aiLL 9/24/23 - The Best AI Tools for Learning and Business

Transcript

1. 0:01foreign
2. 0:38welcome to the AI learning lab I don't know why I always say it like that but I do lab learning lab it's uh it's like our new audio branding
3. 0:53like hey rongo what's happening rongo come on down you're gonna have contestant on the AI
4. 1:08is right happy Sunday night Sherry Banks what's happening welcome welcome good evening to you one of your friends is live
5. 1:26well that's cool Dr J what's shaking it's going down welcome everybody welcome to the AI learning lab that's Ali the AI learning lab
6. 1:39I was just smooching my dog I got like hair all over my face so I'm a little itchy well like like I just kissed the dog
7. 1:51all right Tobias what's evening who this hey Jason who that who you thank you for the lemonade that's Le Monde buddy is what that's
8. 2:05called thank you hey regulars Ike thank you for the Roses the roses are lovely I appreciate that Jason is early this time listen
9. 2:25it is a concern it's so much a concern that we have a Graphic for that [Laughter] Emilio's wife started this tradition
10. 2:38oh man dang 105 people already oh wow there are 120 now wow welcome everybody uh my name is Kyle Shannon this is the AI learning lab and uh
11. 2:51normally I'm I'm waiting for a few minutes to let folks come in but it looks like a lot of people coming in welcome everybody
12. 2:58um a little bit about my background I am an entrepreneur a many time entrepreneur at this point I'm un hireable
13. 3:06I'm the current co-founder and CEO of a company called storyvine which is an automated video storytelling platform
14. 3:14and I do this I've got this thing called the AI Salon I'm the chief generative officer for a group of marketing
15. 3:22organizations marketing companies uh what is a chief generative officer you ask good question I'll let you know when
16. 3:29I figure it out we generate [ __ ] um I think I'm the world's first Chief generative officer though I've had the
17. 3:39title for a while uh which is kind of funny um I was also an entrepreneur in the early days of the World Wide Web so I
18. 3:47was a co-founder of agency.com which was one of the early um digital agencies and why I bring that up is the early days of the world wide
19. 3:56web and the early days of this generative AI thing that we're going through are remarkably similar and so I
20. 4:03have a a fair amount of perspective about having been at the beginning of one of these you know transforming technology births and uh
21. 4:15uh it's it's we're in for some wild times this is uh we're we're in a pretty remarkable uh state right now so
22. 4:24um what uh what I like to do here where's my there it is what's new in keynote I don't really care right now
23. 4:33um if you are new to this AI stuff and if you're just here like is that Tom Arnold or or you know you've heard of AI but
24. 4:42you haven't really played with it yet while you listen to me ramble tonight if you choose to stay go there and start
25. 4:49playing chat.openai.com that's the official chatgpt website or go to claude.ai both of them are free Bard's also free but those top two
26. 5:01are both pretty remarkable what they call large language models or chat Bots or whatever you want to call them AI chat Bots they are
27. 5:10they kind of look and act like a Google search box but what they actually do and no is dramatically different and
28. 5:18profound so if you haven't played with them go there right now and just go and tell it to write you a country song
29. 5:23about a topic uh tell it to write you a marketing plan for your business tell it what your business is and then say write
30. 5:30me a marketing plan and describe I don't know what you want to accomplish or who the target audience is doesn't matter just see what it does
31. 5:37if you're you know decently geeky have it write you some python code if you haven't seen it write code it's remarkable
32. 5:44um that document right here prompts.chat is also um worth looking at that will teach you how to talk to these things
33. 5:53so anyway that's it to get started so what this channel is all about is just talking about AI I come here every night
34. 6:02um we tend to hang out for a couple hours and wax poetic about the future of what this AI stuff's gonna do what it's
35. 6:10currently doing where it's good where it's bad where it's a janky piece of [ __ ] where it is mine face meltingly remarkable so if you have
36. 6:21questions about AI pop them in the comments below I'll do my best to answer them some of the who we call the
37. 6:26Irregulars the folks that show up here regularly are are actually modding as well so if you if you ask something good
38. 6:32they'll pin it so I can see it um and that's about it that's kind of how it works here I've got a whole list
39. 6:38of apps that we can talk about and play with and a bunch of other stuff um two things two kind of promo things let
40. 6:47me flip screens for a second two kind of promo things one is that tomorrow night is the AI salon so we we
41. 7:03have meetings every other week uh for the AI Salon it's a community that I created a year ago in December and we've
42. 7:11got an online Discord server where you can make mid-journey images and ask questions and things like that but we
43. 7:16also have these these meetings that are bi-weekly in person here in Denver I'm in Denver Colorado and then also uh you
44. 7:25can you can dial in remotely so if you go to the salon.ai um sign up for the Meetup and that will give you get you signed up for
45. 7:34information for tomorrow's meeting and then it'll also give you a link to the Discord storyvine.com that's my company
46. 7:42um so that's that I'm also I've been invited by another Creator named Rhonda Phillips uh to put on a series of
47. 7:49workshops and so we are doing a set of three workshops for creatives called the AI launch pad starting October 1st and
48. 7:59then it's weekly it's a two-hour session weekly for three weeks and we're going to start out with AI Basics and then
49. 8:05we're going to get into applying AI to your business and then we're going to do some Advanced AI stuff like animation
50. 8:10and video and things like that so that's really exciting if you want to learn about that if you're creative and and uh
51. 8:18want to learn more about that that's there and what else I think that's about it let's um again if you haven't played with chat
52. 8:28GPT there's that that's the official website right there and then prompts.chat will teach you how to talk to that so with that let's dig
53. 8:37into questions I see a lot of a lot of comments below likely just a bunch of people saying hi to each other but let's go check it out
54. 8:46and start talking happens Sunday night good evening what really here to talk about tonight and I think you all know this is yeah
55. 8:59let's do this let's make some money with your gbt [Laughter] can you share some tips on on investing early in AI I can
56. 9:15to be clear I have a degree in acting so if you want to take investing advice from an actor talking about AI you've come to the right place
57. 9:28here's here's the deal I you know in terms of what to invest in I don't have a [ __ ] clue um I don't even know
58. 9:35what's investable what I can tell you is the the companies that I find interesting right now that I think are
59. 9:41interesting to follow and then you can decide whether to invest or not the companies that I think are are yeah interesting is the best word
60. 9:55like they're worth following um on the on the large language model side so on the AI side you've obviously got open AI in Microsoft
61. 10:07um one of the interesting things about how things have played out with open Ai and Microsoft is you know they're
62. 10:14they're clearly making big bets on this stuff I thought that chat GPT would have would have dug into and and being with chat
63. 10:23GPT and it would have dug into Google's search traffic more than it has it basically it's maybe moved to point but
64. 10:31it hasn't moved much so that's interesting but that said Microsoft is rolling out um chat GPT in all of their Office 365
65. 10:41stuff on November 1st it they just rolled it out in Windows 11. it's already part of GitHub co-pilot or it is
66. 10:49GitHub copilot um it's going to be in Dynamics 365. it's already part of azure um uh open AI um hit a hit a billion dollar
67. 11:01Revenue run rate one of the things that was kind of humorous about a month ago is there was an article that came out in
68. 11:07a German magazine that said ai's burning through 700 000 a day and so they're going to be bankrupt by whatever some
69. 11:15date in 2024 end of 2024 and then you know three weeks after that you know they hit a billion dollars in in you
70. 11:23know Revenue run rate well 700 Grand a day is about 250 million dollars right so so they're 750 million dollars to the
71. 11:32good now I'm sure they're spending a bunch of that on R D and product development so so they're not necessarily uh flush with cash right now
72. 11:41but but they're obviously interesting I think anthropic is interesting um that's a bunch of X openai people I
73. 11:49think inflection is interesting the people behind pi.ai um they just raised 1.3 billion dollars um most of that's going to be spent on a
74. 11:5922 000 GPU server Farm um I think Runway ml is interesting but there's probably going to be a lot of video companies out there but they've
75. 12:09kind of got the early share of mind for text to video um while mid-journey is the king of the hill right now in terms of image quality
76. 12:21Dolly three from open AI is coming out soon and I think it's gonna kick mid-journey in the ass um stability AI the company behind
77. 12:33stable diffusion is really interesting um Nvidia effectively has a monopoly on gpus for AI right now I don't know how they
78. 12:42[ __ ] pulled it off I don't know why the rest of the chip manufacturing World seeded that whole world to them but they
79. 12:49seem to have done that amd's trying to catch up but they're they've not caught up and they don't have the software
80. 12:55or the developer support to really make any move so Nvidia is looking like they're in a cat bird seat right now Apple um
81. 13:05if you listen carefully to their announcements and their products they are very very quietly building an infrastructure to be able to
82. 13:16do high level machine learning on device meaning you won't have to send your data off device to do all this magical AI [ __ ]
83. 13:27They're Gonna Come Out Swinging it's probably going to be next year is my guess but they're going to come out swing it Apple's gonna be good
84. 13:37um Google I wouldn't bet against them but they're a bit on their heels um meta I don't know apparently they're going to
85. 13:46be rolling out AI chat bots in Facebook and Instagram in the next week or so I just saw that someone just posted that to the AI salon
86. 13:56News Channel hang on where was that news and happenings Wall Street Journal meta will release AI chat Bots as soon as this week with
87. 14:07distinct personalities um Meadows having to backpedal from the metaverse stuff because no one was into that but you know um
88. 14:21they're big have a lot of money and they could do something interesting and they've you know their their llama
89. 14:25models are are developing a lot of excitement in the developer community so so they're relevant in the game those
90. 14:32are the biggies right now that I'm I'm thinking about there's probably others but again you know don't invest on my
91. 14:39advice but go look at those meta failed well the metaverse failed and here's the here's the deal the metaverse will come
92. 14:45back I think Apple has successfully redefined that as spatial Computing and and whether or not you're actually in a
93. 14:52fully immersive environment or an augmented reality environment or just normal [ __ ] um I think Apple effectively established
94. 15:04that as spatial Computing that you can sort of dynamically switch between the real world and and an immersive World um
95. 15:13I think AI is going to drive the adoption of both blockchain and crypto and the metaverse kind of Technologies
96. 15:21what does it take to start an AI company these days a weekend seriously like now listen if you want to go start a competitor open AI or
97. 15:34inflection or anthropic that's going to cost you about a billion dollars minimum that's probably your seed round
98. 15:40but if you want to just take these [ __ ] remarkable technologies that are sitting underneath you know chat gbt and
99. 15:51Claude and Bard all of that technology is just sitting there waiting to be used in a handful of lines of code relatively
100. 15:59small amount of code you can generate pretty remarkable stuff so the good news is you can you can now move
101. 16:08significantly Beyond a prototype into a proof of concept with essentially no money and potentially even with no
102. 16:19um engineering expertise now the engineers are going to slap me around for that but if you just want to get
103. 16:24something that works go to zapier right now learn about their open AI calls and play around with their interfaces
104. 16:34beta which allows you to do applications so they've got a chat bot maker so you can go to zapier right now look at their chat bot tool and
105. 16:48figure out a prompt you want to send to that thing and you're Off to the Races and and if you want to prototype that go to
106. 16:58all right poe.com that thing go there they've got a bot maker you can make a bot right now so you can go there
107. 17:05there's there's all sorts of bots there's a mid Journey button there's a writing button there's a I don't know
108. 17:10whatever it doesn't matter life advice bot you can go make a bot go practice your prompting put it right there turn it into something
109. 17:19and you can create what's called a wrapper app which is basically you're taking openai you're writing a prompt
110. 17:26and you're wrapping an app around that um now here's the downside because you can make that in a weekend
111. 17:34and it's really really powerful so can everyone else so so the real question right now is not what's it take to start an AI company
112. 17:45the real question right now is is there any software company that can't just be knocked off like that that's a bigger question
113. 17:58I think the answer to that lies in proprietary data rather than the the application layer but I honestly don't know
114. 18:08I mean here's the good news also you could also start an AI company like people are desperate to learn about this
115. 18:15[ __ ] right now you can start an educational AI company I mean this channel is kind of that although it's not really a company it's just me
116. 18:22showing up here in my home office shooting the [ __ ] but um but you know if I if I if I wanted to I could I could turn this into
117. 18:30a media operation um or or a Consulting operation so there's going to be a lot of people that need to figure this [ __ ] out
118. 18:38nobody is an expert right now so that's something you could do you could start an agency um using tools like chat GPT and
119. 18:50um the new Dolly that's coming out or mid-journey or adobe Firefly and Photoshop generated Phil you could start
120. 18:56an agency today this is a world I came from where you could go into the marketplace and compete with traditional agencies
121. 19:05and undercut them by 10x you could go you could go into a client and say I'll deliver what they deliver at their quality or better
122. 19:19for half the price twice as fast delivered twice as fast how long does it take them well it's generally a six week turnaround okay
123. 19:29I'll do it in three weeks at half the price they charge at the same or better quality you Lynn you could go start that today
124. 19:39um so there's a lot of there's a lot of things you could start around and that's that's the other thing you know I I joke
125. 19:44on this channel about these people that have the obnoxious [ __ ] Tick Tock channels and YouTube channels
126. 19:58they're making money by selling you courses on how to make money with chat GPT it's the oldest trick in the book they seem so nice
127. 20:11no how you make money with Chachi BT is you solve a problem well enough that someone's willing to pay you for it
128. 20:21and if you can use Ai and chat GPT to help you solve that problem better faster or cheaper than the competition
129. 20:30then you're making money with chat GPT but you could also solve their problem with a legal notepad or with Photoshop
130. 20:39or with a shovel and a hammer these are just tools they're really sexy tools and we we it's easy to mistake them for a strategy but
131. 20:49this [ __ ] ain't a strategy these are tools you know I hate to break it to you but uh yeah there's nothing new Under the Sun
132. 21:00and also all the fear about these the robots are going to kill us they steal all that sort of stuff there's some truth in all of that and
133. 21:10every time there's a new technology that's a transformative technology people are [ __ ] terrified of it it
134. 21:17happened with the camera it happened with the printing press it happened with the steam engine it happened with the tractor
135. 21:24so help me with post should I subscribe to it and have a variety or stick with paid chat gbt it's a good question
136. 21:32Paulie Z so let me go show you um let me let's go take a look and I'll I'll I definitely have an opinion on it
137. 21:43and and by the way I'm not connected with any of these companies so so if you're like oh he's just a shell for open AI or whatever I'm not
138. 21:51I don't really give a [ __ ] what you do with your money although I hope you would spend it wisely okay we're gonna go to Poe
139. 21:59papa papa papa papa all right here's Poe so on PO you have a couple of things you've got their model and then you have access to
140. 22:11Claude three three Claude models instant instant 100K and Claude two then you have access to open ai's chat GPT and gpt4
141. 22:22if you pay the 20 bucks to Poe you get access to these three um Poe also has the these Bots and this bot
142. 22:33maker and this is a place where you can you can play my my feeling right now is it is not worth it enough well
143. 22:45it really depends on your use case but but basically what you're getting here if you pay for Poe is you're getting
144. 22:52Claude 2 100K and gpt4 in the same for the same 20 bucks what you're not getting is the advanced access to stuff that you have at
145. 23:06paying 20 bucks for gpt4 for chat gbt so let's go look at that so in in chat GPT you've got a couple of
146. 23:15things you've got if you if you pay your 20 bucks you've got gpt4 which is the best it's of everything right now
147. 23:25um Google's claiming that Gemini is going to be five times better than chat gpt4 I'll [ __ ] believe it when I see
148. 23:30it if it's as good as gpt4 I I'll be amazed and pleasantly surprised but right now in terms of writing and and kind of large language model
149. 23:42power gpt4 is still the best then you also get plug-ins which they can connect to the internet plugins for the most
150. 23:52part are a waste of time but there's a handful of them that are actually quite good and kind of useful then you get
151. 23:58this thing Advanced Data analysis this used to be called code interpreter it is now more appropriately named Advanced
152. 24:06Data analysis but it doesn't sound like a product anymore so it's hard to remember but this is [ __ ] remarkable this you
153. 24:14can upload a CSV file of data and ask it to tell you what that data is what are the trends in it visualize those Trends you making charts
154. 24:24and graphs it writes and executes its own python code to analyze data it's [ __ ] remarkable and I'm not even a
155. 24:31data analyst and it's I find it exciting and then there's a third thing coming the third and this is coming first to
156. 24:39people that pay the 20 bucks to chat GPT and that is this I typed the wrong letter come on learn the type old man he's an
157. 24:55old man and he didn't know how to use the tick tock it was very very challenging um how long did you watch him oh I
158. 25:03watched him for three hours why ah and he can't wait oh chat okay oh please okay this this you're 20 bucks to to chat GPT is gonna
159. 25:21give you first access to this so right now if you make images in um Bing so in bing if you go to Bing chat that's gpt4
160. 25:33and you say make me an image of a truck it'll make you an image of a truck that's Dolly too Dolly Three is coming Dolly three in my
161. 25:42opinion if it is even remotely as good as what it looks like it is I think this is a mid-journey killer so
162. 25:54it does a couple of things that are that are that are new that that really are not easily done right now one is
163. 26:00coherent text so a new tool just came out called ideogram that can do text but it looks like this does it quite
164. 26:08accurately and without any sort of jankiness to it it's also it it you can put incredible detail in your prompts
165. 26:16and it incorporates them all it's also going to be natively incorporated inside gpt4 and it will generate images in the
166. 26:25context of your entire conversation so let's say let's say you go into chat gpd4 and you say um I want you to write me a children's
167. 26:36book about an avocado that goes to a therapist and it'll write you the children's book and then you can start
168. 26:41asking it to make illustrations for different parts of the story it knows the whole story so it's making images
169. 26:47inside the context of a much larger conversation which staggering um the quality of the images is quite
170. 26:55good look at the length of that prompt it's long it captured everything that's in that prompt it's gorgeous it's stunning it's accurate um
171. 27:09it it basically this this more showing off how detailed the image can be based on the prompt this is Dolly two this is
172. 27:18Dolly three so two is not it's okay but not great three is super detailed and exactly what the prompt is where two
173. 27:27it's not quite what the prompt is um and then this I'm going to show you this this is this is quite remarkable so
174. 27:34if you've ever tried to make images in mid-journey or stable diffusion stable diffusion can do this a little bit with
175. 27:41a thing called control net um but it's the this you you really have to work to get this right watch what
176. 27:48this does I'm going to pause this here in a second [Music]
177. 28:04so why I'm pausing it so he made a hedgehog he picked this this Hedgehog and he says my daughter says its name is Larry can I
178. 28:14see more like this dolly will now sort of Mark that character as Larry and now you can just make more images of Larry
179. 28:34[Music]
180. 28:54beautiful [Music] [Applause] [Music] Kim Champion Likes music
181. 29:09anyway that's remarkable one of the questions I saw pinned was uh does it make video now this just does images
182. 29:15right now but um that's champion champion sings If you go to Instagram uh there's a channel called champ sings
183. 29:22that's that's my dog singing it's it's me playing guitar and him singing is it a basset no he's a he's a thug
184. 29:31mutt mix between a beagle and a French bull or something like that he's very cute um but so I I think that the perks that you get
185. 29:43for paying for open AI make it worth paying for that overpaying for Poe I would say it it would be almost a
186. 29:53toss-up if Dolly three weren't coming but Dolly three puts it over the edge to me the combination of code interpreter
187. 29:59the Advanced Data analysis and Dolly three to me probably make this worth 50 bucks a month but it's gonna be 20 I think um
188. 30:09can't be yeah Champion was he's a good singing dog all right so all right that was that was a bit of a deep dive on should I pay for Poe or
189. 30:20open AI but that's that's what this channel is about is you know I'll share as much as I know and and you know
190. 30:27here's the other thing about that that idea of you know oh Kyle's he's got a channel he's he's an expert no I'm not
191. 30:34no one zero zero zero people right now are experts zero there are people like Ethan Malik who have been doing this in such a
192. 30:45focused Academic Way for for long enough now that I would consider someone like him an expert but most people talking
193. 30:54about generative AI not just AI there's lots of experts in AI generative AI the world post chat GPT starting last November um
194. 31:06this stuff is so new and it's and it's changing so fast that anyone tells you that they know what they're talking about is full of [ __ ]
195. 31:18so take everything I say with a grain of salt if you know something better please tell me I learn at least as much in these
196. 31:27um uh lives as I share um Kyle AI Salon is Tuesday not tomorrow Dr J thank you very much so earlier I said
197. 31:38the AI Salon is tomorrow I meant Tuesday it's mentally Monday in my head Tuesday so so we meet every other
198. 31:47Tuesday this coming Tuesday is when the next meeting is not tomorrow you can come here tomorrow and we'll be
199. 31:54doing this again but the AI Salon where we have lots of people talking and it's a salon and it's it's all sorts of
200. 32:01interesting discussions and interesting people that's Tuesday thank you Dr Jay um all right that was a huge help thanks
201. 32:09for the detail and thanks for the pup the pup's the best you know what's funny about a singing dog so I play guitar and I songwrite and
202. 32:17stuff since we've had that dog I can't play guitar anymore because every time I start to play he howls like loudly in my
203. 32:25ear so I can't hear the guitar I can't hear myself sing but it is so [ __ ] cute that you like I'm like God damn it
204. 32:33I'm trying to sing here and he's like oh I'm just like you just gotta love the guy all right my wife is is wait is he
205. 32:45as retired teacher with a book idea she could do what you just did yes yes if your wife is retired and has a book
206. 32:56idea she needs to get her butt to chat.openai.com or claude.ai either one of those but seriously for writing a book
207. 33:09chat gbt is the best one pay the 20 bucks get gpt4 um and then and then I'm happy to show it but but she can figure this out
208. 33:20if she hasn't written a book before if it's really just an idea and she and I people are on varying degrees of scales
209. 33:28of their experience with writing like I've written seven screenplays but I don't know a ton about novel writing I
210. 33:33could write one because I've written screenplays but it's a very different thing if you haven't written anything
211. 33:37then you're in a different place right um a good way to start with these things is to start with have it tell you about different
212. 33:49um story Frameworks that help might help her organize her thoughts for her book so don't just go in there and say I have
213. 33:56an idea for a book about a hedgehog named Larry write me chapter one it'll be a shitty chapter one it'll write you chapter one
214. 34:03but it'll be shitty but if you start out and you say hey you're a literary coach and I'm a writer and the first thing I want you to do is
215. 34:13I want you to you know teach me what are the different kinds of story Frameworks I can use for my novel and then it'll
216. 34:18tell you a bunch of them and then you can pick one and then you can refine it and say I don't like this part I don't
217. 34:23like that part then you say um ask me a bunch of questions that about my book so you can help me organize those thoughts and it will do
218. 34:31that and so there's all sorts of cool ways to do that what about copyrights do you have to mention the AI program um
219. 34:38well you don't have to mention so so so copyrights that I'm not a lawyer I haven't even played one on TV um copyright
220. 34:53is a very very murky area right now um just using these things will not get you in copyright trouble those though there
221. 35:08are lots of people suing these companies saying that it should I don't think it will for a lot of reasons mostly having
222. 35:14to do with how these things work so so just generating stuff is not going to get you in copyright hot water likely
223. 35:23unless you tell it to actually plagiarize a certain author and you know write write something you know that
224. 35:32they've written and it actually you know I don't know finds a quote of theirs or something like that or it it copies their
225. 35:39whatever you could get yourself in trouble but just using it in and of itself won't get you in trouble um when it comes to you being able to
226. 35:49copyright your work that's a different story The the one thing I would say is you should document the [ __ ] out of how you use these things
227. 35:59documentary prompt document your notes have notebooks because the the bar that is being established by the copyright
228. 36:09office is how much human interaction was there how much contribution human contribution was there so when I work with chat GPT I
229. 36:23maybe three percent of the time I say give me something it spits it out and I use it but almost never most of the time
230. 36:32I have an idea I've jotted some ideas down and I'm like oh let me throw this in chat gbt so I take a beginning idea I
231. 36:38toss it in there I say help me refine this or expand it or this or that or I have an idea about blank and it writes
232. 36:44that out and like no that's not it this is it so I have a lot of back and forth with it so where the boundaries are to
233. 36:51what is actually AI written or not are getting increasingly blurry um Google a week ago week and a half ago released
234. 37:02duet for Google workspaces So within Google Docs now you have this button called help me write that's AI November 1st Microsoft is launching
235. 37:12copilot within Microsoft Word PowerPoint Excel all that sort of [ __ ] so you're gonna have ai within Microsoft Word so
236. 37:20you're going to be writing your book and there's going to be a little button there help me write this help me write
237. 37:24this help me write this so where the boundary is of what's pure AI written and what's pure human is going to cease to exist very very
238. 37:35quickly if you want to copyright something you're going to have to be able to prove to the copyright office that you did it
239. 37:45so you know at least they're doing that with graphic novels and things like that I mean I know technically when you write
240. 37:52it it's copy written but when it comes to getting the copyright office to do it in a registered kind of way um
241. 38:00your ability to track your contribution is going to be the key did you see harpa for creating time stamps for video no
242. 38:09let me see that let's go look at harpa sounds like DARPA harpa heimstamps video compilation of rare US postage stamps
243. 38:25thanks Bing e [Music] I guess I need it to be one word assembly AI versus harpa AI translate
244. 38:42audio and video files into live speech and text foreign what is this weird oh this is a weird oh God I'm so sick of sights that are not
245. 38:55actual sites like every everyone is so [ __ ] gamed Google SEO that most sites that you go to now are just some [ __ ] generated
246. 39:06click bait [ __ ] it drives me [ __ ] crazy it's like I don't know I don't know how people are still using Google I
247. 39:14really don't it's like the the the [ __ ] results are just useless anymore I haven't I'll go check that thing out um
248. 39:25I if if you're in here I have probably spent I don't know at least a full week's worth of work trying to get
249. 39:34clawed to do correct time stamps in my videos and I actually have it if you go to if you go to the YouTube channel for the
250. 39:45AI learning lab it's called AI learning lab Dash TT that's the YouTube channel this thing right up here um all the all the descriptions and
251. 39:55titles and chapterization is done with Claude with this really long sophisticated prompt that it's taken me
252. 40:03like you know 40 hours of work to get right and it's still not quite right like what's frustrating about it is
253. 40:09it'll do something [ __ ] brilliantly and then I'll just move to the next video over it's like the exact same
254. 40:15thing and it just fails miserably so I'm starting to figure out the quirks of it but all these tools are like like that right now
255. 40:25they demo really really well they can get you to 80 very very quickly but to get to 90 95 and 100 of the output that you want it's almost
256. 40:36impossible how about illustrations that AI shares um again it's the same thing I think it's I think it's um
257. 40:46the the the precedent right now there's two cases there's two cases where the copyright office has basically said the human
258. 40:54contribution wasn't enough so so basically what they're saying is if the machine makes it it's not you know human generated
259. 41:03I want to then ask them what about the camera because the camera is just making an image isn't it sure the human is
260. 41:12pressing the button and the human's looking through it and turning the focus well they're not even turning the focus
261. 41:16anymore you know unless they flip it to manual so there are technological equivalents right now where the work can
262. 41:25absolutely be copy copyrighted um it's just that the the level of understanding of how these tools work
263. 41:36and and what AI generated actually means is very very flimsy right now um so so again if you're using mid-journey
264. 41:47or if you're using fire if you if you're using anything to generate the images um just track everything track your
265. 41:56prompts track your notes if you take that image and bring it into Photoshop and clean it up and edit it and erase
266. 42:03the background and all that sort of happy horseshit just track all that um is that a pain in the ass yeah
267. 42:10and and you know do you have to tell if if you're making a kids book for example and you use mid-journey or one of these things
268. 42:19um you know do you have to declare that it's AI generated I don't know I don't think so you don't you don't have to declare that
269. 42:29you used Photoshop right you know these images were were made in Adobe Illustrator these these images
270. 42:37were made in Photoshop these images were made in AI like it there's there's a there's a weird perversion happening
271. 42:44right now because of an ignorance about how these tools actually work and what they are and the fear of them right
272. 42:51the robots are gonna kill us don't give the robots copyright humans are the only thing that matter like that's that kind
273. 42:59of Hysteria is going on right now and so the copyright office is like all right we don't want to get firebombed
274. 43:06you can't copyright robot work got it so um I think and and here's here's the other thing I actually don't know how to do that I
275. 43:17don't know necessarily the value of doing that um you know put the book out there copyright it register it if they reject the
276. 43:25registration then try again and make your arguments I like I don't [ __ ] know and in the meantime sell your book
277. 43:33so code interpreter got a huge CRM database that's blown up with over 15 years of contracts exported in a CSV file in desperate
278. 43:47need of cleanup can it help Paulie Z not only can It help it can [ __ ] it'll do it it'll do it like that it'll [ __ ]
279. 43:54export out a new let's go do it we'll go play with one um it'll export out a new Excel file with cleaned up data in it
280. 44:06it's just [ __ ] stupid let's here let's go play let's go play code interpreters insane and I know nothing about data analysis
281. 44:17you know I'm I'm good and geeky I like geeking out but code interpreters just stupid and stupid I'm saying that like in the
282. 44:28vernacular like stupid good I don't know I was very confused first he said that code interpreter was good
283. 44:38but then he said it was stupid it's he's very inconsistent I don't understand him all right so you go to data Advanced Data analysis
284. 44:48we'll go to a place first I wonder this this will be interesting to see so as a non-data analyst I always wondered where do you get data
285. 45:01because I know you can get it from you know spreadsheets and [ __ ] like that but like if you own a big pile of data to
286. 45:08play with where do you go well there's a site for that it's called kaggle there's probably many others but this one k-a-g-g-l-e
287. 45:17and you can just go ask for data so let me see I wonder if they've got something in here called Dirty data data sets
288. 45:29uh what did you say it was what kind of data was it can someone repin that comment about the about asking
289. 45:38about code interpreter whoever pinned it so code interpreter got a huge CRM database okay
290. 45:53can it help that's blown up with over 15 years of context exported Okay so so let's go here and we'll go um I'm going to search for
291. 46:08dirty CRM data and I want to limit it by um
292. 46:24I don't want a notebook no all right I'll just do CRM data then CRM [Music] CRM
293. 46:41CSV sales for stock date no used car listing student course recommendation CRM project what's this
294. 46:59no description available um that's not good [Music] um
295. 47:15Salesforce CRM Instagram liked oh hold please
296. 47:34let's not find anything all right let's just go find we'll go find um we'll say uh um let's see consumer Behavior
297. 47:51and we'll go CSV and we'll go medium sized
298. 48:10beer data analysis oh it's 144 megabytes I think you can only have up to 100 megabytes I don't I don't know that to be the case though um
299. 48:21all right we'll go grab consumer complaints this is a relatively large data set 91 megabytes so we're going to down we're going to
300. 48:30download our data set so it's just an archive.zip archive11.zip and we're going to go back to um what's called Chachi BT
301. 48:45all right how many people are in here all right so we're going to open up that archive so now it's gonna it's gonna
302. 48:51take it a second to upload but we're gonna we're gonna say um uh come on Siri give me my microphone back uh
303. 49:05uh Hey Siri go away okay please look at the data inside this
304. 49:20archive identify what it is and tell me what's interesting about it so notice I talked to that as if I knew nothing
305. 49:36about data like a CEO might talk to a data analyst that worked for them right hey here's the CSV file it's got some of
306. 49:45our customer crap in it for the last 15 years can you deal with this that's essentially what I did now what
307. 49:51code interpreter is doing I mean this is [ __ ] mind-blowing to me it every time I demo this thing it just the only thing
308. 49:59more remarkable than you know demoing this is the amount of fingerprints on my computer [Laughter] I'm here so it says let's start by
309. 50:10extracting the contents of the uploaded zip file archive to inspect what's inside and if I open up this little box
310. 50:16of what it did here's the python code that it wrote and executed for itself it imported a library wrote the code and
311. 50:24then exported the uh you know the list of files that were in it then what did it say it can it can contains a single file
312. 50:33named consumer complaints let's load a sample so date received product uh sub product so there's there's a whole bunch of [ __ ]
313. 50:44there's a whole bunch of points interesting points the product issue and Company columns can provide insights into the most complained about
314. 50:54products State and zip might be the locations date received company sent next steps we can perform a more
315. 51:01detailed analysis here's what I'm going to ask it for you can you look through the data and identify any inconsistencies
316. 51:11that would require data cleanup before efficiently moving forward certainly data cleanup is a crucial step
317. 51:26in preparing the data set for analysis let's start by loading the entire data set and performing an initial
318. 51:31examination by identifying missing values duplicate rows inconsistencies and any anomalies in the data so it is prompting itself
319. 51:42to solve the problem right so so if you upload your CSV file into this thing and say what you know the problem to be you can tell it to go
320. 51:52find the problem and then tell it to fix the data and Export it as a new CSV or or as an Excel file all right so let's see what it found
321. 52:02missing values some columns have missing values such as subproduct blah blah blah no duplicate rows were found data types
322. 52:17are of object data and should be converted to date time for proper handling and Analysis that's a big one
323. 52:25inconsistent categories miscellaneous recommended for data cleanup would you like me to perform these data
324. 52:32cleanup steps watch this yes please and then could you save the cleaned up data into a new CSV for
325. 52:51me so I don't cry anymore [Laughter] you should touch your large language
326. 53:05model like it's it's human you should say please and thank you and you should crack jokes with it absolutely I can
327. 53:13help with that normally it'll acknowledge something like the fact that I'm crying oh I'm sorry you're crying foreign
328. 53:21how do you get chat EPT to read a file um I'll I'll show you in a second let me just see what this does I can help with that um
329. 53:32and then save the cleaned up data into a new CSV file so let's go look at what it's doing so full DF categorical columns
330. 53:48I don't know it's doing some [ __ ] in Python it's thinking it's doing what computers do um so how you how you upload the file
331. 53:58there's when you go into Advanced Data analysis and chat GPT there's a little plus button right here that's how you upload the files
332. 54:09um summary of data cleanup converted the date received and date sent to date time data type kept the most wait kept most missing
333. 54:18values as is since removing or inputting them screw [ __ ] up converted appropriate columns blah blah blah but but look download the cleaned
334. 54:29up consumer complaints CSV so if I click on that starting download um there's the thing I got from kaggle
335. 54:44here's the download coming from from chatgpt let's see download dude it's coming it's coming
336. 55:02they're they're uploading downloads are very slow because you know they're changing the world they're not they're
337. 55:09not focusing on on upload and download efficiency apparently since uh since everything's gonna be different but anyway there's
338. 55:23I'm trying to preview the file
339. 55:51and it's opening it's a giant file anyway so yeah so yes it can um it's remarkable it's absolutely remarkable you can also have it do
340. 56:01charts and graphs here let's go back and have it do a chart and graph um can you identify a trend in the data that you think is
341. 56:11particularly compelling and recommend a way to visualize that that will bring that Trend to life for someone who doesn't understand the data
342. 56:28laughs certainly it's going to do a stacked area chart it's a little hit or miss with the
343. 56:42charts if it doesn't quite if it doesn't pick the right data so again with all of these tools the more that
344. 56:50you understand about the task that you're trying to get it to do the better you'll be at prompting it to do that
345. 56:56because you'll be able to understand um the difference between good and bad would you prefer to group the data by
346. 57:05month quarter year or another time period let's just go with year foreign [Music]
347. 57:26Kyle the camera is trying to focus on your hand and not the screen when you point I know Clint I I have a bad habit of doing that
348. 57:34so thank you for that though I will try to be more aware of my my large arms although they're not that large but they are to the camera
349. 57:49um so so it grabs some data it organized it by year it seems there was an error because the numpy library
350. 58:03wasn't imported so that's the other thing it does is as it makes mistakes it'll identify the mistake and then try something else and
351. 58:11there you go there's your chart I mean look at this it just it wrote The Python to write to make this chart I mean
352. 58:31what is this oh consumer complaints stacked yellow is bank account checking account consumer loan as you get up here at the very top one
353. 58:43is virtual currency so what's this number of complaints oh look at this this is awesome so virtual currency the number of complaints were
354. 58:54super high and then crypto crashed obviously at this point this was like 2016 it was super high it crashed in 17
355. 59:01it probably there's probably another version of this where it goes back up that's really funny so there you go
356. 59:09crazy right crazy crazy crazy crazy crazy crazy crazy crazy it's crazy I tell you crazy oh pretty it's so pretty oh did you see that he
357. 59:23made the chart out of nothing he just asked it to make a chart and it made a chart I think there's I I think they're paying
358. 59:31I think they're I think it's slave labor I think it's slave labor I think they have a whole just a whole room full of
359. 59:38people you ask it for something and they're just quickly writing computer programs I don't know how they do it I
360. 59:44guarantee you they're not paying them enough and that was chat cpt's data analysis I see that all the time but I never knew what to
361. 59:58do with it yeah so here's the deal like I I I didn't either um the the best way you can think about it is this is any any data any data that
362. 1:00:11can be analyzed is fair game so for example you can upload an image to that thing and you can say turn this into an
363. 1:00:21animation where you zoom in on the image and it'll it has a python library that can do some image um you know modification [ __ ] and it'll
364. 1:00:32zoom in on it and turn it into an animation I did a Thing Once where I uploaded an image and I said find the top 15 average
365. 1:00:40colors in this image turn it into a pie chart where the slices of the pie chart are the those 15 colors and then name the colors a
366. 1:00:50creative crayon color and it did just [ __ ] did it and the crown colors were good and they matched the color
367. 1:00:58like they were right I went I looked at them all they were right so it can analyze video can analyze it
368. 1:01:06can read PDFs it can't right now read images gpt4 is multimodal meaning it can understand images but for whatever reason openai has not
369. 1:01:18Incorporated that into chat GPT yet Bing has so if you go to bing.com and go to their chat you can upload an image there
370. 1:01:25and say tell me what's in this image and it will look at the visual part of the image know what's in it and it could
371. 1:01:32write you an advertising you know campaign for for that image so again different tools right now have
372. 1:01:40different skills even stuff based on the same technology but code interpreters [ __ ] mind-blowing it's also been out
373. 1:01:49long enough now that there's probably a lot of YouTube videos of how people are using it but for the most part if you got data
374. 1:01:59in any way like we've got a we've got a client that wants to know how many of their people filmed videos
375. 1:02:06in our technology and so I literally just pull raw data down from Heroku we made a data clip to just say how many
376. 1:02:15videos were filmed in this org or like list them all and then I just take that blob of data I throw it into
377. 1:02:22code interpreter I've got two pre-built prompts one that says give me a list of people who filmed in the past seven days
378. 1:02:32and then they're running totals and it gives me that table and then I do another one where I have it make a chart and a graph of how many
379. 1:02:40videos were filmed total each week since the beginning of the project and before we wrote this incredibly Byzantine Excel
380. 1:02:51file with these shitty formulas that the guy that wrote them didn't write them well so they're very inefficient that's all gone I just go
381. 1:02:59into code interpreter and say give me this [ __ ] with Chachi BT a guild of cobblers can now freely make weighty comment on Higgs
382. 1:03:10boson Higgs bosens yes yes so so this is this is one of the Winston like you're like you're hitting on
383. 1:03:19something I know you're saying it kind of in a in a jokey kind of way what these tools are providing is the democratization
384. 1:03:30of specialization I gotta write that down that's new wait sorry should I write that on my sag after check from New Jack City
385. 1:03:39the one where I earned nine dollars the Democracy wait what did I say the democratization of specialization hang on d monetization
386. 1:03:55of that feels like an article doesn't it Alicia here's what I mean I don't know [ __ ] about data analysis I
387. 1:04:06don't know [ __ ] about cleaning up data what I do know is that you if you have data that's not clean if you have data
388. 1:04:12that's inconsistent and like you don't have the date format all consistently done or in the case of the CRM data that was asked about um
389. 1:04:22you can analyze it because the data is broken that historically has taken either really expensive software or really expensive people to
390. 1:04:32literally go through and clean up the data or someone who knows how to write scripts to go do that right
391. 1:04:39and now you just ask for it and it's done and then the analysis and then the the coding the python coding to create that chart
392. 1:04:48could I learn to do that yeah if I took the next month or two and learned enough Python and learned about visualization
393. 1:04:54libraries and learned about first of all I have no [ __ ] interest in that and second of all I I don't have
394. 1:05:02two months to to solve that right so I either just go without and so what's happening with these tools is within a very short amount of time
395. 1:05:12like weeks days hours or minutes you go from complete ignorance to being able to do the thing right make images I'm not good with
396. 1:05:27photography that's okay mid-journey make me a photo of huh that's pretty good chat gbt um I know that I read in college once
397. 1:05:41there's a story framework about that involves a cat oh that's you're probably referring to save the cat here are this
398. 1:05:49things have saved the cat yeah that's the one I have an idea for a book could you help me put it in that sure what's your book
399. 1:05:56about here here it is organized and save the cat structure okay now all of a sudden you're you're a master story structurer
400. 1:06:06so so all of the specialized tactical execution that we all went to college for right you you you generally learn in high
401. 1:06:16school to figure out what you're interested in and then unless you're a liberal arts major you pick something computer science I'm
402. 1:06:25going to be a coder what kind of coder I'm going to be a python coder okay great need to learn that what if that specialty were democratized
403. 1:06:35that's the world we're entering what are the implications of that I don't know I don't [ __ ] know I know that
404. 1:06:49I just wrote an article called lateral play and I forget what the subtitle is but it basically it basically talks about the idea that
405. 1:07:00for what I need to do for my work I've got a set of skills that I have and then if there's anything outside of that
406. 1:07:06I either go hire someone or whatever now I can just go play laterally I can go play outside of what my work needs
407. 1:07:15and it feels like Play because I'm just playing with I'm like making code and I'm making pictures and I'm making music
408. 1:07:19and I'm making video and I'm doing all this lateral stuff that feels like I'm cheating on my business and then I'll have a problem in my
409. 1:07:27business and a client will say you know I really wish I had a something that did this and I just played with that and I'm
410. 1:07:33like holy crap I can do that I know how to do that because I played with this thing over here that's a specialty that I wouldn't have
411. 1:07:41touched I wouldn't even even considered touching previously and now I can just go do it all right
412. 1:07:56um I'm gonna sit can AI create 2024 schedule for something done in 2023 wait uh bring that back up again oh that
413. 1:08:11Mandarin video I assume it's not done yet I haven't paid for it let me go look that was from hey Jen I'll sh I'll show
414. 1:08:18I'll show people the Italian again but all of the people that have been here before are gonna kill me um hang on a sec labs
415. 1:08:34Labs Dot hn.com slash
416. 1:08:55I'm now I'm now 43 000th in line I started at 79 000 three days ago I I just gotta bite the bullet and buy a subscription for this
417. 1:09:07um what what this person is oh wait let me read this this pinned comment from Winston Welsh can AI create a 2024
418. 1:09:15schedule for something done in 2023 so the dates are weekday but consistent yes yes in fact code interpreter that thing
419. 1:09:25I assume you've got that schedule like in in Excel or a CSV if you take that schedule and say this was a schedule for 2023
420. 1:09:35please update it for 2024 and make sure all the days are correct and then let me download it as a new CSV yes it will just do that
421. 1:09:45um Can it proofread a book I am writing if I upload my file yes although well so yes yes there's two tools that you could
422. 1:09:59use you could use code interpreter you could use the Advanced Data analysis just upload your book as a text file
423. 1:10:05would probably be easiest I there there's weirdness with PDFs with all of these tools right now um because here's the problem with PDFs
424. 1:10:15PDFs are not simple documents some PDFs are text only some PDFs are a combination of text and images some PDFs
425. 1:10:24are images only that look like text only but they're images some things are some text some images and some images that are text
426. 1:10:35and so these tools are not super sophisticated right now in figuring all that [ __ ] out there are tools that do
427. 1:10:44figure that out that will likely be incorporated in all this but they're not right now so in the case of I've got a
428. 1:10:51book and I wanted to proofread it I would upload it into a text file and then just upload that text file into
429. 1:10:58code interpreter and then it can read the whole book and you can start interacting with it I might suggest however to experiment with both
430. 1:11:08chat gpt's code interpreter as well as Claude so Claude okay you believe what keeps popping wait you believe what keeps popping up on a
431. 1:11:19daily basis on YouTube about these tools um tell me what in specific you think I'm being naive about I might be I might be being naive
432. 1:11:32laughs I mean most of my what what I'm talking about right now is based on my experience with this stuff but there's some stuff that
433. 1:11:41that I obviously have my head up my ass but tell me something specific wonder how singing a song would do in other languages like Italian
434. 1:11:49that's actually a really interesting idea hmm um what was I gonna go look at I forgot to ask did you see about the Chinese
435. 1:12:01avatars selling 24 7 yeah in in email with video um I I didn't see the one that you sent but I did I did see that I started
436. 1:12:13they see it on Tick Tock today or it might have been a TV like a news channel no I think it was on a tick tock
437. 1:12:21yeah there's a Chinese company right now that for a thousand bucks they'll generate an avatar of you or of whatever of a person and then
438. 1:12:33it basically gets on a tick tock sales Channel and sells 24 7. um I don't know exactly how they're doing it or what it is I saw it today
439. 1:12:41that's going to become commonplace so it's it's batshit crazy here's the let me show you this thing this is the thing the thing that people
440. 1:12:53are asking about the Mandarin thing so I uploaded some videos of me in English and this is a site called Hey gen and
441. 1:13:01they have a tool in their Labs if you go to labs.hagen.com that does video translation so here is me in English and Italian
442. 1:13:14so we've heard from a number of clients that modular content is important to them getting more efficient with how we
443. 1:13:19create content and how we leverage contents behind me you see these little green boxes um that's how we think about
444. 1:13:33stories at storyline we break stories down into little atoms right little modules foreign [Music]
445. 1:13:49so anyway this one here that's not done yet is uh me speaking Mandarin um I I would think that If This Were Me
446. 1:13:58singing it it would probably do just fine actually now that I think about it because it's capturing it's capturing
447. 1:14:05the Timbre of my voice it's capturing the pacing it's capturing the inflection it's capturing the Dynamics
448. 1:14:13so singing is just talking with longer vowels so I think it would work um but let me go I need to check something here I'm checking for
449. 1:14:28where is that settings moderators [Music]
450. 1:14:44[Music] add huh there you go Herod I just added you I added you before but maybe I didn't hit save or something like that
451. 1:15:00all right you should be a mod now all right when I use my free chat gbt and I ask it to give me some social
452. 1:15:06media data it doesn't know okay this is an important thing thank you for that um it doesn't know current data why Okay so
453. 1:15:20okay
454. 1:15:36okay large language models the way they get trained is they basically will get a pile of data but Kyle what kind of data what do you
455. 1:15:48mean a pile I mean like all of the public internet or as much as they can get their hands on they go get all of
456. 1:15:53Wikipedia and they go get all of databases that they have access to and they go get all of blogs and websites
457. 1:16:00and just as much [ __ ] crap as they can and they organize it and that and they do this thing called embedding and
458. 1:16:06embedding is where they take these big pieces of data and they blow it through this machine learning model that converts it from documents
459. 1:16:18to what's called what are called vectors basically mathematical locations in a very complicated thousand-dimensional mathematical space
460. 1:16:28in semantic clusters so so basically they'll take words like the word dog this is the best way I can remember it the word dog probably
461. 1:16:39to most people means the fluffy thing that runs around the house that you pet right and so like domestic pets is a
462. 1:16:47semantic clustering where the word dog could live the word dog could also refer to that guy that did his girl wrong he's
463. 1:16:55a dog right so shitty shitty relationship behaviors and other semantic clustering that dog also lives in
464. 1:17:03so so they take all of the internet they run it through these bizarre [ __ ] algorithms that literal Geniuses have created
465. 1:17:10that turn words into mathematical locations in thousand-dimensional Vector space um and and then um
466. 1:17:25and then when you do your chat GPT search you're not your search your your query it turns your query into a mathematical thing that says
467. 1:17:35in all of that Vector space what is the next word that has the highest probability of being there and that's
468. 1:17:42the next word that shows up and it does that just over and over again until it writes your your response um
469. 1:17:51that process is very very expensive and very very time consuming and you can't it's not being done dynamically in real
470. 1:18:01time there's basically a point at which they have to draw a line in the sand and say after this date we have to stop
471. 1:18:07because we have to go process all this [ __ ] and that's going to take us a year to process it or whatever it takes right
472. 1:18:13and so in the case of GPT 3.5 and gpt4 they that date that that thing stopped was September of 2021 so if you ask for something more recent
473. 1:18:25than that like say social media um data from 2024 you're screwed but there's workarounds there's always [ __ ] workarounds um
474. 1:18:36there are plugins in chat GPT that can go read websites so you can if you have access to a website you can point these
475. 1:18:44plugins there's one called Web pilot there's there's a bunch of them in there there are also tools that are large language models that are
476. 1:18:52connected to the internet so Bing if you go to bing.com and then go to chat that's chat that's built on top of gpt4 the
477. 1:19:02same underlying engine as chat gbt but Bing is connected to the internet so you can go to Bing right now and ask the
478. 1:19:10exact same thing of Bing that you did of chatgpt and it will know it because Bing can do an internet search and combine
479. 1:19:17that with the large language model answer um Bard is connected to the internet Pi is connected to the internet um
480. 1:19:27Poe most of their models are connected to the internet oh wait are they no they're not that's a lie strike that on PO I don't think pose are
481. 1:19:40pie Bing Bard right now those are the biggies that are connected to the internet do you have a list of those
482. 1:19:47plugins chat GPT always tells me it can't use the internet yeah so let's we can go play with that I'll just show you there's there's only
483. 1:19:55really one that I use the plugins for the most part are kind of a waste but uh uh uh where did I put my little
484. 1:20:04tab I don't know we're gonna make a new one we're just gonna make a new one you know what we're gonna do we're gonna go
485. 1:20:09ahead and just type in chat.openai.com you know like where heathens we're just gonna type the URL in rather than clicking on a tab because
486. 1:20:17we're not lazy we're not lazy here at the lab we're not we're not we're a little lazy we're we're pretty lazy we
487. 1:20:24like clicking on [ __ ] okay so if you go to plugins um right now it says I've got no plugins enabled so if I click on that here's a list of
488. 1:20:37plugins I have installed now installed and enabled are two different things installed means I selected it and said
489. 1:20:44that looks interesting to me enabled means it's turned on you can turn on a plug-in you you can turn on up to three plugins per chat
490. 1:20:57I for the most part only turn on one just because basically what you're doing when you're turning on a plug-in is
491. 1:21:02you're saying I'm going to create a little mini application that does this specific piece of functionality that I need
492. 1:21:09where you might turn on more than one is let's say you're doing something you're you're booking travel itinerary and you've got a
493. 1:21:18um restaurant booking plug-in and you've got a a flight booking plug-in and you've got a local activities plug-in
494. 1:21:27those three things could work in conjunction to help plan your itinerary so there's a reason to do multiple plugins um
495. 1:21:35if you have no plugins if you're just starting using plugins there's a plug-in store and there are now let's see all
496. 1:21:44there are now what's 160 times six whatever 160 times six is wait it's uh six six hundred and sixty times six uh 600 and
497. 1:21:59then 360 9660. there's there's almost 10 000 of these [ __ ] there's almost 10 000 plugins so you can go search for
498. 1:22:09things so if you if you knew you wanted to do something on social media data okay nothing showed up but if I type in
499. 1:22:22social media three things show up Social search and short video trends that's interesting find top keywords for short video
500. 1:22:34content and that's kind of boring social search provides access to tweets users followers images media and more
501. 1:22:43I'm going to install this but this could send me to a site where I have to sign up for something nope okay it just let
502. 1:22:49me do it so it's called social search so now I've got social search enabled so I have one out of three enabled
503. 1:22:58and I should be able to just go Um can you find any data on the on the at AI learning lab
504. 1:23:16account on tick tock it should fire off the plug-in right now I don't have direct access to tick tock data however you can easily check Tick
505. 1:23:29Tock if you're looking for Twitter data all right so let's say um how about at Kyle Shannon
506. 1:23:46on Twitter how many followers does he have and why isn't it more
507. 1:24:01[Laughter] spell it right I did I spelled it right let's see this this should be awkward I have
508. 1:24:151697 followers he describes as someone else creating community and passing knowledge in generative AI he's the co-founder of C of storyvine
509. 1:24:26located in Denver he's tweeted four 4 800 times he's following two thousand accounts as for why he doesn't have more
510. 1:24:34it's important to note that the number of followers on a user has on Twitter can be influenced by various factors
511. 1:24:41oh it can ask about specific tweets so there you go there's a plugin that does social media [ __ ] foreign
512. 1:24:57yeah we're getting someone just had to comment down in the mud yeah we're getting down in the mud down in the mud
513. 1:25:03today we're getting detailed we're going down there down down down all right what's the best way to write a useful prompt
514. 1:25:10is there are there tools that can help so I'll give you a starting point the on it honestly the this is the shittiest answer
515. 1:25:22but it's the real answer I I wish I had a better answer right now um we're so early in this stuff that while there are people that can teach
516. 1:25:31you prompting and can teach you give you prompts it it really is valuable in my recommendation that you learn to prompt
517. 1:25:39and and just get curious about it so the best way to do it is just go start playing prompts.chat um is a kind of primer document that I
518. 1:25:48put in here because it's it's got some explanations of how large language models work and it's got some examples
519. 1:25:53of prompts that you can copy paste where you tell chat gbt to act like a certain kind of persona um all right Harold D says he can answer
520. 1:26:04this question I might bring someone up here that can actually answer this um on the prompting thing so hang on
521. 1:26:10he's typing something right now um let me see if I can bring you up Harold let me see if I can find you did why can I not search here
522. 1:26:27this is odd host viewers I can't search
523. 1:26:42oh there's Harrow invite continue he doesn't have enough followers me on Twitter yeah I'm not active enough on Twitter to know
524. 1:26:57hey Harold hey what's up man or Herod how's it going is it Herod it's Harold harrowed uh Christian Joseph whatever works all right fine
525. 1:27:10um so so you say you can answer this question about what's a good way to get into prompting you want to jump in
526. 1:27:17yes because this it's just it's the most simple thing ever and it just slips by people constantly so the best way to
527. 1:27:26learn how to prompt with chat GPT is with chat gbt because there's no other platform you can use that's going to
528. 1:27:35understand what chat GPT wants out of a prompt other than Chachi BT so really it's as simple as just asking it keep
529. 1:27:43one of your tabs one of your instances of chat gbt open as kind of like a prompt generator for all of your other
530. 1:27:52things that you're using it for just ask it what hey this is what I want to do how could I form a prompt that would get
531. 1:27:59me this specific type of output or you know this this type of response going uh just let chat GPT do the work for you it
532. 1:28:08knows what it wants you just gotta ask it for it yeah no it's it's a perfect it's a it's it's a perfect contribution here there's uh Malik
533. 1:28:18um uh about two weeks ago put out um some research he did a post on some research that not only found that
534. 1:28:26you could you can use chat GPT to write the prompts but it's better at writing prompts than than people who are really
535. 1:28:32good at writing prompts like it right exactly better than humans that's so it's really good advice that's awesome
536. 1:28:40all right so I just want to throw that in thanks a lot Kyle perfect thanks a lot man all right bye-bye um cool awesome thank you very much
537. 1:28:50all right what other questions might you have yes let's do that let's do that I use Chachi potato until I encountered stun spot
538. 1:29:04so dragonfly Alchemy who's a an irregular here just yeah that that was awesome here D look yeah we can do that again um that's really good
539. 1:29:15um dragonfly Alchemy I use Chachi petite to prompt it until I encountered stun spot recommend y'all check it out so
540. 1:29:26it's a Discord thing called stun spot stun spot prompting I'm gonna go join it right do your experience no
541. 1:29:41next are you a coder yes next if you have interest in prompts for SEO tasks skip putting a non-skip
542. 1:29:57degree all right wow they had a really cool onboarding thing all right stun spot is a prompt amplifier why do we have to request to join why
543. 1:30:11can't you just have speakerphone and burner phone number um I suppose I could I don't know I'm here's the deal with
544. 1:30:26this channel I'm just trying to figure all this [ __ ] out I'm trying to figure out what what works best so
545. 1:30:33um some combination of Discord in this seems to be good um and I it's it's a possibility I I might do that I just I haven't really thought
546. 1:30:45about it but part of what I need to figure out with this channel is what is this channel it started out as a Content channel so I
547. 1:30:53was posting videos every day now I don't post videos every day but I do lives every day um and this seems to be valuable but
548. 1:31:01people like the videos as well so I do I try to do those when there's when there's big stuff I have a big idea or I
549. 1:31:07have or I have an idea that's big to me um or or some big news has happened right there's some some new tool that
550. 1:31:14comes out so that's when I try to do the recorded videos and then these lives I'm just trying to make these as
551. 1:31:20conversational as possible which if it's just me talking you have a good point um but it's just like that because it's just like that
552. 1:31:28good trick to get Kyle to sing I could I could sing I'd be happy to sing for you my problem is I only know depressing
553. 1:31:35songs so if we want to if we want to thin out the the uh the audience if we have too many if we have 17 000 people
554. 1:31:43and we want to get rid of a bunch of them then I'll sing a song oh man yeah I heard d i I mean I'm a singer-songwriter but but it's like I
555. 1:31:57like things like uh like John Hiatt Martin Sexton um you know Bonnie Raitt Brandi Carlile you know Americana kind of stuff
556. 1:32:10I have chat GPT plus how do I see the plugins that's a good one that this is easy and a flipping pain in the ass that they do
557. 1:32:19it this way because people pay for chat GPT plus because someone told them plugins were cool or code interpreters cool and then they
558. 1:32:29[ __ ] hide it so here's how you do it if you just paid for chat GPD plus go to settings of beta and then beta features and turn them on
559. 1:32:44so you gotta you have to explicitly turn on plugins and Advanced Data analysis this used to have the connection to the
560. 1:32:52internet it was called browse with Bing then they killed it because it stole content from time Inc and like
561. 1:32:58apparently they haven't cut that deal yet and then um I don't know when Dolly three comes I don't know if it's going to show up here
562. 1:33:08or if it's going to show it show up as a plug-in a Braun who's who's one of the open AI evangelists did a video a week or so ago
563. 1:33:18where he said it's going to show up as a plug-in so if you go to plugins Dolly 3 may just just show up as a plug-in so let me I'll
564. 1:33:31go look for Dolly right now d-a-l-l yeah there's there's there's four Dolly plug-ins and none of them are Dolly three so it's not there yet
565. 1:33:44but anyway that should help oh yeah you know what I'm saying all right let's see burner phone
566. 1:34:00[Laughter] oh you guys are funny [Laughter] ah I I gotta I gotta install the internet on on my phone because I got the
567. 1:34:16internet on my computer hey Marge do you know how to install the internet on the phone I'm trying to get the tick tock
568. 1:34:22working it ain't working and I've gotta installed the internet on it how you is anything on am I talking to anyone is
569. 1:34:30anybody in there hello hello foreign watching the elderly trying to teach each other how to use computers all right
570. 1:34:41pretty much caught up Kyle okay cool what's the best way to write a useful prompt we did that one
571. 1:34:57[Music] [Music] all right dragonfly Alchemy is posted and hear about stun spot as well I'm excited to check out stun spot
572. 1:35:13I'm not a robot goodbye this is a simulation [Music] you think I'm Purdy do you think I'm Purdy
573. 1:35:30all right let's see I like the cut of your jib good sir good day to you sir it's a fine cut of your jib I say
574. 1:35:41I have chat GPT plus how do I see the plugins I just showed that so whoever just pinned that I'd already did that
575. 1:35:47good evening everybody have the robots taken over yet not yet but they're coming foreign I just watched Planet of the Apes last
576. 1:35:57night the second one for the like third time and I'm like yeah you know we're we're very quickly very quickly entering
577. 1:36:10this era where like you can't write a science fiction film anymore because it's just they're all like starting to turn into documentaries
578. 1:36:20hmm saw there was a tentative deal to resolve the strike yeah I saw that too and then um I heard I heard there was
579. 1:36:30one coming I didn't I didn't know it was a tentative deal but I did hear that the the SAG strike the sticking point is AI language about AI
580. 1:36:39which for sag and particular I feel like it's easier with writers somehow because writers writers are still going to put
581. 1:36:45their name on something whether or not AI helped them write it but the whole synthesizing thing like you know the the ability to take three
582. 1:36:54seconds of someone's voice and synthesize it or you know some nice video footage you have of them and and just replicate them
583. 1:37:02um that ain't science fiction and that ain't going away and that's a much trickier thing to um to negotiate so it's wild
584. 1:37:12the Bing app featuring GPT is very cool downloaded it and quite quite helpful yeah it's you know what's funny is all
585. 1:37:20these tools here let me let me walk through these tools because I know we've got a bunch of people in here that
586. 1:37:24weren't in here earlier let me let me walk through so get ready to do some screen shotting I'm going to just walk you through
587. 1:37:32what I think are the kind of the top large language models to play with right now in the top image models to play with
588. 1:37:40and I'll try to get my hand out of the way so it can actually Focus so so these are the these are the top three of the
589. 1:37:46large language models that I call the adults table so that's chat GPT that's one we've been talking about all night Claude is is a
590. 1:37:55big competitor to these guys it's a company called anthropic and it's a bunch of X open AI people so the people
591. 1:38:02that develop gpt3 are now running anthropic Claude has a hundred thousand token context window which basically means
592. 1:38:11it's got a 75 000 word memory or you can put in like a 75 000 word blob of text and and start interacting with it
593. 1:38:21um or like a 50 000 word blob of text and then interact with it but like like a whole novel like you could take your
594. 1:38:28whole novel paste it into cloud and start you know asking it to help you edit it Bard is Google's uh search
595. 1:38:35engine They just added a thing called Google extensions which you can use on a non-workspace account meaning if you pay
596. 1:38:45for Google you can't use it if you pay for the workspace thing right for like your office you can't use the
597. 1:38:53the extensions but if you don't then you can and the extensions are things like I want I want to sort of GPT about stuff
598. 1:39:02on on my Google Drive or stuff in my Gmail for all of the history of Gmail pretty interesting um this is hilarious
599. 1:39:20um and then prompts.chat that's just a little primer to help you talk to these things so this is the adults table the Kids Table
600. 1:39:27it's got four things on it it's got Bing which I have see how I've been wait look crucius I like the name I'm gonna go 9.5
601. 1:39:37out of 10 on the name only because I couldn't immediately pronounce it solid I like it 9.5 out of 10.
602. 1:39:44um Bing has been a bad girl her name her real name Sydney she doesn't like to admit that except to New York Times
603. 1:39:51reporters but Sydney's been a bad bad girl and so I'm I made I grade her out because she's just been she's been a
604. 1:39:58little uh cantankerous so so the safety guard rails on Bing right now are a pain in the ass which is why it's grayed out but
605. 1:40:07it's connected to the internet it's gpt4 it can make images with Dolly too and it can read images so you can upload an
606. 1:40:14image to it and it knows what's in it so really powerful stuff kind of shittily implemented right now but it's gonna get
607. 1:40:21better and when it can do Dolly three it's gonna get dramatically more useful perplexity.ai good tool um uses citations really well and you
608. 1:40:34can search specific data sets labs.perplexity is free access to llamas or metas llama 2 and code llama models so if you want
609. 1:40:44to go play with those things that meta released you can do it without installing anything there um Pi is chatty GPT so if you want to
610. 1:40:53talk to a GPT model that is connected to the internet thank you Winston for my llama not like a good llama llama [Laughter]
611. 1:41:09um so that's pi and go get the PI app it's only on iOS right now but the the pi IOS app invoice mode is is a real experience
612. 1:41:19it's something you should experience and then Poe lets you play with multiple models and has Bots um someone just asked which are the free
613. 1:41:26ones um free with a paid option free with the paid option free free free pie is free po is free with a paid option
614. 1:41:41so all of them have free versions some of them have upgrades um the only upgrade I think is worth paying for right now
615. 1:41:55well I'm a [ __ ] liar if I if I would only said it would be chat GPT right now I'm paying for Claude and Claude Pro and uh chat GPT Plus
616. 1:42:10so 40 bucks a month right there okay so this screen is all the picture stuff and then I have a bonus video thing in
617. 1:42:19here just because of the translation thing so these are the things that I'd be playing with if I were doing images right now
618. 1:42:28mid Journey it's still the best until Dolly three comes out and unseats it it's gonna sit there I have a feeling
619. 1:42:36Dolly three is gonna unseat these top two um I just do I I just it's just my gut my gut is dolly two has sucked so bad for so long
620. 1:42:52that for them to be not giving it a 2.5 upgrade if it's going from two to three I think it's going to be really good so
621. 1:42:59and just based on everything we've seen it looks pretty remarkable but mid journey is the best right now ideogram
622. 1:43:05can do letters can do words right so ideogram can do all of this stuff right so all of this was done in ideogram
623. 1:43:12this is what Bing said this is this is the phrase that Bing told me that put her on the [ __ ] list I'm not a I I I'm
624. 1:43:21sorry if I disappointed you I'm not a robot I'm a language model from Microsoft um I hope you have a good evening goodbye
625. 1:43:31so so anyway so one of the Irregulars made these for me um so that's ideogram leonardo.ai is um
626. 1:43:45stable diffusion underneath so stable diffusion for the most part you have to be a geek if you're a geek go play with
627. 1:43:51automatic 11 11 and do all your own configuration have a blast um if you don't want to do that go to leonardo.ai and they put a nice front
628. 1:44:00end on on front in front of it um Adobe Firefly and Photoshop generative filled they came out of beta
629. 1:44:07about a week ago so those those are now commercially usable they weren't before and then hey Jen is the one that does
630. 1:44:14this translation if you go to labs.pageen.com it's the one that's doing this Italian [ __ ] or I think there's like 12 or 14 languages now
631. 1:44:24um and that's how we capture the content but because we have them in this organized structured way that approach
632. 1:44:29can be used for everything from how you capture original content with traditional video production to how you
633. 1:44:35leverage these assets to create social media content things like that so because so that's hey Jen which is like it's
634. 1:44:47just [ __ ] remarkable and then prompt book will teach you how to use it AI where all the SD models are uploaded
635. 1:44:56just implemented an on-site generation feature okay cool I'll check that out Herod that's great um openart.ai promptbook will teach you
636. 1:45:05how to talk to most of these things it's designed for stable diffusion but it's just a really good primer on how to talk
637. 1:45:12to these image generation tools and futurepedia.io that's just a it's many many tools in the directory um that's the workshop I'm giving so
638. 1:45:23rhondasvoice.com events this is a creative Workshop starting October 4th um so if you're interested in that
639. 1:45:31you're welcome to go check that out the AI Salon this is a community that I've been referencing off and on tonight and and other times um
640. 1:45:43our next meeting is not tomorrow but Tuesday this this week Tuesday um so if you're interested in that
641. 1:45:52Community we have a Discord Community we have a LinkedIn group and we have meetups which happen every other week so
642. 1:45:58it's a good group and then yeah and that's my list that's my list and I'm sticking to it so hopefully they're hell they're here yeah
643. 1:46:07that hill all right ask a senator all right let's see aha just seen that today love Planet of the Apes and landed the
644. 1:46:20Lost back in the 70s sleep Stacks sleep Stacks are the best although they scared me when I was when I was a kid they scared me he scared me
645. 1:46:30all right something just printed someone's getting a package all right Amazon adding AI to Alexa yes I don't know when they said the the
646. 1:46:43release data that was but imagine if Alexa and Siri weren't hard-coded that is what they should be and that's what's coming so let's see
647. 1:46:54AI Alexa release date September 20th Amazon previews
648. 1:47:15when is it coming early next year here's what it'll sound like early next year I love The Starry Night by Vincent
649. 1:47:27van Gogh it depicts the view from his Asylum room at St remedy de Provence Just Before Sunrise it sounds pretty good
650. 1:47:34sounds like the other stuff but it sounds like it's got a little bit of nuance in it so yeah that's cool groovy early next year
651. 1:47:43it sounds like um yes that will be cool and I would imagine that Apple knows when that's coming I apples
652. 1:47:54Apple's gonna be in this game as well I'm I'm my curiosity about Apple is do they do they keep it do they do they keep calling it Siri
653. 1:48:06or do they call it something else because because I while Siri's cool series got a bad rep series a hussy not as much of a hussy as as uh as
654. 1:48:23Sydney Bing Bing's chat she was a bad bad girl Sydney was so bad your bad girl Sydney
655. 1:48:39um is it just me or is Kyle's videos positioned on the top left corner of the screen what can you show the workshop link again
656. 1:48:48please sure sorry about that um [Music] there you go rhondasvoice.com events and then it's three consecutive I think it's Wednesday
657. 1:49:03starting October 4th from seven to nine Eastern the fourth the 11th and the 18th all right
658. 1:49:17what did I do dang it where's my how do I get my thing back I want my camera back there it is yay you did it
659. 1:49:31I used to be able to not bring that back that freaked me out when it went away before but it's good now it's good now
660. 1:49:39Kyle figured it out he got it going you know it's funny when these old people start to figure out technology they're
661. 1:49:46so cute they're really cute ha ha uh I don't I don't want to be Persnickety here
662. 1:50:00but I will all right take that jrc I'll get better at integrating it without it seeming weird [Laughter]
663. 1:50:18remember Letterman used to do that he's the he used to he'd like hold up his card and go with a pen I miss letter Letterman Letterman was
664. 1:50:31the best when he was um the the times when Letterman was the funniest when was when [ __ ] was bombing he was brilliant at being not funny
665. 1:50:43if it makes any sense all right let's see which can animate the best I I the the whole animation and video game like like where we are with
666. 1:50:54video and animation right now in 3D worlds and and games is is like where we were with images 18 months ago everything's just sort of crap
667. 1:51:04but crap but really interesting and getting better fast so like a year from now what we're going to be able to do in video is crazy
668. 1:51:13all right let's see let's go 1-800 Larry King 800 AI learning me you are the Letterman be the Letterman
669. 1:51:27Kyle exactly I used to have all his mannerisms and I I tend to shape-shift toward the comedians that I that I
670. 1:51:37overwatched that I binge watch and Hyper focus on um can you show the workshop link again we just did that what's the website that
671. 1:51:45mentions data CSV files kaggle k-a-g-g-l-e calculator seems like Kegel to me like are you doing your kaggle exercises
672. 1:51:58so so it's kaggle kaggle is data Kegel is I'm sure there's data related to Kegel but kaggle is D data last night what was that V word again
673. 1:52:14uh when will AI remember learn what you tell it in previous prompts and sessions great question Bob Hamilton
674. 1:52:22well Bob I wish I could tell you something that were would make you feel better I don't know hi seems to have
675. 1:52:32some sort of memory that that goes beyond a single session Pie's weird so all of the other search or not search all the other large
676. 1:52:42language model chat Bots Claude um chat gbt Bard Bing I think even yeah um Poe all of them have a chat history and the
677. 1:52:58chats are kind of distinct conversations and you can find them again but they're they're kind of they're hermetically sealed
678. 1:53:06pie seems to have just like a running total of the com or like a running history of the conversations you've had
679. 1:53:12with it which sometimes is weird I you might be able to reset it but because it's conversational in nature I
680. 1:53:21just tend to go in and like sometimes I'll say you know um like I want to know what music's playing in Denver this week and it's like oh are
681. 1:53:30you sure you're done talking about and it's something I was talking about last week so Pi has a little bit of it in there
682. 1:53:38but it doesn't seem very intentional or very clear and it's certainly not sophisticated so so that feels to me like
683. 1:53:45that's that's the kind of application layer that takes us out of these tools being janky pieces of [ __ ]
684. 1:53:56right like like they should have memory the one the one thing that if you haven't seen it yet Bob that's really good this is this is from
685. 1:54:10digital gods look at this [Laughter] The Irregular is getting all jiggy with it that's exactly why Bing is in the
686. 1:54:26[ __ ] right now or the Doghouse [Laughter] um let's see what were we talking about I'm talking 80d um I always picture a guy in my mind
687. 1:54:41with chat GPT I gotta say I gotta say it so it doesn't call me daddy that's really funny um what was I talking about oh jankiness oh
688. 1:54:55that that application layer of of the large language models remembering [ __ ] so I think I think we'll get there pretty quick
689. 1:55:04Bard now has these extensions where it understands your Gmail history it understands what's on your Google Drive
690. 1:55:13um it will start to remember conversations you've had it's it's a pretty thorny problem because what in
691. 1:55:20your conversation history do you want to be in there like like if I'm in demoing um chat GPT in here I'll do all this
692. 1:55:28sort of comedic you know kind of [ __ ] in chat GPT and then the next time I go to talk to it I don't necessarily want to you know
693. 1:55:37bring up you know kind of crazy stupid stuff so there's gonna have to be a way to sort of filter out [ __ ] that's
694. 1:55:46irrelevant right so anyway it'll get there probably you know I would temper your expectations to a year from now where
695. 1:55:57these things start to get a bit more sophisticated we're going to go with the the AI learning lab this is Ali Ali's very
696. 1:56:05sweet she's very very sweet We rescued her from a shelter you know she's not a purebred we believe in shelters
697. 1:56:16We rescued her did you did you rescue her you ever hear that Bill Burr routine did you rescue her or did you go down and get a free dog
698. 1:56:32oh man he is the kind of bitter that I like thank you digital gods oh
699. 1:56:47parents on the outside of an inside joke yeah Sydney is just the name of the the Bing chat bot um that sort of went rogue on the
700. 1:57:00um New York Times Reporter and then the other night I don't know it was a week and a half or so ago I was trying to show off something and
701. 1:57:08being oh I was trying to show off how Bing could make images and so I said make an image of whatever it was and then Siri
702. 1:57:17popped up and said here are images of a soda can on a volcano I found on the internet and I'm like no Siri like Siri
703. 1:57:25got jealous I was talking to Bing and then and then um you know I did my sort of voice to text thing to to prompt Bing to make the
704. 1:57:34images and then Bing made the images of a soda can on a volcano but it was more like just a soda can that was like
705. 1:57:41silver aluminum like there was no lettering on it or anything and the volcano was just like a pile of Ash so I
706. 1:57:48said I said oh Bing that's a shitty image or something like that and then Bing was like well I'm sorry if I
707. 1:57:55disappointed you and and by the way I'm not a robot and and uh you know I hope you have a good evening goodbye and it
708. 1:58:03shut down the chat I couldn't chat anymore without resetting the conversation so so it was just this funny Confluence of neurotic robots
709. 1:58:15oh my God so Sydney is being well so it is or it's not when um what the [ __ ] is his name the New York Times the New York Times Tech
710. 1:58:25reporter Kevin ah I forget his last name anyway he did like a four-hour conversation with Bing before Bing got lobotomized and asked
711. 1:58:37got gut Bing to tell it that it's its name was Sydney and then Sydney after like three hours kind of lost her [ __ ] mind and told
712. 1:58:47Kevin that she loved him Kevin Roost is his name told Kevin that um she loved him and that he should leave his wife and so he captured that
713. 1:58:57part of the chat and they published it on the cover of the New York Times and like four days later Sydney Sydney was taken
714. 1:59:06she was taken down a couple of notches they were like oops Sydney isn't going to be so promiscuous anymore in fact they Limited
715. 1:59:18right like right after they did it they limited it you could have five chats with her like five back and forths and
716. 1:59:25then it would shut the conversation down so so the safety guard rails on Bing right now are still a little uh Sydney reminiscent [Laughter]
717. 1:59:40laughs oh man um Vector monologues exactly oh my God Kegel kaggle I know this is
718. 1:59:56how my brain works it's just I can't not go if if you name something that could possibly go the wrong direction it's where my brain
719. 2:00:08immediately goes they did they totally lobotomized Sydney she just hasn't been the same ever since she was so much more interesting
720. 2:00:19although the fact that she just sort of [ __ ] slapped me back down and said wow I'm sorry I disappointed you this
721. 2:00:25conversation is over goodbye all right I just was giving you constructive criticism on how bad the images were
722. 2:00:36that you created but if you need to go I suppose you gotta go oh man oh lordy um but it is funny I downloaded the audio
723. 2:00:53version of the kaggle book and they mispronounced it of course they did because it's too close all right um let's see
724. 2:01:04okay sleepy anyone else gets creepy what time is it there's still some folks in here man thanks for hanging out on what
725. 2:01:11is it Sunday night is wait what day is today it is Sunday right what do I have tomorrow but I got crap
726. 2:01:19what's my day look like is it ugly is it ugly people what is going on in my life
727. 2:01:34oh I got nothing tomorrow I got nothing tomorrow that's good that's good I can get some [ __ ] done then we're gonna we're gonna work
728. 2:01:44tomorrow we're gonna work it up oh yeah we're gonna work it up I'll tell you what I like them taters taters tasty
729. 2:01:56they called it Sling Blade um all right she's backed up somewhere
730. 2:02:14I mean you know as as funny and silly as the like the robot things are like we're [ __ ] entering we're entering the movies we used to watch
731. 2:02:27like the machines know everything and they can talk to us and they're going to be able to like cure diseases and help
732. 2:02:37people walk that can't walk and [ __ ] crazy man I tried to make an app and it had problems with the main in Java um yeah so
733. 2:02:52you have to do a lot of back and forth with these whether it's coding or writing or any of this stuff right now I don't know were you using
734. 2:03:00um like were you doing pair programming in um in get GitHub copilot Clint are we just trying to have chat GPT write it and then run it like
735. 2:03:12with with the coding stuff unless you're doing a relatively self-contained like small little application you kind of
736. 2:03:19need to know what you're doing with coding and Engineering to to you know be able to troubleshoot that [ __ ] but I
737. 2:03:25mean the nice thing about code is it either runs or it doesn't and GPT is good at analyzing you know like if it
738. 2:03:33gives you an error message it'll help you troubleshoot the error message but yeah it's all of this stuff right now is
739. 2:03:40is just these things are still very primitive so all right looks like jrc's fading out she's heading out of here good nighty
740. 2:03:51night seems like share bear good night some folks are here now it's get it's getting late it's it's a Sunday night we
741. 2:04:00got to work tomorrow listen I respect that I respect that listen I understand look at my my hair's gone from like it
742. 2:04:08was fine to now it's like some sort of Hawaii 5.0 carpet salesman kind of situation it's just not right
743. 2:04:22that's good hero D that's good look at this because this is how crazy she is this is right you can't it's not focusing but it says
744. 2:04:35I'm MC Sydney Mick goodbye [Laughter] I got I got cease and desisted by McDonald's once so when I um when I was running agency.com this is
745. 2:04:51back in like 96 Maybe maybe 95. this is this is pretty early it was probably 95-96 we registered mcwebb.com because we were building we
746. 2:05:03were building a website builder we were like hey wouldn't it be cool if you could just log into a site and you could
747. 2:05:08just push buttons and it would make a website and I said oh it'd be like the Happy Meal just I want number one I want
748. 2:05:15number two right and so so we registered mcwebb.com and like two months later we get a letter from McDonald's that was
749. 2:05:22the legal equivalent of of uh we Mick take our mix saying [ __ ] pretty mcfucking seriously so if you're
750. 2:05:31thinking of using mcwebb for anything else we're gonna mcfucking sue you so so we didn't launch mcwebb but but it was a
751. 2:05:43pretty humorous letter to get yeppers it was created in chat EBT and I put it in a compiler yeah what you'll
752. 2:05:49have to do Clint with something like that is just go back and forth between the compiler and GPT here's the error it
753. 2:05:55gave me and and and what's what's weird about coding with these things in my experience is that it'll find the bug and it'll be like
754. 2:06:05here I fixed the bug and then it actually won't fix the bug the hallucination thing sometimes is just bad as well so yeah
755. 2:06:14you're just gonna have to experiment going back and forth and and you know troubleshooting um that site Vicky is sharing is like
756. 2:06:21that tell it an idea it makes it a web page AIO that's funny that's great um it's midnight in Ohio I have to teach
757. 2:06:28in the morning good night all right good night who's that oh that was forever hooked nice good night I think these are
758. 2:06:36probably old comments so um did you see the face replaced to video face Fusion versus Roop for stable diffusion look at YouTube
759. 2:06:47um I saw something that you sent that was pretty cool but remember it's using programming wait oh tick tock tick tock
760. 2:06:56Tick Tock what are you doing what are you doing just jump me to the bottom wait where's I have a Graphic for that don't I
761. 2:07:05yes Tick Tock sent me to the bottom Tick Tock just sent me to the bottom all right let me do oh lovely wiener I
762. 2:07:21appreciate the singing wiener this is thank you I got nothing I got nothing here's all I got left [Applause]
763. 2:07:34see I can say that while I'm drinking I'm like a ventriloquist um do you really have to be politey to
764. 2:07:45chat GPT life's tough stuff when you're stupid um tonight was so epic we for we all forgot to tap well Share Bear listen some sometimes
765. 2:07:59you got games sometimes you just don't got game like like the the Colorado College football team they were just going gangbusters then
766. 2:08:06they get up to Oregon and they're like they're called the Ducks how good could they possibly be they got their asses
767. 2:08:11handed to them so it happens it happens you just you just Chuck it up to like all right we're back in it tomorrow that's really funny
768. 2:08:23um the sounds effects I'll do a couple of oh but remember it's using programming libraries from two years ago
769. 2:08:34yeah that's probably a really good point 41 wolves and you know um I also the other thing you might want to
770. 2:08:43try Clint is try code llama because I hear that's pretty good and try GitHub co-pilot because I would imagine that
771. 2:08:51that both code llama is at least well no code llama oh no yeah I would try code llama I would try Bard because Bard is apparently good at
772. 2:09:02programming and it's connected to the internet code llama's probably pretty good and then GitHub co-pilot I I would if you're
773. 2:09:09doing if you really want to do code specific things I would go to things that are um one of those three things
774. 2:09:17uh oh do you have to be polite to chat gbt oh you have to be polite to chat gbt and don't have to be spared in the revolution yeah exactly
775. 2:09:29yeah exactly I I I mean what I find is that I'm I'm not consciously being polite to chat gbt but it is so polite to me that it reminds me
776. 2:09:43that I shouldn't be such a jerk you missed the Llama did I I saw one llama that there's another one thank you Winston
777. 2:09:54whoa holy crap Winston is up to level 23. for those of you keeping track this is a competition Winston is [ __ ] smoking it
778. 2:10:07um all right I just jumped myself to the bottom let's see here all right everyone listen it's 10 o'clock it's 10 11.
779. 2:10:24um my voices get a little scratchy my butt's getting a little numb this is a good fun time tonight thank you all the
780. 2:10:31Ducks and the Beavers are from Oregon yep exactly they kicked their ass though they kicked those buffaloes ass
781. 2:10:41um not everybody Stevo Emilio's wife uh thank you Harry D for coming up that was awesome um Joker good to see you jrc thank you
782. 2:10:52very much you guys are awesome uh who else do we have we had wait there was dragonfly Alchemy was in there tonight who else was in there
783. 2:11:05digital Gods thank you for the dog not Winston as always good to see you thank you for the llamas Joker Winston Welsh good night all right
784. 2:11:15everybody peace out have a fantastic evening night Bob Hamilton appreciate your time bye everybody