Jan 19, 2024 - Unlocking AI Magic: How to Create Engaging GPTs for Fun and Learning

Transcript

1. 0:00you ready you [Music] [Music]
2. 0:20ready you and I get all Sunday morning here at home sky blue and C strong
3. 0:35[Music] true then I open my eyes to dream realized in front of me and I haven't got a clue what in the world is
4. 0:50happen [Music] me thank you Robert Rossy think I think I'm happy like first day summer vacation happy got to get a little rest and
5. 1:05relaxation happy on Sunday morning s [Music] TR Oh Sister when you cook that way sweet distraction rules my
6. 1:23day kitchen table just my brain from the sound of you and I rocking up by is yeah I don't understand why wasn't your of a man hundred years
7. 1:39ago if they was going to have to understand tip their hat to the man who [Music] scored think I think I'm
8. 1:51happy yeah CH Shannon champ Shannon on lead vocals fantastic [Music] a minor a minor [Music]
9. 2:15dog dog loves an A Minor no rapping we don't rap we're singing melodious tones
10. 2:31[Music] [Applause] [Music] I got to tune my guitar at some point maybe [Music]
11. 2:47never you [Music] done
12. 3:01[Music] well I heard there was a secret dog David played and he set by log you
13. 3:18don't really care for poopy do you well it goes like this the fourth to fifth M fall
14. 3:32m b king composing hallelujah hallelujah hallelujah you done
15. 3:50hallelujah hallelujah oh yeah Happy Date night everybody cam katkin from Cleveland what's happening cam is a new mod here
16. 4:02tonight so troll away she'll have to deal with it thank you for modding cam sorry so so here's the deal cam asked to mod like I
17. 4:14don't know a week or two ago and I said sure and I so I let her in the The Mighty networks Channel where the mods
18. 4:20hang out and then she couldn't ever mod like she like Tik Tok was broken she could not pin comments it was horrible
19. 4:30not quite as horrible as my hair but it was bad turns out I never set her up as a mod with most computer problems it is it
20. 4:41is usually a user error so all right oh what a fantastic way to spend a Friday night champing Kyle Tunes well that's in
21. 4:51in doses right you know champ you could listen to all night but uh you know it's like it's like an open mic night that no one asked to
22. 5:06have all right so Cam's in we're all good uh oh I don't see the mod icon on comm's Cam's name I think she I think
23. 5:13it's there I think she just pinned something I think she's good to go all right I just got back back in we were out doing a little Gallery
24. 5:26hopping and uh I eat like a quick little meal here so if I get hiccups that's what's going on and my glasses are dirty
25. 5:35looks like champ took a poop on my lenses happy Friday night everybody Welcome to the AI learning lab um the uh
26. 5:46I I'm also giving um uh quaffing advice for any of you uh looking to do something really nice with your hair and
27. 5:56I am the I am the before picture so I'm here for your for your inspiration for your for for your oh my God what the [ __ ] is going on up
28. 6:11there what are you GNA do and I don't is this shirt this shirt nope can't pin huh hang on hang on hold please where are settings
29. 6:26um settings moderators I know I set you up and you're not in there weird Plus cam add
30. 6:44confirm I've already done this Tik Tock try it now cam 1504 here for me Saturday the 20th so 1504 is I used to be a military brat
31. 6:553:4 3:04 in the afternoon Saturday at 3 4 in the afternoon right is that right steo don't use military time we're in America we can't handle
32. 7:07the [ __ ] metric system you think we can handle military time okay let's try it you you pinned that right cam yeah
33. 7:15all right so cam is now pinning away woohoo okay she's happy she's good to go everyone's excited except whoever did my hair they need a
34. 7:32demerit always military time Sherry D yeah I know if you're in the military I was just an Air Force brat I was when I
35. 7:41was a kid I was I was like I'm either want to be a pilot or a stunt man a Hollywood stunt man I went the acting
36. 7:47route my dad went the pilot route I went the acting route so you know and it was funny he always talked about oh I wish
37. 7:53i' had have gone into acting I'm like I wish I could learned how to fly [ __ ] super cool Jets but whatever I've got Sim racing now well I
38. 8:07used to have SIM racing now I've got a Tik Tok Channel this this time slot in my life used to be reserved for about 10
39. 8:15ft that way for my racing rig all right so here's what we're going to do I I don't know what we're going to
40. 8:22do we're we're going to talk about AI [ __ ] that's what we do here it's your gallery hopping hair yeah well it was
41. 8:28not on purpose purpose I just I don't know I don't know what happened I don't know what happened you know whenever my hair looks
42. 8:38this shitty it my fallback position is well at least you got hair you'd think if I had hair I would take some [ __ ] pride in it you know
43. 8:51but no no we've got I like this isn't even like '90s so I I sort of oscillate between 70s Carpet Sales man and and like you know '90s
44. 9:04agency hipster and then this is like some 1980s Flock of Seagulls meets Mad Men it's just bad co-pilot Pro came out oh yeah
45. 9:18co-pilot Pro came out that's interesting so co-pilot Pros 20 bucks a month um this is um being able to have
46. 9:29essentially gp4 inside Microsoft Word Excel all the Office 365 suite for individuals Enterprises had access to it
47. 9:39previously but now they've done General release I think Enterprises it was 30 bucks per seat per person I think
48. 9:46individuals is 20 bucks per seat per person so it seems like all these companies are kind of landed on 20 bucks
49. 9:52a month as the uh as the price so that's what I'm guessing um Gemini Ultra is going to be I think they're going to call it
50. 10:01Bard Bard Advanced is the is is what I remember reading 30 minute Hells indeed this is the time this is the time we do
51. 10:09the 30 minute hellos we'll talk a little bit of uh we'll talk a little bit of of AI stuff in the hellos here um I also
52. 10:19saw an interesting thing that I didn't believe it was real there was a sort of selfie video of Mark Zuckerberg talking about um llama
53. 10:303 and how many h100s they have and how many h100 equivalents they have it's pretty insane what meta is doing if this is
54. 10:39true so we can go look because it may not be true but um they apparently have 300,000 h100 gpus those things are like 20 grand a
55. 10:50piece or 22 Grand a piece or 30 grand I don't know they're a lot of money so they got 300,000 of them and then if you
56. 10:58combine all their other gpus that are not h100s they have sort of the equivalent of 300,000 more h100s just in
57. 11:07raw compute power so they have 600,000 h100 GPU equivalents that they are training llama 3 on that it's [ __ ]
58. 11:23insane like perplexity that just raised a billion dollars they um they bought 22,000 gpus for their server farm and meta
59. 11:38throw in 600,000 at this thing and then what they what Zuckerberg said is they're going to open source llama 3 one
60. 11:46of my thoughts was that they wouldn't who's using going to use a galaxy galaxy AI on the s24 that's another thing we could talk about is
61. 11:55that is the uh the Samsung keyote note where they they announced that new Galaxy um I think people will use
62. 12:04it I mean anyone who buys it the the uh they did this cool thing called Circle to search um with AI so at anything on
63. 12:13your screen you can scribble over it or draw a circle around it and it'll do an immediate search on it what's the best way to get into AI as
64. 12:22a Dev h well I I I would say it depend like a Dev doing what like the the it's
65. 12:37that's such a it's such a broad question it's a good it's a good question do you want to work for a research lab doing
66. 12:44sort of the hardcore math do you want to develop applications um that are Standalone do you want to do kind of front-end Dev and
67. 12:57um you know just just just be doing basic calls like I would think a a really good place to cut your teeth is is start you know
68. 13:09setting up your own open-source servers right set up a Google um collab instance with gpus on it and just go nuts on the open source
69. 13:23world just start learning all the different open source models I think I think as a Dev the more educated you can be like you
70. 13:31should understand what Lang chain is you should understand um some of the agent softwares uh like like agent GPT and auto GPT
71. 13:45um you should understand you know semantic search versus embedding and how you and and Rags retrieval augmented
72. 13:54generation if you're not ragging you're not developing these days you should follow David Shapiro on YouTube backend stuff for Google Etc
73. 14:05yeah I so if you want to do backend stuff I would you should you should be you should be go going deep into
74. 14:12autonomous agents Lang chain Rag and um get into some of the open source projects um David Shapiro's got two that
75. 14:22I think are are really interesting one is a cognitive architecture for AGI total open source you can join that
76. 14:30project and contribute to it um and then the other one is an agent spawning some sort of agent [ __ ] thing when when chat GPT came out with
77. 14:41the assistant API um building a whole bunch of cool [ __ ] on that so I I just go start building go get educated as
78. 14:49[ __ ] on the open source community and build build build understand the open AI apis understand Google's apis because
79. 14:58Google's apis are really interesting right now and they're going to get really interesting when when Ultra comes
80. 15:04out Google has a free course MIT has a free course Stanford has a free course meta has a free course I think meta has one um but I I
81. 15:20wouldn't I wouldn't worry too much about courses if you're already a developer like I just just start building [ __ ] just go
82. 15:29just go to hugging face and GitHub and look at the trending projects everything's AI every project on hugging
83. 15:36face every project on GitHub that are trending it's all AI projects just go [ __ ] pick three and just go dive into
84. 15:44them that's what I do anyway all right we haven't even started yet what are we going to talk about people what are we going to talk about
85. 15:53lovely image here from Danielle very nice little little Japanese zenar Garden or you can buy an AI certification for $997 you can make
86. 16:07money you can make money if you sell $997 certifications right now although I got to say if you're on the technical
87. 16:15side of things some of those certifications you know if you're if you're getting into machine learning and
88. 16:18Ma mathematics and things like that that's that's enough enough established that there are valid courses and
89. 16:24certifications and workshops on that side if you're on the business end of chat gptt where you're not so much a
90. 16:32developer that's where it's a lot squishier um but but even still even even that backend stuff is moving so
91. 16:39fast because what's happening what's happening in the open source Community is you know um uh open Ai and um anthropic and
92. 16:52Google they're all kind of leapfrogging each other these big tech companies are doing what they can to you know one up
93. 16:58one another and then the open source Community is sitting back there taking you know the Llama 2 model and they're
94. 17:06they're sort of mixing and matching all these different approaches that they're seeing all these different companies do
95. 17:12like mixture of experts right that was the whole thing came out on on chat gp4 using mixture of experts model and so
96. 17:20now mixl came out with that and so the open source Community is just going nuts on all these different models and so I
97. 17:27think that's if you're developer I think that's the place to be because open- Source projects are going to start leapfrogging commercial
98. 17:35entities and it's going to get wild wait there's not a plan for tonight's class there's never a plan I never
99. 17:44plan I do not read manuals my son Zach does it drives me [ __ ] baddy Dad we got to read the manual it's
100. 17:54a shovel do you know what that little Notch is for no and I don't care it's usually something cool the
101. 18:02shitty discovers by reading the manual I'm not saying it's not a it's not a positive impact on my life it's just come
102. 18:12on you you just you just you rip open the box you do you do not open it carefully so that you can put it back
103. 18:20together and return it no if you bought the thing you're into it you rip the Box open you take the thing out you you
104. 18:28break it because you didn't read the manual and then you take it back to the store you get in an argument with the
105. 18:35people that they sold you a defective thing and then the guy behind the counter goes you didn't read the manual
106. 18:40did you and you're like no no I read the manual this was defect it was broken like that when I got it and you get in a
107. 18:45fight with them and then you take your broken thing back and try to fix it this is how this is how you live life people
108. 18:53if you read the manual you miss that human interaction you're welcome the these are the kind of life lessons that other channels just
109. 19:05don't give you we keep our pinkies out cuz we're classy can you touch on what hugging face is barely so hugging face okay yes I
110. 19:26can there are kind of two main repositories right now of Codi kind of stuff there's GitHub which is a big code
111. 19:34repository so developers if they're developing stuff put that there hugging face was set up as a model repository so
112. 19:43um with stable diffusion for example which is an image generation tool there's all sorts of different fine-tuned models that people are
113. 19:50building on top of the open source model that stability AI put out and so they save them as these new checkpoints and
114. 19:57models and they upload them to hugging face so it started out as just a directory of AI models that you could
115. 20:04use for different tools it's since sort of grown lots of tentacles they've raised a lot of money there's now
116. 20:12runtime environments there where you can play with models so it's not just about downloading them now you can play with them
117. 20:19there's I think there's code repositories there it it both GitHub and hugging face if you are a non-developer are the shittiest
118. 20:29interfaces on the on in in in the history of the internet they're just there's just links everywhere it's just
119. 20:37like a sea of hyperlinks and if you don't know which one to click on well that's just cuz you're a clueless Noob
120. 20:44um and that's their attitude so um the one cool thing about hugging face though which is kind of fun so if you go to
121. 20:51hugging face as a non-developer if you're a developer you're on your own go go figure that [ __ ] out um let me lower
122. 21:00my brightness there let's go to hugging face so it's hugging face. Co and then when you first get here um here's models and that's like
123. 21:15super geeky just you know models and like the h94 I adapter dface ID you know I I'll tell you what when I was playing
124. 21:26with the h94 D P adapter dface ID it was the funniest thing no one's saying that okay so there's models and then there's data
125. 21:37sets so chat GPT prompts telat right then in the middle this thing with color these These are called spaces and so what the spaces
126. 21:47are are places where you can go play with models that that developers have set up where you don't actually have to
127. 21:54install anything a lot of times they're kind of they're over run with people using them and they're slow and you
128. 22:00can't do it but let's see segment anything web and basically the way they're ordered is whatever are the the ones
129. 22:09that show up at the top are kind of the most popular ones that people are playing with right now so this is
130. 22:16segment anything so we can upload an image here if you haven't seen segment anything it's pretty slick yeah this this will be a good
131. 22:29one and I I haven't played with this before by the way so reset image extracting image embedding left click positive points
132. 22:40right click negative points oh yeah okay so see wait so what segment anything does this is a this is a free model that meta put
133. 22:54out is it automatically goes into this image and it identif ifies different objects within the image so like there's the floor there's
134. 23:05the main subject there's this dude there's that dude's shirt there's that dude's pants let's see there's her dress
135. 23:16there's her shirt that's that lady that's that dude there's his face there's his hat look each of these letters is
136. 23:27segmented now so so you're like yeah but Kyle what is this so this is just basically a thing that meta put
137. 23:33out that every every one of these things that was highlighted a developer would be able to identify things within this
138. 23:40so if you if you see those crazy um you know parking lot things that identify you know here's cars and here's parking
139. 23:49spaces that have come available it's it's using this kind of Technology but this is just a super easy a super easy thing that meta put
140. 23:57out for free so anyway um so that's what hugging face is you can go find models there if you want to geek out it's a
141. 24:03place where you can go you can run deep wi with all of this stuff a good place to start is is with Reddit so go to
142. 24:11Reddit and then go to subreddits for things like stable diffusion or open- Source AI or hugging face or right
143. 24:18within subreddit there's all these different communities that you can play with all right Kyle which is the best
144. 24:23image generator you recommend good question um what's your name Z D um Let me let me go show you something else
145. 24:32first and then and then I'm going to get you there so so hang with me so the first place I want you to go is there
146. 24:40the salon. a okay the salon. a don't use https it doesn't work because I set up the domain as a lazy man and I didn't I didn't check some
147. 24:54stupid [ __ ] box that said use https but where's the manual for AI the salon. a yeah there you go so that's where I'm
148. 25:02taking you the salon. A and then what that's going to take you to is the salong and how you pronounce it is the
149. 25:16Fon a it's not Salon Salon is where you go get a a lovely hairo like mine but the phon is where you go uh get your [ __ ]
150. 25:29together with AI okay so when you come here it's a link tree um the first link is to the community that's where we're
151. 25:34going to go that's we're going to talk about image generation recommendations um there's also a couple of links here
152. 25:40for the GPT for good event directory in the YouTube playlist the GPT for good event um was three weekends ago Ann
153. 25:49Murphy and I um and Vicky Baptist put together this event that we our goal was to have people build 100 gpts in a 24-hour
154. 26:00period for nonprofits this community got together and kicked ass and built more than 150 so that's that's what these are
155. 26:07this are the directory that Vicky built um these are the YouTube playlist of the sessions where Anna and I taught and
156. 26:13then there's also a uh What's this called merch store uh that supports the salon so the salon doesn't cost you
157. 26:19anything the salon community so go to the go to the salon community and sign in join if you're if you haven't joined and accept your
158. 26:27cookies accept your cookies people we've got to track you with your cookies we're going to grab your cookies we've dropped
159. 26:34cookie crumb crumbles on your just just do your cookie just accept all just don't read the terms of service we
160. 26:43everybody owns your data it's gone it's gone you gave it up 15 years ago all right so we're going to log in here
161. 26:53and I am quite a fond I I'm quite fond of this Comm community so this community is growing quickly the first thing I'd
162. 27:00like you to do when you come here is go to welcome to the salon you can ignore that dude I'll show you what that's
163. 27:05about in a second that's some lovingly troll love the construction paper and the cool cutout for the microphone
164. 27:11exactly that's my microphone cut out and here's my trackpad so I can scale things up and down look this is this is some
165. 27:18serious this is some serious construction paper nothing I'm telling nothing but the best production values for 2024 for the AI
166. 27:34salon and look not for nothing for you people look how much black construction paper I ordered just so that I could do that one thing
167. 27:47look at that look at that what am I going to do with this it is this is this the burden this this the burden of doing a worldclass
168. 27:59video production Studio you got to deal with pounds and pounds of black construction paper okay so when you get to the salon go to
169. 28:09welcome to the salon and read read up on who we are and what we're about label it and put it on the merch store this is a high class a
170. 28:19high class Salon if ever there was one exactly exactly people um and then the other thing that I'd like you to do is
171. 28:27please please please um check out our values values what is this some kind of successories poster you gen xer no
172. 28:36values are actually really important because as this community grows like right now this community is quite
173. 28:42remarkable and as it grows I want it to stay remarkable how do we do that well we actually understand who are the
174. 28:49people that are here and we talk about it so like we want people here that are curious and are actively seeking answers
175. 28:56to things not cynical people people that are like it's okay it no we want people that are curious we want people that are
176. 29:02generous generosity is one of the values how do we how do we learn well people within the community they learn [ __ ] and
177. 29:10then because they're generous they share what they learned and this is happening right now in this community there's not information
178. 29:17hoarding going on maybe a little bit you know people got people got to keep their Competitive Edge but for the most
179. 29:24part they're out there going here's what I learned here's what I couldn't figure out here's where this thing sucks and then someone
180. 29:30else goes oh I figured out this piece of it and then someone else goes oh I figured out this piece of it and then they they you know explore
181. 29:41try new tools together so that's exploration is a value then there's collaboration hey what if we work
182. 29:47together hey you want to put something together that's happening a lot in here inclusion because everyone's coming from
183. 29:54everywhere no everyone's being affected by this AI stuff so we don't know where people are going to come from so we
184. 29:59better be open to that and then empathy understanding that right now people are scared of AI they're angry about AI
185. 30:09largely because they don't understand it there's a tremendous amount of misunderstanding and misinformation
186. 30:15about AI good morning from Greece Angelique welcome welcome welcome that's fantastic so empathy is really important
187. 30:27understanding that that people are going through [ __ ] and if we can if we can understand that and have some empathy for where they
188. 30:38are then we can say yeah yeah this stuff is scary I can understand why you're angry want to see something
189. 30:45cool and then you show them something cool and if you do it right they have their first Kevin mallister moment which is this
190. 30:54moment if you remember Home Alone generative AI when you get into it is a series of rolling Kevin mallister
191. 31:06moments where you just get your mind blown like one thing after another and you think you've seen it all and then
192. 31:13you go look at one of these image generation tools and then you think you've seen that all you're like oh I've
193. 31:17got mid Journey mastered and then you go over to Korea or Deco here and it starts rendering images in real time as you're
194. 31:24typing your prompt you're like the hell is going on here so anyway that's what you do when you first
195. 31:38get to the salon is read about who we are and what we're up to and if that resonates with you then this is likely
196. 31:46your people and you can kind of think of the AI Salon as as a support group for people that are curious about AI one of
197. 31:52the things I'm hearing a lot lately is oh my God is this ser seriously a place I can talk about AI cuz every time I
198. 31:59talk about AI with my friends they tell me to shut up and I'm a weirdo and they don't like me anymore is this can I talk
199. 32:05about it here is it going to be okay and we welcome them into the circle Point them to the coffee machine and
200. 32:15tell him to smoke a cigarette and start making images of do it's a support group so so come in introduce yourself
201. 32:25so that's the next thing to do so first thing is check out what we're about the second thing to do is in the community
202. 32:30section the first little tab here is introduce yourself so my request my strong strong strong request is to go introduce yourself wait
203. 32:40there's a coffee machine um and be generous don't just say hey I'm here I'm here eat it losers come in and tell us what you do tell us
204. 32:52what your background is tell us um where you are on the AI Adventure I'm just getting started I've been really into it
205. 33:01for a year but nobody knows it can I talk about it here without being humiliated what wherever you are I've been in it for 30 years but this
206. 33:11generative AI stuff's different so I'm trying to figure out people that are figuring this out better than I
207. 33:20am and then another fun area and I am going to get to the person that asked what's an image tool recommendation
208. 33:28here's one of the things about this channel it's really [ __ ] annoying if you just want the answer the nickname of
209. 33:38this channel well there's oh Teresa just subscribed for the first time thank you very much I really appreciate the
210. 33:43support makes a big difference um one of the nicknames for the channel one of the many nicknames for the
211. 33:49channel is chat add so so we're neuro neuros spicy friendly in here and uh I'm I'm a laminated card carrying member of
212. 33:59the ADHD club and so I'm all over the [ __ ] place but I will eventually get there and if if if it if I was heading in an
213. 34:08important Direction and I get lost the mods actually put me back on track they're like hey Kyle don't forget to
214. 34:13talk about that thing you promised so I'll eventually get there but um the show inel channel in the AI Salon is an absolute [ __ ] blast
215. 34:24because what what's what what the show and tell Chanel has turned into while these lives are on is like a
216. 34:31total [ __ ] jam session where if I say something stupid people will make images about it or if I say something that
217. 34:38deserves being trolled they'll make stupid images about it or or people are sharing gpts that they've made or
218. 34:46they're they're sharing songs that they wrote onso or or they're sharing tips and tricks and things like that um so
219. 34:53it's just a really fun place so it's oh look here here's the AI Salon welcome to the AI salon right this is the salon not the f
220. 35:07um this is bad hair day week thank you Mr it it's it's uh it is it is flattering you got to have thick skin if
221. 35:20you're going to if you're going to have a Tik Tok Channel at 58 because this is the [ __ ] you're going to
222. 35:29deal with the mission starts at 1505 the hell time is that that's how to stump the
223. 35:44Yankees backstory to follow oh Joker Joker J Joker steo please no backstory oh that's a solid one look it's sort of
224. 36:00a what that's sort of what my hair is looking like blandon's debut champy singing Me guitaring and the and the Fantastic ble of
225. 36:10BL that's beautiful a salon Late Show oh man these are cool g so anyway so go go put some some wacky ass stuff in
226. 36:24there beautiful oh that's nice actually like that logo Cindy [ __ ] welcome Cindy [ __ ] what's happening I'm excited about our project
227. 36:38we have a Sydney [ __ ] and I are working on a super cool project that she initiated that is it's going to be so
228. 36:44cool going to be so cool people I'm blind this channel has helped me a lot in my AI Journey Jeffrey Lombardo thank
229. 36:52you so much that's awesome that's awesome I'm I'm listen one of the things I'm so excited about um with AI is accessibility like on all
230. 37:00levels it's we were talking about this it's funny on the in our conversation today talking about the project Cindy
231. 37:07and I are working on accessibility was a huge piece of the conversation and it was everything from you know foreign languages to to
232. 37:17you know if if someone's blind or if they're deaf or if they're cognitively impaired um there's all sorts of
233. 37:23research going on where people are now you know kind of thinking you know thinking things into existence because researchers are
234. 37:33finding ways to take brain patterns and turn those into computer actions um using generative AI like like just the
235. 37:40accessibility stuff with the these tools are just insane and like we we've barely gotten started so so thank you for that
236. 37:47I'm I'm really excited I'm glad this this helps and you know it's like it is it is still kind of random so just hang
237. 37:56in there there I'll eventually get to the point okay so that's show and tell so there's also a water cooler discussion
238. 38:03Channel there's a news channel I'm going to come back to tools and resources um there's a Shameless self-promotion
239. 38:09channel so if you want to if you've got some event you want to do you can put it there and then there's also guilds and
240. 38:14the guilds are really cool what the guilds are within the salon are are sub communities so Stacy and Matt Mr K um
241. 38:22from here on Tik Tok just started um an education Guild there's a video Guild that's they've got their first meeting
242. 38:28coming up um Healthcare Pate who works at Google has got the AI mechanics Guild um if you want to go geek out there's um
243. 38:36a business Guild writer Guild Art Guild and then uh Peter Kaminsky has an AI 101 Guild um and today Peter from the AI 101
244. 38:45Guild and Leah from the Art Guild got together and they put on a mid Journey class and Peter's like I don't know I
245. 38:53don't know how many images he's made but it's probably north of 100,000 using mid Journey so this is a guy that's that's
246. 39:00in hey I need to verify that my monthly subscription is going directly to champy as originally agreed I'll get him in
247. 39:08here champy champy do you get all the chicken treats you need buddy huh and well and listen technically
248. 39:18guitar strings go to champy you because I can't play a chord without hearing that [ __ ] dog sing so so it's chicken strips and it's guitar
249. 39:29strings and an occasional whiskey you know you got to keep the dogs loose all right I've gone from wiring boards and
250. 39:37key punch cards to learning AI what a journey yeah I so so let's see Michael both boor Borth is that right um
251. 39:48yeah so so it's funny my um is Peter's Class A recorded one I assume they recorded it Peter's pretty good about
252. 39:55that stuff so I would say go go into the uh uh I think it was in the Art Guild but go into art or or Peter's 101 Guild and just
253. 40:04ask um talking about Punch Cards so when when I when I first started college at Penn State here I'll flip back over for
254. 40:14a second when I first started college at Penn State I I knew I enjoyed acting but I'm like I can't you know you can't go
255. 40:22into acting like so but I was good at computers and I'm like I'll do computer science and so what I knew of computers
256. 40:29back then was like the Apple 2 and and uh the Mac was was you know rumored to come out and I I learned to program on a
257. 40:40trash 80 you know personal computers so I walk into my first computer compai 101 class at Penn State and I walk into this
258. 40:50room and it's full of like gray and Military green IBM M 3780 mainframe computers and this dweeby this dweeby
259. 41:03little nerd walks in with a stack of Punch Cards like it was like up to his chin he was like holding holding him
260. 41:10down stack of Punch Cards and he walks in and he goes this is a computer program and he goes he goes what what
261. 41:17you'll be doing is you'll be typing lines of code onto these cards and you run them into these machines and then
262. 41:24they generate reports over here and it was the big wide green and white paper and and you'll be running reports for
263. 41:31big corporations for the rest of your life and I was like the hell I will and I changed my major to theater the next
264. 41:38day I was like I am [ __ ] out if that's your definition of a computer and if that's my definition of my next 30 or
265. 41:4540 Years nope I'd rather be [ __ ] broke so anyway Kyle likes to say I work at Google but I also have a PhD in compai exactly
266. 41:59um yeah P's got it going on so so Pat's mechanics guilds really cool if you're into the geeky stuff he can he can he
267. 42:05can get you up to speed and and help explain some of that stuff especially for you know if people are in batch
268. 42:12printing yeah exactly yeah yeah because you had to you had to um you had to time share the computers right and so if you
269. 42:20were a freshman you you had to come in at 10:00 at night cuz the the Elder students got the earlier slots yeah
270. 42:29exactly I'm really interested in seeing these tools in the hands of the elderly oh my God the elderly and and dementia I
271. 42:38lost my dad to Alzheimer's and it's like I just I I just the idea of you know people that are sort of lonely don't
272. 42:46have someone to talk to that can talk to a familiar voice that knows all the stories that is infinitely patient that is there
273. 42:55247 holy [ __ ] it's amazing um okay so back to this so check out the guilds the guilds are super cool let me
274. 43:06go to AI tools and resources so I I am going to get to image generation tools and and what my recommendations are um
275. 43:15this list is a list that I personally curate it's the the AIS Salon uh you know top recommendations where I start is so if
276. 43:24you haven't tried chat GPT if you're in this Channel and you're just here cuz you know my supermodel good looks and
277. 43:32and you know clearly I've got a worldclass Italian hair quaffer um if you're just here for the for the eye candy that's fine but if you
278. 43:42actually want to learn something about Ai and you have not played with chat GPT if you're sitting on the sidelines right now drop everything
279. 43:54seriously go to chat. open.com and start playing while you're listening to me ramble start playing with chat GPT even
280. 44:02chat GPT the free version of chat GPT is still better than most of the other large language models out there it's
281. 44:08incredibly capable it's very very fast it's it's incredibly capable um and then if you can afford the 20 bucks a month I personally think
282. 44:18if you want to be in the AI game at all 20 bucks a month for chat pt4 is the bargain of the century like the amount of stuff it does for 20
283. 44:29bucks a month is insane but you actually have to learn what it is so this list not only tells you you know that it's
284. 44:36there but it says here's all the functions within it that you should be learning and you know I go over those
285. 44:42you you know come back to this channel regularly because I go over this stuff all the time so start with chat
286. 44:49gbt then below that I've got eight different large language models that I think are worth playing with um you've got
287. 44:57the big four that I think are the big four right now Claude Bard perplexity and Microsoft co-pilot which that's
288. 45:04basically chat GPT 4 for free um perplexity is uh really interesting they just raised $75 million from andron
289. 45:15Horowitz B is Google and that's about to get a real kick in the ass quality-wise when Gemini Ultra comes out and Claude
290. 45:23is another very well-funded uh direct competitor to chat GPT uh a bunch of X chat GPT people the people that created
291. 45:32gpt3 went off and created anthropic which makes the Claud and then the specialty llms you got Pi grock llama 2
292. 45:40and po all worth exploring okay so just chat GPT if you haven't played with this stuff before you could spend the next
293. 45:49month or two just learning what all these things are and what I would recommend is just have fun with with them don't take it
294. 45:56seriously just go have fun make it do stupid [ __ ] do what what we call in here parlor tricks make it write poems and
295. 46:03songs and tables comparing ridiculous items and create a meal plan that you know features Gummy Bears I like just
296. 46:13[ __ ] throw [ __ ] at it and see what it does and you'll learn a lot survey is up okay we're going to take a survey now
297. 46:20hang on here people all right we got four minutes so we're asking what gen generation are you so so the the the
298. 46:30survey click the the little it looks like a little blue and white checkbox over in that corner and you got four minutes and so
299. 46:41we want to know are you a boom wait we got Gen X in here twice I assume you wanted gen Z so the D says Gen X but D should be gen
300. 46:53Z I don't know if you can edit if you can edit that but Boomers gen xers gen yers gen zers or I don't
301. 47:03knows all right so just understand like if you're in this channel you should be smart enough to understand that B and D
302. 47:10are not the same B is Gen X D is Gen Z okay glad we cover covered that up definitely Boomer start over uh that's up to you
303. 47:22guys I don't care I could end it you want me to end it and start over all right fine I'll do that hang on I'm going to end this ending
304. 47:32it all right there you go so whoever is doing the uh doing the surveys do it again just make sure gen Z is right I
305. 47:41think you want a gen Z for D um okay so let's keep going when that comes back up I'll pull it up so so you've got chat gbt then you've got
306. 47:52these eight large language models below it then then you've got image generat so here's the deal with image
307. 47:59generation these things are different enough right now that the the answer to which one's best is what are you trying
308. 48:05to do and where do you work and and what are you looking for because they're very very different um dolly3 is the image
309. 48:16generation tool that lives inside chat GPT and inside Microsoft co-pilot so the nice thing about Microsoft co-pilot is
310. 48:23that you get Dolly 3 for free um if you're in chat GPT you need to have the $20 a month subscription to get
311. 48:32Dolly 3 but I would argue that dolly3 inside chat GPT is a significantly better tool than what's inside Microsoft
312. 48:40co-pilot just because of a lot of reasons um the nice thing about Dolly 3 it's good with text it's not perfect but
313. 48:50it can do text so you see a lot of those images in show and tell that have like hey Kyle what's what's going on or
314. 48:56champy this or whatever that's that's dolly3 for the most part mid journey is getting better V6 is better at text but
315. 49:03it's still not great but mid journey is if you want high quality images mid mid journey is still the gold gold standard
316. 49:13okay Korea and Deco here are both leaning into there's there's a new thing that came out from stable diffusion
317. 49:20called sdxl turbo which is instantly rendering images while you type prompts and I I can show those Korea and Deco here are almost more
318. 49:35like like jazz you know that then like like let's how do we do this so all right okay I'm trying to hang hang on
319. 49:49okay okay mid journey is like speed metal you got to you got to be really technically good like uh like uh pfia
320. 49:58kind of you know stuff that's mid Journey um um doll3 is like Americana like simple you know good songs good just singal songs good
321. 50:09storytelling um Korea and Deco here are more like Fusion Jazz they're like jazz they're they're you you know wacky time
322. 50:19signatures and and things like that Leonardo actually so so Leonardo is built on top of a bunch of stable
323. 50:26diffusion tools it can now do realtime image generation it can do traditional stable diffusion stuff you can train models within
324. 50:37Leonardo um they give you 150 credits a day which is it it works out to about 10 images a day it's not as much as it
325. 50:47seems but it's still you can do you can use it for free um and it's also doing video now so uh St stable video
326. 50:55diffusion so Leonardo is a really interesting tool cuz there's a lot there and then you've got Adobe Firefly and Adobe Firefly is
327. 51:08um hasn't been trained on any copywritten images so if you work within a corporation um you can use images that
328. 51:19are generated in Firefly and Photoshop generative fill and all that commercially because there's not there's
329. 51:26not as big a risk as you have with something like mid Journey which clearly has been trained on on you know artists
330. 51:32images because somebody leaked the big long [ __ ] list of artists that they trained that they trained mid Journey on
331. 51:41so so so you know if if if you've got you know ethical issues with how those other things were trained then Firefly
332. 51:49is really good or if you work for a corporation that's super Conservative then it's really good now is it as good good
333. 51:57no why cuz it wasn't trained on as much art so everything's a balance right now now if you're like if you're like but
334. 52:08Kyle that's too much can you just tell me which one's the best did you see my news post about Adobe and mighty
335. 52:14networks I did and it's um what did they I I forget what it was it's hang on I can go see it wait what's that hey I learning lab
336. 52:29um hang on Danielle yeah there was some news about some I think Firefly became commercially available and it's free is isn't that right
337. 52:51Danielle oh mighty networks is being weird it keeps resetting anyway all right whatever I'll figure it
338. 52:58out monetize images free Firefly that's what it was yeah so so go check out there there's some some Adobe
339. 53:05announcements yesterday but one of them is Firefly is now free which is great it didn't used to be you had to have a a a Creative Cloud
340. 53:13subscription so there you go some of these you know some of the prices of these things are trending toward free oh
341. 53:18that's actually a good point okay so Dolly 3 um is free within Microsoft cop pilot it's it's included within the 20 bucks a month for
342. 53:28chat BT mid Journey cost you money I think the minimum is 10 bucks a month Korea is absolutely free which is pretty
343. 53:36amazing because it's a powerful tool and it's got more than their real time engine it's got background removal and
344. 53:42it's got an upscaler and it's all free Deco here is free now these things that are free right now here's a little
345. 53:50software economics lesson for you the reason they're free right now is they're trying to get you to use their tool
346. 53:56instead of mid Journey or instead of some other tool so they're they're free initially and then basically what
347. 54:02happens is this when they burn through their series a or series B money then the series C people come in
348. 54:10and go yeah about how much money you're blowing uh let's go on ahead and start charging people for that so these are
349. 54:16not going to be free forever so so use them while you can um okay beautiful so we have a new oh we got 31 votes okay so in here
350. 54:27tonight out of 31 votes we've got um 30% are boomers 50% are gen xers which makes sense that's who I am 12% are gen yers
351. 54:416% are Jen ear welcome jenz ear U 3% don't know which is one person that's okay um but but welcome that's pretty
352. 54:50cool that's that's kind of what I that's about what I I would expect I would love to see see some some more younger people
353. 54:57in here but you have an old [ __ ] on you know leading the channel so you know I think young people are like why is
354. 55:04grandpa on Tick Tock flick flick flick but you know we have people in here like Rudy from South Africa he's a
355. 55:13young a young sprightly fell he's in here regularly so if if you know some youngans in your life tell them to come
356. 55:20in here tell them this is an old guy he's not bad he's got decent energy in fact do me a favor now we got
357. 55:35139 people in here do me a favor right now share this live either with Gen Y or Jen zers so some young people in your
358. 55:45life especially if you know it's going to piss them off because I've got 23-year-old twin boys upstairs they hate
359. 55:51this AI [ __ ] so so share this thing with with uh with young people in your life or with people in your life that you
360. 56:00know don't want to hear about AI anymore that that's who we want to get in here all right trained um aren't we all
361. 56:11trained on other people's Arts yeah so Brendan you know BT comp KSU what you're talking about is the reality of of what's actually happening
362. 56:22the the the larger these training sets get the less it makes sense that every individual artist within it
363. 56:33got to be compensated because the you know their art was used in the training set there's a point at which these
364. 56:40training sets get so big that they've been trained on all of the output of all of the humanity and and and the the
365. 56:47large language models and the image models and the music models and the animation models and the video models they all essentially
366. 56:56become like the global Library of Congress right or like the global Museum or the global library right that we've all contributed to and
367. 57:08we all get to benefit from so there's absolutely an argument that says that 130 people and eight or so robots
368. 57:16which ones I'll never tell p P's good so so I was in here the other night bragging about how I can absolutely 100%
369. 57:24% guarantee no one will ever jailbreak your GPT and by the time I finished that sentence Pate had jailbroken my GPT
370. 57:32which was really awesome of him to do I'm but I'm actually serious I'm I'm glad he did it because I was not making
371. 57:39a promise what what I was actually saying is assume that anything that you put in your GPT will be had by the
372. 57:46public if they really want it and then and then Pate made it so proved that that was the case I made a Luma to show
373. 57:54what it's capable of but it was at work and scared of HIPPA oh yeah yeah I know the listen the hip the Hippa stuff is
374. 58:04scary right now we've still got I think we've probably still got we we'll see what open Ai and and Google do I I I I don't think Google's
375. 58:19going to address any of this [ __ ] um gp5 could address some of this stuff where they will get more reliable and they will start to have
376. 58:32reasoning and God help us they have some better privacy and security so my my gut is telling me that we have
377. 58:44as as like this Collective little rag tag bunch of people just trying to figure this [ __ ] out probably 18 months
378. 58:52before the tools themselves are stable enough and secure enough that big companies really lean into them healthc care companies are probably
379. 59:03going to be 18 months after that right so so so we're probably three years out from healthcare companies really leaning
380. 59:11into this generative AI stuff they might lean into it on the business process side but anything on the patient side is going to be
381. 59:18slower but what that means for all of us is that the opportunity for all of us if you just get all you need to do in the
382. 59:25next 18 months is simply get AI literate and what that means is be curious like honestly dig through this list that
383. 59:36I have on the AI Salon start with chat GPT learn it play with it have fun with it go make some you know some uh you
384. 59:44know figure out what are the differences between like perplexity is really good at research and Pi's really good at
385. 59:51conversational stuff and grock from X is not good for [ __ ] um and how do you learn that well you just go play with it
386. 59:59and you go figure it out and then and then start playing with the image Generation stuff and you'll figure out
387. 1:00:04what this is good at and you'll have a lot of fun and you'll get inspired and you'll realize oh I've always wanted to
388. 1:00:09be able to be an artist and now I can be or whatever it is and just play and then do the same with the video generation tools Runway
389. 1:00:17ML and pabs or the two two biggies for you know generating content and then hey Jen and D D- ID are good for Avatar
390. 1:00:26creation and then visa and I should probably add what was that one that that we put up the other day um another video tool I don't know but
391. 1:00:39visel is good for now um and then there's music and audio tools and then there's 3D worlds and objects and then there's people to
392. 1:00:46follow and and the the four people that I have here are there more than four to follow [ __ ] yes there's a ton more to
393. 1:00:52follow but these These are like solid like will deliver the goods week after week kind of people all
394. 1:01:02right if you're not a robot click on all the boxes that contain stop lights that's good steo but anyway like just
395. 1:01:11just being in the conversation in the next 18 months and getting curious about this stuff and kind of understanding
396. 1:01:17what these tools make possible is going to yield you opportunities you cannot imagine right now and I can't imagine
397. 1:01:26right now today I had my first live on Tik Tok Mr it how do you do it Kyle I understand now all yeah so here's the deal Mr
398. 1:01:39it when I started doing these lives so so so here I'll tell I'll tell you how I did it um I like this image this this is this
399. 1:01:53is where we all are right now we're we're we're in a we're in a little inner tube uh with with our pet llama uh in
400. 1:02:01the middle of the North Atlantic Ocean in the middle of a storm that's this is where we all are right now um but we're
401. 1:02:10in there together that's what the AI salon's about is about doing it with other people that's what this channel is
402. 1:02:15about here's here's what I did Mr it is I have I have a thing I can do in my brain I suppose it's a talent where I can just flip a switch
403. 1:02:32and then I'll just commit to something and I kind of commit to it blindly ah Winston good to see you thank
404. 1:02:40you for that llama I love that llama the by the reason there are so many llamas in all of these images is because for
405. 1:02:48the first for the early days of this live every time I would say llama 2 or llama one from meta Winstead would throw
406. 1:02:55up that dancing llama and so it became one of the official animals of the uh of the AI learning lab um but the way I did this live was I I
407. 1:03:06said well for the most part when I get home after work I eat dinner watch a little TV and then I go do Sim
408. 1:03:14racing and I was doing like an hour and a half 2 hours of sim racing a night to the point that I got um Sim RAC 's heel
409. 1:03:25which I don't know if you've ever had it but it really sucks it's basically like a super tight Achilles tendon that that
410. 1:03:32makes my heel hurt um and it like it's still pretty bad and I haven't done it in a while um and so I just said what if
411. 1:03:42I give up Sim racing and just do this do these lives and so I said [ __ ] it I said I'll just do it every night and I'll tell you Mr it the
412. 1:03:56probably the first two months I had regular anxiety and panic attacks either before the live or during
413. 1:04:06the live or just and so I just said [ __ ] it I'm going to do another one [ __ ] it I'm going to do another one and
414. 1:04:13sometimes I would sit there for like 10 minutes the little red live button was sitting there and my heart would be
415. 1:04:20racing and I'm like I don't know what to say and I I try to be doing homework and coming up with plans and then I just go oh
416. 1:04:31[ __ ] and that's just just do it like it doesn't [ __ ] matter it's Tik Tock this stuff is ephemeral I mean I'm saving these things
417. 1:04:45on YouTube but I'm I'm really saving them on on YouTube because I know that what's coming from a technology perspective is is going to be
418. 1:04:54the ability to take a body of work like 2 to three hour nightly videos for 6 months or a year and do some really
419. 1:05:02interesting [ __ ] with that I don't know what it is yet but but that I want to play with where do you go to learn about creating gpts do
420. 1:05:13channels I I don't know I do not know what you mean what do you mean channels you're an inspiration Kyle thank you
421. 1:05:29I I was I was on a call today our [ __ ] our story Vine app for this one client is just it's just like Mal behaving it's just it's being an
422. 1:05:42[ __ ] and so we're on the we're on the phone today doing a tech troubleshooting call with this lovely young woman from
423. 1:05:52Texas um who who was just super sweet and trying to help us out and at one point I said something and she said she
424. 1:05:59said um how did she say it she goes she goes um she goes well I hope this helps sir and I was like like don't call me
425. 1:06:09sir she was taking her back cuz she's from Texas they're like that's how I was raised I call people sir uh so yeah when
426. 1:06:17I hear him in Inspiration it hits me kind of like that like sir like custom G P PT resources in the salon I'm going to send you more newbies
427. 1:06:28I talked about wait what was that I don't know what I'd be doing right now if you hadn't started doing this so grateful for you oh thank you
428. 1:06:37very much Danielle I appreciate that um so grateful you pushed on yeah listen I mean here here's the deal here's the H
429. 1:06:44the honest to God truth of why I pushed through that is I've been in this movie before right I started one of the early digital agents in
430. 1:06:551994 and I also started a group called the worldwide web artist Consortium which was a New York City based you know
431. 1:07:02anyone who's interested in the worldwide web join this little group and that little group turned into like a 2500
432. 1:07:08person you know Juggernaut that anyone in the New York City they called it silicon alley at the time right silicon alley reporter that's me
433. 1:07:21and my co-founder of agency.com um anyone who was in at all interested in the web was a part of this group in New York City and
434. 1:07:34we would meet every other week in the Sony building on Fifth Avenue we and Sony gave us this massive it was a conference room that held 300
435. 1:07:45people and we would regularly fill it the the table in the center of the conference room sat 150 people and then
436. 1:07:53we had chairs around it we would fill that room regularly so I know the value of community when technology is this early
437. 1:08:03in raw and so so that's what this was was like I just I feel like I need to get as many people as possible curious
438. 1:08:12about this [ __ ] not afraid of it or if they're afraid of it just willing to go [ __ ] it and play anyway and then I want
439. 1:08:18to find a place for them to coales and we're in a different time now so we can do that online so the AI Salon is now
440. 1:08:24this Global community and it's it's just amazing and and so is this channel it's Global hey dude that's Kyle these Tik
441. 1:08:32toks keep me going ah thank you very much lunatick you're I mean you're kicking butt I'm I'm really just super
442. 1:08:39impressed with what you're doing it's it's just amazing um I remember the first time I was called Sir now about 25 years ago I
443. 1:08:48was looking around who what exactly I definitely have sir you know sir sideburns and a sir hairline and and I
444. 1:08:58this is sort of more like a ma'am haird but you know whatever it did not look this bad for most of the day I will have you know
445. 1:09:10uh I do the same thing when someone calls me ma'am I know it totally threw me like and I turned beat red they were
446. 1:09:17they were both like oh my God you turned red I was like cuz I just it was just such a an instinctive thing don't call me sir man I felt
447. 1:09:27[Laughter] bad oh man uh best tool for flowcharts and org charts um I don't know there's a bunch of
448. 1:09:43plugins for that I don't I don't I don't know that world I I would think at this point there's probably some dedicated
449. 1:09:50tools for that stuff that I've been doing you know flowcharts and or charts for a while that have now added some AI
450. 1:09:57stuff I just don't know them I I would say go to um go to Mike diesel go to futur
451. 1:10:22pedia.com like I like I have yet to see something inside a large language model that does anything sophisticated with
452. 1:10:31flowcharts but I'm not looking for that so there might be something out there I don't know about all right so what do we want to do tonight
453. 1:10:47um so oh someone asked about making gpts maybe we'll make a GPT tonight I can could do a little bit of
454. 1:10:54uh it's already 10:00 damn oh I started late that's why it's Friday night though [ __ ] it let's party get my get my iced coffee going
455. 1:11:06let's get some energy shall we you can make money with ch by the way this thing you can make
456. 1:11:23money what is this about I don't bring it out a lot and it's it's kind of an inside joke at this point but here's why here's why this exists
457. 1:11:38um there's a lot of Tik Tok channels out there and there's a lot of channels out there there's a lot of this on there's
458. 1:11:44not so much on LinkedIn there's a fair amount on Facebook there's a fair amount there's a lot on YouTube and there's a
459. 1:11:51lot on Tik Tok and it's idiots going you can make money with ch and they lean into the screen like this and they twist their [ __ ] heads
460. 1:12:08around and it just it grats in my soul every time I hear it because it I think it it is doing everybody a disservice how they're how they're
461. 1:12:22making their money is selling you courses on how much money you can make with chat GPT they're not making money with chat
462. 1:12:30GPT now they likely use chat GPT to write the shitty course that they're selling you but how are they making money
463. 1:12:41they're not making money with chat GPT they're making money with the fact that people want to get rich
464. 1:12:49quick so they've figured out a problem to solve and they found a new to tool to solve it and they put together these slick
465. 1:12:59[ __ ] channels you make money to sucker people in and so it does a disservice to the tool it does a disservice
466. 1:13:13to I one of the things that I think that that attitude that tool first attitude does is it scares people away
467. 1:13:24it makes generative AI seem like nft Bros you can make money with nfts you can make money with blockchain you can make money with
468. 1:13:41ch no you make money by solving problems you make money by being creative in how you help other people do
469. 1:13:55something and if you do that good enough they pay you money it's called business there is not a [ __ ] shortcut for it period end of
470. 1:14:10story just not occasionally someone will trip onto something and they'll be the first one to generate a 100,000 t-shirt designs
471. 1:14:21and get that up on print of and you know overflood the market for a week and a half and take the screenshot
472. 1:14:30that they probably photoshopped anyway and said I made $100,000 in one week with ch JP just give me $29.95 I'll Teach You How per
473. 1:14:49week and so it makes it it makes it all seem like a sleazy [ __ ] Ponzi and it's not this shit's these tools are like the most
474. 1:14:58profoundly powerful tools in the history of humanity they're deadly serious like the the they're going to change everything
475. 1:15:10everything from science to medicine to the nature of work to the nature of how we value work to what you do to what's
476. 1:15:20possible to what you do like it's going to do all of that like gar [ __ ] teed and it's not going away and it's not overhyped in fact in
477. 1:15:29one of those Corners is a video series I did called underhyped where I've got like 18 videos in there talking about
478. 1:15:37why this is a big deal and why I think it's a bigger deal than people realize so so the you can make money with Chad
479. 1:15:43GPT crowd as far as I'm concerned can go suck an egg I understand why they're doing it you know trying to make a buck it's just
480. 1:15:53a sleazy way to make a buck I don't even want to talk about gpts that killed mine oh that sucks Joker yeah because see there there the
481. 1:16:06so what we're what we're in the middle of right now Joker I would hang in there because because uh and I know you will
482. 1:16:17um open AI right now is trying to figure out how to be be a consumer application company they've been a research lab for
483. 1:16:28five six seven years I don't know when they were founded but for a while they did not expect Chad GPT to be a hit and
484. 1:16:35it was and so they're now dealing with like being a consumer application company and they're really [ __ ] shitty at it um
485. 1:16:50and so so if you look at their terms of service for making gpts there were three areas that they basically
486. 1:17:02said financial services and Healthcare were two of them right and then you know sex stuff was the other one there might have been
487. 1:17:10one other but but basically they're avoiding the Deep liability holes because they don't they're they don't they don't these tools are not
488. 1:17:21sophisticated enough right now for them to put themselves on the liability hook for financial service advice or a bad
489. 1:17:30diagnosis and and a you know a malpractice lawsuit coming at them so they're just not going to let you make those
490. 1:17:38gpts they'll get their [ __ ] together eventually so I'd hang in there but yeah that's really frustrating yeah that sucks I'm sorry about
491. 1:17:45that ah anyway let me look at some comments here let's get let's get some comments going is he is he ever going to
492. 1:17:52is is he ever going to read my comment he just seems to be talking does this man ever stop talking all right I've
493. 1:17:58already got a twin up good yeah just keep just keep [ __ ] flooding them in there just slightly change the language good for you yes do that
494. 1:18:10um but Kyle I want to I haven't got familiar with gp4 please slow down with GPT 5 I know I know I'm right there with you listen I
495. 1:18:24so I I I I am still I would still consider myself in the middle of a semi existential crisis that that started on November 6th of
496. 1:18:412023 and it's funny November 30th 2022 is when I sort of lost my mind and realized oh my God the world just changed this is AI is you know
497. 1:18:53generative AI is like the next worldwide web um and then almost a year later November 6th 2023 open AI has their developer
498. 1:19:05conference where they launch gpts and they also launch multimodal chat GPT and the reason I'm in a bit of an existential crisis
499. 1:19:20is um chat gp4 is signific it's a significantly different tool than the chat GPT 4 that existed in
500. 1:19:37October like like like it changed just into a completely different thing CU it went true multimodal where
501. 1:19:47it can see images and it can make images and it can understand talking and it can talk talk and it can program and it can
502. 1:19:55and it can do all this together and then I'm starting to look at tools like Leonardo that started as a very simple
503. 1:20:03image generation tool and now it's an image generation tool and it's this real-time composer tool and it's a movie
504. 1:20:09making tool and and so I'm watching every one of the individual little simple tools that we've been playing
505. 1:20:16with are getting more and more and more capabilities and they're also starting to blur like this and then and then I looked
506. 1:20:24at the the Google Gemini launch even though it was you know people made fun of that marketing video because they
507. 1:20:32they they accentuated the positives Andor lied as as some people said it's a marketing video so every [ __ ]
508. 1:20:41marketing video is a lie what do you what do you hi you don't think they're going to make it look as good as [ __ ] possible but
509. 1:20:48whatever but if you look at what Gemini Ultra is and what what's coming the one thing that's really clear to me is it's not what we
510. 1:21:00know generative AI is today it's a different thing it's a more visual thing um yep it's everywhere people buy buy it
511. 1:21:13chat GPT prompts for 25 bucks or 47 bucks and a million dollars exactly yeah exactly and and here's the thing Marina those
512. 1:21:23those prompt sets that you buy will the prompts work yes would the prompts work if you just type some [ __ ] in to chat gbt
513. 1:21:33yes would they be as perfect maybe maybe not maybe yours would be better AI is literally magic changed my mind so
514. 1:21:42probably a potato Pate Pate p and I have this uh this debate on a on a regular basis in here um I am I am of the is it
515. 1:21:52Arthur C Clark or wait who said hang on I think it's Arthur C Clark hold please um who
516. 1:22:15said uh Arthur C Clark yeah so Arthur C Clark in 1962 I think three maybe said any
517. 1:22:30sufficiently advanced technology is indistinguishable from Magic that's where we are we are absolutely there what what hit me within
518. 1:22:41within I don't know an hour of using chat GPT was I've been reading about the Turing test for all of my life team magical AI
519. 1:22:55exactly um you know what paint will tell you and P's not incorrect P what Pate will tell you is this is a mathematical
520. 1:23:00equation this is a probability engine he's not wrong my contention is it is such an advanced probability engine that
521. 1:23:10it um performs as magic wait P Pat's banging in wait wait Bard is an AI assistant Gemini is an AI model that now
522. 1:23:21powers behind oh yeah right oh that's true all right yeah it's this is this is the similar so this is the similar
523. 1:23:30conversation to you know here's gp4 here's chat GPT right so yeah I I'll get my terminology correct yes correct are
524. 1:23:40your lives recorded in case I missed one they are recorded if you go to uh YouTube my YouTube channel is UM AI
525. 1:23:48learning lab dtt for- Tik Tock DT TT someone already had AI learning lab on YouTube bastard and I could have gotten it I was
526. 1:23:59there early enough but I was just lazy and didn't grab it so I'm a dumb dumb so it's AI learninglab dtt are the are the the back episodes of
527. 1:24:11these um I didn't put the I did not put the video up of the night that I gave out my credit card info although I could
528. 1:24:20because I've since cancelled that thank you Pate for for being Eagle eyed and going uh hey dum dum oh that was that was some fun that
529. 1:24:34was some fun action um what are we talking about oh any oh this shit's magic convinced me otherwise yeah this
530. 1:24:45is when I when I first started using chat GPT it was instantly what I thought what I was told for years like 10 years that chatbots
531. 1:25:02would be for years we were told that Siri would be this that Alexa would be this that AI powered chatbots would be this
532. 1:25:13and AI powered chatbots were never anything more than logic trees that broke the minute you ask them anything
533. 1:25:20remotely out of BS of their task what they've been programmed to do if you look at a lot of the layoffs that have
534. 1:25:29been happening in the big tech companies a lot of the layoffs are happening in departments where people were writing
535. 1:25:36logic trees like Alexa logic trees and Siri logic trees because large language models just [ __ ] answer the
536. 1:25:45question so the first the the first thing that popped in my head when I used chat GPT was the Turing test is done
537. 1:25:54like we're like we we're we're done and and and what I what it was funny when I first started using it I said you know
538. 1:26:00what's going to happen is there's going to be a bunch of people come out of the woodwork and they're going to be like
539. 1:26:04well we haven't really hit it yet cuz and I saw I saw this article like a month after I was thinking that where
540. 1:26:09this guy was like well I need to be in a conversation for four to five sustained hours and not like he started qualifying
541. 1:26:17like the Turing test hadn't been passed because it hadn't hit his very stringent definition definition of it I'm like
542. 1:26:23okay if we're into semantic arguments now about nuances of the Turing test then we [ __ ] passed it right um
543. 1:26:30weren't we going to build a GPT yeah we probably were I I'll get i'll get back to that and build a GPT
544. 1:26:37um so so Turing test if it hasn't been passed we're like you know I interacting with chat GPT is is more
545. 1:26:48like interacting with an intern than it is interacting with an expert system right you have to guide it you have to
546. 1:26:55coach it it knows a lot it it's it's like got good book smarts but it's you know simpleton so you got to guide it a lot
547. 1:27:04so what do you want champy hang on Champy's had it what do you want you want to go out to Mama oh you want in oh nobody let champ
548. 1:27:21be he wanted in do you want to sing for the people do you want to do a little singing you want to sing for the people
549. 1:27:31huh do you want to sing or you just want to come say [Music]
550. 1:27:51hi [Music]
551. 1:28:07yeah that's a good do oh there's a good doggy oh there's a good doggy did you miss Daddy is that what's going on you miss Daddy
552. 1:28:14oh oh there a good doggy all right chat GPT magic all these things are magic they they are it's so part of it is
553. 1:28:30just we've now got tools that are as good as or better than the technology that was in our Sci-Fi movies when we were
554. 1:28:43kids and Todd Todd really good point just keep in mind it's the worst worst it will ever be right now exactly so we've got in our grubby
555. 1:28:53little hands and we don't have to be technologists to do this the p in GPT stands for pre-trained meaning we don't need to be
556. 1:29:02data scientists to use machine learning that's what it's been for two decades well we've got this large machine
557. 1:29:10learning supercomputer that you're welcome to use if you bring your own data well what data do I bring but all
558. 1:29:15of it bring all of your expert set data and then we will teach how to take your mathematical waiting and hire someone like Pate who actually
559. 1:29:27understands what all that [ __ ] means bring your data come into our data center sweat with the oldies and then we'll make you a little
560. 1:29:37expert system that's what it was for years it's not that anymore Sam mman and now every company on the
561. 1:29:47planet said let's train these things on every piece of data we can get our hands on the more data we throw in it the more
562. 1:29:53compute we throw at it the better it gets let's go and so these things are functionally magic they are technically quite simple
563. 1:30:04and very flawed and janky pieces of [ __ ] but they're Magic Sing champy Sing champy Champy's a rock star it's also per pervasive with the
564. 1:30:20intended unintended to slant politically of its handlers yeah I think I think one of the things that we'll
565. 1:30:28see Sam Alman kind of kind of inferred this in in uh in his talk at Davos this week where you know they were asking him about chat
566. 1:30:40gp5 and what what he Bas so he implied a couple of things he he implied these things are going to get much better
567. 1:30:50reasoning and so so and if if if you followed any of the debacle after they fired Sam Alman about qar and the rumors about
568. 1:31:02qar and and that it was basic it's basically a reasoning engine that sits on top of chat GPT if even a small part of that is true
569. 1:31:12chat GPT is going to be able to start checking its own work and the minute it can start checking its own work it gets
570. 1:31:17better really fast it can start improving itself so not only will it not make mistakes on your stupid little
571. 1:31:24question it's going to train itself to get better and better and better really really fast um so it's it's going to
572. 1:31:32reason and and uh and you know it's it's it's just going to be it's just going to be a completely different thing so
573. 1:31:40anyway how long until we can understand what dogs are actually saying thinking you know it's funny Todd I just saw yesterday or the day
574. 1:31:50before they're they're starting to do that right now so so animal biologists are starting to map you know you know
575. 1:32:03existing knowledge of sounds of birds and animals and calls and things like that and they're training large language
576. 1:32:12models on it and so like it's like it's coming like that sort of shit's coming because understanding understand
577. 1:32:21ing animal language is no different than what they did with there's there's a paraplegic out of Italy um where they trained they had him
578. 1:32:31they they did an fmri scan and they told him to think about moving his legs he was you know had a spinal cord in injury
579. 1:32:39and you know at some point you know below that couldn't move his legs can't move his legs and so but they
580. 1:32:47had him think about moving his legs and then they trained a large language model on that data and then they
581. 1:32:56sent it's [ __ ] insane they they created a Bluetooth bridge that sent the the M the generative AI training of his
582. 1:33:07brain when he thinks about moving his legs around the injury into electrodes at the base of his spine and he's
583. 1:33:19walking so that that's no different than an animal you know communicating something or saying something like it's
584. 1:33:29it's it's pattern recognition right so it's just different data and that's that's why all these tools if if if one
585. 1:33:37of the things you're wondering is like why are we seeing all these tools like it like okay I understand that chat GPT
586. 1:33:44open AI figured out how to make words do this but why are we concurrently seeing images like image generation tools and
587. 1:33:52music generation tools and deep mind just invented 2.2 million new materials 380,000 of which are stable
588. 1:34:02crystals in Material Science it's like an 800-year increase in scientific breakthroughs overnight why are we seeing all these at the same
589. 1:34:15time because it's the same basic technology they're just throwing different data at it so if you throw
590. 1:34:22music data at it it makes music if you throw image data at it it makes images if you throw brain scan data at it it
591. 1:34:30does brain scans it understands them it understands language it understands you know they're taking fmri data right now
592. 1:34:37and having people watch videos and then they're recreating the videos based on the brain waves I wrote a screenplay about
593. 1:34:48this the technology the technology for my science fiction screenplay I wrote in 1992 the foundations for it are being
594. 1:35:00experimented with in Labs today that could actually do what's in my screenplay it's [ __ ] insane it's just it's just [ __ ]
595. 1:35:16Bonkers Tik Tok give me a pat on the back for going 90 minutes sorry chappie due to your father insistence that AI is Magic I cannot
596. 1:35:24support your chicken habit oh Champion you you don't get no more chicken he's going to get skinny Pate you can't let
597. 1:35:32my dog get skinny he's got to be that wide so I can put wine glasses on his back and he can walk around the house
598. 1:35:39carrying wine glasses somehow my parrot used Alexa to call me and leave a message oh that's [ __ ] awesome actually oh you you know what
599. 1:35:51would be cool you've got a Parrot you know what would be cool so I'm in the chat GPT app
600. 1:36:17right I've got a Parrot and I want you to help me train my parrot say things can you give me 10 really interesting
601. 1:36:26ideas of things we could teach the parrot to say absolutely training a parrot can be really fun here are some interesting
602. 1:36:37ideas for phrases one good morning sunshine it's a cheerful way to no I mean whole categories like if I wanted to train a parrot to quote
603. 1:36:49Shakespeare or explain quantum mechanics and I wanted you to train him all day long every day could you do thinking about categories
604. 1:37:00of phrases here are some creative categories one Shakespeare quotes teach your parot so you could just turn Alexa
605. 1:37:09conversation on with your parrot and just have it train it all [ __ ] day and you just come now your parot going
606. 1:37:16to talk like whatever voice you choose oh my God that'd be hilarious you could have the smartest parrot and and you
607. 1:37:27could also hook it up to an uino and and have it you know do do positive reinforcement for the parrot where he gets a little seed every time
608. 1:37:36he says something hook it up to your Arduino how but you're like but I don't know how to program an Arduino sure you do watch chat gp4
609. 1:37:56I'd like to program an Arduino connected to a voice recognition module that every time the voice says good job you learned
610. 1:38:13another phrase it activates a reward system that feeds a snack to a parrot can you please help me write that
611. 1:38:39program you could do this with children you could do this with dogs command good job the command good
612. 1:38:48job constant should correspond with with the ID you assigned a good job command yeah that's right we're hacking pet
613. 1:39:03ownership now kids oh for every hook that now has ADHD to just interrupts when a thought comes hey I resemble that
614. 1:39:18remark wo oh man that's really
615. 1:39:33[Music] funny all right oh we want to talk about gpts let's do that make the parrot talk to a GPT do we
616. 1:39:45have to add parrot to our Zoo not yet if we'll see if this one sticks cuno I think I think I think parrots at this
617. 1:39:53point are just flying around outside the zoo I don't think they've they've decided to uh to roost but we'll we'll see if this one
618. 1:40:03sticks I'm telling you man I the auto parrot trainer because because what you should really do with the auto parrot trainer
619. 1:40:14is not do it on chat GPT but you should do it on um on the unrestricted llama 2 model that that has no safety guard rails on it and just
620. 1:40:24teach it to cuss in every possible language the the most vile things you could imagine saying and and and train it to say those
621. 1:40:39things when people say things like ply want a cracker go [ __ ] yourself this is this is is why you should not have theatricals anywhere
622. 1:40:53near animal training uh opportunities theatricals are just not to be trusted with this kind of [Laughter]
623. 1:41:07power oh I love how much fun you're having with this that's good sometimes sometimes this channel is really just me entertaining myself I
624. 1:41:18think most of the times this channel is just me entertaining myself oh my God that's funny Joker there are two PODS of whales
625. 1:41:33on different coasts in Norway he thinks they speak different dialects and a third went together yeah there you
626. 1:41:43go I mean seriously we're we're in for so Google Deep Mind now Google Deep Mind
627. 1:41:57has they created 400,000 protein folding predictions that are proving to be correct they created 2.2 million new
628. 1:42:07materials 380,000 which of St are stable which is like a MAG order of magnitude increase in materials we
629. 1:42:15didn't know existed so and if you're like well what's a crystal mean and what what does materials mean think
630. 1:42:22batteries there is some sort of room temperature superconductor there was just a sort of false hit on that out of
631. 1:42:30Korea about three months ago that if they hit it it changes everything it changes power transmission it changes
632. 1:42:38Battery Technology like it's the kind of thing like if they get the right crystallin structures going that have
633. 1:42:44the right properties we'll have batteries that you can charge in 30 seconds that last for four weeks under load like that kind of
634. 1:42:57[ __ ] and then deep mind just announced yesterday or was it today might have been today that it's now
635. 1:43:06doing sort of like Advanced High School level geometry it's solving like you know geometry geometry competition questions
636. 1:43:21at like a like some crazy [ __ ] insane hit rate we're it there's just it's just going to be one thing after another where because so many
637. 1:43:38people are focused on this stuff now they're just pointing these machines at all the different problems that we've
638. 1:43:43got and they're going to be like oh here's a pile of solutions go deal with that oh here are a bunch of ways we
639. 1:43:50could cure this genetic disease they I think the FDA today just approved the first Sickle Cell genetic treatment that was AI
640. 1:44:02created AI generated AI assisted something like that is that blandon's caramel K it is exactly what that is Amelio's wife
641. 1:44:15um and I think I from from what I read from the article that that FDA thing was is is the first of its kind right it's the first
642. 1:44:25cure that was generated in this way and we're going to start to see just bang bang dominoes boom boom boom
643. 1:44:34boom I'm messing around with a new AI text to video an image to video website called pix verse PX verse let's go look at
644. 1:44:49it pix verse create breathtaking videos [Music]
645. 1:45:13with there's there's pretty good coherence there this one looks like it's a little janky let's go waves are always weird their waves
646. 1:45:27look okay huh I wonder what they're using I wonder if this is oh is this from
647. 1:45:41X oh that's that's just their X Channel 4K upscale
648. 1:46:08[Music] [Music] feature
649. 1:46:24huh interesting interesting do they talk about does anywhere talk about what it is what's this Discord Community Showcase
650. 1:46:42home Discord Community Showcase home I'm guessing this is sitting on top of stab stable video diffusion but I
651. 1:46:55don't know it looks a little more coherent than that but I I can't tell just because of um just because of
652. 1:47:07uh they're just showing you know they're showing the best examples so you can you can cherry-pick these things to give you
653. 1:47:14really good examples but it's hard it it's hard to get predictable results with any of these things loging successful
654. 1:47:23create can I upload an image image to [Music]
655. 1:47:37video let's take one of these wacky we'll take that dude yeah let's take that
656. 1:47:55dude uploading image to video seed strength of motion prompt see what it does with no prompt the system is operating at full
657. 1:48:11capacity all these free tools are like uh we didn't think anyone would actually use it this much we don't know what we're doing here
658. 1:48:21well we'll come back to that that it looks interesting it was worth worth checking out um I would say I would say
659. 1:48:27if if you want one that it looks like it's a bit more baked um just as an overall tool um Leonardo their their image to video
660. 1:48:38stuff looks pretty good and then um Deco here looks pretty good there's I don't like I don't know man all there's
661. 1:48:49just so much as moving so much is moving so fast it's so hard to know this is kind of the existential crisis it's like I don't know where to
662. 1:48:58put my focus I don't know what to learn right now you're like but Kyle you're supposed to be telling us what to learn
663. 1:49:05I don't know [ __ ] start with chat gbt and just keep going like I don't know I'm I wish I had more better answers for
664. 1:49:17you I wish I could quit you how where can I input and manipulate my own drawings for charts Twisted Crone lead of Vico
665. 1:49:31um I don't quite know what you mean manipulate drawings for charts so so start uploading it into GPT vision and ask it to describe it first
666. 1:49:44yeah that's that's a good place to start but if she's saying she wants it for charts you might not need AI like if you're looking for if you
667. 1:49:53want to do charts that have Turtles stacked on top of one another to represent bar charts there's charting
668. 1:50:01software that'll do that it's not always it's not always um you don't always need ai ai is just a tool charts for artistic
669. 1:50:11per purposes oh yeah then definitely here we let's go do that we'll go play with that right now so here's here's my
670. 1:50:18the the my go-to right now is just keep learning what what chat GPT makes possible so let's see let's go let's go to the internets and we'll
671. 1:50:31say visually compelling chart who's the dude tufty like here's tuft's famous chart
672. 1:50:48right that's the what was this from the Battle of something French Russian campaign I don't know some War but that's the that's the Edward
673. 1:51:02tufty that's from his book Let's see we got anything good in here infographics let's go grab some infographics see if we can grab
674. 1:51:12something interesting um that one's cool all right oops so I'm just going to screen shot it and we'll grab all the text cuz
675. 1:51:26you'll you'll see this thing gets this gets quite it's quite something what it can do all right so we're over here at chat
676. 1:51:38J I'm just going to paste that in you can also upload it oh and I have to upload it my screenshot dumped it to the desktop there it is okay
677. 1:51:51so there it is I'm going to say describe this in as much detail as possible this also starts to get into
678. 1:52:06the accessibility thing I don't know if if if the guy that was in here earlier that said he was blind and this stuff's given him
679. 1:52:12accessibility um the ability to describe stuff like this in details just I it's just amazing um all right countries prize categories
680. 1:52:23chords color coding okay so so now I'm going to say please give me five new ideas for how
681. 1:52:39to make this chart even more visually appealing one of the cool things about having an
682. 1:52:54image generator inside a large language model is that you can have the whole conversation about your ideas about what
683. 1:53:02you want to do before you actually make the image and then when it makes the image it takes all of that [ __ ] into
684. 1:53:08account which is mid uh mid Journey can't do that stable diffusion can't do that but but Dolly can so what's it it
685. 1:53:18say Interactive elements know 3D or animated visualization harmonized color palette so I'm going to say recreate this chart with a
686. 1:53:38harmonized color palette and 3D elements
687. 1:53:55and now it should oh no when I used 3D it thought I wanted it
688. 1:54:09to actually create a 3D thing so I'm going to say recreate it with a harmonized color
689. 1:54:35[Music] palette are we going to do it I'll generate the image circular layout okay here we go so now it's generating the
690. 1:54:46image so this is a way you could do it you can upload anything and you could also upload you know two unlike images a chart and a
691. 1:54:57turtle and tell it to do something like there you go there's a so you know there's where we started and then there's kind of how it
692. 1:55:06interpreted that and you could say well I don't like that or I do like that or add this or add that and just keep just keep
693. 1:55:15iterating Jeffrey Lombardo yeah that's right um all right it's getting to be it's getting to be late here so let's let's let's
694. 1:55:27start let's start winding this thing down I'll start my 30 minute goodbye sooner than [Music]
695. 1:55:43later is this the nature paper with the m materials prediction yeah Jasmine it's it's um yeah it was the nature paper um
696. 1:55:54that had um from Google Deep Mind um that that basically said here are 2.2 million new materials go play with them
697. 1:56:04and here are 380,000 that are stable and as of a month month and a half ago I think 800 of those stable
698. 1:56:13Crystal had been verified in the lab so so because they're all there they know all the crystal structures they know
699. 1:56:20they're stable now the labs can just go test which ones actually can they reproduce and do they exhibit the the
700. 1:56:29properties that exist so like the the the speed of acceleration in Material Science in the next year is going to be [ __ ] Bonkers 30 minute
701. 1:56:39[Laughter] goodbyes all right um what's the difference between GPT and gpts okay let's go do that interesting question so chat
702. 1:57:06GPT I'm trying to think if I if I want to go to my my box metaphor so so so gpts are
703. 1:57:21one of the things that open AI has successfully done and I think them calling their Bots gpts is smart because
704. 1:57:32everyone's going to start calling anything that's a large language Model A GPT right um so from a branding
705. 1:57:43standpoint GPT is going to kind of be like Xerox or Kleenex where it actually was from a brand um but it's going to
706. 1:57:52become synonymous with the generic thing so so right out of the gate there's confusion already between gp4 which is the
707. 1:58:05underlying foundational model right the underlying large language model and chat GPT which is an application sitting on top of
708. 1:58:17gp4 which is different than gpts that are little specialized versions of chat GPT so wait I do need to do my boxes God dang it okay and I
709. 1:58:30and I actually have okay oh God this is this is just painful okay [Laughter] box this is
710. 1:58:44gp4 see it's empty right now all you do is fill it with all of the knowledge of the public internet for
711. 1:58:53the past four decades and close the box and that's gp4 okay the foundational model it's got all this good goodness in it but you the
712. 1:59:06only way you can get it in in into it is through apis application programming interfaces right so you make requests
713. 1:59:13into this thing it does its magic and it sends [ __ ] back out um before chat GPT came out access to the
714. 1:59:24apis for this thing were available for to Developers for years there was gpt1 gpt2 gpt3 there were all different kinds
715. 1:59:34of fine tuned models within them and developers could build [ __ ] on top of these things right so Jasper is one of
716. 1:59:42those companies copy AI is one of those companies notion with all their fancy AI [ __ ] now is one of those companies so
717. 1:59:50there's all sorts of companies that were building stuff on top of this chat GPT it's the blue one that's chat GPT is an app that sits
718. 2:00:04on top of this this app makes it really easy to use all of the functionality within here rather than giving us all of
719. 2:00:12the options they just say here's your little chat box and here's a submit button and now there's like here's a
720. 2:00:18paperclip button to upload [ __ ] and that's what this thing does a GPT is like a little mini version of one of these that just kind of sits
721. 2:00:29on top of it that we get to build so there are now 3 million of these little mini things that are basically just
722. 2:00:38little specialized versions of this main chat GPT all of these are just calling the information sitting inside the Box all right so
723. 2:00:51that's gpts and then like uh um how do crickets make
724. 2:01:07sound and it'll just answer that [ __ ] right so you can ask it anything or you can say uh make me a picture of [Music] um
725. 2:01:25of a caramel corn kernel on a tongue and it'll make you
726. 2:01:41that and so it can kind of do anything which is one of the intimidating things about it is you don't necessarily know
727. 2:01:50what to do we do need to wait for the this is a ridiculously helpful visual that's good Amanda h i i i i mean this is a professional video
728. 2:02:02Studio don't get me wrong um we Cho we choose to use Simple props like garbage uh because it it is easier for people to
729. 2:02:13understand not because this is an amateur hack operation oh wait wait why could I not try
730. 2:02:34again um so anyway so so chat GPT is the general tool and then we'll go we'll go look at a specific GPT no it's for whatever reason it's not
731. 2:02:45it apparently a caramel corn kernel on a tongue is too pornographic or something whatever um okay so I'm going to go look
732. 2:02:55at uh 70s let's see 70 7s muscle car machine so this is a GPT this is this is the second GPT I ever made gpts are really easy to
733. 2:03:09make um I'll show you how to make them in a second and this one's sole purpose in life is to make you a post
734. 2:03:19of a 70s muscle car you can hang in your bedroom you're like why would anyone want to do that and then you might also
735. 2:03:29ask well wait couldn't you just do that in chat GPT you could but who would do that no one so this is something you
736. 2:03:35stumble on my metaphor for gpts right now it it hit me today what gpts are like because they're so early and
737. 2:03:43they're so janky and they're so easy to make gpts right now are kind of like like when you're leaving Walmart and
738. 2:03:50they have the vending machine where you can get the little plastic cup that has a plastic toy in it you get it for your
739. 2:03:57kid to shut him up here take the little thing don't choke on it you little bastard that little toy that's what gpts
740. 2:04:05are like like chat GPT is the whole vending machine and and our the little gpts we're making there's like one or
741. 2:04:12two cool ones in the whole machine and the rest of them are just janky pieces of [ __ ] so that's what that's what this
742. 2:04:17is this is a janky piece of [ __ ] but it's a fun one and so I've got one little button here and it says I'm ready
743. 2:04:23to design my dream car so I click that button and it says fantastic let's start designing your dream car the first step
744. 2:04:31is to to grab your preferred model so let me see I'll do a I'll do a uh I'll do a 71 Corvair which is not a muscle car but it's kind of
745. 2:04:43funny the Chevrolet cor Corvair is an interesting choice especially for a muscle car since it's known for its
746. 2:04:51unique rear rear engine layout and the fact that it can't take Corners at speed um it doesn't exactly fit in the typical
747. 2:05:01muscle car category would you like to go with one of these I'm going to say um I want a pumped up Resto mod
748. 2:05:14Corvair as if it were a muscle car from the 70s it was also only produced until
749. 2:05:351969 I'd like it to be mint green and glossy with black stripes
750. 2:05:50it should have drag slicks and chrome mags and Fire coming out of the exhaust pipes and so now it's going to make us a little
751. 2:06:04poster I think yeah there we go so that's a GPT a GPT is just a little specialized version of a thing
752. 2:06:16the cool things you can do with them you can train them you can train them on your own data so if you had a PDF of drink oh
753. 2:06:29that's actually really [ __ ] cool come on now sure it's a it's a 16 cylinder it's it's apparently a 16 cylinder engine in it but you
754. 2:06:45know and apparently it's going in reverse cuz the Flames are shooting forward the Corvair the only drag dragster that that drags
755. 2:07:00backwards but I mean doesn't that remind you of like you know those 70s stickers of the crazy Dragsters and it look it almost spelled
756. 2:07:08Corvair [Music] right oh man but but if you had a uh a PD F of you know a book you wrote or of specific recipes for certain kinds of
757. 2:07:24food your grandma made you could make a GPT that just searches on the knowledge that you put in it right so so super
758. 2:07:33super powerful stuff okay I just turned 59 today and I'm still excited about the GPT explosion dude I'm 58 and here's the
759. 2:07:42deal like we are absolutely in a revenge of the Gen xers moment where one of the things I'm consistently seeing I'm so discouraged from word
760. 2:07:55wrestling with chat gp4 oh with it with its safety guard rails I know it's a pain in the ass um one of the things I'm
761. 2:08:03consistently seeing is that is that you know all Elder Millennials and gen xers are [ __ ] lit on fire about this
762. 2:08:13generative AI stuff and um the younger Generations don't seem to care as much about it I the only thing I can think of
763. 2:08:21is we were all there we like we all got to experience what happened with the worldwide web and and you know we watch
764. 2:08:29the internet come onto the scene and transform everything and so I think there's I think there's a palpable visceral holy [ __ ] this is one
765. 2:08:38of those things um but uh but but yeah you you should absolutely get excited about this stuff because it's it's crazy simple to do so
766. 2:08:47here's how you make them you're like you're not going to believe how easy it is to make these things um
767. 2:08:53so explore gpts right and then this is the GPT store so you can go find things in here um and then you can go look at your gpts
768. 2:09:05but in the upper right hand corner it says create so you click on Create and then this interface there
769. 2:09:12there's sort of two modes on this left hand side there's create and configure I'll get to configure in a second and
770. 2:09:18then over here is the preview pane or they also call this the playground which is confusing as [ __ ] cuz they also call
771. 2:09:24the developer area the playground so this is a different playground it says preview they call it pre playground what
772. 2:09:31it doesn't [ __ ] matter anyway go back to evidence a that that um open AI does not know how to be a
773. 2:09:38consumer software company yet so the way you make gpts when you're in create mode you just talk them into existence and
774. 2:09:48because I'm using uh Voice to Text for it to do the typing I'm literally going to talk this into existence so
775. 2:10:06um okay I want to create a GPT wait where's my cursor hang on I want to create a GPT that asks
776. 2:10:25someone for how they're feeling emotionally and then responds
777. 2:10:39with describing a car from history that best represents that emotion and why and then renders a photograph of that
778. 2:10:55car that looks like it should be in a magazine that's it it's now building it so so like literally anything that
779. 2:11:12you like it's done um tell me about your current mood oh wait it's so we we've got to fix
780. 2:11:26these things but um so I'll say um hang on I'm feeling a little blue and would love a little pickme
781. 2:11:41up feeling blue how about a 1963 Volkswagen Beetle this car is Iconic for its unique contended positive free spirit
782. 2:11:50Adventure let's create a photorealistic image of a 1963 Volkswagen Beetle oh wait did it oh regenerate I think I broke
783. 2:12:06something is it still going I can't tell if it's still rendering
784. 2:12:24um I'm feeling blue something's
785. 2:12:40broke oh I might have shitty internet right now so now what's it going to do the Bugatti type 57cc from the late
786. 2:12:541930s is going to make us feel better do you think as soon as it hits FIFA games kids will start releasing
787. 2:13:10AI I don't I don't quite understand the question you you mean when AI gets integrated into like EA Sports kind of
788. 2:13:21games um I think what's going to what's more likely to happen that's pretty slick looking car I think what's likely going to
789. 2:13:30happen is that um everyone's going to be able to make custom games that are like AAA quality games in the next 3 to five
790. 2:13:38years by just doing what I'm doing like speaking this into existence so anyway so so now that this thing's made now I
791. 2:13:45can so let's see it says great choice let's give this GPT a name how about emotion Wheels how about let's see
792. 2:14:05um let's call it your Automotive therapist so we're going to give it a name and now it's going to generate a
793. 2:14:21profile picture so like it's doing all the work right it gave it a name it's going to give it the description it
794. 2:14:27wrote those pre-done questions there's a little C thing I'm going to say that image is really boring I want
795. 2:14:39it to be super sexy and make Fe people feel good so it's going to make us another image
796. 2:15:10perfect okay now it's going to ask us to start refining this thing and we're going to stop here and I'm going to tell
797. 2:15:15you why in a second so I'm going to go over here I'm going to reload this so now we have the cool icon your
798. 2:15:21Automotive therapist I empathize with emotions and describe a historical car that matches with a photo that's a
799. 2:15:28horrible description we're going to go fix that but first we're going to save this thing and so if I save it to
800. 2:15:35everyone that's going to make it available in the GPT store if I save it to anyone with a link then I can you
801. 2:15:40know just send it to people only me means only this account can play with it so I'll save this as open we're going to
802. 2:15:46give it a category like uh lifestyle good enough we confirm that so there we go so that works now I'm going to go fix some things in it
803. 2:16:01I'm going to go edit GPT I'm going to go to configure now and we're going to change the description tell me your emotion and
804. 2:16:16aisle give you a picture of a car that will make you feel better and then I'm going to go um hash
805. 2:16:32therapy hashtag um uh Gearhead hashtag um uh
806. 2:16:50uh whatever Gearheads fine okay so now in my instructions I'm going to say wait well we're going to do one other thing so we've got these
807. 2:17:00pre-written questions and it it it wrote them in such a way that it's like when you click on one of those buttons you're
808. 2:17:06actually asking the GPT that question so when you ask how are you feeling today it will actually answer that so
809. 2:17:15we're going to have it say we're GNA click here and say um let's get started and then I'm going to get rid of all these other
810. 2:17:29things so the only thing someone's going to be able to click on over here is let's get started okay and we'll do
811. 2:17:35exclamation point or how about let's get rolling because it's a car get it so let's get rolling and then I'm going to
812. 2:17:46say up up here this is our instruction so this is our prompt so I'm going to say follow these rules if
813. 2:18:03someone says quote let's get rolling ask them how
814. 2:18:18they are feeling all right I'm going to put a bullet point in front of this because we're going to give it a couple of
815. 2:18:30rules oops did I mess that up yes okay uh ask them how they're feeling and then we're going to also go bullet point um after they tell
816. 2:18:45you make the image first and then describe the car you gave them to
817. 2:19:03reflect their emotion and why and then I'll give it another bullet point I'll say um provide a a
818. 2:19:19Freudian analysis of how their emotion is tied to that car model
819. 2:19:36color and year and your reasoning in three simple bullet points
820. 2:19:51points okay all right so we gave it some stupid rules but but so so now what I'm doing again what I'm doing here is I'm
821. 2:20:00programming but I'm programming in the English language and you can do this in whatever language you speak right our native
822. 2:20:09language is now going to become the primary interface for all computer systems all of them and it and it won't just be language it
823. 2:20:21will be take a picture of something it will be you know talk into a computer it will be look you know point the camera
824. 2:20:30somewhere it will be um a real-time conversation where it will dynamically um identify different languages being spoken and automatically
825. 2:20:42translate between all of them all right so so now I'm going to update our
826. 2:21:01GPT Brandon usage guidelines oh Freudian analysis so let's see provide a uh humorous humorous mock
827. 2:21:19anal uh analysis of how their emotion is tied to that car model okay this should work so what I just ran into there was uh in in
828. 2:21:33their usage guidelines you can't do anything that is um medical or financial because those are regulated environments right so Freudian
829. 2:21:44analysis okay so let's get rolling so now it should ask us how we're feeling glad to hear that how are you feeling today let's explore those
830. 2:21:53emotions and find a historical car that resonates with them I'm feeling poopy but hopeful uh uh uh uh uh creating image
831. 2:22:09error creating image why oh poopy God dang it oh maybe not let's see there's an issue generating at the moment however
832. 2:22:20let me still describe the car it CHS a 67 Beetle I think Beetle Volkswagen Beatles are going to be solid solid
833. 2:22:29recommendations here um try again with a different model creating image error creating image okay so something's broken so
834. 2:22:42let's go back to our little GPT what did we say I said said if someone says let's get roll okay after they tell you
835. 2:22:58um choose a car and make a 16 by9
836. 2:23:13wide photo of that car as if it were photographed for a glossy
837. 2:23:34magazine after you make the photo comma oops photo
838. 2:23:51describe the car you get blah blah blah BL okay let's see if that works I don't know why sometimes they just Peter
839. 2:23:58out this is this goes back to the the um one of the technical definitions of a lot of these applications that we talk
840. 2:24:07about a lot in this channel these things are janky pieces of [ __ ] right now let's get rolling and just accept that as part of
841. 2:24:16the experience because they're going to get so much better so fast that you're going to be really bored a year from now
842. 2:24:22when these things are perfect you're going to be like remember when this things used to suck and we actually had
843. 2:24:26to use our brains to make them do cool [ __ ] um I'm a little blue but hopeful your Automotive therapist feeling a bit blue requires a d d d
844. 2:24:42creating an image and like it told me it's going to you know styled like a glossy magazine photograph so I'd go in and tell it not
845. 2:24:52to do that so so testing these things is an important part of this right make it test it make it test it make it test it
846. 2:24:58just just keep playing keep playing so oh nice 65 you're kidding me 65 baby blue Mustang convertible I actually had
847. 2:25:07a baby blue 65 Mustang Fastback that was very close to that color humorous mock analysis the emotional horsepower just like your
848. 2:25:17emotions the Mustang is powerful convertible top represents your openness to brighter days ahead retro Vibe the
849. 2:25:24Vintage aspect of the car hints it Nostalgia a subtle reminder of every emotion like every car all right so that's
850. 2:25:32gpts good Lord people I said I was saying goodbye like half an hour ago and then you made me make gpts what are you doing what are you
851. 2:25:40doing to me look at my hair that has not gotten better all right people we going to do this I'm going to
852. 2:25:53get out of here okay so here here's the thing if you're new here go there if you're new to AI Show and Tell
853. 2:26:02please okay if you're new to ai go to the salon. click on the first link it says the AI Salon Community this is a
854. 2:26:11community that that where the people that hang out in this channel also hang out here when when I'm not live and it's
855. 2:26:20a really remarkable Community where we help each other you can go there you can talk about AI no one thinks you're a
856. 2:26:25weirdo you can be a beginner you can be a 30-year vet it doesn't matter anyone's welcome how many PDFs can I load load at
857. 2:26:32once I think five uh no I think 20 20 for for the knowledge base I don't know the size limit I think it's 100 megabytes per or 100 megabytes
858. 2:26:45total um and I'm going to go check out the show and tell Channel R don't choke oh
859. 2:27:06wait see I can't say anything in here where it turns into evidence that's lovely I supplied a photograph of a GPT to produce a
860. 2:27:18professional photo apparently I'm not recognized as a 55-year-old that's good Becky looking looking sweet it's it's also not very
861. 2:27:28photo it's not a photograph right it's an illustration but it's beautiful lovely that's great a i Salon Zoo oh that's from uh
862. 2:27:39parrots want to be included in the zoo that's from uh Firefly Adobe Firefly very cool actually yeah we we should be playing
863. 2:27:49with Adobe Firefly to get some of the groovy uh groovy titles in here you know what would be fun is to is to pop some
864. 2:27:56of some of these things from Adobe Firefly on top of images made in Del Kyle stuck in one of his many rabbit
865. 2:28:06holes that's that is just too [ __ ] close to reality time to go RV in nice oh that's beautiful all a conspiracy I tell you
866. 2:28:18who's who gets in next really good why does his chest look like a butt hole sort of or what I don't know what
867. 2:28:31that is but it's not I think I can get permanently banned from this channel for that bird that's beautiful actually
868. 2:28:39ooh Cory Sandler I think you're right I think I think parrots might need to get in the show but you're right the other
869. 2:28:46animals are going to going to be very jealous that's that's like that's like a serious [ __ ] showoff if I were a llama I'd be pissed
870. 2:28:54at that for one thing it's it's about six times the size of a llama oh those are beautiful actually I I'm thinking AI Salon taken over by
871. 2:29:05parrots yeah I think it's good I think they're in yeah it's funny the baby parrot's not as cute I think the oh that's beautiful
872. 2:29:15Dag all right all right what was it like before 2024 we had to type type search the web poly want a
873. 2:29:31GPT oh my God sweating with the oldies oh man all right you guys are awesome awesome awesome awesome okay
874. 2:29:41everybody sleepy Mr it I know me too okay I'm going to get out of here I'm going to go Hy hdate and get some sleep
875. 2:29:47good night Kyle thank you for all you give us absolutely um you're a great friend to us all thank you I appreciate
876. 2:29:53that Dr J Becky I'm so done typing Beretta oh Beretta but it was a white bird yeah that's right n never mind
877. 2:30:03let's see good night Kyle thank you for all you give us yep absolutely uh all right who's in here forever
878. 2:30:11hooked Robert Rossy thank you as always Rudy good to see you Rudy saigo hope all's well down there in South Africa I
879. 2:30:17guess you're just you're getting ready for D or something right or lunch I don't know what time it is there but
880. 2:30:24it's it's not word it is here U no never mind it's still rendering oh yeah it's something rendering what was it it doesn't
881. 2:30:33matter oh what was the show it was uh Columbo I survived my first mod night Cam kin you did great awesome thank you
882. 2:30:41very much for modding thank you to the mods thank you to the Irregulars thank you to my subscribers thank you to
883. 2:30:45anyone that hangs out on a Friday [ __ ] night we had like 165 170 people in here there's 147 in here right now
884. 2:30:54that blows my mind um thanks for hanging out I hope this is worth your time um Winston good to see you uh Brendan thank
885. 2:31:03you for what you do I appreciate it Joey pie steo you to man peace out dude uto good to see you Lord digital Gods
886. 2:31:14sporting the turtles tonight very ni Zack and allaha have a great night thank you thank you Tobias to you
887. 2:31:20too Tracy thank you thank you for uh coming and hanging out and doing all that stuff Pate as always there's
888. 2:31:28analogy here between parrots and AI models well don't you know this pate parrots are magic parrots are harbingers of magic
889. 2:31:40Pate that's why they have to be in the ai ai learning lab Zoo all right everyone I'm out of here peace out have a good night I'll see you
890. 2:31:52tomorrow