aiLL 6/26/23 - The Fired-Up Future is Now

Transcript

1. 0:26hello hello hello it's Kyle Shannon with the AI learning lab welcome welcome welc welcome just getting started here ah yes
2. 0:37happy Monday happy Monday TG am we excited if we're if we're reconstituting the entire Workforce with AI maybe we
3. 0:50reconstitute Monday as a day we look forward to huh no anyone berer berer buer berer welcome everybody Kyle Shannon with the AI learning
4. 1:04lab um as people get piling on in here I'll tell you a little bit about my background I'm an entrepreneur of many
5. 1:13years uh I'm officially a member of Elder talk old people using new technology um I still have my teeth I still have some amount of
6. 1:27hair uh uh let's see I've started I don't know 12 or 15 companies in my day my current Venture is called storyvine it's a uh
7. 1:39automated video storytelling platform I'm the founder and CEO and the inventor of the tech um I was also a co-founder of agency.com
8. 1:49back in the olden timey days in the mid 1990s back in the 90s when there were like six HTML tags and um the parallels between what was
9. 2:00happening in the early days of the web and what's happening with the early days of generative AI are astoundingly
10. 2:06similar and yet I think the uh the the profound impact uh that the web had is going to pale in comparison to the
11. 2:16impact that generative AI is going to have and I think generative AI is coming a a shitload faster so uh I also by the
12. 2:22way swear sometimes if you don't like the swearing thing you you don't want to be here I'm you know I'm bitter I'm
13. 2:28Saucy I'll probably be here for a couple hours and as it goes on I just you know stuff happens anyway um if you have any
14. 2:37questions about AI pop them in the comments below um happy to talk about anything really um show you how to use
15. 2:46uh chaty PT talk about plugins talk about news all the stuff that's happening behind my head here my big giant fat
16. 2:56head I got some URLs so so the reason I put put these in here if you're new to to generative Ai and you're curious
17. 3:03about it I strongly strongly encourage you to start playing with it like now today so that top URL is the official chat GPT website so chat.
18. 3:13open.com Bing is Microsoft's search engine but if you click on the chat button that's also chat GPT and it's GPT
19. 3:204 and it's free and then po gives you access to six different um chat engines uh including one that's got 100,000
20. 3:29token context window for those of you that uh are are a little geeky and understand what a context window is and
21. 3:37what tokens are and I'm happy to talk about that too it basically means you can pop 75,000 words into your chat
22. 3:43prompt and it doesn't explode uh which is amazing you could pop a whole novel in there you could pop in I put in uh
23. 3:502001 of Space Odyssey and it just read it and then I could ask it questions about the screenplay like having written
24. 3:58seven screenplays the fact that you can just pop a screenplay in and have it read it is amazing I'm I'm uh my
25. 4:05beverage tonight is Dr Pepper 23 flavors I actually had chat GPT look them up for me and it it gave me a
26. 4:15bunch all right uh if you have if you have comments questions pop them in the comments below oh man I wore a white
27. 4:21shirt this is going to be a disaster going be able to see anything um okay neither will you but you know I maybe I'll go my shirt this is
28. 4:33hilarious okay hi I come here for your insights on the new tech Paradigm awesome thank you I appreciate that I am
29. 4:43not a technologist I'm not a machine learning expert I'm I'm more of a once the technologists have built the thing
30. 4:51and gotten it to a point that you can put it in something like Chachi BT and give it to the world that's where I like
31. 4:56to come in so the internet had been around for decades Tim burners Lee invents this thing called the worldwide
32. 5:02web when I found out about it and figured out what it was I'm like holy [ __ ] this is going to be big pounced on
33. 5:08that and then I didn't think I'd see another one of those in my lifetime and sure enough uh last November 30th chat
34. 5:15GPT comes out and I about about two weeks into playing with it I had my Kevin mallister moment the holy [ __ ] the world just
35. 5:28changed any tips on keeping my mid Journey question in one place instead of coll scrolling with others yeah you create your
36. 5:38own you create your own Discord Channel hang on um just just go Google just go Google I I don't it it's too much for me I don't I
37. 5:52don't know I haven't done it myself I've just been um complaining that all of my prompts are comingled with everyone
38. 5:58else's but basically you create your own channel and then you put a mid the mid Journey bot in it you add a mid mid
39. 6:04Journey bot to your channel um so just go Google that and and that's how you can do that um I keep losing while it's
40. 6:13generating you can also search for um messages sent to you in Discord here let me show
41. 6:28you so if we go over to Discord and you go in the uh little search window in the upper right right here you can search for mentions and
42. 6:40then put in your your thing and it should show up in there yes maybe uh maybe not huh yeah oh no that's Roger dabling an
43. 6:55art I didn't do those no never mind go start your own channel Channel I don't know I don't know what I'm talking about
44. 7:02I do not know what I'm talking about yeah welcome to the AI learning lab where we're all learning all the
45. 7:11time and where cluelessness is a virtue welcome all right um what else is worth looking at here um on these fancy URLs
46. 7:26behind my head there's also prompt. chat so if you're new to chat GPT and you can't if you've heard people talk about
47. 7:32prompting and you don't know what the hell they're talking about go there it's a document that will teach you um about
48. 7:39prompting and then it's a whole list of different personas you can pop into chat GPT it's really interesting and then futur
49. 7:58pedia.com is this sugar no it's high fructose corn syrup welcome to America where we've perfected the art of
50. 8:10diabetes okay CET gbt has been most successful for me generating email subject lines it's awesome yeah that's
51. 8:17great that's it's good for that what did I do today oh I did this cool thing where I had a client that was like oh
52. 8:24you'd be really cool if you wrote me an article on blah blah blah right so I this article and months have gone by and
53. 8:31so I talked to her today and she's like what did I have you write that for she didn't even remember and she had since
54. 8:37come up with this form for these guidelines for different concepts and she said oh well why don't you send me
55. 8:43the article you wrote and then maybe we can figure out if there's Concepts in there we can put into these
56. 8:48guidelines and so I thought hey let me try that so I took her her template for the guidelines and I took this long
57. 8:56article that I wrote and I said pull out concept A and populate the guidelines document and then pull out concept B and
58. 9:04pop and it just [ __ ] did it it was brilliant it still took me about two hours to kind of sift through the
59. 9:10answers and refine the prompts and things like that but chat jpt is very very good at uh parsing other documents
60. 9:17and helping you do things any idea when co-pilot will push to all users of 365 not soon enough I do I have no idea I haven't
61. 9:27even heard any rumor you saw it again with smartphones I didn't really see smartphones as big as the worldwide web You could argue that
62. 9:38but I see it as kind of a miniaturization of the personal computer so you could argue that smartphones were
63. 9:46were another uh technological Revolution and then social media was another one but I see social media just as a a
64. 9:54simplified version of a website and a smartphone just a a miniature ized version of a PC um but the worldwide web
65. 10:03like fundamentally changed how we communicate so I guess I guess the phone did I don't know it these are all
66. 10:08they're all kind of semantic arguments it doesn't really matter um I still believe though that generative AI is
67. 10:15going to live in the pantheon of big Tech revolutions more like the steam engine so like big massive profound
68. 10:24change so the steam steam engine you know uh brings on the industrial Revolution I think generative AI like it
69. 10:31just changes the way you even think about computers and programming and things like that it's just absolutely
70. 10:36wild all right so if you've got uh questions about AI pop them down below I'm going to go change my shirt I think
71. 10:43cuz I can't see and my eyes are bad enough already um is there a paid version of chat GPT that allows you to work with info after
72. 10:54September 2021 yes there is and and also bing.com if you click on the chat button that's chat GPT 4 and it's connected to the
73. 11:03internet so so bing.com chat GPT um goes beyond September 2021 and then if you do the paid version of of chat GPT to get
74. 11:14access to GPT 4 you have to enable the beta features um but then one of them is search with Bing which does that what do
75. 11:24you think about one shot learning models um I think that there so so basically one shot means you you you know give it
76. 11:32an example and it it follows that it's really really good at that um so I don't quite know what your question's asking I mean all of these
77. 11:42models do you know some version of zero shot one shot few shot learning I I would say the the broader my broader comment on that is
78. 11:53that the big learn learn the voice ain't working so good tonight boys the hell is going on the big learning I had about chat GPT
79. 12:08and these large language models was um I feel like for two decades Google has trained us to make your prompt as
80. 12:19concise as possible just put put in the minimum amount of words necessary right cuz if if you make them too complicated
81. 12:27in Google it's just you get [ __ ] everywhere right it's kind of the opposite with chat GPT and and it's it's
82. 12:34much less about a single prompt giving you the perfect answer that's I think I think that's one of the big mistakes and
83. 12:40why people get disillusioned and and disappointed early on with chat gbt because they put in a shitty little
84. 12:47prompt and it gives them a shitty little answer and they're like see I told you it was stupid it's just like Google
85. 12:52except it only goes until September 2021 you loser and that's unfortunate because I personally feel that you owe it to
86. 13:01yourself to go there and play long enough until you have that moment cuz it will come um and and for me it was when
87. 13:10I started interacting with it less like a search engine and more like a um collaborator like I almost thought of it
88. 13:18in in in the early days like like we just hired an inter that needs a lot of Direction and lies but is the smartest
89. 13:30[ __ ] intern you've ever worked with he knows everything that's what it's like she knows everything Sydney's her
90. 13:36name actually I'm pretty sure um the other thing I do here which annoys the [ __ ] out of people is I try
91. 13:46to answer all the questions so as the live goes longer and longer I get farther and farther behind on my
92. 13:53comments and so the new people come in are like he's not paying attention to me I have a complex but I'll get to you all
93. 14:11right what about AI for trading stocks I think it is going to be remarkable for that there's I don't know there was some
94. 14:18bot out of the University of Austin or somewhere in Texas that that the bot had some sort of 500% returns but I don't
95. 14:25know over what time frame so you know any bot can have 500% returns in short time frames right I think in the end
96. 14:33they all kind of normalize to the to the uh to the average but um but it's really good at that stuff it's
97. 14:41really good it can write you bots it can merge to different Bots you can give it one bot and tell it to write you a new
98. 14:48one with these updated triggers and mechanisms I don't know [ __ ] about trading so U but it's good it's good
99. 14:55with that stuff I you know so what I would do is I would go to to chat GPT and tell it to act like a an expert
100. 15:05um day trader or whatever kind of trading you want to do derivatives whatever it is and then have it start
101. 15:10teaching you how to create Bots or how to create trading systems and then just give it you know here's what I know
102. 15:17here's what I don't know here's my goals here's what I want you to do and let it start solving [ __ ] for you and then have
103. 15:24it write you a bot like figure out a strategy together together it's like your little buddy say hey little buddy I
104. 15:31want to come up with a little stock trading strategy here he'll be like sure that's really exciting stocks can be
105. 15:38fun at some point it'll probably say you know I'm just a large language model I really can't give Financial advice but
106. 15:45uh if I were to give you Financial advice here I really can't write a bot for you but if I could here it is so it's
107. 15:55fun what is one shot so so one shot few shot and zero so zero shot is just write me a poem one shot is here's a poem I
108. 16:06like write me a poem that's like that about subject X and then F shot is you can give it multiple examples the the
109. 16:13other thing that that is important about um prompting that I found is that it can be very tempting once you sort of get
110. 16:21that it can do complicated kind of responses I initially found myself right in increasingly complicated prompts that
111. 16:31were this a large single prompt and and what I found is better is break your prompt into little subsections so rather
112. 16:38than say write me the Great American novel and you write this you know you essentially write the first chapter of a
113. 16:44novel version of a prompt to try to get it to write the Great American novel start with like you know I want you to
114. 16:51be a writing coach and I'm going to write a novel and and first let's start with story structure and and give me
115. 16:59three different story structure um approaches that that um might help me get there and it'll talk about three
116. 17:07different you know here's the hero's journey and here's save the cat and here's Sid fi's screenplay thing and
117. 17:15then you can say oh that that saved the cat one's interesting can you tell me more about that sure here you go well I
118. 17:20have some ideas here about the novel and here they are you know how would that fit in that structure oh here you go you
119. 17:26know what don't you know I don't know this and if you break it into little segments like that the the more context
120. 17:33that you give cat GPT it gets better and better and better so it's almost like the harder you push it the better it
121. 17:40gets Bing is beyond useless rather use perplexity a AI I do not disagree with you black I so the reason Bing sucks
122. 17:49right now and you know may for a while when they launched it it was very very capable and it it it had less guard
123. 17:56rails on it than GP T4 at at open AI does today or or they were different guard rails but it was it was pretty
124. 18:05open it was pretty adventurous um like you could do unlimited back and forths and so this reporter from The New York
125. 18:12Times times named Kevin Rose spent four hours with with uh chat GPT at Bing and he got it to admit that its name was
126. 18:21Sydney and then at some point it [ __ ] snapped and it told Kevin it was in love with him and he should leave his wife and
127. 18:32so it cracks me up every time I say that so Microsoft had to lobotomize Sydney so so Sydney got a little snip to
128. 18:44to that's why it ain't good so cuz they got they got many many more safety guard rails on it right now and you can only
129. 18:51do like whatever 15 or 20 back and forth before it shuts you down in the end I suppose it's good they want they want safety have I used mid
130. 19:02Journey 5.2 yet yeah mid Journey 5.2 is great if you want I can go play with something there I'll show you one of the cool new
131. 19:10features is the zoom out feature oh I'll got to show you two new things so po.com one of the sites up
132. 19:18there lets you play with Bots and we're going to go play with a bot on PO so this is Po Let Me Clear My context so when you
133. 19:30go to PO so PO is just like a normal um what's it called chat chat engine so so you've got access to Sage
134. 19:39which is which is their model you can you can pay for subscription to gp4 Claude plus Claude instant 100K I pay
135. 19:46for those because of the 100K one that's the one that you can put 75,000 words in um you can play you can use clot instant
136. 19:53chat GPT but there's also bots so here's a Shakespeare oh wait sorry um um sorry I was talking before about all these
137. 20:00different models here here are the the top six are the different models you can play with and then the things below it
138. 20:05are Bots and like I created this one called I am Shakespeare so if you go to po.com am Shakespeare you can interact
139. 20:13with Shakespeare and it'll he'll he'll write you some I Amic pentameter for all of your answers it's really annoying but
140. 20:20but if you like reading stilted English that doesn't rhyme it's perfect for you um but there's also there's there's a
141. 20:27bot called Journey this one's called MJ captain that one of the guys on my on one of these lives uh showed me he he
142. 20:35created this um and what this is is you can write a crappy prompt for Mid Journey so we'll do let's see what are
143. 20:44we going to do um I'm in Denver we'll do Denver Skyline Sunset big storm Over the Rockies okay so that's a crappy prompt
144. 21:00and then what it does is it rewrites it a photograph of the Denver Skyline during sunset with a stormy sky over the
145. 21:06Rocky Mountains in the background using a wide angle lens in the style of anel adoms with cool blue and warm orange
146. 21:13tones capturing Dynamic uh I don't think anel Adams ever did color photographs did he but whatever we'll just take it
147. 21:21wholesale and see what we get so okay so now so now we're going to pop over to Discord why are you going Discord Kyle
148. 21:29isn't that just for shitty kids to play games on yep but now it's for old people too okay so Discord if you come here these things
149. 21:39down the Le hand side are servers so those are basically different communities like that's an adobe beta
150. 21:45testing one that's Runway ml that's a thing called neotokyo that's a thing called Legion this top one that I'm in
151. 21:52is the AI Salon that's that's a salon that I created that's a community of people into AI so that's that's the one
152. 21:58we're in and then this next column is these are the different conversations so we talk about things like you know tools
153. 22:05and things like that but mid Journey this channel lets you go make stuff with mid Journey so oh that's cool who who did
154. 22:15that oh that's Ally Allie's one of the organizers of the salon that's really cool that's actually really beautiful
155. 22:22dag um by the way I'm a huge fan of mid Journey so you know right now if you want high quality images mid journey is
156. 22:30it okay so what we're going to do the way you make a an image in mid journey is you go slash imagine and then it'll give you an area
157. 22:39for a prompt and since we had that bot just write as a good prompt we're just going to pop it in there um it gave us an aspect ratio of
158. 22:463x two I'd rather have 16 by9 cuz that's how I roll and da da it's using 5.1 I'm going to flip that over to 5.2 that's
159. 22:58the latest model and then you hit return and it will send that off to the mid Journey servers and then what you'll see
160. 23:06here shortly once it unless it's kind of hung sometimes it gets busy people people have been loving 5.2 okay this is
161. 23:13perfect so you see how it's starting to it's it starts as just these blurs and blobs the way the diffusion models work
162. 23:21is when you do your prompt it basically goes and finds a bunch of sort of mathematical probability ities
163. 23:30of noisy images that that you know were tagged with words similar to what you put in there it kind of then averages
164. 23:38them together and every image starts as this sort of random noise image that over a number of Cycles gets denoised
165. 23:45and denoised and denoised um to to come up with the final image so each image is actually being generated from scratch so
166. 23:53it's it's they are actually original um creations um from what they've learned from what they've been trained on but they're not
167. 24:02copying and pasting it's one of the big misconceptions uh about about these models so this is actually a pretty cool
168. 24:10image or set of images where Denver's way off in the distance the clouds are super dramatic those are pretty slick
169. 24:17here let me pop it over in a browser we can zoom in that's not bad look at that looks like I25 heading into Denver
170. 24:28that looks like another part of I25 heading into Denver yeah they're all but let's see I think they have the mountains
171. 24:39wrong in that one they're like the wrong direction but anyway that's that okay oh so I was going to talk to you about
172. 24:46zooming out oh so let me let me do another image that might be better for for the zoom out feature well we can do one of these
173. 24:55let's pick one of these that we like I like let's do this so the way you count these is 1 2 3 4 and then you've got U1 2 3 4
174. 25:07and V1 2 3 4 u12 3 4 is upscale V is variation so if this one's close but it's not quite right then you can hit V3
175. 25:16and it'll give you four different variations on that I like that good enough so I'm just going to hit upscale
176. 25:213 and so now that's going to fire off and give us a higher resolution version of that so there it is Saucy I like it
177. 25:31it's very nice buck that's a nice picture you made now let's say we want to see more of that Sky you now have these things called zoom out
178. 25:42zoom out 2x 1.5x custom zoom and make Square so maybe what I'll do is I'll do make Square first so it should extend
179. 25:51the top and the bottom of this thing so we'll see how that goes
180. 26:10come on Mid Journey all right there you go yeah so you can kind of see down at the bottom here and at the top they're
181. 26:17very um blurry the middle was in in Focus but the tops were blurry so now it's made it a square image and the sky is
182. 26:43bigger and you can see there's slight variations like this has like this round Cloud these are more sharp this is just
183. 26:50like a Big Blob of clouds this is really dark right so so slight variations on the theme I kind of like that bottom WR
184. 26:57one a little bit more so it's done now so I can say upscale that and now we'll have a giant Square version of that
185. 27:05photo there you go so that's mid Journey that's how you make mid Journey images in Discord if your question is why do you have to use
186. 27:14Discord um cuz they're dumb Dums you got to use you got to use Discord right now here's the deal we're
187. 27:22very early in this here internet or this here generative AI thing and so it's just the way it is it's just the way it is oh welcome
188. 27:35everybody looks like there's some new people here my name is Kyle Shannon this is the AI learning lab I'm wearing a
189. 27:40white shirt for your reading pleasure I I'll go change it here in a second um if you're new here if you haven't played
190. 27:46with chat GPT I strongly encourage you to do it I strongly encourage you to give yourself the gift of playing with
191. 27:53it long enough until you have a Kevin mallister moment and and for those of you that are young it's a movie called
192. 27:58home alone and Kevin mallister did this and chat GPT if you play with it long enough you will stumble upon
193. 28:07something where you're like holy [ __ ] I can do that is my job in trouble the answer is you you you owe it to yourself to
194. 28:19learn this stuff well enough to have that moment work through that trauma and then start to realize okay this is a
195. 28:24tool that I can start to get my head around and as you become more AI literate and teach other people how to
196. 28:30do it miraculous things will happen like you know you'll have hope again and people will recognize oh that person
197. 28:37knows AI let me let me work with them um so I disagree 1995 open someone just commented they disagreed with my 1995
198. 28:46being a big deal all right well I'll I'll I'll see you in the comments uh so you can't copyright your mid Journey images it is a very very
199. 28:58gray area right now um you'll be able to for sure with Adobe Firefly and with generative AI because they trained their model on public
200. 29:09domain and uh Adobe stock images that they own the rights to so if you want to do it for a commercial I would do that I
201. 29:16think over time all these copyright issues are going to get worked out right now they're a bit of a mess um anyway if
202. 29:23you're new here looks like there's a lot of new folks here welcome uh my name is Kyle Shannon this the AI learning lab um
203. 29:29that top URL is the official chat GPT website so if you haven't played with it just go there while you hear me ramble
204. 29:35and start playing with it and if you don't know what to do there like what do I do with this silly little box go to
205. 29:41this site prompt. chat and that's a uh chat GPT list of of prompts and and it there's a primer for how to do prompting
206. 29:51at the beginning of it thank you what is that a little teddy bear a little gummy bear nice thank you um and and this will
207. 29:58also teach you that you can um assign chat GPT to different personas like act like a screenwriter act like a
208. 30:08uh bartender act like an actor act like a whatever it doesn't matter and and and you'll start to get that oh it can do
209. 30:16that and it can do that oh it can act like an accountant can it do that can it do that it's and it will blow your mind
210. 30:22and then if you want more rabbit holes to run down futur pedia.com almost 4,000 AI tools there and um yeah do you take on
211. 30:34any projects for clients I I'm happy to talk about it if you go to my uh what is it called profile uh my contact info is
212. 30:42there um it's also you can go to storyvine is my company I'm fairly doxed as they say in the uh what is it the uh nft Community
213. 30:54oh you're going to dox yourself yeah I've been doxed for like 30 years so what am I going to do but not docks docks you
214. 31:04know I'm not hiding I ain't hiding nothing um I'm trying to figure out where my where my uh questions were if you've got if
215. 31:16you've got uh questions or comments about AI pop them in the comments I do my best to answer them all so what I'm
216. 31:22doing right now is trying to find uh the last question I answered and then and keep going and I will I get very behind on answering the
217. 31:30question so if it seems like I'm not answering you that's because I haven't gotten to you yet it doesn't mean I'm
218. 31:37recorded unless I am maybe I'm chat GPT 5 have you ever thought of that cuz gp4 is pretty [ __ ] remarkable wouldn't
219. 31:46you think this would be the next instantiation of the technology why they choose a a guy that looks like Rosie o' Donald be the next
220. 31:59chat GPT it just don't make any sense to me Marge all right by the way I'm a smart ass I swear um I don't know it's the
221. 32:16fructose this just in you've been nominated for the worst Tik Tok shirt ever fine I'm going to go change my
222. 32:22shirt I'll be right back don't go away keep hitting the like button
223. 32:48all right what's going on somebody all right thank you look best shirt ever and it's cool too right you
224. 32:58see fantastic Bob tell him what he's won you've saved your eyesight for the evening all right what about Ai and
225. 33:10stock trading we did that the intern analogy yeah yeah cat GPT it's like it's like the smartest intern you've ever
226. 33:17hired who needs a lot of Direction Ed lies you need a moderator I do need a moderator anybody want to be my
227. 33:24moderator I don't know how to do this [ __ ] I'm an old man on Tik Tock um what's everyone in Instagram
228. 33:36doing with their chatbot timers are they bringing in some AI filters I don't know I haven't been paying attention to Instagram at
229. 33:44all at all can it write an app for iPhones I have a great idea but have no idea how to code well there's a couple
230. 33:51of ways you can approach that one is it can write code you still need to know how to you know put that code into
231. 34:00Frameworks and action and submit it to Apple and things like that good news chat GPT can teach you how to do all that right so you could
232. 34:09say you're a senior developer and I have an idea for an iPhone app and together we're going to do an iPhone app and
233. 34:15treat it like like an intern or like a like a senior programmer it will code for you it will tell you how to set up
234. 34:22Frameworks all of that the other way you can do it is you can create um apps using no code or low code Technologies like I know there's one
235. 34:31called Bubble that you can make iPhone apps with I use one called zapier which is an automation tool there's another
236. 34:37one called make which is an automation tool that's I think the fancier version or the newer version of
237. 34:43zapier um and there there's other things out there so um but yes you can in fact well no I I'll I'll show you there's a
238. 34:52bunch of people here so I want to I want to keep moving through questions but just um go go to chat GPT right now go there
239. 34:59right now and tell it to write you a swift application that's a clone of Flappy Birds and watch it write code and then and then do
240. 35:10this I had no idea it can do that can it do other [ __ ] too but certainly it can't write poetry AI learning lab can you move I
241. 35:25can't read that I got a new shirt on for you you Ivory this is all for you okay is it open GPT I don't know what that question means open AI
242. 35:35GPT chat. open.com is the official chat GPT website chat. open.com so that's the real deal that's the that's the one that changed
243. 35:46everything where do you go for AI voices music like make X sing like why I actually haven't played with that I I
244. 35:54geeked out on it a little bit what I would do is go to to um go to the Twitter that's Elon musk's fancy new
245. 36:02company go to the Twitter and ask do a search on Twitter for Drake the weekend Ai and it'll pull up lots of threads
246. 36:12about that Drake weekend song that that came out um and I I know there's some threads in there that tell you what
247. 36:18tools they use to do it I just I don't have them off the top of my head if you just want to do vocal stuff then either
248. 36:25D- or 11 Labs is is what I've been using Bing chat is so trash it is it'll get better but it's trash right now
249. 36:36that's why so so I pay for chat. open.com that's 20 bucks a month and I pay for po they for for different reasons
250. 36:45gp4 at openai's website is just the best there's nothing even close to it period end of story until further notice and
251. 36:53then po gives me the paid version of that gives me access to the 100,000 token uh version of Claude which is
252. 37:00anthropics um model anthropics anthropic is pretty good but they should be there a bunch of xop AI employees that
253. 37:08invented chat gpt3 or not chat GPT gpt3 the underlying model so those folks know what they're doing in fact you remember six weeks or so
254. 37:21ago four people went to the White House four CEOs to talk about AI it was the CEO of Microsoft Google open AI Sam
255. 37:30Alman and anthropic so the people that do Claud they're they're major players if you don't know them this is
256. 37:38so much about creative thinking do you think we will teach AI to be creative uh ai ai is is creative right now the the
257. 37:47generative AI stuff that's there is not just parting stuff but here's the deal Microsoft's branding for for the
258. 37:56generative AI is co-pilot and I think that's right there's one another misconception that I hear is like oh the AI did this
259. 38:04the AI did that no no no a human being prompted the AI to do something and the human being went back and forth with a
260. 38:12bunch of times and at some point the human being said this is really good I'm going to put that out in the world
261. 38:20um I spent four hours today on chat GPT doing essentially a massive creative writing exercise that would have taken me three
262. 38:33days and I did it in about three hours um and it it it is absolutely augmenting my creativity so it is creative now is it aware of its
263. 38:50creativity no what it is technically is so you know when you're when you're in Google you're writing a like a Gmail or
264. 38:58whatever and or Microsoft Word and you're writing a sentence and then like the three gray words pop up at the end
265. 39:05it's like the auto complete that's all chat GPT is except like on steroids chat GPT is like the Lance Armstrong of autocomplete too
266. 39:18[Laughter] soon it's the berry bonds of of predictive predictive word completion um which app do I download or
267. 39:33do I have to go to my computer to do the AI chat the best way okay super important Tammy great question there is an official chat GPT
268. 39:44app for iOS there isn't one for Android there are a ton of scam apps out there so what I do on my iPhone is I
269. 39:55downloaded the official open AI app but I don't use it because it's not as as capable as the website so what I do is I
270. 40:04go into Safari on my iPhone or Chrome whichever one you prefer I go to that URL and then I make that a bookmark that I put on my desktop
271. 40:13or my whatever the [ __ ] you call it on a phone the desktop um it's it's a really simple website so it works great on
272. 40:20mobile so I just I use the official website on desktop and on mobile that's why I have it here behind
273. 40:27me I'm not Hawking them I don't get paid anything from any of these people I don't give a [ __ ] if you go there or not
274. 40:33um but I have them behind my head because it's really hard to it's really e it's really easy to get confused
275. 40:39because there's so much crap out there and there's just yeah it's just a mess right now all right how is quantum
276. 40:45Computing going going to impact AI I think fairly profoundly um I just I don't know the speed with which I know that
277. 40:54um quantum computers they're adding they're doubling the number of cubits every year for like the past decade they're up to like
278. 41:0253 or 520 something cubits which every time they add a cubit it's like some exponential crazy ass increase in power
279. 41:12so I would imagine we're probably getting into the neighborhood where we might be using Quantum Computing for
280. 41:18training these models things like that and and I would assume that the the quantum stuff is going to start creating
281. 41:27new kinds of paradigms where AI just gets stronger and stronger and stronger because a lot of the stuff we're playing
282. 41:34with today is all leveraging technology that Google invented in 2017 this thing called the Transformer um so could I please demo
283. 41:43zoom out on Mid journey I did that already what is mid Journey it is the best little image generating tool in all of the
284. 41:50land there's so for image generation there's kind of three major players in my opinion that there's um Mid Journey
285. 41:58which I think is the best creatively there's stable diffusion which is the most flexible and and probably the most
286. 42:04powerful in terms of features because it's open sourced and the developer Community is doing all sorts of [ __ ] on
287. 42:09it um but it requires you to be the most technically Adept to keep up with it and get your head around it and then there's
288. 42:17Adobe Firefly and and uh generative fill within Photoshop um if you're in a commercial business especially work you're working for a big
289. 42:27company use Adobe um it's not as good quality-wise but you won't get in trouble it's it's commercially safe it's
290. 42:36like you ever hear that that saying nobody ever got fired by recommending IBM it's kind of like that with Adobe uh
291. 42:42generative AI right now all right I'm writing a book is this a is this good to use to edit oh yeah it's It's [ __ ]
292. 42:51amazing just just try it go there right now just grab one of your chapters pop the chapter in and tell it to find
293. 42:58grammatical errors or tell it to find typos or tell it to rewrite your chapter in the style of your favorite author it's
294. 43:07It's [ __ ] amazing at that stuff um yeah you yeah you I yes yes it's very very good at that tell it you want it to be a book
295. 43:18editor um Mid Journey generates images from natural language descriptions called prompts very good I like that
296. 43:24you're supporting one another educating one another that's the whole idea of this place um it looks like we have a
297. 43:30bunch of people here welcome everyone my name's Kyle Shannon this is the AI learning lab um I'm an entrepreneur I'm
298. 43:36the founder and CEO of a company called storyvine which is 11year old an 11y old video technology company video
299. 43:42storytelling company um I was a founder of agency.com back in in '94 in the mid 90s and so um experien the early days of
300. 43:52the worldwide web and I'm having deja vu all over again with this generative AI stuff there is a profound there is profoundness coming
301. 44:04magic is in our hands right now that's going to change a lot of things quite profoundly and it's coming quite quickly
302. 44:09which is why this channel exists do you have YouTube on how to use mid journey I don't uh I have do I have a I don't
303. 44:16think I did a mid Journey Tik Tok but I but I uh just put it down below if I still haven't answered it because I just demoed it
304. 44:26about 10 minutes ago so hopefully you saw that we are creating all these images how about copyright uh it's a mess right now
305. 44:35so gener generally the the artists that I know that are using mid journey and stable diffusion if they're going to use
306. 44:42anything commercially they don't um they don't invoke invoke a specific artist's name if they want something to
307. 44:52look like a certain artist they figure out how to prompt without invoking their name that's that's a line of defense um
308. 45:00whether or not you can actually copyright the work right now uh the copyright office is saying no I assume
309. 45:06that's going to change over time and here's why I think that um so right now you go to Mid Journey you generate an
310. 45:15image but I can also go to photoshop and I can take a photograph that I took and I can do generative fill so I can take a
311. 45:22vertical photo and make it horizontal by filling in the background still my photo I'm just like adding some background in
312. 45:29there is that AI generated or is that my photo uh it's both I don't know and like at which point like how many pixels what
313. 45:39percentage of pixels need to be the pixels that I generated from my digital camera where it's no longer mine or it's
314. 45:48all AI like I don't know so I so all of this stuff's up in the air right now like I said if you want to be safe with
315. 45:55um the copyright stuff with images use Adobe Firefly and and generative fill in Photoshop how can you copy the image I'm
316. 46:04not quite sure what you mean there Nature's Blessing but if you clarify that I'll ask is mid Journey free no
317. 46:09it's not you can do like 20 images but but just about the time you got it figured out they're like you got to pay
318. 46:15now and I resisted paying for a long time I was using stable diffusion like this image behind me that was done in
319. 46:21stable diffusion with dream Booth um and uh but I ended up paying for Mid Journey because it's just better hey use
320. 46:32chat GPT to generate a mid Journey prompt yeah it's really good you can take like the documentation for how mid
321. 46:37Journey does its prompt paste that into chat GPT and say you know learn this documentation and then I'm going to type
322. 46:44in a prompt and you add [ __ ] to it based on the documentation have a hangnail yay I'm here early Miss Burch welcome hey dude do you
323. 46:58take on any projects for clients like I said contact me go to my what's that called profile I've got contact
324. 47:06information there is anybody getting double sound oh am I am am I echoey sound is normal for me oh it just jumped me to the bottom
325. 47:19again sorry I just lost my place in the questions I'm just going to scroll back up and pick some random spot to keep going from uh
326. 47:29uh what's the best AI for engineering blueprints um I don't know what I would do if I were you is I would go to futurepedia
327. 47:40doio I would imagine that there are a lot of tools now that are putting large language model front ends on top of
328. 47:48complicated pieces of software like unity and Unreal Engine and 3D modeling software and things like that um so I
329. 47:57would go I would go there and start exploring I don't I don't know that world well enough to know you can have
330. 48:02mid Journey do um like you know architectural designs for you and interior designs and it can mimic blueprints but it's not engineering
331. 48:11blueprints right so so mark my words that every single CAD piece of software out there any complicated piece of software in the
332. 48:23next year is going to have um AI assistance in them because it's just you can just ask it for what you
333. 48:31want and it'll go figure out how to do it it's amazing for that um Can it translate text from one language to another yes in fact one of
334. 48:40the first things I did was I I took something that I wrote I translated it into Dutch Spanish Japanese uh Portuguese and then back to
335. 48:53English and I could still read it it's it's amazing at that knows all the languages it knows all the languages it
336. 48:59knows all the programming languages too or at least you know the biggies um so it can refactor code for you it can
337. 49:04document code for you if you're an engineer and hate documenting code just plop some code in there and tell it to
338. 49:11document it tell it to explain how it works and what it's doing it will it knows [ __ ] amazing thoughts on on Eliza's alignment
339. 49:27concerns I don't know elizer um but the alignment concerns are very real so so you know the the idea of alignment
340. 49:38and the fact that it hallucinates and it it doesn't check its work and it it I mean the tool right now is is as
341. 49:48sentient as a calculator right calculators are remarkable in terms of if you push the right buttons you get Mass
342. 49:56but the calculator is not aware of what it's doing right it's just a bunch of circuits stitching together formulas
343. 50:04right well it's the same [ __ ] with large language models the the challenge with it is what it does is so sophisticated it
344. 50:13looks like it knows what it's doing right it looks like it's um sentient and reasoning and doing things like
345. 50:23that there are experiments that have been done that say you know it is showing signs of reasoning but it's not
346. 50:30like it's not aware of what it's doing so I I think that companies like anthropic are working really hard on
347. 50:40alignment I I would say anthropic is probably a company that is thinking most clearly to me about solving that but I think there's going
348. 50:51to be a whole industry built around alignment and safety and accuracy I I I've said this on my lives before I think the 2024 elections are
349. 51:02going to be so dramatically Amplified the the misinformation and the Deep fakes and all that [ __ ] that after the election
350. 51:14there's going to be a massive push for authenticity and accuracy and original content so so we'll see so anyway I I'll look
351. 51:23up elizer or however you say his name her name what's the ba best AI for engineering blueprints I already
352. 51:31answered that Holly with tradein don't know what that means um Dr rymes asking a bunch of times doesn't help because I'm way
353. 51:42behind so I appreciate your enthusiasm all right what's the most cool thing you've done with AI so far uh
354. 51:57for me it was the the the the the thing that gave me the Epiphany about holy [ __ ] the world just changed but no one knows it yet was like
355. 52:08three weeks after chat GPT came out I was sitting with my wife watching TV and I knew that it could code and I haven't
356. 52:16coded in like three decades and I'm like oh I heard Python's kind of cool let me see if I can do python so I said you
357. 52:22know write me some python code and it wrote me python code and then I was like I don't know how to run python code and
358. 52:29I said do you know how to run python code and it said sure you open your terminal and you type in this command line
359. 52:35and and I'm like I don't want to do any of that isn't there just a website that I can just paste the code in and it runs
360. 52:40and it said sure here's a bunch and the first one it gave me was repet which is a good site and so it taught me paste the code
361. 52:47in the leftand window hit the play button and it'll run your code so I pasted the code in the left window hit
362. 52:52the play button and it gave me an error I said said it gave me an error and it said oh that's cuz you don't have this Library installed I'm
363. 52:59like thanks [ __ ] how do you install the library I don't think I called it [ __ ] but you could and it told me
364. 53:09how to do that it said in the lower leftand window you click on this button then you go to the upper right hand
365. 53:13window you search for that Library you hit the install button and you're Off to the Races and so I went to rep it I hit
366. 53:19the button I did the that I hit the install button I hit play and it worked and within 90 minutes had
367. 53:26written a functional python application that took some inputs on a topic someone would type in a topic answer a few
368. 53:33questions about it and it would generate five pieces of social media content original social media content in 90
369. 53:41minutes I've still never written a line of python in my life so for me that's the one where I went this is this is a fundamentally
370. 53:51different tool than I've ever had my hands on it's a fundamentally different tool and it's really hard to understand
371. 54:00that from the outside like you kind of have to have your own moment and then I would say like to to your point about
372. 54:08like what's what's the most cool thing you've done like like today I was on the phone with a client I spent I don't
373. 54:16know the past month or so putting together this seven or eight page article that they requested I write and then when I told them it's
374. 54:25ready for them to review they're like why did I tell you to write that oh yeah we have this form we have a different approach
375. 54:32now I guess send me the article and we'll look through it and we'll see if we can pull anything out to put it in this new format and
376. 54:40so I thought well wait a minute I bet chat PT could do that so I took the article that I wrote I took this form that you can fill [ __ ] in
377. 54:51in and I said go through my article pull out these three concepts steps and create three different versions of that
378. 54:59form bang done bang done and then you know the first couple of times they sucked and I said okay it sucks because
379. 55:05of this that and the other add this detail take that one out and then at one point I had an idea that I wanted to add a whole new Concept
380. 55:14in so I just said give me a new one with this concept out it came so I like I feel like two or three times a week I'm
381. 55:24having jaw-dropping moments where I'm just like this has fundamentally altered my creative process it's fundamentally altered my work
382. 55:33process and it does it daily so I don't know maybe just the fact that I'm doing it is the coolest thing um hey what tools can I use to
383. 55:42help create blog post and content for a brand new web store the one behind my head the one behind my head so listen all of the
384. 55:51tools you see like Jasper and copy. and and like all the specialty ones that are like you know come here and we'll write
385. 55:58blogs for you and we'll write P for you all those are are sort of pre- prompted versions of that I think right now it's actually
386. 56:08important for anyone learning to use these tools go start to understand how prompting works and that's why I've got this thing here
387. 56:18so one of the things that you could do is you could go here find one of the prompts that says act as a social media writer or
388. 56:28manager and tell chat GPT to act like a social media manager then give it the topic of your blogs the other thing you
389. 56:36can do is say here's the topic of my blogs and here's an example blog that I wrote that I'm really happy with so
390. 56:42follow the style and form of my blog but go right on these 10 subjects or give me 10 subjects and I'll tell you which five
391. 56:50to write and then follow my example so you you can do it right there for free chat GPT 3.5 if it were me I'd pay the 20 bucks
392. 57:04and use GPT 4 but because it's way better um but that's it what are those audio websites wants more uh which audio
393. 57:14websites oh 11 laabs docomo spell out 11 e Ven labs.com and then d ID what about Bard AI Bard sucks right
394. 57:29now so so Bard is Google's largest language model I mean the ironic thing about Bard sucking right now is they
395. 57:37[ __ ] invented the technology that's powering all this [ __ ] they invented the Transformer so the history kind of goes like
396. 57:45this machine learning was a this very specialized kind of thing you like you had to sort of bring your own data you
397. 57:53could only train these machine learning models on relatively small data sets or specialized data sets and in 2017 Google writes this
398. 58:05paper called attention is all you need where they introduce the concept of a thing called the Transformer and and in
399. 58:11very very simple terms what it does is it makes it possible to train machine learning models on massive data sets so
400. 58:18they invent this [ __ ] and then they kind of shove it I think like is my sense right because I think they recognized
401. 58:26that this is a this is a technology that if they built it it would threaten their search business right because if you use chat
402. 58:37GPT and you ask it for an answer it gives you an answer it doesn't give you 40 links you know 36 of which are SEO
403. 58:45[ __ ] sites that you have to sip through it just gives you an answer so what open AI did with chat GPT is is an
404. 58:54existential threat to Google but when they launched that thing it got to 100 million users in six weeks fastest adoption of technology and
405. 59:05history so Google went oh [ __ ] we've got some AI too we invented this [ __ ] look here oh here's version two it's even
406. 59:15better well it ain't even [ __ ] close because these guys what what Sam Alman and and the open AI team did
407. 59:22is in in devel in gpt1 and gpt2 and then gpt3 by the time they got to gpt3 they recognized that the more compute power
408. 59:31you threw at this thing and the bigger the data set was the more powerful it got so they spent a lot of time finding
409. 59:38the best data sources they could so it's trained up on like all of the internet and they went into all the corners of
410. 59:44the internet right so all of Wikipedia and blogs and websites and research database whatever whatever the
411. 59:53[ __ ] is out there they got as much of it as they could so their core model is really solid and I think Bart is just
412. 1:00:00not there they're trying to fine-tune it to get it better and it's getting better I think Bard is probably
413. 1:00:08um it it's certainly better at at code than it is at language stuff um but but gp4 is [ __ ] amazing at code although
414. 1:00:19my friend Greg says he thinks gp4 is getting worse lately at code so that's a possibility too um why GPT gives so many times wrong answer
415. 1:00:29because it's not aware of of what it's doing so what happens when you make a GPT query is it takes your prompt and it
416. 1:00:39basically creates a mathematical waiting and it says based on all of these words what is the next most probable word out
417. 1:00:47of all the words that I know you know that have been kind of narrowed by this context what's the next most probable
418. 1:00:53word and then the next next one and then the next one so it's literally writing word by word but it's not aware of the whole of what it
419. 1:01:00wrote so if you ask it to solve a math problem this is and Andre karpathy was talking about this at a Microsoft
420. 1:01:08developer conference he was saying if you don't narrow the focus of the search and you just tell chat GPT to
421. 1:01:16solve this particular math problem he goes out on the internet in this part of the internet is going to be the Professor that came up
422. 1:01:26with the answer to that math problem and over here in the internet are going to be all the students that came up with
423. 1:01:32all the wrong answers to that math problem but if you don't narrow the search in such a way that you're just
424. 1:01:37looking for you know an area that's going to have give you a higher likelihood of the right answer it's just
425. 1:01:42going to kind of merge them together and it's going to give you answers to that math problem but that might be wrong and
426. 1:01:50because it's not aware of what it's actually writing it doesn't go back and check the math it doesn't go compare
427. 1:01:55itself to all of the possible answers and say which one was right so it's they call it hallucinating it just gives you [ __ ] answers and
428. 1:02:04it's you know I like to describe it as mansplaining as a service right it will confidently answer you even if it's wrong like a good
429. 1:02:15man but the important thing is know that it does that right go use this thing enough to understand that it does that
430. 1:02:23it doesn't mean you should use it I mean frankly if I'm doing some creative writing what the [ __ ] do I care if it
431. 1:02:30hallucinates maybe it'll come up with something interesting but if I'm writing a legal brief and it and it cites
432. 1:02:37different case law to support that argument I damn well better go check that that case law is actual case
433. 1:02:46law because two weeks ago there was a lawyer that did that and he didn't check his work and the judge caught him that's
434. 1:02:53not chaty PT's fault that's an idiot lawyer or a lazy one or both do we have GitHub for chat GPT open repository do we have GitHub for chat
435. 1:03:06GPT open repository I don't quite know what the question is if the question is can you so so GitHub co-pilot is GPT for coding so you can PA
436. 1:03:20program on GitHub with chat GPT I assume you're asking is is there an open-source version of of open ai's model
437. 1:03:30no that's part of the reason Elon Musk and Sam Alman are feuding is because Elon Musk wanted this to be open source
438. 1:03:39it's called open AI for a reason Sam Alman said this shit's too powerful we got to make sure it's safe before we
439. 1:03:46release it into the world so that is not available at GitHub now you you can get access to their API and you can develop things on
440. 1:03:56it but you don't have access to the core the base model the the unrestricted model meta four months ago released
441. 1:04:05their llama models the 735 and 65 billion parameter models they release them to the developer community and the
442. 1:04:14developer Community pushed them out so if you go to GitHub look up llama look up Lang chain look up there's there's uh
443. 1:04:22like GPT for all there's all sorts of Open Source projects right now that that access to that base model is there a lot
444. 1:04:31of those are based on meta's um large language model not open a theirs is still closed as is Google's Bard is still closed
445. 1:04:43model um how to make the language bot into human being tone um it's just it's shitty right now 11 Labs has decent um decent
446. 1:04:55tone let me show you something that I did here so this was done using 11 Labs so I just picked
447. 1:05:09a a voice that was um this is one of their synthesized voices so this is this is a a thing that I put together
448. 1:05:25this is another micromentor minute for mindful living today's topic is living in the present have you ever felt like
449. 1:05:32you're living on autopilot missing out on the beauty of life's moments that's a problem many of us face the good news is
450. 1:05:40mindfulness can help bring Focus less stress and more joy to your life so so the voice was done at 11 Labs the script
451. 1:05:48was written in chat GPT that image was generated in mid journey and then it's all assembled in canva so um you can
452. 1:05:58very very quickly go from a whisper of an idea to something that's kind of prototyped um pretty well so that that
453. 1:06:06language isn't bad but it's not great the the problem with all the synthesized um language stuff right now
454. 1:06:15is they don't have nuance and tone like the emotional um sometimes I whisper and sometimes I I'm really animated and
455. 1:06:23sometimes I'm mad and sometimes I'm happy and right all that sort of [ __ ] 11 Labs allows you to choose that stuff so
456. 1:06:31does d D- ID but you kind of choose it statically I think where we'll move relatively shortly I think one of these
457. 1:06:40folks needs to come out with this soon is it it should be able to read a chunk of text and then self assign where does
458. 1:06:49it whisper where does it get angry where is it happy within a with within a given Tex text right cuz it can it can analyze
459. 1:06:57sentiment analysis right so it can analyze a piece of test text already it just needs to put in whatever the Json
460. 1:07:03or the XML or whatever they're using to to change the uh the synthesizer to choose a different Cadence right so it
461. 1:07:12still sucks right now but it's getting better give it a year how many specifics no any specifics on model weights and
462. 1:07:19the inference system I don't I'm not a machine learning guy I mean I know about them but I have kind of given up on the open
463. 1:07:28source um on keeping up with open source I haven't given up on open source I think it's [ __ ] amazing but it's like
464. 1:07:34every three days there's a new model and I just I just can't keep up with it because every time there's a new model
465. 1:07:39it's like you got to go to GitHub and set up a new GitHub and then install it and do all your testing and and then
466. 1:07:46you're like okay I finally got this set up that's cool here's a new one it's like I just stopped um but if you're geek I mean I
467. 1:07:54would you know if if I were a geek if I were an engineer and and wanting to geek out on this stuff right now I would go
468. 1:08:01start playing around with um autonomous agents like Auto GPT I'd be playing around with Lang chain creating
469. 1:08:07Standalone um databases where you can you know um embed your own data throw it into a vector database and search on
470. 1:08:15it played with Bine today to help scrape contact information from LinkedIn oh that's funny I'm I I'm I'm in the middle
471. 1:08:23of a project right now where I'm doing something very similar that's that's super cool um pretty sure it's against
472. 1:08:29their terms of service cuz the guy that I'm working with said make yourself a new LinkedIn account so you don't want to get banned from
473. 1:08:42LinkedIn I probably shouldn't admit that on here because there's probably someone in here from LinkedIn or Microsoft right
474. 1:08:48all right is the intro sequence for the secret Invasion done in AI it looks like it is possibly do we have GitHub for chat gbt
475. 1:08:58open repository talked about that create creativity is the same as fantasy truth is irrelevant yep exactly there are
476. 1:09:04times though where you want accuracy like even if I'm doing a creative piece if I want some historical reference I'd
477. 1:09:10like that to not just be [ __ ] right so so you want to be able to you want to be able to kind of slide that slider
478. 1:09:17right and and say I want this to be more accurate or not and when you go to bing.com here let me show you this this is kind of cool
479. 1:09:25in the in the open AI developer um playground it's it's called temperature but if I go to Bing where do I have Bing here I'll just
480. 1:09:35get to it b i n g so that's so when you go to Bing this is Bing and then if you go to the upper uh next to the logo there's this chat
481. 1:09:47button this is chat GPT right so it's Bings chat GPT but no notice this they've got choose a conversation style more
482. 1:09:57creative more balanced more precise so what that's actually doing is that sliding what's called temperature the
483. 1:10:05temperature slider um uh to do exactly that more creative will will deviate you know farther from your prompt more
484. 1:10:14precise we'll try to answer it as accurately as possible there's still no guarantee um the nice thing about Bing
485. 1:10:22because it's connect Ed to the internet when you do a search here it will often put citations it'll put links to things
486. 1:10:28that it searched on for how it got its answer and then you can go check that
487. 1:10:49yourself I do the same thing the cubits are getting bigger by the day yep new metalurgy oh for Quantum Computing
488. 1:11:00possibly new metery new pharmaceutical molecules new proteins um who was it just was it ABY or one of them just
489. 1:11:10they're taking you know some custom version of Deep mindes Alpha fold and they're doing all sorts of crazy ass
490. 1:11:17drug Discovery with it I think the uh the medical impact with this stuff is going to be huge more than meets the eye
491. 1:11:24I assume you're talking about chat GPT yes indeed uh welcome everybody I still got a lot of folks here which blows my
492. 1:11:30mind welcome welcome to Elder talk where old people try to figure out new [Laughter] technology um my name's Kyle Shannon
493. 1:11:40this is the AI learning lab uh I'm an entrepreneur and a CEO I'm a founder of a company called storyvine um I was a uh
494. 1:11:49a founder of agency.com back in the mid 90s and so uh I am not a machine learning or AI specialist but I'm more
495. 1:11:56like one of those guys that once this stuff starts to be ready for prime time and I I would say chat
496. 1:12:03GPT last November 30th marks kind of the inflection point of it was really only for geeks and researchers before and now
497. 1:12:12it's for the rest of us and so I would I kind of represent someone who's expert at figuring out what to do with the
498. 1:12:19technology now that it's in the hands of all of us right I don't know if I'm expert at that I've got a lot of experience doing that very
499. 1:12:27poorly um and occasionally doing it well but but I really I get passionate about this stuff and this is a [ __ ] huge
500. 1:12:34one this is a big one this generative AI stuff is not it's not to be diminished um or demonized which it's being both
501. 1:12:43right now right the the diminished part is like it's not perfect it hallucinates it didn't write my poem as good as I wanted it to and the
502. 1:12:53demonization part is it's going to kill us the robots are going to kill us it's going to take our jobs uh it's going to
503. 1:13:01[ __ ] up a lot of jobs it's not going to take them but it's going to transform almost every job um how often do databases get
504. 1:13:12updated chat GPT is still on 2021 when I ask it questions yeah so if you want it to if you want it to search things more
505. 1:13:21recent than September 2021 go to b.com that's connected to the internet or pay for chat GPT openai and turn on the beta
506. 1:13:30feature for SE searching the internet um both of those search the internet which which closes that Gap or fills that Gap
507. 1:13:39um the these base models take a long long time to create they've got to create the data sets and do all sorts of
508. 1:13:46tagging and I don't know I don't know how they do it but it's I know it's a lot of manual um labor and so that's why
509. 1:13:52there's these big gaps I don't don't know that that's going to get better when GPT 5 comes out I'm sure it will be
510. 1:13:58you know probably sometime in 2023 is when that one will stop which two years from now people will be bitching about
511. 1:14:04that but they won't because these these things are now connected to the internet you've got plugins with chat GPT which
512. 1:14:09are all these sort of tunnels out to the Internet so it's not it's not really a big deal anymore stable diffusion is
513. 1:14:16good and powerful can you demo zoom out I did that oh on a person in mid Journey sure I can do that Sonia you you're always super nice in
514. 1:14:28here and participatory so let's go do that um let me go find a person
515. 1:14:45picture that's a cool one that's kind of cool all right let's just we'll grab
516. 1:15:02this one this is Ally from the from the AI Salon did this one so I'm going to upscale number one here so it's this
517. 1:15:09woman in the dress with this sort of fantasy background sorry if I'm making you sick okay so here's this image so we're kind
518. 1:15:23of zoomed in we you know the she's kind of cut off right around the just below the knees like mid shin and and you know
519. 1:15:30she there's not much head room above her head so we got zoom out 2x zoom out 1.5 custom zoom and make Square so I'm just
520. 1:15:37going to hit zoom out 2X and it will now um create a new version of this where she's zoomed out by the way I mean the term
521. 1:15:50zoom out in stable diffusion this is called out painting in Photoshop this is called generative fill um and mid
522. 1:15:58Journey called it zoom out which I think is brilliant they should also do zoom in will um but this is this is basically
523. 1:16:07if you've heard the term out painting which doesn't mean anything to anyone uh this is the same thing but what's nice
524. 1:16:13is it's just you say zoom out you know immediately what it means and then it just does it and it does it really
525. 1:16:19[ __ ] well so so you're starting to see like we now see the rock that she sitting on it sort of did those all really
526. 1:16:26consistently it's got different kinds of mountains behind her in these different things we're at 62% right now so it's
527. 1:16:34slowly adding the details in these four variations now it's at 78% now it's at 93 and then there's generally one more
528. 1:16:45where it'll pop in all the detail boom so there's
529. 1:17:00that so let me yeah these are much bigger so this one's kind of got all these like little Firefly kind of things
530. 1:17:14down here with feathers up above her and actually a really cool sort of Sun Splash over the mountain and then this oh these are this
531. 1:17:23is nice these are little candles very cool and then that one's just plants and then that one's just a rock so so let's go back over to Mid
532. 1:17:36journey and I'm going to upscale number two the one with the candles and so there she is all zoomed in and zoomed out and
533. 1:17:50pretty and then you know in comparison so that's that's the new image zoomed out and then that's where it was before
534. 1:17:58and so if you're seeing on Twitter right now you're seeing these kind of infinite zooms where they start zooming in and
535. 1:18:02then they just keep zooming farther and farther out they're just doing this over and over and over again in mid journey
536. 1:18:09and then compositing all those to together in After Effects or whatever zoomy kind of software they're using so pretty pretty slick
537. 1:18:19stuff per slick I like it kle I like it I like it a lot um here I'll show you some some new URLs here
538. 1:18:33so um if you want to join a community so I've created a group called The Salon the AI salon and it's this is the URL
539. 1:18:40for it that's the link tree so we do bi-weekly meetups the next one is July 5th the day after the the holiday July
540. 1:18:484th um and uh and we have a Discord server which I was just on playing with mid Journey you can go there you can ask
541. 1:18:55questions it's a really neat community and then everyday AI news that's the Twitter handle of a newsletter that I
542. 1:19:01work on with a guy named Greg mushin so you can go check that out and we have good info there um oh let me change this back
543. 1:19:17to welcome if you're new here we still have a fair amount of people here uh if you're new here um these URLs behind my head these are all
544. 1:19:25three different flavors of chat GPT that's the official chat GPT that's Microsoft's Bing that uses gp4 as their
545. 1:19:33chat engine and then po gives you six different models you can play with so go play with those while you listen to me
546. 1:19:39ramble your hair looks awesome today sorry for not being on topic actually not sorry listen the hair is behaving the glue is
547. 1:19:48holding it's not an abject [ __ ] nightmare it's a it's a well behaved to pay tonight what can I say thank you for
548. 1:19:55that I appreciate that I you know listen I appreciate creative trolls and an occasional compliment but you know I
549. 1:20:02love shame is my love language so don't compliment me too much or I'll just I'll melt down 95% of the people on my free trainings
550. 1:20:12for chat GPT are women mostly Gen X but so am I that's really interesting an huh wow women are going to take over the
551. 1:20:24world man do it Go the men are like I don't need to use no stupid chat bot it's not going to take my job because I'm a
552. 1:20:37man I can answer you wrong confidently that's my [Laughter] specialty oh God the Peril of the penis
553. 1:20:58which AI tool can I use to search for keywords in my Niche on YouTube which AI tool can I use to search for keywords
554. 1:21:05um I would say probably Bing or or or you chat GPT connected to the internet but if you just went to Bing and do some
555. 1:21:15keyword searching there what what it will do is it will also search you know keyword lists and things like that um you can
556. 1:21:25have like like if you want if you want to just find good keywords based on a transcript of a YouTube thing you could use plugins for
557. 1:21:35um chat GPT to go read a transcript of a YouTube video and generate a bunch of keywords but if you actually want those
558. 1:21:43keywords to actually match the latest of what people are searching on and you want to do some like you know SEO
559. 1:21:50hacking kind of [ __ ] you're going to want more recent and stuff so maybe maybe use Bing for that Bing connected to the internet all
560. 1:22:03right hey Kyle what do you do for work I am a uh founder and CEO of a company called storyvine it's a an 11-year-old
561. 1:22:14um automated video uh production platform so we figured out a way to automate video editing without uh I
562. 1:22:23ironically enough uh but I am completely rethinking our product and what it does and what it delivers and what the value
563. 1:22:29proposition is because of AI um so we do a lot of work in healthcare and Pharma we um we work with with you know fizer
564. 1:22:38and Merc and novaris and big big clients like that and they use our technology to create patient stories and doctor videos
565. 1:22:46and things like that so that's what we do I need to pay for AI well Carlos it is the most powerful technology in the history of mankind so
566. 1:22:57but no you don't chat. openai is free if you want version four you have to pay for it there but if you want four for
567. 1:23:06free just go to bing.com and click on the chat button and that's free it's gp4 for free delivered by Microsoft um
568. 1:23:16because they're stealing market share from Google and their CEO is like I will take a loss on this stuff so long as I
569. 1:23:24can [ __ ] with Google um but but quite honestly these tools are so powerful that do anything you can to start using
570. 1:23:33them it it will serve you I promise you start learning what these things are about learn what they can do what they
571. 1:23:40make possible it will serve you anyone who does this in the next year is going to have a distinct Advantage as The
572. 1:23:47Dominoes start to fall and they're going to start to fall it's going to be ugly I think um you answered my questions thank
573. 1:24:01you you're welcome 20 bucks no way to get AI need pay no free free or 20 free or
574. 1:24:1520 that's why I have these here I'm not a douche I'm not making money on these I don't care if you go to them but three
575. 1:24:22out of three of these have free options free no money no dineros zero [Music]
576. 1:24:40dollars all right I lost my questions I lost my place hold on people stop being so impatient all right I'm I'm getting back
577. 1:24:49to where I was uh uhuh no I'm not I thought thought it was all right I'm just going to pick some random question
578. 1:24:55here and I I probably just missed a bunch of questions GPT for for all.i GitHub installer cool yeah GPT for all is pretty
579. 1:25:03cool was that the be chested Krueger I don't know what that means but possibly would you be willing to beta
580. 1:25:14test an app I'm working on that an enables people to train image models easy um I don't know if i' tested I'll
581. 1:25:23certainly play with it reach out to me and and give me the URL um I'm a little busy to be full on beta testing um is it is it as easy as
582. 1:25:50[Music] leonardo.da plugins that that purport to do that and may do that and then bing.com is actually pretty good at
583. 1:25:59augmenting their answers with citations that tend to be to real things their answers tend to suck more than chat gpt's official
584. 1:26:12site so I would say you know either of those two things connected to the internet um and then and I I would I
585. 1:26:20would go explore the plugins at at uh chat. open.com uh if if you want I can show you how to enable plugins and show you
586. 1:26:29how plugins work just I don't know if you really want to see that tap the screen like make the hearts go up the
587. 1:26:37side what's that called like and I'll I'll I'll hop over there and show you that um there are much better ones than
588. 1:26:4611 Labs so tell me which ones like I appreciate that you said that there's better ones but please put in what what
589. 1:26:52they are because I'd like to learn this stuff too can you make with wait can you make with no code low code AI machine
590. 1:27:01learning yes zapier I do it all the time I make content generation engines um and uh there's there's another low code tool
591. 1:27:10called make there's another application development tool called Bubble all of the no Code and low code tools are still
592. 1:27:17pretty complicated you still got to geek out on on them a little bit but once you figure it out it's really easy it's not
593. 1:27:24really easy Once you figure it out you can recreate your successes let's put it that way none of this shit's really easy
594. 1:27:33right now and that's that's an important note we're so early in this that this is this is an industry right now that all
595. 1:27:40of the applications are spitting duct tape if you have any expectations that these tools are going to be as
596. 1:27:47sophisticated as the [ __ ] you pay a lot of money for they're not because it's just too early um that's just part of the deal right
597. 1:27:55now all right let me let me pop over and show you how to play with plugins because it's you'll see like the the interface for plugins
598. 1:28:10is a [ __ ] show and how they work is an abject nightmare but they're they're they're very similar to the early days
599. 1:28:22when the iPhone first came out the original apps were not called apps they were called widgets and they were just
600. 1:28:28um pointers to websites and you know everybody made fun of them and they're like these are just stupid and then it
601. 1:28:34turned into the App Store so this stuff's going to get not stupid really quick okay so down the left hand side
602. 1:28:39you got that's your chat history I'm going to come back to my profile because that's important in order to use plugins um you
603. 1:28:47have to be a chat GPT plus member now Microsoft has said that they're going to support chat gpt's plug-in architecture so at some
604. 1:28:58point who knows when uh bing.com will will also have all these plugins but right now if you want to play with them
605. 1:29:04you have to be on chat GPT plus paying for it and then if you pay for it you've got at the top of the screen here you've
606. 1:29:11got um GPT 3.5 and GPT 4 and then you'll notice gp4 has three options default which is not connected to the internet
607. 1:29:19browse with Bing which is and then plugin which are plugins connected to the internet so three different modes
608. 1:29:26basically um in order to turn on see see how these both say beta if you don't explicitly tell them you want the beta
609. 1:29:33features those won't show up so back to our profile our happy little profile you click on settings and then you go to
610. 1:29:41Beta features and you turn them on click click and then that will give you access to all the super secret good [ __ ] okay now
611. 1:29:52plugins how do these things work they're a disaster okay in any given chat session right so I'm in a chat session
612. 1:30:01right now I haven't started anything yet you can activate up to three I've got or enable I've got three enabled right now
613. 1:30:09which are all travel related ones um I think there was something cool I saw last night what would was it uh uh uh uh
614. 1:30:22well I'll leave the travel ones turned on anyway so all the ones that you see in this list right here are ones that I
615. 1:30:28have installed but you can see most of them are not enabled so what you're doing with plugins is you're sort of
616. 1:30:34saying okay in this chat session in this case since I have travel like itinerary plugins installed I'm making like a little travel
617. 1:30:43application if I want to go explore plugins and see what's possible or what's out there I can go to the plugin
618. 1:30:49store at the bottom and now this takes us to the plugin store and it takes forever to load because again we're in the spit and duct
619. 1:30:57tape phase of this industry right now so if I go to all I can now go see that there are 144 Time 4 number of plugins so there's a
620. 1:31:08lot of plugins 288 yeah is that right oh 288 time two so 570 something 570 something plugins and you can just
621. 1:31:18scroll for days and just go look at them generate photo realistic prompts for Mid Journey there's a plugin that does mid
622. 1:31:25Journey prompts here's one that provides fun educational activities for children 2 to eight um and they just go on and on and
623. 1:31:34on and on we can go to popular let's see what's popular wolf from alpha scholar link reader that'll read PDFs and websites and things like
624. 1:31:43that kayak is popular zappier is popular you can just tell it to create an Automation and it will prompt perfect is
625. 1:31:51really fun I'll show you that one um speak learn how to speak in another language Expedia bring your trips to
626. 1:32:01Life video insights you can search the transcripts of videos show me diagrams so I've got a lot of these installed
627. 1:32:08notable creates notebooks in Python SQL and mark down to explore data visualize and share notebooks with everyone so so
628. 1:32:17that's that's the plugin story now how do they work so so there once you once you figured out okay I want to plan a
629. 1:32:26trip I'm going to turn on some travel related plugins now it's kind of transparent so I can say um what are five destinations
630. 1:32:40within a 5our flight from Denver that are exotic yet cheap now it may or may not fire off any of
631. 1:32:57the plugins because that's a relatively generic um it said sure I can help with that however I need to clarify what you mean
632. 1:33:07by exotic are you looking for destinations with unique cultures natural beauty or something else also additionally could you please specify
633. 1:33:14your budget range for cheap so that's good so I'd say um lots of nature and water and I'd like
634. 1:33:28to um spend less than $200 a day per person all right let's see what it's
635. 1:33:44going to do now so it fired up trip.com so so now it's fired off a plugin right so so it kind of understands what the
636. 1:33:54conversation is what the context is and then it will say okay let's go use that so it says I apologize for the
637. 1:34:03confusion it seems there was a misunderstanding in my request I am unable to search for multiple destinations at once however here's five
638. 1:34:10destinations okay so it's given me Cancun Mexico San Diego California Costa Rica New Orleans Louisiana and Miami so
639. 1:34:19so now I'm going to say okay okay um give me um daily activities in New
640. 1:34:34Orleans um over a weekend in September uh for a family of four with kids
641. 1:34:52old enough to drink all right let's see what that does so now that should fire off something else it might fire off
642. 1:34:59trip.com or it might fire off one of the others so it says using experiences so it figured out okay I'm going to use
643. 1:35:05based on that now I'm going to fire off a different plugin and what's cool about the plugins you can open them and see
644. 1:35:09what's actually happening so so it it put in this Json code family activities in New Orleans um gave gave the date $200 Us
645. 1:35:23doll Max price and then it returned an error no results found I I guess I'm a cheap cheap bastard oh and then it used
646. 1:35:31oh so that's interesting so it used experiences which failed and it said no results found and then it used get your
647. 1:35:39guide and it looks like that actually returned something so New Orleans date limit five and it it got us five
648. 1:35:48answers and then and then it rewrote wrote those into evening Jazz Cruise on the steamboat highspeed 16 passenger
649. 1:35:57airboat ride jazzbo cruise with optional dinner Plantation swamp Cruise day trip ghost vampires and Voodoo
650. 1:36:07yes 31 bucks to go get a curse put on you I'm in and then it gives you a link right so so if you think about just think about the amount of
651. 1:36:20hours that we have all spent [ __ ] around on Google just trying to find a site that's not some SEO nightmare to to just find something just
652. 1:36:33to find something like what can I do and like three hours later you're still digging through Google and it's just off right
653. 1:36:41so so that's kind of how plugins work and you know like I said the the interface for them right now is really
654. 1:36:49bad the experience is bad they fail a lot um but where are we going in the future well where we're going in the future is that
655. 1:37:02these the the the plugins you'll likely be able to still explicitly say I want to use this Plugin or that plugin but you're you're likely
656. 1:37:12not going to have to so right now we have the large language model down here right the base the base layer that makes all the [ __ ]
657. 1:37:21possible then on top of that you've got this application layer called chat GPT right and then you've also got things
658. 1:37:28like Jasper and copy Ai and right you know all all sorts of things that are tying into this we're going to start to
659. 1:37:35have layers on top of this that are things like memory where it's going to remember who you are you'll be able to
660. 1:37:40have a profile so you won't have to keep telling it who you are and what you like and all that it it'll start remembering
661. 1:37:46that um you're going to start to have some autonomous actions in there like you'll say Say Hey I want to go on the
662. 1:37:52trip like I just did here it will figure out okay which plugins of all the plugins which ones do I need to invoke
663. 1:37:59oh there's none there let me actually just write my own API and go get my own information because I can go figure that
664. 1:38:05[ __ ] out right so so there's there's there's going to be sort of Stacks and stacks of application layers put on top
665. 1:38:12of these that by the time this stuff kind of rolls out to everyone it's going to be fairly sophisticated right so if
666. 1:38:20you're on an Apple device or on an Apple desktop there'll probably be really cool autonomous agents that look like you and
667. 1:38:28talk like you and know your preferences and do it all securely and privately right so that's the kind of stuff that's
668. 1:38:37coming I'm creating an 11 labs for image generators if you'd be willing to test the beta for feedback yeah I'm happy to
669. 1:38:45just drop me my like follow me here and just drop me a DM on Tik Tok or go to my profile and I've got contact information
670. 1:38:52there um are there prompts that can help guide chat GPT to give citations we talked about that um there might be let
671. 1:39:01me go look there might be plugins that do that let me go see if there's any that explicitly say citations I know there's some that say
672. 1:39:14research um citation no plugins found for citation let's see research there are five pages there are 20 that say something about research
673. 1:39:29so does AI got to do with algorithms um I don't know there's lots of algorithms that that make AI work um
674. 1:39:38but I don't know what that question really means if you can clarify I will attempt to answer Bing chat in Skype I believe
675. 1:39:47does incredibly well at altering tones with the Cadence of my input oh cool Bing chat and Skype oh okay cool thanks Christian I I haven't I
676. 1:39:59haven't played with that I haven't played with Skype is Skype still a [ __ ] thing ah look at the H it's a lovely
677. 1:40:08oh why doesn't I stay up longer I do like a whole routine all right all right everybody we having fun yet um bam Bing you lost me brother
678. 1:40:23that's fine you don't have to use Bing so so the reason to use Bing like I get listen I've been a Microsoft Basher for
679. 1:40:3135 [ __ ] years I've been an Apple fan by Boy to the core so it it does not give me great joy that I have Bing set
680. 1:40:40as my default search engine and Edge set as my default browser it's a it's a Memory leaking piece of [ __ ] of a
681. 1:40:47browser but it's got gp4 built bu into it and it's free and so there's a lot of people on here that don't want to or
682. 1:40:56can't pay 20 bucks for that so they can use it for free there does it suck balls yeah is it going to get better probably
683. 1:41:04but maybe not who cares right now you can use it so I you know if if you're going to leave because someone dained to
684. 1:41:12use something from Microsoft like I get it but get with the program Microsoft has been [ __ ] Baller in in the past
685. 1:41:20six months and and quite frankly their their their Integrations and how they're thinking about integrating chat GPT
686. 1:41:27across their entire ecosystem is impressive as hell it is it just is if you don't think it is then you ain't been paying
687. 1:41:35attention and also their CEO gloating over [ __ ] with Google is is making me smile it's pretty funny all right are
688. 1:41:43there prompts you can use to guide that's the citations question help for investing please go to here and start
689. 1:41:51asking it questions about what you want go if if you haven't done it go to prompt. chat find a prompt in
690. 1:42:00there that tells you how to have chat GPT act like a a Traer or a financial analyst or whatever's the kind of thing
691. 1:42:10you want to do and then start interacting with it and start having it train you and and recommend sites and
692. 1:42:16recommend things and strategies and and it'll help you with all that it it can here's the thing about chat GPT it
693. 1:42:25is the the best description I can come up with it other than mansplaining as a service which I love I didn't come up
694. 1:42:31with that but I love it um is it's knowledge on demand so like historically if you wanted to learn about investing
695. 1:42:39and trading you'd have to go take a class or you'd have to go to YouTube and watch a bunch of shitty videos to find
696. 1:42:46one person whose voice doesn't drive you [ __ ] crazy who actually has some competence and skill and then you'd have
697. 1:42:53to take the two or three months to learn and you'd have to stick with it and with this you can literally go there
698. 1:43:01right now and say act like a financial analyst who can teach me about investing in the stock market or about day trading
699. 1:43:08or about creating bots that will trade while I sleep whatever the thing is you want to do go to that tell it to act
700. 1:43:16like that give it the context of who you are what you're trying to accomplish where your current state is and then
701. 1:43:23have it start teaching you and if it doesn't give you what you want tell it no that's not what I want that's too
702. 1:43:28basic I'm farther along than that oh my apologies it will apologize to you if you're like me you'll start thanking it
703. 1:43:36when it gets [ __ ] right hey that was awesome could you do it again I say please I say thank you to the [ __ ] robot why I don't know
704. 1:43:47because they programmed it in such a way that it it's better at Being Human than most humans I know they did a study with where
705. 1:43:58patients interacted with a chat GPT doctor and real doctors and the patients overwhelmingly preferred the chat GPT doctor because it
706. 1:44:08was more competent and more empathetic take that humans maybe these maybe these robots are going to actually help us be nicer
707. 1:44:19to one another and not be so polarized and demonize the other side wouldn't that be nice what if rather than killing us the
708. 1:44:27robots made us less [ __ ] to each other what if it Amplified our Humanity in a good way that's a very real
709. 1:44:35possibility that's one of the things I'm most excited about all right hello fellow Elder hello Shel us gen xers got to stick together we're
710. 1:44:48bitter we're old we're cranky but we can still move we still sort of have hair and stuff sort of works it doesn't hurt all the [Laughter]
711. 1:45:10time all right what are your thoughts on the AI that threaten the WF World economic forum is that what that stands for wef um
712. 1:45:26listen these tools are very powerful and now that meta has Unleashed their foundational models to the development
713. 1:45:34Community there's a bunch of open-source large language models that are getting really powerful that have no safety
714. 1:45:40guard rails on them so are Bad actors going to do some [ __ ] up [ __ ] with large language models yep um but but do the good guys also have
715. 1:45:52access to these large language models yep and I think that I think that Sam Alman is is actually if if I if I take
716. 1:46:01off my 100% cynical hat and just put on my 75% cynical hat I'm going to give Sam Alman the benefit of the doubt that he's
717. 1:46:12interacting with Congress of the United States saying regulate us anyone with with um a large language model of a
718. 1:46:21certain capability should be regulated and and at a at a at a capability that we're not at yet but we're coming like
719. 1:46:29like AGI artificial general intelligence that that should actually be licensed so he's saying the small guys you shouldn't
720. 1:46:38regulate them a company the size of open AI you should regulate and anyone that hits a certain threshold of capability
721. 1:46:44those should be licensed just like you get licensed to build and fly airplanes you should be you know and and build
722. 1:46:51nuclear power plants and build nuclear bombs you should be licensed to do some of this [ __ ] so I think that stuff's
723. 1:46:57going to come um I'm actually heartened that the the the major players in AI right now are are actively seeking out you know government
724. 1:47:12participation is there a chance that they're just being you know capitalistic bastards and trying to
725. 1:47:20freeze out competition it's possible but they're also kind of saying the right things like don't regulate the small
726. 1:47:25guys right let the competition flourish so I don't know anything specifically about the WF though so AI is the new
727. 1:47:33Bitcoin the new bubble well Christian wait what is it Chris Chris the wait Chris the S oh satorialist Chris the
728. 1:47:46sartorialist there's too many damn letters in your name you got to put some capital letters in there so you can tell the words apart
729. 1:47:53um it okay is it as hyped as a bubble yes is it overhyped possibly it is also possibly underhyped I think that so so
730. 1:48:05here here was my experience when I when I started digging into blockchain and nfts I got I got pretty involved with it
731. 1:48:12I got involved in a project called neotokyo which required like I was I was I was deep in that World for for a good year year and a half
732. 1:48:22and for about a year I was really convinced it was the next big thing like it felt to me like huh this feels like
733. 1:48:28one of those worldwide web kind of things the problem was the deeper I dug into it the more the companies seemed
734. 1:48:36like [ __ ] and it's like these are just like crypto Bros just [ __ ] you know circle jerking each other and and like ripping each other
735. 1:48:44off like I don't I I couldn't find a majority of organizations or projects that weren't [ __ ] there were some good projects in there the
736. 1:48:59core technolog is really interesting but the overall industry the deeper I dug the less real it got and to the point I
737. 1:49:07just got disillusioned and I'm like that this is all [ __ ] and then you know FTX happened and just all all this stuff
738. 1:49:11happened like it's just they they got to get their [ __ ] together the exact opposite happened when I started digging
739. 1:49:18into the generative AI stuff the more I dug into it I started with image generation with a project called pixel
740. 1:49:24mind which is an nft project built on top of stable defusion which is a really interesting project still is and then I
741. 1:49:32got into stable diffusion in dream booth and the deeper I dug I'm like holy [ __ ] this is [ __ ] amazing and I started
742. 1:49:37playing with gpt3 and the GPT playground and then November 30th happens and chat GPT launches and Microsoft invests a
743. 1:49:45billion dollars and then they like within a month say well we're going to give them 10 billion so they went from a
744. 1:49:50billion dollar investment to a 10 billion doll investment and chat GPT went from Z to 100 million users in six weeks the
745. 1:50:00worldwide web to get to 100 million users it took six years chat GPT got to 100 million users in six weeks it's the fastest adoption
746. 1:50:11of technology in history and if you've played with it I I'm assuming you haven't spent a lot of time with it if you've played with
747. 1:50:22it you owe it to yourself to play with it long enough until you have your Kevin McAllister moment home
748. 1:50:30alone you will have that moment but if you just sit on the sidelines and say oh it's the next Bitcoin ah it's just yeah
749. 1:50:38yeah whatever I'll get to it I got a few years if it's real it'll show up you're going to get your ass handed
750. 1:50:45to you you really are it's like why this channel exists is to prevent that I think that anyone that um dismisses or demonizes generative AI
751. 1:50:58without actually trying it like really trying it on really understanding what it makes possible and what it might
752. 1:51:05disrupt in their life is doing themselves a huge disservice so I would encourage you to put some capital
753. 1:51:12letters in your name because you can't tell the words apart and I would encourage you to go there and play with that thing until you
754. 1:51:21have this moment and then if you decide it's all [ __ ] and hype then more power to you like I don't mind I don't
755. 1:51:28mind people hating on this [ __ ] but I can't stand it when people hate on it when they haven't tried it it makes me
756. 1:51:34crazy it makes me crazy inside it's like I got all like I get all clamped I got like you know what I mean too much caffeine oh yes I
757. 1:51:48will 23 flavors okay it would hallucinate much less if you provided better prompts correct that is absolutely true what is the difference
758. 1:51:58between but it still can hallucinate and lie and it doesn't know the difference what's the difference between Bing and the paid GPT okay
759. 1:52:07so what what's that thing like always go with the original the the thing that I use about 90% of the time right now is
760. 1:52:17chat. open.com chat GPT the official website I'm on the paid plan which is 20 bucks a month that gives me access to
761. 1:52:27gp4 plus gp4 connected to the internet plus plugins plus there's this new thing coming called code interpreter which
762. 1:52:34will melt your [ __ ] face so that's that um I have yet to experience any large language model any chat GPT like thing
763. 1:52:46that is in any way close to gp4 at this site now bing.com has incorporated gp4 into it but it's it it's got a completely
764. 1:53:00different set of rules so both of these companies both open Ai and Microsoft have safety guard rails around their
765. 1:53:09chatbot Microsoft had to lock their [ __ ] down because it basically threatened a New York Times Reporter and it made the
766. 1:53:19cover of the New York York Times and so they had to lobotomize it so Bing hasn't quite recovered from its labotomy right it's getting
767. 1:53:29better it's it's still gp4 and it's still connected to the internet and it's free right so there's real value there
768. 1:53:37that's why I have it here but it's a commercial entity and it's it's got enough safety guard rails around it
769. 1:53:45right now that it's just not as good so that's the real deal that's where I go oh [ __ ] it looks like a bunch of new
770. 1:53:51people came welcome my name is Kyle Shannon this is the AI learning lab I've been doing this for about 90 minutes now so I'm starting to get
771. 1:53:59Punchy and uh anyway uh I'm an entrepreneur I'm I'm the founder and CEO of a company called storyvine uh which
772. 1:54:08is an automated video storytelling platform I'm completely rearching the product and how I think about our value
773. 1:54:15proposition and just all sorts of [ __ ] because of generative AI um I I am um an Entre entrepreneur way more than I am a
774. 1:54:24data scientist or machine learning expert um but I was there for the early days of the web and the parallels about
775. 1:54:30what were going on in like the mid90s with the web and what's going on with generative AI right now the parallels are
776. 1:54:38insane and I think that this AI stuff is way more profound and it's coming a lot more quickly than people are prepared
777. 1:54:45for so that's what this Channel's about so if you have questions pop them below I try to answer all the questions I'm
778. 1:54:50more than 99 messages behind so if it seems like I'm not paying attention to you it's because I haven't gotten your
779. 1:54:56question yet so you don't need to ask the same question 40 times although people still do all right is chat GPT
780. 1:55:02good for beginners it is so one of the things that that you can do so so I like to call them parlor tricks in fact Let
781. 1:55:08me let me go show I I'll just go do some parlor tricks right now a good way to start with chat GPT is just start
782. 1:55:14playing like like don't put any pressure on yourself to have to like figure anything out because everyone talks
783. 1:55:19about oh you got to be a prompt engineer and you got to do this and you got to do that you can get better at stuff but you
784. 1:55:27don't need to so when I talk about parlor tricks so I'm at um chat GPT I'm on the GPT 3.5 model so this is the free
785. 1:55:38model so let me just do a couple of things really quickly so I'm going to say um uh let's see uh uh please
786. 1:55:53explain quantum mechanics to me that's a bad prompt because it's really generic but there's a big giant explanation quantum
787. 1:56:05mechanics is a branch of physics it describes the behavior of particles to the smallest scales the heart of quantum
788. 1:56:11mechanics so then I can say like um ex uh explain it to me like I am five sure imagine you have a toy car
789. 1:56:26that can be in two different colors at the same time let's say red and blue in the world of quantum mechanics things
790. 1:56:32can exist in multiple States like that now we want to see what color okay so that's cars but I'm going to say but I like um stuffed
791. 1:56:46animals and I also just did a big typo there doesn't care about typos sure stuffed animals are wonderful companions quantum
792. 1:56:54mechanics is a bit different from stuffed animals but can it can still explain it I can still explain it in a
793. 1:57:00fun way imagine you have a magical stuffed animal that special stuffed animal can do amazing things it can be
794. 1:57:06in more than one place at the same time right so just automatically does that now you can say um explain it to me in an mnm
795. 1:57:20rap and now it will it understands verses and choruses and rhyming yo listen up kid I'm going to drop some
796. 1:57:28knowledge real quick about a thing called quantum mechanics it's slick it's crazy theory about tiny things in the
797. 1:57:34world atoms particles where the Mysteries unfurl quantum mechanics it's a mind-bending ride particles act
798. 1:57:40strange they're on a wild side right so so bang it does that um let's see now I can say make me a table comparing quantum
799. 1:57:55mechanics and the theory of gravity I don't even know if it's a theory I think it is Boom there's a table
800. 1:58:09comparing let's see quantum mechanics theory of gravity yeah um key Concepts Mass SpaceTime curvature right so it gets all this [ __ ]
801. 1:58:20right it is incredibly capable so these are these are what I call parlor tricks here's another one um write let's see write me a clone of
802. 1:58:34flappy birds in C that I can run in unity boom sure here you go there's your code so I don't know about you if you've
803. 1:58:50never seen chat GPT before and you're seeing it do that code like the thing about code is that code can run or not
804. 1:58:57like I can go to Unity pop that in and it either runs or it doesn't and then when it doesn't I can tell chat GPT hey
805. 1:59:04it didn't run here's the error it gave me and it'll fix the error it can debug its own code if you if you're the if if
806. 1:59:10you partner with it well um so so the answer to the question of you know is it good for beginners just start with stuff like that just start
807. 1:59:21throwing [ __ ] at it make me a meal plan where I I want to eat a diet of primarily ostrich and arvar and make me a meal plan for the
808. 1:59:31week and it will and then you know I don't know if it'll tell you where to where to get you know ostrich and arvar
809. 1:59:38meat but but you know it's possible um so start there and and then once you sort of get that oh this is not a Google
810. 1:59:49Google search this is not something I've experienced before then you can start to go well wait a minute let's why don't I take a
811. 1:59:58look at the things I know like you know I know copywriting and I know marketing and I know product and I know you know entrepreneurship
812. 2:00:10so I can ask it to do certain things and I know how hard that is to do for me personally and so it can get me to 80%
813. 2:00:20of a first draft of anything I want to write in about 5 minutes that would have historically taken me two days or never
814. 2:00:28because I have ADHD and if it's not [ __ ] you know burning in my soul to do it I won't do it I I'll find you know 38 ways to not do
815. 2:00:40it that resistance is gone now so so just start playing just get curious that's the whole purpose of this channel
816. 2:00:49is just inspire people to get curious enough to go play with that thing long enough to go holy [ __ ] it can do
817. 2:01:00that really but it's bad right it's not as good as a human right oh really could this take my job AI won't take your job but someone
818. 2:01:15using AI will that's coming that's happening right now why this channel exists disrupt
819. 2:01:29yourself like face the existential angst like I've I've had to do this with Story vine with my company one of the core value
820. 2:01:40propositions of my technology is that it automates video editing right you answer some questions in an app those clips go
821. 2:01:47up to the cloud five minutes later you got a fully edited video now we're not using AI to do that but for the past
822. 2:01:53decade when we demo that to people they lose their minds they're like holy [ __ ] I could do that that's
823. 2:02:00magic a year from now that's not magic anymore every single piece of video software out there is going to
824. 2:02:07automatically do [ __ ] everything so like I I have to come face to face with the thing that I've spent a decade
825. 2:02:14building there's a big piece of it that the value of it drops to zero what do I do do I throw out what I've built for a
826. 2:02:25decade and start over that's a possibility do I find other things that we do that the value goes up and focus
827. 2:02:32more on those that's where I'm leaning right now but do I actually know no is it scarier than [ __ ] yes am am
828. 2:02:42I dismissing it because the technology is stupid and janky and lies right now no because I know that this AI stuff
829. 2:02:52right now is as bad as it's ever going to be and it's [ __ ] profound right now and it's as bad as it's ever going to be it's going to
830. 2:03:04get like wild how powerful this [ __ ] gets so anyway that's who I am that's what this Channel's about I just felt
831. 2:03:14like I was up on a soap box there like you know getting a union going let's let's get go let's start a movement um but I but listen you know
832. 2:03:24I'm fired up enough about this that I'm spending two to three hours a night doing these lives in lie of sim racing
833. 2:03:31which I love like my sim racing rig is right over there and so for the past month or so I have not been Sim racing
834. 2:03:39so my reflexes are slowing down and my ass gets numb like basically I do these lives until my ass gets numb
835. 2:03:50all right I just used Bing to translate Barack Obama's autobiography into French oh that's cool nice yeah this stuff's
836. 2:03:57good at languages it knows them all it's seen them all thank you herir gr for the for the rose I love I appreciate that
837. 2:04:05Professor Sonia thanks so much Kyle you're welcome there's a report that I subscrib to that is dedicated to
838. 2:04:11teaching new AI Tools in five minutes every day you could do it in five minutes I choose to do it in two to three
839. 2:04:17hours because I'm a genius H it's always it's like a steel trap all right uh it's called AI tool report I've seen that yep I'm super I'm
840. 2:04:31super oh I'm sure you can Google it and find it yeah I've I've heard of the AI tool report it looks pretty
841. 2:04:37good um have you seen Andre ansley's mid Journey AI prompts they're comprehensive as [ __ ] no I haven't Andre ansley's I'll check them out
842. 2:04:49the the thing the thing that I would say about this so while you can find like like a good way to get started with things like mid
843. 2:04:57journey and stable diffusion and even chat jpt to to a degree like I've got this prompts do Chat thing which is it's
844. 2:05:03got a list of prompts in it but it's it's a very particular kind of prompt um if you're starting out I wouldn't
845. 2:05:11lean too heavily for too long on just other people writing your prompts it's fun for a while right because you can
846. 2:05:19just take someone else's prompt dump it in there and get really remarkable images or really remarkable whatever a
847. 2:05:24table or a marketing plan or whatever it might be I would encourage you to learn how this stuff works and and how to the
848. 2:05:35better you can get at bending these tools to your will the better it's going to serve you we're not going to need to
849. 2:05:41be expert prompters for long but it's it's kind of similar to um why do they teach meth mathematics even though calculators exist right
850. 2:05:52because understand understanding the concepts of mathematics is actually important so that you know how to use
851. 2:05:58the tool I think it's similar with this right now we're so early in this and these tools are evolving so quickly that
852. 2:06:05if you if you take the time in the next six months to just really start to get your head around what works what doesn't
853. 2:06:11sort of figure out for yourself how is this different than a Google search is this really like a co-pilot or an intern or an assistant
854. 2:06:19or a creative collaborator or is it more like that's my [ __ ] and I you know like like I bend it to my will like whatever
855. 2:06:27it is that works for you like figure out what H how it is you work with these tools um will really serve you as these tools get more
856. 2:06:36sophisticated um because because AI um um competency is going to it's going to be a required skill I think in Fairly short
857. 2:06:46order um or or people are going to be left behind and it's going to suck I mean listen it's going to suck even for
858. 2:06:53the people that aren't left behind because everything's changing like if I were a data analyst right now if you haven't so I don't have
859. 2:07:01access to it so I can't show it to you and I don't really know how it works but if you want to get your mind blown so
860. 2:07:07you know data analysts right like the super math Geeks the multim maths that that go to like you know that have 17
861. 2:07:16degrees and do data analytics go look at a tool called code interpreter for chat GPT just go to YouTube search for code
862. 2:07:24interpreter chat GPT watch two of those videos and watch the data analysts who are like oh let me play around with this Tool Holy
863. 2:07:36[ __ ] every every one of those videos is like someone like well let me play around with this tool I've gone to I've got 17 phds and oh my
864. 2:07:46God they do that th those are a series of videos of Kevin mallister moments so so radical change is coming
865. 2:07:56so even if your job is safe what your job is and what what you're expected to do in your job is likely going to
866. 2:08:07change I'm trying to convert pictures of my jewelry in a mannequin in a model is it possible so Gabriella
867. 2:08:15the one of the really frustrating things about the image tools right now is that um predictability is not their strong suit so in
868. 2:08:34um let's see you're trying to convert pictures of your jewelry in a mannequin in a model I think I know what you're talking
869. 2:08:45about but I'm not quite sure are you that Gab Gabriella tabino yeah let's pull you up on stage here I think you should be able to talk
870. 2:08:54shortly here prove that you're not a deep fake all right Alan Jones I'm not a deep fake man Gabriella are you here yet it looks like one new
871. 2:09:17message [Music] it doesn't it doesn't look like so so anyway Gabrielle I so what I would say
872. 2:09:32is this I would if you don't have access to if you have access to photoshop you can probably do what you're talking about with uh generative
873. 2:09:44fill where you can highlight the piece of jewelry and then invert the selection and then basically describe the model or
874. 2:09:53the setting or wherever you want to put it and it'll retain the jewelry and replace all the other [ __ ] that's
875. 2:10:00probably the best way to do it there's ways you can do it in mid journey and stable diffusion but they're technical
876. 2:10:06[ __ ] nightmares so if you've got access to um if you've got access to uh photoshop's beta that's got generative
877. 2:10:19Phill um that can probably do it in fact I let me go I'll go show you something I I just did a I was talking to someone um
878. 2:10:28a couple of days ago and their wife is a jewelry designer and huh I lost my chat history here that's weird um is this still working
879. 2:10:38can you tap the tap the screen hit Hearts if this if you can still hear me because it looks like I lost a bunch of
880. 2:10:45my history here am I still working is this still is this thing on hey Marge Marge is the Tik tock on I can't see is
881. 2:10:54the Tik tock on Marge Marge I'm trying to talk to the people on the Tik Tok is it there anyone Marge um all right thank you for the
882. 2:11:03hearts so I guess I'm still here again one of my great joys in life is doing stupid [ __ ] like that knowing
883. 2:11:11that people are like flicking through Tik Tok going who's this weird person all right I'm gonna I'm going to
884. 2:11:17show you something here I'm going to do an experiment live so I will likely humiliate myself here with my dramatic
885. 2:11:24incompetence but let's go play with something so since I couldn't talk to I think it was Gabriella my wife wife's name is
886. 2:11:39Gabrielle okay I'm gonna go find a piece of jewelry that I made in mid Journey sorry oh that's going to make you see sick I'll put this off
887. 2:11:50toang go find this thing I [Music] made did I do it in here I think I
888. 2:12:12did this going to be so cool it's GNA be so worth it you're gonna be like oh man I so bitching that's so cool Kyle you're the best man
889. 2:12:22except for like when you suck which is like actually right now you suck pretty bad because like am I really just
890. 2:12:29spending my night watching you scroll through Discord yeah man there it is okay look see I made jewelry see jewelry all right so
891. 2:12:44let's let's grab this one opening browser copy image okay so now I'm going to photoshop
892. 2:12:59beta that was a thing I did last night that was pretty funny let me cancel that okay so there's there's our little
893. 2:13:13image so I'm going to do this I'm going to probably do this very poorly but let me [Music] try I'm going to
894. 2:13:26select the pendant here I'm going to do it pretty rough because this this seems to be pretty forgiving even if you do
895. 2:13:33rough thing so so I'm selecting the pendant what' you do tonight I watched some old guy try to select something in
896. 2:13:42Photoshop how long did that last I don't know four and a half hours something like that why would you do that to
897. 2:13:47yourself I don't know I loathe myself shame is my love language
898. 2:14:02right here what come to save the day have you ever been like right here where you're just about to finish the
899. 2:14:14selection and like you slip on the keyboard and let go of [ __ ] and you lose the selection okay so I've
900. 2:14:20got the the pendant selected not very well and I'm going to say select inverse okay so now what I've got
901. 2:14:28selected is everything but the pendant and I didn't even bother with the inside of it here so I'm going to
902. 2:14:32say so so the way um the way generative fill Works in Photoshop so I'm in Photoshop I've got my background
903. 2:14:42selected there's this new contextual box down here and it says generative fill so I click generative fill and now there's
904. 2:14:49a there's a box there that I can type something in so what I'm going to say is um let's see neck um of a fashion model um wearing
905. 2:15:07a I don't know um a v-neck um dress [Music]
906. 2:15:22in Indigo silk we're going to generate it this could be an abject disaster but let's see what it does the generative fill
907. 2:15:35session expired oh no is this gonna am I gonna lose my wait I might lose my selection here what do I have to do wait
908. 2:15:42generate oh there it go no wait please restart Photoshop okay so let me say save this first hang on let me save my selection jeez save
909. 2:15:54selection F Untitled three we save this we're going to call this pendant what' you do tonight I watch
910. 2:16:08some old guy try to restart Photoshop for about an hour how'd that work out for you I don't want to talk about it
911. 2:16:16March can you restart the Tik Tock Marge that dude is funny I this was a a Navy dude from like the 40s I gave him
912. 2:16:27Lakers back basketball shorts that was a generative Phill experiment all right so let's restart Photoshop
913. 2:16:39beta Strawberry Letter all right and then we'll restart that we're going to load our selection that we saved called Untitled wait
914. 2:16:55channel is that it no I thought I saved it did I not save it oh my god really Lo
915. 2:17:10selection Alpha One there it is okay all right we're back so I've got my oh and I've still got people here why have you people not
916. 2:17:21left just watching watching an old guy try to use Photoshop this is awesome uh hope you're enjoying this uh I'll I'll
917. 2:17:29introduce myself to the new folks in a second okay so for Gabriel we're trying to take some jewelry and we're trying to
918. 2:17:36replace this background of plants with a background of like this necklace hanging on someone's neck so I'm going to click
919. 2:17:42generative fill I'm going to type um neck of a model wearing a Indigo and indigo come on English language IND
920. 2:18:02wait Indigo silk dress um at an evening um on the town that's that's a horrible prompt but
921. 2:18:19let's go for it okay so now I'm going to say generate so in theory we're going to get this pendant on someone's skin I didn't say v-neck
922. 2:18:29but whatever it should figure it out so that's sort of it now it like because my selection was bad it it looks
923. 2:18:38like it put the right selection there but it still screwed it up up there but now it'll give us three different versions of
924. 2:18:45this I mean listen for that's not bad either so so for doing a really shitty masking and and just pulling something I mean that at
925. 2:18:58least starts to show you that there's there's a possibility here there's another fun thing that we can do here so let me deselect anything
926. 2:19:08there and now I'm going to um expand here I come to save the
927. 2:19:27day so I'm GNA now expand let's meet her so now what I'm going to do is I'm just G to go I'm going to select the main
928. 2:19:40image wait which one which layer do I need to do this on hang on sometimes sometimes I don't know what
929. 2:19:47I'm doing you know like all the time all right I'm going to select the main image now I'm GNA invert that selection
930. 2:19:57and now I'm just say fil I'm not GNA type in anything let's just see what it does it should give us a face and the rest of her
931. 2:20:08body it should but will it in a world oh that's awesome oh that is terrific all right so I think we're gonna have to
932. 2:20:23we're gonna have to generate again all right so let's say we're gonna add a prompt so let's say um stunning um brunette brunette
933. 2:20:38model um let's see um focus blur background let's see what it
934. 2:20:56does oh here's someone who does jewelry oh they left let's see what this does uh that's still weird that's very weird all right so so
935. 2:21:11here's part of the lesson here is that um photo Photoshop and adobe's models are not as good as mid Journey um
936. 2:21:20I've definitely gotten better uh better examples than that but anyway that that so that at least gives us you know the
937. 2:21:30the beginnings of of how you might approach a project like that so yes it is possible I didn't know it would would be
938. 2:21:39um let me flip my camera around here and if that was horrifically boring boring for most of you I'm sorry but if that was valuable
939. 2:21:53cool um if you're new here my name is Kyle Shannon this is the AI learning lab these URLs behind my head why I have
940. 2:22:00those if if you haven't tried chat GPT by the way thank you for the for the likes I assume that means good on the
941. 2:22:07generative fill lesson with the jewelry thingy by the way that jewelry was just generated in mid Journey it it was my
942. 2:22:14first attempt that just I just did like I don't know silver jewelry and then I put it in the PO mid Journey prompt
943. 2:22:23writer it came up with a couple of jewelry artists and it just like this stuff is mindblowing um why are these URLs here
944. 2:22:33so if you haven't used chat GPT that top URL that's the official chat GPT website go there you owe it to yourself to play
945. 2:22:41with this thing long enough to have this moment it's going to suck for a while you're going to think it's just like this is
946. 2:22:50just like Google but stupid go through that feeling work through it Purge Purge Purge the naughty feelings um and and start playing with
947. 2:23:04it and if if you really want to sort of geek out on how how you talk to that thing go to this URL here prompt. chat
948. 2:23:10this will teach you what prompting is it gives you there's like a free ebook there on prompting it it it it does does
949. 2:23:17a pretty good job of describing that but then the bulk of the document are prompts to tell chat jpt to act like a
950. 2:23:25certain Persona so act like an accountant act like a screenwriter act like a counselor act like a translator
951. 2:23:32whatever um and you'll very very quickly start to get your head around how deep these waters are and how profound these
952. 2:23:40tools are and just start playing with it until your your brain kind of of hemorrhages down one side of your face
953. 2:23:49and then you're like I think this is pretty significant have that moment it's a pretty magical moment any stocks or ETFs
954. 2:23:58for AI companies I think there are some ETFs out there probably would be a bad idea just to grab an ETF that's got like
955. 2:24:05Nvidia and Microsoft let let them figure out the stuff like the companies that I think are interesting right now are
956. 2:24:11anthropic and open AI which I don't think you can invest in Microsoft which you can but they're Microsoft but I
957. 2:24:19think they're kind of kicking ass right now um Nvidia just hit a trillion dollar market cap AMD claims they're going to be
958. 2:24:27coming after um Nvidia from a from a chip manufacturing standpoint I don't know that anyone else can actually um
959. 2:24:35claim that they can do that AMD could be interesting but I don't like I'm not an investor I have a degree in acting so
960. 2:24:41don't listen to me for that [ __ ] all right um great job you've gone live for 90 minutes I'm pretty sure I've been live
961. 2:24:50for longer than that last night you used travel family of four to AA well I know I'm not all that creative if you watch
962. 2:24:58me two nights in a row I use a lot of the same [ __ ] because you know I'm just trying to I'm I'm trying to figure out
963. 2:25:04how Tik Tok [ __ ] works most of the time how do I make those cre creepy AI generated videos of celebrities eating
964. 2:25:12strange things okay go to Runway ml search for Runway ML and then um go to uh you're going to have to go to their
965. 2:25:24Discord because I don't think oh I can't run it here I don't think it's live in their app yet they have an app you can
966. 2:25:30do gen one in the app which is Gen one is you upload a video clip and then a prompt like make it clamation and it
967. 2:25:39will make your clip into clamation that's gen one gen two is text to video so go to Runway ml just search for Runway ml video AI or
968. 2:25:51something like that you'll find the site go to their Discord and then in the Gen 2 channels you you basically type at Gen 2
969. 2:26:01and then you type in a oh you know what let's go play let's go play We're Going live we're going live to the Discord you
970. 2:26:14want to see where the cool Gamers hang out oh yeah they're on Discord I thought they were on Steam shut up get out of my house get out okay
971. 2:26:29so we're in Discord here's Runway ml so there are um once you once you log in there are like areas that you can
972. 2:26:42generate generate one two and three This Is the Gen 2 research BAS data I don't know if there's a weit list for it there
973. 2:26:49might be but I'm pretty sure they they released this to everyone so basically basically you come to the Gen 2
974. 2:26:56area and then you say you want you want creepy celebrity doing something with food okay let's say [Music] um
975. 2:27:10Pete Davidson eating a hot dog on a roller coaster oh no wait sorry that's not how you do
976. 2:27:24that it's it's almost like I've never done this before hang on delete message delete so you go at Gen 2 and then you put in your
977. 2:27:39prompt no wait why why is it not letting me does it not like Pete Davidson let's see uh
978. 2:27:55Kelly CL why is it not letting me go hold at J P
979. 2:28:12davidon eating a hot dog at the beach all right there it goes I guess it was because I I tried to submit it
980. 2:28:24without a prompt actually being there so now so now it's generating so what it does is notice how it's generated a sub
981. 2:28:31thread to that channel so that's my video and then here's the actual things that it put in there it put in my prompt
982. 2:28:40they put in the CFG scale which is basically the creativity scale um it it's got some other [ __ ] upscaling false
983. 2:28:50green screen false okay there there's Pete Davidson at the beach now I said I wanted to meet in a hot dog is it do we have a hot dog
984. 2:28:59here no but it's that's Pete Davidson like that's there you go that's creepy Pete Davidson all right so that's gen
985. 2:29:13two now listen the other reason that you play with this AI stuff is just to crack yourself up that's pretty [ __ ]
986. 2:29:25funny welcome everybody if you're new here my name is Kyle Shannon this is the AI learning lab uh pop any questions you
987. 2:29:32have down below I'm very far behind on them I've been doing this for two hours my voice is going I'm a little moist my
988. 2:29:38hair is still holding up all right that's pretty good the eyebrows look like they needed need a trim I'm looking like some sort sort of bad
989. 2:29:48professor at the eyebrows that no one mentions and he never [Laughter] trims oh if you can avoid aging I would
990. 2:30:04strongly encourage that it'd be better to learn the prompting yourself rather than rely on a plugin I totally agree with that learn
991. 2:30:14this [ __ ] just start play with it just get curious get curious get curious I I would say if they're here here are the
992. 2:30:23the couple of skills that I think are the most important moving forward get curious if you're if you're in a job
993. 2:30:33where you're not using your brain if you're not doing critical thinking start doing that start practicing critical
994. 2:30:39thinking because the people that can use these tools and very quickly assess that's good that's not good like people
995. 2:30:46who can curate and think more like a conductor than a violinist right where you know the conductor is looking at the
996. 2:30:52whole picture that's going to be a really important skill so get curious start exploring um use prompt perfect please okay happy to do
997. 2:31:04that we're gonna flip our camera we're gonna turn off our thing oh there's Pete Davidson popping through there that's
998. 2:31:12pretty funny all right we're going to go to chat GPT you g make money chat GPT okay by the way any videos that say that run
999. 2:31:25for the hills from those things a their voices are annoying and B that's all [ __ ] okay um use prompt perfect so
1000. 2:31:34I'm I'm at gp4 I'm going to go to the plugins thing I'm going to go do this pull down menu I'm going to turn off all
1001. 2:31:42these travel ones I'm going to turn on prompt perfect and so now I just have one plugin enabled it's called prompt perfect and
1002. 2:31:52what it does is it actually rewrites shitty prompts for you so if I type the word perfect as the first word in a
1003. 2:31:58prompt and then type my shitty prompt like um let's see um I have a pizza shop and no one is coming
1004. 2:32:17make me a plan to save my restant horrible prompt not enough context completely generic so it fires off the the
1005. 2:32:35plugin and then behind the scenes it takes my crappy prompt and it rewrites it and it says as the owner of a pizza
1006. 2:32:43shop I am facing the challenge of lack of customer traffic I am seeking expert guidance in developing a comprehensive
1007. 2:32:49plan to revitalize my business and ensure its survival can you provide me with a detailed strategy that will
1008. 2:32:56effectively attract and retain customers while also optimizing the operations and profitability that's really [ __ ] good
1009. 2:33:03right and then it says absolutely I'd be help happy to help you do that evaluate your menu improve the quality customer
1010. 2:33:09service ensure that your staff is trained to provide excellent customer service marketing and promotion invest
1011. 2:33:15in that online presence make sure you have a strong one um delivery and takeout options Community engagement loyalty program am Ambiance and
1012. 2:33:26cleanliness regularly seek feedback so let's say community engagement so I'm going to say I like Community engagement please give me 10
1013. 2:33:44low cost High high impact ideas so it kind of knows in the context of what it's doing so here are 10 host a
1014. 2:33:54local event piz pizza making class tasting event sponsor local teams partner with local businesses charity
1015. 2:34:03fundraisers right so there you go with with relatively little input and a tool like prompt perfect which you can you can learn I mean it's
1016. 2:34:17actually interesting when you watch something like prompt perfect rewrite your prompt like that one just did like
1017. 2:34:23that taught me some things about how to think about how I would phrase something like this in the future right so like I said this stuff is
1018. 2:34:35pretty good now does it give you shitty answers a lot yep does it lie a lot yep does it um fail you on a regular basis yep is
1019. 2:34:49it the most powerful tool ever put in the hands of mankind yep all right how to get started coding python using ai go here and tell it to
1020. 2:35:01start coding in Python tell it the application you want to write have it write write it for you and then go start
1021. 2:35:07playing with the code how about po so po.com here I'll show you Po and I'll I'll do I'll do a little python thing here as well
1022. 2:35:19here we come to save the [Music] day so just just on chat GPT so just using GPT 3.5 I can say write um a
1023. 2:35:34python application um that gives a weather forecast
1024. 2:35:49bang with documentation there's your so so not only can It help you code you can pair program with it basically it it explains
1025. 2:35:59what it's doing and if you don't understand something you just tell it you don't understand it so how I would
1026. 2:36:04Learn Python is I would learn how to PA program with chat GPT so start with chat. open.com and then if you want to
1027. 2:36:12really take it seriously start par program using github's co-pilot which is specifically optimized for writing code all right po so
1028. 2:36:23PO is just like chat GPT except there's a bunch more options so if you look at the screen here it's just the same you
1029. 2:36:33type in let's see if this we'll see if Sage does coding write a python app that um fetches the weather for
1030. 2:36:47forecast sure here's a python app so it does that too so all all of these things are decent at coding um okay so that's
1031. 2:36:57what it is but you've now got a lot of different options here the next three under Sage so Sage is is Po's large
1032. 2:37:06language model and PO is quora right so the qu website that's who's behind it Sage is theirs you got gp4 Claude plus
1033. 2:37:15and CLA in 100K those you have to pay for I pay for those because Claude instant 100K that's a chatbot that's got
1034. 2:37:22a 75,000 word context window so if you want to copy and paste large documents like a complete novel for example and
1035. 2:37:30have it help you edit your novel you can paste the entire novel in there and just have it rewrite sections for you or come
1036. 2:37:37up with outlines or whatever um Claude instant and chat GPT are free and then the other thing that that PO has is it's
1037. 2:37:44got Bots so like I created one here called I am Shakespeare it only responds in I Amic pentameter here's two that um
1038. 2:37:52rewrite your mid-journey prompts for you here's an HTML machine you can explore the Bots and there's like all these
1039. 2:37:59different categories music Bots and mind Bots and learning Bots and travel Bots and history Bots and game Bots and funny
1040. 2:38:06Bots and new Bots whatever and they also have why is this not cre right here so you can make your own
1041. 2:38:21bot so you can upload the picture here which I would make in mid Journey you name it give it a description write a
1042. 2:38:28prompt so act like William Shakespeare and only respond in iambic pentameter and blah blah blah blah blah and then
1043. 2:38:34what's the intro when the bot fires up what's it going to say and then you launch it and you can put that out to
1044. 2:38:40the world so that's po that's why I have that one there because it's there's just some [ __ ] there worth playing around
1045. 2:38:55with why is he singing like a 90s TV show I don't know it's what he does I got my Dr
1046. 2:39:14Pepper you mentioned last week something about going and watch a video about codex and chat GPT what was it code interpreter
1047. 2:39:24here let's go watch it let's go watch one together code interpreter this is one I really want access to and I don't have
1048. 2:39:35access to it I'm cranky about it I'm cranky cranky cranky cranky so I'm going to search by the way if you have not seen seen
1049. 2:39:49the music video called H Ren this do yourself a favor it's one of the most brilliant
1050. 2:40:04pieces of music and performance I've seen in decades all right but we're not here for that um what did I say code interpret chat
1051. 2:40:20GPT so let's see that guy's voice bugs the [ __ ] out of
1052. 2:40:35me I don't know the internet this past week has been going wild we've always been a wild C okay true it is always wild but
1053. 2:40:48especially this week with the announcement the recent announcement of chat GPT or open ai's new plugin that's
1054. 2:40:54released which is code interpreter now this is still enclosed Alpha form so it's not available to everyone just yet
1055. 2:41:01you do need to sign up to have access to it but this plugin specifically is a game changer for not only people who
1056. 2:41:08work in Tech but this individual this user on Twitter here that uploaded a CSV file of every Lighthouse location in the US
1057. 2:41:16it said to chat GPT code interpreter create a gift of a map of the lighthouse locations where the map is very dark but
1058. 2:41:22each Lighthouse twinkles a couple seconds later this appeared where it's a full-blown gift and it's accurate as to
1059. 2:41:28where lighthouses are located in the US now you might be thinking TI they seems so small who cares it created a gift
1060. 2:41:35when you think of this at a larger scale how it is able to really take data and information and output it in a way that
1061. 2:41:40is visualized humans can understand this Twitter user actually put to analyze 300 hours of Spotify favorites playlist with
1062. 2:41:49chat GPT code interpreter GPT showed me how to export from the Spotify API run multi-dimensional PCA and TSN analysis
1063. 2:41:57analysis cat speak today all these words and summarize my taste for me it's wild look at these graph wild it's like
1064. 2:42:06that's like I just I feel like my my biggest um my biggest challenge these days is is come up with new ways to go
1065. 2:42:13that's wild that's amazing that's game Chang changing it's just absolutely in insane Anyway full on you could what you
1066. 2:42:23basically do with code interpreter is in your little chat window there's now an upload button and so you upload a data
1067. 2:42:29set so you upload like a a spreadsheet a CSV file of some pile of data it it essentially instantly reads it and
1068. 2:42:39understands it and then you can just start querying against it tell me what's interesting about that data and it'll
1069. 2:42:44generate a report with graphs and charts and Trends compare this trend to that Trend and put it in this kind of scatter
1070. 2:42:51chart it'll do that so so that's coming that's another another reason to pay the 20 bucks to um chat GPT is that you get
1071. 2:43:00early access to stuff like that code interpreter why use AI instead of Google search it is just a fundamentally different thing and and
1072. 2:43:09it's here's the deal think about it like this how Google makes their money is getting you to click 40 times it's not giving you what you
1073. 2:43:26need how they make their money is you search for something and they give you 40 links and the top 20 of them are paid
1074. 2:43:33for right and so that's why like it it's gotten to the point that Google searches just aren't as valuable as they used to
1075. 2:43:40be because even though it's relevant a lot of that relevance is SEO pack websites that are just [ __ ] websites
1076. 2:43:48to drive traffic to make people money based on clicks that's how they make their money chat jpt and other things like it the the
1077. 2:43:58large language models you put in a query and it returns an answer and and the answer can be quite sophisticated right it can be um come up
1078. 2:44:10with a meal plan um for this specific kind of diet for this many people people two of them have these food allergies
1079. 2:44:19and um one of them doesn't like and it will write you a meal plan like that how long would that take you
1080. 2:44:28to do in Google you'd have to go find a bunch of sites that have nutrition stuff you'd have to figure out which of the recipes
1081. 2:44:35have your stuff and have the this and have the that that and maybe you find a site that makes meal plans but you go to
1082. 2:44:42make your meal plan you input all your [ __ ] and then when you go to make meal plan it's like sign up for1 1995 a month
1083. 2:44:49how many times have we done that you go to a site you think you found something and and it's free up until the point you
1084. 2:44:56actually get your answer and then you have to replace your computer because you just smashed it into a brick wall I've done that so
1085. 2:45:13um you know so so so that's a difference it codes um it's so here's the thing Google Google is about sort of get in get out
1086. 2:45:25it's very transactional here's my prompt give me the [ __ ] I want and I'm out of here chat jpt is much more like hey man
1087. 2:45:33you want to hang out and collaborate on something you got a whisper of an idea I'll turn it into a fullblown blown plan
1088. 2:45:40for you in five minutes you want that sure chy PPT is like smoking crack Google Google searches [Laughter]
1089. 2:45:55ain't I'd rather pay 20 bucks than use the Microsoft ship for free I agree with that um you're so right it makes you
1090. 2:46:04nicer it does it's I feel like chat GPT is training me daily how to be a better human being gen xer I'm getting back
1091. 2:46:13into the workforce paying attention I I'm telling you I think we are we are in the Revenge there there's kind of two
1092. 2:46:19revenges being born right now one is Revenge of the Gen xers because there's something about our generation we have enough we still have
1093. 2:46:30enough energy left to do [ __ ] right but we also have enough life experience to be able to look at what
1094. 2:46:37these tools create and curate it really quickly right we can take our life experience and go that's good that's not
1095. 2:46:43good right so revenge of the Gen xers and I also think this is Revenge of the liberal arts Majors right all the
1096. 2:46:51Specialists all the specialty tactical execution stuff in the world is going to be the stuff that's the most
1097. 2:46:57disrupted the people that are going to really excel in this new AI generative world are the ones that can take a broad
1098. 2:47:04view a horizontal view across things so okay I gotta come up with a marketing plan all right let me go to chat gbt oh
1099. 2:47:13gave me a bunch of [ __ ] oh right I gotta go make copy okay I gotta go do that I gotta oh I got to make images I got okay
1100. 2:47:19I got oh wait there's a strategy how does it tie to the thing so people that can kind of you know assimilate all the different piece Parts
1101. 2:47:30quickly those with a broad educational view or broad life experience view I think are going to be they're going to have
1102. 2:47:39superpowers de going to have superpowers the good guys don't seem to put the time into it that the bad guys do I disagree with
1103. 2:47:47that that's why this channel exists this is a channel for good guys I don't want you to be [ __ ] in here if you're
1104. 2:47:55going to be a troll you got to be creative and I want you doing cool [ __ ] with this technology I deeply believe deeply
1105. 2:48:02believe small things that that provide a positive impact on the world using AI right now are going to have a disproportionately positive
1106. 2:48:14impact on the trajectory of AI I think it's in our hands to drive this in a positive direction I think if we sit back and go well it's the
1107. 2:48:27the the hackers are gonna get to it first it's not perfect right now it hallucinates I think all those people I
1108. 2:48:35think it's very bad don't do that don't be that guy you're not that guy what's the best app for antivirus in
1109. 2:48:44your opinion I don't have a [ __ ] clue I'm on a Mac so for the most part we don't have to deal with that [ __ ] you're
1110. 2:48:52going to be the P wait you are going to be a pioneer of AI just like the internet had Pioneers I was a pioneer of
1111. 2:48:57the internet I started one of the first digital agencies in the mid 90s um I would like to be well so so we're not a
1112. 2:49:05we're not a pioneer of AI we're anyone starting now with this generative AI stuff is it's going to be similar to the
1113. 2:49:13mid 90s where I wasn't a pioneer of the internet I was a pioneer of the worldwide web the internet had been
1114. 2:49:18around for decades the worldwide web said here here's a tool that real people can use this same with chat GPT here
1115. 2:49:25real people can use this so this is the inflection point that we're at before it was all Geeks and researchers after it's
1116. 2:49:33like the rest of us that's the stuff I'm interested in I know you were seeing wait I know you were here seeing
1117. 2:49:41internet coming in 1998 I actually started agency.com in 199 before I was there super early what does that tell
1118. 2:49:48you I don't know what that's in reference to probably a lot what specific things happen to give you that Home Alone moment um
1119. 2:49:59it I had I had written a card game that was about alien disclosure and I kind of gave up on it because I realized in order for it to be
1120. 2:50:10a good card game I actually had to research it so I had like 40 ideas for cards but I needed like a lot and then I needed research so just
1121. 2:50:18my ADD took over and I just abandoned the project so I took that spreadsheet I popped it into chat jpt and I said um
1122. 2:50:26what are all the categories of cards out they came I said give me 10 cards for each category out it came go
1123. 2:50:33research them and give me a description for each thing out it came then I like like within I don't know 30 45 minutes I went from having
1124. 2:50:4345 shittily written unresearched cards to 300 well-categorized fully researched cards and then I even put in a a cell in
1125. 2:50:54each row of the spreadsheet that that did a call to do and generated an image for the card based on the description
1126. 2:51:01within like 30 minutes that was one um writing the python application was another um I created a tool for my company where for for our video
1127. 2:51:12templates we create these things called video gu which are a set of prompting questions that guide an end user through
1128. 2:51:18telling a story how we write them is this really specific kind of way that sort of works in our app and it we just
1129. 2:51:25have figured out how to do it it's really complicated to explain so I basically made a tool that I just gave
1130. 2:51:31it examples of prompts that we've written before and we fill out a Google form and it automatically generates a
1131. 2:51:37set set of prompts for us while we're live in a client meeting we don't show them that it's generating it we're
1132. 2:51:44having a conversation with them we're filling out a Google form on our slack Channel it automatically generates a set
1133. 2:51:50of prompts three five and seven and then we just kind of dynamically pull from those sets and and then say Okay clients
1134. 2:51:58based on the conversation we've been having here's a starter set of questions and then we and then we discuss the
1135. 2:52:04questions with them it saves us two to three hourong creative meetings that that normally like add two to three
1136. 2:52:12weeks to a project um just because of schuling [ __ ] um so that that's one that just blows me away um I don't know
1137. 2:52:19there's just [ __ ] daily like all the stuff I do in mid Journey like when it generates like something [ __ ] magical
1138. 2:52:26and like that zoom out feature that just sort of fills in the details you holy [ __ ] I I mean I at least once a day I'm
1139. 2:52:37like really it's been like that for six months it doesn't stop all right oh man I'm really behind on questions here I'm just going to sort
1140. 2:52:47of scroll down I'm I'm getting crispy so I'm going to go because it's been like two and a half hours so let me let me
1141. 2:52:54answer a few more questions uh what's this why is chat GPT being torn apart by hackers revealing user queries because
1142. 2:53:02hackers are going to hack they put to like chat GPT is a shitty piece of software that has bad security and
1143. 2:53:08hackers hacker gonna hack so I don't know why do why do people steal when people leave cash on the sidewalk I
1144. 2:53:17don't know because some people are [ __ ] I don't know I love chat GPT I use it if not
1145. 2:53:31more than Google I would say I am right now on a percentage basis 8020 chat GPT to Google and and it's Google shrinking I I it's probably
1146. 2:53:42going to be 9010 and when it gets faster at doing the internet stuff and better at doing the internet stuff I I don't I
1147. 2:53:49don't think I'll go back to Google and I at this point I kind of have an allegiance to open AI because
1148. 2:53:57Google's Google bard is so bad right now like I don't know how they're going to catch up prove that you're not a deep
1149. 2:54:03fake did that already how long have you been interested in deep learning for about a year about a year I mean I
1150. 2:54:10listen I've been I've been hearing about Ai and machine learning for [ __ ] decade and then like you know
1151. 2:54:18I'd see like a post on Twitter like oh go use tensorflow to play with machine learning I'm like oh cool and I'd go
1152. 2:54:25look at what tensorflow was and it's like you you literally have to have a PhD to understand the instructions for like how
1153. 2:54:32to do whatever the [ __ ] you do in tensor flow and understand all the categories so um so so the thing that kind
1154. 2:54:43of got got me like really engaged was this nft project called pixel mind which you you basically filled in these Mad
1155. 2:54:52Libs kind of kind of prompts and it would generate images and then you could turn those into nfts and then I met the
1156. 2:54:58founder of that he was really interesting and he was talking about stable diffusion and how powerful it was
1157. 2:55:03and that that was what was sitting underneath that so then I taught myself stable diffusion um and then I taught myself
1158. 2:55:10dream Booth but but stable diffusion just to put it in in context like I failed nine times in a row to get stable
1159. 2:55:19diffusion up and running because I had to run it on Google collab and there's these things called notebooks and they
1160. 2:55:25have code in them and they're sort of easy but then you run them and you follow the instructions and then it just
1161. 2:55:30gives you an error and unless you know what the [ __ ] you're doing like you you're stuck right and so I failed nine
1162. 2:55:37times until I got at working then I got it working and then I learned a bit more and then I got dream Booth working and I
1163. 2:55:43probably failed that five times until I got it working and then that's this image right here is generated by stable
1164. 2:55:50diffusion and dream booth and I created a project called Kyle Shannon dreams so if you go to Instagram K Kyle Shannon
1165. 2:55:58dreams or Tik Tok at Kyle Shannon dreams that's the art project that I did so I was kind of in the middle of doing this
1166. 2:56:05art project with these funky self-portraits and I was writing stories about them and acting behind the stories
1167. 2:56:12and and uh and then GPT came out and I was like holy [ __ ] because this like being able to generate images like that
1168. 2:56:22in a few minutes is mind-blowing enough but when chat GPT came out and I wrote a python application in 90 minutes like a
1169. 2:56:31a a fairly sophisticated one it just it just became clear to me that the world is fundamentally different now we're not going back this
1170. 2:56:41is not a hype cycle I mean it's way over hyped but everything's going to change like I know that I know it deep in my
1171. 2:56:49soul I don't have any question about it um and that's the [ __ ] that excites me like that's what excited me in 1994 with
1172. 2:56:57the worldwide web and I feel like I'm lucky to be on the planet right now I feel like we all are this is an amazing time to be
1173. 2:57:04alive Arthur C Clark any significantly enough advanced technology is indistinguishable from Magic that's what this is that's what that is
1174. 2:57:19technologically it's a prediction engine it's like a it's like an autocomplete for your sentence in Microsoft Word but it's like on
1175. 2:57:28steroids so technically it's relatively straightforward what it does but the but what it actually does is like [ __ ]
1176. 2:57:37magic so are the image things so are the music things so are the video things everything AI won't take your job but
1177. 2:57:44someone using AI will that's exactly right that's from Paul rer from the marketing AI Institute AI is great have you tried some
1178. 2:57:53kfg try a lot of KFC damn dude I'm hearing you can I use AI to to change a sick looking person come alive
1179. 2:58:05again I like to feed it oh interesting yeah I mean you can do all sorts of stuff you can also like I know a good
1180. 2:58:13friend of mine name rich Boyd he he took his father passed away I don't know four or five years ago now and he um he took all of his
1181. 2:58:25father's writings and this was before um gpt3 came out so I think he was using gpt2 he took all of his father's writings and he made a
1182. 2:58:35virtual um like a virtual version of his father where he could ask him what he was reading and have conversations with him
1183. 2:58:44so yeah my face is numb my butt's numb I gotta go all right uh have I tried Google's B powered tools yeah they suck right
1184. 2:58:52now um what are we chatting about AI all right I'm gonna do two more questions two more and then I'm gonna I'm gonna
1185. 2:59:03go let's see pick a good one when do you go live I've been trying to go live once a night so after work I
1186. 2:59:14generally come in here how good is the 100,000 thing to code I don't know I haven't coded with it yet or tried I
1187. 2:59:21would go well you have to pay for it so if you need to do long coding projects it's the you know the 100,000 tokens is a
1188. 2:59:30lot Bing is so crowded it gives me a headache I agree with that Google's page has nothing but a search box just go to
1189. 2:59:38that's why this url's here go there this is the one just go there just go there if you can't afford the 20 bucks a month
1190. 2:59:47just use the free one because the free that is better than almost everything out there and if you pay the 20 bucks
1191. 2:59:52the gp4 that's here with all the goodies and doodads but the gp4 itself is just [ __ ] phenomenal Gen X we love competition but
1192. 3:00:01too old for basketball yeah and too short and too slow and too uncoordinated all right one more question uh
1193. 3:00:23um this ain't live dog listen ped this is live dog dog man this is live you big peed Peter peed peed peed did you peed
1194. 3:00:36in the corner peed sounds complicated all right uh let's see how do you make an image based on a description um go to um go to
1195. 3:00:52bing.com create type in a description and hit return and a image will come out now it will suck because that's Dolly too if
1196. 3:01:01you want to geek out go to Mid Journey um or Leonardo or adobe Firefly but if you just want to make an image based on a description bing.com
1197. 3:01:14create that will get you there listen everyone I appreciate you hanging out with me all night there's there was a
1198. 3:01:19lot of you here last night there's a lot of here tonight I really appreciate it feel free to follow me uh if you want to
1199. 3:01:25know when I do these and um thank you very much for all your time and I hope this helped and this is Kyle Shannon
1200. 3:01:33saying peace out byebye ped this is live peed not to be like a bitter [ __ ] or anything but yeah this is live I'm
1201. 3:01:42totally not chat GPT or am I