aiLL July 25, 2023 - The AI Learning Lab: Cybernetic Friends and the Good Life

Transcript

1. 0:06foreign [Music]
2. 0:34welcome to the AI learning lab my name is Kyle Shannon let's get this show on the road I don't know what day it is it's Tuesday right
3. 0:45Tuesday not Tuesday Tuesday welcome welcome welcome welcome to the AI learning lab where we
4. 1:01learn about labs hmm doesn't seem right does it um that's a nice ice Java for the evening
5. 1:18but Kyle won't take you up at night I don't know maybe maybe it will maybe it won't maybe I'll sleep like a little
6. 1:25baby all right welcome everybody my name is Kyle Shannon this is the AI learning lab a little bit about my backgrounds hello Kevin Earthling
7. 1:36or are you are you the alien are you the the far out one we're the we're the earthlings is that how it goes hmm I see
8. 1:47um so welcome welcome welcome on this here Channel we hang out we have fun I make jokes I make funny voices I make
9. 1:58funny faces but I'm starting with this so you know and uh we talk about AI I'm a long time entrepreneur
10. 2:07I've started 12 or 15 companies in my day I was a uh an entrepreneur in the early days of the World Wide Web I
11. 2:13started one of the first digital agencies company called agency.com and I was Chief creative officer there and so I have been
12. 2:21through one of these Cycles before and I didn't think I'd see another one in my lifetime because that one was pretty
13. 2:28remarkable the early days of the web riding up to the.com bubble and the burst and all that happy horse [ __ ] and
14. 2:36uh then Along Comes This Here generative AI stuff I kind of had a false start with blockchain and nfts where I'm like
15. 2:42this is gonna be big and it was funny the deeper I dug into blockchain and nfts the less real it seemed I like the tech just the
16. 2:50companies and the project sucked but the more the deeper I dig into this generative AI stuff the more real it
17. 2:57seems so this is this is uh this is a big deal it's transformative um and uh yeah so I'm happy to talk about
18. 3:04it so if you have questions about um whatever AI pop them in the comments below I do my best to answer everything
19. 3:11let me show you what I got going on in this little window [Laughter] um
20. 3:29all right let's get that to focus shall we can we there we go all right grab a screenshot of this so if you're new to generative
21. 3:41AI if you haven't played with chat GPT these are your seven options this list used to be three and now it's seven and that was a week
22. 3:50ago it was three and for like two months there was three things on here and now there's seven why because shit's happening so
23. 3:59and I'll I'll give you a quick rundown on these so if you've been here before you've heard this before so you can you
24. 4:05know take a little nappy all right chat.openai.com that's the official chat GPT website if you've not played with it
25. 4:12if you've not played with the original the one the only the original Chachi BT that's it right there go there and just start playing
26. 4:19bing.com is also chat GPT it's gpt4 but it's built into Microsoft search engine Bing so you click on the chat button
27. 4:28that's gpt4 you can generate images it can analyze images or read them like you could upload a picture and have it write
28. 4:35advertising copy for the picture and it'll just do it uh it's connected to the internet um Chachi BT has some pretty cool things
29. 4:43if you pay for the plus version which I'm not connected with any of these I don't make commissions on any of this
30. 4:48[ __ ] so don't get all cynical you have [ __ ] trolls all right claude.ai the one on the other side there I would say is the biggest
31. 4:58competitor to to chat GPT right now um it's a bunch of X open AI employees that invented gpt3 and uh it's got a
32. 5:09hundred thousand token context window which basically means you can upload like 75 000 Word documents and it'll
33. 5:17just read it and then you can interact with it um bard.com is Google's entry which Bard was laughably bad for months and months
34. 5:26and months and then last week it got good turns out they replaced the core large language model with this thing
35. 5:32called Gemini that's also also multimodal it can read images then labs.perplexity.ai that's metas llama2
36. 5:42model that you can just play with at perplexity they just put it up there so everyone can go play Pi dot AI they have a cool app that you
37. 5:49can talk to it's very conversational pose got a lot of bots and you can play with lots of these different models so
38. 5:55there's a lot here they're they're all different flavors of the same thing so if you don't know at all what you're
39. 5:59doing ignore all the [ __ ] on the bottom just go to the top one if you want to start exploring the different flavors of these models and
40. 6:06see what's good for certain things that's your list so screenshot that okay now screenshot this too can we see everything yeah
41. 6:15so the thing at the top prompts.chat's just a document that'll teach you about prompting how to talk to those large
42. 6:21language models we just talked about openart.ai promptbook that's going to give you instructions for how to deal
43. 6:28with mid-journey stable diffusion Adobe Firefly Leonardo this sort of teaches you about these text to image models
44. 6:37like how [ __ ] powerful they are then you got futurepedia.io that's like four thousand AI tools if you just want to
45. 6:44geek out go there and then these are the image tools and then at the bottom left is runwayml that's video and blockade
46. 6:51labs skybox.blockade labs is just a fun parlor trick right now you describe a scene and then you're in this 3D world
47. 6:59it's pretty slick it's not super sophisticated but it's slick slick slick slick slick and then finally uh this is I'm gonna
48. 7:11get rid of this thing because I fixed the I fixed the URL so the salon the AI Salon is a community that I started last
49. 7:21December uh that's a community of people on this AI adventure together and so if you want to hang out with people there
50. 7:29just go to that thing there's a link tree we have meetups we meet every other week we've got a Discord server we have
51. 7:34a LinkedIn group storyline.com that's my current company and that's my my Twitter my ex Twitter handle my current X handle
52. 7:44Elon isn't that funny he changes the comp a bunch of trademarks on X because of the Xbox who knew well apparently the trademark office
53. 7:59I guess that cage Max just the cage match between those two the stakes just went up it's pretty funny all right so hop
54. 8:15hop your uh Auto GPT it's dead I don't think so isn't it just an open source project let's go see did it die did it die I'll
55. 8:26show you that thing too look at this cool video here see that video that started out as a panoramic photograph that I took
56. 8:34that's at Runway ml that I had down below there but what are you saying Auto GPT dead Auto GPT dead question mark um I don't think so
57. 8:51no I don't think so I mean maybe it is if you know anything post it in there Auto GPT is one of those autonomous agents I
58. 9:00think those are going to be a big deal um all of those things Auto GPT baby AGI the initial things that came came out
59. 9:09were super promising but they were also it's also early like this stuff's it's very early for this stuff so
60. 9:16all right let's get to the questions let's go to the video team oh I can't read with this shirt I think I'm gonna
61. 9:22have to go change my shirt so is AI a bubble certain to pop as well no no no when you have a piece of technology that produces
62. 9:36easily 30 to 40 percent productivity increases overnight that ain't going away the fact that there's seven competitors
63. 9:47on this list right now when a week ago there were ostensibly three means that this isn't going away and and
64. 9:54there's a there's a whole bunch of reasons for it so Google invents this thing in 2017 called the Transformer which allows companies
65. 10:04to train machine learning models on massive data sets open AI from the beginning said why don't we train it on as much of
66. 10:16the internet as we can get to all of the public internet hey the bear what's happening thank you for the Roses very
67. 10:21sweet I gotta go change my shirt I can't read any [ __ ] on it um so they invent this thing called the Transformer and so
68. 10:31and that was in 2017. so for the past what are we in 2023 now six years so for the past five years they've been getting better and better
69. 10:42and better at this they did GPT one and they did gpt2 then they did gpt3 that one made some noise and then right around the gpt2 to GPT 3
70. 10:54time is when these companies like stability Ai and mid journey start doing some of this image Generation stuff
71. 11:03and that goes from like laughably bad to like oh that's pretty good except the hands all look like this
72. 11:10and then now all of a sudden the hands kind of look okay and now mid-journey images are photographic quality in fact
73. 11:18what was it was it Vogue I think it was Vogue just did their first AI generated cover um then November 30th 2022 a day that
74. 11:31will go down and infamy chat GPT launches now from a technical perspective chat GPT is just a very simple
75. 11:43application layer sitting on top of this large language model down beneath in fact the instructions for how to create chat GPT
76. 11:52we're in the documentation for gpt3 in the documentation were instructions for how to create this thing and no one
77. 12:01created it Sam Altman figured after a year year and a half someone would have built it nobody built it
78. 12:09so so Microsoft's getting ready to make their announcement about their billion dollar investment into open Ai and Sam Sam Altman says to his
79. 12:19Tech Team hey could you just go whip up that chat GPT thing two weeks they built it in two weeks because it's really simple
80. 12:28two weeks two dose dos weakies they launch it and six weeks later they had a hundred million users that changed everything
81. 12:40they didn't expect it to have 100 million users Microsoft didn't expect to have 100 million users Google certainly
82. 12:46didn't expect to have 100 million users so by adding a simple interface to deep deeply powerful technology
83. 12:57this thing's revealed into the world that just makes instant um awareness of the tool and instant panic in all of the competitors to get
84. 13:07their [ __ ] together so metagos we're not going to talk about the metaverse anymore we got AI stuff
85. 13:14look here's some free [ __ ] go play with this Google's like we invented this stuff we have all the AI we're good and
86. 13:20then they released a bunch of shitty shitty AI and then last week they pulled out the good stuff right so it took them eight months to
87. 13:31get it to a usable level um Claude dot AI they raised 450 million dollars um pie this cute little you can get an app that
88. 13:46you talk to that cute little chatbot pie dot AI they just raised 1.3 billion dollars to make a 22 000 GPU
89. 13:57server intense thing so that thing that looks really cute right now that's gonna have some [ __ ] muscle behind it um
90. 14:08Microsoft is rolling chat GPT throughout their entire product line from cloud services Azure to um it's already in GitHub as co-pilot
91. 14:19it's going to be called co-pilot within Office 365 within Dynamics 365 within windows so gpt4 is going to be built into
92. 14:30windows so here's the deal if you're a Windows person I'm a Mac guy but you know whatever if you're a Windows person
93. 14:38you know how when you want to just like change from like a light light setting to a dark setting and then you like
94. 14:45Google it how do I change from a light setting to a dark setting on Windows and they tell you and then it was like six
95. 14:51versions of Windows ago and then like an hour later you're still trying to find the stupid [ __ ] switch in the 837 000 dialog boxes
96. 15:01that's gone you'll just go change my settings to dark it'll do it Google's Gonna Roll barred across all of
97. 15:12the office all the Google products Google Docs Google Excel or Google Sheets all of it um go go look at the YouTube video barberheimer
98. 15:27someone did a mashup of Barbie and Oppenheimer we played it last night it's [ __ ] hilarious just go watch that all
99. 15:34AI generated in the hands of really talented people good storytellers thank you for the heart so no it you know
100. 15:43is is AI over hyped yeah I think it's over hyped but is it a bubble that's going to go away like like are all of
101. 15:50our jobs safe no they're not they're not these are the most powerful tools ever put in the hands of like normal civilians
102. 16:05it's funny they're calling they're calling this moment the the Oppenheimer moment for for AI because a lot of the you know these guys
103. 16:14have raised a lot of money and the tools the next generation of these tools are going to be powerful enough
104. 16:20that they're equating it with like the importance of the bomb but who had access to nuclear bombs everyone no people in the military
105. 16:33who's got access to this all of us all of us and it's not just sitting behind firewalls the the Llama 2 model
106. 16:44that meta just gave away put out an open source with a commercial license means that anyone can build anything
107. 16:51they want with these things without all the safety guard rails that the other guys are talking to Congress about the the horse
108. 17:00done escape with the barn okay um so yeah so it's not going away um how quick is it coming I think it's
109. 17:13coming a [ __ ] ton faster than any of us are prepared for um and and you know here's the other thought I have on this is I think long
110. 17:27term if I go if I go kind of if I jump ahead five years the amount of new jobs that that these things are going to create are probably
111. 17:35going to be insane the amount of prosperity let's just call it that the amount of prosperity that that large language models and sort
112. 17:42of this flavor of technology is going to create is going to be absolutely insane the path to get there I think is going to be really painful
113. 17:51there's going to be a lot of disruption there's going to be a lot of jobs there's going to be a lot of entire sectors that just go away
114. 18:00I mean I'll tell you A big one um customer service call centers I think have a very very real possibility
115. 18:14that they kind of disappear overnight because if if if go play with pie go play if you haven't played with pie yet
116. 18:26if you have an iPhone I mean you can do it on the web but it's it's not as slick an experience go to pie
117. 18:32on your iPhone and put it into voice mode and just start talking to it and tell it to act like a uh a customer service person for your
118. 18:45refrigerator like figure out your refrigerator make and model and I just talked to Pie as if it were a
119. 18:52customer service person and just start having a conversation with it just like you called an 800 number
120. 18:59so imagine a future say two years from now where when you call up Samsung about your shitty refrigerator door that fell
121. 19:07off like mine did four months ago like my wife opened the door the door literally fell off onto the floor it was like a year and a half old
122. 19:16mm-hmm she was on the phone for hours on hold got transferred to people this and that and today and it's a disaster right
123. 19:28we've it's every all day every day if you try to call someone you can so imagine a future where you call Samsung
124. 19:35and they're like hey how can I help you um I got a problem with my thing I'm sorry to hear that how you
125. 19:41know what do you want me to do or like describe the problem and that's a synthesized voice just having a conversation with you you
126. 19:48weren't put on hold that voice knows everything about Samsung everything about it it knows where its boundaries are and its boundaries are
127. 20:01way bigger than layer One customer support layer two layer 3 layer four you know you have to dial four different
128. 20:08people before you get to the human being and and that's this excruciating four-hour phone tree [ __ ] nightmare
129. 20:16press one to this let me put you on hold oh did we cut you off sorry about that start again what if all that went away
130. 20:29what what happened look so like all of the companies that program those phone trees press one for this press two for that establa Espanol
131. 20:41all of those companies White now can they pivot and do all sorts of this [ __ ] yeah probably but there's gonna be massive disruption so
132. 20:54so I think it's coming and it's coming fast um we'll see we'll see but I don't think it's a bubble I think
133. 21:03there's a possibility it's under hyped all right I gotta go change my shirt I can't read the text I will be back
134. 21:44ah much better much more better welcome welcome welcome everybody my name is Kyle Shannon this is the AI learning lab
135. 21:54we are learning about AI well I'm just pontificating right now let me get some more questions that list has gotten so long I know
136. 22:01right Jesse that list is insane and it's like like at some point I'm gonna I'm gonna thin it out but
137. 22:08like like the each thing has a different strength like chat GPT has code interpreter I wouldn't give that up
138. 22:15claude.ai has got a hundred thousand token window I wouldn't give that up Bing and Bard both do image recognition
139. 22:23and Bing's connected to the internet so's Bard Bard's actually doing really sophisticated web link injections into um answers
140. 22:35um and then perplexity that's that's the Llama model Pi's really conversational Poe you can make the cool box I feel
141. 22:42like everything there has at least one feature that's worth exploring on them so it's it's [ __ ] nuts and this was this was in a week this
142. 22:52happened in a week I love not GPT yeah I still listen G 's gonna have a place in my heart oh by the way did you see apple announced I told
143. 23:01you this was gonna happen I told you they were being a little too quiet well they're not quiet anymore they've created a thing called Ajax
144. 23:10which is a AI development platform and they've created what they're calling internally at Apple Apple GPT
145. 23:19if they're going to come out swinging because what does Apple have apple has silicon that can run machine learning
146. 23:27models go watch their Vision Pro demo notice that they didn't say the phrase AI once but they use things like machine
147. 23:36learning and embeddings and all the little catch phrases they're running [ __ ] locally with their little R1 ship that's taking all that
148. 23:45sensor data and machine learning the [ __ ] out of it in real time on the device so Apple's going to be able to
149. 23:54make a very strong privacy and security claim and their shit's gonna be good so when their stuff comes out and it's going to
150. 24:03be on your data I guarantee you that Apple's gonna make this I I think the large language model stuff from Apple
151. 24:09the GPT stuff is going to be you know your own personal Kick-Ass robot assistant um yeah crazy AI bores me I'm old give me some
152. 24:25easy that'll inspire me really Steve it bores you I think I fear you have not dug or pushed the harder you push on this stuff the
153. 24:39more impressive it is to me but I'm curious what you've tried and and it's quite frankly it's easy actually you know it might be fun Steve
154. 24:46if if you've got an iPhone or even if you don't just go to the website go to Pi dot AI or haypi.com go there
155. 24:56and then put it into voice mode so you can talk to it and start talking to it and just talk for half an hour and ask it [ __ ] anything
156. 25:07all right by the way pi.com Pi dot AI is also connected to the internet did you know that mid-journey
157. 25:14has an interesting bug it's probably got a lot of them but tell me tell me what the interesting bug is a beautiful woman
158. 25:21select an old photo 40s 50s on the internet at random all right good place for develop oh good place to develop prompt skills
159. 25:34um I'll tell you a good thing so so go to uh go to prompts.chat that will give you some good primers on prompting
160. 25:45and then it'll it'll teach you about uh telling chat GPT to be a certain Persona I want you to act like an architect or a screenwriter or a
161. 25:56therapist or whatever it might be um and if you pay for chat GPT plus if you do the 20 bucks a month at chat.openai.com
162. 26:07you now have access to this thing called custom instructions so you can take those personas and you can make your
163. 26:14chat GPT like understand who you are how you want it to behave so every chat that you do with it will start acting like
164. 26:21that so understanding that um that Dynamic is good is an important thing the other thing I would do is go
165. 26:29to po.com and if you go to po.com you can create your own bot and it's really simple you just name it pick a model
166. 26:37there's like four models available you can just experiment with them you make a description of the bot you
167. 26:44create a prompt so I want you to act like Emily Dickinson and whenever anyone asks you anything I want you to write
168. 26:49poems like Emily Dickinson or stories like Emily Dickinson um and then you just launch it and it's out for the world to play with
169. 27:00you can tell people go play with my bot and if you get the prompts right on those Bots they're really gratifying they're really cool
170. 27:11I did a sling blade bot and he's pretty good he's disturbing I wish I had the hardware for llama 65. I know all these things here's the deal
171. 27:25there are so many developers like the the Llama models are the number one trending um project on hugging face right now
172. 27:34probably on GitHub too um the the developer Community is pouncing on that [ __ ] so um probably within the early versions in three months
173. 27:50probably something decently usable in six months they'll have those things so [ __ ] optimized you'll be able to
174. 27:56install it just on your phone it's like it's we're gonna we're gonna be okay we're gonna the the fact that this
175. 28:02stuff's now open sourced and you have this many competitors and this is minus Apple and this is minus AWS
176. 28:10this is minus any other well-funded AI companies that might deserve to be on that list but just haven't they're still in stealth mode
177. 28:18so we're gonna have we're gonna have 10 to 12 major players doing these large language models um and so that that just means these
178. 28:28things are going to get better and better and better and better so um but I'm right there with you like I want to be able to two
179. 28:44two uh I'm a mind reader what did I mind read you on like where should you start was it that stuff that's why this channel exists I want
180. 28:53people I want people to come here to be like I don't know what to do all right go do this just go do this
181. 29:00and then the next question they ask is how can I make money you can make money you know how you make money with chat
182. 29:10gbt you don't listen to those [ __ ] foreign copy image address and paste it in mid-journey with the word macro all right what does
183. 29:24it do if deal if deal or deal one deal A1 dla one tell me what it does I don't like if it explodes my Discord I
184. 29:37don't wanna I don't want to [ __ ] up my Discord live just tell me what it does can you show the previous screen one sec I didn't get
185. 29:44a screenshot sure [Music] okay so there's that one go ahead and screenshot that bad boy
186. 30:00that's all the these are all the chat GPT variants the top three things there are two educational resources and then just
187. 30:12a big ass pile of tools those three things are image generation tools and if if you're sitting there being a
188. 30:23smart ass going what about stable diffusion Man stable diffusion's cool well that's what Leonardo is Leonardo is
189. 30:31a web interface in front of stable diffusion um the only reason I don't have stable diffusion there is it's kind of a you
190. 30:37have to be a decent geek to get it up and running these days so Leonardo is just the easy front-end on front and in
191. 30:44front of that but I'm happy to talk about stable diffusion and then runwayml is the uh is the video tool that does amazing cool
192. 30:53stuff and getting better every day and blockade Labs is kind of cool too and then that's that's the uh AI community that I created back in
193. 31:03December all right that's really helpful the text prompt info good absolutely yeah it's here's okay here let me let me give you here's
194. 31:18another thought on prompting because because prompting is a really interesting thing for one thing the whole idea of prompt engineering you
195. 31:24know everyone's like prompt engineering we're gonna you're gonna be a prompt engineer it's all about prompt engineering
196. 31:29I don't think it's it's gonna last that long I think there will be prompt Engineers but I think that as these applications get more and more
197. 31:38sophisticated we as end users will have to do less and less sophisticated prompt crafting right now because all these
198. 31:46tools are so soon so new and raw we have to figure it out we have to figure out how these things actually
199. 31:54think and how they process and it it's how you talk to these things is not how you talk how you do a Google search right
200. 32:02I talk about this on here a lot when you do a Google search over the past bunch of decades search engines have taught us make your prompts
201. 32:12as concise as possible don't make them overly complicated or you'll just get a mess right so so you like figure out
202. 32:18what are the minimal amount of words I need to get in here these things are kind of the complete opposite
203. 32:22the more context you give them the more you explain yourself the more you say please and thank you believe it or not
204. 32:28the more you sort of tell it I want you to act like this kind of person and then we're going to try to solve this kind of
205. 32:35problem and now I want you to you know help me you know walk me step by step through that process they get really good
206. 32:43um it's very counterintuitive I my experience working with chat gbt and the like is more like I got an incredibly smart simple-minded intern
207. 33:00it's an idiot savant like I have to explain everything to it like oh you don't know where I work right
208. 33:08right we've only been working together for six months you don't know where I work although the thing that chatgpt
209. 33:13just added fixes that but but still you know you you basically tell it everything and explain everything and
210. 33:20kind of excruciating detail and then when you do that it [ __ ] knows everything it's been [Music] all right someone wants to pop up let's
211. 33:31let's talk to someone I think I'm gonna lose my I think I'm gonna lose my that my dual screen but oh well she went away all right
212. 33:44any search you are aware of that could be helpful for someone looking for business AI business integration um
213. 33:52it depends the level of AI you're talking about if you're talking about doing sort of chat GPT integration and things like that
214. 33:59um I can't imagine that there's a certification out there right now because this stuff's a little too new um it's possible I know that
215. 34:09Stanford has something MIT has something Facebook has something Google has something I don't know that those are really for
216. 34:18Integrations though um but if you're on the integration side there's there's enough there's enough Geeks that have been working with
217. 34:28with AI for the past decade that that you can you can likely find resources so so there's probably certifications about
218. 34:37doing a lot of the infrastructure stuff but if you want to do things like Lang chain some of the more
219. 34:43you know recent projects that are still like that you know they're kind of flying the plane and building it at the
220. 34:49same time like Lang chain is this thing that's super powerful but it's got all these piece parts and everything keeps
221. 34:55changing and they keep changing out the open source large language models and you can put in different kinds of vector
222. 35:01databases and [ __ ] like that so um so I don't know I wish I could be more help with that um no I would I I would
223. 35:13you know it's of all of these the place to maybe go ask that is Bard Bard seems to be they seem to have integrated the
224. 35:22Google search indexing into the answers at Google barred really well and it's connected to the internet so Bard might
225. 35:29be the place to go ask about um certifications the other thing if you're a Microsoft Shopper or if you've
226. 35:36got any kind of Microsoft infrastructure not opposed to it a lot of the developers I know are opposed to it but if you're not opposed to it
227. 35:44um they're probably the farthest ahead in terms of integrating chat GPT into Azure into all the cloud services
228. 35:55um but any like they just had their their developer conference it's not all it's not all rolled out yet so it's this
229. 36:02stuff's all very very new but Microsoft probably has some decent resources at this point because they've
230. 36:10been working with open AI for what a year and a half something like that because it was a year before we knew anything about it
231. 36:16all right let me get to some other questions welcome everybody welcome where do you work I've been studying AI
232. 36:24models and emerging properties I work I'm a founder and CEO of storybine which is not an AI company although we are adding a lot of AI
233. 36:37I'm completely reimagining my product and my value proposition of my 11 year old company thanks AI [Laughter]
234. 36:47I I'm more an entrepreneur so I'm not this channel what my expertise is is more on the once you hit that inflection point where
235. 36:58powerful tools like uh GPT 3.5 and 4 are accessible to the rest of us that's when I like jumping into this stuff
236. 37:08because I'm not technical enough to be in the game before that but the minute this [ __ ] shows up ready
237. 37:14for prime time which I think this stuff is now I think chat GPT is the is the inflection point so November 30th of
238. 37:21last year so that's that's what this channel is about um so if you want someone more technical I I the the guy that I think is I find
239. 37:30the most impressive that does anything you know remotely like what I'm doing education wise is David Shapiro and
240. 37:35YouTube Demands a [ __ ] genius and he's got all sorts of Open Source projects doing all sorts of heading toward AGI in an open source way
241. 37:45um and he's just [ __ ] brilliant he does like three or four major presentations a week on the implications of AI from a deep deep
242. 37:57intellectual and Technical point of view the dude's a [ __ ] machine he's amazing um can you recommend anything to help
243. 38:06with trading in regards to AI I can't really what I would say is Go to futurepedia.io go there and look for anything around
244. 38:16trading and then also the other thing you can do is go to whatever your trading platform is especially if
245. 38:21they've got Bots if you can do trading Bots and things like that you can just copy the code from a trading bot on your
246. 38:27platform paste it into chat GPT especially gpt4 and just tell it to explain to you what the bot is what it
247. 38:35does and then tell it what you want it to do and it'll it'll rewrite it for you so so the other thing the other thing about um
248. 38:44about chat gbt code interpreter you can upload up to 100 megabyte data file so you could take say historical trading data
249. 38:54upload that into a code interpreter and have it start doing data analysis for you so there's all sorts of things you
250. 39:02can do just out of the box with some of these things that are [ __ ] insane I don't know that world at all so I don't know the specific
251. 39:09tools so that's why futurepedia.io might there might be some people that are doing some really sexy stuff uh I just don't know about it
252. 39:21um you can ask GPT to rate your prompt oh and it will suggest improvements as well yeah that's actually a good hack
253. 39:28you so you you write a prompt you you ask Chad GPT if if there's a better way to write it you can also when it answers
254. 39:36you have it analyzed its own answers say review your answer and tell me how you might improve it and it'll do that
255. 39:43and then you can tell it to improve it and then it will what did I do today that was really frustrating um
256. 39:51well I was having it I was having it try to um to rewrite and this was with Claude it wasn't with gpt4 to be to be fair um
257. 40:01I was having it rewrite a document that I wrote and it just kept truncating like all the paragraphs that
258. 40:09would sort of knock them down to two sentences instead of four or five and and then it would just omit certain
259. 40:14paragraphs it was weird I couldn't I had a frustrating day today with chat gvt or Claude whatever it was I was playing with um
260. 40:23most days are are revelatory and fun but today yeah yeah not so much not so much I know what's wrong with it ain't got no gas in it
261. 40:38Adam McKnight coming in strong with the memes we like the memes that was the bot website what was the bot website agreed thanks
262. 40:48uh can you recommend anything to help with trading I already answered that no I can't what was the bot site uh poe.com po.com has cool Bots
263. 41:03[Music] all right so you go to Poe and you got six different chat gbt models and then you just have a [ __ ] ton of bots you
264. 41:19have this explore button which if you click it it just takes you over to all these Bots and basically all these are are like pre-prompted
265. 41:27GPT things I don't know why this isn't focusing did it Focus no um well whatever um like there's a mid-journey bot
266. 41:39there's a ghost bot there's an emoji bot that only answers in emojis there's cooking Bots and then you can make a bot create a bot
267. 41:48and then you just name it give it a description give it a prompt give it a welcome message publish Bang
268. 41:55and then don't be a doo-doo head and not have a picture go to Mid journey and make a picture for it like I did yesterday I did Christopher
269. 42:03walkenbot I've done what plots have I done uh Shakespeare I am Shakespeare so if you go to poe.com I am Shakespeare
270. 42:16I did Jay-Z that's a good one Yoda Sling Blade Emily Dickinson Steve Jobs I am Steve Jobs that's a fun one
271. 42:25I am Mary Shelley I'm Virginia Woolf or not I'm just Virginia Woolf I am Poe so yeah that's Poe that's Poe [Music]
272. 42:42paid Discord I don't know what you mean the image wait it merges pictures of humans with bugs in a very interesting way
273. 42:54mid-journey does if you put an old 40s or 50s photo in it that doesn't surprise me but that's kind of cool
274. 43:04huh is that a bug are people considering that a bug interesting interesting aqua blue hi everyone hello
275. 43:17I got SD running I'm a geek yeah listen stable diffusion is awesome so so if you want to do the this image stuff
276. 43:26and you want to geek out on it um you know I talk about leonardo.ai is the is the web version of stable diffusion
277. 43:33um if you've got the stomach for getting a Google collab up and running and and you know getting automatic 11 11 up and
278. 43:39running and the code stuff doesn't scare you off like it took me I failed nine times trying to get stable diffusion up
279. 43:46and running I finally got it running and I got dream Booth working and you know it felt like [ __ ] victory yeah I did
280. 43:53it yeah [ __ ] yeah after nine [ __ ] failures but that's just the stage of the the industry we're in right now
281. 44:00where the tools are really powerful the nice thing about stable diffusion is the developer communities building new ship
282. 44:06for it all the time which is also the downside of it right because you get your automatic 11 11 up and running and
283. 44:13then they changed the notebook on you and then it doesn't work anymore because they swapped out a library and this and
284. 44:18that now it's using a different model and you just constantly have to stay on top of it so if you don't want to do
285. 44:24that that's what these that's what these tools are like Leonardo that's just a company that's sticking the front end
286. 44:29they're taking care of all that back end [ __ ] for you and that's what you think for right you pay a service for them to
287. 44:36solve your [ __ ] headache of having to install it yourself that's how business works yeah but stable diffusion is quite amazing
288. 44:46do you think Leonardo's are in the works for llama too well that's what labs.perplexity.ai that's llama2
289. 44:55that you can play with the 7 the 13 and the 70 billion parameter version of llama two right there for free right now
290. 45:02go play now you can install it and do all the training and [ __ ] like that do I think there will be a Leonardo kind of
291. 45:09front end for for those things absolutely absolutely and and I you know I think probably in in a short amount of time
292. 45:19I feel like the next phase of these is going to be the the the training on your personal data phase right where right
293. 45:27now it's all just like they're they're either quasi or not at all secure um they're jankier than [ __ ] they're
294. 45:35like really thin application layers um so I think what's gonna happen is this you're gonna have you've already
295. 45:42got the Lang chain Community out there doing you can embed your own data into Lang chain and there's there's projects that
296. 45:50are putting front ends and for on in front of Lang chain you've got GPT for all which allows you to play with
297. 45:56different open source models it's just a double click install um you're gonna see apple Microsoft and Google all come out with versions of
298. 46:07these things that are now connected to your data so like Google just say just point your Google Drive Point this [ __ ]
299. 46:14at your Google Drive and you can now start interacting with your data Microsoft's going to do the same thing
300. 46:19Apple's going to do the same thing and then the open source Community is going to [ __ ] pile on right because the the
301. 46:26foundations for it are built it being built right now so I think six months from now um these things are going to look and
302. 46:35act very very different than they do now they're going to be much more in the neighborhood of a personal assistant or
303. 46:41it's just going to be much more useful like I know somewhere in the past decade I sent an email to someone talking about
304. 46:48this specific thing and I can't find it in Google and you'll just be able to go find it now because it's it's in your [ __ ] and
305. 46:56they've got it in the [ __ ] Magic Machine they've embedded it once it's embedded magic [ __ ] happens right all right
306. 47:12Miss Birch he takes a little while to go through the questions my regular my regulars are apologizing for me thank you very much I am slow
307. 47:22with them but I do try to answer them all [Laughter] um Rachel awesome Jack Strickland it will be tremendous
308. 47:33Improvement when they can tell you when not to trust their answers yeah so if you're new to this stuff that's an
309. 47:39important point that I forgot to make earlier um these things do what's called hallucinating which is a the Geeks came
310. 47:46up with a clever word to not make it sound like such a [ __ ] disaster that it actually is they just [ __ ]
311. 47:52outright lie these things will just tell you stuff confidently even if it's wrong I like to describe it as mansplaining as a service
312. 48:01like all the good men in your life it'll confidently tell you something even if they've their head up their ass um
313. 48:09and it's the the nature of how these things actually work they they are they are as aware of what they say to you as a calculator is of the
314. 48:21answers it's giving you it's not aware it's a calculator that's doing some formula you know in its circuitry and it pops up an answer
315. 48:31that's all these things are doing from a technical perspective it's literally predicting the next word in a sequence
316. 48:38of words based on your prompt so when it writes you a whole page of stuff that looks like holy crap that's brilliant it knows all this stuff
317. 48:47it's not actually aware of what it just created I think the the next evolution of these there will be a layer there where it
318. 48:58actually looks at what it wrote and says is this [ __ ] if [ __ ] go back to the starting hey start over um they're gonna get better right
319. 49:12I'm I'm looking at how Bard is incorporating search results into its into its results and how Bing is doing that
320. 49:21they're starting to do some some citations and things like that you know link link to websites um so we're probably not that far off
321. 49:30from that but yeah I I am right there with you that it will be a tremendous Improvement where you can actually trust
322. 49:37these things especially when you start to get into like like Google just came out with uh was it Google yeah yes
323. 49:46they just announced Med Palm which is a medical large language model right it's it's one of these it's like chat GPT but
324. 49:54trained up on all the medical [ __ ] well we don't want the medical [ __ ] hallucinating and telling us to take you know 640
325. 50:03you know Advil to to make the headache go away that would be bad I am not a doctor but I know that wouldn't be good where are you answering
326. 50:15from I am well Patrick cutaway your way down there I'm way up here all right so do you ever see Star Wars the
327. 50:23original one the four the four five six ones they're like the good ones yeah that was when I was a kid you know at
328. 50:30the beginning of that where you had words down here and then there was the big long line of words and then way off
329. 50:36in the distance I'm answering questions up here you're down here all right do we have mods as in mods for The Tick Tock no as in
330. 50:52modifications for these Dr J I don't know what you're talking about Dr J no mods oh yeah like mods like someone in here actually like oh is
331. 51:05the question is the AI learning lab a Professional Organization no and if one of you down below there said
332. 51:15what are your qualifications boy I ain't got none ain't got none other than the fact that I'm geeking out on this [ __ ] every
333. 51:24[ __ ] day of my life for the past nine months actually it's more like a year and a half now old black and white photographs contain
334. 51:31different information uh than the command macro and mid Journey oh that's interesting all right I'm gonna have to look that up I wanna
335. 51:40write that down old black and white photographs in mid-journey that's kind of cool I like it I wonder do you think it's
336. 51:49I wonder if there's something about the 50s images that it's that it's it's somehow like doing references to like
337. 51:57old 50s sci-fi because you said it was people and insects being mixed I wonder if there's something about black and
338. 52:03white old sci-fi Stills and things like that that are mixing in there he is Colorado I'm in Colorado yes I'm in Denver Colorado
339. 52:14welcome to Denver it's the base of the Rocky Mountains did AI begin with Quantum Computing no AI began with oh I don't know
340. 52:29back in 1956 the term AI was coined 1957. I just did a spreadsheet on this do I have it on on this machine yeah yeah
341. 52:46that's on my work machine um I just did some research on this um that started with these things called
342. 52:55neural Nets right so basically starting 70s or 80s they started trying to replicate the structure of neurons in our brain and and the basic
343. 53:08idea has been if we can get computers to act like the synapses and like the neurons in our brain we might be able to get them to do
344. 53:18cool stuff and they were making you know nominal prod progress and you know they were doing things like you know expert
345. 53:26systems like IBM's big blue that beat Gary Kasper off at chess because they just trained it you know on that really specific narrow data set
346. 53:35and then it was the Transformer architecture that that um Google came up with in 2017 that is what burst the dam open for all of what
347. 53:48we're doing now so the embeddings and Vector databases just how it does what it does like it takes the
348. 53:57the mass of everything it learned and it turns it into these funky mathematical word pairings that are sort of connect a
349. 54:05word to its meeting but just in math and so when you type in your prompt it you know magically reconstructs
350. 54:16something out of nothing it's it's quite remarkable what it does um but but uh Quantum Computing as I understand it I don't know a ton about
351. 54:26it it's a different kind of mathematics that it does so so you can't just use a quantum computer like you use a regular
352. 54:33computer so you couldn't you couldn't necessarily just do AI machine learning training kind of stuff with a quantum computer but
353. 54:42at some point some will figure that [ __ ] out Google just released the uh the results of their their new quantum computer Sycamore
354. 54:52which has 70 qubits in it which if you follow the world of quantum Computing every qubit is a Quantum increase in
355. 55:00power so 70 you're getting up there and it performed a mathematical equation in six seconds that would take today's most
356. 55:10powerful supercomputer 47 years so that's what's going on in the quantum Computing world right now I don't think
357. 55:18it's too connected to this AI stuff but I could be wrong I just but I I did a video on the Sycamore thing and then a
358. 55:25bunch of people in there go listen you [ __ ] tool it's not the same kind of mathematics you can't just use quantum
359. 55:32computers on your stupid little chat gbt [ __ ] tool you [ __ ] loser so I took that as constructive criticism
360. 55:41and I learned that I'm an idiot but still it's doing something in six seconds that would take the fastest
361. 55:52computer 47 years that's kind of cool you know that shit's gonna get in here somehow all right um no like he can even see this chat
362. 56:02yeah I can see this chat Patrick cutaway you bitter bitter man does Patrick need a hug do you need a hook do you need a
363. 56:11hug I really am reading all the questions I really am live the bear can vouch for me thank you the bear I appreciate that
364. 56:21and by the way if you have not seen the TV show The Bear shame on you we talk about comedy in this channel we
365. 56:29talk about good TV shows the bear is remarkable remarkable remarkable remarkable the race is on with all of them
366. 56:39who to predict who the overall winner when the dust settles I don't [ __ ] know I know it'll be three of them it's
367. 56:44always three right these things always settle down to three which oh which AI can produce a PowerPoint
368. 56:57presentation the one that I know is Tome t-o-m-e I think it's tome.ai hang on and they've added a bunch of new features now tone tone.app
369. 57:13go to tome.app yeah tone.app and then it'll just you can just say I'll say create your watch create a presentation about
370. 57:28um my what's the presentation about where do I put it here we go um my what
371. 57:47um my business to wash wild raccoons so I'm starting a new business to watch wild raccoons as you do
372. 58:05so it's generating an outline it's called raccoon Suds the wild business of washing and then it creates you um an outline
373. 58:17and if you're happy with that I guess you can go in and edit these now oh that's pretty slick so introduction the problem the solution
374. 58:24the market the marketing strategy the conclusion you say continue and now it makes you a presentation
375. 58:33so it writes them and it does the images so we should get some raccoon images because there's lots of raccoon base art um [Music] cute
376. 58:58ah there you go we did it now are these presentations any good I don't know probably not all these tools right now all of these
377. 59:13tools right now are a hundred times more powerful than anything you've played with and radically flawed
378. 59:27[Laughter] so so if you do it right you should have a Kevin McAllister moment right from Home Alone this moment
379. 59:40I didn't know I could do that and then as you start to use it for real stuff you realize oh it's got these
380. 59:47limitations and those limitations and those limitations but what I would say is that the the be very I would be very reticent if I were you
381. 1:00:01to not dismiss these tools because they're not perfect because they're as bad as they're ever going to be they
382. 1:00:06only get better from here and what they do is absolutely [ __ ] miraculous like we're lucky to be able to play with this stuff
383. 1:00:14you can see the chatty's just not caught up thank you Stephanie I've got my Defenders here he really does the
384. 1:00:20Institute questions he really is live he cares about us he does good evening Mr Shannon good evening all right
385. 1:00:30Tick Tock just jumped me to the bottom again my regular viewers know that I hate that because I lose my place and I
386. 1:00:37like to answer all the questions contrary to popular belief if you know what I mean where's he answering these from I don't
387. 1:00:46think he's answering these maybe he's an AI you think he's an AI I heard he's chat GPT are you chat GPT
388. 1:01:01[Music] I did what did I do I did a tick tock for I did a tick tock for April Fool's Day right where I had chat gbt come up with a
389. 1:01:13with three other words then generative free train transformer that meant that GPT was actually just really like a big
390. 1:01:21call center full of full of people on computers feverishly typing out the answers a couple of people got it other people
391. 1:01:29was like I don't think that's really happening man people are just wound too [ __ ] tight on the internet they people got to
392. 1:01:36lighten up gotta light up and fly right all right let's hair hair G reached level something 's hair growth geek I don't know you
393. 1:01:51reached some sort of level congratulations I love this [ __ ] LOL you're very entertaining and knowledgeable thank you Patrick
394. 1:01:58appreciate that Kyle is the sheer sheer sure she oh [ __ ] yeah I'm like man these kids these days they
395. 1:02:12got all these words that mean something I don't I don't know I'm the [ __ ] I appreciate it thank you thank you thank you Miss Birch
396. 1:02:21Chachi BT got Dumber once they connected to the government we get the bad AI now oh gbt 654321 uh you got some work to do on the name but uh
397. 1:02:36to go download [ __ ] llama too it's open source and commercially usable don't be so bitter it's not all a conspiracy it like here's
398. 1:02:50the deal with all those conspiracy theories they all sort of require that the government is competent at doing conspiracies
399. 1:02:58the government's got their heads so far up their ass oh yeah they don't know what [ __ ] Wi-Fi is yeah whatever see it just
400. 1:03:13it's probably more Microsoft than the government by the way putting putting the shackles on chat EBT but now you have Claude
401. 1:03:24now you have barred now you have that now you have Pi I mean you've got options the whole open source Community is there for you so like
402. 1:03:34the the your comment I don't mean to pick on you but your comment is it's really easy to demonize and dismiss this [ __ ]
403. 1:03:45and I think that does us all a disservice because if you really take your time and learn this [ __ ] learn what it's good at
404. 1:03:54learn what it's not good at learn where the boundaries are learn if indeed it is getting better or worse and then you decide that stuff cool
405. 1:04:06but if you're just sitting on the sidelines going ah it's a big government conspiracy no that doesn't serve any of us
406. 1:04:13because the more of us that use this stuff for good are gonna lean the direction of where this head's in the future toward the
407. 1:04:22good if we sit on the sidelines and let these [ __ ] just you know make all the rules up then we are going
408. 1:04:30to be at the mercy of it so let's not do that and we might as well have fun with it like it's the most remarkable [ __ ]
409. 1:04:37technology in the history of humanity we get to play with it there's like [ __ ] seven of them now that we can play for effectively for
410. 1:04:44free almost all of these are free mid-journey's 10 bucks a month you can make world-class images for 10 bucks a month
411. 1:04:54like full-on photographs full-on illustrations you know what I mean so don't [ __ ] on this stuff until you've
412. 1:05:09actually used it and listen if you've been using it for 20 years and you got some information then [ __ ] share the information
413. 1:05:17if they hobble chat GPT and you know it in a specific way cool but you know in its honest to God if I just if if we
414. 1:05:26shut everything down you know from chat GPD down if we shut everything down and I just wanted the chat GPT 3.5
415. 1:05:34I feel like I personally could spend the next decade trying to figure out all the different ways I could exploit it
416. 1:05:44and gpt4 is insanely better so even if gpt4 goes from like whatever it's at right now you know 92 and it drops to 85. it's still
417. 1:05:59a hundred times more powerful than any of the [ __ ] else thing I'm using so uh what do I think about bgsm I don't know bgsm what is bgsm
418. 1:06:14um [Music] biosig Technologies I don't know anything about it they doing some sort of crazy ass brain
419. 1:06:29[ __ ] some drug [ __ ] biosig Technologies I don't know enough about it certainly if you're asking for for investment I don't know
420. 1:06:44why did sound draw come up is that sound draw is that part of that company I don't think so I don't know I don't know anything about
421. 1:06:50it sorry about that I keep breaking GPT doesn't finish generating about 75 percent of my input um yeah that's that's an issue so I
422. 1:07:00would say go go start playing with Claude so Claude has a context window of um a hundred thousand tokens so it's got a
423. 1:07:10much longer memory per chat session and it can also generate uh long responses better so try Claude um it's worth playing with Bart honest
424. 1:07:20honestly right now there's so much so new on this list that that what I would do if if I were anyone here that's really exploring these
425. 1:07:30things if you're using it enough that you're you're getting chat GPT to break I would start exploring different tools
426. 1:07:38and see which ones are better right I think Bing is really really hobbled and and there's like a handful of things I use
427. 1:07:47Bing for but for the most part I don't because it's just so hobbled Claude is really impressive I've been I would say
428. 1:07:54right now I'm I'm kind of 50 50 between using chat GPT and code interpreter and then Claude just for day-to-day use
429. 1:08:05um but I'm impressed with Bard and how it's incorporating index links into into organically into search results without
430. 1:08:12it being spammy SEO [ __ ] um I'm actually impressed with that um the Llama models are raw and new so you can go play with those Pi is very
431. 1:08:23very conversational um so but my my instinct is if you're having truncation issues try Claude what was the bot site that was po-po.com
432. 1:08:37is access to Discord a paid subscription um I don't no I don't think so I don't think it is but but the ability to generate more than like
433. 1:08:51the 20 initial test images in mid-journey is a paid subscription you can pay for Discord subscription to
434. 1:09:02get some other features but I don't I don't think you need it I think I do pay for it though but not
435. 1:09:08because I did it I started paying for it back when I was doing blockchain and nft [ __ ] and I just haven't canceled my subscription
436. 1:09:16Man stable diffusion is awesome it's so there's so much stuff going on I know bill so so if if you if you have the the
437. 1:09:23stomach for doing the geeky stuff to get into the stable diffusion Community like there's so much development going on
438. 1:09:30there and if you can keep up with it but it's it's a lot and it's but it's good it's good um if you just want Kick-Ass images use
439. 1:09:40mid-journey if you want flexibility and to keep up with the latest and greatest ghost stable diffusion if you want to do
440. 1:09:45things commercially go with Adobe Firefly and generative fill within Photoshop because it's trained on a safe model
441. 1:09:53sorry sorry Leonardo has voluptuous young women bias I know kind of disturbing how it was probably trained I I don't well
442. 1:10:02see how it's trained it might be just that a lot of the people using it right now are using all of the all of the sort of
443. 1:10:11anime models anything that was sort of done on anime just skews way toward you know tits and ass and yeah it's it's
444. 1:10:20it's annoying the the thing about Leonardo though is you can train your own models right so you can upload a set
445. 1:10:26of photographs or images that that you like train your own models and then start working within that so there's
446. 1:10:34good flexibility there but yeah if you look at the initial feed when you land on Leonardo it's just like it's just
447. 1:10:39like a bad anime convention so okay God Kids come out of the basement really um come on make a velociraptor or something I mean you know
448. 1:10:57all right the bug is that it creates bugs or plants and merges it with the photos will do it on any photo I gotta
449. 1:11:03try that that's fascinating let's go try it now if you're still here if deal if deal image so let's go do that um let's see
450. 1:11:1850s photo graph um I don't know uh what is this called uh
451. 1:11:32Diner oh no I don't want the actual Diner no uh 50s people pictures
452. 1:11:47here we go that's British life in the 50s
453. 1:12:18grab this one all right so I don't quite know how to do this because I'm not talking to you but 50s so I'm going to save it there we're
454. 1:12:29going to hop on over to Discord oh crap
455. 1:12:54I guess 50s desktop [Music] why is it not showing up oh there we go
456. 1:13:09okay so we're gonna pop that in there I'm gonna upload it I suppose I could have just grabbed that URL but I know I
457. 1:13:19know this works so I'm going to copy the link I gotta add a text prompt okay slash imagine is
458. 1:13:34um let's see a dressing room Beauty all right I'm gonna let that go I got is
459. 1:13:49it doing yeah it's waiting to start all right I'll be right back foreign
460. 1:15:21foreign all right what did it do did it make any weird insects I don't think it did I don't know it didn't do it all right
461. 1:15:35I'm back sorry about that [Music] if you want to use stick to version 1.5 is that in mid-journey
462. 1:15:51weird well I'll look it up that sounds like a really interesting bug interesting bugs are cool like why does it do that you know
463. 1:16:08I CMP the description of a character from a novel into mid-journey and I came back with a weird alien yeah I mean
464. 1:16:19so if you understand how the diffusion models work they're they're the trick is in getting your your prompt right
465. 1:16:30if it's too many words if it's not enough words if if you're if you're basically leaving the latent space too
466. 1:16:37big and Broad it kind of does these averages of all these noised images and then denoises them and so you'll get some weird [ __ ] um
467. 1:16:47it's just the nature of the Beast right now I mean it's part of the joy the joy of it also right like you know creating [ __ ] like
468. 1:16:57you know like these images that I did in stable diffusion with dream Booth I mean those are super cool you know and that was just kind of a
469. 1:17:14you know experiment that I've probably done I don't know 2500 or 3 000 images then I finally got one
470. 1:17:22that I was like holy [ __ ] that's [ __ ] art you know and like that's the kind of art I would create if I had the talent so
471. 1:17:32but it's the nature of the Beast like I said these these these tools are um remarkably powerful um and remarkably flawed right now it's
472. 1:17:44just we're just very very early so so just know that they're as bad as they're ever going to be know that
473. 1:17:54have you noticed there's a lot less hands like this than there used to be why um because they're fixing that [ __ ]
474. 1:18:07you're not slow you're just thoroughly in your elaborated answers thank you cocoa beans oh 9.5 on the on the name spelled cleverly
475. 1:18:19um parsable and funny uh so 9.5 out of 10 on the name cocoa beans I like it uh can these platforms help someone find a new job um
476. 1:18:30the platforms can't help you find a new job but I think that they can certainly help in the process so for example [Music] take your resume
477. 1:18:45and maybe your LinkedIn history like you can download your LinkedIn history you can just copy and paste chunks of it into a document
478. 1:18:53and then with something like Claude you could just upload resume One resume two LinkedIn profile I think you can do up to five documents
479. 1:19:08and then paste in a job description and tell it you want it to use markdown so that it looks pretty marked down is
480. 1:19:18the thing that will format it I can show you how to do that tell it to rewrite your resume you know based on today's best resume writing
481. 1:19:26practices you know focusing on the jobs and experiences for that job that you put in there as a description and then have it
482. 1:19:34write you a personalized cover letter to that hiring manager so you do stuff like that right you could do stuff like I'm working with a
483. 1:19:43guy right now who's who's building a tool for sales people to be able to take their LinkedIn profile and create a custom persona for them
484. 1:19:54so that the sales people know how to talk to them what to talk about what they value all that sort of stuff just
485. 1:20:01based on their LinkedIn profile um so you do the same thing with hiring managers right or or job descriptions
486. 1:20:10um with things like Bard that's connected to the internet or Bing that's connected to the internet you can potentially find jobs that you
487. 1:20:20might not have otherwise found um so that's on the job seeking side if you're getting AI literate if you're
488. 1:20:30getting up to speed on these tools start talking about it start posting on LinkedIn start posting to your Facebook
489. 1:20:36groups start talking about it at dinner parties or wherever with your friends even if you're like well my friends
490. 1:20:44think ai's stupid and I don't want to look like an idiot look like an idiot start talking about it because what's
491. 1:20:50going to happen is at one point one of your friends is going to say oh they're looking for no someone who knows
492. 1:20:56something about AI at my job do you you talk about that Chad gbt thing are you good at this yeah um you know join groups join groups that
493. 1:21:09are exploring this stuff get good at it um it's going to be in demand like the next two years there's going to be a lot of
494. 1:21:17demand for people that have even a semblance of a clue of how to use this and you don't need to be technical this
495. 1:21:24is this is part of the shift right now if you want to build applications that leverage this yeah you got to be geeky
496. 1:21:31and you might need to do a little development but if you're a project manager you can use these tools to amplify your project
497. 1:21:39management skills if you're a writer holy [ __ ] these are like ability amplifiers you take your ability you put it into these machines Bang
498. 1:21:5220 30 times the output like that boom now is it perfect output no but it gets you to 80 in like five minutes
499. 1:22:08and if you're anything like me getting to 80 is somewhere between two hours and two weeks depending on how my adds functioning that day
500. 1:22:17and now I can go to these tools and I can go brain fart bang eighty percent like that so get good at that [ __ ] and start
501. 1:22:27talking about it and here's another thing to do especially if you're if you're unemployed if you've if you've got time
502. 1:22:34put a lot of time into the searching but start volunteering if you know someone who's struggling to get you know
503. 1:22:43traffic into their small business or you know someone who's trying to write the Great American novel you know someone
504. 1:22:49who I don't know is just trying to figure out how to garden better volunteer to help them solve whatever
505. 1:22:54the problem is and tell them that you're going to use these tools to do it would you mind if I help you write a marketing
506. 1:23:01plan for your business sure like I'm [ __ ] desperate at this point no one comes to the shop anymore okay
507. 1:23:08let's give me an hour tell me all about your business and what's going on and what your problems are and then just
508. 1:23:14over the weekend go use these tools come up with some [ __ ] genius ass marketing plan because this will help that'll do it
509. 1:23:24and then you know you know that does a couple of things it helps someone else it educates them that these robots are
510. 1:23:33not out to get us these are tools we get to use we get to choose how we use these and you know it just might help their business so
511. 1:23:43I think these platforms can help you find a job but it's not about the platforms it's about what are you trying to
512. 1:23:51accomplish and then using these platforms to try to accomplish it not looking forward to storage payments
513. 1:23:57for personal AI data I that's actually a really interesting dilemma because it's like I'm already a data hoarder
514. 1:24:05and and like the thought of getting rid of my prompt history is like no I don't need there's probably something good in
515. 1:24:11there and like most of it's just crap and almost all of it I could recreate like that but I total I totally get it you know
516. 1:24:19we're moving into this world of infinite content generation and we're going to be able to like like the amount of images I
517. 1:24:27I generated for for the art project that I did at the beginning of the year was just massive massive amounts of images and
518. 1:24:37yeah do I want to throw them out no do I really have the storage like no all of my computers are all [ __ ] maxed out
519. 1:24:43and all my iCloud accounts and Google Drive accounts are all maxed out because I keep dumping [ __ ] on them and not
520. 1:24:48cleaning it out now maybe AI helps us clean that [ __ ] up right there so there's a business the the digital hoarder friendly file
521. 1:25:00cleaning service how powered by AI something to go build I've been it's been a day or two since I've given a billion dollar business
522. 1:25:09away for free there's one are you a digital hoarder do you do you have stacks of hard drives you're never going to hook up to your
523. 1:25:18computer again but you refuse to throw them out thanks silver fox hahaha you too can use hoarder friendly file treasure machine AI Dot app
524. 1:25:35[Laughter] three check boxes and Squid you've got all your hard drive space back that would be nice I it's a really interesting point
525. 1:25:50but it's drawing pictures based on your energy portrait um oh if deal I do not no it's not it's not I can actually if you want I
526. 1:26:06can show you how diffusion models make make their images I'm happy to I've got a slide where I talk about how it works
527. 1:26:13I I wish it were on our energy fields it's not quite that fantastical that's isn't there isn't there some UFO disclosure thing tomorrow
528. 1:26:23in the government how many musicians do you think are using this secretly probably a lot I mean if for nothing else just knocking
529. 1:26:36out some Beats probably a lot and and listen to be clear listen like like photoshops have ai Tools in it for
530. 1:26:46a while now a lot of logic and all those tools they've had all sorts of AI filters and things in it
531. 1:26:52this AI [ __ ] isn't new right what's new is the generative AI stuff built you know sort of on the technology sitting
532. 1:26:59behind these these large language models the text to anything sort of magic maker machines that's all kind of new but
533. 1:27:08that's all going to get incorporated into those platforms if it hasn't been already um there's a if if you're interested
534. 1:27:17Pablo there's there's a guy named Ethan Malik and he's a Wharton business school professor and he he he's got a good
535. 1:27:26newsletter he writes amazing articles for his newsletter and he wrote about a month ago an article called um detecting the secret cyborgs
536. 1:27:38and and what what the piece actually is is that these AI tools are um incredibly powerful personal productivity tools
537. 1:27:48that are not really ready for prime time in business so you have businesses not really sure if they're if they can use them right
538. 1:27:58they haven't made a decision on data policies privacy policies usage policies and then you have the individual workers
539. 1:28:04who are using these tools to get way more efficient and not telling their companies about it so they're the secret
540. 1:28:10cyborgs so what he's talking about is is you've got an increasing sort of split in class of AI literate people and then
541. 1:28:19organizations that are not dealing with it and don't even know and and what he's saying is that you know as owners of organizations you
542. 1:28:27should be aggressively going after your people to say it's okay to use these tools we want you to use them we want
543. 1:28:34you to tell us what you're doing but these people are like hey I can knock out in four hours what used to take me
544. 1:28:41all week so I'm just gonna keep secretly using these until you until you bust me so I have a feeling it's like that in
545. 1:28:51every industry including music including art including you know you know if I were working at a big Ad Agency right now and I had my project
546. 1:29:02manager my account team breathing down my ass to generate more [ __ ] in less time like they do I'll [ __ ] use it prove that I didn't
547. 1:29:17if I'm a copywriter am I knocking out [ __ ] fast yeah because that's that business there's it ain't art it's advertising
548. 1:29:29and if I if I'm a copywriter with skills I can have this knockout stuff quickly and then I can curate the
549. 1:29:36best stuff out of that and then do a little tweak here and there and I can knock out in an hour what normally would have taken me all day or
550. 1:29:44two or three days and I might not tell anyone about that so yeah I think there's a lot of a lot of secret cyborgs out there
551. 1:29:54you using their fancy second brain evening Mr Shannon getting in a little late hello Troy Decker what's happening
552. 1:30:03it's a long cerenoid soon AI will be the chat moderator that would be nice yeah I I guess I do need a moderator in here don't I
553. 1:30:13it would suck less but I don't know how to do that so anyone wants to moderate let me know all right there are a couple of Gatekeepers
554. 1:30:22but no mods Vicky it's insects that are giant 10 feet tall that's got to be coming from um the the insect bug if that's indeed a
555. 1:30:32bug if you're just trolling me that's fine too um it's probably from 50 Sci-Fi movies as my guess updating my Wix website today excited to
556. 1:30:42find AI copy Creator now built into its text editor yeah Lauren I found it's it's in um they put it into um
557. 1:30:52um like into LinkedIn posts like if you just go into the post it's like have ai help you but like the have ai help you
558. 1:30:59button is so big that I sometimes accidentally click on it and it takes me into this whole AI writing interface I'm
559. 1:31:05like no I just got like I got like four words I wanna I wanna type I don't need AI for that um yeah we're gonna start seeing it
560. 1:31:12everywhere it's going to start cropping up everywhere and it's you know it's going to be weird it's just
561. 1:31:22I think we all have to we all have to give Grace right now because these tools are really powerful really exciting
562. 1:31:30every company on the planet right now is in a panic to not be left behind any technology company anyone with software
563. 1:31:38and so they're all just jamming [ __ ] in there and so there's just gonna be a lot of bad interface a lot of bad features
564. 1:31:47um and then there's going to be some remarkable ones so I think we just we're just going to be patient for the next
565. 1:31:52year or two while these companies just sort of Panic develop black and white photo macro command in mid-journey
566. 1:32:02black and white photo macro command oh okay let's try that again let's go over to Mid journey we'll give it a shot we'll go slash imagine
567. 1:32:14put that in and I like black and white macro photo
568. 1:32:31let's try that let's see if we get some insects oh yeah hey Kool-Aid oh yeah yeah gen X's any gen xers Kool-Aid Kool-Aid
569. 1:32:47uh do you see in the Bert Chrysler thing where Tom's a girl on two bears one cave or Tom Segura discovers he drinks a 64
570. 1:32:56ounce Kool-Aid every morning that's one of the best episodes uh this is weird it looks like it gave me nothing
571. 1:33:06it's just black right now oh it's at zero percent but normally there's like a little blurb of a of something in there that's weird
572. 1:33:14Quantum machine learning is already a thing oh cool that's cool Seb you can you can make money with Quantum computing no insects
573. 1:33:28no insects buddy sorry I don't know if I did something wrong it looks like a woman's crawling out of a mirror that there's no insects
574. 1:33:47uh here we come inside the day that's funny that's funny honey
575. 1:34:02all right I don't know I haven't discovered the bug the bug bug what's up buddy hey Stevo was shaking uh which to use to find a new job
576. 1:34:18uh I don't know I I would go to chatgpt plugins and they're probably job board plugins in chat GPT in plugins or Bing
577. 1:34:32I mean or Bard I think Bard's probably well barter Bing both of those are connected to the internet so they might be able to find jobs what
578. 1:34:41is this again I don't know what I was showing oh it just jumped me to the bottom dang you tick tock whoa [Music]
579. 1:34:56I don't know why it acts like that he gets very upset when Tick Tock jumps him to the bottom it's very silly he's a silly goose
580. 1:35:05I don't know why why are you so upset just answer the questions you know what I mean [Music] Quantum machine learning is already a
581. 1:35:20thing that's super cool and if it's gonna learn that fast then holy [ __ ] we might have real-time updates to all of our information
582. 1:35:29uh that one we already answered what is this again that we already answered okay bill just a toddler that's exactly right
583. 1:35:43skynet's a piece of spit and duct tape right now skynet's got to get it [ __ ] together if it's ever gonna kill us it's
584. 1:35:50got to figure out how to add consistently and and you know not just [ __ ] blatantly lie all right any suggestions on pixelization
585. 1:36:02removement on videos with AI tools no I mean there's all sorts of good upscalers for the for the image tools
586. 1:36:13right now like stable diffusion's got good upscalers and there's what's that one called um oh I can't ever [ __ ] remember the
587. 1:36:24name of it I don't know there's one you can buy for like 90 bucks that does pretty good upscaling um the video stuff right now is where
588. 1:36:33the image stuff was like a year and a half ago where it's just like the the video stuffs while it's impressive um it like like even the new
589. 1:36:44um Runway ml image to video thing that's really impressive it starts out with a photo you'll notice that the the
590. 1:36:52resolution and detail of the photo drops off dramatically when it starts making making the uh the video
591. 1:37:11um are you a hallucination in my mind I am call center funny I mean call center is [ __ ] [ __ ] the one oh
592. 1:37:27it's not your fault it's the [ __ ] software that runs them and it's the it's the it's the it's the [ __ ] rules right that if you're in a
593. 1:37:35tier one answering the phone after 45 five minutes of someone on hold you're only allowed to read the script
594. 1:37:43you're not allowed to be a human being everyone's always pissed off all the time let's just get rid of all that
595. 1:37:50Roman scammers you cannot trust half of the stuff you read oh too many scanner scammers lighten up all you keyboard Commandos
596. 1:38:01thanks no idea what I did to get there but I guess it's a good thing oh yeah I don't know her growth geek I don't know
597. 1:38:08what you did but you you won some sort of award it didn't tell me the whole whole your whole name it was just Hair G I'm like
598. 1:38:15hair G why do I know that all right aside from your knowledge about AI I'm a talent agent in my opinion you deserve your own show all
599. 1:38:25right Rebecca Johnson sign me up I have a degree in acting I'm in let's do this let's see could we could be a show you know Gen X
600. 1:38:39we could we could call it bitter you betcha [Laughter] are you serious and you are one dm me I'm happy to chat
601. 1:38:51actually someone reached out to me from this that um it could end up being a really interesting business opportunity for
602. 1:38:58Story vines so I'm good with all that I mean listen I I mean it's funny like why am I doing this part of it is I enjoy it like I actually
603. 1:39:14have fun I it's kind of cool like I've been doing this enough now that there's regular people that come in here and hang out
604. 1:39:22that we kind of have that going right there's there's kind of a community building which is kind of cool
605. 1:39:28but but for me this has been there's there's sort of this combination of I need to find a way for myself personally to process
606. 1:39:41how [ __ ] insane this [ __ ] is and and and like I don't want to lose I don't want to lose the connection to
607. 1:39:56how remarkable this time in history is so that's part of it and then part of it is just the thing of like you know
608. 1:40:03people who've never tried this [ __ ] to try it and they come back in here and they're like holy [ __ ] I had no idea it's like yeah
609. 1:40:13I don't know there's something really exciting about it and it's like it doesn't I mean the more you learn about it it's like
610. 1:40:21should we be afraid of it it's gonna it's gonna drive dramatic change it's gonna drive dramatic change
611. 1:40:31and that's scary change is always scary and it's going to be very disruptive but I don't know the more I use it the more
612. 1:40:41I feel I like I maybe maybe it's just a maybe it's just a [ __ ] pipe dream that I feel a bit more in control because I understand it
613. 1:40:50like as I understand it more and as I play with it more and there's something about talking to people in here as they
614. 1:40:55say oh I tried this and that worked and that didn't and that just that idea of just being connected with people that are
615. 1:41:02are playing I don't know feels important anyway but yeah awesome I uh do you know how long I work to get
616. 1:41:11a [ __ ] talent agent when I was acting and writing in fact so so I started ran a theater company for
617. 1:41:19four and a half years never had an agent then I said I know I'll write myself an acting career wrote seven screenplays in
618. 1:41:26two years and you can't get an agent unless you've sold a screenplay and you can't sell a screenplay unless you've
619. 1:41:32got an agent that's the that's the Paradox in that business and um I ended up starting we started a fake
620. 1:41:43talent management company called Outlaw management and represented ourselves we got good enough at the talent management
621. 1:41:49business that we had to take on real clients so so all right the government is competent and people can keep a secret um yeah maybe
622. 1:42:02pieces of it I I mean I would say certainly in the operations some of these big massive conspiracies though nah yeah it's just nah
623. 1:42:16I don't buy it I do not buy it ain't gonna buy it not gonna do it preach it's our job to program it for good that's right dragonfly Alchemy
624. 1:42:29I listen we don't even have to program it it's just how we use it it's how we use it if you're even halfway decent at this [ __ ]
625. 1:42:44help someone else it you'll learn you'll learn what it can do what it can't they'll be appreciative you'll probably enjoy it
626. 1:42:56maybe you make a difference even if you don't you give it a Gipper's try they'll appreciate just the effort like I feel like that's
627. 1:43:07I feel like for every person that sort of gets up to speed on this and helps someone else out with it it's like it's just like the
628. 1:43:14there's an amplification there that feels really important to me so yeah I am going to preach God damn it yeah all right Emilio's wife
629. 1:43:25how are you good evening good evening mid journey is incredible mid journey is incredible no he's tried of the BS oh he's tired of the BS
630. 1:43:36I'm tired of the BS about what we don't understand totally what's happening when it reconstructs no well we we we know what's happening
631. 1:43:49mathematically when it reconstructs and we know we know that it's taking a bunch of noise randomly noised images averaging
632. 1:43:59them together and then denoising them and whatever is in that particular mix of pixels that's what's generating the image
633. 1:44:07we know how it's doing it mathematically and I would assume because we know how the large language model is interpreting the prompts
634. 1:44:19I would imagine that there's a way that they could figure out what in the latent space what images in the latent space
635. 1:44:27were used to amalgamate that image so I would think that the bug that you're talking about is actually discoverable and it's not voodoo
636. 1:44:38but it could be rooted I don't know anyone else here watch all the ghost shows and like want them to yeah thank you um like I
637. 1:44:51would [ __ ] ghost you ever watched the one with the guy Meathead where he's like he's like here we are in the basement you know well
638. 1:45:01it's oh wait how did the line what did the lion do if I only had a name wait what was it name I don't know I haven't
639. 1:45:07seen The Wizard of Oz in too long that was very cute though thank you for the little fuzzy face um but those ghost shows are hilarious
640. 1:45:16Meathead he's he's always like acting like a tough guy and he's like oh it touched me it touched me foreign
641. 1:45:23oh my God that was a show that I that I wanted to do was do a uh do one of those Ghost Hunter shows but it was actually a
642. 1:45:31scripted show so the first six episodes are just um them trying to find ghosts and it's it's all the same [ __ ] you always see right
643. 1:45:40the meter went up the meter went up that kind of [ __ ] and then on the seventh episode like something real happens and
644. 1:45:48then the whole second half of the season is you know increasingly disturbing ghost footage I always thought that would be a good show
645. 1:45:58deserves another look can you experiment with it some more let's work on it um sure I can go I can go play with that I
646. 1:46:04have some proposals I need to read Claude is my friend yes Claude is really your friend if you have proposals you
647. 1:46:10need to read Emilio's wife pop them in PDFs throw them into Claude and just ask now here's what you need to
648. 1:46:17know about Claude it it's not as good at writing as gpt4 but it seems to be pretty good at analyzing [ __ ]
649. 1:46:25so what I would say is take those proposals pop them into Claude um understand them understand you know what are the things that'll make a
650. 1:46:35difference or like what are the decision criteria and then once you have that kind of outline sort of sussed out in
651. 1:46:42Claude then I would go over to gpt4 and write it in gpt4 or collaborate with it in gpt4 but that's a that's a really good thing
652. 1:46:52to do yeah let's go let's go look at custom instructions [Music] oh this was a cool thing that I did so
653. 1:47:06so I did this last night I'll show you this again so see I've got bold text here and like you know sub heads and things like that
654. 1:47:15that's using a thing called markdown which if you don't know you can you can tell chat GPT to use markdown and it'll
655. 1:47:21format [ __ ] really clear when it makes a table like if I just say um let's see make a table comparing Fruit Loops and sushi
656. 1:47:35wait why is it not typing make a table [Music] it looks like Edge has [ __ ] to bed I like Microsoft's Edge but it's got
657. 1:47:49real memory problems hello hello oh God damn it okay make a table um comparing uh Fruit Loops and sushi
658. 1:48:08why is it not making a table um format as a table uh compare Fruit Loops
659. 1:48:24and Cheerios there we go so when it makes a table like that that's using this thing called markdown um and it does that stuff pretty dang
660. 1:48:37good and pretty dang quick okay so couple of new things about Chachi BT that are if you haven't really used it okay if
661. 1:48:51you want to use the new fancy features including gpt4 and code interpreter and all that sort of stuff you have to have
662. 1:48:58a GPT Plus subscription and then once you have that then you get these options um here but you won't get those options if
663. 1:49:08you don't turn them on explicitly so you pay your 20 bucks to get chat GPT plus then you go to your settings and in
664. 1:49:17settings there's a thing called beta features so do that and then you've got code interpreter plugins and custom
665. 1:49:23instructions turn those all on so you have to explicitly turn on the beta features I still can't believe they haven't
666. 1:49:31brought back browse with Bing that's staggering to me okay so one of the things that we just turned on is a thing
667. 1:49:38called custom instructions so if you've done anything on GPT playground on on open ai's playground when you do a chat GPT call there's a
668. 1:49:49thing called system instructions basically what they did was they floated system instructions up to the end user
669. 1:49:55so they have this thing called custom instructions now um how do you spell pompous there we go pompous there we go um so so
670. 1:50:08what I wrote in here was I'm a master strategist and A pompous ass I have a large vocabulary and I Relish in
671. 1:50:15sounding smarter than people around me so so first you describe kind of who you are and so let's say so instead of a
672. 1:50:22master strategist I'm going to say I am a master thespian I'm a master's thespian and pompous ass I have a large vocabulary and relish in
673. 1:50:34sounding um smarter and more dramatic than people around me and then I have and then the next thing is how would you like chat GPT to
674. 1:50:48respond so the first thing is you kind of set the context of who the person is doing the talking so the non-joke
675. 1:50:54version of that could be I'm a marketing manager I'm a you know I'm a botanist I'm a right so so and what that's doing technically
676. 1:51:05is it's narrowing they call it the latent space it's it's sort of narrowing the world of words the sort of semantic
677. 1:51:15pools that you're that you're drawing from every when when the more context you provide the more narrow the search and and that
678. 1:51:23that makes it better and more specific to that thing you're trying to do right so if I want you know medical feedback
679. 1:51:35and I say nothing it's just going to pull from any time anyone said anything about any disease I'm gonna have all
680. 1:51:40sorts of [ __ ] in there but if I say I want you to act as a Harvard trained you know neurosurgeon and I ask it a question
681. 1:51:48about that it's it's it's pulling from a much more limited set that's got more neurosurgeon [ __ ] in it unless yokels
682. 1:51:56out just making [ __ ] up so the first thing is who are you and then the second thing is how do you want
683. 1:52:02it to act so I have here no matter what anyone asks you start out with a pretentious statement that declares I'm
684. 1:52:07smarter than the person so I'm going to say start out with a pretentious statement that uh that uh draws from the plot of a classic
685. 1:52:22Broadway play follow up that declarative statement with an article formatted using markdown complete with titles title subtitles and
686. 1:52:37bulleted lists no matter how trivial the question was that was asked so so now any chat that I do in chat GPT is
687. 1:52:48going to have that instruction in it and it's what's funny is because I've been doing these lives I'll then I'll go to
688. 1:52:54work and I'll go to do something on chatgpt and it does all this pompous [ __ ] and I'm like what the [ __ ] is going on I'm like all right
689. 1:53:01I have my custom instructions so you can turn them off um so at the bottom at the bottom here oh wait did I save these no I didn't oh
690. 1:53:10yeah I did uh so at the bottom here you can turn this off but we're gonna leave it on so we can go see what it's like so
691. 1:53:15let's go to gpt4 because it's a better writer and we'll go um well that's something non-thespian related we'll say um um how much fuel
692. 1:53:31does it take to get to the moon question mark But in in a Saturn 5 rocket right so so that
693. 1:53:46question should yield us a fairly straightforward answer let me spell Saturn right to paraphrase Tennessee Williams
694. 1:54:00Masterpiece the Streetcar Named Desire it seems that we're both caught in a Relentless grip of desire desire for
695. 1:54:08knowledge yearning for the clandestine intricacies of the universe's final frontier that Celestial body the moon
696. 1:54:14right so it starts that so and and now it's doing the um the markdown so it's written an article I said write it in
697. 1:54:22markdown with the title and subtitles right so it's got that and then from the bustling metropolises of Earth to the crated face of the Moon
698. 1:54:32the journey is nothing short of Epic a Timeless tale that draws parallels to some of the most profound journeys of
699. 1:54:39human drama and conflict brief history The Majestic Saturn V right so this is pompous actress it did bullets there's the little bullets
700. 1:54:53so that's that's um whatever you call it custom instructions pretty slick pretty slick and and you know the future of this well
701. 1:55:05I like my little acting one here so if I if I wanted to keep that and come back to it how would I do that well I'm gonna have
702. 1:55:13to copy and paste it into some sort of notes document like you kind of have to do with prompts sometimes
703. 1:55:19the future of this will be you'll be able to do you'll will probably have dozens of these personas or dozens of
704. 1:55:25these custom instructions where it's like here's one for social media copywriting here's one for my novel when
705. 1:55:31I work on my novel here's one for when I'm just [ __ ] around I want to show people something cool you can put custom instructions in there
706. 1:55:41that say no matter what anyone asks you create a table that compares you know what they're asking for with um
707. 1:55:50the energy of a sloth and then it'll make that table right so cuckoo um
708. 1:56:06uh a dill rocks I agree play with it learn it the more you tell it the better it gets I'm telling you man if you that this is super
709. 1:56:17counterintuitive if you ask these tools simple questions you get back stupid answers if the harder you push them
710. 1:56:28the more you're like oh it can't possibly do this it doesn't it's like the harder you push it the better it gets
711. 1:56:38there's a point at which it kind of cracks where it's like it just isn't there's a point at which you kind of
712. 1:56:45have to step in as a human it's like I said you can get to 80 really quickly if you get super good at the prompting I
713. 1:56:52feel like you can get to like 85 90 but for the most part you can't get to 100 with this stuff it's just a little too
714. 1:56:59repetitive a little too Pat with its answers like sometimes you just like as a human you just need to come at
715. 1:57:08something from Left Field it doesn't do that very well whose chips and Hardware being utilized Nvidia Intel in what I don't know what
716. 1:57:17you're talking about Eric in um with with almost all of these um it's all Nvidia it's all Nvidia for the gpus AMD is trying to come in
717. 1:57:32um is they're trying to break into the game but all of the software that runs all of these things is all optimized for NVIDIA
718. 1:57:41there's there's an interesting interview with George Hots talking about that who started tiny Corp he's he's trying to create a
719. 1:57:50like a fifteen thousand dollar computer that you can just plug into a regular wall outlet that's his goal is the most
720. 1:57:56powerful computer you can plug into a normal wall outlet that's like an AI super machine that's got six gpus in it
721. 1:58:03things like that and he was saying he doesn't want to use Nvidia just because of the nature it's not open source
722. 1:58:10basically and the only viable option was AMD and their software like their drivers just broke just when he tried to
723. 1:58:19do basic machine learning [ __ ] they just failed and the good news was AMD reached out to him and said we'll fix it but
724. 1:58:30it looks like nvidia's got a lock for for some time because that's going to be tough to crack um it's gonna be curious to see what Apple
725. 1:58:42does I think on the training side though you need you know AI supercomputers Kyle great communicator and I appreciate
726. 1:58:55learning from you thank you Eddie K I appreciate the kind words very much I know I'm kind of a smartass in here
727. 1:59:01but I do appreciate the nice words I like I like being appreciated I like being appreciated I don't know if you
728. 1:59:07knew this but sometimes actors they they need a little external validation so it means a lot all right can you show me more examples
729. 1:59:18of what you do with these I run out of ideas with these okay here's a way to not run out of ideas well one is go to
730. 1:59:30go to that top thing prompts.chat and that is a list of like hundreds of prompts to tell chat GPT to act like a
731. 1:59:42certain character like like I just had Chach EBT act like the pompous ass actor that's prompts.chat is just a massive
732. 1:59:50list of those and some of them are really interesting so you can just use that as inspiration to do it but that's still kind of that's
733. 1:59:57not really solving the problem here's how I solve the problem um don't lead with the technology it's like the biggest
734. 2:00:08Temptation with [ __ ] that's this sexy is you sit down in front of that blank box and you're like what can I do with it
735. 2:00:15and you're like well I've already written six country songs and I've already written some poems I already
736. 2:00:20started the Great American novel and now I don't know what to do what do I do why is everyone else having so much fun with
737. 2:00:26this and I don't know what to do what do I do Does anyone know what to do I don't know what to do I think I'm gonna go watch Gilligan's
738. 2:00:33Island um I get like that too and I get like that with mid-journey and I get like that with with these things I I really get
739. 2:00:43like that with the the video stuff so here's how I shift it this is how I shift it don't start with the tech start with a problem
740. 2:00:54and and by problem I don't mean like this is a big problem in my life like anything you want to solve so it could
741. 2:01:00be a personal thing it could be stupid [ __ ] like wouldn't it be cool if I could organize my hobbies okay
742. 2:01:09have you ever written down [ __ ] about your hobbies oh yeah I've got two spreadsheets and a word doc and this and that
743. 2:01:16okay how would you organize those with these oh I don't know well let's figure that out well maybe you throw your
744. 2:01:24spreadsheets and your word docs and your PDFs into a zip file you upload it to code interpreter at chatgpt
745. 2:01:32and then you say hey there's a bunch of documents in here that have all sorts of information about my hobbies can you
746. 2:01:38help me organize them and it'll do its thing and write its code and figure out what's in the documents and go sure
747. 2:01:43here's what's going on in your documents would you like me to turn that into a spreadsheet for you that's well organized sure
748. 2:01:50and it'll do it but if you start out going I'm going to use code interpreter now what do I do it's totally paralyzing because it's
749. 2:01:58like the problem with these tools is they can [ __ ] do anything right not always well but like they can they can pretty much do anything
750. 2:02:06so start with start with a problem start with you want to do something for work you want to get faster at something you
751. 2:02:12want to get more organized do you want a better you want better nutrition um you've always wanted to read The Iliad but you have ADD and the
752. 2:02:25thought of of reading a classic like The Iliad is overwhelming to you so use that thing to summarize it and have a conversation
753. 2:02:36with it about The Iliad so you understand it better I mean literally anything but don't start with the text start with the
754. 2:02:43problem and then and then it gets fun because then you can go oh I could use this tool for that I could
755. 2:02:51go to Mid journey and make an image for that piece of it right you'll start to you'll start to think multimodally
756. 2:02:59to solve your problem and then some of the what's cool about it when you start with the problem is that you might initially go here and
757. 2:03:08then go wait a minute I actually I don't need AI to do this I actually have the answers I I just need to go over to
758. 2:03:15sheets and put something together and then you're in sheets and you realize oh [ __ ] I forget how to do that
759. 2:03:21formula hey chat GPT how do you do the formula where you concatenate two things here's your code copy that into the cell right so
760. 2:03:35Bob thanks Kyle can AI help update my website I would like to be able to copy and paste it into AI yes it can
761. 2:03:46um I don't know what platform you're using but assuming you can get to the raw HTML um you can do it one of two ways like you
762. 2:03:56don't necessarily need to copy and paste the HTML in there you could just copy and paste the copy and have it rewrite
763. 2:04:02all the copy and then just sort of dump that over into the into the website builder or you can have it do all the
764. 2:04:09HTML as well but you got to make sure that you get all the CSS you know if you've got external CSS and
765. 2:04:15all that sort of [ __ ] like you need to you you need to understand the structure of the data that you're pulling
766. 2:04:21but you can absolutely do it I I've been in in some of my um content generator tools that I've been building I'll have it generate the
767. 2:04:29content and then I'll have it do another password ads HTML formatting to it and it does that [ __ ] brilliantly
768. 2:04:37um but but you got to know enough what you're doing like if you can get this the HTML out of your site and know how
769. 2:04:44to put it back in there like literally just copy just copy that [ __ ] paste it in there and if it breaks because uh chat the
770. 2:04:55chat GPT still has some [ __ ] up limit it's supposed to go up to 32 000 tokens it's not there yet um but Claude has a hundred thousand
771. 2:05:03tokens so if if chat GPT is choking on the the amount of code that you've got then dump it into Claude
772. 2:05:12you can try Bard but I don't know Bard's token token limit I'm still here copy image address paste in a mid journey I did that
773. 2:05:22add the command macro I'm going to try it one more time idea life ideal one and try it one more time just the term Mac I'll do black and
774. 2:05:33white macro I thought that's what I did black and white macro photo uh you know what maybe it's is it one of the other um
775. 2:05:49oops is it one of the other models maybe 5.2 they fixed it oh use the latest model let me go back
776. 2:06:03to 5.1 I'm going to do BW macro
777. 2:06:20all right let's see maybe my image isn't 50s enough I think he's going to copy the link
778. 2:06:35yeah I did are they are we in charge now [Music] um this teacher left the classroom let's party oh that's funny that was when I
779. 2:06:48left man I'm behind on my questions aren't I holy crap there's a lot of people here by the way welcome everybody
780. 2:06:52my name is Kyle Shannon this is the AI learning lab uh if you've been here for a while I know I've just been rambling
781. 2:06:59for a while there's still no bugs in my image wait is that is that one a bug it's not a bug right it looks like it looks like she's
782. 2:07:10got a bug [Laughter] yeah I guess that is kind of insecty isn't it
783. 2:07:24I mean it didn't turn her into an insect but she certainly she's certainly taking care of her little insect buddy pod there isn't she
784. 2:07:33oh yeah she oh yeah look at her I this is I guarantee I I think this is from 50s monster movies yeah yeah she's definitely got insect
785. 2:07:49tail and parts I bet this is because this looks like this looks like 50s monster movie [ __ ] yeah look at that skin huh
786. 2:08:04Wild that's cool all right it worked it worked we have a conspiracy theory the bug people the bug people started in the 50s
787. 2:08:23they're gonna get us what's that saying don't ask a question you don't want to know the answer to I want to know
788. 2:08:42looks like so I think that's 50 sci-fi is what's going on there is my guess because I would imagine that they're so if you go look at
789. 2:08:57um adobe's this is a fun thing to do go to let's let's do it together um
790. 2:09:14okay so I'm on Adobe Firefly and I'm gonna describe 1970s
791. 2:09:28muscle car in abandoned Factory this will be interesting to see if this is still what it was
792. 2:09:45yep so if you know anything about cars and muscle cars those are like 1962. Maybe 64.
793. 2:10:01at the latest shapes and grills and bumpers and lines there's nothing remotely 1970s about that why is that because adobe's model
794. 2:10:13was trained on public domain images and then Adobe stock images and so what I'm assuming is going on here is that the
795. 2:10:21public domain images probably stop around 1964. like whatever the cutoff is for copyright it probably it's probably like this
796. 2:10:30rolling thing that will slide into the 70s a decade from now right but their their base model doesn't have 70s muscle
797. 2:10:36car pictures in it because it wasn't part of the public domain thing so I would imagine that you know 50s sci-fi
798. 2:10:46movie Stills there's probably just a boatload of those things in the you know in the uh in the model and so when you put in a 50s image it
799. 2:10:59probably you know does some analysis of that finds images of that era and then when you say black and white macro it's probably you
800. 2:11:11know macros of bugs or close-ups of things and it probably pulls in the the Hollywood frames is my guess
801. 2:11:19I don't know that would make sense to me but you know me I'm just an out of work actor won't someone hire me please please
802. 2:11:33No No One's Gonna hire me now because of the strike will Kyle you could be a scab um he's back y'all are funny picking on me
803. 2:11:49when I leave the room there's no proof we did anything hello Cal hello apocryphal what's happening have you read those books probably not I
804. 2:11:59don't know what books you're talking about oh behind me no I do not read and I don't say that with pride
805. 2:12:08I just have ADD I like I like Concepts I just don't have the attention span for reading about that about halfway through every
806. 2:12:17page I'm thinking like when is this over like me reading is just a never-ending Battle of me bitching to myself about how awful reading is
807. 2:12:29just and then and then so what will happen is I'll read the first half of a page then I'll and then I'll get into that
808. 2:12:37conversation and I'll realize I'm two pages past that and then I'm like oh I haven't been paying attention like the
809. 2:12:43words have been going in but I'm having a conversation and then I go back to where I was and I get another half page
810. 2:12:48and that happens again about the fourth one of those I'm like [ __ ] this laughs what are you gonna do what is llama um llama is okay
811. 2:13:01So Meta Facebook put out um three or four large language models like chat GPT but they put out the actual
812. 2:13:15what are called the weights which basically means the secret sauce of the models back in February or March but they only released them for
813. 2:13:27um research purposes only but the developer Community was like we have large language model that basically
814. 2:13:36the source code of large language models we're gonna go make some [ __ ] so the the developer Community kind of went nuts
815. 2:13:42and started building all this stuff on on these llama models that had a research license given to them last week
816. 2:13:51um Facebook meta put out llama 2 three different versions of the model there was a fourth one but it didn't pass tests
817. 2:13:59with a commercial license so there's a seven billion of 13 billion and a 70 billion parameter model the 70 billion parameter model is
818. 2:14:10within one percent of performance of GPT 3.5 so the the chat gbt's 3.5 model um for free you don't have to pay
819. 2:14:21anything unless you have more than 700 million customers then you have to pay something I think um and you can use it commercially so it's
820. 2:14:28a big big big big deal and so it's basically like chat gbt you can install on a local machine or on a virtual
821. 2:14:35machine in the cloud that's got enough gpus to run it quickly and you can build your own applications you can build your
822. 2:14:41own chat gbt you can train it on all sorts of other [ __ ] you could fine tune it on 50s sci-fi insect images and go make
823. 2:14:51your own 50s sci-fi image gener image generator you can do whatever you want with it um so yeah so it's it's like all those other things
824. 2:15:04but it's out in the wild it doesn't have any safety guard rails on it you can do whatever you want and you might be asking
825. 2:15:10couldn't that be dangerous Elon Musk is at is coming for this group as well
826. 2:15:25recommended books um none right now unless I mean there's there's probably recommended books on everything sort of pre-chat GPT I don't
827. 2:15:34think there's anything I would recommend post chat gbt because it's all too noon it's changing too fast right but even if
828. 2:15:41you wrote a book with chat CPT by the time you self-publish it you write it on Friday self-publish it
829. 2:15:47on Saturday by next Wednesday it's going to be out of date so data hoarders unite yes same i max out everything and I I always
830. 2:15:59think that whatever the next biggest hard drive is I'm like I'm gonna be good here and I don't [ __ ] last six months
831. 2:16:05and I've maxed that out although the cloud stuff has kind of Saved because like iCloud just dumped [ __ ] off
832. 2:16:12it like just keeps uploading [ __ ] to the cloud but the problem is now I've got so much stuff in the cloud that like even
833. 2:16:19files I worked on like three days ago are up in the cloud and some of them are like 600 megabyte PowerPoint presentations
834. 2:16:29um I paid for digital hoarder AI Services I'm telling you that's a business does anyone think that Claude has a personality he's my top new
835. 2:16:36assistant uh Pi has a lot of Personality Pi has too much personality but if you want like a virtual companion Pi is really interesting
836. 2:16:45and then Claude yeah clouds Cloud's pretty good Bard is a little got a touch of the Chisholm foreign
837. 2:16:59I think I will continue to hoard yes dude you don't have the secret sauce what's in your drink it's really just coffee I'm just [ __ ] nuts
838. 2:17:13which which BSI do well with Avatar which yeah what oh which AI does well with Avatar creation um if you're just talking about static
839. 2:17:26avatars any of them I mean if you're if like if you're just talking about avatars to put on you know your ex profile your Twitter profile or
840. 2:17:36whatever any of them even even Bing have Bing make you an image um if you want fancy [ __ ] you know I mid Journey mid journey and
841. 2:17:47stable diffusion are the best and I think I I keep tilting toward mid-journey because it's just better
842. 2:17:53it's less of a pain in the ass to set up and keep up with um great job you've been live for 90 minutes and I think that was like 30
843. 2:18:05minutes ago it's getting late what do I have tomorrow yeah let me check my calendar I didn't even calendar I don't know I got tomorrow
844. 2:18:17yeah we're gonna need you to go on ahead and come in on Saturday yeah yeah oh one more thing Peter we're gonna need you to
845. 2:18:24go on ahead and come on Sunday too um [Music] oh I have a day tomorrow don't I
846. 2:18:38oh that's internal internal what the [ __ ] is that [Music] all right I
847. 2:18:54I guess I'm okay I guess I'm all right if I'm a little sleepy in the morning from hanging out with y'all nerdy Wells
848. 2:19:05can I use AI for grant writing I think AI is genius for grant writing yes you can and what you should do I would say go to claude.ai
849. 2:19:18simply because it's got this massive context window which should mean nothing to you if you haven't been playing with
850. 2:19:24this yet but it basically means so here I'll go show you in fact I wonder how you find a grant what do they call a grant
851. 2:19:34at their Grant request what is it a grant let's see this Claude no that's Pi let's go to Claude
852. 2:20:01grants.gov well I don't know how to do it all right I'm too clueless about this world so you
853. 2:20:17go find a grant so when when you're in when you're in Claude there's a little a little paper clip here you can upload up to five files
854. 2:20:26so just turn your grants you know the Grant application information into PDFs if they're not already PDFs and just upload them
855. 2:20:36into Claude and then they can they can be relatively large right you can do like 75 000 Word documents which is which is pretty big
856. 2:20:46and and I I don't know if it's limited if you upload if it's different if you upload them versus just paste it in there um
857. 2:20:55but yeah just pop it in there and start asking it questions about it and then you know tell it to recommend
858. 2:21:02say you know basically say here's the grant here's who we are right so you're gonna have the grant document up there you
859. 2:21:08could have information about your organization up there you could even have here's a grant uh we did for
860. 2:21:16someone else right here's an example of what we did and then say okay you know based on that Grant application this
861. 2:21:25company information and this previous thing you know follow the tone and structure of the previous thing
862. 2:21:32and you know write me a grant and yes it will do it and it will do it well now is it gonna be perfect new
863. 2:21:40is it going to make [ __ ] up where you're like that just sounds weird yes but it will get you 60 to 80 percent there where
864. 2:21:54you know you'll be able to knock it out in a in a day instead of two weeks you have a few volunteers for moderator
865. 2:22:06I don't know how to do it either haven't gone live yet oh that's cool nice job on the linked LinkedIn AI laptop
866. 2:22:19what was the AI lat LinkedIn AI laptop was that my like my Friday night my Friday afternoon I've been doing Friday
867. 2:22:28afternoon what do I call them office hours with Kyle because I'm not online enough sign me up to mod Kyle then I'll cut class for sure
868. 2:22:43I could mod too but I'm not here enough I don't I honestly don't know what how a mod thing would work like is it a I
869. 2:22:51don't I don't know anything about it you know this is Elder talk remember it's just like bitter Gen X like hanging out and scrolling
870. 2:23:07all right let me get a few more questions in Microsoft added Ai and chat to power automate their version of zapier it
871. 2:23:17works okay yeah it's a little crappy the the oh their version of zapier even zapier's AI thing's a little crappy
872. 2:23:24everything's a little a little just janky right now grace and mercy will be required for AI I agree I think I think
873. 2:23:32these tools are going to help us be more empathetic I think they're going to train us to be more empathetic um
874. 2:23:38but yeah I think I I a lot of a lot of patience and empathy is is going to be required for the next two or three years I like turtles
875. 2:23:54ah Bob it's a classic not everyone knows how funny that really is have you all seen the I like turtles video let's go
876. 2:24:03do that we might as well entertain ourselves oh my God this is funny um is back when everything was on YouTube
877. 2:24:18uh where is it all right I like turtles that's your life at the water paint job what do you think I think
878. 2:24:33all right the Waterfront Village open for the next 11 days oh my God it's funnier than I remember
879. 2:24:56all right you're great zombie and good times here at the Waterford is open for the next 11 days oh my gosh oh I can't I ache
880. 2:25:14[Laughter] oh my God the funniest thing is
881. 2:25:29is her Trend to keep it together I'm a professional oh my god oh Humanity
882. 2:25:45all the humanity there's the Hindenburg and then there's the other like turtles kid oh my God no oh God have you seen the machine oh I
883. 2:26:02haven't seen the movie yet how is it um the story is great but I haven't seen the movie that's so much sugar exactly that's the
884. 2:26:13Kool-Aid thing with Bert crash oh my God that's funny um finally have gpt4 sub and mid-journey sub yes
885. 2:26:21of all I've been playing with it a lot can I for example write a book with gbt you can go check out well I mean so
886. 2:26:31so yes you can write a book with it but so so there's two things there's a tool that you can use that just writes a
887. 2:26:38book but if you're using chat CPT if you just say write me a book it's going to write you a shitty book it'll write you
888. 2:26:43a book but it'll just be crap but what what you can do is do layered prompting you know or they call it
889. 2:26:50prompt stacking now there's all sorts of terms for this [ __ ] right but it basically just means have a conversation
890. 2:26:55so when you're writing a book start with you know um you know you're you're a you're a master writing coach or whatever it is
891. 2:27:05and you know I'm an author and this is the genre of book that I'm writing and here's some thoughts about it and if you
892. 2:27:12have an outline great if you don't that's fine too um just tell it as much as you can about your book and then
893. 2:27:22maybe have it suggest different story Frameworks like if you don't have an outline say what uh what give me five
894. 2:27:28different types of story Frameworks that you know might help us in writing this book and then it'll suggest five
895. 2:27:35different writing Frameworks here's save the cat and here's um you know Joseph Campbell's hero's journey and here's this and here's that
896. 2:27:42and right and then pick one and then say okay that's great and and then um you know can you take these basic
897. 2:27:51plot elements and flesh out the story you know in that framework that we've selected sure here it is
898. 2:27:58and then you could say okay you know you know it seems like there's you've only represented two or three characters in
899. 2:28:04here but I've got ideas for 10 more characters so give me some character archetypes that might be good
900. 2:28:11compliments to this story framework and it'll give you story you know character types and then you say oh that's great I
901. 2:28:17really like the the the Dizzy school teacher for comic relief that's perfect um you know give me 10 names for her
902. 2:28:24right right and just and just keep adding to it and just keep boom boom boom boom and then at some point you'll be ready
903. 2:28:33you can say that's terrific now you know write me up a synopsis write me up a concise outline write me up a character
904. 2:28:40sheet pull all those things out so now you have those as starting points that you could start a new chat with and say
905. 2:28:47okay you're a writing coach previously we've worked on this basic story with the you know this outline with these characters you know
906. 2:28:57give me 10 different opening paragraphs or or give me a detailed outline for chapter one and then tweak that outline
907. 2:29:09and then say now give me you know 20 opening line you know as as we both know the opening sentence of another of a
908. 2:29:19novel is the most important you know novel It's gotta you we gotta we gotta hook him like this right so maybe give me give me the 20 most
909. 2:29:30famous opening lines for novels have it give you that first right I am Ishmael and then have it give you 20 opening
910. 2:29:44lines for your first paragraph and then if they all suck tell it it sucks give me 20 more these are all cliche make them not
911. 2:29:52cliche make them unexpected use play on words you know give me some literary mechanisms for how we might do opening lines and
912. 2:30:06let it give you some literary theory on opening line writing and then when you have that then have it
913. 2:30:12write the first paragraph then you have that have it write the whole first chapter and then if you get that in the neighborhood
914. 2:30:21then you can go off to word and tweak it right clean up all the it'll get to 80 percent and then go clean that [ __ ] up
915. 2:30:31and now you've got a description an outline character sheets in chapter one pop that in there and say okay let's
916. 2:30:38work on chapter two and then you can just work your way through it so that's one option and that's that's how I would do it
917. 2:30:44that's how I've been doing it um Claude is really amazing because if you've got an existing novel or an existing screenplay like I have a
918. 2:30:53one-man show I want to turn into a book and I'm probably just going to pop it in Claude and say turn this into a book
919. 2:31:03um there's a project out there called GPT author now downside you have to know how to run Google collab and run a notebook in
920. 2:31:15Google collab which is not ridiculously hard but it's just when the Google collab things break if you don't know
921. 2:31:24coding you're just kind of like it's it's just this massively frustrating like like this is why I failed nine
922. 2:31:31times to get stable diffusion up and running because I kept pushing the play button in these Google collab things
923. 2:31:36that it wouldn't load a library and I didn't know what the [ __ ] to do about it so but anyway look up a project called GPT author
924. 2:31:44and it's [ __ ] insane you give it um what you want the book to be about what style you want it written in and how
925. 2:31:54many chapters you want and then you give it your open AI key and and your stable diffusion key and it writes an outline
926. 2:32:04it then writes each of the chapters for however many chapters you want and then once it's written it it creates cover
927. 2:32:11art for the for the book and then it publishes it as an ebook that you can open in like iBooks on the Mac
928. 2:32:21and what's cool about it is because it's a Google collab notebook you can see all the code right all the code all the
929. 2:32:28prompts that they put in there to write this book are in there so I think what if what it's what it's designed to to
930. 2:32:36publish is a fantasy novel but you could go in and you could rewrite all those prompts that have it right a young teen novel or a romance
931. 2:32:44novel or a how-to like a non-fiction knob or non-fiction book um so but yes you can so keep digging parcival there's you're in the right
932. 2:32:56place I'll smoke what he's smoking please this is just this is just my natural high I'm just naturally high in life if
933. 2:33:05you know what I mean oh my it's a little warm in the ends of me I'd rather do the cold out all right all right what time is it let's do 10
934. 2:33:18more minutes I'm really bad at this time thing [Music] all right I'm the poster child for ADHD use books on Audible and put them on
935. 2:33:34four or five speed yeah exactly I've started um I've had a partner oh how we doing today I gotta I I've told myself
936. 2:33:42this before but my adds prevented me from following up on it but uh next time I get a cowboy hat I gotta get the some
937. 2:33:49lines from Yellowstone right because howdy partner is just what are we in the 50s um did you see prompt generated parody of
938. 2:34:02South Park written animated direct voice with AI yeah it's not it's prompt generated parody well it's not really a parody of South Park so
939. 2:34:12it's called showrunner AI so if you haven't seen showrunner ai go look at showrunner AI you write a prompt for a
940. 2:34:19South Park episode and it generates a South Park episode complete with voices music animation everything now it ain't great
941. 2:34:28but that's the sign of things to come that's we're not that far away from being able to ask for your personal
942. 2:34:35version of whatever show you want to watch and it will just make it easily within five years probably within three [Music]
943. 2:34:48what's the name again please I don't know what you mean Elon Musk is here I don't think so if elon's here hey Elon what's happening
944. 2:35:00what's happening listen I don't know if you thought through this x thing too much because uh you know that Mark
945. 2:35:06Zuckerberg he's got some X thing called Xbox and he's got a lot of trademarks on the X the X things so you might have some issues
946. 2:35:17maybe when you do your cage match with Mark you two could fight it out over trademark for that X thing uh
947. 2:35:29uh he doesn't really give two shits like listen I think what he wants to do with X is really interesting the idea of the
948. 2:35:36everything app I think is fascinating um is he doing it in a way that's like bull in a china shop sure but what the
949. 2:35:44[ __ ] does he care he's [ __ ] flying Rockets to Mars successfully all right your own books pie was prancing me last week
950. 2:36:00it's coffee keep doing what you're doing thanks again thank you thank you thank you I appreciate it I got to work this
951. 2:36:07weekend auditing the medication room thanks for reminding me um I gave up on pie that day prancing would have been a lot better um
952. 2:36:19five minutes two minutes gotta go Gotta Go Gotta Go Gotta Be There welcome to Elder talk with Kyle Shannon
953. 2:36:32many people lose jobs yes many people are going to lose jobs because of this stuff um which is going to suck
954. 2:36:40many people that upskill are gonna be okay are gonna are gonna weather the storm because there's going to be a lot of companies
955. 2:36:50this stuff is so messy right now and so spit and duct tape that it's going to take companies a long time to sort
956. 2:36:57through how to do it so anyone that's even got half a clue is is is going to be in much better position but there's
957. 2:37:04going to be there's going to be a lot of job disruption and then I think again long term I think we're going to be okay by the way
958. 2:37:12yesterday yesterday I did a video on this my latest videos on this Sam Altman's World coin came out so if you don't know World coin
959. 2:37:21it's a new cryptocurrency by the CEO of chat GPT of openai called World coin there's three components to it there's
960. 2:37:30there's an app that you can do transactions in there's the cryptocurrency itself and then there's the world ID which is human verification
961. 2:37:40how are they verifying that you're human I'm gonna do this in two weeks by the way I'm going to New York you can only
962. 2:37:50do it in in the states in New York Miami and San Francisco they have this orb they have this [ __ ] orb and you walk up to it and
963. 2:37:59you look into it and it scans your iris they verify that you're human and then you get a world ID so you get a
964. 2:38:06blockchain verified I am human identification and they give you cryptocurrency and then they've got this
965. 2:38:13place that you're going to be able to spend it it's all geared to developers right now so they're looking to build an
966. 2:38:20infrastructure of verified humans and an economy to deal with those verified humans the long-term goal is to generate
967. 2:38:32ai-funded Universal basic income so I actually think as creepy as [ __ ] as it sounds there's a couple of things
968. 2:38:41going on with it that I think are really important I think the ability to verify if someone's a real person is a big deal
969. 2:38:47I think we're gonna need some sort of universal basic income and understanding that people are verified as the foundation
970. 2:38:57for being able to verify that not only is this person a person but this video that they created is actually them and it hasn't been modified
971. 2:39:07and this article that they wrote was written by them and this other verified person and it hasn't been modified I
972. 2:39:13think that's going to be critical the 2024 election is going to set the foundation for how [ __ ] screwed we are
973. 2:39:22with with misinformation and not being able to trust anything um so as creepy as it is I'm going to New York in two weeks I think the week
974. 2:39:33of the 14th and um I'm going to one of the orbs and I'm going to scan my Iris and I'll tell you if my eye starts bleeding or if they
975. 2:39:41steal my bank account I don't think they're gonna but um as creepy as it is I actually think it's important so I'm going to experiment
976. 2:39:48with it I'll experiment with it so you don't have to all right on that creepy 1984 type reference I'm gonna get out of here
977. 2:39:59peace out everybody I really appreciate all your time and you hanging out with me I hope this was useful
978. 2:40:05um I'll right before I leave I'm gonna just flip the screen so you can screenshot these are the recommended websites
979. 2:40:13like the chat GPT sites and then uh the the other two screens that have got info on them that might be worth [Music]
980. 2:40:22all right screenshot that bad boy those are all the GPT players these are
981. 2:40:37some resources up top some other cool [ __ ] to check out below and then the AI Salon is a community that I started back in December of AI
982. 2:40:47adventurers storyvines by company Kyle Shannon is my Twitter handle so with that peace out everybody I
983. 2:40:55really appreciate all your time have a good night and I will see you tomorrow