata based on the provided transcript from 10:13 so I just changed this date 1013 do the following task craft an
432. 1:15:37interesting and click worthy title and then format it a certain way write a single paragraph description that
433. 1:15:43follows these rules that are like should be in the first person should not be self- congratulatory like when it first
434. 1:15:50started doing these it was like I'm the most most awesomest person in the world I'm like no no no no no no um so gave it
435. 1:15:59some just some boundaries there and then and then identify the the third one is this was the toughest one this is the
436. 1:16:05one that really took me a lot of time to figure out identify the topics Kyle talked about most passionately and
437. 1:16:11follow these rules topics should be determined by a combination of what he says and how long he talked on the topic
438. 1:16:19longer sections imply they are more important um select the topics that will appeal to a broad audience list the
439. 1:16:25topics chronologically in this format and then be sure to capture the topics for the entire length of this transcript
440. 1:16:32and it does it follows these rules about 20% of the time and then sometimes it just goes off the [ __ ] rails and it's
441. 1:16:41just useless and what I found that's weird is um sometimes if I just change the length of like the copied block by
442. 1:16:51like a word it'll it'll complet completely behave again so there's something about how How It Ends on a
443. 1:16:56certain word I think [ __ ] it up I don't I don't know how or why have you tried to upload a picture of yourself to
444. 1:17:03gp4 and ask it to make you Pixar like you can't do that right now be not afraid so there so that's what I was saying the
445. 1:17:13um chat GPT right now has broken itself into these really distinct modes so the image upload mode is not connected to
446. 1:17:22the do mode so what I could do is I could upload a picture of myself into the vision mode and have it describe me
447. 1:17:33and then I could take that description to the image mode and then have it you know take that description and then add
448. 1:17:39I want you to render this like a a Pixar character we can try that um so let's see what this did it um dissecting the AI revolution in Plain
449. 1:17:54Talk on October 13th 2023 I hosted a wide- ranging three-hour conversation in my AI learning lab on Tik Tok so I got
450. 1:18:01that part right it didn't break the chapters into individual lines but like oh yeah here's a good example so it
451. 1:18:09didn't break them into lines and it only went up to 5 minutes and 20 seconds so I'll go back and look at the video it's
452. 1:18:15a 3our it like it knows it's a three-hour thing I'm going to say it looks like you only got topics for
453. 1:18:27the first five minutes or so this is a three-hour transcript try again so so I just yell at it sometimes I'm like start over
454. 1:18:41[Laughter] loser it should apologize and tell me oh you're absolutely right I only got for the first three minutes five minutes and
455. 1:18:53then it'll say let me do that again and then it'll do it for the same 5 minutes and that's the the other thing
456. 1:19:01about I talk about this a lot that these tools right now are janky pieces of [ __ ] and this is a perfect example of that is
457. 1:19:07that I've got this very very specifically crafted prompt and then sometimes it works completely predictably and then it'll just go off
458. 1:19:17the rails and and that's what I I hear really consistently from people so this is now not bad it's it's a little too much
459. 1:19:30detail but that's fine like the whole purpose of these chapters is as much for like SEO as anything right like I just
460. 1:19:37want I just want to be able to if someone's searching for um autonomous agents like there it mentions autonomous agents and whether
461. 1:19:46or not it actually jumps to that I don't really care if I were doing this for a media organization the the accuracy of these
462. 1:19:54things would be more important but then what I do is I just copy this go over to YouTube and paste it in there so that's
463. 1:19:59the that's the prompt that was being asked about there pretty slick is't it AI learning
464. 1:20:17lab what time is it what time hey Eddie what time is why you wash your Dicky in I think uh uh uh uh it's 10:22 Do You Know
465. 1:20:28Where Your Children Are I do wide ranging is that a euphemism for rambling exactly it's absolutely is wide
466. 1:20:41ranging plot do.ai check it out yeah plot do.ai is pretty cool other than their annoying name of plot dot dot
467. 1:20:50having to say dot dot twice um but yeah it's that's pretty slick it's a if you haven't played with plot do.ai go play
468. 1:20:59with it if if you want to write a screenplay it'll it'll write you um it's basically a storytelling framework like
469. 1:21:06all the storytelling Frameworks but they just threw the AI generate button within it um the interface isn't great but it's
470. 1:21:14it's a pretty cool tool give it some items out of your pantry and ask for a dinner recipe oh I could do that let's see how can
471. 1:21:32I uh we can we don't need I don't need to take an actual picture let's go it's a good thing to play with um I'm going to say
472. 1:21:46um picture of the inside of a refrigerator this looks sad what's in there I can't see
473. 1:22:06it actually this is good all right so this this is a UK UK refrigerator let me so let me save this image as
474. 1:22:21fridge I'll save it to the desktop okay so this is going to be good watch this well I don't know if it's going to
475. 1:22:30be good this is going to be an experiment these tools in chat GPT right now are so so new like I don't
476. 1:22:40know like now that we have chat GPT that has Vision that that it can see and interpret things like what do you
477. 1:22:47[ __ ] do with that I don't know yet but we're going to we're going to to see if we can make a recipe out of our
478. 1:22:56Strange kitchen okay so the so the way so so Dolly 3 will make images there's nothing here that says I want to see
479. 1:23:05images so if you go to default mode it adds this little picture icon and then if you click that now I can attach an
480. 1:23:14image so let me go find that fridge image and I'm going to say
481. 1:23:29can you make a recipe for dinner using what I have in the fridge and the recipe should be appropriate to the country that I live
482. 1:23:45in cuz it looked like it was from the UK how did I that's cranberry juice oh maybe not we're lazy this is not
483. 1:24:04us okay of course let's take a look at the contents of the fridge there's some dairy products possibly milk yogurt and
484. 1:24:12cream there's eggs grapes some condiments a juice box possibly cranberry juice some other items which are not completely visible however I'm
485. 1:24:20not certain about the country based on the image can you please specify your country I'll do my best um oh heck just be adventurous and make
486. 1:24:32me something tasty it's going to be disgusting all right let's get creative with what I can see creamy eggs Scramble
487. 1:24:43with grapes and toast ingredients eggs from the fridge dairy product could be milk yogurt or cream grapes salt and pepper assuming you have
488. 1:24:55these bread for toast assuming you have some of the pantry condiments all right prepare the bread all right that's any what other ideas do
489. 1:25:06you have for what we could do with vision we decided we're going to make an irregular AI recipe to to replace
490. 1:25:19pie but we only counsel each other I think that's good I think that's good forever hooked yeah that's solid Kyle thoughts on J Jasper
491. 1:25:29AI so so my thoughts on Jasper and copy Ai and all of the things that are that are like these sort of Mega prompt
492. 1:25:37libraries is is what I is the best I can describe them um they've been around for a while so so Jasper and copy AI were being built I
493. 1:25:49don't know a year or two before chat GPT now that chat GPT is here I think it probably serves you better to learn how to prompt chat GPT
494. 1:26:03yourself to to do what Jasper or what copy AI can do now when would I use a copy AI or a Jasper if I'm like if I don't have
495. 1:26:14anyone in my company who's got any ambition to learn this [ __ ] and I'm just trying to get more efficient in having
496. 1:26:23my copy team crank [ __ ] out then what I might say is go figure out if Jasper or copy AI is better and just you know
497. 1:26:32click on buttons and go make the [ __ ] they that they've prompted for you but like now that now that chat gpt's gone
498. 1:26:47multimodal is copy AI going to keep up is it going to go multimodal well it might but they're going to be behind the
499. 1:26:53eightball so I so I just kind of feel like things are changing so fast right now it's a better fundamental skill to
500. 1:27:02understand how these things do what they do and where the boundaries are than it is to be able to push a [ __ ] button
501. 1:27:09on you know copy Ai and just type in a topic and have it write a blog post so I don't know I like I understand why
502. 1:27:17they're there I think that there are Absolut there's absolutely a market there people are going to want the easy button you can make money
503. 1:27:27with but um I don't know I don't have a lot of interest in it right now so pesto mushroom zucchini barbecue
504. 1:27:42chicken it won't do much with mine peanut butter pickles and romen almond kale oh this is good everyone's giving their their inventory
505. 1:27:52Becky almond milk kale spinach cherry tomatoes wor theer sauce ketchup mustard and eggs that that's like you got
506. 1:28:00Gourmet action going there I think I can finally run a bot locally sort of interesting that's cool pie is important
507. 1:28:06yeah parmesan and cream cheese actually those two together sound pretty good right now um hello everyone hello cam catkin from
508. 1:28:16Cleveland how are you cam catkin from Cleveland had cataracts actually you know what I did this um in the workshop I
509. 1:28:33taught tonight we we made I showed Automation and so for the automation because a lot of the people in the
510. 1:28:38workshop are voice actors so we made we made an automated tool that um you give it a topic and then it will write you a tongue twister
511. 1:28:51so write me a tongue twister for my friend cam catkin from
512. 1:29:12Cleveland clever cam katkin from Cleveland carefully concocted a colossal cluster of crisp cranberry cotton candy
513. 1:29:20continuously confusing shattering comrades with his chronically clever cranberry concoction Charisma oh this is going to get
514. 1:29:32obnoxious I'm going to do this every time I see you now wait the AI learning lab getting obnoxious never I see so many AI apps
515. 1:29:44that are just riffing on top of chat GPT yeah so Mary Mary I call I call them rapper apps or I don't they're called rapper app
516. 1:29:52so it's basically just wrapping you take a prompt and you wrap an application around it and it's basically just
517. 1:29:59someone pre-ot a prompt and is calling the open AI API like most of those apps you could just build your own in in
518. 1:30:08zappier and and and I think you should because you should like like just learn how this [ __ ] Works hey mind repeating the plot plot
519. 1:30:20do.ai so it's it's hang on it's plot p l o t plot d o t dot period AI so
520. 1:30:38the dot dot is is one is the word and one is the what were you saying about plot dot so what I was saying about plot do.ai is um
521. 1:30:53um it's so if you if you know story Frameworks like save the cat or or any any of those um screenwriting story
522. 1:31:01assistant tools where you have to like fill out all the forms right like you know what's the opening inciting
523. 1:31:08incident and what's this and you know what are your character arcs right it g it gives you all of the [ __ ] you have to fill
524. 1:31:14out and if you're like me if you've got add you start those story Frameworks about 17 times and you never finish them
525. 1:31:22because they're [ __ ] exhausting right so what plot dot has done is they've taken one of those basic story
526. 1:31:29Frameworks and every one of the modules you have to fill in there's a write it for me button so you can just go give me
527. 1:31:37ideas and it'll write [ __ ] and then you can edit them and then you move on to the next segment and it will incorporate
528. 1:31:44what you wrote before for this segment and so you basically just walk through the um the the story framework
529. 1:31:53development and then I think it even moves into writing you know scenes and dialogue but I'm not I'm not quite sure
530. 1:31:59about that but at a minimum you'll get a fully fleshed out um story framework with you know character breakdowns and
531. 1:32:08plot arcs and you know just all the [ __ ] you need to write a good screenplay like all that structure is really important
532. 1:32:13it's just a a pain in the nads to deal with but not anymore thank you very much appreciate that appreciate thees what are your thoughts on
533. 1:32:26short-term memory of chat GPT how far can it remember in a conversation my thoughts are it really sucks and I I
534. 1:32:32want them I want it to have longer memory um these things will have memory I I think the memory thing is really
535. 1:32:39interesting like do you want it to remember everything you've ever told it or just certain things at certain times
536. 1:32:45and it's probably certain things at certain times and then how do you actually deal with that so there's an
537. 1:32:50interface challenge with memory um and and then I'm sure there's a technical challenge with memory but the way it works is here let me let
538. 1:33:01me talk with my hands let me talk with my hands and and we'll explain it with my hands but I got to get a good picture up
539. 1:33:08here hang on hang on people AI improv we'll do that welcome to chat add I'll be here all week don't forget to tip your waitress try the
540. 1:33:26chili does he have tourettes um okay so I need a visual hang
541. 1:33:42on okay this is good all right so this box of Lego pens that I bought in Prague let's say this represents um 3,000 words
542. 1:33:59okay so chat gbt has a context window of what they call 4,000 tokens and you're like but Kyle what's a token it doesn't
543. 1:34:07matter the Geeks have decided that a period and a word are each a token and so it's 4,000 tokens which about 3/4 of
544. 1:34:17that is words so you have 3,000 word memory so so think like you have a box that you can stick 3,000 words in so
545. 1:34:25when you write your first little prompt that's you know some amount of words and then it responds and that's
546. 1:34:31some amount of words and then you write another prompt and that's amount of words and then and then basically it keeps filling up that
547. 1:34:37memory and if you've used chat GPT long enough sometimes you'll be working with chat GPT and you're like you know you
548. 1:34:43told it that your idea for the Great American novel and then it's going along Gang Busters and then all of a sudden it
549. 1:34:49seems like it forgot what the hell the novels about that's because when you told it about your idea for the Great
550. 1:34:54American novel as your prompts fill up that bucket they go out of the memory and so it's it's only sort of prompting this
551. 1:35:04amount of [ __ ] back to back to open AI so all of this [ __ ] gets lost so that's why it kind of loses its mind sometimes now um
552. 1:35:15Claude has a memory of 100,000 tokens so chat gpt's 4,000 you know Claude is like 100,000 so it's 75,000 words so to put that in context
553. 1:35:28The Great Gatsby is a 50,000-word novel so in Claude you could take those 50,000 words paste it into the prompt and have
554. 1:35:36it write an epilog for The Great Gatsby you couldn't do that with chat GPT you couldn't take an entire screenplay and paste it in here what
555. 1:35:45happens when you paste a big document in is it immediately blows out the the memory and it says oh sorry that's
556. 1:35:51that's too long I can't deal with it so natively gp4 is supposed to have uh a 32,000 token context limit so gp4 is
557. 1:36:05going to be going from about 3,000 words to about 25,000 words so you know much more um but it's not anywhere close to
558. 1:36:14100,000 tokens so that's how that [ __ ] works and then and then that memory is only live within the cont text of a
559. 1:36:22single chat so when you chat about your great American novel here and then you go into a new chat it doesn't have any
560. 1:36:28memory of that old thing it's not stringing your conversations together across conversations it can and you know people
561. 1:36:37that are building applications with Lang chain and things like that where it's just dynamically reembedding it's like taking all the
562. 1:36:46answers from this conversation embedding it throwing it into the large language model so then you can GP P that later
563. 1:36:52like there's ways to do it but it's not Auto Magic right now it's not it's just we're we're in the janky piece of [ __ ]
564. 1:37:01stage of the industry all these tools right now like what you're asking for is absolutely reasonable shouldn't these
565. 1:37:08things have a little better memory why does it forget my story we were going along so good and it didn't know who Bob was Bob's the main
566. 1:37:17character how could it forget Bob well Bob was an [ __ ] Sydney didn't like Bob Sydney Sydney killed Bob off what are you gonna
567. 1:37:35do oh man H I got catfish to Cleveland but not by cam katkin exactly that's great um I know better from you and this
568. 1:37:51this community that's awesome Mary Mary thank you that's appreciate yeah this community is turning out to be pretty
569. 1:37:57[ __ ] swell I'm digging it I need to learn about how to use zapier zapier's okay hey so here look I'll give you a
570. 1:38:05little sneak peek of what we so oh this is wild I'll show you this too if you haven't seen this um oh I got to probably download it from
571. 1:38:20the cloud yep I do hang on let me download this from the cloud cloud look at the look at the fingerprint
572. 1:38:29collection it's a little managerie of my little fingerprints do you like it I've made it just for you I I point I talk
573. 1:38:37I'm I'm a hand talker what are you gonna do man some someone said I forget who I was talking to they were like Kyle's a hand
574. 1:38:48talker isn't he and they're like how did you know it's like like every thumbnail on my YouTube channel I've got my hands all up in my
575. 1:38:57face oh man what are you gonna do what are you gonna
576. 1:39:12[Music] do okay I'm not going to show you a lot of this cuz cuz this the zappier [ __ ] is is um really complicated it it it it's it's
577. 1:39:28not so much that it's complicated it's that it's um it's really janky it's it's the problem with no code tools is
578. 1:39:39that in order to make them powerful enough they've got to put a lot of interface in front of you um and
579. 1:39:51and what you effectively do when you're dealing with no code tools is you're still doing programming you just don't
580. 1:39:57have to learn the actual syntax of programming language but you're still having to think about variables and how
581. 1:40:03to insert variables and [ __ ] like that but I'll try to break it down for you simple so this was one of the one of the
582. 1:40:08slides from a section I taught tonight on automation so uh image was generated in um Dolly 3 of course because you know why not
583. 1:40:24um okay so the tools the tools that we used are um zappier and open AI playground um and the way it kind of
584. 1:40:37works is this when you're doing these automation tools there's basically there's basically three modes there's an input mode which serves
585. 1:40:47as a trigger so something happens that starts the automation and that could be you add a line to a Google sheet you add
586. 1:40:55a row to a Google sheet right uh and maybe that comes from you know someone entering something in some system
587. 1:41:03somewhere and when they enter it over there it gets added to a Google sheet and that can trigger the automation so
588. 1:41:09that's your input another common thing might be you create a Google form somebody fills in a Google form and when
589. 1:41:17they hit submit that submission becomes the input and the trigger for the automation so they call those
590. 1:41:24triggers and then just as a as a really simplified thing I'm I'm doing a super simplified example here the middle
591. 1:41:33section the processing section you're going to take some data from that input and then you're going to mix it into a
592. 1:41:40prompt and you're going to send that to open AI to to do some chat GP J chat GPT in now you don't have to use open AI
593. 1:41:49right this is just a thing that you can do in zapier zapier's got like 5,000 modules so think of these as like Lego
594. 1:41:57blocks so in this case Lego block one is Google forms that's the input Lego block two is the chat GPT Lego block that
595. 1:42:06we're going to send a form or send a prompt with some input from the form over to chat GPT and we're going to get
596. 1:42:12back a result right so that's our second Lego block and then the third Lego block is we're going to stick that in the
597. 1:42:19Gmail and we're going to email it and so um I'm trying to think where if I go to zier well no let me go
598. 1:42:34to Google forms forms. google.com forms. google.com so I made this form called the tongue twister maker and it's a
599. 1:42:47single answer form what subject would you like your tongue twister are about um let's see um oh wait I got to make this a hang
600. 1:43:02on I want this tongue twister about [Music] um Lamar The Llama oh no we'll do cam
601. 1:43:18katkin from Cleveland [Laughter] this is never gonna end cam all right so cam cat from Kat can from Cleveland is
602. 1:43:29what we want our tongue twister about so I'm going to submit that okay so that was just a Google form I just submitted
603. 1:43:35that in about 20 seconds I'm going to have an email in my inbox with the tongue twister but how did it get there so let's go to
604. 1:43:45zappier so you go to zapier and you got to log in cuz if if you don't log in you're a loser don't be a loser log [Music]
605. 1:44:00in all right so within here here's my zaps so a zap is you know a combination of these tools so if I go into this
606. 1:44:09simple example zap that we use tonight here's the Google form here's the chat gbt thing and here's the Gmail so if I click on this
607. 1:44:19Google form over here on the right hand side are the details for it so if I I can go up here and I can say edit but basically what
608. 1:44:28happens is you you connect your Google account to to zapier and basically you you define in this form that when someone submits the
609. 1:44:41form that begins a trigger and you take whatever they submitted in the form and that now becomes available for your conversation ation in
610. 1:44:51chat GPT and so I've got a prompt in here that says what what subject uh that says write a tongue twister about and then
611. 1:45:01it's the input from the form so the test in here was 150b lovable dog but what I just entered was was cam Katen from
612. 1:45:09Cleveland so it's going to take that with Cam catkin from Cleveland it's going to send it to open Ai and then
613. 1:45:16it's going to return a tongue twister and then in Gmail I've got here's the subject here's your tongue twister about cam katkin from
614. 1:45:26Cleveland that's the subject and then here's the body of the email hey there you asked about a tongue twister about
615. 1:45:32cam katkin from Cleveland here it is and then here's the reply that came from chat GPT so that's how you build the
616. 1:45:41email so now if I head over to my email we'll get to do this again with Cam katkin it's our favorite thing to
617. 1:46:02do so here's your tongue twister right so here's my email and Gmail and it says here's your tongue twister about
618. 1:46:13cam katkin from Cleveland cam Ken from cam Ken from Cleveland cunningly caught a cluster of colorful caterpillars calmly consuming chamomile
619. 1:46:24on a cool cloudy day in my email bang like that that's Automation and that's that's kind of the most simple version of it but now imagine
620. 1:46:35that where you can put in branching logic and you can do you you could do a thing where you ask someone five or six
621. 1:46:42questions about a topic and then you take some combination of their answers and you send a bunch of different calls
622. 1:46:49to open Ai and and you you generate a blog post and you generate tweets and you generate you know so you could make
623. 1:46:58yourself content engines with this right so this is essentially what Jasper and copy AI are is they're they're basically like
624. 1:47:08automations like this this is all they're doing they take a prompt they take your input they they put your input
625. 1:47:15whatever you typed in they wrap it inside the prompt they wrote send that to open AI and then return that as you
626. 1:47:22return that to you as their magic thing well you can go make your own and what you'll learn in making your
627. 1:47:29own is super valuable right now because everything's changing so fast so you're going to be able to be more Nimble if
628. 1:47:37you're actually learning the concepts all right people listen listen listen we need to talk okay I got my
629. 1:47:46vitamin C I got my vitamin C right here I know what you want you can make money with was that AI generated
630. 1:48:01maybe um okay can you turn that tongue twister into an image oh that's brilliant okay yes as a matter of fact we can cam Kat can from
631. 1:48:16Cleveland all right we're going to do new chat we're going to go to Dolly 3 oh wait let me do this
632. 1:48:34here I'm creating a children's book of tongue twisters for every letter of the alphabet and I want your help creating clever
633. 1:48:50images that bring the tongue twisters to life period the first image I'd like you to create is for this tongue
634. 1:49:18twister please give me four images in distinct styles that would be appropriate for children's book and make the images landscape
635. 1:49:32orientation boom cam cat cam katkin from Cleveland cunningly cunningly caught a cluster of
636. 1:49:48colorful caterpillars calmly consuming chamomile on a cool cloudy day I mean the thing about Del because
637. 1:49:57it's got prompt coherence where where like it'll take those details I would be very surprised if all of our images didn't have it might have
638. 1:50:06Cleveland it'll have colorful caterpillars yeah there's Cleveland in the background I I didn't get Cam cat of course cam katk can's a
639. 1:50:33cat I think that's the one I think I think this is you Cam cat cuz while that one's cute this this feels this feels more tongue twister
640. 1:50:48appropriate now you want to see something something insane okay great I like style number four a lot period let's make sure that
641. 1:50:59anything we do from here forward uses style number four please write tongue twisters for a
642. 1:51:14b c and d or no not c a b
643. 1:51:31d and e all right of course Andy alligator at apples and avocados a top an antill and April Bella bunny balanced okay um
644. 1:51:46that's one two three four uh uh great can you make those images now and it should remember the style and so now we now we've got the
645. 1:52:03first we've got the first five images of a 26 image book with tongue twisters I mean how what world are we in when did this
646. 1:52:17happen like two weeks ago that's the the answer is two weeks ago it's when it happened the cat body may disappear only smile and eyes
647. 1:52:32[Laughter] appearing so let's see how we do with our with our A's and our B's oh you know what we should also tell it to do can
648. 1:52:41creating the image be incorporated into the zap output uh I think it can I don't know I don't know if uh if um Dolly 3 is incorporated into
649. 1:52:57zapier yet it probably is I just haven't had time to look at it yet all right these that's that's the style that's not
650. 1:53:06that's not for sure that is so two of the styles are wrong but anyway I mean without even really trying
651. 1:53:15we could have something to do lazy baby llamas m for L lazy baby llamas all right let's we'll do we'll do L all right we'll do
652. 1:53:27um okay one more period lazy baby mam wait oh
653. 1:53:46llamas I said write a sassy twister Lulu and Lulu and Luna the lazy Little Llamas Lounge leery lick licking lemony
654. 1:53:56L lollipops and laughing at Looney lullabies okay perfect perfect let me see four variations and make two of
655. 1:54:13them darker in tone uh darker in emotional tone something went wrong why why
656. 1:54:28okay now it's good couldn't Thelonius Google sheet of the lab's URLs be automated with zappier eg possibly yeah I don't know what you do
657. 1:54:43with it um yeah was well it it depends what you want to do like his Google sheet we we could certainly set up a Google sheet
658. 1:54:52that every time he adds a line to that it posts it to a specific channel in Discord for example you could do that
659. 1:55:06um that's perfect fun live tonight can creating the image be incorporated into a zap output it I think it
660. 1:55:17can but but I don't know if if Dolly 3's been um put in there yet those do not look like llamas the mysterious Melodies of past Lis that's
661. 1:55:36cute all right people is this the same site you can use to update your profile photo on social media is this the same
662. 1:55:44site you can use to oh no this is different um a lot of those profile picture updaters on social media are
663. 1:55:53just apps that are built on top of stable diffusion and dream Booth um there's one that I did I did a
664. 1:56:01bunch of images here you see some stupid images of me um sure Kyle like we can't get enough of your face let's see some more of your
665. 1:56:16face oh my God here we go these are so
666. 1:56:30dumb this this app's called reminy it does like these you basically sort of just pick a model and you upload your
667. 1:56:39stupid pictures to it and and then like there's me as the Connecticut boat club dude look I'm an advertising
668. 1:56:53guy I like hot dogs M hot dogs is tasty more hot dogs look life's tough it's tougher when
669. 1:57:08you're stupid this my it's my Kevin cner this is my best Kevin cner oh
670. 1:57:23God all right I think I think we should we I think that's got to be enough we're not going to be able to unsee [Laughter] that oh
671. 1:57:34man all right people listen this was fun tonight let me get on Get on Up out of here let me get on up out of here lordy
672. 1:57:44exactly Sher D thank you for the rose Sherry D I appreciate that good hanging out with y'all I think your cowboy needs
673. 1:57:51to be your new look I should probably go there I think I think having that look involves me losing about 40 lbs and
674. 1:57:59gaining about 6 Ines love you guys um do you want me to start writing kids books what would be your oh do you want me to
675. 1:58:08start writing kids books what would be your process from A to Z I mean I kind of just did it the chat GPT is really
676. 1:58:16good at writing kids books so I you know if especially if you have do inside chat GPT I'd just sort of write the book and
677. 1:58:26then start making the images um throw it up on Amazon self-publishing and go right I think you should set yourself
678. 1:58:38a goal you want to write kids books I you should set yourself a goal here's here's your goal if you want to write
679. 1:58:44kids books write a kids book and publish it to Amazon before Sunday night cuz what's stopping you well I
680. 1:58:56don't know an artist well you don't need one now well I don't know how to write well you don't need to
681. 1:59:02now Irregulars and Milds present Turtles and [Laughter] llamas all right everybody let's see who we got Amelio's wife good to see you
682. 1:59:14tonight Amelia's wife was in the uh in the workshop the the uh the what was it called the AI Launchpad workshop with Ronda um thanks
683. 1:59:23to sherbear who who couldn't make it and gave her her seat to to uh Amelio's wife that was super awesome rock paper
684. 1:59:32scissors good to have you Clint Copperhead of course have a good night Source Camp good to see you Don Smith
685. 1:59:38who else we got here who else we got here Joker cam katkin of course how can we forget cam katkin from Cleveland uh um what's up talk Dylan
686. 1:59:53Tobias he Tobias is no longer muted he's happy he was not have me the other night Corey good to see you Source Camp Karen
687. 2:00:03White good to see you all right everybody um what night's tonight Wednesday so tomorrow's Thursday you know what that means don't you we're
688. 2:00:13going to be here at the AI learning lab you know what we're going to be doing learning bring your notebook books
689. 2:00:19people knowledge doesn't grow on trees it grows in chat jpt peace out everybody