aiLL 8/16/23 - The Future Is Here - AI Is Already Changing Lives

Transcript

1. 0:06mm-hmm hey Silver Fox what's happening back in the home office
2. 0:21back in Denver headquarters back home I mean his life was happening I just figured out how to how to actually invite you guys
3. 0:32so all you regulars I should do that so that you're not just relying on tick tock take Doug's a bunch of dum-dums
4. 0:46encourage more viewers to share yeah what she said share share the live we're just getting started so
5. 0:57people are just coming in but if you're willing to share the live pop that out there you'll be happy to know I reset my Apple watch
6. 1:10and it's not full anymore I put in a beta I have a feeling that's what that that's what happened I put in a beta and I think it it
7. 1:19exploded it exploded ah I got up at uh four this morning in New York and wait hang on
8. 1:33the partner just sent me something that I don't think is right oh yeah it is yeah
9. 1:49yeah yeah yeah that's correct that is correct ma'am you are correct all right let's see let's get back here I did some
10. 2:04housekeeping on the old chart the old charity chart cleaned some [ __ ] up we're gonna have that we're gonna have that we're gonna have that
11. 2:14I guess we'll still have that oh yeah we got that oh as as people get in here uh for wait for a second I thought you were
12. 2:22an Indian oh hang on there's it's probably the lighting is changed and this filter is making me look weird
13. 2:31it's pretty apparently making me look weird enhanced filters filters [Applause] many people say there we go is that better
14. 2:46a little a little more better you better you better you bet all right welcome everybody um what you call it so we've been scammed
15. 3:02the AI learning lab is just a humble little Channel are we all we want to do is learn about Ai and the scammers are like no
16. 3:11we're going to take your nice people and we're gonna get money from them we're going to have them invest in our
17. 3:18multi-level marketing scheme so if you see an AI learning lab asking you to join some [ __ ] first of all if I ever do that I'll tell
18. 3:29you here that I'm gonna do it and then we'll it'll be like an agreement so first of all I've never done that and
19. 3:36I won't do that and if I you know if I do I'll tell you and so if someone's asking you check the name if it's got two L's for the word
20. 3:46learning that's not how you spell it so not only are they scammers they're not very smart no but they are because
21. 3:55the A's next to the L and then the L's next to the L and they're all lower case so it all just looks like the same thing
22. 4:02so two L's at the beginning scammer hopefully that's the only one uh have I reported it to tick tock yes have Tick Tock done anything no
23. 4:12does that make me cranky yes do I think Tick Tock are dum-dums yes are they not really doing it by intent
24. 4:22yeah yeah no they aren't and and the other thing I think they accidentally slipped up on was they they accidentally
25. 4:28copied over videos of me talking about AI that was that was probably just a slight oversight as well although Tick Tock sees nothing wrong
26. 4:37with it now how do they don't how they don't I I'm reporting and I'm like these guys are copying me there's pictures of me
27. 4:47and then on both accounts and I'm reporting theirs anyway that's that um let's look at some other stuff so welcome everyone my name is Kyle
28. 5:00Shannon um this is the AI learning lab my uh my remit in life is to hang out and talk about AI stuff lately this is a
29. 5:11self-imposed remit uh I think this stuff's really important my background I'm an entrepreneur I've started 12 or 15 companies in my day
30. 5:20um if you're new to generative AI if you've not played with chat GPT go there right now that top URL chat.openai.com that's chatgbt
31. 5:31and you may hear people on Tick Tock going you can make money you won't hear that from me except in that stupid voice because that's what
32. 5:38they sound like and they're idiots so we'll talk about that but this shit's really important so go play go there and
33. 5:47play and if you don't know what to say there you can go to prompts.chat and it'll tell you what to say there okay so
34. 5:53I I switched things up here if you've been here I this was this list had seven on it before I've knocked it down to
35. 6:01these three with the prompts.chat thing this is like if you're just getting started the reason I have the other two chat GPT here
36. 6:10if you pay the 20 bucks a month you get gpt4 you get code interpreter and you get plug-ins I pay the 20 bucks I think it's worth it
37. 6:19but I'm not getting paid by them so I don't give a [ __ ] if you do it or not these two Bing is connected to the
38. 6:25internet it can also make images for free thanks dragonfly I appreciate that and then Claude the one over there
39. 6:31that's a competitor to these two from ex-employees of this first one a bunch of ex-openai people started anthropic
40. 6:39and their Claude model has a hundred thousand token context window so you could paste in a let's say you wrote The Great American
41. 6:46novel and you just wanted a little editing pass done on it you could paste the whole thing into the Claude prompt
42. 6:52and start interacting with it it's amazing so so the first one is if you haven't done anything just go there
43. 7:00if you want to sort of play with other things like connected to the internet or images or long long chat contexts as they call them go there
44. 7:10now there's also some other opportunities feel free to uh what do they call it screenshot this um so Bard is Google's entry into this mess
45. 7:22um they are not great but not bad but it is connected to the internet is pretty good at that it can also read images so
46. 7:30you can upload images and have it talk to you about it then you got labs.perplexity.ai that's uh meta's
47. 7:38llama 2 models you can play with pi is from a company called inflection Poe is from quora all of those these are just
48. 7:46interesting other things to experiment with and then if you want to do pitches pretty pitches I would love to do some
49. 7:52pretty pitches Kyle where do I go what do I do there's four of them the bottom two are stable diffusion but they're kind of
50. 8:00friendly versions of stable diffusion mid-journey in my opinion's the best but you got to make it in Discord which is a
51. 8:07pain in the ass but it's the best that's the one I use I pay for it that's 10 bucks a month Adobe flyer fly fly or fly Adobe flyer fly
52. 8:16uh charges you you have to have a creative uh whatever it's called Creative Cloud license uh but that's adobe's version that's commercially
53. 8:28um it's not trained on copywritten stuff so if you're in a corporation that one's probably going to be the one you'll end
54. 8:32up playing with Leonardo's a web interface clip drop is from stability AI themselves there's some pretty cool
55. 8:38tools at clip drop and then if you really want to geek out um stable diffusion like automatic 11 11 just go go to Reddit and go to the
56. 8:47stable diffusion subreddit and see all the geeky projects there there's a lot of development going on with stable
57. 8:53diffusion uh which makes it super cool but you gotta you gotta be a little geeky to do it so these these allow you
58. 8:58to do it without being super geeky and then openart.ai promptbook look it's designed for stable diffusion but this
59. 9:05is a great document and it'll teach you about how how these things work in general um and and how deep they are thank you
60. 9:11Adam appreciate the Rose The Rose is sweet oh the other thing we do here which I think is kind of cool is I answer questions
61. 9:24so you give me your questions you put them in the comments there and then I'll answer them and then what's going to
62. 9:30happen is this see I make predictions too I'm going to predict that before the night is up someone's gonna go why aren't you
63. 9:38answering my question Kyle this is very frustrating a little feedback on the channel this is what we got last night I
64. 9:45like the content here it's fantastic but let me give you a little feedback on the channel I asked my question hours ago
65. 9:51and you haven't answered it yet so that's a little you know I would go ahead and not do that if I were you I'd
66. 9:56go ahead and answer my question so that person probably came in like an hour after I started and then I was
67. 10:03already behind because I answer all the questions I go through and answer them all so I there might be a better way to
68. 10:08do this and at some point I'll do it but at this point I'm just I'm I'm right now an old man stuck in my ways and I'm I'm
69. 10:15just a curmudgeon and it's just the way I'm doing it and uh at the point at which I take twerking lessons things are
70. 10:22going to change in this channel it's going to be a lot more spinning lights probably a lot more strappy garbs
71. 10:34I would think some spandex would be in order but right now you get t-shirts and a shitty haircut um another big topic conversation what the
72. 10:44[ __ ] you drinking Kyle that's just water but what do I have in it it's not booze no no apple cider vinegar
73. 10:54all right here's something this is not an AI tip but this is [ __ ] insane if you've got acid reflux the reason you have acid reflux is not
74. 11:05because you have too much acid it's because you don't have enough try if you have it well partially it's what you eat like if I eat
75. 11:14really like like uh something like fatty with dairy like ice cream will do it just I don't know certain combinations of food it's always
76. 11:24shitty food certain combinations of shitty food I'll get it and uh if you drink apple cider vinegar it adds
77. 11:32acid to your stomach the little flap that makes that happen it opens if your stomach is too low in acid it's
78. 11:39counterintuitive because as it comes up you're like I got too much acid no to little acid not too much I didn't know that
79. 11:48I didn't learn that from chat gbt but I did learn that from the internet and if it's on the internet it's true that's the law
80. 11:58and you know does it taste like [ __ ] oh yeah that's why I washed it down with a little coffee mmm all right so pop your questions in the
81. 12:08comments below I will do my best to answer them and uh keep up but I will fall behind on the keeping up
82. 12:15there are some regulars in here welcome welcome welcome good to see you all right what have you all said while I
83. 12:23have been rambling on Sonny and Cher I don't know what I was singing but it was probably sunny and Cher for a second
84. 12:29there I thought you were an Indian not really eyes look too old to see that well oh our eyes too old to see that well the bear hello
85. 12:40she bath cyber security 101 damned impersonator are they not really doing it by Insane I reported it to but the only choice
86. 12:50choices are they're impersonating me or a celebrity wait a minute I'm a celebrity tell them I'm a celebrity
87. 12:58in my own mind they don't have that option but you don't need to tell them that I was a I was in New Jack City I played white boy number one
88. 13:11not white boy number two I got the I got the lead when it came to all the white boys in New Jack City white boy number
89. 13:17one right here I was driving the car it's true I still get checks from uh who's it from
90. 13:34um Warner Brothers Warner Brothers Holt yeah Warner Brothers I got I got a residual check for seven dollars and ninety cents
91. 13:47for what was the period it doesn't say does it say residual is 10. year to date I've made twenty two dollars from New Jack City
92. 14:01that's that's a celebrity right all right so Kyle was a celebrity exactly oh yesterday wait yesterday was sheep wait oh sure
93. 14:12bear not she bear working on a name yeah sure bear yeah I I okay sure bear not she bear well I don't I'm reading fast what are
94. 14:23you gonna do all right well you come up with a name and then I'll keep saying Bears she bear sure bear or the bear
95. 14:31depending on my mood and when you get it nailed down we'll go there we'll go with that Sam corn hello good evening good evening
96. 14:39what day is today Wednesday witness day Wednesday what's interesting is gpt4 is a 32-2 token window normally but code
97. 14:50interpreter must have more yeah well Sam so gpt4 chat gpt4 still only has four thousand maybe it has eight the
98. 15:03API I don't think it has 32 yet I think the 80 the API only has 8 or 16 it might have 32 I don't know but then yeah code interpreter you can
99. 15:16upload up to 100 megabyte file so code interpreters got a significantly bigger context window here's what here's what I think is going
100. 15:23on with with uh with code interpreter is it spins up a virtual machine inside your chat session and I have a feeling
101. 15:32it's just it's like embedding whatever your document is into a little mini Vector database which is basically it's basically gptizing your
102. 15:43whatever documents you upload into its sort of local memory for that virtual machine I think that's what's going on there
103. 15:51um so so it doesn't have to play by the same rules it can it can sort of Chunk Up what's what's things are in the vector
104. 15:58database there there can be a hell of a lot more of it there than just the the context window can do in a single chat so
105. 16:06that's speculation but I think that's what's going on there but yes the the net effect for you as an end user is code interpreter
106. 16:14you can have a lot more data in there so you can either go to Claude and do up to 60 or 70 000 words just in
107. 16:23the chat or with an upload or you can use code interpreter the challenge with code interpreter is because it's doing the autonomous writing
108. 16:32um the writing Python and interpreting it and then seeing if it got it right and it it sometimes you want it just to do a
109. 16:40simple thing and it starts a whole chain of things and it takes a while so it can be a pain in the ass so I've been
110. 16:45finding Claude pretty useful to just pop a long piece of text in there and do something with it all right so I had nine million data points
111. 16:55spreadsheet I wanted to put in code interpreter wait had a 9 million data point spreadsheet I wanted to put in
112. 17:00code interpreter so bad earlier did you do it Jesse or is that too I don't know what that translates into in terms of
113. 17:07megabytes maybe that's too much for it if you got that sort of stuff like start playing around Lang chain
114. 17:14because you can pop that [ __ ] in Lang chain and go go to town fun suggestion is Moises which which is an AI music tool for splitting out
115. 17:25tracks and guessing chords so that's cool I like music I like music music's cool let's go to Moises Moises have you
116. 17:32seen Moises no but I used Moises isn't that a lotion all right Moises wait Moises the musicians vocal remover and much much more
117. 17:48You've Got Mail start free yeah yeah same yeah you palooka oh sorry um [Music]
118. 18:05all right what do I do new your playlist is currently empty you're gonna upload a separate track a master track demo playlist
119. 18:21mode [Music]
120. 18:42now listen [Music] let's champion you singing Buddy you singing champion
121. 18:57[Music] foreign here let's see
122. 19:13oh mute oh see oh let's see so so it splits it into stems basically oh that's pretty slick so wait I can solo this
123. 19:27is soon [Music]
124. 19:52it's pretty slick that's pretty slick it's a little muddy but it's that's pretty slick yeah like it didn't separate out the violin
125. 20:07that's cool all right I'll definitely check that out moises.ai Studio Dot moises.ai Moises it really just does sound like a
126. 20:17skin condition but but thank you for that who's who gave me that uh Sam corn thanks you Sam that was cool that was very cute got coffee I do
127. 20:30got coffee Tobias was shaking I got coffee and I got apple cider vinegar water because I know how to live
128. 20:40I finished my board deck I got back from New York this morning at like I don't know nine Denver time and uh took a little nappy then I did a
129. 20:53little grocery shopping and I did a little work and then I got on the I got on a tick tock I got on a tick tock with all you
130. 21:03fine fine people do me a favor if you would we just tick Tick Tock didn't blow out the algorithm tonight because they
131. 21:09don't like us so if you would share the live and hit the like button see if we can get some more people in here oh
132. 21:15they're just a few more just popped in but still let's get some more folks in here let's we're gonna make it a party
133. 21:20we want to make it a party and if you haven't followed me what the [ __ ] is your problem follow follow share like I don't bite it just
134. 21:29takes me a while to answer questions and I talk stupid is he really live he hasn't answered my question I don't think he's really live
135. 21:41why does he act like a dumb dumb all right is this even live why won't you answer me they won't do it time check 813 Colorado
136. 21:58813 Denver Colorado do you know where your fat dog is yes I do yeah why is this Geezer Gen X not answering me yeah yeah yeah see yeah
137. 22:11yeah see ya paluka he won't answer any of these questions I don't think he's really answering it he's probably an AI I heard his GPT five
138. 22:24welcome to Moviefone anyone live in New York in the 80s welcome to Moviefone press one for the Terminator Perez toe
139. 22:37wow is it it plus Health course two and one oh is this it plus Health course 201 you know I'm versatile I'll give I'll
140. 22:46give advice on anything as as we like to say about chat GPT and all these large language models they're mansplaining as a service
141. 22:55they will confidently answer you even if they're wrong aren't men the best which we're just we're just [ __ ]
142. 23:05insufferable but you know we try we do we do what we can nope Tick Tock just jumped me to the bottom because it knows I love that so much
143. 23:16all right we don't have too many questions so it didn't it's not too painful yet uh I think that's true though it's
144. 23:23healthy yes indeed ahem we are the Irregulars oh that's good I like that the followers of the AI learning lab welcome Irregulars
145. 23:34that's true we're probably all old and half of us are constipated half of us are just struggling to make it to the
146. 23:40toilet that's good that's solid that's solid I think I think I think you did it Emilio's life I think we've got the name
147. 23:48for the for the Posse the Irregulars it's good I think that one's gonna stick just like just like the poop in your underwear [Laughter]
148. 24:04he's a celebrity oh I'm a celebrity oh wait Sonny and Cher was my response to tick tock asking us to share oh yeah
149. 24:13okay you are freaking awesome thank you Bruce B R Bruce B extrapolate is the wrong word LOL but you know what I mean
150. 24:21okay I don't know when did I use extrapolate I'm normally pretty good with using words in the right context so
151. 24:27if I used it in the wrong context don't believe you about New Jack City we'll need the orb to con to confirm all the
152. 24:36orb actually the orb thing do I have it here I think I um did I do a screenshot hang on let me see if I can show it here if I go effects
153. 24:51and I go green screen and I go that I can show you my leaves that yeah it's here look uh
154. 25:07all right so there's the there's the uh World ID verified at orb Anonymous it's one of the cool things about this technology is
155. 25:19I'm I'm verified but this is this is like what they call zero proof so I don't need to provide anything other
156. 25:28than this in whatever systems developed to support this but it doesn't have my name on it it just says if if you're the
157. 25:37owner of this particular wallet you know that blockchain address if you flip it over it's got the block the
158. 25:43public blockchain address on the back um as long as you have that you're verified human anonymously which is pretty slick
159. 25:51it's pretty slick I like it it's slick six six click how do I get rid of this thing oh there you go did it ah look at that old man
160. 26:03figured out the Tic Tac he figured out that technology he's pretty good at it it was surprising I don't know I don't
161. 26:11even think his fingers work how's he tapping on the screen I just subscribed to YouTube excited great thank you Grandma sing where are
162. 26:24you where are you like I was I was gonna say where are you calling in from this isn't radio in the 70s hey girl where are you calling in
163. 26:31from Fantastic Poughkeepsie oh wow I went to school there once got kicked out if you know what I mean is chat GPT creative were you in here
164. 26:44the other night and asked that same damn question and I answered it I I answered that question two nights in a row for like 45 minutes
165. 26:54don't be [ __ ] trolling me I don't I I'm thank you Wes oh wesn radio wsn radio welcome you still need dashes in your name
166. 27:10um because it's Westerner Westerner wrestner ad IO and cure [Laughter] wsn radio the sound of smooth jazz here in the
167. 27:25evening pull up a nice glass of apple cider water sip it back feel the smooth smooth tones of Kenny G here we go
168. 27:42I'm gonna get me a radio DJ Gig I'm gonna do it we're gonna bring back radio so here's what I did I put it it just it
169. 27:53it was too [ __ ] complicated to have seven things on a thing so I put Claude and Bing because I I just feel
170. 28:02like those three are the big three right now plus the prompts.chat I put them here so for people that are just getting
171. 28:08started those are the things to look at the alternates barred um the Llama models pie and Poe I just
172. 28:18put them on their own separate things so this is like if you want if you want to dig deeper or or as Apple says you want
173. 28:25to try some deep Cuts you know you know here's your hits let's go to the b-sides and then I put the image things with the
174. 28:37prompt book for images all on the same page because that made sense and then I just stuck future PD at the bottom because
175. 28:44future PD is okay but like it's just I don't pay attention enough to it so that's future PD is just the
176. 28:52thing that I do when people ask me [ __ ] I don't know anything about I'm like go research it for yourself it's right
177. 28:58there so anyway that's what's going on with that so good eye good eye good eye The Irregular is coming in hot with the uh
178. 29:06with the observations telling me I'm using words wrong pointing out changes in the what is this called the little window
179. 29:18um I dropped the Trump indictment PDF into code interpreter which was great best way to learn about oh that's a
180. 29:24great idea actually yeah those big long indictments there's a lot of them there's uh what is it 91 total charges
181. 29:30now yeah pop those indictments in there and have it summarize it asking all sorts of [ __ ] that's a great idea
182. 29:37you know what I've been doing with it is um I've been doing terms of service I you know I I have I am not a manual reader
183. 29:45as we've talked about on this channel before and so terms of service are even less interesting me than manuals and um
184. 29:54and so every time one of these especially these AI things one every time they they say we've updated our
185. 30:01terms of service which I normally would ignore that now I'm like let me see what you [ __ ] are up to and I go pop it
186. 30:07into code interpreter or uh Claude and and uh and and ask I just say summarize it in like you know the the most important
187. 30:1810 things in in 10 short bullet points and it'll do it it's pretty pretty dang good what's line chain oh wait Bing is
188. 30:25sometimes gets very slow yeah exactly um so I'll get the Lang chain in a second Bing gets slow uh well I think all of these things as
189. 30:37people use them they're using up lots and lots of gpus that might be just Microsoft kind of throttling the amount
190. 30:43of uh compute power they're willing to throw at the free service they haven't launched um co-pilot the the chat GPT rolled into
191. 30:53Microsoft Office and Microsoft Windows that hasn't launched yet they're gonna charge 30 bucks a month per head per 30 bucks a month per head
192. 31:00for for that service um which is decently reasonable for businesses but for individuals it might be a bit steep on top of you know Office
193. 31:09365 subscription um that might just be a throttling thing so that that's probably what that is uh the thing was spelled moisey m-o-i-s-e-y
194. 31:18yeah that's correct um Lang chain is a you have to uh you have to be a little geeky to use it but what Lang chain is is it's a chain
195. 31:29of different um components that kind of string some things together basically what happens with Lang chain is
196. 31:35the the simplest way I can describe it is you can chat GP you can gptize your own data so you embed your data and then you
197. 31:44store it in what's called a vector database and and the vector database is basically where you can search like chat GPT
198. 31:52um and then there's other components to it like you can search the web and things like that so basically you're
199. 31:56making a little application that you can upload your own data to and then you can have your own version of chat GPT where
200. 32:02your gpting your own data so if you've got all of your customer service documentation for your company you could
201. 32:08pop it into line chain and create yourself a little chat bot for either customer service people or even for your
202. 32:14customers to actually find answers to [ __ ] based on your documentation so it's pretty powerful pretty popular
203. 32:21um and and very very flexible it does require you to use your Noggin and your coding skills it's not insurmountable
204. 32:29like it's it's in the neighborhood of complexity of getting stable diffusion up and running except there's a few more
205. 32:35components to it but it's not bad if you want to learn about it just go to Reddit and go to the Lang chain subreddit and
206. 32:41I'm sure there's all sorts of tutorials there and you know Google collab notebooks that you can just fire up and and turn it on I'm sure so
207. 32:49all right so happy your flight brought you safely back to home base thank you Firefly I appreciate that my dogs are
208. 32:56singers too it's just so amazing like you know we've had a bunch of dogs some dogs can't watch TV um some dogs can't hear music we had one
209. 33:05dog that watched TV it's just like when the TV it goes on the dog was like hmm dog and run across the TV but like oh
210. 33:15and then uh champ just if it's acoustic guitar um or brass bands like like when they play the the national anthem in the F1
211. 33:24races on the weekend he sinks to those two so he likes marching bands and singer songwriter he's he's an Americana
212. 33:31kind of dog with the exception of the international uh national anthem exception and the other kind of music he just
213. 33:41doesn't care about classical music occasionally if it's got if it's mixed in a way where it sounds a little
214. 33:46acoustic he'll sing to that it's got some some uh what's it called uh if it just sounds acoustic e all right
215. 33:55let's see sup dude hey Stevo what's happening good morning hello you Hi how are you welcome welcome welcome
216. 34:02is that puppy here on the T-shirt it is indeed I was taking a little nap on the puppy's bed where he he looks out the window and so it is uh
217. 34:15not a good idea in a black shirt come for the information stay for the silly voices you've come to the AI learning lab but I
218. 34:23haven't learned anything yet shut up it's gonna be good soon it's gonna get better it may not love the dog hair on your shirt it's a
219. 34:32fashion accessory it is how are your board meetings usually oh how wait how are your board meetings usually go how many fidget toys do you
220. 34:40bring Well normally I'm presenting some of the stuff and then uh and I've got my screen share on so I'm usually pretty busy and you know
221. 34:49actually having to pay attention um they usually go about an hour and a half um and we started we're like an 11 year
222. 34:56old company at this point so our board members know us we know them so we can be pretty efficient with them
223. 35:03and it's not a contentious thing we had a we had a a contentious investor for a while but we bought him out so
224. 35:10that's good that's gone that that headache is gone so we just got people trying to do the rat thing we're trying to figure some [ __ ] out
225. 35:21all right is oh just jump me to the bottom King gang it just yeah freaking Rock Em Sock Em [Music]
226. 35:38is that really apple cider vinegar or AI juice oh wouldn't you like to know wouldn't you like to know now that I've
227. 35:46had my Iris scanned maybe I got special drinks maybe I do maybe I'm part of a club you're not a part of you have to
228. 35:55you have to have your own Anonymous World ID to be part of the club I'm a part of whoa these large chain molecules taste delicious
229. 36:09all right I feel like we live in a world a weird time warp or something no one around me wants to talk about AI oh
230. 36:16April strangeways so yeah so here's the deal with that [ __ ] them I like I've got I've got I mean here's
231. 36:24the thing I was an entrepreneur in the early days of the world wide web and I got I got this crack addicted on
232. 36:31the early days of the web this is like 94.95 and I was a screenwriter and I was in theater companies like I had you know
233. 36:38actor and writer friends and they all just thought I was [ __ ] weird like what the [ __ ] are you doing what are you
234. 36:42talking about why do you what is this world wide web artist Consortium thing why do you always like why are you
235. 36:47always talking about this and then I I started agency.com with this big agency and it got really big and it got really
236. 36:53famous and and everyone was like ah hey you know I I should have listened to you like five years later they're like this
237. 36:59is really amazing this sense of this stuff's amazing I should have listened to you I'm like okay yeah I mean yeah you could I was
238. 37:07I was talking about it incessantly all you had to do was not be a douche all those same people why are you
239. 37:14talking about this AI stuff this AI stuff's awful it was terrible I don't know why you're talking about this I mean [ __ ] them [ __ ] him
240. 37:22yeah it's weird it's weird but here's here's the here's Here's my thought on it is we've got you know 45 or 50 years
241. 37:30of Hollywood is essentially telling us that the robots are going to kill us we've got technology that is going to be
242. 37:38disruptive is gonna is gonna um cost people jobs I think it's also going to generate a lot of jobs it's
243. 37:44going to generate a lot of prosperity but it's going to be disruptive it's going to be a mess um and and it is
244. 37:52um it's a lot of change and it's very existential right if you're a copywriter the first time us chat GPT to do something that you
245. 38:02know how to do that maybe you spend many hours of your day doing and it does it like that and then you start getting the prompting
246. 38:10dialed in and it does it even better and then at some point it does something in five seconds better than you can do it
247. 38:16and at that point you go [ __ ] what what's my role it's like it's it's terrifying right and it's you know I
248. 38:24think about voice actors like all of the you know they can now take three seconds of your voice and synthesize that and
249. 38:32right now the synthesized voices are a little you know robotic and monotone and shitty um but but you know I saw one that came
250. 38:40out I don't know a week or so ago that's got all these different emotions in it and when they start dynamically
251. 38:46assigning emotions based on the script this stuff's gonna start getting more and more realistic it's going to start
252. 38:52having the Cadence of a human it might not have the Cadence of an add freak like me talking a million miles an hour
253. 38:59but at some point it will and so so like I get why it's scary and and here's here's my other thought like
254. 39:08I get I get this scared thing hey thank you very much for the rose um I get oh w-e-s-n w-e-s-n now with
255. 39:16underscores I've changed my name I gave you a rose and I'm hoping you call me out welcome to wesn it's the hotline
256. 39:26how are we doing today what's your request Jane wants to know all right um so thank you dragonfly so I get the um I get the fear thing
257. 39:41um I also get that some people after having tried it are going to decide not to use it I totally respect
258. 39:51that like what I'm not down with right now is people that are sitting on the sidelines [ __ ] all over it that have
259. 39:57strong opinions about Ai and they haven't actually used it if you've got strong opinions about AI but you've actually used it cool like I
260. 40:06that's actually dialogue I'm really interested in but most of the people that have strong opinions about AI haven't played with it
261. 40:13my thought is this this shit's coming whether we want it or not like I kind of think of it like um you know people deciding that they're
262. 40:23not going to use AI like like the writers saying you know don't use AI to write screenplays or the actors saying
263. 40:29you know you can never use AI in a feature film it's kind of like the State of Florida saying we're no longer accepting hurricanes
264. 40:36it's a new law we just passed it okay great nice law great sentiment but the hurricanes are still coming
265. 40:46I'm gonna prepare for those at all and that's kind of my that's that's where I am with this stuff is is like I don't
266. 40:52have judgment whether people like it or not I don't give a [ __ ] like I get why some people might [ __ ] hate it I
267. 40:58think especially creative people you know a lot of creative people they're they're self-worth doesn't come from the output it comes from the
268. 41:07process if you love illustrating if you love 3D modeling this [ __ ] is gonna piss you off like I get no enjoyment out of that it's just a
269. 41:16machine that squirts out [ __ ] even if it's as good as what I do or better than what I do I get my enjoyment out of the process
270. 41:24great you won't use this stuff that's cool but just know you're going to be competing against people that are
271. 41:31and then but and and everyone can make that choice and and you know there'll be people that are not using AI that still
272. 41:37have good flourishing careers but those careers are just going to be you know narrowed you know increasingly down
273. 41:44um but yeah it is it's we're in a weird weird time warpy kind of here's where we are April we're in the beginning right
274. 41:52chat EPT is only nine months old um this [ __ ] ain't getting uninvented there's there's way too much investment
275. 42:02way too much competition the last couple of nights people have been saying oh there's been this article that Chachi PT
276. 42:08may be out of money by the end of the year okay they might be or by 2024 whatever it was um but I don't think that's gonna happen
277. 42:16and there's there's 40 companies sitting in the wings waiting to get their technology out there um so so there's you know it's coming
278. 42:26it's coming so um the other thing you can do April is find community so I created a community called the AI Salon we actually had some
279. 42:33people from in here came to the salon and last night they were like hey I came to the salon it was awesome it is it's
280. 42:38really cool so if you go to the salon.ai there's a it's a link tree and the first link talks about you know what the salon
281. 42:48is the second link is the Meetup the third link is the Discord um so you could join a group like that it doesn't cost anything it's free
282. 42:56um and uh you could join a group like that or find something locally find people that are talking about this stuff
283. 43:02make new friends who who are exploring this stuff and getting their heads around it because we're all trying to
284. 43:08[ __ ] figure it out and if you've got people that are like I don't know why you do that that's a waste of time
285. 43:15you're enabling Beelzebub if you if you got people like that just [ __ ] them all right it's called Moises uh I don't
286. 43:27know the URL because I use the app usually Moises is moises.ai that music app Studio dot m-o-i-s-e-s dot a i
287. 43:44Silver Fox um yeah same in my surroundings yeah where people are like no we you're a dum-dum if you talk about AI
288. 43:54That's stupid that's stupid that's stupid I like I like I like real work I like I I'm a writer and I like
289. 44:03real words if yeah if you got those people um yeah the glitches is in the Mage Matrix raffle um hello everyone how far behind are we
290. 44:16well root source I started this I don't know maybe half an hour yeah half an hour-ish behind but you asked that question probably 15 minutes
291. 44:28ago so when you ask the question we we had just gotten started but when I'm answering it now we're behind I have 56
292. 44:36messages down below that I have not seen yet so so for the new people here welcome yeah we got a few more people here my
293. 44:42name is Kyle Shannon AI learning lab we talk about [ __ ] here if you haven't played with chat GPT there's the URL that's the official URL
294. 44:49I'm not getting paid by them I don't give a [ __ ] if you use it but if you haven't used it go use it go play right
295. 44:55now if you don't know at all what to put in there go to this one prompts.chat that'll teach you how to use it
296. 45:01and then these other two are just the kind of two main alternatives to chat GPT Claude and Bing Bing is chat gbt but
297. 45:09it's connected to the internet from Microsoft and it's it's been hobbled a bit so it's it's Bing is a pain in the
298. 45:17ass right now but it is connected to the internet and Claude you can put in giant text blocks and it reads it all right so
299. 45:24those are go play there if you're new um so welcome everybody welcome welcome welcome and then if you have questions
300. 45:30about AI pop them in the comments below and I am behind but I do try to get to everything it's just I I never succeed
301. 45:37this is like life you know you try to get to your to-do list you have six things on it you're like this seems
302. 45:44reasonable you get to the end of the day there's 17 things on your to-do list and you didn't get two of the six done the [ __ ]
303. 45:52this is why I don't like to-do lists I I see them as monuments of Shame I have a philosophy philosophy about
304. 46:01email inbox zero and to-do lists if it's important enough it'll come back uh I have a quick question is there an
305. 46:13add-on to help AI summarize articles and books they're so yeah um articles so so probably let's here let's
306. 46:23go play I'll go show you how plugins work so if you pay your 20 bucks a month to chat gbt you can make money look they added back the little
307. 46:34suggestions here that's all fancy all right so now we're in chat GPT if you don't pay your 20 bucks a month you get
308. 46:41GPT 3.5 if you do pay your 20 bucks a month you get gpt4 Plus Code interpreter plus plugins and at some point they may
309. 46:49bring back connection to the internet but they haven't but there's plugins that are connected to the internet okay um so
310. 47:00if we go to plugins uh let me let me show you one other thing because someone's gonna someone because of this
311. 47:07probably gonna go okay I'll sign up for the 20 Buck a month thing I don't see plugins where are the plugins you have
312. 47:13to actually enable the plugins and and code interpreter so you go to your settings and then you go to Beta
313. 47:19features and you actually have to turn them on so if you don't see them do that okay now that what did I just do what the hell did I just do
314. 47:30um okay we go to plugins so when you when you have plugins you've got this drop down menu and here's all the plugins I have
315. 47:38installed now just because they're installed I'm not actually using them in any given chat I can enable up to three of them
316. 47:46so the one I have enabled right now is web pilot um but what plugins do is they're like little micro services that
317. 47:56they've basically described what the micro service does and then when you interact with chat GPT it it basically
318. 48:02says are any of the enabled plugins you know needed to to service this query and then if it is it'll fire off the
319. 48:12plug-in and it does what it does so web Pilot's a good one so what we can do is let's go find let's go find an article well here's
320. 48:21here's a good one so so this is zapier's terms of service so this could be it could be a link to a PDF of a book it
321. 48:29could be a link to an article like this um for a PDF of a book you you I I I'll give you another option but if it's
322. 48:39articles or in this case terms of service I'm going to copy the URL here um I'm gonna say um
323. 48:55[Music] please give me the five most important things I am agreeing to in these terms of service comma in a
324. 49:09short bulleted list period and so now what it's going to do is it should fire off the the web pilot plug-in oh wait no hang on I didn't give
325. 49:20it the URL oh wait oh actually yeah please it's good it said of course please provide the link yeah dumb dumb
326. 49:27okay there's the link now I'll give it the link all right so now fired off the web pilot plugin and it's going off and it's reading that
327. 49:36article and now it should just give us those five bullet points back based on the provided content acceptance
328. 49:45of updated terms thank you for the hearty heart Clint appreciate that to access the service you must register for a zapier account
329. 49:55some features may require payment use of your service you can only use the service for lawful purposes
330. 50:03and then data ownership and privacy this is the big one you retain the ownership of your content including content from third-party
331. 50:11services however by transmitting the content through the service you grant zapier a limited license to access use process and store
332. 50:22content for specific purposes like providing and improving service to zapier's privacy policy and data processing addendum
333. 50:30also applied to your data okay um so I'm going to say please write a snarky poem about number five foreign
334. 50:44this is also how to use chat gbt it's just for your own entertainment in the realm of debt data fast and wide
335. 50:52number five stands with swagger and pride you think you think you own all with a click and a type oh dear user
336. 51:01that's just hype you upload and share thinking it's free but number five Chuckles it all comes to me with every
337. 51:08bite every tweet every song your data is with five all days all day long um you might think it's yours that cute
338. 51:17little meme but fives in the shadows plotting a scheme privacy at scoffs a little slide in a grin in the digital
339. 51:23age that's paper thin so next time you upload think twice maybe Thrice for number fives watching and it won't
340. 51:32be nice in the world of data the king of the hive beware dear user of sneaky number five oh chat EPT you're just awesome all
341. 51:42right so so that's that's a way to do it another way to do it is you can use code interpreter when you use code interpreter
342. 51:51you're actually going to upload a file so you're gonna hit that plus button and you can upload that's where you could
343. 51:56upload say to a PDF of a book or you could go to Claude claude.ai
344. 52:11and then they've also got a little upload thing here you can just paste an entire novel right in your chat window and go ahead and do that
345. 52:20so that's how I would do that if I were you hope that helps I hope that helps I don't know if I help sometimes I know I
346. 52:27help other times I don't know if I help because I don't I don't I can't see I don't follow to the bottom I don't know
347. 52:34what people saying until later and I realized oh I [ __ ] up all right I really enjoyed the AI Salon from Tobias it was incredibly
348. 52:42informative there were so many brainiacs yeah there's a lot of smart people there one of the things we do in the salon
349. 52:48is we have values and we we reinforce them a lot so things like generosity and curiosity and uh and empathy
350. 52:57like having empathy for people like a lot of times people come to the salon and they're like I just learned about to
351. 53:03say I should I'm [ __ ] scared like am I gonna lose my job like it's it's terrifying and and then you know the the
352. 53:10people that kind of lean into it and say okay I'm willing to be curious and I'm willing to and then and then those are
353. 53:16the ones that usually end up like sharing what they learned like here's what I learned and here's what I'm doing
354. 53:20now and now I'm really excited and you know seeing that transformation is amazing so thanks for sharing that
355. 53:25Tobias did you see the Lance Elliott article on Forbes 8 15 about mega personas no but that sounds interesting
356. 53:32let me go find it so I can read it later thank you um Lance Elliott Forbes Mega personas
357. 53:48I think I think personas are interesting the Bold promise of Mega personas oh let's go you know what we could do yeah
358. 53:55we can this is this isn't going to pay well is it no good okay so we're gonna go do what we just talked about
359. 54:03we're going to plugins I've got web pilot selected and they're gonna go um based on this and then I put in the URL oh crap stop generating
360. 54:19um based on the above article um let's say please give me three short bullet points
361. 54:34that capture the main ideas of the piece only one message at a time please oh I [ __ ] up hang on a
362. 54:55based on the article below please give me three short bullet points to capture the essence of the article
363. 55:17I don't know what that's from that might be for my my smooth jazz Channel wesn how can I help you oh smooth jazz 88.
364. 55:33it's midnight do you know where your doobie is based on the article by Forbes introduction of introduction of Mega personas in AI
365. 55:44imprompt engineering this scales up the concept of multi-personas where wait in prompt engineering for generative AI
366. 55:54the approach scales up the concept of multi-personas where instead of instructing the AI to pretend to be a
367. 56:01few distinct entities it could be directed to simulate hundreds or even thousands of personas simultaneously holy [ __ ] that sounds cool
368. 56:10Mega personas can be particularly loose God what is with my tongue Mega personas can be particularly useful for
369. 56:18simulating large-scale surveys or group oriented analysis I might need this for a project I'm working on right now
370. 56:27who's the April strangeways why thank you for sharing this I had not seen this this is really good caveats and
371. 56:33considerations how does it do it does it say I'm sure it does in the article uh but uh one could instruct the AI to
372. 56:41simulate responses from a large group Saving Times and research yeah I'm actually getting ready to do this I'm
373. 56:46creating a bunch of synthesized data so if if this project comes through um they want to take input from 2 000
374. 56:53people and sort of process it in real time so in order to develop the prompts for that I want to I want to sort of describe a
375. 57:02bunch of different kinds of employees in a company and describe the company and then have it generate you know simulated
376. 57:10responses from 2 000 people and turn that into a CSV file that I'll then use to prompt and dial in my prompts
377. 57:18um so I'm gonna go read that article thank you for that that's really awesome see you see what's happening here it's
378. 57:24there's a level of generosity here that just you don't find this any old wear this isn't if you give me 19.95 1995 today I'll
379. 57:34teach you how to make money oh well I suppose I could at this point I could put together you can make money
380. 57:46with Chad gbt instruction booklet and probably sell it for 19.95 because I'm a celebrity as we've discussed earlier
381. 57:54this evening yeah super cool dude with excellent stage or video presence super thank you Firefly that's awesome
382. 58:01um it's a Chrome extension what is oh wait Oh Thomas curl check out give me summary.ai it's a Chrome extension cool
383. 58:10okay great yeah yeah so if if uh if folks have written extensions that do like article summary book summary in
384. 58:17fact Bing sort of does it too although it it was shitty for the longest time let me see if that's gotten better I'll show you this
385. 58:26so I think I've still got the article here um so let me go all right so I'm in Edge so in Edge there's right up here is a Bing sidebar
386. 58:39so I can I can flip open the Bing sidebar we're doing this real time so this may or may not work so this is a
387. 58:45little mini version of chat GPT in theory it's aware of this this web window so in this web window I'm going to put in that Forbes article
388. 58:57all right so there's the article so in theory what's it now what's it saying here new Bing with images you can tell it to make
389. 59:06images generate page summary all right so I just clicked a button it pre-prompted me to generate a page summary
390. 59:12the early days of Bing this was all [ __ ] up sometimes it would recognize what was in that page sometimes it wasn't
391. 59:19Mega personas are a new technique for prompt engineering the author introduces things the challenges in caveats so clearly the
392. 59:27author used that because that's the same term we heard before so that's working pretty good so I'm going to say um
393. 59:37write me an example prompt that would generate one of these Mega personas based on what the author says
394. 59:55if it gives me a safety guidelines thing I'm going to friggin smack it one possible example prompt that could generate a set of these
395. 1:00:09you are to pretend that you're a lawyer who is a personality type you are blank age-old in a location you have been practicing for years
396. 1:00:23oh I see what it's doing okay so what it's okay so the oh I've done this before okay I this is actually really cool can I make this bigger
397. 1:00:32um no I can't so so basically what they did is they wrote kind of a Mad Libs prompt you know you're a lawyer of personality type of age
398. 1:00:43you're you're in a location you've been doing it for so many years you've got this satisfaction level you're this gender you're this race
399. 1:00:53ethnicity and then each one of those things is a variable they put them in square brackets and then basically you say you know
400. 1:01:02create you know variables for for all of those things and then answer that question for a lot of different folks
401. 1:01:09and stick that in a in a CSV file yep I could see that working that's super cool that's super cool so another option here
402. 1:01:15other than the Gimme summary AI is use the edge browser with their Bing sidebar which now looks like it works again so
403. 1:01:24Bing getting better that I didn't use that for a bunch of months like six months because it just sucked but that's the thing about this this
404. 1:01:33technology right now kids all you Irregulars out there um is this stuff is changing all the time so the thing that was laughably bad
405. 1:01:44like like Google's barred was embarrassingly bad for months six seven months and now it's not and that feature in Microsoft Edge was
406. 1:01:57really bad and now it's not um do me a favor if you would just go ahead and share this live if you're
407. 1:02:03willing to do that if you're willing to be weird in front of your friends that are like why are you sharing some old
408. 1:02:08man talking about AI is he even qualified he's he seemed like he was a little touched when I watched it what are you
409. 1:02:16doing with your life if you're willing to put yourself out there then more people can come see this okay I'll do like the I'll be like this
410. 1:02:26okay we're gonna talk about the artificial intelligence because it's I I like the artificial intelligence it's
411. 1:02:34all right well I like intelligences you can probably imagine from what was I saying all right yeah it's the Chrome extension hola buenos
412. 1:02:50noches Julio H welcome I I said that with kind of a fake actor accent I don't know how I did
413. 1:03:06um but but I think it says good evening salute so hello or good day or uh from Colombia hello from Colombia that's
414. 1:03:14awesome welcome welcome welcome welcome I do not speak Spanish I just speak there's a famous Steve Martin from one
415. 1:03:21of Steve Martin's early albums he goes those French have a different word for everything he's like he's like he's like I was I I
416. 1:03:28get to the hotel I get in the cab I would like to go to the hotel please driver turns around and he goes so then you try to ask them
417. 1:03:39with a French accent I would like to go to the hotel please [Laughter] oh Stevie Stevie was funny back in the day
418. 1:03:50before he got all too wrapped up in that banjo I mean he's good at the banjo but you know it's not quite as funny as the
419. 1:03:57arrow through the head come on all right you have a great radio voice thank you very much Alyssa thank you very much welcome
420. 1:04:09welcome to wesn smooth jazz88 don't ask the same DM questions yeah don't do it I'm too subscribing for you thank you very much
421. 1:04:23appreciate that you high if we could be wait if we could be just be in real time then other people wouldn't need to ask
422. 1:04:31over and over I know but that would require me to change how I do things and I'm not willing to do that right now if
423. 1:04:37someone someone I think was Emilio's wife said there's probably a way to do this where if we could pull the API feed
424. 1:04:44this would actually be cool if we could pull the API feed from the chat send it to open Ai and summarize what
425. 1:04:53the questions are especially if it's like you know they're kind of Stack ranked like here's the ones people are
426. 1:04:58talking about the most and maybe you have one column with things I should answer and another column was snarky ass
427. 1:05:06comments that I can make fun of see that would be cool and you know Tick Tock could do that but you know they're
428. 1:05:12not gonna so someone on here wants to do a little development project you go go have chat GPT teach you how to code then
429. 1:05:19you can figure out if there's a tick tock API then you can write that application it could be I don't know it
430. 1:05:25could be piped into my ear through some sort of Bluetooth connectivity device I'm telling you we could billion dollar
431. 1:05:32idea right there you just go nuts with that thing and then well it wouldn't really solve anyone else's problem but
432. 1:05:38mine avocado jumpsuit saw a startup that uses AI to read and evaluate business plan decks for VCS VC Associates scared yes Ted yes um
433. 1:05:50the the uh the folks that read screenplays like coverage on a screenplay it's like all the associate agents at
434. 1:06:02William Morris like their job's done like that that that ship has probably sailed because you could take the
435. 1:06:10coverage documents that have already been written one of the things these things are really good at is mimicking
436. 1:06:15other documents so you could just say um you know take this screenplay put it into one of the uh plugins or
437. 1:06:24code interpreter take this screenplay follow this format for tone and structure and give me coverage of the screenplay
438. 1:06:32you could knock out 100 of those in an hour so here so here's a lesson so if you're one of those people at William Morris that
439. 1:06:41reads screenplays be the first one to get good at this [ __ ] because at some point someone's gonna go who knows how to do coverage
440. 1:06:50with AI I do okay come here you're high you you get to keep your job AI will not replace your job someone using AI will
441. 1:07:05it's from Paul rateser from the marketing AI Institute that's the thing that's the thing it's a thing is it a good thing nope is it a bad
442. 1:07:16thing not for the person that keeps their job that's why this is here get curious get learning but what if I don't like it Kyle I don't
443. 1:07:28know then don't just don't be surprised when this [ __ ] comes and smacks you upside the head I didn't think it come
444. 1:07:35out could come after consulting jobs I have a master's degree I get paid 200 000 a year sorry not gonna need your
445. 1:07:45services anymore this thing just does it now is that really true no but some amount of people in every knowledge job
446. 1:07:53in every creative job are going to be displaced within five years there'll be lots of new jobs so a lot of those people are
447. 1:08:02going to migrate to new jobs but the ones that don't make the transition won't and it's going to be really painful
448. 1:08:10it's gonna be really painful and I don't like it I don't like that part of it but I'm just messenger are you even qualified to teach this
449. 1:08:18class when's the test oh there will be tests there will be tests shortly after the twerking launch is when we start the
450. 1:08:27testing Google vertex AI has some amazing image question answering that I haven't seen anywhere else I haven't
451. 1:08:33played with vertex AI let me see that this is tonight I'm like learning more from you than I'm teaching which is awesome um Google vertex AI
452. 1:08:47yeah I went on to watch him he's supposed to be some sort of expert or something and then everyone in the in
453. 1:08:53chat do more than he did every time they said have you heard of this he's like no I haven't heard of that I don't even know what he does
454. 1:09:01and then it seemed it seemed like it someone said he was an actor it seemed like he's an actor he doesn't know anything about
455. 1:09:10technology I think he's acting like an expert that's what I think it is I think he's acting like an expert
456. 1:09:17and he's pulling a wool over everyone's eye acting like an expert and I bet he's I bet he's signing them up for a
457. 1:09:23multi-level marketing scheme I bet this is a pyramid scheme I knew it I knew it I didn't trust him when I saw him there
458. 1:09:31was something about him I didn't like it not one bit okay let's go look at ver vertex uh uh unified data platform that offers
459. 1:09:42fully managed ml tools for text so I assume this is something where you can upload your own data new customers get 300 in free credits
460. 1:09:51cool build deploy and scale machine learning models for faster fully managed ml tools and use case go to the console
461. 1:10:00huh oh sorry this is machine learning developers data scientists and data Engineers yeah see there's where you lost me a bit
462. 1:10:07this is probably really powerful but it's you're gonna generative AI Studio try now language speech and vision that this is
463. 1:10:17I'm starting to have PTSD of of the past five years when whenever I would heard about tensorflow like Google would say
464. 1:10:24there's a new version of tensorflow that's really easy and I'd be like oh good I can go play with machine learning
465. 1:10:29[ __ ] and I'd go there and be like I don't have a math degree I'm not an engineer test tube text prompt code prompt
466. 1:10:42oh this looks pretty cool so this is basically just like open ai's playground so yeah this looks pretty good all right I'll mess around
467. 1:10:50with that so I got I've now got three things to go check out this is cool tonight thank you for that Sam
468. 1:10:56uh I'm 50 and about blind LOL can we see a list a little larger for the moment absolutely sorry about that so I understand we've got we got
469. 1:11:08limitations but we got some superpowers there what's your name oh April um April we got some superpowers like
470. 1:11:14we've been on the planet long enough that we can use these tools in powerful ways so just screenshot that give that a
471. 1:11:20screenshot so there's your there's your basics here's your bonus category of uh large language models Bard which is connected
472. 1:11:31to the internet that's Google's entry uh perplexity AI is good on its own especially if you're doing research and
473. 1:11:38want citations and then lab stop perplexity is meta's llama 2 models that you can play with pi is super chatty so screenshot this
474. 1:11:46po you can create Bots you can play with multiple large language models within the same interface so that's that
475. 1:11:55then this is all of your image not all of them this is a handful of image things mid journey is the big one the
476. 1:12:02that I think is still the leader of the pack Firefly is not trained on copywritten data that's from Adobe and
477. 1:12:09their generative fill is incredibly powerful um I mean Firefly is not bad and they've got text tools there that are pretty
478. 1:12:16interesting Leonardo is stable diffusion under the hood but a really cool web interface on top clip drop is from
479. 1:12:23stability AI so from the makers of stable diffusion this is like their playground of experimental tools and then the prompt book openart.ai
480. 1:12:33prompt book that'll teach you how to talk to these image generation things it's designed for stable diffusion but
481. 1:12:38the basic things still apply and then futurepedia.io is um whatever you call it it's just a big directory of AI tools all right so
482. 1:12:48screenshot that and then the final little piece is uh that's the link to the AI Salon that I um host
483. 1:12:56um there's a a whatever you call it a uh link tree there with with links the Meetup we meet every other week and then
484. 1:13:05there's a Discord link there and then storyvine is uh my company and that's my social media tag my tag my my app you
485. 1:13:15can at me there you can add me there hang on the dog's line and I'll be right back what are you doing
486. 1:14:21foreign one of my boys microwaved me some chicken tenders
487. 1:14:35with blue cheese sweet mm-hmm okay don't eat with your mouth full [ __ ] rude
488. 1:14:56okay how's it going with the workshop planning um I had a big client meeting this week and I got a board meeting next week so I did
489. 1:15:09the board meeting I've got a 15-minute thing I got to record for virtual Work Workshop I've got a talk that I'm giving to a big
490. 1:15:23Pharma company on the third I here's my commitment I think um I knew I shouldn't have put crazy glue
491. 1:15:39in my drink um I think the last week in August I can probably start thinking about a workshop like I've already done a workshop for um
492. 1:15:51for a big big company so let me put something together let me put something together I gotta get my [ __ ] together
493. 1:15:57and I'll just do it thank you for asking thank you for keeping me honest writing them down cool only voted three times in my life not
494. 1:16:06into politics yeah I'm I've voted but I'm I'm more into politics now that the country is this polarized I think we
495. 1:16:17gotta we gotta fix some [ __ ] I don't like where we are at all okay y'all I'm glad he spends time and
496. 1:16:28actually answers questions in depth oh thank you Tobias I appreciate the kind words gpt4 is incredibly expensive to
497. 1:16:35run they spend seven hundred thousand dollars a month hosting it I think it's 700 000 a day would that be right that could be right
498. 1:16:45yeah I think it's I think it's has Sam Altman described it it is eye-wateringly expensive wonder if they will include it with my
499. 1:16:56with M3 licenses and up I don't know what M3 licenses are Chris Ward I assume you're talking about Microsoft
500. 1:17:03bubble IO is creating apps but how to get higher for SAS developer um I don't know
501. 1:17:20you know if you're if you're a SAS developer I mean I would I would say just lean into like get good at really quickly spinning
502. 1:17:28things up right because you know you can use these these tools to speed up your coding you know I would I would I would
503. 1:17:39definitely play around with Lang chain if you're if you're a SAS developer I would play around with Lang chain and
504. 1:17:44get that up and going because what you can do then is you can you could find a sector of businesses that really want
505. 1:17:52to be able to use their own data but don't have the technical chops to do it so you could either build your own SAS
506. 1:17:58product for that or you could go to those businesses and say hey I can build you a your business GPT really really
507. 1:18:04quickly that might be a way to do it again my My Philosophy on you know you can make money is it's it's it's completely 180 degrees
508. 1:18:16the wrong approach you don't make money with the tool the tool is not the thing right you can make money with a shovel
509. 1:18:26how do you make money with a shovel I don't know you figure out who needs a hole what kind of hole do they need what kind
510. 1:18:33of shovel do you need and then if you can dig that hole better than other people can dig that oh they'll give you money and then you can
511. 1:18:39make money with a shovel so so if you're already a SAS developer what I would say is you know figure out where there's a business need
512. 1:18:50and then if you can find a way to you know demonstrate that you're good at that you can either go get a job with
513. 1:18:57the company or you can start your own thing or you can start a Services business or you can start an education business
514. 1:19:03right do workshops do put put stuff out there all right Sam corn wow how can they afford to pay 700k a month no it's is it
515. 1:19:14let's see how much it is I think it's 700k a day from that article is that right let's see how much does Google spend it or not
516. 1:19:30hang on what is going on here okay how much does open AI spend why is that not typing in there good
517. 1:19:45damn it how much does open AI spend her day in compute
518. 1:19:59100K per day 3 million a month according to estimations chat GPT
519. 1:20:15yes 700 000 per day that's from The Business Insider article from April 20th seven hundred thousand dollars a day
520. 1:20:23now they've lost users so maybe it's down to just half a million a day so so you thought your electric bill was high
521. 1:20:38seven hundred thousand dollars a day a day a day so let that do the people come in and they say I want
522. 1:20:48you to write me a little poem about my doggy it's like watching four dollars [Laughter] the other thing in addition to Lang
523. 1:20:59chain that's super useful if you're interested in coding with AI is hugging face yeah so hugging face is a site that
524. 1:21:05it's like GitHub where but there's lots and lots of Open Source models like lots of different Frameworks AI Frameworks
525. 1:21:12for everything from images to animation to music and everything in between um both GitHub and hugging face assume
526. 1:21:20that you know what the [ __ ] you're doing with tech if you don't you're gonna have to get over that hurdle the good news is
527. 1:21:27you can use Chachi PT to create tutorials for you so like one thing you can do is you can take the readme file
528. 1:21:33from a GitHub repository or from a hugging face thing take the readme file pop it into chat EBT and tell it to give
529. 1:21:39you a step-by-step instruction for how to get that thing up and running and it'll do it you can even say I'm on a
530. 1:21:45Mac or I'm on a PC or I'm on Linux whatever it might be if you're on Linux you're probably okay with command line [ __ ]
531. 1:21:52um but yeah hugging faces is is really good and hugging face also lets you um they host a lot of these models so you
532. 1:21:58can actually go play with different models there my experience has been that a lot of times the hosted models are sometimes
533. 1:22:07maintained and sometimes not and they sometimes work and they sometimes don't and it's really confusing to sift
534. 1:22:13through all of the [ __ ] so it it's definitely experience to get up to speed but it's it's an amazing uh resource so
535. 1:22:21uh if you if you're doing any like the person that was talking about SAS before like yeah you should be you should be
536. 1:22:27all over the sub subreddits for all these different projects um you should be all over GitHub you should be all over hugging face you
537. 1:22:34should be finding people like David Shapiro um who's who's you know building really sophisticated um AI tool AI systems out of these tools
538. 1:22:44and doing YouTube videos about it he's amazing Julio H okay bye friend Salon Julio I'm sure you've already gone but thank you
539. 1:22:52for coming Chris word trying to fly away I can create a team's bot that can do my work for me um I I think you could probably do that like
540. 1:23:04I I don't think you're that far from that so um I mean let's let's go let's go where do we want to go I'm
541. 1:23:14gonna try Bard for that one oh wait did I just delete one of my slides I did that was stupid that was dumb dumb
542. 1:23:22he was a dumb he was he had a little slide there had good information on it and he just deleted it I don't know what he did
543. 1:23:29um what am I doing oh teams okay so I'm gonna go to bard.google.com let's see
544. 1:23:50hello hello dang it um can you find latest
545. 1:24:08documentation on creating teams Bots I'm getting lots of hearty Hearts I don't know I'm getting lots of hearty
546. 1:24:18Hearts but thank you for the hearty Hearts I appreciate the hearts the heart makes me happy and I'm happy
547. 1:24:25I'm not programmed to assist with that well that sucks jerk let's go try Bing that I'm not gonna use you
548. 1:24:33so that's what you would call a safety guard rail where these systems are designed to not give you what you want because the lawyer
549. 1:24:42said it's bad all right um can you find me the
550. 1:24:59[Music] um all right it's it's not like in the the oh wait I would like access to your microphone hang on allow let's try this again
551. 1:25:11can you find me the latest documentation for the teams API as I want to write oh well it just took that
552. 1:25:27um sure here's the latest documentation all right if I wanted to if I wanted to write a bot could you
553. 1:25:42assist me with that if I wanted to write a bot could you assist me with that
554. 1:25:58oh of course Microsoft has lots of resources oh this is Microsoft right this is Bing msteams chat bot in 10 minutes creating
555. 1:26:10a chat bot with the power of teams virtual agents tips and tricks um I would like to simulate being at my desk oh stop responding
556. 1:26:27I would like to simulate I would like to simulate being at my desk and pretending that I am working how would I go about creating such a bot
557. 1:26:39this it might this it might kick back and go oh wait oh God damn it I am thinking about writing about that
558. 1:26:49can simulate someone sitting at their desk working period how might I go about that all right let's see what it does
559. 1:27:00looks like something went wrong start over yeah I had a feeling
560. 1:27:18yeah it I'd like to write a bot for Microsoft teams that simulates someone sitting at their desk doing work how do I do that
561. 1:27:38I just have a feeling anything and kind of like it
562. 1:27:57can you help me write the code for the first version no it said no
563. 1:28:12um yeah I I what what I would do what I would do is I would go to um
564. 1:28:26I would I would go find the go find the documentation and then turn it into a PDF and pop it into code interpreter and and it could help you
565. 1:28:39write it could probably help you write those Bots or go to Claude um or just take enough of the documentation into into gpt4 and have it
566. 1:28:48have it code the bot for you I think you could probably spin something like that up in less than an hour I would think
567. 1:28:55so I like it is leonardo.ai any good yeah it's really good it's really good I'll go show it to you so leonardo.ai is one of the ones I have
568. 1:29:05on my little list it's one of the ones I like because it's it's just it you don't have to program anything so it's all the fancy stable
569. 1:29:14diffusion stuff um but it's if it has a wizard it's got to be good it's got a wizard I like a wizard everybody's a good wizard
570. 1:29:24you like a wizard uh um welcome to leonardo.a.i
571. 1:29:38so it's got a cool dark background gives you a platform overview this is all new I haven't seen this sort of teaches you
572. 1:29:44about the Leonardo Academy let's go so what you got here is across the top you have different styles so like there's a cool sort of anime
573. 1:29:58pastel dream um absolute Beauty here's like deliberate that looks like 50s poster art right here's like a little you know Pixar or
574. 1:30:12you know cartoony kind of thing isometric sci-fi buildings paper art cute anime vintage photography right so
575. 1:30:21you got different styles so you can generate within that then you have like here's all the recent [ __ ] that's been
576. 1:30:27generated any one of these you can see what prompt they used right so you can just scroll down so you don't you don't even really need to
577. 1:30:35do like retro style theme so I can click on this and then can I generate with this no it's not letting me some of them you can generate with
578. 1:30:48like here's a seamless pattern why am I not allowed to generate with this model I don't get it it's that
579. 1:31:00text to image generation that did I do that let's see Future model um let's say seamless pattern of
580. 1:31:15chubby black and white cartoon dogs on a on an azure background
581. 1:31:30see what it generates you're currently on the free plan well you don't need to shame me about it good Lord
582. 1:31:39I'm just trying to do my thing here so the I think this is some things I did before did I do these yeah like I did this 50s wallpaper
583. 1:31:49pattern I did all these puppies those puppies is cute oh look it did it chubby black and white dog um editing canvas crisp upscaler
584. 1:32:02let's see this one's cute look that's cute I just made that I just made that and then you know what else you can do Leonardo no Kyle what
585. 1:32:10that's why we're here if if we knew what to do on Leonardo we wouldn't be asking you you dumb dumb oh yeah right okay
586. 1:32:15fine so you can go to training and data sets and you can upload your own pictures and train your own data sets
587. 1:32:23and so if I go to future or fine-tuned models and so here's all the models other people have done that you can go play with
588. 1:32:32but then I can go to your models and I've got two of my giant fat face but I have one of these cloud pictures that I took
589. 1:32:41generate with this model and so now it's my model that I'm generating in so I can do um
590. 1:32:57skyline of Manhattan why is it not typing Skyline or all right skyline of Manhattan
591. 1:33:13from the Brooklyn Bridge let's do Lower Manhattan because I was just there the lower Manhattan from the Brooklyn Bridge
592. 1:33:28during um Manhattan hinge do you all know what Manhattan hinges it's so cool it's the one day a year
593. 1:33:43where the sunset um sets exactly in line with the with the cross streets of Manhattan it's pretty cool foreign
594. 1:34:00oh Stevo 42 you were chat gpt4 that's pretty funny those are good pictures but it didn't really it's sort of used my skylines but not
595. 1:34:16really it it more used my editing style it's kind of my editing Style but it's not quite yeah it's actually
596. 1:34:22not bad that's my my clouds tend to be a bit more dramatic than that but those are good they're not really the Brooklyn Bridge at all
597. 1:34:35well that one is that's well that's one part of the Brooklyn Bridge but the middle like the other things right here like that's just wrong
598. 1:34:44and that just missed it entirely but anyway so you can do that um what else can you do here I don't know you can do lots like what's nice is
599. 1:34:54so Leonardo gives you sort of all of the bells and whistles of stable diffusion or kind of the most recent most stable
600. 1:35:01ones but but just that the whole idea of being able to start with you know a pre-existing model or start with
601. 1:35:08pre-existing images and then render within them like that's a super cool style like it would take you a while to
602. 1:35:13figure out how to do that now you can just go here and you know jump in and and do something in that style right so you could do um scary
603. 1:35:25Kitty um yawning we'll see if it
604. 1:35:40is the right uh model no it didn't do the drippy paint thing but it is a scary kitty yawning so anyway yep I like Leonardo
605. 1:35:59um I like it a lot I got a lot time is it 9 39 I got I got like 20 more minutes you know I gotta I could get some shut eye I've been
606. 1:36:10traveling all day but I still come here for my Irregulars we call them the AI learning lab Irregulars because we're older
607. 1:36:21and [ __ ] just don't work anymore I love that Louis CK routine where he's like I went into the doctor my ankle
608. 1:36:30hurt he goes yeah just wore out it's doesn't really mean it work because yeah it's just that's how it's gonna be now
609. 1:36:43it's so true it's so true just ship breaks it just breaks Why is my got my little camera a little too high there we go that that
610. 1:36:59all right hey oh just jump me to the bottom dang it dang it welcome everybody my name is Kyle Shannon this is the AI learning lab if
611. 1:37:08you're new here um if you have not played with chat GPT go there go to chat.openai.com that's the official chat gbt website
612. 1:37:19and if you want to learn how to prompt and you don't even know what prompting is and you're like what the hell you've
613. 1:37:24been talking about go to prompts.chat and that will teach you how to talk to these things and then these other two things are cool things to
614. 1:37:33um to compare chat GPT too all right I lost my place in my comments so I'm
615. 1:37:48trying to find out where I was in comments it is per day the seven hundred thousand dollars yes let's just start right there I think I
616. 1:37:59just skipped some questions but that's okay bear thank you so much for the dangly lights I love the dangly lights they make me happy
617. 1:38:09Lang chain great to expound on I've not actually set up Lang chain um I think I did expound on it but basically you can with Lang chain you
618. 1:38:19can create a gptization of your own data is the is the sort of short way to say it there's lots it can do but that's kind
619. 1:38:28of it in a nutshell you can take your own data you embed it you put it in a vector database and then you can search
620. 1:38:34on it just like you search with other stuff so it's quite powerful um if if you don't want to geek out
621. 1:38:41that sort of capability is is going to be coming right Google's gonna provide that with Google Drive um Microsoft will provide that with
622. 1:38:52their Office 365 suite and their cloud storage um AWS will probably provide that Apple will probably provide that so
623. 1:39:01right now if you're geeky and and want to figure this [ __ ] out go play with Lang chain because because
624. 1:39:08if nothing else you'll learn a lot about the different language models and how you can do that and what embedding is
625. 1:39:13and what the different Vector databases do and combining semantic search with Vector search and there's all sorts of
626. 1:39:20permutations you can do with that um but but don't feel like you have to do that right now that's sort of you know Advanced generative AI
627. 1:39:33but if you're geeky I think it's time well spent even if other stuff's coming yep hire the folks that use this to use that shovel
628. 1:39:42that seems too high what the seven hundred thousand dollars a day is it on Azure it is on Azure yes seven
629. 1:39:49hundred thousand dollars a day a day yeah well they raised 10 billion dollars for Microsoft that wasn't in cash that was in compute
630. 1:40:03so that might even be the non-real retail price right we need solar power for AI wow what a burn rate yeah it's crazy this stuff's
631. 1:40:15crazy but here's the thing like I you know I've got these split on different um things now but you've got those those three
632. 1:40:24you got those four plus you've got whatever Elon Musk is going to do plus cohere plus Apple plus AWS might come out with something
633. 1:40:36kind of consumer like they might do something with Alexa so that's 11. and then there's probably three or four
634. 1:40:46waiting in the wing so you so I I kind of feel like by the end of the year there's going to be 15 major viable large language models
635. 1:41:00trying to figure out how to get people to use their service so they're going to keep one-upping each other in terms of of services um
636. 1:41:11so even though that's a ridiculous amount of money so here's the stat for you Apple just said that they have spent the bulk of their
637. 1:41:2124.6 billion dollar research budget for this year on generative AI the market cap of open AI is less than Apple's research
638. 1:41:38budget foreign billion dollar research budget is what Apple's just toying around with so I think when Apple comes into this
639. 1:41:51Market they're going to come in quite powerfully like they do and they and they will likely redefine
640. 1:42:01when Apple comes in it ain't going to be this geeky it's gonna be really slick as as Jobs would say it'll just work
641. 1:42:11right these things right now kind of work if you know how to prompt and you know how to do this and you haven't done
642. 1:42:17that you don't go over your token rate and you you got to learn this one does that and that one does that and that
643. 1:42:21one's connected to the internet but if those don't work then you can go to these three and then you go to the that
644. 1:42:26won't be apple Apple's gonna be like we don't talk about this [ __ ] until it's good let the caveat that the past decades
645. 1:42:34kind of sucked for them but you know other than that uh how does this dovetail with blockchain and Bitcoin so um I showed earlier
646. 1:42:46I'll do it again I did it again I'll just do it I'll just do it right now so I let me how do I get rid of that
647. 1:42:58there you go so this is my this is my world ID so so it says verified at orb World ID Anonymous so this is my verified blockchain
648. 1:43:10I am a verified human on the blockchain this is Sam Altman's company called World coin and you verify your humanness on
649. 1:43:21this little orb that scans your irises which everyone's like you're an [ __ ] why would you do that I had a couple of
650. 1:43:29people Taco said well you just earned an unfollow buddy I'm like okay like I don't think I did anything to you but whatever um um
651. 1:43:44so why did I do that why did I get myself you know scanned and use my biometric data for that because I think that um moving forward
652. 1:43:57it's gonna be so easy to um fraudulently do everything represent someone else synthesize their voice make it seem like the New York Times
653. 1:44:09wrote a scathing piece on someone that they did they really didn't write it's it's just going to be all over the place
654. 1:44:15and it's going to be really hard to discern and so I think you need a foundation of trust right and so this is a system it's
655. 1:44:22Sam Altman's company so the CEO of chat GPT just launched this two weeks ago and I was in New York City you can you can get
656. 1:44:29this done in New York Miami and San Francisco right now but I assume they'll have more things coming um and
657. 1:44:40you know the ability to be able to say like imagine a Twitter where you know that everyone on Twitter is blockchain verified as a human
658. 1:44:48and then every post that they post is blockchain verified that yeah they wrote that and it hasn't been modified in any
659. 1:44:55way and yeah this video that Kyle put out is of Kyle and it hasn't been duped um so I think that AI is going to drive the
660. 1:45:06adoption of I don't know if it's Bitcoin and blockchain but you know some sort of tokenization Technology you know
661. 1:45:12blockchain maybe makes sense um and then some sort of payment infrastructure like like uh this the the world coin thing that Sam Altman did
662. 1:45:22this thing they've got three components they've got a cryptocurrency so that's the cash they've got the ID which is the
663. 1:45:30verification and then they've got a transaction platform like you know stripe or or PayPal or [ __ ] like that
664. 1:45:36but it's it's designed to use with this so you could imagine building so every e-commerce site that's out there right that all takes cash
665. 1:45:47you could imagine the eBay version of this where everyone on eBay is a verified human or Etsy right Etsy right now is a cesspool of just
666. 1:45:56Chinese crap right because no one's really verified and everyone can just do these fake accounts imagine having an
667. 1:46:02infrastructure where you could say yeah no this is you know the the AI learning lab Irregulars like we verify that these
668. 1:46:10are these are people here not Bots we can do that with this so I think it's going to drive it um I don't it's not
669. 1:46:15going to do it anytime soon it'll probably be three years out but um but I actually think it's important
670. 1:46:20and uh yeah so there you have it there you have it yeah yeah there you have it yeah see what I'm saying yeah yeah see
671. 1:46:28yeah yeah see oh who's that that's the uh that's the uh that's the brother from Napoleon Dynamite doesn't he talk like this with
672. 1:46:37his lower jaw and he just moves his bottom lip yeah yeah you know it's really good I threw the football I see the football isn't it him
673. 1:46:47eat the food dude I'm 57 year old Dev laid off last week and you've inspired me again thank you MK thank you for that um 58 here
674. 1:46:57welcome welcome gen xer let's go [ __ ] some [ __ ] up um yeah here's the deal I I just I I think that any time spent
675. 1:47:08in the next 18 months getting your head around this [ __ ] and getting articulate with it and the thing that I would
676. 1:47:13encourage you to do MK is as you're learning talk about it talk about it on LinkedIn talk about it with your friends
677. 1:47:19at dinner parties like let them think you're a [ __ ] weirdo aren't you too old to be learning something new yeah go [ __ ] yourself
678. 1:47:27learn this stuff because with your life experience you'll you'll be able to take the the sort of coding
679. 1:47:36superpowers the gpt4 does and you'll be able to use them in a way that that'll you know obliterate someone who doesn't
680. 1:47:45know how to do that like myself like I can make code with this but I can't look at the code and know if it's any good
681. 1:47:52right so if you can look at the code and go oh that's structurally [ __ ] up you're going to be able to just quickly
682. 1:47:56go oh so fix that and it will um or or anything it doesn't have to be coding like one of the things that you
683. 1:48:04can do with these tools is you can kind of reinvent yourself there's in the AI salon there's a guy Roger that was a he's a
684. 1:48:12retired software engineer and the the salon originally started out as the AI artist salon and you know and
685. 1:48:19he was like I'm not an artist and he was sort of curmudgeoning about it and then he ended up with his engineering
686. 1:48:25mind and his you know developer mind he got really good at prompting mid-journey and so he started making amazing art and
687. 1:48:34now he's like now he's become an artist right using these tools um so so be open to wherever it might lead you but just start getting
688. 1:48:42articulate with it or uh literate with it and and go that's that's awesome but sorry to hear you're laid off that sucks
689. 1:48:50that sucks McKenzie has introduced an internal AI chatbot that is the totality of their knowledge and data yep I think
690. 1:48:58every company is going to have this it's going to be quite easy to do gonna be quite easy to do so Kyle can you enable SS
691. 1:49:11I don't know what FF is but tell me what it is and I'll sure maybe single single sign-on I don't know what you're talking about
692. 1:49:20no e-commerce for Corporate America don't know what that means front end guy I found an article on it uh uh
693. 1:49:34felonious MK now's your chance to Pivot to helping Services love that some of us are wondering what Bots are um okay so I bought um
694. 1:49:49so you know in slack when people type [ __ ] you could you could create a bot a bot is basically just a script like a
695. 1:50:00computer it's computer code you could you could write a bot that anytime someone asks a question in a certain Channel
696. 1:50:13it sends that question to open ai's API and sends you back the answer from openai so bot is just kind of a service call to
697. 1:50:21an API that can that can do some things so you could create you could create a bot if you're doing the teams thing where um
698. 1:50:35randomly three times an hour your Bot could ask a question of one of your colleagues and and you could make it just ask a
699. 1:50:46name questions like hey do you know where the report is from last Friday so it seems like you're working and it
700. 1:50:53seems like you're acting or you know you're acting like you're working and you could be off you know having chicken nuggets [Laughter]
701. 1:51:05um but they can do more sophisticated things you could have a bot that every time a new article is written on a
702. 1:51:14particular RSS feed you summarize that article and paste the summary into a slack Channel or paste the summary into
703. 1:51:23an Excel worksheet or or whatever it might be so so just think of them like automations they're they're automations
704. 1:51:30where something happens it triggers the bot the bot goes and executes code and comes back and delivers something
705. 1:51:41a Bots like an Excel macro exactly simply recording of things for the computer to do but Auto yep that's it
706. 1:51:47and and Firefly zapier is a tool that allows you to create automation so they've got what's called a trigger so
707. 1:51:54the trigger might be someone types something into a Google form and then I'm going to take the input of
708. 1:51:59that Google form put it in a prompt send it to open AI it's going to generate an image then I'm going to send another
709. 1:52:04prompt and it's going to generate a paragraph and then I'm going to stick that in a I don't know Word document and then my
710. 1:52:14social media person is gonna check that document once a day for new ideas so zapier can do all of that the bot
711. 1:52:23won't do any good if they monitor the screen uh yeah but you'll be a there was a there was a dude that did this without um
712. 1:52:35without AI where he just he just recorded like a loop of himself for like 30 SEC no it was longer than that like 30 minutes like he just
713. 1:52:46recorded like 30 minutes like and then he just played that on a loop in the zoom calls and I think it took
714. 1:52:55people three weeks to figure out he wasn't actually on the calls he was actually watching the zoom calls but just it was his bot that was on
715. 1:53:03camera not him so yeah you'll be able to do that someone will come up with that I've used chat CPT to write VBA code yep
716. 1:53:15I use chat EBT to write python programs and then run them on Android using pyroid 3 super cool I then convert them
717. 1:53:22into Android apps yep super groovy um no experience in Python but a later the Chachi BT can generate code and
718. 1:53:31teach me yeah it really is good at that you can you can point it at documentation and have it um like like using the uh the web pilot plugin
719. 1:53:43point it to the documentation of something that you want to learn and then have it come up with a custom lesson for you
720. 1:53:50and and do it interactively where it it teaches you something and tell it not to move on to the next step until you've
721. 1:53:57completed the first step and you tell it that and it'll just create a tutorial for you and it'll just walk you through
722. 1:54:03it and if you get stuck on something you're like I'm stuck here and they'll say oh well you need to install this Library I don't know how to
723. 1:54:09install that Library oh you do this I don't know how to do that well then you do it like this right so great way to it's it's a great way to
724. 1:54:18learn just what these things are capable of is just pick a project like that it generates working code it's great yep yep ion China
725. 1:54:29I'm gonna dive in MK I was on the co-pilot beta but had no idea how big the ecosystem got now yeah it's
726. 1:54:41it's it's it's big and you know MK one of the things that I think would probably be worth you checking out is um
727. 1:54:50just go to GitHub or go to go to uh Reddit and just dig into the whole llama 2 Community because the the whole open
728. 1:54:58source Community with these large language models llama2 just came out two three weeks ago which is meta's you know base large
729. 1:55:06language model they've got a 7 a 13 and a 70 billion parameter model but they're licensed for commercial use so so
730. 1:55:15because of that the developer Community has flocked to that so a lot of the top trending I don't know what's going on my
731. 1:55:22throat a lot of the top trending projects on GitHub are llama 2 projects so that's worth checking out
732. 1:55:30Mr Kyle Shannon you're home I am indeed I am home to General I recommend Motorola or OnePlus as affordable Android phone obtuse even
733. 1:55:43I don't know what that means someone give me a nice compliment thank you very much what website is that I don't know where I was [Laughter]
734. 1:55:56all right let me I'm going to scroll to the bottom here I'm jumping a bunch of questions but I just want to pop to the
735. 1:56:01bottom I gotta I gotta take off soon it's getting late and my my throat is season up something's going something's
736. 1:56:07going down I don't know what it is man hope I didn't catch the crew all right
737. 1:56:21SS screenshot oh screenshot I can't do it either ever since he got out of tick-tock jail that's weird wait can I screenshot my own thing
738. 1:56:34I can screenshot my own thing you can't screenshot what do you want to buy us are you on an iPhone that's weird what your reply to explain
739. 1:56:49of what a bot is is amazing oh great thank you Firefly I'm glad that helped I need to know how to code and if so
740. 1:56:55what language would be helpful python seems to be the thing of the day and um G uh code interpreter at chat GPT is
741. 1:57:04writing and executing python um Crossland David I don't necessarily know that you need to learn how to code
742. 1:57:14these things are going to do the coding they're not quite there yet but they're close so if you want to learn how to
743. 1:57:20code I'd learn how to code in conjunction with these tools so go to GitHub use copilot it's a pair programming tool
744. 1:57:29from openai that's designed specifically to work with a coder um Firefly is that like your zapier demo
745. 1:57:37you did last week yep that was that was exactly what it was that was that zapier's kind of Bot making the bot just
746. 1:57:44basically means you're sort of logging into something like Discord or slack and interacting is as if you were a person
747. 1:57:55um just tell it you want to Learn Python exactly so do all AI tools make Bots no I I mean think of bots as just you know they're
748. 1:58:10they're just little automations you you can here's here's what's the big difference historically if you'd make a bot
749. 1:58:18you would you would say Okay I want to do some sort of categorization bot and you would say when someone writes an article
750. 1:58:30um I'm gonna I'm gonna pull the metadata from that article that's got a category metadata in it and then I'm going to put the link of the
751. 1:58:38article the title of the article and the category of the article in a spreadsheet and so you would do very kind of manual
752. 1:58:46logical connections right you know bot making and automations was very much like Lego blocks and logic trees and
753. 1:58:53things like that the difference now with large language models with the AI is that you can take unstructured data or
754. 1:59:02inconsistently structured data and just ask it [ __ ] and it will do it for you right so you can just say go read that
755. 1:59:10article and write me a summary or go read that article and write me five tweets or a tweet Thread about that article
756. 1:59:17and then you know let's post that tweet thread to Twitter um that wasn't possible before I mean you could you could do it with with gpt3
757. 1:59:28for a while but but the popularity of these things the the large language models make what you can do with those
758. 1:59:35Bots much more sophisticated with essentially You Know This Much programming you're you're just sort of you go grab a pile of data
759. 1:59:46you throw it over to one of these large language models and if your prompt is clever enough it will throw you back
760. 1:59:53what you need it might throw you back structured Json data that goes into your tool that you know writes a blog post
761. 2:00:01for your company website and it does that twice a day automatically so um and you can do that without a ton of
762. 2:00:07programming so so the difference is the sophistication of the data processing that these tools make possible
763. 2:00:14that's that's why it's such a big deal all right 126 people on board this is crazy it's Wednesday night it's late you
764. 2:00:23people should be in bed what are you doing I got to be in bed my throat's going all right I'm gonna go I'm gonna go oh
765. 2:00:31my goodness the education system needs yeah especially with teacher shortages and post pandemic I think these things
766. 2:00:36are going to be if you're interested in the education system Firefly go check out um Khan Academy which you probably know
767. 2:00:43about it's been around for years the flip the classroom thing three four months ago they released this thing called conmigo okay
768. 2:00:51h-a-n or k-a-h-n um conmigo as in Amigo and it's a gpt4 teacher teaching assistant that will teach you a student how to learn
769. 2:01:06something it's really amazing worth worth checking out videos on YouTube Real Money strategies about chat EBT and making money not
770. 2:01:14showing your face yes or no that's a joke right [Laughter] okay okay here's the deal
771. 2:01:29anyone charging you money to teach you how to make money or making money on the teaching not on the chat GPT now they've probably used
772. 2:01:38chat CPT to write the course um you could start a YouTube channel doing you know text only animations with
773. 2:01:49music and you could completely automate that you could do that right now in Google Sheets with GPT uh for uh gbt for
774. 2:01:59sheets and you could you could automatically generate a giant spreadsheet of copy lines and calls to action and URLs and pop that into canva
775. 2:02:10bulk create and create a bunch of videos so you could absolutely do that um here's the challenge so can everyone else
776. 2:02:20so it comes back to do you have a good idea if you can come up with an angle that people would actually want to pay
777. 2:02:27attention to that stuff great but just know you're increasingly competing with companies and organizations that are
778. 2:02:35going to be publishing 10 000 videos a day doing the exact same thing so what I would say Kathy is the experience of learning how to do
779. 2:02:45that with chat GPT and doing it and putting it out there if you put it in that context not like I want to make
780. 2:02:51money but like I want to learn how to do this then I think it's interesting but again I think that what what they're
781. 2:02:58saying what their approach is is if you automate video production you'll be rich and that's wrong that's like saying if you know how to
782. 2:03:10use Photoshop you're a designer no no photoshop's the tool you got to go to the other side and say what are you doing with that automated
783. 2:03:21video production okay I know that cat people that like to crochet love inspirational poems about yarn and
784. 2:03:33kitties I'm gonna do a YouTube channel About Yarn and kitties great okay perfect there's your idea now which tool are you going to use are
785. 2:03:44you going to go film a bunch of kitties probably not but could you use mid-journey to make a bunch of images of
786. 2:03:50kitties with balls of yarn yeah you could do that and then could you use chat to write all sorts of clever little
787. 2:03:56headlines and then could you use canva bulk create to assemble them into videos yeah you can do all that
788. 2:04:03so so don't mistake the tool for the strategy the strategy is where is how you make your money solving the problem is how you make your
789. 2:04:12money the tool is just the tool so yeah that ranch dressing gotcha yeah exactly it's blue cheese by the way from
790. 2:04:22Fire on the Mountain and it's delicious you can't leave now yeah well actually you said you can't leave now
791. 2:04:30it appears that Apple phones are not able to screenshot and Androids are oh Apple phones are able to screenshot and
792. 2:04:38Androids are not that's weird I use code interpreter with unity 3D I bet that's amazing so Unity 3D is like a
793. 2:04:46you know game development environment you can take code from an existing game in unity pop it into chat EBT and just
794. 2:04:54tell it to change all the shapes to spheres instead of cubes and it'll do it crazy ass [ __ ] all right I'm gonna go
795. 2:05:04I'm winding this down I'm winding this down people I'm doing it I still have a final paper to write I still have a
796. 2:05:09bunch of [ __ ] to do it's 12 a.m on the east coast thank you for hanging out he show you how to no cost wait he show you how to know cost
797. 2:05:18I don't know what that means but I don't charge uh I did front end payment for ClickBank oh interesting that's MK um halfway to her
798. 2:05:32the average person no no Androids can screenshot all right so tomorrow all the Irregulars we're gonna have to figure out what the
799. 2:05:42screenshot issue is and solve it for people because it would be nice if you could screenshot I watched The Tick Tock with scuba
800. 2:05:49divers they hardwire the Wi-Fi link underwater Oh that's interesting wow that's interesting huh that's cool at least I don't have to
801. 2:05:59do that for the average person in the office wait for the average person in the office is to learn it is to learn about AI
802. 2:06:08I don't know what you mean there like what should they learn for the average person in the office right now I would say just like personal
803. 2:06:17productivity [ __ ] use chat EPT to write emails use chatgpt to summarize call notes like when you do a zoom call or a Google meet
804. 2:06:28turn on recording and transcriptions and then take that transcription and pop it into chat GPT or Claude because Claude
805. 2:06:35can take bigger documents and have it summarize your call notes and create a to-do list for every person that was on the call
806. 2:06:43stuff like that just just really start using these tools to to extend what you already do make what you do
807. 2:06:52better faster smarter than you could have done it without it and just start with small stuff you don't you don't have to like
808. 2:07:01you don't have to start programming and doing all this [ __ ] like that's the whole point of this channel is
809. 2:07:06we're talking about chat GPT which is sort of post when the Geeks when only the Geeks could play right chat gbt is the first tool
810. 2:07:16that makes these tools accessible to real people so if you're in an office just start playing and if you're one of the people in the
811. 2:07:22office that's good at this don't keep it secret talk about it say to people oh I did this with chat EBT look what I did and just be careful
812. 2:07:31with you data you gotta if you gotta anonymize your data if you want to be safe in a corporate environment but
813. 2:07:36anyway listen this is fantastic thanks for hanging out with me everybody I'm gonna I'm gonna take off I'm gonna go
814. 2:07:43peace out peace out peace out thanks everyone I'll see you tomorrow