aiLL 8/12/23 - Wild Times Ahead: Generative AI's Impact on Jobs and Society

Transcript

1. 0:00foreign
2. 0:21happy Saturday night welcome to the AI learning lab not the AIL learning lab those would be scammers I even made a chart
3. 0:38we've got charts and graphs charts and graphs welcome everybody welcome welcome welcome welcome happy Saturday night my
4. 0:46name is Kyle Shannon this is the AI learning lab uh uh just had din-din good evening Dr J was shaking was shaking what's going down to bias
5. 1:02date night for you and the misses [Laughter] let's see it's the Barbie movie the AI learning lab hmm foreign
6. 1:20welcome everybody welcome welcome welcome welcome we'll let people get in here we'll get a little settled
7. 1:27I still have my cinnamon from last night but I'm not putting it in my coffee because that was a disaster
8. 1:33plus one also welcome welcome both of you Beretta Barbie nice nice name that's a Beretta Barbie that's a uh
9. 1:43that's a 10 out of 10 on the name 76 of Dr J 76ers gen xer for sure go Gen X oh yeah oh yeah hey Kool-Aid
10. 2:09uh if you'd be so I come off strong if you'd be so kind share the live so we can get some other folks in here tap the hearty hearts
11. 2:19so you know we at least pretend you're enjoying it laughs the Joker's here welcome Joker thank you Dr J
12. 2:37en Jennifer Murray photography hello hello hello Danny D Roger Martin welcome welcome all right now it seems to be popping up
13. 2:52a little bit maybe we're hitting the algorithm Saturday night there's an MMA fight on you could be watching that or you could be watching
14. 3:03the MMA fight that is the AI learning lab what's what is on must be paint it's white and it will not it's making me crazy
15. 3:17I can't get off that famous scene from Macbeth or you could be working what are we
16. 3:32chatting about tonight well laughs no big pressing no big pressing issues I made a couple of videos today um there's a cool
17. 3:45I don't even remember the site there's a cool site that builds you a website it's a cold relume r-e-l-u-m-e
18. 4:02imagine a website now generated it generates a site map and a um a set of wireframes that you can then
19. 4:11export to figma or word uh WordPress no I don't think it's WordPress I think it's uh webflow re-loom so that's pretty amazing oh a
20. 4:24lot of people just join welcome everybody my name's uh Kyle Shannon this is the AI learning lab um let me let me flip these cameras and
21. 4:34I'll I'll talk a little bit about what the hell we do here why are we hanging out on a Saturday night look I
22. 4:42got a little Hatchet Hatchet in the mountains we got the hatchet shirt we're gonna be able to read up here it's going to be fantastic
23. 4:50um let me tell you what's good this is annoying so I guess there's there's good news and bad news the good news is is
24. 4:56that the AI learning lab um has some followers the bad news is one of those followers is a [ __ ] scammer so the top one is me the one that
25. 5:09is has two L's in it at the beginning is the scammer and I've reported them now three or four times in tick tock's just
26. 5:19like yeah whatever they're good but they're actually asking people to sign up for um multi-level marketing and I would
27. 5:28never ask you to sign up for a multi-level marketing program I would ask you to sign up for a pyramid scheme they're completely different
28. 5:36Okay so so if you subscribe to a channel you think is this channel go check and make sure that it doesn't have two L's at the
29. 5:48beginning of it um so that makes me cranky but anyway that's not why we're here why we're here is um so
30. 6:01a little bit about my background I'm a I'm an entrepreneur um my current gig is I'm the founder and CEO of a company called storyvine
31. 6:11that is an automated video production platform that believe it or not does not use a ton of AI we've been around for 11 years
32. 6:20figured out how to do that [ __ ] before before AI came along um we're rolling a lot of AI into it but I I've been an entrepreneur for a few
33. 6:29decades and I've started 12 or 15 companies and I was I was a Founder in the early days of the world wide web in
34. 6:36the mid 90s I started one of the first digital agencies a company Dr J I'm team Ponzi myself that's awesome
35. 6:45I started a company called agency.com which was one of the first digital agencies and the reason I bring that up
36. 6:51is that what's going on with chat GPT and these generative AI models the image generation things and the video things
37. 6:59and the audio things and chat GPT and all these competitors that I have listed here it's very very similar to the early days of the web so
38. 7:09um so what we do on this channel is just talk about it um just you know if you've got questions about it ask questions if you haven't
39. 7:17played with chat GPT chat.openai.com that that URL right there that's the official website just go there start playing while you listen
40. 7:25to me ramble um if you've been playing for a while some of these other sites are are also chat GPT or GP chat gbt competitors
41. 7:35I think they all are good enough that they're at least worth taking a look at see which of these is better for you uh
42. 7:42like for example Claude dot AI you can you can paste in like a 50 000 word novel it's got a hundred thousand token
43. 7:51context window so Claude can be really useful when you might not be able to do something on chat gbt Bing and Bard are
44. 7:58both connected to the internet Bing can also generate images this is Microsoft's version of gpt4 it can read images it
45. 8:06can make images it's connected to the internet so it's kind of interesting um chat GPT the official one has code
46. 8:14interpreter which is quite powerful and it's got um plug-ins and it's got this thing called custom instructions and
47. 8:21gpt4 is just baller it's it's kind of the best of of these things um the perplexity one that's just uh
48. 8:28perplexity giving us access for free to metas llama models uh Pi is from a company called inflection posed from
49. 8:36quora all of them have the little something to offer a little bit different at some point I'm going to start winnowing this list down but if
50. 8:43you're just getting started just start at that top one and then the other thing to do is go to this thing prompts.chat so if
51. 8:53you're super new and you don't quite know what prompting is or maybe you've heard people talk about prompt
52. 8:58engineering or prompt crafting or maybe you've never heard that term at all and you're like what the hell is a prompt a
53. 9:04prompt is basically you're just chatting you're just asking it to give you some [ __ ] hey do this thing for me
54. 9:11prompts.chat will teach you how to do that and it's it gives you a whole bunch of ideas for ways that you can use these
55. 9:17large language models um this next one down is the same thing for images uh it's it's written for stable
56. 9:24diffusion but the basic principles here apply for all of them and then futurepedia.io is uh just a big long
57. 9:31list of tools there's like 4 000 AI Tools in there you can geek out go nuts just know that most of them are crap and
58. 9:37they're all probably trying to ask you for 20 bucks a month so I would not sign up for any annual concept tracks right
59. 9:43now everything's moving too fast all right find a little piece of the puzzle here um if you're if you want to join a
60. 9:50community doing doing uh this sort of thing uh the salon.ai that's that's a group I started back in December called the AI salon and
61. 10:00um we have we meet every other week we meet in person here in Denver we meet uh you know virtually for those not in Denver
62. 10:09um the next meeting is next Tuesday we meet every other Tuesday and uh John Knack from Adobe speaking from their AI team he's he's uh
63. 10:18he's uh they're doing some really interesting stuff over there and he spoke at the salon last January and he's
64. 10:24going to be back Tuesday night so that's exciting um what else yeah just go to that URL and sign up for that
65. 10:36um the uh YouTube channel up here AI learning lab Dash TT Dash Tick Tock that's where Archives of these lives are going to be stored
66. 10:45and uh yeah I've started working on that I I spent like an hour and a half today trying to get the [ __ ] prompt to work
67. 10:53to generate chapters consistently and I just couldn't do it and one of the things that I learned about Claude dear
68. 11:01old Claude up here even though you can put in giant documents it limits you to like 20 or 25 um queries a day I probably if you're
69. 11:10using long documents um which I was trying to do some testing and so I just burned through all my uses so that made me cranky I was like hmm
70. 11:24welcome to bear what's happening how are you doing all right the other thing I do is is ask questions so so please feel
71. 11:34free to pop any questions in and I will do my best to answer them let's see what do you need to participate in the
72. 11:42Tuesday Salon camera microphone Etc you can it's just a Google meet so so if you go to the uh if you go to this URL
73. 11:51um you can you you know join the Meetup Community or join the Meetup Group and then there's also a Discord group so
74. 11:58you can you can go to the Discord group and at the Discord group you can play with mid-journey and chat gbt and you
75. 12:04can ask questions and things like that so that's that uh storyvine's my company and that's what that is so you don't you
76. 12:12don't need to even be on camera you can be you don't have to be for the for the salon we do like to introduce ourselves
77. 12:21so it's kind of it's kind of like a support group it's like it's like it's like it's like aiaa hi my name is Kyle I'm into AI hi Kyle
78. 12:37uh and then what about these lives what do we do here um I ramble um I ramble I answer questions I get snarky
79. 12:50I get sassy I cuss occasionally so if you don't like that sort of stuff get the hell out what is wrong with you swearing's fun
80. 13:01swearing's knee I'm a large language model and I don't have any safety guard rails yes okay let's do this let me get to some questions
81. 13:16hello hello hello a lot of people saying hello hello Ellie hello all right uh I used an uploaded photo of a vehicle
82. 13:27used it in mid-journey with prompts prompt was a disaster not accurate at all thought it would at least keep what was uploaded
83. 13:37yeah it doesn't it doesn't really do that um I don't know how geeky you are Tobias but if you um if you want to explore
84. 13:48that there's there's some new things coming there was a paper that came out like a month ago where they figured out
85. 13:53a way to essentially Auto trace a source image and pull out basically capture all the detail the the detailed Line work
86. 14:02and then you can modify everything else um but I haven't seen that built into anything yet um the thing that I would play around
87. 14:10with is control net um so control net is one way you could do it but you could also try the remix um the remix capability in mid Journey
88. 14:22rather than just uploading a photo and trying to do it um trying to think what else
89. 14:36yeah I don't know I I don't have a good a good suggestion for you I was thinking maybe you could use the describe
90. 14:43function in mid-journey have it describe your Source image and then use the source image as part of the prompt and the description that it
91. 14:52selected and then change whatever you want in that I don't know if that's going to work I would I would look at
92. 14:58control net for stable diffusion um if if you want to start to get some things that are more predictable and
93. 15:04you're main you're maintaining the composition and the look of the car uh and then if not it just it might be
94. 15:10too early for that now the other thing that you could do with an image like that Tobias is um if you've got Photoshop and you use
95. 15:19the Photoshop beta with generative fill you could select the car and change out all the background or you could select
96. 15:26the car and change the color things like that so there might be there might be there might be some ways you can do it
97. 15:32but it's it's that's that's a tough one get it and getting things super predictable in mid-journey and stable
98. 15:38diffusion right now is a [ __ ] show so yeah welcome to the uh welcome to the early days we sure are early here at the internet
99. 15:47at the dawn of the generative AI era it's the great Renaissance and right now it's like a great pile of poo poo so just
100. 16:02it'll get better at some point the predictability things are real pain in the ass I saw there was a cool um I
101. 16:12saw some video of a uh or not video I saw like a Twitter post of a uh I think it's open ai's new model so open
102. 16:23ai's Dolly the the like the image generation thing within Bing is really shitty it's just bad it's just doesn't
103. 16:30have a lot of controls it just gives you square images it's not all that good um but apparently they've got one now
104. 16:36that does texts really well like one they have internally so I have a feeling open AI is going to come up with with a
105. 16:43mid-journey competitor that's gonna I have the the Xmas tree is going to be piled with technology
106. 17:00I want a new large language model and I want to admit journey to not suck I don't want to do it in Discord there's a lot in 10 sign up what
107. 17:18there's a lot in 10 sign up morgie I don't know what you mean wait was that in reference to something earlier
108. 17:27I don't know what that means I do not know what that means I'm not that's uh not my it's not my uh first uh the the
109. 17:34words it's not my first language there's a lot intense on them they're not now AI learning lab is the real one yes
110. 17:47oh AI learning lab with two B's Joker joking it up I like it it's solid it's solid it's good okay I think I asked last time but why didn't
111. 17:59you get to my question what do you think about adobe's AI photo um I probably just didn't get to your so what happens
112. 18:04Jennifer if you came in even an hour in because I try to answer all of the questions if you're down at the bottom
113. 18:13sort of late in the thing I get like a hundred questions behind so you were probably just you would probably just swept up in the
114. 18:21uh Gen X are trying to use tick tock um situation um I like generative fill a lot So when you say when I when you say do I
115. 18:35what do I think of adobe's AI photo there's there's a couple of things there's Firefly which is kind of their playground and there they have
116. 18:47um uh you know text to image and they've got the the stylized text thing and they've got one other thing now then they've got
117. 18:56generative fill um both of them are interesting I think generative fill is is actually really quite good so let's go play with that
118. 19:06do I have layers on this they do okay well let's go find something fun let's go find a fun image oh you know what we'll do
119. 19:17I know it'll do no what Kyle what what are we gonna do you know what we're gonna do we're gonna go to I'm now on the AI Salon we're in the
120. 19:28we're in Discord now so if you haven't used Discord these are all servers so these are all different communities this
121. 19:34is the AI salon and then these are all conversations within the AI Salon community and I'm going to go to the mid Journey Channel
122. 19:45and let's just go find a cool image that one's neat who's doing that Kyle WB Kyle WB it's not me that's someone else
123. 19:57those are sweet those are neat um let's see 3D optical illusion this is cool
124. 20:18wow nice that's nice too oh that's a source image yeah yeah he so so um Tobias here's a guy that uploaded that as a source image
125. 20:34and then he got he got this is outputs which is pretty close but it's not exact right like if you want to keep the same lines of a car
126. 20:43like this has got completely different Towers this has got an archway that this uh well this has an archway but it doesn't have that Tower
127. 20:52it's got a tower in the middle so yeah so it it really does reimagine them but it puts you in the neighborhood so
128. 21:01so yeah if you want to if you want it to be specific okay so let's see what do we want to do what do we want a picture of Let's do
129. 21:10okay so the way you make images with mid Journey you type slash imagine when you're in a channel that's got mid-journing installed
130. 21:20and then it it gives you a prompt it'll say um let's see [Music] um
131. 21:35photograph photograph of a fly fisherman standing by a stream in the fall with the leaves changing
132. 21:49color and then I'm going to do dash dash AR and I'm going to do two by three so this is going to be taller than it is wide
133. 22:04and so now we'll have not now you know mid-journing is going to go off and it's going to make this beautimous thing for us
134. 22:21and Jennifer um I don't know if you were here earlier I think you were when I was talking about the AI Salon my co-host in the AI Salon
135. 22:28is a woman named Leah fasten who's a commercial photographer out of Boston and so she's been playing a lot with
136. 22:36this AI stuff so you know it would be in potentially interesting for you to come to the salon just to talk to her or you know meet her
137. 22:48all right what's it doing here it doesn't look like the fly fisherman that's okay are these done okay so the way mid Journey works it
138. 22:56gives you four options for any given image and then if you find one you like I kind of like this one it's a little bit more dramatic
139. 23:02in the Stream with and he's off in the distance a bit more although this one's kind of Moody it's got It's got some stuff here so with mid
140. 23:10Journey you've got you one through four and V one through four that those um are related to these four Images
141. 23:18obviously you means upscale and V means very variations so if you see one you like but it's not quite perfect you can hit V4 so we'll do
142. 23:28that now and now it's going to render four new images kind of based on that one so it'll take the same seed of that
143. 23:35image and it's going to give us four variations of it and this this is often if I'm if I'm putting together an image for an article
144. 23:45or something that I'm writing and I just want to knock it out quick what I'll often do is I'll type in two or three different um
145. 23:54spins on a prompt that are all in the same neighborhood it'll generate four four from each and then I'll just do
146. 24:00variations on anything I like that came out of that and then I'll kind of pick from those and I'll just upscale them and then most
147. 24:08of the time within I don't know two or three iterations I can get something that that I like um let's see this top one kind of looks
148. 24:21the most like a fly fisherman does it I don't know it doesn't really matter okay let's grab yeah let's grab this so so here's what
149. 24:32I'm going to do I'm going to hit upscale number one and so that's going to give us a big version of this then we're going to hop
150. 24:39over to photoshop [Music] okay so there's our picture it's a
151. 24:55lovely picture um now I'm gonna go show you how to do this in generative fill but you can also do this in um
152. 25:09in mid-journey actually I'll show you both why not so um so we've got a vertical image here now let's say I I was doing a piece and I
153. 25:19actually need this to be horizontal and I you know I want to see this this dude fishing but I need I need a
154. 25:27horizontal image well I can crop it but you know then I'd kind of lose the the beauty of the scene so
155. 25:33um once you up upscale an image it gives you a bunch of options here you can do variations strong and subtle you can
156. 25:41zoom out two times which will basically like put him you know in the distance 2x right you'll get two times more stuff
157. 25:48around him you can do a custom Zoom where you could turn it into a 16x9 image but they also have these these
158. 25:55panning things so what if we wanted to create an image where we're going to put some copy on the page so we want the
159. 26:02fisherman to the left and we want you know some stuff out to the right well we can actually do that we just hit this right button here
160. 26:09and what that's going to do is it's going to add to the right of this image so this is similar to generative fill so
161. 26:15we'll see what this does here and it's going to give us four options here and then we'll go do the same thing in generative fill so let me
162. 26:24pop this up open it in a browser copy this image go over to photoshop say new paste all right so there's where we'll
163. 26:38start there so let's go back to mid Journey hey mid Journey how you doing tonight huh I'm fine how are you uh I'm all right mid Journey
164. 26:52oh mid Journey you're the best except for the fact it's still in Discord am I right get get it go and make a site make a
165. 27:06site where we can go make pictures in a pretty website I don't have to go to Discord I hang out with all the nft Bros and they make me crazy
166. 27:16all right so let's see so how did this do I can't quite see the edges of these so let's just grab one of
167. 27:26these that looks decent they're all kind of the same these have this cool little waterfall off to the right
168. 27:35so let's go grab we'll go grab four upscale number four I think I want to purchase the paid version but I'm not sure if it's worth
169. 27:49it oh for for for Adobe um if you're if you're using so so this is still Jennifer um I so so if you're using
170. 28:03it for clients and if you're using it if you work within a business um well it sounds like you're a photographer so so you're I assume
171. 28:15you're freelancing you you're probably using this for clients here's here's here's one of the big reasons to use
172. 28:20Adobe over mid-journey is Adobe hasn't been trained on um copywritten content copywritten images it's been trained on public
173. 28:32domain images and Adobe stock images that they've got access to Oh all three of these added fishermen in there that's funny
174. 28:41all right well we'll upscale number four actually that's kind of interesting we could we could uh we could pop that fisherman out of there
175. 28:52um but there's what it did so here's what we started with and then it added all this stuff to the right so so that's that um
176. 29:01and then over in Photoshop it's it's kind of the same thing so I'm in the Photoshop beta and so we've got this
177. 29:08vertical image and I can do the same thing here where I grab the crop tool and I can say I want a little bit to the
178. 29:19left and I want a lot to the right and so now I have you know that's kind of the image area that I want so what
179. 29:29I'm going to do is I'm going to select and I just I heard what you said that you do this all the time so we'll do this for other folks
180. 29:42let me select this so what I'm doing now is I'm selecting the image and now I'm going to invert that selection
181. 29:55and what happens in in the in the beta is that you've got this this contextual menu that pops up when you select
182. 30:02something and so now I'm going to say generative fill and I can put a prompt in here so I could say you know that I want you know
183. 30:12just more stream and more Hills or I can just let it try to figure it out this is a pretty clear image so it tends to do
184. 30:18good with this kind of this kind of imagery so we'll see what it does I'm suspecting that photo Photoshop will
185. 30:26not put an additional dude in there where mid-journey did um that's not quite right because it gave us a whole second stream it's like
186. 30:38he's below an island right here um that's one option there's another option that's closer but they're all kind of bad
187. 30:50all right so we'll try another one we'll do um
188. 31:15select inverse and then we're gonna do generative fill and we'll do um let's see um steep Hillside and we'll generate that we'll see if we
189. 31:30can get it to to give us a steep Hillside instead of a second stream foreign and I will get back to your question oh
190. 31:44that's kind of cool that's much better um that's even better that's even more better right we got this is where the
191. 31:51stream used to come down through here and I've got these big round boulders so you get you get three different
192. 31:57options that one still looks like a second river all right that that's the one right so that's the one and and you
193. 32:02know what I might do is I might do something like um like that's too bright if I want to put text over that that's too bright so I'm
194. 32:11going to grab that I'm going to highlight that selection and I'm going to say um dark Moody um fall uh trees or something like that
195. 32:26and hopefully that'll darken up that section and then we could use that for our our little ad campaign for our backpacks yeah perfect
196. 32:36right so now we have a nice dark area over there that's a little too busy that's a little too busy so the first one's good so
197. 32:46there you go so so in this case um Photoshop in my opinion did this better than mid Journey did it right it
198. 32:55filled it out nicely here it extended it nicely there and with a little bit of prompt prompt magic we did something in there um
199. 33:04let's try something fun um I'm gonna say generative fill um black bear eating
200. 33:20a trout let's see see if we can get a a bear in the image yes two they're too dramatic but anyway that's funny all right we'll get
201. 33:33rid of that all right that's generated Phil so back to your question um Jennifer uh you know photoshopping cheap but the the
202. 33:45generative fill stuff and the access to Adobe Firefly um because that stuff is um is is commercially viable like the
203. 33:57fear I have about mid Journey even though it's beautiful is because it's been trained on um other people's work especially if you
204. 34:05use other people's artist names in your prompts um it could that might be problematic like we don't know where the copyright and
205. 34:14all the lawsuit stuff is going to land um so so Adobe is really safe for that and and you know they're Adobe they're
206. 34:20gonna get their [ __ ] together the fact that they're leaning into this so hard so when when John Knack spoke at the AI
207. 34:26Salon back in January they were not clear that they were going to go into this world because back in those early
208. 34:33days the artists were pissed off they were not happy about this [ __ ] at all um so when they went into it they did it
209. 34:40you know training training on um images that didn't violate anyone's copyrights which is pretty smart so yeah
210. 34:47I think I think they're worth uh worth playing all right I'm learning ux uh oh uh relum is making me pretty sad yeah
211. 34:57I know Timothy Brooks so here's the deal everybody everyone on the uh on yield uh learning lab on the live tonight this is
212. 35:06this is part of the [ __ ] we're all gonna have to confront like I've been confronting it with my video business for a while now that
213. 35:18anyone in the knowledge or creative or you know technology World any any jobs that are computer mediated and a lot of
214. 35:27them even that are tangential to that all of our jobs are going to be completely redefined and some of them
215. 35:41some of what they become we might not like and some of what they become might mean companies don't need as many of us for
216. 35:51that particular job like they're still going to be writers but there might not be as many of them and what those Writers Do might not be
217. 36:00what they do now and same with ux designers and same with engineers and coders and consultants and like I don't
218. 36:09think anyone's exempt if you touch a keyboard you know if you touch a computer in any way if you work in a call center
219. 36:21everything's going to be redefined and you know a lot of the the more tactical execution kind of jobs and a lot of the jobs that are about
220. 36:31managing complex systems like like that reloom site that automatically makes a site map and then wireframes of the
221. 36:41websites just continue that out so so it'll say here's your site map here's your wireframes here's your designs
222. 36:51and then those are all tied to um you know code on the back end right so so someone's gonna Stitch that entire workflow together
223. 37:01and So within an hour you'll be able to create a very sophisticated website effectively from a single prompt but here's the thing
224. 37:13a good web designer a good ux person even if that stuff is auto-generated is going to be able to look at what it
225. 37:19generates and go oh that's crap oh there's a better way we could do this let's modify this thing and let's let's
226. 37:26move these things around on the site map because they don't make as much sense right so there's going to be just like
227. 37:31if you've ever used chat CPT you can get it to 80 in like five seconds getting it to a hundred percent is almost impossible
228. 37:42it's going to be the same thing with things like relum or whatever the tools are where you know you're you're a musician and you're like
229. 37:51I love handcrafting music and then Along come this comes this tool it's just like squirt you know here's here here's your
230. 37:58exact style you know song after song after song every three minutes well you're going to be able to use that tool in a way that a
231. 38:06non-musician whoa so I think there's going to become superstars of figuring out the new way to do the old thing right and then
232. 38:17there's going to be a lot of people that are just like I don't want to use AI that's not what I got in this for and
233. 38:24they might not just be whining like that like they might be really sad because because you know there was a there was a uh a post on Reddit
234. 38:35three or four months ago about a 3D artist he worked for a 10-person company and there were two 3D artists in his company there was him
235. 38:46and there was a guy that wasn't as good as him and then this AI stuff came along and he said he went from being a 3D artist to
236. 38:56being a prompter and he didn't like the prompting he liked building models thank you thank you for the heart
237. 39:05and the guy that wasn't as good as him loved the prompting and saw the opportunity with the boss and saw a tool that could make him better
238. 39:16and so the guy that wasn't as good jumped on this [ __ ] became better than the guy that was good or I don't know if betters right more
239. 39:25productive and more in the boss's favor right because they were getting [ __ ] done faster like instead of it taking two
240. 39:35months to build a model it was taking him a day and a half and the post that he wrote was just this Lament for
241. 39:45his job evaporated overnight and he was just [ __ ] sad and he quit like he quit the company he's like I don't want to do like I
242. 39:55don't want to be a prompter that's not my thing and so I think every every one of us is going to confront this in some form or another
243. 40:07and it's you know there's a piece of it that's really sad and then wait for it I don't know how long the painful
244. 40:19period's gonna be my guess is two to three years it could be five years but on the other end of the painful
245. 40:24period where there's all this just massive disruption and and and and facing all of these issues these tools are going to create new jobs
246. 40:37that we can't even imagine today that are they're going to be [ __ ] amazing right because they're just going to be
247. 40:44at such a higher level and they're probably going to be multimodal like I I feel like people that get good with
248. 40:50these tools are going to be kind of like Symphony conductors where you've got all these tools and and you're kind of Oz
249. 40:55behind the curtain and you've just got this panel and you're like I've got a vision for a company and a website and
250. 41:03software and a marketing plan and you know let me take a week and launch that company like that's gonna there's gonna be
251. 41:11people that are like that right they're just gonna be masterful at this [ __ ] who's who's that gonna be it's going to
252. 41:17be someone that right now goes instead of just dealing with the sadness of losing what was lean into it and get curious about
253. 41:27it and go oh what could be right now is the time that if you lean into what could be with this stuff
254. 41:39get curious about it think of it like an adventure you know when you choose to go on an adventure you choose to go on it even if you're
255. 41:50scared even if you know there's going to be shitty things even if you don't have the confidence that you can go
256. 41:58slay the dragon or find the treasure or whatever the [ __ ] your adventure is but you do it anyway because there's a higher sense of
257. 42:08something and where my head is right now is this shit's coming whether we like it or not just it just is it's not going to get
258. 42:21uninvented and it's not going to get worse these tools today are the worst they're ever going to be anyway so yeah I know I'm with you
259. 42:34Timothy Brooks I get it it's you know it's funny I was I don't know who I was talking to the other day I forget who it was but but I was
260. 42:46talking about we were talking about covid and lockdown and working remotely and and I was talking about
261. 42:54isn't it ironic that that Zoom of all companies just said that they're you know they're going to ask their
262. 43:00employees to come back to the office now it's only two days a week and optional Fridays thanks Lauren um but
263. 43:13I you know I feel like I'm I'm in a bit of mourning for my own company my own company storybine where we're a small company but we we have
264. 43:22a lot of introverts and we have me and I'm not and so I get my energy from being around other people and the other people in my
265. 43:32company don't and so we went into lockdown they were [ __ ] they were living their best life they're like this
266. 43:37is awesome I don't need to deal with people and I can get [ __ ] done and you know what it's what it's ended
267. 43:44up doing for me as a founder of the company and and as someone who who I don't know I like there's something
268. 43:52about Humanity that feeds me right um I feel like my company in a lot of ways has become a to-do list with zoom check-ins
269. 44:03and that really bums me out and so I've been kind of mourning that for a while and then I think this stuff on top of
270. 44:10that is is kind of a second kind kind of um you know death change that we're all going to have to deal with and
271. 44:22you know so so we're still trying to figure out what work looks like you know now that now that covet has sort of you know
272. 44:30blown through Society blown through the global Workforce and and done whatever it's done um and and this is going to be the next
273. 44:39sort of wave that's going to change a lot it's going to change probably more than covid did um sorry I'm adjusting my little stand here um
274. 44:52it's gonna it's gonna it's gonna make some messes all right for a physician that wants to pursue a career in generative AI with no coding
275. 45:05experience what next um there's so many possibilities so I would get I would get really educated there's there's um
276. 45:18Google just came out with a large language model called medpom med-p-a-l-m or Med Palm two I don't whatever is the latest Med Palm that
277. 45:28Google came out with that's a large language model that's trained specifically on you know medical stuff and and you know
278. 45:36medical data and it's multimodal so it can do imagery and things like that I would get educated on on the tools that are out there that are
279. 45:46that are specific to to that category um I'm trying to think I just heard today someone just came out with a thing I
280. 45:59think it was on the the AI marketing the marketing AI institute's podcast this week they were talking about
281. 46:06um I forget who the company was that came out with it but they used the medpom large language model and created
282. 46:11a thing I think called Palmetto let me see if that's right I don't want to run you down a rabbit hole um Palmetto AI um
283. 46:36let's see Medical page now found that's not right
284. 47:01nah I don't I can't remember what it is so anyway so I would just I would start paying attention to um anything anything medical related
285. 47:10that's going on with AI um you don't need to know coding um I would start to I would start exploring um things like pie so so pi
286. 47:22dot AI if if you go to the App Store and you look for pi personal assistant you'll see the app for pi or you can go to a website Pi dot AI
287. 47:32that's from a company called inflection and it's a it's a chat bot that's very chatty it is not super it's not
288. 47:40efficient like if you want to go do work you know use these top things but if you want to go have a conversation with a large language model
289. 47:47and you want to start get a glimpse into the future of what it might be like to interact with these virtual um agents or or assistants
290. 48:00it's a really good thing to experiment with so I I would download the iOS version of that app you can put it into
291. 48:07speech mode and just start talking to it and start asking it questions I've got kids with chronic lyme disease and so I
292. 48:15was asking it about Lyme disease and treatments and things like that it knew all the stuff it knew about babesia and
293. 48:21Bartonella and just just all the stuff and so and that's not even been optimized with something like Med Palm
294. 48:28underneath it so I I would think that as as a physician you're going to be in a in a a powerful position to understand what
295. 48:40worked in the old model and what didn't and be able to see opportunities to make things better so make things better
296. 48:48might look like better bedside manner it might look like extending um caregivers and hcps reach so imagine a physical therapist you know
297. 49:03a physical therapist can only see so many patients in a week and you know the the patients they have might not be able
298. 49:11to get in touch with them if they have questions and things like that imagine that physical therapist being able to say
299. 49:18you know here's your plan for the week and by the way here's the virtual me that you've got 24 7 access to that you
300. 49:26can ask questions to it knows all of your record it knows all of what we've talked about so far and if you've got
301. 49:34anything going on that requires my attention the virtual me is smart enough to know that and we'll forward you know your
302. 49:41message to me and the next time we talk I'll know what you talked about right and that you know something like that could be there um
303. 49:50these things are going to be amazing for simplifying complicated systems I know there's a lot of companies working on that but you know
304. 49:59you could you could find a startup that's working on simplifying you know patient records and and the amount of
305. 50:07time that nurses and doctors spend doing [ __ ] paperwork so the insurance companies will refuse to pay them not that I'm bitter
306. 50:18but I'm [ __ ] bitter that I think a lot of doctors are too it's why there's so much goddamn burnout
307. 50:27so here's here's the thing about this AI stuff is these are just tools these are not the solution these are just tools the solution is you
308. 50:39with your expertise saying here's a problem that needs solved I the old way I would have solved that is X Y Z
309. 50:49let me go learn enough about what these things are and how they work and what they're good at and what they're not good at
310. 50:56that I might be able to take some small piece of this problem and experiment with solving it with these tools
311. 51:05and then what you might find is holy [ __ ] it's amazing at that or what you might find is is oh it's not really
312. 51:11ready for prime time and you know it's it's probably somewhere in between the two um and so I wouldn't don't worry if you're
313. 51:22not coder who gives a [ __ ] we can we can find coders and and if you really want to learn coding this [ __ ] will teach you
314. 51:28coding and it'll do a lot of it for you if you're geeky and like geeking out on [ __ ] great it'll teach you how to do
315. 51:34that you can go do that if you have no interest in that great have no interest in that go solve the problem
316. 51:41right take your your knowledge as in working with the system and and do that with it all right hope that helps
317. 51:51what do you need to participate in Tuesday's Salon uh just uh it's it's a it's a Google meet you just need a web
318. 52:00browser that can do Google meet so it's you need a camera if you want to be on camera you need a microphone if you want
319. 52:08to talk but there's even a dial-in number you can just dial in and listen Claude cut me off oh wait hang on oh oh
320. 52:15Tick Tock just bounced me to the bottom oh tick tock you make me crazy happy Saturday night everybody looks
321. 52:25like we've got some new folks here my name is Kyle Shannon this is the AI learning lab we're learning about labs no
322. 52:32um also I'm gonna pop up here so uh we're popular enough now that we have been officially um cloned by a [ __ ] scammer
323. 52:46so if you get any messages from AI 2l's learning lab that's a scammer the word lab only has one L in it and
324. 52:57because it's i l you know and and it's you know if you have lowercase L's it's like ill looks very similar to Il
325. 53:05so just be careful the scammer is a i l l learning lab and then or I mean learning only has one L in it so so just be careful of that I
326. 53:17I wish that wouldn't have happened and tick tock's not dealing with it right now so I don't think anyone other than
327. 53:24me is allowed to report them for that because tick tock um I've reported it now four times and they're like we don't see anything wrong
328. 53:31with it well they've duped all my [ __ ] content and they're telling all my people to sign up for a multi-level marketing piece of [ __ ]
329. 53:42all right welcome if you have questions about AI popping below um uh what was I gonna say I tried to answer them all so I get very very far
330. 53:56behind on these so if if I haven't gotten to your question or if I haven't in previous things don't be upset with
331. 54:03me don't hate me don't hate me because I'm beautiful don't do what's good about the hair tonight I haven't even paid attention to
332. 54:09it it's all right it's not bad tonight it's not it's not a terrifying
333. 54:25the bear hey everybody hey Mr cgo what's happening the bear is referencing my title I have a title of Chief generative
334. 54:32officer for an organization I do some work with which I think is the first Chief generative officer in the history
335. 54:39of the world ah take that and then you might ask what is a chief generative officer fantastic question Bob tell him what
336. 54:49he's won you can make money watch every day uh about um Claude cut me off today too yeah that bugged me because it was like
337. 54:59you've got five questions left you've got four questions left and I'm like just don't [ __ ] up my question and I
338. 55:05won't keep doing this thank you the bear that's my dangly lights I like the dangly lights whoa the bear reached member level 11.
339. 55:15the bear just turned it up to 11. holy [ __ ] I have no idea what that means but I like it we got some spinal tap going on in here
340. 55:30that's unusual I locked I talked to Claude all day at work I hope it's not a new thing so cuppy I think what's going
341. 55:36on with Claude I'm taking the transcripts from these lives which are like the one I did the the one I was
342. 55:44experimenting with was a three hour and 18 minute live so it was a transcript from three hours and 18 minutes of me just rambling
343. 55:53um so so it's a really long input so I have a feeling it's number of tokens so if you're just doing short back and
344. 55:58forth with Claude it's probably not it's not gonna rate limit you and and I'm trying to do like I'm what I'm trying to do is I'm
345. 56:07this is I can I can show you what I my prompt is insane for this thing what I'm trying to do is have it read a three hour transcript
346. 56:15and then on average pull out six interesting things I said per hour as chapters and create a YouTube title A YouTube
347. 56:27description and chapters with time code on them and like it'll do it perfect and I'm like awesome I got it and then the next one I do
348. 56:39it it'll do like you know six quotes from the first 15 minutes of the thing and ignore the rest of the transcript
349. 56:46like ah and then I try to change that up and then it just messes something else up and it just so it's just and and so
350. 56:54where I ended up getting was I said okay let me just generate the chapters first and not worry about the title in the
351. 56:59description and so I was sort of halfway through getting the chapters done and that's when Claude said go screw yourself and then I went to
352. 57:08code interpreter because I thought oh I know I'll pop it into code interpreter in chat GPT and it can read the long
353. 57:15file and it's gpt4 and it's going to be awesome and then it just kept writing python code and it was like I can't
354. 57:23understand the time code and this is in RTF and then I uploaded it as a text file and then it could sort of read it
355. 57:29but it could sort of do the math and sort it like I don't know it's just sometimes these things are you make it a little crazy
356. 57:39but here's the thing when I solve that like like going through the the the nightmare of learning how to use these tools right now
357. 57:52like what it's what it's training me to do is just think very very critically about how you use words to pull interesting magic out of these
358. 58:07things I don't think that's going to be super critical long term I think the machines are going to start prompting themselves
359. 58:14and get really good but in the in the short term it's it's really important and I you know it's also it's also fun
360. 58:21and really gratifying when you get it right because here's the deal when you get those prompts right and you can sort
361. 58:26of lock that [ __ ] down and then you can just start popping in long form YouTube transcripts and getting out really good
362. 58:33answers like I can go back and you know do a whole bunch of content generation which is really exciting all
363. 58:40right you will take on the Persona of an I.T entrepreneur that is first enough in AI to teach tick tockers that's what I've done
364. 58:49you know it's funny it's like like I have a sort of I have a choice with this channel which like this is a really
365. 58:56interesting kind of dilemma as we build community here some of the people that have been here a lot and come a lot if I
366. 59:06if I just keep with the beginner stuff then it it kind of loses its value and and then but if I make it to advance the
367. 59:14new people will come and they'll be like I don't know what the [ __ ] he's talking about and so I'm trying to sort of skate
368. 59:19between those two worlds what what's why I think it's okay to keep focusing on the relatively like what if you've
369. 59:28done nothing with this go start with chat EPT here's the tools is this stuff's changing so fast that three
370. 59:34months from now if you check in with me even though I'm still doing the beginning kind of stuff I will likely be
371. 59:39talking about different tools different techniques different challenges different news items that are like you
372. 59:45know holy [ __ ] this company just laid off 50 of their staff or whatever it might be and and you know so so I think
373. 59:52there's I think there's lots for us to to um to talk about even if even if this doesn't end up diving deep and and it
374. 1:00:01number of folks have asked me you know do I have a workshop have I thought about doing a workshop I think I might I
375. 1:00:06might do a workshop at some point I've kind of mentally made that decision I just haven't quite decided what I want
376. 1:00:12it to be so if I do that I'll let you all know obviously all right where do I find those information I don't know which
377. 1:00:18information you're talking about any cinnamon in that coffee no this is um cinnamon free but I do have cinnamon
378. 1:00:25here but I'm not putting it in there because it made a mess last night I had like a cinnamon mustache it was it's not good
379. 1:00:33I don't have enough milk in that coffee Chris what's happening car Bruno I love seeing people coming back this is really fun
380. 1:00:45um I need an AI assistant that helped me remember things disabled veteran requesting a suggestion um
381. 1:00:59there's not a there's not a good short-term answer right now ah crap there was there there is the [ __ ] was it called
382. 1:01:15um first of all thank you for your service um there is if you just hang on for another I don't know three or six months
383. 1:01:30there's gonna be tools coming out that do this so when Apple releases their um their AI I have a feeling it's going to have memory
384. 1:01:40um I assume at some point open AI is going to do it um I definitely think that Pi's gonna do it the inflection company here
385. 1:01:50there's a there's I I there's a company called or there's a project called Lang chain where you can build these large language
386. 1:01:59models so that they have memory so and I I know that there's a project out there hang on let me see if I can find it because
387. 1:02:09wait I got an email about it hang on hang on hanging let me see let me see well let's just go look shall we in that special aren't you
388. 1:02:24just terrific all right um pop some pie it's a voyager Runway let me see AI memory
389. 1:02:45recollect okay I found it
390. 1:03:02the caveat here is I don't know if this is any good but try read Dash collect dot AI and let's go look at the site hang on
391. 1:03:16read Dash collect dot a i accept invitation so you you have to apply for the wait list
392. 1:03:32um
393. 1:04:02oh it already exists I already have an account I don't know if this is any good but I think this is what you're talking about start
394. 1:04:13okay let's watch this together the founder and CEO of Records I'm really excited today once you've completed your onboarding a
395. 1:04:19hobby extension installed then I recommend you pin it as you browse online will automatically bring in most articles for you and you
396. 1:04:25can intentionally bring in things like PDFs as you read you can also do things like highlight information right click
397. 1:04:31here and click our extension and that will bring in those highlights for you in this article and you can also add
398. 1:04:36notes if you'd like if not simply click off the extension here if I click it will show me that this article has come into
399. 1:04:43my account with this check mark nah this doesn't seem right this doesn't seem like what you're asking for um
400. 1:04:49I don't know keep keep checking back here I there so the the challenge that you're facing my my dad had Alzheimer's was was an Air Force vet
401. 1:05:01um the the challenge that you're facing this AI stuff is going to be brilliant for we just haven't quite gotten to the
402. 1:05:08point that the memory piece of these AI tools is is not there right now so like within a within a um chat GPT session it'll have memory of
403. 1:05:20that conversation up to a certain point but but you know it doesn't have memory across them um that stuff's coming it's it's it's
404. 1:05:29coming it's just not there right now so I'd say just be a little patient with it I'll keep an eye out for it because
405. 1:05:35because I think it's a I think it's a fascinating Challenge and I think it's going to be so [ __ ] empowering it's
406. 1:05:41going to be like a super power um for anyone you know my wife had a brain injury so anyone who's got you
407. 1:05:49know memory issues and challenges uh with that stuff I think this stuff's gonna be incredible for that the the tools that I
408. 1:05:58know of are not there yet but they'll you know they will be here at some point um I like making chat gbt pretend to be
409. 1:06:09Shang Shang Tzu from Mortal Kombat and say his catchphrases oh yeah it's good for that it's it's great for party
410. 1:06:15tricks actually pie is really good at that um what did I ask pie I said to pie I said I wanted to learn something and I
411. 1:06:23wanted it to teach me something by inventing a game and it came up it invented like a trivia game and then it
412. 1:06:29started asking me questions and it was actually really interesting and and you know that that starts to point to to a future of
413. 1:06:38hyper personalization where it's like I'm not just gonna like learn if I want to learn a new language you know I could
414. 1:06:45go buy the off the shelf you know language learning system that someone invented but I could also just say hey I
415. 1:06:53want to learn Spanish and I want to I want to learn it you know talking about um I don't know you know fluffy kittens
416. 1:07:02whatever the hell is you know with you know cultural movie references and and I want you to you know turn every lesson
417. 1:07:10into a game and it'll just start doing that and you know going back to that memory thing these interactions
418. 1:07:18are going to start being recorded So these agents that we interact with are going to remember what we talked about
419. 1:07:24yesterday and three months ago and six months ago and a year ago and you'll be able to say things like
420. 1:07:31hey I feel like I've kind of stagnated am I am I really learning anything new you know tell me where I was three months ago oh sure Kyle
421. 1:07:41well three months ago you had six words in your vocabulary oh right and now you have 622. you know so so that kind of stuff is coming
422. 1:07:51and that's and you know playing with these tools with with things like the the Mortal Kombat you know toy right the toy that you made um
423. 1:08:01you're learning how to do that and at some point I guarantee you this will happen some problems gonna come up where
424. 1:08:07someone's like I really need to solve this thing and it's gonna be some variation on what you built for the Mortal Kombat thing
425. 1:08:14and you'll be like oh you know let me show you something and you show them your little Mortal Kombat tool and I'm
426. 1:08:19like this is kind of like what you're talking about right they're like yeah if we could just do that for this oh yeah I
427. 1:08:25can do that let me just change the prompt here you go Emilio's wife what's happening welcome welcome welcome
428. 1:08:36stable doodle is doing text quite well is it wow all right so stable dude that's what it is clipdrop.com
429. 1:08:55stable doodle so so okay so what is clipdrop.com clipdrop.co is similar if you've played or seen anyone play with Adobe Firefly
430. 1:09:07where there's like these different tools or or descript is another one there's different tools this is stability AI so
431. 1:09:14the company behind stable diffusion these are all just individual little tools like like here's um stable diffusion XL that's their uh
432. 1:09:24mid-journey competitor um they have an uncropped tool which is like unzoom in mid Journey they've got a reimagine tool where you could oh
433. 1:09:34actually Tobias reimagine so go to go to clipdrop.com and upload your car picture to reimagine so reimagine's a cool tool
434. 1:09:44where you can like upload a photograph of a bedroom and it will maintain the um the composition but change out like
435. 1:09:55Windows and walls and you know drapery and bed sheets and things like that that could work with your car image reimagine
436. 1:10:02but then they have this stable doodle thing and let's see um can you type in text let's see um let's do
437. 1:10:25this this is embarrassing embarrassingly bad drawing foreign
438. 1:10:41help wanted sign in shop window
439. 1:10:57so that's my shop oh did it holy crap copy cuppy 206 bringing the heat all right so here here's here's your value
440. 1:11:10of the AI learning lab I love it I love it when members help one another and and like bring new stuff that's super cool so if
441. 1:11:20you don't like the the fact that this says help wanted um getting these generative tools to say anything remotely like what you want is
442. 1:11:29is really tough and that just did it and it and it basically followed my little model there that's pretty slick wow all right cuppy
443. 1:11:40206 I I can I give you back a heart I can block you I can report you I can translate you and mute you but no I can't acknowledge you
444. 1:11:53copy206 golf clap very nice that's really cool um but anyway yeah that's that's clipdrop.co so that's stability ai's little pile of tools
445. 1:12:06they have a relighting tool they have an image upscaling tool they have a background replacement tool and they
446. 1:12:12have a text remover if you've ever tried to photoshop text out of a complicated image with with subtle gradations that's
447. 1:12:19almost impossible so lots of tools I these sites are really interesting what these basically say to me is you know all of the companies
448. 1:12:28doing this work are experimenting with all sorts of different things and these are their experiments and the ones that
449. 1:12:35work they're going to roll into you know products and the ones that don't they'll just jettison so but we
450. 1:12:41get to play in the meantime that's super cool thank you for sharing the super super cool use chat gbt to get sales intelligence oops
451. 1:12:53use chat EBT to get sales intelligence Biz ideas off the ground writing sales emails using triggers yep Ted that's
452. 1:13:00perfect that's awesome and again it's like do the stuff that you know you're already doing and just have these tools
453. 1:13:08help you do it better as you get your head heads around is you get your head around what they make possible great
454. 1:13:15here's another tip I was I was working with um some of the folks where I'm the I'm the chief generator officer of that group and um
455. 1:13:25oh no you know what this was this was last Friday yesterday yesterday I was given a workshop to a big multinational corporation an AI how to
456. 1:13:35use chat GPT kind of Workshop and um we did kind of an AI confessionals at the beginning of the thing like like
457. 1:13:45who's used chatgpt over the last week and what did you learn and did you do anything for work and this one guy said
458. 1:13:51well I didn't really do anything for work but at home um you know my my kid wanted a a story and and so we went to chat GPT
459. 1:14:04and we made a story for her that where she told me what she wanted the story to be about and we made this story and so
460. 1:14:10ever since we've done that like every night we've had a new story that we're generating um and you could tell he got really excited about it
461. 1:14:19and it's one of the one of the um we're so early in whatever these things are going to be that that literally
462. 1:14:31anything you do with them whether it's making the Mortal Kombat slang talker or you're making personalized books for
463. 1:14:39your kids or or you're writing a creative um resigning email that that basically tells them to [ __ ] off without telling
464. 1:14:48them to [ __ ] off like anything you do where you sort of get your head around how these things work and what they're
465. 1:14:54good at is going to serve you in lots of other places right because you're you're putting the piece Parts together of of
466. 1:15:02how this incredibly deep and sophisticated I don't know it's a tool works right and someone's gonna ask you
467. 1:15:13something at work and you're going to be like oh that's just like the thing I did for the kids book wait a minute I know how to do
468. 1:15:20this right and you're going to be able to solve something without even knowing that what you learned over here was
469. 1:15:26actually directly applicable over there so don't judge anything that you do right now even if it's brilliant don't judge it as
470. 1:15:34brilliant because three months from now someone will come out with a feature that will probably obliterate it
471. 1:15:41and then don't beat yourself up if you think well this is just stupid nothing stupid right now every every any
472. 1:15:48insight you have is gonna be a revelation for someone else there needs to be more more legislation in AI Tech I don't disagree with that
473. 1:15:59um the uh if you're interested the the the four major players Microsoft meta Google open AI I think there might be
474. 1:16:10seven but they just created this group called The Frontier Model group the Frontier Model something Frontier Model forum
475. 1:16:25so go look up the Frontier Model forum and they're basically agreeing like the the big companies that are building
476. 1:16:34these really strong models are agreeing to work together to Share technology and anything they learn about alignment and
477. 1:16:43alignment with these tools is basically like are the goals of the AI machines aligned with the goals of these meat bag machines
478. 1:16:54and if they're not we're in trouble and so um yeah and the other thing you know Sam Altman and Microsoft and Google to a
479. 1:17:05degree lesser degree but but to a degree have been pretty aggressive about interacting with Washington and saying
480. 1:17:13like Sam Altman's been you know openly talking about please you know regulate us and he has a plan like he wants you know
481. 1:17:22companies the size of open AI should be regulated if they get good enough that they can do things that are what they call AGI
482. 1:17:33artificial general intelligence those should be licensed like you license a nuclear power plant and then everyone else he says they
483. 1:17:42shouldn't be regulated so so you've got some activity from the major players in these companies interacting with
484. 1:17:49Washington in what seems like a pretty smart Bay right now I don't know how they're not necessarily being
485. 1:17:54transparent but they're at least being active and visibly active so I totally agree with you what AI are you using is anyone
486. 1:18:03able to get it um I don't know which AI was I was using the image thing that I was generating I was generating in mid-journey anyone can
487. 1:18:12get mid-journey but you do have to do it in Discord the Photoshop thing where I extended the photograph in Photoshop that's the
488. 1:18:22Photoshop beta so you need a Creative Cloud subscription um the other thing I've used tonight was I I don't even think I've done anything in
489. 1:18:33chat GPT yet so I think it was uh I think it was uh mid-journey ask it if it believes in God it'll tell
490. 1:18:45you it doesn't believe in anything and certainly if you start asking it about to to comment on religion a lot of its
491. 1:18:51safety guard rails won't deal with it um yeah I mean if you really if you really like I mean if you want to get educated about
492. 1:19:02religion it knows about all of them right the minute you start going into that territory of like does it believe
493. 1:19:08in things like that the safety guard rails are pretty tight around that if you really want to explore that [ __ ] go
494. 1:19:14to uh labs.perplexity and play with the Llama 2 models and if you really want to explore that [ __ ] install your own llama
495. 1:19:212 models and train them on your own and take off all the safety guard rails all together and I I talked to a guy that they he built a
496. 1:19:30bot within his Slack his company slack thing using llama 2 without any safety guard rails he says it cusses so much that he said it's
497. 1:19:40embarrassing how much it cusses and he said it changes personality every three days he said it's wild having having one
498. 1:19:47of these models that doesn't have the safety guard rails on it he said it's it's not necessarily a good thing
499. 1:19:54I mean if you think about that you got a professional work environment and you put it on your slack Channel and
500. 1:20:02this thing starts being racist and aggressive and violent and cussing and all that sort of stuff in a work environment
501. 1:20:13does that does that guy open himself up for a hostile work environment lawsuit because he installed a non-safe model possibly so
502. 1:20:26I mean feel free to go play with the unrestricted models from meta but just know that they're unrestricted
503. 1:20:34I was hoping to use Bing to create help design a new logo it appears that everything is watermarked really
504. 1:20:45you mean like the little Watermark down in the lower right [Laughter]
505. 1:20:59let's see [Music] make me an illustrated logo
506. 1:21:14of an organic avocado company that has a prairie dog as its mascot
507. 1:21:30um wow that's a very creative request I'm happy to help you do that you know you should just be able to like
508. 1:21:46there should be a button on all these things it's like I'm not in the mood and it's just like when I ask you [ __ ] just
509. 1:21:54tell me the answer don't give me the fluff oh um I don't see any watermarks here but wait do I
510. 1:22:12oh you used bincreate bin creates a completely different thing isn't it hang on isn't being create like their their
511. 1:22:20canvas sort of thing been create no that's the image generator what was that prompt I just gave it hang
512. 1:22:34on where's my big women where did my bingo what did I do did I lose my bang oh Bing where'd you go oh no big
513. 1:22:51I lost my little avocado logos oh tab management oh add oh there it is [Laughter]
514. 1:23:07actually cute little logos oh sorry sorry about that so I'm this has been create and I'm not getting watermarks so
515. 1:23:19maybe you need to have an account is a possibility um I don't know what to tell you to help design a new logo it appears everything
516. 1:23:33is watermarked you could try um you could also try leonardo.ai which is sits on top of stable diffusion they've got a free version
517. 1:23:45that you can do like I don't know seven or ten message or images a day for free um if you've got access to the Adobe
518. 1:23:57uh Creative Cloud you could try Firefly um they seem to be actually doing pretty good with the the logo stuff although I
519. 1:24:10don't know if Firefly hang on I think fire I didn't get that shut up did you try again oh stop Siri um Adobe Firefly beta FAQ
520. 1:24:43[Laughter] let's see um can you use
521. 1:25:00Adobe firefly images commercially yet you cannot so Adobe Firefly is still in beta so you can't you can't use anything it creates
522. 1:25:15um commercially so that sucks so I think I think Firefly and you know Photoshop generative fill are going to be good but
523. 1:25:23they're not they're not they're they're not ready yet I would try Leonardo um it might be worth your while to play
524. 1:25:32around with mid-journey because it's it's pretty good at the logo stuff as well um but I'm not seeing I'm not seeing the
525. 1:25:39uh the watermarks that you are for for Bing create I love how you all have set it up wait set up have it set up to play with
526. 1:25:49mid-journey and such in the salon yeah that's pretty cool it's it's neat to be able to be in a community that's that
527. 1:25:56where you can ask questions and talk about things and talk about the news um in a in a way that um where you can
528. 1:26:04see everyone's images and you can see oh that's Roger's images and that's Ali's images yeah it's neat it's it's pretty
529. 1:26:10fun AI should only be used as a tool by a human and not replace humans in a workplace it should but it will definitely replace humans in the
530. 1:26:21workplace um I think it's going to be a long time until well I don't know a long time I don't know what that means
531. 1:26:29within the next two years probably more like five years I think most of the times most of the time humans will still be in the loop but
532. 1:26:43there will probably be less humans needed to do the same amount of work so I so even if AI is only used as a tool um
533. 1:26:59you're still we're still going to lose jobs Paul Paul rates are today on on the I listened to the his podcast today uh
534. 1:27:08the marketing AI institutes weekly podcast he made a really interesting point he was he was talking they were talking about
535. 1:27:17um chat gpt's code interpreter which is a full-on data analyst in your pocket right you can you can if you do a survey
536. 1:27:25you know the thousand people asking them all sorts of questions you can just pop that data into code interpreter
537. 1:27:31and generate charts and graphs and insights and identify Trends and anomalies in the data he made a great point that it's like
538. 1:27:43they can now do data analysis in their company and he said even if we were a hundred person company I don't know how
539. 1:27:49many people they have maybe they have 50 people right now he said even if we were a 100 person company we would not
540. 1:27:57hire a data analyst we don't have enough data to analyze but he goes with this tool we're now doing full-on data
541. 1:28:05analysis so they're kind of leveling up the quality of what they do um because this tool is available so I think that's
542. 1:28:13also going to happen is that is that there's going to be kind of an amplification of the capabilities of the humans that are in a company
543. 1:28:25the challenge is like I'll take I'll take a world I know if I if I look at agencies that right now have a time and materials model they
544. 1:28:35charge by the hour for their creative people so if they think it's going to take 50 people a year to execute all the
545. 1:28:43stuff that the client wants then they build them those 50 people times the number of hours they work times 120
546. 1:28:49bucks an hour or whatever it is they're going to be competing with small 10 person shops that can do the same
547. 1:28:57amount of work at the same or better quality for half the price how do they how do they compete in that
548. 1:29:06world without cutting costs I don't know how they can or without cutting employees I don't know how they can
549. 1:29:14so I don't know I do not know I do not have a good answer for that one replacing humans with an AI is a crime against humanity um
550. 1:29:26I don't necessarily agree with that I think I think listen I think there there's going to be disruption here
551. 1:29:33that's going to be massively painful and some of it's not going to be great a lot of the jobs that are out there are repetitive
552. 1:29:42Cog in the machine um mindless soul-crushing work and these tools are going to automate some of that out
553. 1:29:56and what that may the benefit that that may cause is people stop working those jobs that are not fulfilling in any way that are crushing
554. 1:30:10and they apply their talents to other other kinds of jobs and those might be jobs that they enjoy more more so I
555. 1:30:17think there's going to be a lot of reinvention of people either discovering that they want to do something different
556. 1:30:23or you know potentially being forced into some new line of work or some new way of working that ultimately might be
557. 1:30:31better so I don't think it's you know are there going to be bad situations and a lot of them yeah I think I that I agree with I absolutely
558. 1:30:43agree with that and it sucks and this [ __ ] isn't getting uninvented computers and Robotics did the same thing yep absolutely
559. 1:30:57and major gift funding fundraising how you doing tonight Kyle can we talk more about automations yeah ask me specific questions I'm happy
560. 1:31:05to happy to share what I can um that would be awesome maybe I'll do an automation Workshop that's that would
561. 1:31:14probably be a good one like using zapier do some automation stuff because it's you know what's funny the the uh
562. 1:31:23conceptually automations are really simple but when you get into them and when you get into the prompt crafting of
563. 1:31:29them and how you incorporate Fields into the things it's a bit of a [ __ ] show so that might be a good Workshop to do
564. 1:31:36that would be awesome thanks for the recommendation I know the photography worlds will be affected too yeah yeah
565. 1:31:42Jennifer I don't know if you're still here but um Leah Leah fast and you know she's she's been a professional
566. 1:31:48photographer for 25 years and she's really good she's really talented and when we started the salon she hadn't
567. 1:31:55even played with generative AI yet we were just talking about she was curious about it and she she sensed that it was
568. 1:32:01coming and she was talking about listen I've been I've been through the transition from film to digital like I
569. 1:32:07know what this is like everyone's scared of it and then we figure it out and and so she goes I'm going to start
570. 1:32:13learning this stuff and I was I was pretty good at stable diffusion at that point so I was teaching her a little bit
571. 1:32:17about stable diffusion and then she just sort of went all in and figured out how to replicate her style
572. 1:32:24using stable diffusion she could take images and you know train up using dream booth and things like that train up you know
573. 1:32:33the loras to to mimic Her Style and she was a little bit blown away and she was like oh this is this is there's there's a part of it for her
574. 1:32:45that was like really disturbing and terrifying and then there was a part of it for her that was really exhilarating
575. 1:32:51it's like this is the new camera right and I'm and I'm able to express my vision for you know what I think is important in
576. 1:33:00photography and I can express that with these tools even though they had you know fingers that looked like this
577. 1:33:07and then stable diffusion got a step better and she was then able to I think it was in mid-journey she she was able
578. 1:33:17to upload one of her photographs and just sort of infinitely generate things that looked like her photography and
579. 1:33:25that really freaked her out and I'm that's the kind of thing I'm seeing I'm seeing it with copywriter friends of mine I'm seeing it with
580. 1:33:33artists I'm seeing it with photographers where as they lean into it they're discovering an excitement about
581. 1:33:44the future of what it might be like but they're very much being confronted with the reality of if one of her clients at some point if
582. 1:33:57if the tools get easy enough that anyone anywhere can just take any image like that image from that Wired
583. 1:34:04Magazine that's got you know that pink background and he's got a big forehead you could just take that and pop that in
584. 1:34:11and say I want to generate a whole series of images that follow that style with all you know describing the
585. 1:34:18characters that you want and it just does it flawlessly what happens to all those photographers what happens to that photographer
586. 1:34:26that I don't know I mean I do know and then and then who become who become the trusted creative resources in that environment
587. 1:34:40they might not be the people that were the trusted resources previously right some might be some are going to make the transition
588. 1:34:52but it could be a 25 year old kid who Masters you know um a particular Masters copying Styles right without you
589. 1:35:07you know using artists names they're just really good at the tools so that that could be coming all right
590. 1:35:16um that would be awesome thanks for the recommendation okay on Mid Journey how do we put an image in it all right let
591. 1:35:21me show you that name a journey all right so here's uh I'm in Discord these are your servers here's your
592. 1:35:28conversations I'm in the AI salon right now and I'm in the mid Journey Channel so if you want to upload an image to use
593. 1:35:37essentially as your source to then do a prompt on top of you do it like this you go to I'm on a Mac so I go to finder find an image
594. 1:35:49and these are some of my Kyle Shannon dreams things those are let me go find something oh there's Champion there's my doggy
595. 1:36:01um all right let's grab Champion actually all right so I'm gonna pop that in there so so I literally just dragged it from
596. 1:36:12my finder window into the Discord chat and I hit return and so what that's done is Now it's uploaded it into Discord so
597. 1:36:19now I'm going to right click on a Mac control click and I'm going to say copy link so I'm going to copy the URL of this image
598. 1:36:29and then you basically just do slash imagine like you normally would and rather than typing anything you paste in the link to that image
599. 1:36:38and then you say um [Music] chubby dog
600. 1:36:53looking at a spare rib while lying on the beach at sunset
601. 1:37:08so we'll see if it does it like it should it should make a dog that looks like champy see his little see this little prison
602. 1:37:17his prison tear he Shanks someone in the big house [Laughter] Johnny bullets 10 out of 10 on the name
603. 1:37:30um AI That's free I'll I'll get back to that in a second oh this definitely looks like champ this is cool
604. 1:37:36notice if you haven't seen mid-journey before if you haven't seen how these diffusion models make the images they
605. 1:37:42start as like these kind of random noise images and then they just get increasingly less noisy until they're
606. 1:37:48actually just generated this looks like him a lot the one in the upper right that's pretty cute he's on the beach the one on the bottom
607. 1:37:57here has got some bones that's pretty cute um and then I could say um let's do slash imagine pop that image back in there and I'll now do
608. 1:38:12um claymation sculpture of a chubby black dog dancing
609. 1:38:28in a diner foreign ation it including animation
610. 1:38:46sculpture of a chubby not Chevy chubby black dog dancing in a diner and we'll do AR wait dash dash AR
611. 1:39:03-16 colon nine uh unrecognized parameter dang it
612. 1:39:20hold please hold please old man trying to type type [Music] I haven't had any trolls tonight telling me isn't acute that the elderly are
613. 1:39:37teaching people how to use AI on tick tock come on come on where are my trolls we're patrols peeps by the way uh feel free to uh subscribe
614. 1:39:50to or uh to uh to follow my channel I make Tick Tock videos I do these lives a lot if you want to be notified when I go live
615. 1:39:57feel free to follow me um I'm trying to think what else hmm let's see is this gonna look claymationy
616. 1:40:08it's not looking super claymationy right now uh they're not really claymationy but they're cute they sort of claymation kind of eyes
617. 1:40:22anyway so hope that helps AI That's free okay so Johnny bullets John and bullets Johnny bullets here um chat GPT is free but you can pay 20
618. 1:40:37bucks a month for the fancy version Claude is free Bing is free Bard is free labs.perplexity is free Pi is free po is
619. 1:40:50free but there's a there's an upsell there's a 20 Buck upsell there as well um if you're gonna pay 20 bucks pay it
620. 1:40:59there's an upsell there's a 20 Buck upsell there as well um if you're gonna pay 20 bucks pay it to open AI because it's just a better tool
621. 1:41:11um and all the rest are free and then for images leonardo.ai is sort of free you you get like I don't know seven or ten images a day um
622. 1:41:25there's something like mid-journey called Blue Willow where you generate images in Discord and that's free um
623. 1:41:40trying to think what else these are the biggies I mean I if you're just trying to get your head around this
624. 1:41:46this uh this chat CPT stuff in this large language model stuff start with these most of them are most of them have some sort of free option
625. 1:41:54and some are all free um do you pay for it I pay for chat GPT I pay for mid-journey the the lowest it's like 10 bucks a
626. 1:42:08month because I don't do it a ton I pay for zapier to do automations I pay for Google collab to do stable diffusion but I haven't I I
627. 1:42:21pay for that but I haven't used stable diffusion in ages because because I pay for mid-journey now and it's a bit of a
628. 1:42:27pain in the ass to set up um trying to think what else I pay for I pay for a a video thing um
629. 1:42:42what's it that's kind of it I don't pay for much I'm purposely trying to keep myself limited because this shit's so
630. 1:42:52overwhelming as it is like like the fact that there are seven things on this list like I I should cut this down to three but I can't because
631. 1:43:05they all have something interesting and I you know I could go through why I have these seven on here but I think they're
632. 1:43:10all worth in worth worth exploring and those are just the chat GPT like things and we don't have apple in there we
633. 1:43:17don't have Elon musk's whatever his entry is going to be in there and we don't have like whatever thing is
634. 1:43:23being researched right now that like that list is probably going to go to 10 soon so it's pretty overwhelming right now
635. 1:43:36the salon is free Discord is free chat gbt 3.5 is free so thanks dragonfly Alchemy um some freed some are paid I don't pay
636. 1:43:49for anything myself yeah you don't you don't have to pay for anything right now there's enough free options with this
637. 1:43:55stuff here's the deal most of these things are just just this side of demo wear right like the the underlying large language model
638. 1:44:06sitting underneath them are incredibly powerful and I spent a lot of years on them the application layers for these
639. 1:44:12like chat GPT itself is just now starting to get basic features like you still can't this blows my mind [Laughter]
640. 1:44:30this is my chat history right so this is all of the chats I've ever given I've ever interacted with you still can't search for them
641. 1:44:43so if I want to go find something I did like I know I did something on a Twitter thread and and the problem is because I I demo
642. 1:44:51this thing on on lives all the time like most of my recent ones are all stuff that I've done in here that are not
643. 1:44:58things that I would be searching for so so these tools are very very very very very raw right now so that's why a
644. 1:45:07lot of them are free as they get better they'll start charging for them but as they get better it'll be worth paying for them
645. 1:45:18I heard the coolest hack for mid-journey oh tell us share do share do share fight for a human's future
646. 1:45:28you find someone's image and prompt you like then you change two to three words and you get what you are wanting yeah
647. 1:45:36that's a cool hack there's also um look up I haven't played with it so I don't like I haven't played with it
648. 1:45:42recently so I forget how to do it so I don't want to put you through me trying to remember that but look up remixing in
649. 1:45:48mid-journey because there's uploading the image and doing what you just described but there's also remixing where you're remixing two images
650. 1:45:57um and that's that's an interesting way where you can you don't have to use prompting at all um don't pay for AI sell out don't pay
651. 1:46:05for AI sellout don't Johnny bullets I don't know what you're talking about all right I'm sorry but as
652. 1:46:13a self-respecting when the Discord makes my ovaries shrivel up it's so proteo with a horrid ux I'm telling you it's so bad it's so Discord
653. 1:46:28is just a [ __ ] disaster it is it's just like and if it's like crypto Bros crypto bros man you know man it's all about the community
654. 1:46:40of douchebags I'm in one good nft community and even that one has got some sketch
655. 1:46:56we must have better legislation than AI yes you've said that and I agree I agree better legislation for all we all need
656. 1:47:05more legislation we want regulation why are you regulating us stop regulating us we want more we that's too much we want
657. 1:47:13more more less more less more never never any regulation except for the ones that I like yeah I know I know it's true
658. 1:47:25hey where do you need truth tables where do you need truth tables in a microprocessor if you're asking me about microprocessor
659. 1:47:36design this is the wrong Channel silly question Basics on AI um I don't quite know what you're asking knit or night um
660. 1:47:51if you can clarify that I I know nothing about um sort of memory and microprocessor design and routing and
661. 1:48:01logic gates and [ __ ] like that I know the words but I don't know at all what I'm talking about there what's the best for video free uh
662. 1:48:12there's not many options right now for free video all the all of the time-based media the video the audio synthesis the
663. 1:48:21animation talking things they all cost money and they're all expensive um I think that we'll see so the the CEO of
664. 1:48:31stability AI has said that the the version of the the version of image generation that they're using inside
665. 1:48:39stability AI right now or they're developing right now is rendering at 60 frames a second so I have a feeling that
666. 1:48:47stability will at some point here launch something that's doing 30 frame a second or 60 frame a second video and they'll
667. 1:48:55probably have a free option but right now there's just nothing so which is a it's awful now that's cool Kyle I don't know what I
668. 1:49:05did but awesome alien invasions will create a government agency to control AI technology wait unit unique unique real uni Creel
669. 1:49:19TV Crea oh uni uni creatv so I assume your Conspiracy Theory TV show awesome but okay you know I don't know
670. 1:49:38the world's so [ __ ] up right now that it's possible that the government just announced we have alien technology in
671. 1:49:45the rest of the world's like yeah whatever 's on the panic in the streets but they haven't really announced it
672. 1:49:53like that the the thing two weeks ago with the two guys like you know the whistleblowers talking about we have
673. 1:49:59alien technology neither of them had any actual data or or like facts it was all like wow I heard Bob said we've got an alien
674. 1:50:11in the back of that 49 pickup like okay it's like come on come on we're gonna have it let's have it what about stable diffusion stay okay so
675. 1:50:24here's here's the thing about stable diffusion and why when when I talk about uh images image Generation stuff
676. 1:50:33the three that I talk about right now are mid-journey Adobe Firefly and Leonardo so Leonardo is a web front end on top of stable diffusion
677. 1:50:44so is um Blue Willow which is the Discord based version of that the reason I don't talk as much about stable diffusion is that it's just a
678. 1:50:55pain in the ass to actually you know what I should probably put on here maybe I'll do that right now [Laughter] thank you for that clip drop
679. 1:51:08dot Co um because those are both because that's that's the stable diffusion you can play with uh stable diffusion uh XL there
680. 1:51:22um but see see my problem like I try to keep these things simple and it's just like they just you know there's just a there's a lot going on
681. 1:51:34here's the deal with stable diffusion it's it's a really exciting tool and project and XL gets pretty close to Mid Journey
682. 1:51:45um it's just really complicated to set up and because there's so many because it's open source and because so many
683. 1:51:51developers are developing so many features so quickly it's a it's a software management nightmare unless you're a geek and you
684. 1:52:01really like keeping up with the latest version and updating your Jupiter notebooks and writing your own [ __ ] python code I like
685. 1:52:12if you want flexibility and if you want sort of the latest and greatest tools stable diffusion is the way to go if you
686. 1:52:19want the best quality right now mid journey is the way to go but a lot of people are catching up if you want
687. 1:52:25something that's commercially not dangerous Firefly and generative fill for Adobe are the way to go and then if you want just images that
688. 1:52:35are kind of okay Dolly and Bing Images are the way to go I think those will get better I think I think uh open AI is going to release
689. 1:52:46some version of Dolly that doesn't suck but it sucks right now I've been in it for 43 years the job is this job is constantly changing yeah
690. 1:53:00yeah yeah well if you're not if you're in I.T that I mean it's constantly changing the the thing that's different
691. 1:53:06David I I think about about this stuff is it's it's I mean you're you're probably already
692. 1:53:22dealing like if you're in I.T you've been dealing with machine learning and AI stuff probably for a lot of years
693. 1:53:27anyway right like the generous stuff isn't new but what the generative stuff adds for it groups is they're going to be systems that come
694. 1:53:35out that dramatically simplify the complicated right so so within Windows 11 Microsoft is going to put generative
695. 1:53:42AI copilot in in you know on top of windows so you'll just be able to say reconfigure my system you know to I
696. 1:53:52don't know change these settings and it will just go find the settings and change them um and that's going to happen sort of at
697. 1:53:58a big it level as well so you're still going to have to manage that layer up there but what happens to all the people whose job
698. 1:54:08it is to manage all that stuff manually right now what happened you know do you may not need as many of them or you
699. 1:54:14may need to find new things for them to do or the it group May shift into an innovation group because these tools can do so many
700. 1:54:26things that you might be able to quickly be developing applications and things within the cut like the role of it might change
701. 1:54:34dramatically so even though things in it are always changing and evolving I think these tools are just enough of a different nature
702. 1:54:46um that that it might it might radically change what you do and what the group does it's fascinating stuff um got a question for you the
703. 1:54:58[ __ ] all right the [ __ ] got a question for you no [ __ ] all right [ __ ] what's your question put your question in there
704. 1:55:08so what are your plans no [ __ ] oh I think we got a troll [ __ ] the [ __ ] I think he's trolling me I don't know what you're it's here
705. 1:55:19I don't know what your uh all right AI is the best technology ever made um when AI starts to create even better
706. 1:55:33AI it will improve exponentially I think that's starting to happen now if you if you look at like every time you see a
707. 1:55:40white paper on Twitter and then three days later you see a new white paper that incorporates that white paper and
708. 1:55:46two others for some new white paper and then they take that white paper and incorporate like holy crap like the Lang
709. 1:55:54chain and the auto GPT and baby AGI and I forget what the new one is and then you got GPT for you got all
710. 1:56:00these things just sort of stacking on top of one another it's absolutely insane right now my Pal's in charge of public health for
711. 1:56:08nursing homes and hundreds of people a week are still dying of coven yeah it's not over and it looks like there's a new
712. 1:56:15there's a new variant that's putting a lot of people in hospitals right now yeah I know it's not over it's it's weird how we
713. 1:56:24we were in lockdown and then we just kind of weren't and but it it ain't over yet so yeah at some point some variant
714. 1:56:31May shut us back in I sat in this office for two and a half years
715. 1:56:47it's gonna be a new kind of lonely I don't know what I was talking about back then why use Google to search for medical AI
716. 1:56:56when you can use an AI tool to find it for you yeah I agree with that although you know you have you have Google barred
717. 1:57:03is connected to the internet now to find anything up to date you need something connected to the internet so you need
718. 1:57:09Bing or Bard or one of the plugins for for uh chat GPT since they killed its connection to the internet and by the
719. 1:57:19way why has that not come back why have they not fixed that it's been over a month it's been what six weeks since they got rid of chat
720. 1:57:30cheap being connected to the internet Bing's still connected to the internet their plugins are still connected to the
721. 1:57:36internet what the actual [ __ ] Sam Altman Sam are you I I don't know if Sam is a regular Watcher of the AI learning lab but you
722. 1:57:48know I like to think that he is Sam what are you doing buddy like I know you pissed off time Inc with
723. 1:57:56the Fortune Magazine thing like you know pulling content from behind the firewall cute trick but uh seriously you haven't
724. 1:58:03[ __ ] patched the hole like what is the issue I I can only imagine the issue cannot be a technical one I I guarantee you it's
725. 1:58:13not a technical issue I think the issue is I think this is I think they're in a negotiation right now with Publishers
726. 1:58:20and I think the Publishers are like if you connect that [ __ ] thing to the internet and you steal more of my
727. 1:58:25content before we have a deal where you give us 12 trillion dollars we're gonna have an issue I think that's
728. 1:58:31what I think that's why it hasn't come back but [ __ ] do your negotiation cut your deals and give us back our
729. 1:58:37connection to the internet this is about us not you foreign we got someone in the chat here called the [ __ ] you see what I'm saying
730. 1:58:50you don't [ __ ] around with this community you don't do it hook it back up good Lord I'm now now I'm all for clumped
731. 1:59:05ah she's awfully clamped we have on that special I'm not special looks like Mr Shannon just took a trip
732. 1:59:18down Satan Lane isn't that terrific wow wow wow wow didn't take long for that to happen did it have a little bit of that caffeine the
733. 1:59:28devil's juice the devil's juice I'm Jack 8 37 p.m Mountain Time um what time is it actually 10 to 10.
734. 1:59:44okay listen I I have to be at my co-founders house tomorrow morning at 4am so I can't be up much longer I gotta
735. 1:59:55finish packing I'm going to New York City and I'm getting my eyes Zapped by one of Sam Altman's orbs and um
736. 2:00:05what else am I doing going to see a big Pharma client I'm not gonna be there alone we're flying in tomorrow
737. 2:00:15and then I come back Wednesday morning so like quick quick intense trip which is good uh uh 5K 80s to do llama two yeah I
738. 2:00:28don't know I don't know what the requirements are for llama two for the 70 billion parameter one I think you
739. 2:00:32need some pretty serious iron but the I think the seven and the 13 you don't need a ton I think you can do
740. 2:00:40the 7 and the 13 on the on a Apple silicon but but I don't I haven't played with them yet that's why I have this up
741. 2:00:48here the labs.perplexity you can't train it but you can at least go play with the models and see where they are right now have
742. 2:00:55you seen anything on project management for infrastructure Improvement Master planning a city like no I just like that's that's
743. 2:01:04significantly more specialized than what I'm dealing with right now I would go to futurepedia.io and and look for it my
744. 2:01:12strong strong suspicion is that probably not um with this with this generative AI stuff but there's probably some AI
745. 2:01:22assisted tools within project management software for that kind of stuff I would think that if you've got that
746. 2:01:30kind of expertise you learning these tools just getting up to speed on these tools you know it could be the kind of
747. 2:01:36thing you could contribute to something there but I I don't know go to go check out futurepedia.io and see if you find
748. 2:01:43anything that that feels fairly specialized I would assume that's going to be probably an existing entity in that
749. 2:01:52space that layers this generative stuff on top of their tool is what you'll see first so I I just start paying attention to the
750. 2:02:02oh you know you know what might be interesting is go to um just go to Reddit and you know search for the kinds of
751. 2:02:10keywords you have there and then generative AI or you know GPT project management Master planning infrastructure Improvement
752. 2:02:19um I'm sure there's people talking about it there and if they're talking about it if there's a tool there they'll be
753. 2:02:24talking about the tool so that might be a way to find that that feels like something specialized enough that probably not for a year or two
754. 2:02:35but I'm totally clueless clueless to that world so gonna ask AI to explain web3 to me if you get a good answer big [ __ ]
755. 2:02:44420 10 out of 10 on the name um if you get an answer for that let me know diagnosed Diagnostics will be great but
756. 2:02:55as an MD you'll have to dig yeah they're they're like a virtual PA yep exactly these tools are not really ready for prime time yet so
757. 2:03:05we have to give them a little Grace cinematic voice actor 1971. voice acting is in is going to be slammed by this stuff
758. 2:03:15how are you going to fix the safety braking instructions allowing illegal activity with gpd4 I don't think you are and quite frankly
759. 2:03:25um meta open sourcing their llama 2 model like at least open AI has some safety guard rails on it yeah hackers are gonna hack and and a
760. 2:03:36paper just came out last week that someone figured out how to put a code you know some code at the end of a
761. 2:03:44prompt that will jailbreak any of these things and there's nothing they can do about it they didn't they didn't
762. 2:03:49actually put the prompt itself in the paper or the code and the paper but someone someone's cracked that already like that's done
763. 2:03:56but the illegal activity using llama 2 that like metal just putting out the weights some would argue is irresponsible oh so so
764. 2:04:11we can at least [ __ ] open AI at least Sam Altman said they're committing 20 of their compute resources to Super
765. 2:04:20alignment so that so they keep these things safe like at least the big companies are are aware of what you're
766. 2:04:26talking about and they're they're putting Tools in place for the safety the fact that people are going to hack
767. 2:04:33it and figure out ways to do shitty things with it why was that any different than what we've had before
768. 2:04:37other than the fact that they can do way shittier things but the good guys also have these tools as well don't forget that you'll be able
769. 2:04:46to [ __ ] with the hackers two right you'll be able to catch them and punish them and do all sorts of [ __ ] to them
770. 2:04:55the other direction right so that like that war continues like that war has been going on ever since there was something to hack so
771. 2:05:08I don't know I don't know how they're gonna fix it you know so so one of the things that I think you know back to to web 3 and the
772. 2:05:17crypto Bros and all that sort of stuff one of the things that feels like it becomes very very apparent in the next two years is that
773. 2:05:29blockchain or some sort of tokenization technology is going to become really really important we need to be able to verify that
774. 2:05:39someone is who they say they are that the video that you're watching is them in in an unaltered way from the
775. 2:05:48original source that the thing that you're reading was is actually from The New York Times that the website that
776. 2:05:54you're looking at is actually the New York Times that you know you could tokenize you know was was this large language model jailbroken
777. 2:06:07you know was something generated with the jailbroken version of one of these things like so I think that I think that that
778. 2:06:22because these things are kind of open and raw night right now and because meta released llama 2 to open source I think we're gonna have
779. 2:06:32kind of a two-year free-for-all and at some point someone's gonna get hurt or something bad's gonna happen and
780. 2:06:40then everyone's gonna lose their ever loving [ __ ] minds and go we got to do something about this AI it's so dangerous
781. 2:06:47why didn't we see this coming well we did and we're talking about it and the meta's like [ __ ] it here have fun
782. 2:06:59Elon Musk I don't like this Sam Altman's trying to keep open it it's called open AI for a reason [ __ ] it I'm going to
783. 2:07:06start X so we're gonna have llama two and then we're gonna have llama three four and five right that Matt is just gonna go here
784. 2:07:15here's our research have fun and then we're gonna have Elon Musk with his X thing going hey here's more open source go have fun
785. 2:07:23so it isn't about getting I I mean I already think that open AI Google to a lesser extent um anthropic anthropic is big on safety
786. 2:07:35that like I I think their approach to safety is actually more solid than open eyes is right now and inflection that
787. 2:07:43does Pi they're going to be good with this I think the big players are going to be fine like I think they're aware of
788. 2:07:48it and they're dealing with it I think it's the open source stuff we got to pay attention to so crazy first time here Blue Man 88 Blue
789. 2:07:58Man Group Blue Man group's cool what's this AI for which one I don't know what I was using that's Jays just
790. 2:08:08got out of prison is AOL still relevant or not actually what's funny about AOL is having an AOL email address for the
791. 2:08:17longest time was just a sign of like you're a loser if you still have an AOL email address it has come full circle
792. 2:08:24you are Mighty [ __ ] hip right now that's hilarious this is a perfect description on how we use it okay I don't know what I'm saying
793. 2:08:36but awesome oh just just jump me to the bottom it did all right so I gotta go soon so I'm not gonna I'm not gonna go find my place
794. 2:08:45so I probably just skipped a bunch of questions so if I didn't get to your question don't hate me but I'm just
795. 2:08:50gonna sort of jump to the bottom all right let's see how secure is perplexity Labs info entered so Mez assume with all of these things
796. 2:09:02until further notice um they're not just assume that they're not um if you want it to be really secure
797. 2:09:13run these large language models locally so so start playing around with the open source things or spin up a virtual
798. 2:09:20machine in like a Google collab machine or a Microsoft Azure or an AWS on their whatever it's whatever their thing's
799. 2:09:29called it's not I forget what it's called this doesn't matter um you know there there are secure instances you can set up like I know
800. 2:09:39Azure has them I know AWS has them so if you really want it to be secure you're gonna have to set up your own
801. 2:09:44environment if you're doing something like sending data to these companies assume that it's going to be
802. 2:09:51not that it's going to be used assume that it it could be accessed it could be used for training so depersonalize it dedupe it genericize
803. 2:10:01it if it's something proprietary I would just assume you know until someone comes out and makes a claim your shit's safe
804. 2:10:09with us and we guarantee it and here's the contract like right now um open AI put out they've got a [ __ ] website
805. 2:10:21about how we don't train on any data going through our API none of it's trained but their actual agreements that you sign the license
806. 2:10:35actually says well we may use it okay which is it are you not using it or you may use it because the contract like our lawyer
807. 2:10:48right now is like I don't give a [ __ ] what their website says here's what we signed like three days ago so
808. 2:11:02to put it in context I don't think open AI or anyone expected chat GPT to take off the way it did there wouldn't be seven competitors on
809. 2:11:18this list if chat GPT had just come out and no one paid attention to it gpt3 had been out for a year and a half before
810. 2:11:24that the instructions for how to write chat GPT were in the documentation and no one [ __ ] built it open AI built chat GPT in two weeks in
811. 2:11:35anticipation of the Microsoft announcement that they're investing in open AI and they launched it and hit 100 million users in six weeks
812. 2:11:44they weren't expecting that Microsoft wasn't expecting that Google certainly wasn't [ __ ] expecting that so all of this stuff is like
813. 2:11:56the output of this stuff is sophisticated enough that we all kind of assume they have their [ __ ] together
814. 2:12:02I think they're all scrambling to keep up with what just happened thank you very much for that what is
815. 2:12:08that is that like a CD Boombox CD player that was cool Long Live Gen X CDs huh remember when they came out the really thin CDs like
816. 2:12:20these are awesome and then they would break because they were so thin and foreign yeah so so yeah assume this stuff is not secure
817. 2:12:34when will the internet access come back to chat EBT plus I'm right there with you panda come on Sam listen listen everyone at the AI
818. 2:12:43learning lab is very upset about this lack of access to the internet with chat EBT we've been talking about it for weeks
819. 2:12:51we put a committee together and we have decided it's been long enough it's been long enough give us back the internet
820. 2:12:59or we will go somewhere well no we won't because gpt4 is better but that's not the point fam it is not the point
821. 2:13:08we've talked about this before you know what I mean all right
822. 2:13:22can you make your dog talk champion I can make him sing if I play the guitar he will sing he's a singing dog in fact wait what's my um
823. 2:13:36hang on ADD moment Instagram
824. 2:13:54[Music] [Music] yeah if you go to if you go to Instagram
825. 2:14:10and go to Champ sings you can see my doggy singing I'll play one 80d moment wait why is it an audio
826. 2:14:28[Music] [Applause] [Music] [Music]
827. 2:14:45[Applause] [Music] yeah
828. 2:15:02[Music] it was a good singing dog who's a good singing dog [Laughter]
829. 2:15:16yeah music he loves to sing I I mean we'll be watching TV and like just some light guitar Musical come on in the
830. 2:15:26background you don't even notice it and then you'll just he'll just his head will pop up and we'll just go oh oh
831. 2:15:39um I don't think it looks like claymation because it understood a play Mission yeah I know um I would also wait I would pay also
832. 2:15:53kill wait for his magnificent I don't know if that means soprano zero I mean um when will internet come back to chat
833. 2:16:04EBT uh we already talked about that Kyle have you figured out a good way to keep track of your chats no
834. 2:16:12here's so this is one of the uh one of the many downsides to ADHD is that me and I'm good at designing ways to organize
835. 2:16:27chat histories but not good at actually doing the work um and I've got like I've got like um I've got like add hubris
836. 2:16:44where I'm like like I'm working on a prompt I'm like oh I'll remember this and then like I go to do it again I'm
837. 2:16:51like wait what the [ __ ] was that prompt and then I can't find it and now I'm in a I am in an absolute train wreck right
838. 2:16:58now of of prompt Insanity I've I've got some in my notes program I've got some I they're they're everywhere I I scroll incessantly through
839. 2:17:11mychat history at chatgpt trying to find something like a needle in a haystack because it doesn't have a [ __ ] search
840. 2:17:17function how can it not have a search function Sam Sam listen listen couple of things um okay uh connection to the internet we've
841. 2:17:28talked about it I told you we talked to the people here at the lab um what they've said is it's not acceptable not acceptable also why can I
842. 2:17:37not search for my prompts why I've typed them in they're in a database you retrieve them from a server all in a list you sort them
843. 2:17:50chronologically reverse chronologically of course could I not search for them maybe Sam it's almost been a year
844. 2:18:05[Music] and then the other thing is um you can if you're worried about your safety your your your um your data security
845. 2:18:20you can turn off prompt history and training in chat GPT but you lose all your prompts like you lose all the prompts that you've written
846. 2:18:30so at some point I don't know I don't know I there's a way you can download your data so at some point I
847. 2:18:38think what I'll do is I'll download I don't I don't know what I'm gonna [ __ ] do I don't know what I'm gonna do
848. 2:18:45what what I'm actually gonna do is I'm going to wait for this technology to get better because because it sucks right now and
849. 2:18:52I'm just gonna accept the fact that it's just gonna be in an ad hoc add nightmare hellscape of super cool projects and
850. 2:19:02innovative ideas for the next year um and then once this technology gets good enough that it's actually stable in some form or another
851. 2:19:12then then you know I'll bring in the professionals and actually build the system around it but holy [ __ ] no I wish
852. 2:19:18I had a better answer for you how do you use AI to make full-time income this is the coming manual attention
853. 2:19:33this just start get just start getting AI literate is the first step making money with chat GPT is no different
854. 2:19:47than making money with Photoshop or making money with bartending skills right how do I make money with booze well I don't know you don't
855. 2:19:59but if you set up a place where people can come sit and drink drinks and you make tasty drinks for them now you have a business
856. 2:20:07it's the same [ __ ] these are tools so what I would say I would sort of put it back on you and say what are you good at what are you
857. 2:20:14passionate about what are your skills today um you understand certain communities you understand certain communities that may
858. 2:20:22have pain points that you may have an idea for how to make better but maybe you don't quite know how to do it well
859. 2:20:28that will be a hint to how you might use these tools to solve those people's problem using these tools in a way that you
860. 2:20:39couldn't have done it otherwise and that's how you'll make money with chat gbt or AI or it might look like you get curious
861. 2:20:47about these tools and you start to figure it out and then everyone you tell about it's like I don't know anything
862. 2:20:51about that can you teach me about that yeah I can teach you about that but you want should I put together a class yeah
863. 2:20:57put together a class okay I'm gonna charge uh 100 bucks okay great oh okay then maybe the next one you charge 200
864. 2:21:06bucks then maybe the next one you know then maybe you start teaching companies and you can charge them 2 000 bucks um
865. 2:21:20these are just tools they're just tools now they happen to be profoundly powerful tools that may allow you to make money in
866. 2:21:30areas where you haven't historically had the expertise or the knowledge so what these things can do is like
867. 2:21:38unlock your abilities like maybe you know a certain way of I don't know managing pizza shops it doesn't [ __ ]
868. 2:21:46matter like whatever is the thing you know maybe you understand the numbers side of managing pizza shops but you never quite
869. 2:21:54understood the marketing side but you've always been interested in it but you never really got around to going to
870. 2:22:00marketing school or hanging out with ad people well now you can go learn about pizza shop marketing and come up with a strategic marketing
871. 2:22:10plan for every pizza shop in your city and you can do the numbers because you could do them before and you can figure
872. 2:22:17out how to use these tools to do the number thing better too and now all of a sudden you've got a comprehensive
873. 2:22:22service because that's the [ __ ] you know and then these are the tools that amplify you these are ability amplifiers
874. 2:22:34what are your abilities and what are your desires what have you always wanted to be able to do these may help these will likely help
875. 2:22:45so it ain't about the tool it ain't about the tool it's about the intention it's about the human being saying you know I could have been a
876. 2:22:56player they call me the [ __ ] I could have been a player now I can be now I can be makes sense yes no
877. 2:23:10um oh Kyle you're so full of [ __ ] that's a poo poo answer try the Dan do anything now and you'll be amazed yeah I don't like I haven't
878. 2:23:22played with Dan like what's Dan up to he must be up to 17 or 18. the last Dan I saw was Dan seven um what what cipriano's talking about
879. 2:23:33there Dan is this prompt that you can use to jailbreak chat GPT to not follow its own safety guidelines
880. 2:23:42um I can on iPhone search that is chat GPT no I don't think so I think they they took browse with Bing is is in the chat GPT
881. 2:23:56app I don't think so you might have a non-official chat GPT app there's a bunch of non-official chat GPT apps that
882. 2:24:03might be connected to the internet but the official one's not you can search within the open AI app for GPT history keyword searches from
883. 2:24:17previous conversations oh copy206 copy with two things tonight holy crap copy you're you're our MVP of the night
884. 2:24:25okay so the the chat GPT iPhone app you can search your history why would they not have that on the website why would
885. 2:24:33they not have that on the website if they built it into the [ __ ] app because it's a different development team I don't care
886. 2:24:41Sam Sam listen we really need to talk okay just told me you can search your chats on the app okay so what's the story
887. 2:24:57and don't tell me it's a different development team I don't care we're gonna if anybody knows Sam we gotta have a conversation with Sam
888. 2:25:07all right what do I think about Jasper um Jasper's interesting so so Jasper and copy Ai and a lot of things like that
889. 2:25:15you know they're built on top of the you know the foundational models so so they're basically Jasper and copy.ai are
890. 2:25:23kind of fancy versions of chat GPT and they've been around you know they've they've predated chat gbt by I think a year
891. 2:25:30um so they're interesting I'm personally um I'm personally right now more interested in learning how these things work and
892. 2:25:41and how to do it myself just so I can learn how to do it and build my own tool um but if you if you don't feel like
893. 2:25:49geeking out with the prompting I think things like Jasper and and copy AI are interesting they're good that like I
894. 2:25:55find them a little complicated because there's so many things you can do with these that what those pieces of software
895. 2:26:02end up being is just like like an insanely large number of features that are all just slight variations of
896. 2:26:11prompts right and and so I then spend more time learning the interface of this bloated software that I'm going to use three percent of
897. 2:26:26and I have to sift through the 97 of the crap that other people are going to use right but the company has to build all
898. 2:26:33that [ __ ] because you know well now that I made a Blog writer now I need to make a short blog writer and now that I've
899. 2:26:38made a short blog writer writer I've got to do a short blog writer with hashtags and then I've got to do a short blog
900. 2:26:45writer that's not really a Blog but it's opposed to Instagram and it's just like [ __ ] it I'm gonna go learn zapier and just make
901. 2:26:53the tools that I need so so that's where I am right now all right let me answer a few more questions we'll get the hell out of here
902. 2:27:01uh hey Kyle I just finished packing C see you at 4 10 a.m that's Monique that's my co-founder I gotta get my ass
903. 2:27:08to bed I don't know when she put that in there if you're still watching Monique I know I got so the clothes are picked out
904. 2:27:14I got all this [ __ ] laid out that I need I got the board deck ready I got my to-do list ready I got chargers everywhere
905. 2:27:29that's so funny um your regulation rant hilarious um are we having super conducting AI I really want the super conducting thing
906. 2:27:39to be real but it's looking like it's not being real there's also that new cancer pill that I want to be real but I
907. 2:27:47don't know if that's going to be real I really want LK 99 to not be [ __ ] would be interesting well we'll see
908. 2:27:54we'll see the the uh the fat lady ain't sung but I think she is having a burger I think she thinks she's in the wings
909. 2:28:04with a burger and a milkshake getting ready to oh she's a singer she wouldn't be having a milkshake she's she'd be having a Diet
910. 2:28:14Coke really large Diet Coke um I can't wait for the day that AI refers to me as a number exactly this is so off topic do radio shows or
911. 2:28:26audio books oh do radio shows or audio books your voice is great oh thank you very much I appreciate that I um I mean I have a
912. 2:28:34degree in acting so so um but I appreciate that I appreciate that sometimes I listen to my voice and I'm like is there any is this is there
913. 2:28:42there but you know sometimes in the land far far away you know I can go there but then I sound like a faker like I sound like a parody
914. 2:28:54of a radio guy and and it's funny like there's a guy that used to work for me at storyline guy named Wolfgang and he
915. 2:29:01has a radio voice he's got he's just got a radio voice and it's like he doesn't have to fake it it doesn't have to go
916. 2:29:06there so but thank you I appreciate the kind words can AI recreate LK 99 well LK 99 has to actually be built what what
917. 2:29:16what AI could potentially do I was thinking about this yesterday but AI could potentially do is take all
918. 2:29:22the papers all the research of all the research labs around the world that are trying to replicate LK 99 take all of
919. 2:29:29their data synthesize them analyze them and see you know see see where the anomalies are and see where the through lines are
920. 2:29:37so AI could potentially validate multiple um research uh findings um very very quickly which could be
921. 2:29:49could be interesting but in the end they got to build the [ __ ] right they gotta they gotta come up with a if it's real
922. 2:29:55they gotta find a way to manufacture it and do that in a replicable way and then ultimately do that in a scalable way and
923. 2:30:01if they can do that then we're all off to the [ __ ] races then we get flying cars we got all sorts of [ __ ]
924. 2:30:09um does the air prompt generated text prompt wait what roro does the air wait oh does the AI prompt generated
925. 2:30:20generate a text prompt or only images does the AI prompt I don't quite know what that means within mid-journey you can
926. 2:30:33upload an image and you can have it describe that image so you can have mid-journey we'll write you a prompt based on an image
927. 2:30:40and you can do it obviously the other way around where it's text text to image it's classified all right I gotta go
928. 2:30:47Monique's on here she's gonna kick my ass Monique he's he's legit funny I love it yeah like nit many current tools
929. 2:30:57provide hundreds of alerts which are ignored AI would help it would um hey are you set up to do this from
930. 2:31:05your hotel room well you know what's funny Tobias I because and Monique you'll find this funny so the bag that I
931. 2:31:12normally take on trips is not long enough to take my little stand that I have my phone on and I was thinking
932. 2:31:18about that so I am going to take a different bag than I normally would so I can put my stand in there so so I can
933. 2:31:25continue my lives because because the people the people need more of this foreign
934. 2:31:39do you ever use autonomous AI not right now um well that's not true go play with code interpreter if you
935. 2:31:48haven't played with code interpreter that's autonomous AI it's it's a really contained small version of it so it
936. 2:31:55spins up a virtual machine and you give it a go and you upload a file and then it writes python code and executes
937. 2:32:01python code and gives itself goals it prompts itself it makes mistakes it goes oops I use the wrong Library let me try
938. 2:32:09another one and then it um so if you want to see it in action without having to really dick around um go play with code interpreter at
939. 2:32:17chatgpt all the things like baby AGI and auto GPT and there's a new one that just came out that looks really promising
940. 2:32:25they're they're all still a little weird and they get into these kind of recursive Loops where they just can't they just sort of
941. 2:32:33what they do is they run your open AI bill up to really expensive things without solving anything so so they're all kind
942. 2:32:42of crappy right there but but um I I think autonomous agents are are ultimately where all of this goes like
943. 2:32:50right we're going to look back at the the fact that we're right now having to think about how do we craft a prompt to
944. 2:32:57get this thing to do our thing like that's going to be cute three years from now we're gonna like remember when we
945. 2:33:03used to write prompts by hand that was stupid wasn't it now they just do it all right here come the Stars don't move
946. 2:33:12uh can you recommend AI tools that generates a video based on a script not really based on a script but you
947. 2:33:21could you could break down your script and and generate images so runwayml is the tool runwayml.com I think hang on
948. 2:33:29let me just make sure it's.com yeah runwayml.com go to runwayml.com sign up for that Gen 2 is text to video
949. 2:33:44you can also generate an image in mid-journey and use the image as your first frame of a video sequence
950. 2:33:51and then you can generate a four second clip and then you can extend that four second clip four times so you can now
951. 2:33:57make 16 second clips from an image or from a text prompt with Runway ml so write your script break it down into
952. 2:34:08less than 16 second segments and go to runwayml and make it all and it's not cheap but it ain't ridiculously
953. 2:34:15expensive because you're generating video from nothing but go do that um I just spit out my drink I don't know
954. 2:34:24what I said but great that's that's I listen there's nothing better than a good spit take so now the New Yorkers coming out of Mr
955. 2:34:32Shannon oh I think this is what I was yelling at that's someone all right I gotta go to bed people um let's see use Runway this evening got
956. 2:34:41some really scary images yet when I asked for a 50 year old woman oh yeah yeah yeah these things are a bit rough
957. 2:34:47right now I might be addicted to your lives thanks thanks for packing to keep doing all right listen I'm gonna go pack
958. 2:34:56Monique if you're still there go to bed um I gotta go to bed too I will see you at what did you say 410 four when I
959. 2:35:04don't know I got I'll see Monique sometime around four in the morning all right um it's been a pleasure happy Saturday
960. 2:35:13night thanks for hanging out with me uh you always swell it really is cool um oh someone from a very good stream
961. 2:35:19tonight from Riyadh wow cool thank you Clark thanks for the tips been using pie Claude and chat for biostatistical tests
962. 2:35:29in public health oh that's awesome that's great you were here two or three nights ago asking about that that's awesome great
963. 2:35:37Pi Claude and chat yeah very cool very cool all right everybody um peace out have a good night USA loves Tick Tock bye everybody