AI Learning Lab

Feb 15, 2024 - (2 of 2) Sora: OpenAI's Mind-Blowing Text-to-Video Model

iIGW1AptXqc
Video2024-03-102:46:168 views

Description

In the vibrant world of AI and creativity, Kyle Shannon explores the intersection of technology and artistic expression on his TikTok channel, AI Learning Lab. His discussions delve into the transformative potential of AI tools like Sora, which can generate videos from text prompts, revolutionizing how we create and consume media. Kyle emphasizes the importance of community engagement, encouraging viewers to embrace curiosity and collaboration as they navigate the rapidly evolving landscape of artificial intelligence. He highlights the profound implications of these technologies for storytelling, accessibility, and personal reinvention, inviting everyone to participate in this exciting journey. For more insights and to join the conversation, visit his TikTok channel: [AI Learning Lab](https://tiktok.com/@aiLearningLab). #ArtificialIntelligence #Creativity #AI #Sora #Community #Innovation #Storytelling #TechForGood Chapters 00:00:00 AI Learning Lab Intro 00:08:10 OpenAI Sora Announcement 00:09:14 Retweeting Kyle's Screenplay Prompt 00:15:50 Importance of Community 00:21:30 Adobe Animate From Audio & Frank The Hotdog 00:24:53 The AI Salon Community 00:27:02 AI Salon Values Explained 00:34:20 AI Tools and Resources 00:36:51 The Irregulars Club 00:40:08 Creativity Workshop Promo 00:44:35 OpenAI Sora Deep Dive 00:47:28 Sora Capabilities Explained 01:00:37 Sora's Understanding of Physics 01:10:46 Sora's Weaknesses & Limitations 01:15:32 Sora Safety Features 01:18:20 Sora's Technical Details 01:21:09 Sora & the Path to AGI 01:27:26 The Coming AI Revolution 01:31:48 AI's Impact on Work & Creativity 01:37:12 AI Rekindles Creativity 01:40:22 AI vs. The World Wide Web 01:46:52 The Importance of Experimentation 02:01:36 AI Salon Membership Drive 02:03:22 Confronting AI Cynicism 02:06:19 Kyle's Python Coding Breakthrough 02:13:14 Sam Altman & Sora Requests 02:14:50 Examples of Sora Videos 02:19:00 AI Optimism vs. Cynicism 02:22:52 AI Generated Video Games 02:24:45 VR and AI Integration 02:27:18 AI Agents & The Future 02:30:19 AI Accessibility & Empowerment 02:31:41 Real-time Language Translation 02:40:44 AI Salon Promotion & Closing 02:46:12 Outro

Chapters

Transcript

0:02 how are we going to do this what do you
0:03 want to
0:04 [Music]
0:08 do and we to
0:12 Direction this car is my
0:17 train I've been driving I've been
0:21 wondering what it is I'm running from
0:24 [Music]
0:26 again feel like an 8-year-old man
0:31 hold it on to
0:35 29 and a on that
0:40 Horizon California
0:43 [Music]
0:49 L up aead of trucks carrying a wide
0:53 load three house couldn't
0:57 have cute little front door and knows my
1:01 love ain't sure whether the crash should
1:06 left you see a broke a home up myself
1:10 once there I stumbl to that door thank
1:14 you Robert Rossy read a note by the DA's
1:19 light said don't you come around
1:24 here
1:27 anymore well I thought enough
1:32 was freedom on the
1:35 RO
1:38 decision that's what I've been
1:45 [Music]
1:51 told that was good he did
1:53 [Music]
1:59 good
2:03 [Music]
2:13 you better be hanging out here tonight
2:15 we got some Sora to talk about I'm Sora
2:19 with
2:20 [Music]
2:24 you even Sam alman's all excited about
2:29 it
2:31 [Music]
2:54 I just quit Peter zoomed to see you Kyle
2:56 oh poor Peter Peter Kaminsky yeah
2:59 there's a lot going on
3:00 um Leah and Jane were supposed to be
3:03 tonight at this
3:04 time but they got something messed up
3:07 with the time so they're not going to be
3:08 tomorrow at 8:00 p.m. Eastern so I'll
3:10 talk about that event that I you should
3:14 if you want to get in on this if you
3:16 want to get in on this here creativity
3:19 and artificial intelligence stuff you
3:20 should be a part of the Jane and Leah's
3:23 uh
3:24 creativity it's like a six- week program
3:27 they're putting together with daily
3:29 video prompts it's [ __ ] nuts it's so
3:33 [Music]
3:50 cool you and I
3:54 Al Sunday morning here at
3:59 home Sky's blue and the car is strong
4:03 it's
4:06 true then I open my eyes to dream
4:11 realized in front of
4:15 me and I haven't got a clue what in the
4:18 world is happened to
4:22 me think I think I'm
4:26 happy like first Sumer vacation happy
4:30 got to get a little rest and
4:32 relaxation
4:34 happy like Wi on Sunday morning singing
4:42 [Music]
4:44 true all right Champion you
4:46 [Music]
4:58 done what do you think you
5:01 done a
5:05 minor yeah a minor
5:07 [Music]
5:20 right all right all right all right all
5:22 right all right all right all right all
5:23 right all right fine fine fine fine I
5:26 understand you're not getting paid
5:27 overtime he normally does half a song
5:30 he's done
5:31 like two nearly half a
5:37 songs he's exhausted he's not the
5:40 Rolling Stones people he can't do
5:42 three-hour
5:44 tours it's not
5:48 Gilligan that was an old people joke see
5:51 there's this TV sh doesn't matter all
5:53 right
5:54 listen welcome welcome to the AI
5:57 learning lab if you haven't been here
5:59 before I'm flipping crazy but you can
6:03 make money
6:04 with um I'm crazy I have a crazy
6:08 dog um crazy people are in here we call
6:12 them the Irregulars they're not regular
6:14 they're not
6:15 weird they're
6:17 irregular on many levels and they're
6:20 swell so if you see people in here like
6:23 answering questions or like you know
6:26 steering you around or telling you
6:27 what's going on that's cuz they come
6:29 here night after night so I I do these
6:33 lives let me count 7 days a
6:37 week for a couple hours each
6:40 night because this AI stuff is is cuckoo
6:45 and what's been interesting is I
6:48 promised this community that 2024 was
6:51 going to be Bonkers batshit
6:55 crazy and I'll be honest with you it's
6:57 been a bit of a [ __ ] snore it's been
6:59 been it's been a little
7:02 sleepy until
7:06 today and it wasn't Google Gemini Ultra
7:09 that did it that that [ __ ] landed
7:12 like
7:13 a brick made of
7:15 turds but but Google did have a really
7:19 interest there was actually two big
7:21 announcements today and one of them was
7:23 from Google and it wasn't from Gemini
7:25 Ultra it was from Gemini
7:28 Pro which went from 1.0 to 1.5 and looks
7:31 like it's really [ __ ] good like a
7:33 million tokens you can upload like long
7:36 videos and it will summarize them [ __ ]
7:39 like
7:41 that so that's a big
7:43 deal and then and then our our favorite
7:47 old our favorite topof the
7:51 list AI company the little company that
7:56 could open AI launched a text to video
8:00 model today which is actually
8:04 Sora
8:08 s oh and I and I want your help on
8:11 something I want your help I just I I
8:14 just sent a tweet to Sam Alman so he's
8:17 doing something really cool right now so
8:18 I want you to I want you to boost
8:20 something for me actually let's go do it
8:21 right now um I want to be an a regular
8:23 Detroit J Holden you've been showing up
8:25 here here's the requirement of being an
8:27 a regular you just show up to this tick
8:29 talk on a regular basis and you're
8:31 [ __ ] in but you should also join the
8:33 AI salon and go hang out in the
8:36 Irregulars club we have a channel on the
8:39 AI Salon um
8:41 Network and in there we've got our own
8:43 private club so like when we're hanging
8:45 out here at night you can share pictures
8:46 and we can have inside jokes and things
8:48 like that it's really cool uh so if
8:50 you're not part of that you got to be
8:51 part of that did you make a request to
8:54 Sam I did so here's what we're going to
8:56 do I want you to go to the site for
8:59 formerly known as Twitter
9:02 and all
9:04 right so Sam posted we'd like to show
9:07 you what Sora can do so so Sam Alman
9:11 samama so go find at samama and then
9:14 find this post we'd like to show you
9:16 what Sora can do please reply with
9:19 captions for videos you'd like to see
9:21 and we'll start making some so Sam is
9:23 personally taking taking requests it's
9:27 like the AI learning lab he's personally
9:29 taking requests to make videos of
9:33 prompts and then he said don't hold back
9:35 on detail or difficulty so I posted
9:39 opening shots from my feature screenplay
9:41 I put me feature screenplay but whatever
9:43 I'm Irish opening shots from Miss
9:45 feature screenplay salmon flying East um
9:49 so so fade in exterior Gig Harbor
9:52 Washington docks 1991 pre-dawn foggy
9:55 Northwest air hangs over the glassy
9:57 water of the Puget Sound cut to a white
10:00 fishing boat with blue stripes rest in a
10:02 creaking slip interior fishing boat
10:05 cabin docks Gig Harbor following Jack a
10:08 weathered regular right so the first uh
10:10 like four visuals from this movie I
10:13 wrote that open the movie and then I put
10:15 bonus here's a GPT where you can play
10:18 with the screenplay so my request is go
10:21 just boost the [ __ ] out of this tweet
10:23 let's see if we can get Sam to make make
10:25 this image or make this movie this this
10:27 opening sequence all right right go what
10:31 are you waiting for all right here's
10:33 what I want you to do go like retweet it
10:35 or hard it or whatever the hell you do
10:37 or tag Sam that's it repost it and tag
10:39 Sam and like you got to make the opening
10:41 scene of this movie and then how do I
10:44 find it go to um here's how you find it
10:47 you go to well you could go search for
10:49 my go to Kyle
10:51 Shannon and then so on
10:54 Twitter if you go to my
10:58 profile um you go to replies so I
11:01 replied to his thing so go to Kyle
11:04 Shannon go to
11:06 replies and then it's like the fourth
11:09 one yeah the fourth one down right here
11:12 Sam mman don't hold back on the detail
11:14 or difficulty and then I have opening
11:16 shots from M feature
11:20 screenplay all right and yeah exactly
11:23 hard it repost it tag Sama at
11:26 Sama go go
11:32 alman's going to be like the [ __ ] is all
11:34 these people tweeting this stupid tweet
11:36 with a fragment of a screenplay all
11:39 right did you find it everybody find it
11:41 so so here's what I want you to do go
11:44 there do that do give it loving and then
11:46 come back here and tap the screen show
11:47 me the hearts we'll get some get some
11:49 some loving going here and then then
11:52 I'll tell you what the hell this
11:53 Channel's all about listen here's what's
11:56 important all of the self-serving stuff
11:59 for me personally has to go first all
12:02 right it's just the way it
12:04 is this is about me people it's not
12:07 about you okay I I know I know I've said
12:11 I appreciate you in the past it was all
12:13 a ruse it was all for this
12:16 moment I want my tweet tweeted to
12:20 Sama what did I miss oh cuno kruno so
12:24 look go to my go to my Twitter account
12:27 go on Twitter and go to act Kyle Shannon
12:30 that's me see that's an AI generated
12:32 version of me in front of a bunch of AI
12:34 generated versions of me cuz I can't get
12:37 enough of me see how this works and then
12:41 scroll down and go to the replies so the
12:44 second little tag
12:46 there and then scroll past all this [ __ ]
12:49 and get to the Sam mman reply so Sam
12:53 mman said don't hold back in detail or
12:55 difficulty I'll tell you what this is in
12:56 a second if you're just getting here and
12:59 then my tweet is the opening of my
13:01 screenplay so I want here's I'm I'm
13:04 trying to get Sam Alman to to put this
13:08 prompt in his Sora video generation
13:11 model so I can get the opening of my
13:13 screenplay done CU then I'll put it on
13:16 the internet and I'll sell my screenplay
13:19 and then I'll be in Hollywood I'll be in
13:20 Hollywood by the end of the week which
13:22 is
13:23 tomorrow and I and that'll be it for
13:27 this the the AI learning lab will be
13:30 over I'll be like taking calls from Matt
13:33 Damon and Scorsese it's going to be it's
13:35 going to be a rough breakup but I just
13:37 this is what I need it's what I need
13:39 right now all right what's Altman's uh
13:44 Twitter handle is s s m
13:48 a what's going on why is my did my mouse
13:51 die I think my mouse just
13:55 died hello Mouse oh come on really did
13:58 you just die
14:03 it dead it dead hang on
14:07 no now it's okay I've got a trackpad but
14:10 let me just charge I'll just charge it
14:11 up all
14:13 right I think I'll charge it up I just
14:15 dropped it
14:17 um all right so cool you cool on that
14:20 all right
14:24 so all
14:26 right so go go go go make Sam happy or
14:30 go make go make me happy I don't ask for
14:33 much here do I well I ask you to show up
14:36 here and you just do that anyway you
14:39 weirdos I'm not playing this game
14:41 Amelia's wife she shows up late and then
14:43 she's not going to retweet my tweet and
14:45 then you'll forget about us of course
14:46 I'll forget about you are you kidding me
14:49 when I'm sitting at the [ __ ]
14:51 Oscars getting ready to go up there to
14:53 receive my special Oscar for AI
14:55 generated feature
14:58 film you think I'm going to talk to you
15:02 people jeez you don't know anything
15:05 about how this world Works do you it's
15:07 so sad then we're not playing this
15:12 game I can't find your Tweet it's been a
15:14 while since I've used X go to Kyle
15:18 Shannon and then on my channel you'll
15:20 see tweets or posts I think they're
15:22 called post now and right next to it
15:24 you'll see replies and then scroll down
15:27 there it's like the fifth reply and
15:29 replying to Sam Alman no uhuh not doing
15:31 it oh forever hooked that's so that's so
15:34 petty after all I've given to you the
15:37 depth of generosity from this Channel
15:40 and I can't get one little feature film
15:42 made I don't understand you
15:45 people all right so if you're brand new
15:49 here to the channel my name is Kyle
15:51 Shannon we talk about AI stuff and
15:53 occasionally I will beg people to
15:55 retweet something I think it's the first
15:57 time I've ever done that
16:01 H I do try to get you to join the AI
16:04 Salon but that's you know that's for you
16:08 I'm salty Irregulars what the heck is
16:11 that Oscar winner Kyle Shannon exactly
16:13 there'll be banners in the back of the
16:15 Oscars the Irregulars I'll be like could
16:18 you please have those people removed
16:21 that's I don't know what they're talking
16:23 about I I don't know why they have
16:26 posters of llamas I I don't know I don't
16:29 know these
16:30 people anyway welcome
16:34 Hollywood to my spe wait oh I've got to
16:36 go okay yep yep no I'll just I'll oh I I
16:40 haven't received an award oh so I should
16:43 not be on stage got it got
16:46 it all right well I'll uh I'll meet the
16:50 Irregulars in the parking lot then okay
16:53 so um Oscar winner Kyle shanon
17:03 I don't know you know what may maybe if
17:06 if media keeps being
17:08 weird maybe at some point the AI
17:11 learning lab wins some sort of special
17:13 Emmy for for television in
17:18 Innovation he took a $20 million
17:20 Production Studio and make made it look
17:22 like a shitty home office and the
17:25 production values of the show were so
17:27 low that people thought it was just an
17:29 old guy sitting in an office was
17:33 remarkable
17:36 revolutionary peasants Kyle
17:38 2025
17:40 exactly oh all right so yeah so this
17:46 this uh open AI um sorus stuff is really
17:50 really exciting next up egot you're damn
17:53 right we're we're going to get a Grammy
17:56 a special Grammy the first Grammy given
17:58 to a a dog human
18:01 combo we're going to we're going to
18:02 start there we're going to turn that
18:05 into a a Broadway
18:07 show
18:09 where we're going to use Kaa realtime
18:11 image
18:12 generation and and we we'll be do it'll
18:15 be like an improv show with real time
18:17 image
18:20 generation Oscar Tony oh Emmy yeah this
18:24 Tik Tok channel will get an Emmy because
18:27 in 2027
18:29 the Emmys are going to open up Tik Tok
18:31 as a new
18:34 network and then of course my
18:37 oscar it's it's you think it's a [ __ ]
18:41 joke people it is not out of the realm
18:43 of
18:44 possibility shit's about to get
18:49 weird maybe the academy also likes
18:52 turtles exactly everything will be
18:55 Turtle
18:57 themed
19:00 Sydney [ __ ] what's happening you do look
19:03 like a celebrity I was surprised you
19:04 weren't one yet oh thank you very
19:07 much I uh I I don't know which level of
19:10 celebrity D CD dless dless
19:14 celebrity one of my favorites was some
19:17 someone said I was scrolling through
19:19 through Tik Tok and I thought Alec
19:21 Baldwin got fat so I stopped like nice
19:25 that's solid that's a good
19:27 troll
19:32 all
19:33 right oh man you'll get to use Alex
19:36 Baldwin for your stand in yeah exactly
19:39 exactly he can be my stunt double my
19:41 body double Cory sandley Pottery Corey
19:44 Sandler Pottery can you summarize what
19:46 Sora is I missed the beginning yes I
19:48 will I'm going to go through the I'm
19:50 going to go through the announcement in
19:53 detail and then I'm going to go to I'm
19:55 going to go to Twitter and and show you
19:56 some of the stuff that Sam alman's doing
19:59 people are giving impr prompts Kyle is
20:01 the coffee with lche it is with
20:07 lche it's
20:12 tasty all right let's get so here's the
20:15 deal if you have questions about AI pop
20:17 them in the comments below We've Got
20:19 Mods in here who will pin things I want
20:21 an animation eventually of orange head
20:23 and Frank the hot dog be be like an
20:26 adult swim swim show I know Jim I should
20:30 probably I should probably take that
20:32 [ __ ] more seriously I'm there's
20:34 something about Frank that's just he
20:36 just cracks me the hell up if you don't
20:38 know Frank by the
20:41 way we'll show you a little bit of
20:52 Frank here's
20:54 Frank okay I'm going to try this
20:56 retweeting thing thank you side hustle
20:57 Mimi
20:59 I appreciate it yeah retweet it repost
21:01 it make sure you tag
21:04 SS ma that's Sam alman's Twitter handle
21:08 so tag him and say make the beginning of
21:12 Kyle's thing he'll talk about on a
21:13 stupid Tik Tok
21:15 Channel comments are flying tonight oh
21:18 yeah fantastic B tell him what he's
21:21 won you make
21:24 money so listen things got a little slow
21:28 in a i learning lab you mean you mean
21:30 the Frank that's on my refrigerator this
21:33 is that Frank this Frank was made
21:36 Popular by Source Camp who screenshotted
21:40 this printed this out and Frank is
21:42 hanging on her refrigerator like some
21:45 sort of Twisted grade
21:48 school creativity Pride
21:53 session okay done but you better not
21:55 quit us oh Amelia's wife I would never
21:59 quit you I was just joshing I'll just
22:02 when I when I receive my oscar I'll uh
22:05 what will I do I'll wear I'll wear a
22:06 turtle baseball
22:10 cap um Okay
22:14 so [ __ ] was a little slow like the month
22:18 of January came and went there wasn't
22:20 any really big announcements February
22:22 kicked off we're like ooh Ultra is going
22:23 to happen Ultra happened and it was a
22:25 piece of
22:27 poop and then in the AI Salon Lord
22:31 digital Gods um discovered this thing
22:34 from Adobe Express called animate from
22:38 audio where you can just record audio
22:42 you you pick you pick a character which
22:44 are just these hor horribly horribly
22:49 Illustrated characters and then you pick
22:51 a background it's you know it's like
22:53 Windows 95 [ __ ] clip art but anyway
22:56 it's fine it's fine they're cute they're
22:59 cuish uh you pick a character you pick a
23:02 background and then you pick an aspect
23:03 ratio like do you want it for Tik Tok do
23:05 you want it wide whatever you want and
23:06 then you talk and then it does a little
23:08 processing and it makes a thing so like
23:11 there was nothing going on we didn't
23:12 have anything going on so we made this I
23:15 put the Twitter Link in Show and Tell oh
23:17 cool okay so I'll Point people there as
23:19 well thanks thanks Daniel that's great
23:21 but but we've got to get them to meet
23:22 Frank for so so this is this was this
23:25 was the second of these characters that
23:27 I did the first one was very dull pencil
23:29 and this was
23:30 Frank so I come down to the field to
23:33 play the game and they're like hey Frank
23:34 Frank what are you doing I'm like I'm
23:37 here for the game they're like Frank
23:39 it's Thursday there's no game today I'm
23:41 like you got to be [ __ ] kidding me
23:43 there's no game today I'm ready for the
23:45 game they're like Frank Frank you look
23:48 [ __ ] delicious I'm like hey hey hey
23:50 back up hey whoa whoa whoa whoa who the
23:52 fell get your hands off my
23:54 bun oh my goodness so so I leave I go to
23:58 down to the the Piggly Wiggly there I
24:00 get myself a soda I come back they're
24:01 like they're like they're like Frank uh
24:03 we we got an issue like what's that it's
24:05 like Tommy's here oh what do you mean
24:08 Tommy's here yeah Tommy's here what's
24:11 going on with Tommy Tommy's hungry oh no
24:14 no no no no Frank it's going to be okay
24:19 it's not going to be okay you know D
24:21 doesn't stop a [ __ ] eaten oh my oh
24:24 Tommy Tommy no don't oh God oh God
24:28 that's it it's stupid it's stupid so
24:31 this this is this is how we were
24:33 entertaining ourselves for for for like
24:36 this is couple of nights so but now we
24:39 have actual entertainment okay
24:42 so let me go show you what the hell's
24:45 going on well should I should I yeah you
24:47 know what let me let me okay first thing
24:50 we're going to do we're going to do a
24:51 little little something here I want you
24:53 to go if you're new to
24:56 AI one of the most important important
24:58 things you can do right now is be
25:02 curious and get connected into Community
25:05 you can show up here every night the the
25:07 community that shows up here every night
25:09 is amazing and and this is kind of an
25:12 extension community of this this group
25:14 called the AI salon so the AI Salon I
25:18 started the week after chat GPT came out
25:20 so it's been around for about a year um
25:23 we just launched our web Community our
25:26 web-based community platform
25:28 um like 3 months ago and we're we're up
25:30 to almost 1,200 members and so this
25:33 group is amazing so you go to um the
25:36 salon. A and that's going to take you to
25:41 um this link
25:42 tree um which is got it's got a merch
25:45 store so so the first of all the AI
25:46 Salon doesn't cost anything and it
25:48 doesn't cost anything because we have
25:50 people that support us like Sherry D is
25:51 supporting uh sponsoring the month of
25:54 February also if you buy any merch that
25:57 that supports the salon but it doesn't
25:58 cost anything but the first link here AI
26:01 Salon Community will take you to the AI
26:05 salon and it should drop you right into
26:08 this page so welcome to the AI salon and
26:12 my request if this sounds interesting to
26:14 you this is kind of
26:16 like if you're that person in your life
26:19 that you you're like I think this AI
26:21 stuff's really cool but every everyone I
26:23 talk to about it's like could you shut
26:24 up about AI I don't want to hear about
26:26 your stupid AI what's this AI thing
26:28 aren't the robots going to kill us it's
26:30 stealing isn't it thievery it's the
26:31 world's largest plagiarism
26:34 engine if you're tired of hearing people
26:36 like that in your life this group's the
26:38 opposite of that this group is like AI
26:41 Enthusiast people are choosing to be on
26:42 this adventure
26:44 together and understand that nobody
26:47 knows what's going on so there's like a
26:50 real camarad and enthusiasm to
26:56 learn and and it's a remarkable group
26:59 and so what I would love for you to do
27:01 is read about who we are why we do what
27:03 we do how we do it and then I really
27:05 want you to look at the values of the
27:07 community because this is who this
27:09 community already is and as it grows I
27:13 don't want to lose
27:15 this so the people in here are curious
27:18 curiosity is one of the values they're
27:20 also generous meaning they learn
27:23 something and and and you're like oh how
27:24 did you do that and you're like oh I did
27:25 it like this and I did it like that and
27:27 like just people are helping one another
27:29 new people come in and introduce
27:30 themselves they're like oh welcome you
27:31 should go check out this go check out
27:35 that um exploration is one of the values
27:39 people in here are adventurous they're
27:40 like choosing to be on this adventure
27:42 they're choosing to figure it out when
27:44 something like Sora comes out this new
27:47 open AI video
27:49 model this group is like on it they're
27:51 like oh there a new thing go look is it
27:54 is that as big a deal as I think it is
27:56 oh [ __ ] what we oh no it's not that big
27:59 oh yeah it's
28:00 big it's like
28:03 that um collaboration is one of the
28:05 values like like a lot of people in here
28:08 starting things together uh doing
28:10 workshops together we have guilds now so
28:12 we have guilds and the guilds are now
28:14 starting to do mashups I'm going to talk
28:15 about one of those tonight from the from
28:17 the art and the writing Guild are
28:19 getting together to do like a six- week
28:22 creativity Workshop that you should
28:24 absolutely do inclusion is one of the
28:27 values like people are coming from
28:29 everywhere everyone's going to be
28:30 impacted by
28:32 Ai and because no one knows what they're
28:34 doing with this stuff yet people are
28:36 coming from everywhere so like Embrace
28:38 that because someone who comes from some
28:41 bizarre point of view where you're like
28:42 that doesn't make sense that they're
28:43 doing anything with AI they're going to
28:45 give you the big
28:48 breakthroughs check out a dude named Jim
28:51 Ross who I think's in here tonight I saw
28:52 you earlier I don't know if you're still
28:54 here Jim Ross is in the Self Storage
28:57 business Self Storage management and he
28:58 trains other people how to do that he
29:01 started using cat gbt the week it came
29:04 out this is a guy that's just like every
29:06 time there's a new tool he's like I'll
29:08 try it he's like I'll put it out there
29:11 my f my favorite story of Jim is I I was
29:15 I was showing how to make sunno songs
29:17 not Sora suo sunno songs one night and I
29:21 said look you can make stupid songs and
29:22 he had just met with a prospective
29:24 client and he goes why don't I make him
29:27 a song about self storage and like the
29:29 name of their company so so he goes make
29:31 a song about you know this these people
29:33 at this company talking about self
29:35 storage he sent him the song five
29:37 minutes later he gets back an email that
29:39 says you're hired let's start working
29:41 together so it's like who would even
29:43 think to do that well people that you
29:46 don't expect that and they're in this
29:48 community um empathy is another one of
29:51 the
29:52 values this shit's scary to a lot of
29:54 people I posted the thing about um I I
29:58 did a I did a video before I came on
30:00 live tonight and some of the comments on
30:03 there people are scared right people are
30:05 angry this one dude's like yeah it's
30:08 stealing I'm never going to use it I'm
30:09 like you know dare to understand how it
30:11 works and dare to
30:13 understand you know how to use it well
30:15 enough to understand where the
30:17 boundaries of it are and he said well I
30:19 might be willing to try as long as it's
30:21 ethical I'm like well step in start
30:23 playing start understanding what it is
30:25 you can have a voice in that in fact one
30:28 of the things that when we talk about
30:30 Sora tonight they're red teaming it
30:32 right now so they're doing testing for
30:34 safety but they're also um they're
30:37 giving access to it to creative
30:39 Professionals in the film industry
30:40 because they want input from them and so
30:43 that guy instead of just bitching and
30:44 complaining if he got involved could
30:46 actually have a voice in the future
30:48 direction of that like that's how early
30:50 we are with this stuff so anyway have
30:53 empathy for people understand that
30:54 people are in a lot of different places
30:56 with this and the final one is gratitude
30:57 ude this is our newest value we just
31:00 added it like a week ago thanks to
31:02 Danielle who's in here be not afraid it
31:05 was her idea and I think it's a really
31:07 good one so I just explicitly put it out
31:10 there you know what we should
31:11 acknowledge one another
31:14 more even if we're generous right and so
31:18 what we wrote here is We Believe
31:19 gratitude is a practice gratitude is not
31:21 just a feeling but an action take the
31:24 time to acknowledge the efforts support
31:26 and successes of the EXT extraordinary
31:28 people in this community especially if
31:30 someone inspires you go out of your way
31:32 to acknowledge them tag them in your
31:34 post reach out and let them know they
31:36 made a difference holy [ __ ] that's a
31:39 cool Community to be a part of imagine
31:41 if people did that on a regular basis
31:43 that's who this community is all right
31:45 so if those values sound bitching and
31:48 awesome great these are your people if
31:51 they sound awful then they're not right
31:53 and that's cool too but that's what this
31:56 this community is up to and then my
31:58 first request is go to the introduce
32:00 yourself
32:01 section and and introduce yourself
32:04 Embrace that that
32:06 um that value of
32:09 generosity and tell us who you are tell
32:11 us what you're up to tell us where you
32:12 are with the AI stuff even if you're
32:14 brand new or even if you've been in it
32:16 for 30 years and you're cynical about it
32:18 now and you're like I've been telling
32:20 people about AI for 30 years and all of
32:22 a sudden you little Whipper Snappers
32:23 coming up here talking about AI like
32:25 it's a new thing it ain't a new thing
32:28 jump
32:29 in tell us what's going
32:32 on and then the show and tell Channel as
32:35 someone just said that there's there's
32:36 cool things here this is a a fun place
32:39 where you can share
32:41 things oh that's really cool marrying
32:43 the elements love it oh cool oh that's
32:47 really cool water and fire that's a good
32:50 series that's going to be a good series
32:52 beautiful beautiful beautiful all right
32:55 here's the link so the third link thank
32:58 you Daniel the third link in Show and
33:01 Tell right now is the link to my Twitter
33:04 post that I want you to go tag Sam mman
33:07 in so wait here let's see you gotta uh
33:10 I'll put in here um don't forget to
33:16 tag
33:18 Sama on
33:20 Twitter and then I'll put x.com for
33:26 you people who fall for their branding
33:40 nonsense all right so so go do
33:44 that and then and then yeah and then
33:47 maybe Sam will make the uh the first
33:49 scene of my
33:50 screenplay um the other thing to to to
33:53 pay attention to here at the salon is
33:56 there's a news channel there's also Al
33:57 AI tools and resources this
34:00 list um is where you should start like
34:05 if you're brand new with AI and you're
34:06 like I don't but Kyle I don't even know
34:08 where to
34:09 start I I hate to break it to you Kyle
34:11 but I'm not very computer
34:14 literate I don't know how to use the Tik
34:16 Tock how do you turn this thing off um
34:20 if you like that it's okay here's where
34:23 you
34:24 start you go to chat. open i.com you
34:28 start with chat gbt and you can start
34:30 with the free version there's like it it
34:33 does
34:34 enough that if you haven't used a large
34:36 language model before just start playing
34:39 with it and watch what it does and just
34:41 your mind will be
34:43 blown hopefully you'll have your first
34:45 what we call your Kevin McAllister
34:48 moment remember home
34:50 alone and then if you can afford it pay
34:53 the 20 bucks for chat GPT
34:56 Plus why
34:59 because it is the best deal in software
35:02 right now and Sama is not paying me to
35:04 say that he doesn't even know exist well
35:07 you might know now now that you've been
35:09 tagging him I'm a dumb
35:14 tweet but he he don't know nothing I
35:17 don't make any money if you go buy 20
35:19 bucks a month
35:22 there but then below that are a bunch of
35:25 is this working yet where
35:28 okay connected yeah it's back okay Mouse
35:29 is back good okay this is also a list of
35:33 other things to check out other large
35:35 language models Claud Gemini Advanced
35:37 perplexity AI Microsoft co-pilot
35:40 specialty llms Pi grock llama 2 po image
35:44 creation Dolly 3 mid Journey Crea did
35:47 you see the the mid Journey 6 text thing
35:50 the the text that mid Journey 6 can do I
35:53 did a post about that on LinkedIn today
35:55 I thought that was the big news of the
35:56 day
35:58 and then um Gemini Pro 1.5 dropped and
36:02 then um open a is saur dropped so it
36:06 it's it was a big day um and there's
36:10 video creation Tools in there I will add
36:13 in parentheticals open AI Sora at some
36:16 point here um then there's audio tools
36:19 3D tools and then there's people to
36:21 follow so if you're just getting started
36:23 what I can promise you about this list I
36:26 personally maintain this
36:27 list there's no [ __ ] on this list
36:31 there's nothing here that will waste
36:32 your time there's nothing here where
36:34 it's like ridiculously like I don't know
36:38 it's not overly janky it's not overly
36:40 you know expensive a lot of these have
36:42 free tear some don't but most
36:46 do but none of these will waste your
36:48 time Irregulars New Image let's go check
36:51 okay so then let's keep talking about
36:53 this let's keep going down I'm going to
36:55 come back to guilds because we're going
36:57 to talk about a specific one but but if
37:00 you go past the guilds which the guilds
37:02 are really cool and go to clubs there's
37:05 this club called a i l Irregulars that's
37:08 us we have our own club Cory Sandler oh
37:12 wow this was the prompt create an image
37:14 of beautiful spirals fractals God that's
37:19 gorgeous
37:22 unbelievable remember in the 90s when
37:25 when like fractals first hit the scenes
37:28 and you could get like the you could get
37:31 the mathematical formulas and you could
37:32 put them into programs and it would like
37:34 draw the
37:35 fractals and now you can just ask for
37:38 like an incredibly
37:43 sophisticated an image of beautiful
37:45 spirals fractals metal filigree waves
37:48 rain and it just makes that
37:51 [ __ ] I'm telling you man this is we are
37:54 just in such a
37:56 different
37:57 era G that's
38:00 gorgeous oh by the way so this is Corey
38:03 Sandler so on Tik Tok at Corey Sandler
38:07 Pottery follow
38:10 her she's
38:12 remarkable she does Pottery which is
38:14 beautiful but she like generously just
38:16 shows you working in her
38:19 studio and then she sings and then she
38:22 talks and
38:23 cooks and now she's a flipping GPT
38:28 developer she's out there making
38:32 gpts and she'll show you how she uses
38:35 these these AI tools and it's
38:37 fascinating watching an artist Embrace
38:41 these tools and see what they do with it
38:42 and she so she's one to watch you'll
38:44 learn a lot you'll learn a lot from her
38:47 oh that's nice cam so much going on
38:48 today yeah exactly we don't have enough
38:50 eyes do
38:53 we gorgeous gorgeous gorgeous there I am
38:56 look at me winning my my oscar my turtle
38:58 my turtle infused Oscar very nice I'm
39:01 pretty sure that's Alec Baldwin but you
39:03 know that's all
39:04 right is that these were Cory Sandler
39:07 too yeah the uh the circus ones really
39:10 cool Sora Sora Sora Sora Sora the
39:18 Irregulars you wish you wish it would be
39:21 like that although hey you never know
39:24 man maybe maybe this becomes the uh
39:28 maybe this becomes the thing that goes
39:29 up on the Marquee why not nice thanks
39:32 Sam champ champy and the Irregulars
39:35 that's me that's good yeah exactly after
39:38 I'm a Hollywood bigwig you know you know
39:41 I'm just going to be eating barbecue
39:44 every like breakfast lunch and dinner
39:47 that's totally what I'm going to look
39:54 like all right so that is the
40:01 um that's the AI
40:03 salon and
40:05 then if you go under announcements at
40:08 the very top Salon
40:10 announcements I just put a new thing in
40:12 here um so this is a link to the prompt
40:16 creativity live kickoff Jam so Leah
40:20 faston who runs the Art Guild she also
40:22 co-founded the AI Salon with me so Leah
40:24 and I co-founded this December 7th was
40:27 the first meeting of the salon in
40:31 2022 and so Leah has teamed up with Jane
40:35 Endicott who runs the Writers Guild and
40:38 they're doing like this six week daily
40:42 exercise to break you out of if you
40:45 think you can't be creative or you don't
40:47 have ideas or don't know what to do with
40:48 these AI things or can't quite think
40:50 about it they're doing this crazy thing
40:53 that's kicking off tomorrow at 8:00 p.m.
40:58 Eastern 6:00 p.m. my time 5:00 p.m.
41:03 Pacific so go sign up for that [ __ ] do
41:05 you know a lot about funnel building AI
41:08 such as gamma I don't Kayla um anything
41:12 that's super tactical like that I tend
41:13 to I don't I stay I don't I stay away
41:16 from it but I don't stay away from it
41:18 because it's bad I stay away from it
41:19 because there's too much of it there
41:20 there's too many specific little niches
41:23 so what I would say is go to at the
41:24 bottom of the resources list is is a
41:27 directory called futurepedia
41:29 doio so what I would do is go to Future
41:55 pedia.com
41:57 okay now that's
42:01 that so go to the go to the salon
42:04 announcements join
42:06 that um go to the Irregulars club and
42:10 post anything you're working on tonight
42:12 songs
42:13 whatever
42:15 um go to show and
42:18 tell and somewhere down in there is the
42:21 link to the tweet that I want you to tag
42:23 Sam Alman
42:25 in
42:28 that's just for me and then AI tools and
42:31 resources if you're just getting started
42:34 and you want to figure out well where do
42:35 I start what do I do what I tried to do
42:37 with chat gbt was I tried to put a
42:39 bullet point for everything you should
42:41 learn about
42:42 it how do I learn about it you just
42:45 start playing you just play play play
42:50 and then if you're like B cow but cow
42:53 cow cow cow cow cow cow cal cal how do I
42:58 make
43:01 money don't think about that right now
43:04 you can make money
43:05 with it's not from these buttheads that
43:07 make those videos I can promise you that
43:10 they're making money by getting you to
43:12 pay
43:13 them to tell you how to make money with
43:16 chat GPT for a course that they probably
43:18 made with chat
43:20 GPT that's how they're making their
43:22 money it is not with Chachi
43:25 PT
43:28 um all right my live stopped playing in
43:31 the background just
43:43 now going to try to stay awake forever
43:46 hooked is Sleepy there's no sleeping in
43:51 AI all right here we are look
43:55 puppies
43:57 okay so here's where we are people I
44:01 panicked because I thought it was
44:02 tonight and I missed it so excited to
44:04 join tomorrow night so Cindy [ __ ] it was
44:06 supposed to be tonight and and they got
44:08 their wires crossed on timing so it's
44:11 actually going to be tomorrow night two
44:13 hours before I go
44:15 live so that's pretty cool yeah that's
44:19 that's a good one
44:21 um whoa picture and pictur is off the
44:24 screen for me I know Tik Tok just freaks
44:26 out every now and again
44:29 lately okay
44:32 so you ready to jump into this [ __ ] this
44:35 shit's
44:36 cockoo I'm gonna do what I did I'm going
44:39 to do what I did uh earlier
44:41 today um which is I'm just going to go
44:43 through this document then we'll go play
44:45 on Twitter maybe by the time we get
44:47 there Sam Alman will have created my uh
44:49 my Puget Sound
44:53 movie
44:55 um
44:58 okay so if if you want to follow along
45:00 if you want to play along you don't have
45:01 to wait for
45:02 me go to open ai.com
45:07 a SORA s o r a and that's that's the
45:12 site I'm on and so I'm going to read
45:14 through this I'm going to show you all
45:16 the videos I'm going to talk about stuff
45:18 I'm happy to answer any questions um
45:21 first question I'll answer this is not
45:22 accessible right now it's in red teaming
45:25 which basically means they're doing
45:27 safety um safety testing and they're
45:31 also giving it to a handful of creative
45:33 people to to test and and sort of figure
45:35 out like what are the implications of
45:37 this for the creative industry because I
45:39 think they're pretty profound there's
45:41 also some
45:42 hints in what they're saying here that I
45:45 think are pretty
45:48 crazy this for for me this is the first
45:51 shoota drop in
45:52 2024 that is proof that shit's going to
45:55 get weird pretty quick
45:57 because this is
45:59 cool but what they talk about here is
46:02 Way Beyond video so the the first
46:04 paragraph we're teaching AI to
46:07 understand and simulate the physical
46:09 world in motion with the goal of
46:12 training models that help people solve
46:14 problems that require real world
46:21 interaction that didn't say anything
46:23 about video we're teaching the AI
46:26 teaching AI to understand and simulate
46:29 the physical world in motion with the
46:32 goal of training models that help people
46:34 solve problems that require real world
46:38 interaction that's a big [ __ ] that's
46:41 not we made a cool video
46:44 tool what you'll see as we go through
46:47 this is the reason the video is so good
46:50 is that the model understands the
46:54 physical world context that the objects
46:57 are interacting with which is
46:59 mindblowing if you understand how
47:01 diffusion models work it's like holy
47:05 crap all right so after that little
47:08 provocative opening then they get to
47:11 introducing
47:13 Sora so so we're trying to figure out
47:17 how the world works and we're training
47:19 AI models to do that [ __ ] oh and by the
47:23 way here's a cute little video tool all
47:26 text to video model Sora can generate
47:28 videos up to a minute long while
47:30 maintaining visual quality and adherence
47:33 to the user's prompt to be to to put
47:36 that in context the state-of-the-art as
47:39 of today Runway ML and pabs let you
47:44 render 4 seconds of
47:46 video this is up to a
47:49 minute very rarely in in even a feature
47:52 film do you have a segment that's longer
47:55 than a minute right right you're usually
47:56 cutting but it's usually longer than 4
47:59 seconds so you can have a minute of
48:02 video generated from a single prompt
48:04 here okay so let's go look at some of
48:06 these things I'm going to read the
48:07 prompts the the first one you've
48:08 probably seen if you've been paying
48:10 attention at all
48:13 today so this
48:18 is the prompt here
48:21 says a stylish
48:24 woman walks down a Tokyo Street filled
48:27 with warm glowing neon and animated by
48:30 City signage she wears a black leather
48:32 jacket a long red dress black boots
48:35 carries a black purse she wears
48:37 sunglasses and red lipstick she's
48:39 walking confidently and
48:41 casually um the street is damp and
48:45 reflective creating a mirror effect of
48:47 the colorful lights now watch see that
48:49 cut right there that's not an edit of a
48:53 separate generation that's all in a
48:56 single generation it's it's doing
48:59 internal
49:00 scene generation and and cutting itself
49:04 this is a 59 second clip out of that
49:07 single
49:12 prompt
49:14 Bonkers and then the cool thing about
49:17 this site is every one of these little
49:18 sections there's these this is a
49:20 carousel so so there's nine videos here
49:22 so that was the first one and each
49:24 little section has its own video
49:27 so here's some woolly mammoths I think
49:29 that one's okay not great this one's
49:33 stunning The Prompt here
49:37 is a movie trailer this this is what
49:40 gave me the idea of putting salmon
49:41 flying East in there a movie trailer
49:43 featuring The Adventures of a
49:45 30-year-old Spaceman wearing a red wool
49:48 knitted motorcycle helmet Blue Sky salt
49:52 desert cinematic style shot on 35 m M
49:56 millimeter film Vivid
49:59 colors all of these edits
50:02 are the this the clip that it generated
50:05 is all these different edits in a single
50:11 clip like it's telling a
50:17 story [ __ ]
50:20 crazy crazy
50:22 right oh me thinks Hollywood is in
50:24 trouble still need cons
50:26 consistency yeah but but wait yeah
50:30 consistency yeah but the the consistency
50:33 here is actually quite good like
50:35 it's it knows this character like it's
50:39 it's keeping track of what's Happening
50:41 happening with that character off screen
50:44 as he comes on
50:46 screen wait I'm a local cinematographer
50:48 what was
50:49 that
50:51 um and I'm equally curious as I am
50:54 terrified so m Mike the Hamman here's
50:57 here's my my strong
50:59 recommendation absolutely you know feel
51:02 the
51:03 fear but jump in and start playing you
51:06 can't play with this now but I if you're
51:07 not playing with Runway ML and P and all
51:09 that sort of stuff start playing with
51:11 the with the the um AI video generation
51:15 tools so that it just becomes another
51:17 tool in your in Your Arsenal right that
51:21 you know you can pick up a a red 4k
51:24 camera and you can pick up one of these
51:26 tools and you can mix those together
51:28 right storytelling is still storytelling
51:31 um you choosing how to prompt this and
51:34 how to put these things together and
51:36 understanding which tools to use where
51:38 and which ones are valuable and which
51:40 ones aren't that will be every bit as
51:42 much a skill as your current skill of
51:44 being able to choose a lens um do color
51:47 grading all the stuff that you've you
51:49 know you've learned to do in your
51:51 current craft this just going to be
51:52 another set of tools so I would say lean
51:54 into it um someone uh pinned P's thing
51:59 there I didn't see it so if you can
52:01 repin
52:02 that Mike the Hamman oh I definitely am
52:05 LOL great awesome yeah yeah yeah good
52:08 yeah jump in jump in okay um so that's
52:11 that one I thought that one this one in
52:13 particular I thought was really
52:14 impressive um this one for a drone shot
52:16 if if I'm a if I've got a drone video
52:19 business this one would scare the piss
52:21 out of me this is a 8sec long drone
52:25 panning shot and like if if if you're
52:28 used to watching AI generated video wait
52:31 since when have we decided that AI
52:34 marketing things was
52:35 trustworthy just anytime it's open AI I
52:39 don't I don't know that this is
52:40 trustworthy and I you'll see you'll see
52:42 p i I'll show you some things in here
52:44 they even call out things that this
52:45 isn't good at um I would
52:50 say Okay pay so here's the deal I would
52:54 say if if we're going to compare two
52:57 random companies like say Google as just
53:02 I'm just picking that one just kind of
53:03 randomly and open AI just cuz you know
53:06 you mentioned it if I look at what
53:08 they've presented in marketing and what
53:10 they've delivered in the marketplace I
53:12 would say that open AI has a better
53:14 track record from my experience over the
53:16 past year your fan fanboying hard for
53:20 openai I I am yep preview of
53:24 open.com Sora I am because I think that
53:28 this
53:30 is well particularly with with how
53:34 they're talking about why why it's as
53:36 good as it
53:38 is um I'm reading between the lines and
53:40 assuming that there's a hell of a lot
53:42 more underneath this now is it possible
53:46 this [ __ ] doesn't work or it's going to
53:47 cost us you know $80 you know for 100
53:50 credits and that's going to get us 100
53:52 seconds a video it's possible but I
53:54 doubt it
53:57 um all right let's see this one's cute
54:00 the the flame on the candle's really
54:01 weird and not good but the dude is
54:05 good like really good
54:08 animation um this one's pretty wild The
54:10 Prompt is a gorgeously rendered
54:13 papercraft world of a coral reef Rife
54:16 with colorful fish and sea creatures by
54:18 the way Pate to defend myself a bit here
54:22 when the when the Gemini Ultra marketing
54:24 [ __ ] came out
54:26 I did two lives that day and then I did
54:31 I I spent two or three nights unpacking
54:34 and watching every one of those [ __ ]
54:36 videos so I I can't help it that that uh
54:41 Gemini Advance did not live up to what
54:44 the marketing did there I would still be
54:46 talking about it if it if it did but it
54:49 didn't so if this is a pile of [ __ ] when
54:51 they launch it I'll absolutely call it
54:54 out all right um this is a paper a reef
54:58 of paper creatures wait a gorgeously
55:00 rendered paper craft world of a coral
55:04 reef with colorful
55:06 fish um cool
55:10 bird if you seen
55:15 um uh what's her name Kelly Bosch um she
55:18 does she's got this bird video it's very
55:21 much like
55:23 that look at the subtle little movements
55:26 there that's
55:27 amazing cool all right this one I
55:31 thought's pretty good pirate ships in a
55:33 in a cup of coffee come on that's pretty
55:36 good there's a little bit of coherence
55:37 weirdness there but it's not
55:40 bad um this is a dude on a cloud okay
55:43 today Sor will become Avail is becoming
55:46 available to Red teamers to access
55:47 critical areas for harms or risks so red
55:51 teamers are um looking at this from a
55:54 from a safety stand standpoint they're
55:56 trying to break it figure out where the
55:58 risks and vulnerabilities and you know
56:00 how people might use it in a [ __ ] up
56:01 way um we're also granting I think this
56:04 is important we're also granting access
56:07 to a number of visual artists designers
56:09 and filmmakers to gain feedback on how
56:11 to advance the model to be the most
56:13 helpful for Creative
56:15 professionals we're sharing our outside
56:17 research progress early to start working
56:20 with and getting feedback from people
56:22 outside open AI which they have a
56:23 history of and give the public a sense
56:25 of what AI capabilities are on the
56:27 horizon
56:29 right so here's a pile more of
56:33 these and some of these are weird I'll
56:35 show you a weird one here that one's
56:37 really good actually so this is a The
56:40 Prompt here historical footage of
56:41 California during the Gold
56:44 Rush it's so cool that it color graded
56:46 at like you know
56:48 50s like sort of 50s you know
56:54 recoloring
56:56 this is amazing I know isn't it it's
56:58 just nuts is Sora only for paying
57:00 customers Sora is not live right now
57:02 this this is it's in testing right now
57:05 sad thing about Gemini Advanced and
57:06 Gemini had gotten quite good Advanced is
57:09 a bit like earlyish SP yeah so so black
57:12 ice I'm G to after we do Sora here if
57:15 I've got time I'm going to head over and
57:17 look at I'm going to see if I can find
57:18 the uh the release
57:20 for Gemini Pro 1.5 which also launched
57:24 today which looks really good so so
57:28 Gemini Ultra 1.5 or whatever is going to
57:31 be the number will probably be even
57:32 better than that so um it that looks
57:36 pretty pretty pretty strong um all right
57:40 here is closeup view of a glass sphere
57:42 that has a Zen Garden in it in it
57:44 there's a small dwarf in the sphere
57:47 who's raking the Zen Garden creating
57:49 patterns in the
57:50 sand I mean come on look at the the you
57:54 know the reflection
57:56 right crazy this one okay The Prompt
58:01 here let me make this oh wait let me
58:04 read The Prompt first extreme closeup of
58:08 a 24 year-old woman's eye blinking
58:11 standing in maresh during magic hour now
58:16 watch right there that's the reflection
58:20 of maresh at Golden hour in her eye I'm
58:23 like holy [ __ ]
58:28 cinematic film shot depth of field like
58:31 that's a cool
58:32 one um this one's
58:37 okay not bad like you know the physics
58:39 are a little weird they admit that their
58:41 physics are weird his feet get
58:45 weird um this one's a beautiful homemade
58:49 video showing
58:51 people of of Lagos Nigeria in the year
58:55 2056 shot with a mobile phone camera a
58:59 little bit of jankiness in that one but
59:01 it's not
59:02 bad um this one's very weird so what
59:06 what you see in this one let me make
59:08 this
59:09 bigger is the these it starts out with
59:13 none of these dudes they just
59:15 materialize out of the petri dish which
59:19 could be a creative choice but it's
59:21 really weird that one's really weird
59:22 also all the little dudes only have two
59:24 legs
59:25 so you see them from the front they have
59:27 two legs and then when they turn
59:28 sideways they still only have two legs
59:31 so that one that one is one that I
59:33 looked at it and I'm like why would they
59:34 include that but you know whatever um
59:38 this one's this one's quite good the
59:40 camera rotates around a large stack of
59:42 vintage television showing different
59:44 programs 1950 Sci-Fi movies horror
59:48 movies news static a 1970 sitcom
59:53 Etc um set in
59:55 side I think it's a New York City Museum
59:58 yeah a large New York museum
1:00:00 Gallery right so you're in kind of this
1:00:03 Museum space it's you know some of them
1:00:06 are off some of them are on some of them
1:00:07 are
1:00:08 static some of them are horror flick I
1:00:11 mean it's like the prompt coherence here
1:00:13 is pretty
1:00:16 tight I think that looks this one looks
1:00:18 dumb so we'll go past
1:00:21 it um Sora is able to generate complex
1:00:24 scenes with multiple characters specific
1:00:27 types of motions and accurate details of
1:00:31 um subject and background the model
1:00:34 understands this is important the model
1:00:37 understands not only what the user has
1:00:39 asked for in The Prompt but also how
1:00:41 those things exist in the physical
1:00:44 world and so the thing that I'm thinking
1:00:46 of
1:00:48 is you know how when you do runway
1:00:52 sometimes like it'll be a car and you
1:00:54 like have a car on a 45° angle and
1:00:57 instead of the car like going the
1:00:58 direction a car would go it sort of
1:01:00 slides towards you or it slides away on
1:01:02 a 45° angle what this is saying is it
1:01:06 understands like the orientation of the
1:01:07 object how it would inter interact with
1:01:09 the physical world and then the first
1:01:11 example is the camera following behind a
1:01:14 white vintage SUV with black roof rack
1:01:17 as it speeds up a steep dirt road
1:01:19 surrounded by pine trees on a steep
1:01:22 Mountain slope dust kicks up
1:01:26 from its tires the sunlight Shines on
1:01:29 the SUV as it speeds along in the dirt
1:01:31 road casting a h a warm glow over the
1:01:34 scene the dirt road curves gently into
1:01:37 the distance with no other cars or
1:01:39 vehicles in sight like the The Prompt
1:01:42 coherence here is
1:01:43 tight um the trees on the other side of
1:01:46 the road are redwoods with patches of
1:01:49 greenery scattered throughout the car is
1:01:52 seen from following the Curve making it
1:01:55 seem as if it's on a rugged drive-
1:01:58 through rugged terrain the dirt road
1:02:00 itself is surrounded by steep hills and
1:02:03 mountains in a clear blue sky above with
1:02:06 whis wispy
1:02:07 clouds the the dust kicking up from the
1:02:10 the vehicle looks a little fake and
1:02:12 weird but man the like the the realism
1:02:17 like if you just lose the if you just
1:02:19 sort of cover that
1:02:22 up like it's that's that's in the that's
1:02:27 in the hood right that's definitely in
1:02:30 the Hood um this one blew me
1:02:34 away this one is
1:02:37 staggering so so the prompt
1:02:39 here Reflections in the window of a
1:02:42 train traveling through the Tokyo
1:02:44 suburbs and then look look at that
1:02:47 reflection as it passes a
1:02:50 building and it shows the
1:02:53 reflection of the woman you know filming
1:02:55 this on her phone are you kidding me and
1:02:59 again how how is it doing that well it
1:03:01 sounds like this is not just a video
1:03:04 model this understands that there are
1:03:07 different objects in the scene that are
1:03:10 in are in a reality like it understands
1:03:14 the physical world and how objects in
1:03:16 the physical world interact with the
1:03:18 other things in that
1:03:20 environment this isn't just about
1:03:23 video
1:03:26 there's another drone
1:03:28 shot um this one's got some janky ass
1:03:31 [ __ ] like the crabs just weird and
1:03:32 starts to act like an octopus the the
1:03:34 little suction cups on the octopus thing
1:03:37 sort of animate on their own that one's
1:03:39 weird I that doesn't seem like one they
1:03:41 should have included but
1:03:44 whatever um this one's really janky
1:03:46 watch her
1:03:48 arm her arm turns into a
1:03:51 pillow and then the cat gets an extra
1:03:53 paw
1:03:56 watch
1:03:58 this look cat now here's another paw
1:04:01 coming oh
1:04:04 wait okay her arm turns into a pillow
1:04:07 and then there's a second paw right
1:04:12 there all right so the Jank is still
1:04:15 there but still remarkably good like
1:04:18 this
1:04:19 one look at
1:04:22 the when it when it zooms back out at
1:04:24 the beginning of this look at the water
1:04:26 down here so you know how in Runway the
1:04:29 water will sometimes just sort of morph
1:04:31 into objects near it and it'll it'll get
1:04:34 weird [ __ ] in it it's got It's got a
1:04:36 coherence to it right again because if
1:04:39 they're doing what they say they're
1:04:41 doing where it understands the objects
1:04:44 in this environment and how they
1:04:45 interact with each other like you notice
1:04:48 the coherence of these three birds is
1:04:50 staying pretty tight they're not getting
1:04:52 weird morphe heads
1:04:57 you know the leaves are just kind of
1:04:59 moving
1:05:01 naturally
1:05:04 wild um this one was pretty good too
1:05:06 this is a kind of a natural scene a
1:05:09 Chinese Lunar New Year celebration video
1:05:12 with Chinese dragon like like this see
1:05:15 how like it maintains its shape it
1:05:18 doesn't go all f [ __ ] janky and weird
1:05:20 like all the the runway ml stuff does
1:05:25 all right so that's that the model has
1:05:27 deep understanding of language enabling
1:05:29 it to accurately interpret prompts and
1:05:32 generate compelling characters that
1:05:33 Express vibrant emotions Sora can also
1:05:37 create multiple shots with Sora can also
1:05:40 generate multiple shots within a single
1:05:42 generated video that accurately persist
1:05:45 characters and visual
1:05:48 style so it's doing those multiple edits
1:05:53 in a single generation
1:05:56 so here we are in one room in a in a
1:05:58 gallery oh no it starts in a different
1:05:59 room starts in that
1:06:03 room realize that our actual reality is
1:06:06 just layers of probability layers yeah
1:06:08 it's I know 101 This Is We I know we
1:06:12 it's weird we're getting into weirdness
1:06:14 I told you 2024 yeah gp4 is best
1:06:18 understood as a preview that's right Dr
1:06:22 bod can you imagine having this well we
1:06:25 don't have to imagine it this is very
1:06:27 likely going to be inside chat gp4 this
1:06:31 capability so right now if you haven't
1:06:33 played with do inside chat
1:06:37 gp4 you really should you should really
1:06:40 understand that it can see that it can
1:06:46 generate because this
1:06:48 capability you know you could be talking
1:06:52 about marketing for your business and
1:06:55 come up with a campaign and then just
1:06:57 say make me that
1:06:59 ad and then off it will
1:07:02 go and by the way the the goo the Gemini
1:07:07 Pro 1.5 one of the things they demo is
1:07:11 it's got a million token context window
1:07:16 and it's multimodal so you can upload
1:07:19 videos and it understands the videos and
1:07:21 cut can jump right to a specific scene
1:07:23 in the video
1:07:25 just with a prompt I wonder how long it
1:07:27 takes to render this this is all the
1:07:28 stuff we don't know
1:07:30 Joseph you know it it there's there's
1:07:33 two there's two really big unknowns how
1:07:35 long does it take to render these things
1:07:37 and how much will they cost right is is
1:07:40 this a $50 subscription on top of the
1:07:43 $20 subscription or is this included in
1:07:45 the 20 bucks if they include even a
1:07:49 limited version of this in the $20
1:07:51 version holy [ __ ] uh I I doubt that just
1:07:56 looking at the prices of other video
1:07:59 generation tools and and looking at how
1:08:01 long how long these things
1:08:04 are um I would think that they're going
1:08:06 to be expensive this one this one blows
1:08:08 me
1:08:09 away like the coherence here look at the
1:08:12 legs you know you know how legs are in
1:08:14 Runway where the the legs look like
1:08:15 they're just floating and morphing and
1:08:17 then there's a third
1:08:19 leg this unless you look for really
1:08:22 subtle specific things
1:08:26 doesn't look like AI generated
1:08:38 content like there's a weirdness
1:08:42 where she's above this awning and then
1:08:46 as they
1:08:47 walk like there there's other stands
1:08:50 underneath that awning like now it's an
1:08:51 appropriate height right there's some
1:08:53 weird subtle things but if you weren't
1:08:57 looking for it right if you weren't
1:08:58 paying tight
1:09:01 attention um this one's pretty good it's
1:09:04 a little it's a little underwhelming but
1:09:07 the coherence there is beautiful camera
1:09:09 movement's
1:09:10 interesting this one I love the prompt
1:09:12 on this one the story of A robot's life
1:09:15 in a cyberpunk setting so it's
1:09:17 maintaining the coherence of that
1:09:19 character putting it in different
1:09:21 settings but it's it's all in a
1:09:23 cyberpunk
1:09:30 environments three different scenes as
1:09:33 as one generation a 19-second
1:09:35 clip with three different scenes in
1:09:38 it the humans are too big where in this
1:09:42 one or in the previous one yeah the
1:09:44 humans were too big in the previous one
1:09:47 yeah
1:09:50 uh this one's quite good although it
1:09:52 doesn't quite do the the end of this
1:09:54 prompt is he smiles Riley as if he's
1:09:56 understood the secrets of the universe
1:09:58 and that doesn't really happen
1:10:01 but
1:10:03 that's pretty
1:10:12 stunning um that one's janky that one's
1:10:16 janky this one's this one's a little
1:10:19 weird the snow's a little weird but
1:10:21 other than that like this is I could see
1:10:24 this on a Purina dogchow commercial
1:10:27 right you cut to the puppies frolicking
1:10:29 in the
1:10:41 snow that's funny Silver
1:10:44 Fox my current model okay okay the
1:10:47 current model has weaknesses so this is
1:10:49 and and Pate I think this is I think
1:10:52 there's a lesson here for for any
1:10:53 company doing marketing
1:10:55 things so they showed you a bunch of
1:10:57 cool stuff whether or not they
1:10:59 purposefully put in some Jank in in in
1:11:02 some of the collections you know for one
1:11:04 thing it's not just one or two videos
1:11:07 right that like each one of these
1:11:08 sections has eight or nine videos in it
1:11:10 so it's a lot of
1:11:11 content and then this section the
1:11:14 current model has weaknesses it may
1:11:17 struggle to act it it may struggle with
1:11:19 AC accurately simulating the physics of
1:11:22 a complex scene and may not understand
1:11:25 specific instances of cause and effect
1:11:27 for example a person might take a bite
1:11:30 out of a cookie but afterward the cookie
1:11:32 might not have the bite mark the model
1:11:34 also confuses spatial details of a
1:11:36 prompt for example mixing up left and
1:11:38 right and may struggle to pre precisely
1:11:42 uh to struggle with precise descriptions
1:11:45 of events that take place over time like
1:11:48 following a specific camera trajectory
1:11:50 and then they give you examples of it
1:11:52 right so this guy's on a tread Mill and
1:11:54 then the treadmill is like moving like a
1:11:56 car right like it's weird and then they
1:11:59 actually put in here
1:12:01 weakness some uh Sora sometimes creates
1:12:05 physically implausible
1:12:07 motion so they're actually calling out
1:12:10 the weaknesses of the model which I
1:12:11 think I think it's a good call this
1:12:14 one's just weird this is kind of like
1:12:16 those other things look look look at all
1:12:17 the the the the puppies just
1:12:21 spontaneously generating out of one
1:12:23 another
1:12:24 other and it says weakness animals or
1:12:27 people can spontaneously appear
1:12:30 especially in scenes containing multiple
1:12:33 entities right that's just trippy but
1:12:35 they call it out as that's a
1:12:38 weakness um this one's wild so that
1:12:41 bounces that looks good and then it goes
1:12:43 in here and explodes and then boom and
1:12:45 then just another ball shows up right so
1:12:48 prompt is basketball hoop basketball
1:12:51 through hoop then
1:12:52 explodes weakness an example of
1:12:55 inaccurate physics modeling and
1:12:57 unnatural object
1:13:00 morphing right so that one's
1:13:04 weird this one's super trippy so so
1:13:08 there's a there's a thing that's sort of
1:13:11 like a oh no it's a
1:13:16 chair and then they're trying to carry
1:13:18 it now it's just floating on its own oh
1:13:21 my God where it's haunted
1:13:25 the weakness in this example Sora fails
1:13:27 to model the chair as a rigid object
1:13:30 leading to inaccurate physical
1:13:32 interactions you think it's like a
1:13:34 [ __ ]
1:13:37 ghost um oh and then this one she's just
1:13:40 weird she just she goes to blow out the
1:13:42 candles and like is not in the same
1:13:44 neighborhood although I've been at
1:13:46 birthday parties where people could not
1:13:48 figure out how to blow out candles so I
1:13:51 don't think this one's that far off
1:13:53 simulating complex interactions between
1:13:55 objects and multiple characters is often
1:13:58 challenging for the model sometimes
1:14:00 resulting in humorous
1:14:09 Generations n Nana can't blow out the
1:14:12 candles again Jimmy Jimmy could you get
1:14:14 up here and blow out the candles for
1:14:15 Nana n Nana's missing
1:14:19 again shouldn't they wait until they
1:14:21 have a working product gez
1:14:24 no I don't think so the shadow I I don't
1:14:27 think so at
1:14:28 all
1:14:31 um they're putting this out to let us
1:14:34 know what's
1:14:35 coming and they're also saying hey we're
1:14:37 red teaming this right
1:14:39 now I would I would imagine
1:14:43 so so even though
1:14:46 Google's
1:14:47 Ultra launch kind of
1:14:50 fizzled what they launched today with
1:14:53 with 1. five is really impressive with
1:14:55 with Pro 1.5 is really
1:14:57 impressive I think open AI recognizes
1:15:01 that they can't sit on their
1:15:02 Laurels that at some point here someone
1:15:05 is going to knock them off their
1:15:07 perch and so by previewing this it it
1:15:11 keeps them in the capat bird's seat to
1:15:13 say you all have no idea what's coming
1:15:16 right like Sam mman can say it all he
1:15:18 wants that chat gp4 is best understood
1:15:21 as a preview progress here is not Linear
1:15:25 by releasing this we start to get a
1:15:27 sense of what the [ __ ] he's talking
1:15:28 about so um all right so that's that
1:15:33 okay safety we'll be taking several
1:15:35 important safety steps ahead of making
1:15:37 Sora available in open ai's products we
1:15:39 are working with red teamers domain
1:15:42 experts in areas like misinformation
1:15:44 hateful content and bias and we and will
1:15:47 be adversarially testing the model we're
1:15:50 also building tools that help detect
1:15:52 misleading content such which is
1:15:54 detection uh a detection classifier that
1:15:57 can tell when a video was generated by
1:15:58 Sora we plan to include CT
1:16:02 c2p metadata in the future if we depl if
1:16:05 we deploy the model in an open AI
1:16:07 product right so it'll have water
1:16:09 marking in it based on that standard uh
1:16:12 which it links to um in addition for us
1:16:15 developing new te techniques to prepare
1:16:18 for deployment we're leveraging the
1:16:19 existing safety methods that we built
1:16:21 into other products like DOI which
1:16:23 basically is when you try to get DAL to
1:16:25 do something nasty it won't do it right
1:16:27 so they're basically saying we're going
1:16:28 to do the same thing with
1:16:30 video for example once an open AI
1:16:33 product once in an open AI product our
1:16:36 text classifier will check and reject
1:16:39 text input prompts that are in violation
1:16:41 of our usage policies like those that
1:16:44 request extreme violence sexual content
1:16:46 hateful imagery celebrity likeness or
1:16:48 the IP of others we've also developed
1:16:51 robust image class classifiers um that
1:16:55 are used to review the frames of every
1:16:57 video generated to help ensure that it
1:16:59 adheres to their usage policy right so
1:17:02 so they're at least thinking about this
1:17:04 stuff we'll be engaging with policy
1:17:06 makers Educators and artists around the
1:17:08 world to understand their concerns and
1:17:10 identify positive use cases of this new
1:17:12 technology despite extensive research
1:17:15 and testing we cannot predict all of the
1:17:17 beneficial ways people will use our
1:17:19 technology nor all the ways people will
1:17:22 abuse it right and that's true like
1:17:24 that's
1:17:25 why they release chat gp4 is they want
1:17:28 people using it to see where the
1:17:30 boundaries
1:17:31 are that's why we believe that learning
1:17:34 from Real World use is a critical
1:17:36 component to creating and releasing
1:17:37 increasingly safe AI systems over time
1:17:41 this one's weird there's a physics
1:17:43 disaster right there um but there's
1:17:45 other good there's other good examples
1:17:48 here this one's pretty
1:17:52 cute that one's not bad but
1:17:55 boring that one's weird watch The
1:17:57 Seagull that flies in here it's just
1:18:00 bizarre uh physics
1:18:03 nah the cat's weird that's fine but too
1:18:08 dark that's weird that's weird that's
1:18:12 weird this one's pretty
1:18:21 good research techniques
1:18:24 this is geeky but important Sora is a
1:18:26 diffusion model which which generates a
1:18:30 video by starting off with one that
1:18:31 looks like static noise and gradually
1:18:33 transforms it by removing the noise over
1:18:35 many steps right just like diffusion
1:18:37 models create images same same
1:18:40 technology Sora is capable of generating
1:18:43 entire videos all at once or extending
1:18:46 generated videos to make them longer by
1:18:48 giving the model foresight of many
1:18:51 frames at a time we've solved a
1:18:52 challenging problem of making sure a
1:18:56 subject stays in the same stays the same
1:18:58 even if it goes out of view
1:19:01 temporarily similar to the GPT model
1:19:04 Sora uses Transformer architecture
1:19:06 unlocking Superior scaling performances
1:19:09 um we represent videos and images so
1:19:12 this this is this starts to get into
1:19:15 some of the technical stuff that that
1:19:17 you know more technical folks will
1:19:19 understand we represent videos and
1:19:21 images as collections of smaller units
1:19:24 of data called patches Each of which is
1:19:27 akin to a token in GPT by unifying how
1:19:31 we represent how how we represent data
1:19:33 we can train diffusion Transformers on a
1:19:36 wider range of visual data than was
1:19:38 possible before spanning different
1:19:40 durations resolutions and aspect ratios
1:19:43 don't quite know what it means but it
1:19:44 basically sounds like they're taking a
1:19:46 modular approach to how they assemble
1:19:49 this stuff and then it says Sora Builds
1:19:51 on past research in dal and G GPT models
1:19:54 it uses Rec captioning technique from
1:19:56 Dolly 3 which involves generating highly
1:19:58 description captions for the visual
1:20:01 training data so this is where it writes
1:20:03 its own prompts so so basically you're
1:20:06 going to write a prompt it's going to
1:20:07 rewrite it so that it it's confident
1:20:09 it'll do something good for you as a
1:20:11 result the model is able to follow the
1:20:13 user's text instructions in the
1:20:15 generated video more faith Faithfully in
1:20:17 addition to being able to generate a
1:20:19 video solely from text instructions the
1:20:22 model is able to take an existing still
1:20:24 image and generate a video from it we
1:20:25 would expect that animating the im's
1:20:28 contents with accuracy and attention to
1:20:30 small detail the model can also take an
1:20:33 existing video and extend it or fill in
1:20:35 missing frames learn more in our tech
1:20:37 technical paper coming later
1:20:40 today which I haven't seen yet but it's
1:20:42 probably out by
1:20:43 now Sora serve okay the last sentence
1:20:47 here Sora serves as a foundation for
1:20:51 models that can understand and
1:20:54 simulate the real world that sentence
1:20:58 right there Sora serves as a foundation
1:21:01 for models that can understand and
1:21:03 simulate the real world a capability We
1:21:06 Believe will be an important milestone
1:21:07 for achieving
1:21:09 AGI this isn't just about
1:21:12 video video is just the most
1:21:18 simple expression of what they're
1:21:20 talking about in that last sentence
1:21:23 serves as a foundation for models that
1:21:25 can understand and simulate the real
1:21:27 world right which which is the bookend
1:21:30 to the opening
1:21:31 paragraph we're teaching AI to
1:21:34 understand and simulate the physical
1:21:36 world in motion with the goal of
1:21:38 training models that help people solve
1:21:41 problems that require real world real
1:21:44 world
1:21:46 interaction crazy right cucko
1:21:52 nuts wild thoughts
1:21:59 questions she's blowing crumbs off the
1:22:02 table someone has a
1:22:04 [Laughter]
1:22:07 grandma imagine people with a augmented
1:22:09 reality AI generated faces they wear
1:22:13 that AI generate live and alter you will
1:22:16 always be happy I well this is this is
1:22:18 one of the things
1:22:20 that we're getting we're getting into
1:22:23 weird territory here like imagine the
1:22:25 endorphin rush like we get endorphin one
1:22:28 of the one of the reasons these lives
1:22:30 are
1:22:31 fun is if I flip over to the
1:22:36 AI salon and go to the AI Irregulars
1:22:40 Channel I can flip down through these
1:22:43 images look now starring
1:22:45 Sora um there's a TV accepting an Oscar
1:22:49 there's there's a big movie camera
1:22:51 that's called sora like there's there's
1:22:53 a bunch of things in here where people
1:22:57 are in real time you know putting up
1:23:01 images that are in response to this
1:23:03 conversation right there's like this
1:23:05 endorphin rush of like oh I can make
1:23:08 that image I can make that image let me
1:23:09 share it let me connect it let me put it
1:23:11 in the
1:23:13 World Imagine The endorphin rush of
1:23:16 being able to speak a movie into
1:23:21 existence not just like a video clip but
1:23:24 a
1:23:29 story and if you don't
1:23:31 think Google and open Ai and
1:23:37 meta have versions of these models for
1:23:39 all the different media types you're
1:23:41 high on crack right they're going to
1:23:44 have video animation audio
1:23:47 music um character
1:23:51 voices game me mechanics all of that
1:23:54 stuff is going to be a
1:23:58 stream in these in these engines that we
1:24:01 get to play with so you'll ask for
1:24:04 something and right now they're all
1:24:06 these discrete tools like there was no
1:24:08 audio for any of that video we looked
1:24:11 at well when they when they release the
1:24:15 the fully true multimodal
1:24:18 engine you're not just going to have
1:24:20 video you're going to have video with a
1:24:22 soundtrack
1:24:24 and sound effects and dialogue and
1:24:32 characters and then it's going to
1:24:35 be 60 seconds and then it's going to be
1:24:38 3 minutes and then it's going to be 10
1:24:40 minutes and then it's going to be speak
1:24:42 a feature film into
1:24:49 existence and it do you remember
1:24:52 remember the um remember if you're if
1:24:54 you're old enough there's there's a
1:24:55 bunch of you
1:24:57 know old and timey types in here
1:24:59 remember the the early days well not I
1:25:02 guess remember the middle days of the CD
1:25:04 ROM
1:25:05 era where there were a lot of these CD
1:25:08 ROMs that had that had multiple Pathways
1:25:10 right there would Choose Your Own
1:25:11 Adventure where you'd get to a certain
1:25:14 point in the
1:25:17 movie then you could choose a
1:25:20 different prescripted pre rendered path
1:25:26 right well imagine saying I want to
1:25:28 watch a
1:25:29 romcom that's got these kind of
1:25:31 characters in it and one of them
1:25:33 collects Turtles and one of them's a
1:25:34 llama
1:25:37 farmer and they start doing things and
1:25:39 you're watching it you're like I'm kind
1:25:40 of bored you just go yeah let's let's
1:25:42 have this take a a darker turn let's
1:25:45 have it let's have this turn into a a
1:25:50 thriller and then it just
1:25:56 shifts
1:25:58 like we are not like a decade away from
1:26:02 that we're like 18 months to three years
1:26:06 away from
1:26:11 that they going to be perfect no are
1:26:13 they going to be janky as [ __ ]
1:26:17 yeah but did you ever watch a Roger
1:26:20 Corman slasher flick
1:26:24 that he'd make for
1:26:27 [Laughter]
1:26:28 $20,000 were they perfect no were some
1:26:32 of them hits
1:26:33 yes it's going to be like that when I
1:26:35 remember my dreams after I wake up I can
1:26:38 create a video of it yeah Genie n Rob
1:26:41 jeie have you seen my go uh look up uh
1:26:45 Dream Weaver on gpts I created a a dream
1:26:49 interpreting um GPT that will ask you a
1:26:53 few questions about your dream it'll
1:26:55 then generate a visualization a single
1:26:58 visualization and and then it gives you
1:27:00 um a uh a yian analysis of your dream
1:27:06 it's called Dream Weaver in the
1:27:08 gpts but yeah like the the future
1:27:11 version of that is yeah your dream comes
1:27:13 to
1:27:15 life you get to recreate the visuals and
1:27:18 maybe the audio and the mood of
1:27:21 it
1:27:27 Bonkers I like to put my image in the
1:27:29 movie yep we'll be able to do that
1:27:31 imagine when AI can monitor our dreams
1:27:33 and then realize them in video back to
1:27:35 us yeah in the morning you wake up
1:27:37 you're like oh yeah that was it o
1:27:39 [Laughter]
1:27:43 delete the cynics have entered the live
1:27:45 that's all right I don't mind
1:27:48 cynic it's
1:27:50 listen this stuff is scary scary I get
1:27:54 it it's absolutely scary and listen you
1:27:57 can be
1:27:58 cynical I would just strongly
1:28:03 encourage don't be cynical from the
1:28:07 sidelines be cynical while you're using
1:28:10 the tools my experience has been you'll
1:28:13 probably get less
1:28:17 cynical you'll see how [ __ ] powerful
1:28:19 they are and and quite honestly how
1:28:22 inspiring they
1:28:25 are because these tools are not yet at
1:28:29 the point where you just push a button
1:28:31 and good [ __ ] comes
1:28:33 out you put in a prompt and something
1:28:35 shitty comes out and you're like a
1:28:36 that's [ __ ] awful and if that's where
1:28:38 you stop you're absolutely robbing
1:28:43 yourself of what's
1:28:46 possible the art right now with
1:28:49 exploring all these tools is figuring
1:28:51 out where the boundaries are
1:28:54 and if you're a creative person and you
1:28:56 can navigate those you can do some
1:28:57 remarkable [ __ ] like these tools
1:29:00 amplify your vision they amplify your
1:29:03 ideas they amplify your
1:29:06 capabilities but you have to be willing
1:29:08 to embrace the imperfection embrace the
1:29:14 janky and then if in doing all
1:29:17 that you're still like he [ __ ] it all
1:29:20 right at least you tried it but if
1:29:22 you're on the sidelines going these
1:29:24 things
1:29:33 are I feel
1:29:36 bad it's why I do this
1:29:38 channel I don't want people sitting on
1:29:41 the
1:29:46 sidelines cuz here's the deal with these
1:29:49 tools they're not going away you can be
1:29:52 pissed off you can be cynical you can be
1:29:57 fatalistic you can say humans are better
1:30:00 than the
1:30:05 machines and they're still
1:30:11 coming it's like being pissed off at the
1:30:13 printing press or the steam engine or
1:30:15 the personal computer or the worldwide
1:30:20 web College professors
1:30:24 picketed when the calculator was
1:30:30 invented did it stop did does anyone
1:30:33 even know that that happened
1:30:36 no did it stop it
1:30:41 no these are the most remarkable
1:30:45 tools in the history of
1:30:48 humanity and we get to play with
1:30:51 them
1:30:54 we get to play with them we're lucky
1:30:57 enough to be on the
1:31:00 planet when someone like Sam
1:31:04 Alman has enough of a vision how to take
1:31:06 something like a
1:31:08 Transformer and jam enough [ __ ] data
1:31:10 in
1:31:13 it that
1:31:15 it spits out crazy cool
1:31:21 [ __ ]
1:31:23 and it's cool enough that a 100 million
1:31:25 people used it in the first six weeks of
1:31:27 its existence with no
1:31:29 advertising Word of Mouth
1:31:34 only that's why all the other companies
1:31:38 are doing all this AI
1:31:40 [ __ ] because of that November 30th
1:31:45 2022 we're like a year and change into
1:31:48 whatever this thing is becoming it's not
1:31:51 going away
1:31:53 and we we're not even we're not even
1:31:56 we're not even at the tip of the iceberg
1:31:57 like we're at
1:31:59 the like at the it's like this is the
1:32:02 tip of the iceberg we're like like way
1:32:04 up here like we don't even know what the
1:32:06 [ __ ] tip of the iceberg is right
1:32:08 now this Tech will allow me to show show
1:32:11 the world a better way money based on
1:32:13 each other's infinite worth yeah it we I
1:32:16 think we move into an incredibly
1:32:18 prosperous
1:32:20 time an incredibly prosperous time but
1:32:23 to get to that we're going to have to go
1:32:25 through some [ __ ] changes and what a
1:32:28 human's hate
1:32:31 change all right we're all going to have
1:32:34 links to our social media and our own
1:32:35 little movies absolutely once upon a
1:32:38 time you will have an AR VR version of
1:32:41 yourself we talked about that trippy
1:32:42 music videos Lori Ruskin I'm already on
1:32:44 the other side of it now it's 4:23 a.m.
1:32:48 yes I remember yeah missed yeah missed
1:32:52 exactly Mist was one of those Do Your
1:32:55 Own Adventure in the end what you
1:32:57 realized with mist though was that there
1:33:01 were like four sections with like what
1:33:05 five things each you could do in each of
1:33:06 four sections like when you get done
1:33:08 with it you're like oh like you could
1:33:11 map it on a wall and it would be a
1:33:13 finite very very finite world with
1:33:15 finite
1:33:18 things imagine playing
1:33:21 mist
1:33:22 that is just as compelling or more
1:33:24 compelling because it'll be completely
1:33:27 immersive but it's
1:33:30 infinite you can explore
1:33:34 everywhere and every time you try a
1:33:36 puzzle it will be
1:33:38 different AI can help us generate
1:33:41 creativity not not replace it that's
1:33:43 that's exactly the point Michael Janette
1:33:46 is when I first started using this stuff
1:33:49 where my brain immediately went is where
1:33:51 I where I've always gone with computers
1:33:53 which is oh these tools will make me
1:33:57 more
1:33:59 efficient these tools will make me more
1:34:03 productive that's a very 19th century
1:34:06 Factory floor kind of mentality isn't it
1:34:10 because what have computers served
1:34:11 served as for us over the 20th century
1:34:15 and into this
1:34:16 Century a more efficient way to do the
1:34:20 Cog the the shitty work of
1:34:23 business but it's still on us to do the
1:34:26 shitty [ __ ]
1:34:28 work and what I've learned as
1:34:31 I've dug in deep to this AI
1:34:35 [ __ ] is
1:34:37 that it it does the shitty work for me
1:34:42 and lets me live at a creatively higher
1:34:45 plane for more time and here's what I
1:34:48 mean by that in the olden timey days
1:34:52 like October
1:34:55 2022 in the olden timey days if I had an
1:34:59 idea I'd have to go okay I got an idea
1:35:02 all right what do I got to do to bring
1:35:04 it to life well I should probably write
1:35:06 an
1:35:07 outline oh then I start writing the
1:35:09 outline and I'm like wait do I use
1:35:10 Microsoft Word or did you get one of
1:35:12 those fancy outlining things or do I do
1:35:13 I want a flowcharting thing right like
1:35:15 you spend all this time just just
1:35:17 [ __ ] trying to figure out even how
1:35:19 you're going to think about thinking
1:35:21 about your
1:35:22 idea and then you're like oh I got to
1:35:24 actually go research something you go to
1:35:26 Google and like 45 minutes later you're
1:35:27 down some [ __ ] rabbit hole and
1:35:29 there's all these just [ __ ] shitty
1:35:31 sites and now you don't even know what
1:35:33 the idea was anymore and you're trying
1:35:35 to [ __ ] claw your way back up to the
1:35:38 idea and two or three hours later you
1:35:41 maybe have the idea articulated to some
1:35:46 level of completeness that you can stand
1:35:47 back and look at it and go ah that was a
1:35:50 shitty idea or yeah that's a good idea
1:35:53 and then you go further right but that
1:35:55 was like two to three hours of like
1:35:58 Creative
1:36:00 Energy just getting the idea out of your
1:36:04 head and what I'm finding now is I have
1:36:08 that same shitty idea and I go oh okay I
1:36:12 I got to create an outline and this I'm
1:36:14 like oh no I
1:36:16 don't dear chat
1:36:18 GPT could you see if this is a shitty
1:36:21 idea or not
1:36:23 Ah that's not bad like in 30
1:36:27 seconds it's
1:36:29 articulated and then I could go oh give
1:36:32 me three more ideas like that boom
1:36:34 there's three more ideas oh that's kind
1:36:36 of cool oh if I took that idea and that
1:36:38 idea and I put them together what if I
1:36:39 put these two ideas
1:36:41 together huh what would that look like
1:36:44 as a um Transformer
1:36:48 machine up comes an image oh that's
1:36:51 that's pretty good could you make it
1:36:53 more steampunk boom there it
1:36:55 is it's like bang bang bang bang Idea
1:36:59 Idea Idea
1:37:01 Idea if you're not [ __ ] playing with
1:37:03 these tools you're missing that
1:37:12 gift AI literally saved my life I found
1:37:15 a good reason to live and kept on
1:37:17 learning oh that's amazing Cosmic lever
1:37:19 that's awesome yeah
1:37:24 it's I am watching on a daily basis and
1:37:28 nightly
1:37:30 basis people in this
1:37:32 community fullon Reinventing themselves
1:37:36 rediscovering
1:37:38 themselves
1:37:40 rediscovering the
1:37:42 joy of their
1:37:45 5-year-old
1:37:48 selves remember when you were five
1:37:50 before the world knocked the [ __ ] out
1:37:51 out of you you're like I could do this
1:37:53 and I could do that and I could do that
1:37:55 and then the world's like [ __ ] you you
1:37:56 can't do that you're not talented and
1:37:58 you don't have this and you got shitty
1:37:59 grades and eat your [ __ ] baloney
1:38:02 sandwich of little [ __ ] [ __ ] you're
1:38:04 like yeah [ __ ] I didn't realize the
1:38:06 world was this
1:38:08 scary and so you shut all that [ __ ] down
1:38:11 right okay I'll be an
1:38:19 accountant and then this AI stuff comes
1:38:21 along and you're like okay I'm going to
1:38:23 do accounting work I'm going to be more
1:38:25 efficient with accounting and then what
1:38:27 it can make
1:38:28 pictures all right well make a picture
1:38:31 of a calculator oh that's that's pretty
1:38:35 clever you know I could probably make a
1:38:38 a picture of a of a Pegasus too could it
1:38:41 wouldn't do a pegasus would it I could
1:38:43 understand how it did a calculator it's
1:38:45 it's simple lines it just draws it that
1:38:48 make a make a pegasus oh God I could I
1:38:52 can make
1:38:53 Pegasus you know I had that idea for
1:38:56 that game could it helped me make a game
1:38:59 and then all of a
1:39:00 sudden that person that was shut down is
1:39:03 rediscovering that little kid inside
1:39:05 them was like
1:39:07 oh I can make [ __ ] again I Can
1:39:11 Dream I can take these Dusty
1:39:15 ideas lunar stick in here Daisy talks
1:39:18 about she had she had boxes of index
1:39:21 cards of
1:39:23 ideas she did standup comedy she did
1:39:25 politics she she captures her
1:39:29 ideas and AI came along she's like
1:39:33 oh not only do I have ideas here I have
1:39:37 boxes of ideas pull out an index card
1:39:40 put it in AI oh there's a GPT she's made
1:39:43 105 or 110 gpts at this
1:39:48 point she's a software developer now
1:39:52 did she learn to code
1:39:55 no are some of those gpts that she
1:39:57 created like giving her incredible joy
1:40:00 and giving incredible joy to other
1:40:02 people that are using them
1:40:05 yes thank you power thank you very
1:40:10 much I mean be cynical but holy [ __ ] are
1:40:13 you missing out on some fun I also
1:40:16 compare AI to gutenberg's printing press
1:40:18 it's it is it's it's on that level like
1:40:21 this
1:40:22 is so I was lucky enough to be an
1:40:25 entrepreneur in the early days of the
1:40:27 worldwide web I started one of the first
1:40:28 digital agencies in
1:40:30 1994 company called agency.com
1:40:33 it was [ __ ] remarkable we grew to
1:40:36 2200 people in 5 years it was just it
1:40:39 was absolute [ __ ] Insanity we built a
1:40:41 lot of the websites for the Fortune
1:40:46 500 this AI
1:40:50 [ __ ]
1:40:52 is a thousand times more profound than
1:40:54 the worldwide web
1:40:57 was and nobody knows
1:41:01 it like the amount of people who've paid
1:41:05 for chat gp4 and use it on a regular
1:41:09 basis is
1:41:11 like four five maybe
1:41:14 6% probably in the
1:41:19 US 95% of the people don't have a
1:41:23 [ __ ] clue what's coming don't have a
1:41:25 clue what's
1:41:31 coming and it's coming so fast it took
1:41:34 the worldwide web to get to 100 million
1:41:37 users took six years there was no
1:41:40 infrastructure nobody had internet
1:41:42 access there was like three HTML tags
1:41:45 you couldn't even Center images on a
1:41:47 page when I started took six years to
1:41:50 get to 100 million users it took 6 years
1:41:52 for just that janky ass
1:41:55 [ __ ] to turn into
1:42:00 something chat GPT took six weeks to get
1:42:04 to 100 million users the infrastructure
1:42:07 is all there the processors are all
1:42:09 there the server Farms of tens of
1:42:12 thousands of gpus to run the inferences
1:42:15 to do this [ __ ] are all
1:42:18 there the Transformer architect that
1:42:21 Google invented in
1:42:24 20177 has been being developed on for 7
1:42:30 years chat GPT is just a simple
1:42:33 interface on top of all this work that's
1:42:36 been going on for two three four five
1:42:40 decades and chat GPT similar to the
1:42:43 worldwide web is a Simplicity layer
1:42:47 sitting on top of complex
1:42:50 systems
1:42:52 so it's just like the little Google
1:42:53 search box you just type in words and
1:42:55 pring out comes
1:42:58 [ __ ] and as janky as it is and as [ __ ]
1:43:01 up as it
1:43:03 is it's already changing people's lives
1:43:08 the worldwide
1:43:11 web people could
1:43:14 publish that they couldn't before it
1:43:16 changed some
1:43:17 lives but not like this level of
1:43:20 profound
1:43:26 so yeah so the
1:43:29 cynicism well I guess that's the deal
1:43:31 with
1:43:32 cynicism you end up spiting yourself in
1:43:35 the
1:43:36 end and I really hope you don't I hope
1:43:39 you try it play don't take this [ __ ] so
1:43:44 seriously I've made a few llms with IQs
1:43:48 around 60 or
1:43:50 less
1:43:52 uh uh uh I disagree a little awesome
1:43:55 fantastic without the internet AI most
1:43:57 likely wouldn't
1:43:58 exist oh don't get me wrong Brandon I
1:44:02 don't I'm not saying that the internet
1:44:04 wasn't important I'm saying that
1:44:08 the the the the the impact that it
1:44:13 had took a long time to
1:44:16 formulate and and relatively speak
1:44:19 speaking was very very simple what the
1:44:21 worldwide web did was it allowed you to
1:44:24 connect networked computers with the
1:44:27 click of a hyperlink right that that's
1:44:29 the that's the
1:44:31 Innovation and all of the [ __ ] that
1:44:33 we've seen you know from the mid90s
1:44:37 until
1:44:39 2020 mobile phone and social media and
1:44:43 websites and e-commerce and Global
1:44:45 Communications all that stuff was made
1:44:49 possible because of this
1:44:51 and yeah that's absolutely the
1:44:53 foundation for this AI stuff what I'm
1:44:56 saying about the AI stuff is the tools
1:44:58 themselves are profoundly more
1:45:01 powerful it's not just the Simplicity of
1:45:03 clicking a link right it's we just
1:45:06 looked at it it's make me a movie write
1:45:10 me a story do this data analysis and
1:45:14 it's doing the
1:45:16 work it's not a Google search engine
1:45:19 that's like I want to learn pip Pyon and
1:45:21 it and it connects me with a YouTube
1:45:24 video that teaches me how to write
1:45:26 python it's writing the [ __ ]
1:45:33 python it's a fundamentally different
1:45:37 thing the internet just allowed us to be
1:45:40 more efficient at the [ __ ] we were
1:45:42 already doing this AI shits doing the
1:45:46 work we were
1:45:48 doing it's a different era it's a
1:45:51 different
1:45:52 era so it's much more like like I go
1:45:56 fire like fire
1:46:00 language written
1:46:02 language whe wheel was somewhere in
1:46:05 there probably wheel was before language
1:46:06 right or I don't know somewhere wheel
1:46:09 wheel written
1:46:13 language printing
1:46:16 press steam engine like like there
1:46:19 aren't many of these right this AI
1:46:22 shit's one of them I think the worldwide
1:46:25 web is a subcategory of digital
1:46:29 Communications right
1:46:33 computers this one's a big one this
1:46:35 one's one of the
1:46:37 biggies it's going to fundamentally
1:46:39 alter how we Define
1:46:42 work what we value in work what our
1:46:46 contributions are in
1:46:49 work
1:46:52 you you you may hear this phrase
1:46:55 AGI artificial general intelligence and
1:46:57 they batted around I we're going to do
1:46:59 AGI we're going to do AGI we're going to
1:47:00 do you know what that means when we get
1:47:01 to
1:47:02 AGI that these machines are as good as
1:47:05 people at most
1:47:09 jobs not like can add some [ __ ] can do
1:47:13 your entire
1:47:19 job
1:47:21 two block Tom Kyle Your Enthusiasm is
1:47:23 contagious as [ __ ] thanks I get fired up
1:47:26 like when I see stuff like this like the
1:47:29 the the the the [ __ ] voice of like
1:47:32 yeah it's not perfect
1:47:35 and take a look at what's there don't
1:47:38 sit on the [ __ ]
1:47:43 Sidelines try
1:47:48 it hey Mikey he likes
1:47:53 it remember that
1:47:55 commercial is anyone honestly still
1:47:57 working well that's a whole another
1:47:59 question AI will reach human Levels by
1:48:02 around 2029 Ray Kurtz um go listen to
1:48:05 David Shapiro he thinks we're going to
1:48:07 hit some clear definable version of AGI
1:48:10 by the end of this year so sometime
1:48:12 between 2024 and
1:48:15 2029 and kurts was 2030 for the longest
1:48:18 time he's he's accelerated it a year
1:48:20 kurts accelerated a year go go watch
1:48:22 videos uh from David Shapiro on YouTube
1:48:25 about AGI he's written five books on it
1:48:28 he's got a cognitive architecture
1:48:30 project that is open sourced he thinks
1:48:33 we're essentially
1:48:36 there I'm going to go
1:48:39 with not this year but this year is
1:48:42 going to get really [ __ ] weird it's
1:48:44 going to get weird and I think Sora is
1:48:47 the first the first the first Domino to
1:48:49 fall um of of at least seeing something
1:48:52 that's like oh
1:48:55 [ __ ] I'll tell you so I've been on this
1:48:59 here Tik Tock [ __ ] for
1:49:02 uh year and three
1:49:06 months and I got 23y old twin boys
1:49:09 they're right up there they could give
1:49:12 two shits about AI they they don't care
1:49:15 they're not interested like yeah
1:49:16 whatever dad you should check this out
1:49:19 chat GPT is cool yeah whatever dad you
1:49:21 should see gpts yeah whatever dad I
1:49:24 showed him Sora tonight that first video
1:49:27 of the of the woman walking through
1:49:29 Japan Rainy Streets and I said look what
1:49:32 this thing can do both of them
1:49:35 said
1:49:39 oh oh
1:49:41 that's huh they've never done
1:49:47 that some shit's coming I I hope you'll
1:49:50 be live streaming the birth of AGI
1:49:53 [ __ ] right I
1:49:57 will certainly if it comes this
1:50:04 year is this a different David Shapiro
1:50:07 it's uh he's uh at Dave shap on YouTube
1:50:12 it's on if you go to the AI salon and go
1:50:15 to AI tools and resources scroll to the
1:50:18 bottom and David Shiro's there let me
1:50:21 see if the link is still
1:50:24 right that page isn't available he
1:50:26 changed his fraking name
1:50:29 again hang
1:50:31 on he did this once before make up your
1:50:34 mind
1:50:39 Dave
1:50:41 David
1:50:45 Shapiro David Shapiro what did he change
1:50:48 his name to at Dave shop that's right
1:50:50 why is it [ __ ] up hang
1:50:53 on hang on people we're going to fix
1:50:56 this nonsense right
1:51:00 now let me edit
1:51:03 this
1:51:10 edit edit
1:51:13 oh wait
1:51:19 URL
1:51:22 update I had a bad URL bad URL in there
1:51:27 that was that
1:51:31 was AI learn that was that was AI
1:51:34 learning lab user
1:51:35 error dum dum didn't have
1:51:39 a let's see that work there we go all
1:51:42 right go to uh go to the AI Salon tools
1:51:46 and resources and go go watch it and
1:51:48 then look up his videos on AGI
1:51:51 post AGI economics AGI within 18 months
1:51:55 yeah watch his video on AGI in 18 months
1:51:57 Dave's latest video was interesting
1:52:00 which one
1:52:02 99.97%
1:52:05 cheaper yeah sometimes his his things
1:52:07 are out there and a little weird but he
1:52:09 is he is incredibly well educated on
1:52:14 both the technology side but also the
1:52:16 economic impact
1:52:18 side when we see an AG I talk just watch
1:52:22 why did you choose Tik Tok over
1:52:24 Instagram um so I had an in well I mean
1:52:28 I still have I have Kyle Shannon on
1:52:32 Instagram and Kyle Shannon dreams so I
1:52:33 had an art project on
1:52:37 Instagram and then I decided to
1:52:41 do to turn my photos into
1:52:46 videos and I just found Tik Tock like I
1:52:49 just started on Tik Tok I just tried
1:52:52 it and then it made it really easy to
1:52:54 bring those things to life Kyle Kyle KY
1:52:57 can you encourage the nice people to
1:52:58 register for the weekend GPT Jam I can
1:53:01 of course an Murphy if you don't know
1:53:03 Ann Murphy you should team up with an
1:53:06 Murphy go to the AI
1:53:09 salon and if you don't know what the
1:53:12 [ __ ] I'm talking about with the AI Salon
1:53:13 oh there there's a lot of new people in
1:53:15 here so you don't go
1:53:17 here this community
1:53:21 the AI Salon is a group that I started
1:53:23 with Leah faston who's a photographer
1:53:25 out of Boston we started this the week
1:53:28 after chat
1:53:29 GPT we just moved the community to a
1:53:32 web-based platform like three months ago
1:53:35 and that's that this will take you there
1:53:37 so the salon. is going to take you to a
1:53:40 link
1:53:41 tree that looks like this and the first
1:53:45 link is to the AI Salon community and
1:53:47 when you go to the AI Salon
1:53:49 community
1:53:51 it's going to drop you
1:53:52 here and go read about what we're up to
1:53:55 a community of extraordinary adventurers
1:53:58 exploring the AI unknown the AI Salon is
1:54:02 a ragtag group of Misfits Nero Wells and
1:54:04 miscreant who have gathered to explore
1:54:06 the Uncharted Wilderness of generative
1:54:08 AI we are everyone from entrepreneurs to
1:54:11 Educators from healthcare workers to
1:54:12 creatives and everyone in between and it
1:54:14 goes on and on and not on and on cuz I
1:54:17 have ADD I can't read that much [ __ ]
1:54:18 it's short but go read about who we are
1:54:20 what we're about and then check out our
1:54:23 values
1:54:24 curiosity like be curious don't be
1:54:26 cynical be like huh how do you figure
1:54:28 that [ __ ] out and one ways you figure
1:54:31 that out is you go to Ann's GPT jams
1:54:33 which I'm going to talk about in a
1:54:34 second generosity share what you're
1:54:38 learning an started learning learning
1:54:40 this stuff got really excited about it
1:54:43 started a company and set a mission for
1:54:44 herself to
1:54:46 train more than 1,500 people in
1:54:49 nonprofit to learn about AI in I I guess
1:54:53 in a year I don't know what her goal was
1:54:54 but she accomplished it like in no time
1:54:57 cuz she's just
1:54:59 remarkable so
1:55:02 generosity be
1:55:04 generous share what you know talk to
1:55:06 people answer questions exploration
1:55:09 people in in here on an adventure
1:55:12 they're willing to explore they're
1:55:13 willing to go out there collaboration
1:55:15 inclusion people are coming from all
1:55:17 over the place welcome them in their
1:55:19 point of view will teach you [ __ ] you
1:55:22 never knew was possible I see it every
1:55:24 [ __ ] day
1:55:26 empathy even the people that are in here
1:55:29 that are cynical and pissed off and like
1:55:30 Kyle you're just a
1:55:33 Optimist I have deep deep empathy for
1:55:37 them cuz this shit's scary it's weird
1:55:41 it's moving fast people are feeling out
1:55:43 of control they're feeling like
1:55:44 existential [ __ ]
1:55:49 crisis understand that understand that
1:55:52 everyone's struggling to get this [ __ ]
1:55:55 together and then the last one is
1:55:57 gratitude like acknowledge one another
1:55:59 so anyway that's the value so go check
1:56:01 that [ __ ] out then if you like those
1:56:03 join us sign up for the you know get an
1:56:07 account and the first place I would
1:56:10 strongly encourage you to do is go to
1:56:12 the community section and introduce
1:56:15 yourself you're like but c c c that's
1:56:20 scary people are going to judge me
1:56:22 people always judge me you know k k
1:56:26 people say I'm a dumb Dum I don't want
1:56:28 them to say I'm a dumb dumb I don't know
1:56:30 anything about AI I'm so far behind k
1:56:33 k this community won't do that that's
1:56:36 not what this is
1:56:38 about everyone in this community
1:56:41 actually one of my one one of my
1:56:43 missions lately is to try to get people
1:56:45 in this community to stop [ __ ]
1:56:47 apologizing for how non-expert they are
1:56:50 everyone right now is feeling like they
1:56:52 don't have a [ __ ]
1:56:54 clue if you're listening to me Babble on
1:56:58 tick tock on a Thursday night at [ __ ]
1:57:02 midnight in New York
1:57:05 City you're farther ahead than most
1:57:08 people on the planet with this AI stuff
1:57:10 so introduce yourself and let people in
1:57:13 this community welcome you in and kind
1:57:15 of steer you
1:57:17 around and then you know hang out in the
1:57:19 water cooler see what people are talking
1:57:21 about go to showand tell and just share
1:57:22 some of your images you made you know
1:57:24 there's news if you've heard interesting
1:57:26 news put it up there AI tools and
1:57:28 resources that was that thing I just
1:57:29 fixed go there and fix that up
1:57:32 okay and then there's guilds there's a
1:57:34 bunch of different guilds which are
1:57:35 subcommunities
1:57:37 led by remarkable people in our
1:57:39 community and then down below that are
1:57:42 clubs and one of the clubs is Ann's and
1:57:45 Murphy's weekend GPT jams and so go
1:57:49 click on
1:57:52 that and go register for her GPT Jam
1:57:56 this Saturday she does these every
1:57:58 Saturday for two
1:58:02 hours and she brings in people to talk
1:58:05 people to share what they're doing this
1:58:07 weekend it's a it's a it's an innovation
1:58:09 lab where people Corey Sandler's going
1:58:11 to be in there talking about how she
1:58:13 does what she
1:58:14 does you can go learn how to do
1:58:16 gpts 3 or 4 weeks ago um
1:58:20 um Scott I forget his last name Fletcher
1:58:23 Fleming something like that I don't know
1:58:25 Scott came in and taught people how to
1:58:27 connect their gpts to
1:58:30 zapier and it was a [ __ ] train wreck
1:58:32 cuz it's so janky everyone no it was
1:58:35 just it was a it was a fraking mess and
1:58:37 he had the patience of a
1:58:39 saint and people who have never
1:58:42 programmed
1:58:44 before successfully connected their gpts
1:58:48 to an API to Google Google
1:58:51 services and and got their GPT to like
1:58:53 send a [ __ ]
1:58:56 email was
1:59:00 incredible so go sign up for that
1:59:04 [ __ ] all
1:59:06 right and if nothing else you'll meet
1:59:09 some really cool people they're really
1:59:11 nice imagine a community where people
1:59:15 weren't [ __ ] to one
1:59:18 another
1:59:20 that's kind of what this
1:59:21 group has evolved
1:59:24 into and it's one of the reasons I talk
1:59:27 about the values if you're like why do
1:59:28 you talk about the values that's kind of
1:59:30 isn't that kind of soft and snowflakey
1:59:32 what is it some some some sort of woke
1:59:36 situation
1:59:38 well if you want to be surrounded by
1:59:40 people that don't treat you like a piece
1:59:42 of
1:59:42 [Laughter]
1:59:46 [ __ ] then then yeah maybe you're woke
1:59:48 maybe this is a woke group group but I
1:59:50 it's just like what what if what if what
1:59:55 if while we're trying to figure this
1:59:57 [ __ ]
1:59:58 out while we're feeling clueless and out
2:00:01 of control and very far behind and like
2:00:04 we don't know if we're going to have a
2:00:05 job next year cuz we're a copywriter and
2:00:08 holy [ __ ] what's happening and oh my God
2:00:10 I'm a cinematographer and I just looked
2:00:12 at that open AI [ __ ] and now it's
2:00:15 getting [ __ ]
2:00:16 real What If instead of having to do
2:00:18 that on your own you come into a
2:00:21 community that weren't
2:00:23 [ __ ] and we're
2:00:26 supportive and we're willing to be
2:00:28 vulnerable enough to say yeah I don't
2:00:32 know what's going on either I'm scared
2:00:35 too but you know what I figured out this
2:00:37 week I figured out how to make a GPT
2:00:39 what's that oh you don't know what a GPT
2:00:41 is want to see something cool
2:00:45 yeah imagine a community like that
2:00:47 that's what this is
2:00:50 not what it wants to be what it is why I
2:00:52 talk about these values all the time
2:00:55 is these values were derived for how the
2:00:58 people are behaving in this group not
2:01:02 the other way around it's not like we
2:01:04 want to be like this it's like this is
2:01:07 who we are and I want to keep it like
2:01:09 this as we
2:01:11 grow so if you look at these values and
2:01:13 they feel like foreign territory to you
2:01:16 and they turn your stomach this isn't
2:01:18 the people for you
2:01:20 if you look at these values and you're
2:01:21 like holy [ __ ] that would be cool if
2:01:23 that
2:01:24 existed guess
2:01:25 what it
2:01:29 exists all right so if you dig that join
2:01:33 step one introduce yourself go introduce
2:01:36 yourself let's see
2:01:40 last introduction
2:01:43 Kayla gager gag gager gagnier introduced
2:01:48 your nine minutes ago Kudos
2:01:51 yay and then just scroll right down to
2:01:53 Anns GPT jams and then also the other
2:01:56 thing you should do go look at I think
2:01:59 you can see it both in Leah's Art Guild
2:02:02 should be right at the top
2:02:09 Here sign up for the club yeah it's
2:02:12 right at the top there it's also in
2:02:14 James writer
2:02:17 Circle and if you go to salon
2:02:23 announcements the first announcement is
2:02:26 about tomorrow night Leah and Jane So
2:02:29 Leah is the co-founder of the salon with
2:02:31 me and Jane runs the writer Guild
2:02:33 they're doing our first official Guild
2:02:36 mashup and they're putting together a
2:02:39 six week daily course to
2:02:43 to put you on a a creativity Journey
2:02:47 with prompting an image generation
2:02:49 that's gonna blow your flipping M blow
2:02:52 your flipping
2:02:55 mind the daily crunch go Kyle go yeah
2:02:59 damn
2:03:04 it I'm telling you
2:03:07 man it was funny I just looked over at
2:03:10 at the mod channel so the the mods that
2:03:13 that help me make sense of what's going
2:03:15 on in here there's one little comment in
2:03:16 there that said what does this say the
2:03:19 cynics have entered the
2:03:23 live
2:03:25 welcome listen if you're a cynic and
2:03:28 you're bitter about this
2:03:30 [ __ ] but you're still
2:03:34 here you're
2:03:38 curious maybe just tonight let the guard
2:03:41 down a little
2:03:43 bit you can come back to your cynicism I
2:03:46 promise you take your
2:03:48 cynicism like you can make money with J
2:03:52 take it like a little thing and just set
2:03:54 it over there and just go play with AI
2:03:57 tonight go to chat. open.com if you
2:04:00 haven't or if you have like dig deeper
2:04:04 tonight right
2:04:06 now if you're still
2:04:12 here and just do what I call parlor
2:04:15 tricks have it write you a song have it
2:04:18 um
2:04:19 create a meal plan for your family um
2:04:24 the focus focuses primarily on gummy
2:04:26 bears as a main ingredient like just
2:04:29 like really try to have chat GPT do
2:04:35 anything have it write code for you but
2:04:37 I don't code do it anyway and then ask
2:04:41 it how you run the code it just wrote
2:04:43 for you and it'll tell you that that was
2:04:46 my big Epiphany with chat GPT
2:04:49 was I was sitting like like a week and a
2:04:52 half after it came out I had heard that
2:04:54 it can do coding and I haven't coded in
2:04:56 for a [ __ ] ever and I've never really
2:04:58 coded all that well cuz I just did HTML
2:05:00 when there were like six HTML
2:05:03 tags and so I'm sitting on the wife my
2:05:06 my or I'm sitting on the wife I'm
2:05:08 sitting on the couch my wife's watching
2:05:11 TV and and I say oh chat GPT make me
2:05:14 some python
2:05:16 code and up came some python code I was
2:05:19 like ooo that's really cool looks pretty
2:05:21 it looks like I've seen people code
2:05:23 python that's what it looks like and
2:05:26 then I realized oh [ __ ] I don't know how
2:05:28 to run python
2:05:30 code and it was funny what immediately
2:05:33 hit me when I said I don't know how to
2:05:34 run python code was like this Panic of
2:05:37 like oh [ __ ] I'm clueless I would have
2:05:40 to call someone and that would be
2:05:41 embarrassing and I'd have to admit that
2:05:43 you know I own a tech company but I
2:05:44 don't know how to run python code guess
2:05:47 what I didn't have to do that hey Chad
2:05:49 gbt I don't know how to run python code
2:05:52 not a problem just open your terminal
2:05:54 put in this command
2:05:56 line and I was like you know what I
2:05:59 don't want to do command line [ __ ] I
2:06:01 hate command line [ __ ] I'm a Mac guy
2:06:03 because I don't like command line [ __ ] I
2:06:04 want to point and [ __ ]
2:06:06 click and so I said can I do this on a
2:06:09 website can I just copy and paste the
2:06:10 code to some website and have it do it
2:06:13 and it saids sure here's repet hang on
2:06:16 doggy's doggy's wanting to pee I think
2:06:20 come on how you
2:06:35 go
2:06:37 so I copy and paste my code into
2:06:40 repet and it gives me an error and again
2:06:44 I have that [ __ ] Panic of like Shame
2:06:47 Shame you're a bad you're a bad man you
2:06:50 know nothing about technology they will
2:06:52 judge you they will judge
2:06:54 you I said give me this error and I
2:06:57 copied and pasted it and put it in there
2:06:58 and it goes oh that's not a problem you
2:07:00 just install this Library I was like
2:07:02 [ __ ] how do you install the
2:07:05 library and it never judged me it never
2:07:08 got like cranky and pissy and like okay
2:07:11 I thought you would have known this by
2:07:14 now it said on repet in the lower left
2:07:18 hand corner you click on on this button
2:07:19 and in the upper right hand corner in
2:07:21 the little search box you search for
2:07:22 this name of the library and you hit
2:07:25 install and I did that and I hit the
2:07:27 play button and it ran and I was like
2:07:30 holy
2:07:32 [ __ ] that was like a little [ __ ]
2:07:38 Miracle Within 90
2:07:40 minutes I had written a fully functional
2:07:43 python application that was making calls
2:07:45 to the open AI API if you don't know
2:07:48 what that means chat gbt will tell you
2:07:50 but basically my Python program was
2:07:53 talking to the back end of open AI
2:07:56 generating content I made a python
2:07:59 application that asked someone about a
2:08:01 topic and then asked them three
2:08:02 questions about it and then generated
2:08:04 five pieces of social media content I
2:08:06 did it 90 minutes I've never written a
2:08:09 line of python in my
2:08:10 life if you haven't had an experience
2:08:13 like that where this thing blows your
2:08:15 [ __ ] mind to the point that you're
2:08:17 like what does that even [ __ ]
2:08:24 mean you're not playing enough
2:08:28 play play dick around with it go make
2:08:32 images go make songs go make a [ __ ]
2:08:35 movie it doesn't matter it literally
2:08:37 doesn't matter go to go to um GitHub
2:08:42 co-pilot and have it pair program with
2:08:47 you
2:08:49 I saw in the news advances have been
2:08:51 made in Alzheimer's due to AI the the
2:08:54 amount
2:08:56 of the quality of life improvements that
2:08:59 AI is about to unleash on us in in in
2:09:03 health and science and Discovery is
2:09:05 going to be [ __ ] incredible I lost my
2:09:08 dad to
2:09:12 alzheimer's I mean that one in
2:09:14 particular these large language models
2:09:16 are going to be spectacular
2:09:24 for so if you've got that cynical Edge
2:09:26 and you're sitting on the sidelines and
2:09:28 you're
2:09:29 like yeah I want to hate
2:09:33 this there's a lot of good reasons to
2:09:35 hate it Hollywood's tot us that the
2:09:37 robots are going to kill us apparently
2:09:40 these things are just stealing from
2:09:42 everyone it's the world's greatest
2:09:44 plagiarism
2:09:46 engine it's hard it's technical
2:09:49 there's all like the the excuse board
2:09:53 for why not to do AI is a [ __ ] mile
2:09:55 long it's so easy right now to be
2:09:59 cynical and on the
2:10:04 sidelines what if for
2:10:07 tonight you step on the
2:10:13 field what's the worst could
2:10:15 happen you prove yourself right and you
2:10:18 realize hey is a piece of [ __ ] and I'm
2:10:19 never going to use it again great at
2:10:22 least you tried
2:10:25 it oh my God if I get my mind blow do I
2:10:27 end up with Kyle hair you can you can
2:10:30 you can go to the AI salon and ask for
2:10:32 the
2:10:36 Kyle I need a haircut so [ __ ] bad
2:10:39 it's so
2:10:40 bad Kyle go check out the IR regulars
2:10:43 Channel I posted some images gen turbo
2:10:45 speed it's fast oh yeah the Gen the
2:10:48 turbo turbo stuff the the sdxl turbo
2:10:50 stuff's wild isn't
2:10:52 it let's go look at the the Irregulars
2:11:00 channel oh image generation no it's not
2:11:03 sped up oh just playing with stable
2:11:06 diffusion this this is X sdxl
2:11:17 stuff oh yeah yeah this that's
2:11:22 crazy yeah this is this is s sdxl turbo
2:11:27 is this comfy UI I don't know what this
2:11:29 what this interface is
2:11:35 but yeah this is nuts yeah so you can't
2:11:39 really see it here but as he's typing in
2:11:41 the
2:11:42 prompt it's generating it in real time
2:11:45 and you can go play with this if you go
2:11:47 to kore.ai it's got got one of these if
2:11:49 you go to Leonardo a it's got one of
2:11:52 these so you don't have to do the
2:11:54 programming interface you can just go
2:11:55 play with this [ __ ] right
2:11:59 now it's absolute
2:12:02 Insanity absolute
2:12:04 Insanity like well what would I do with
2:12:06 that I don't know it doesn't matter just
2:12:08 go play with it just go play with it
2:12:10 just go play just go play anything you
2:12:12 learn today is going to be useless in 3
2:12:14 months cuz the tools are changing that
2:12:16 fast so don't get married to
2:12:20 Perfection embrace the
2:12:24 Jank that's super cool thanks for
2:12:26 sharing
2:12:33 [Music]
2:12:35 that split view image that contrast the
2:12:38 experience of being at the front seat of
2:12:39 a high-speed roller coaster and the
2:12:42 transformative Journey that it is
2:12:47 yeah
2:12:53 the boys on the ceiling that's pretty
2:13:01 good so
2:13:04 cool so cool people well I didn't even
2:13:08 get to let me go see if I can find the
2:13:11 the Google oh you know what else I
2:13:12 didn't get to hang on let me go back to
2:13:14 chat add for a second let me go look at
2:13:16 Sam Alman here
2:13:19 so this is
2:13:21 Sam's look let's see did I get anyone
2:13:24 opening shots TW Robin and 29 others
2:13:28 liked my my thing I have to re reiterate
2:13:31 something theand sends people to open AI
2:13:33 every s night let's see it Sam
2:13:36 nice let's make this video on Sora I'd
2:13:39 like to see one oh you guys are awesome
2:13:40 that's great that's
2:13:46 great oh you guys are [ __ ]
2:13:51 awesome well who knows if Sam will
2:13:54 notice but if so that would be cool
2:13:57 um so let's go look at
2:14:05 Sam's at
2:14:08 s so Sam if you don't know
2:14:15 oops that's not where I want to do that
2:14:17 Sam is the
2:14:19 uh CEO of open
2:14:30 Ai and
2:14:34 he has
2:14:36 been taking requests to use the the uh
2:14:41 Sora video
2:14:45 model and he's been you know he's been
2:14:47 building some some things for people so
2:14:50 here's Sora our video generation model
2:14:52 we'd like to show you what it can do
2:14:54 please reply with captions for videos
2:14:56 you'd like to see and we'll start making
2:14:58 some don't hold back on the detailer
2:15:01 difficulty that that uh tweet right
2:15:04 there don't hold back on the detail or
2:15:06 difficulty that's the one I responded to
2:15:10 with my opening scene of one of the
2:15:12 screenplay my first screenplay I
2:15:15 wrote um and so I asked Sam to do it and
2:15:17 so if if you want to if you want to help
2:15:19 out if you want to help a guy
2:15:21 out go find go to Kyle Shannon go go to
2:15:25 my reply so how you do this you go to
2:15:26 Kyle Shannon go to
2:15:30 replies and then you'll
2:15:33 see you'll see this tweet where I'm
2:15:35 replying to Sam Alman and if you would
2:15:39 um like this and then um retweet it or
2:15:43 whatever the hell they call it now
2:15:44 repost it with a quote and then when you
2:15:47 quote it quote Sam
2:15:53 Alman Tik Tok just quit on me are we
2:15:55 still good am I still
2:16:00 live um but anyway let's go look at look
2:16:03 at what some of Sam's Sam's little
2:16:06 videos are let's go look at what he
2:16:10 built that one's cool I saw that one
2:16:12 before that one's
2:16:15 weird that's a half duck half dragon
2:16:18 flies through a beautiful sunset with a
2:16:20 hamster dressed in Adventure
2:16:25 Gear this one's pretty gorgeous
2:16:31 actually how long is this 10
2:16:37 seconds
2:16:41 gorgeous there's the duck here's a
2:16:43 better
2:16:45 one
2:16:47 nice oh that one's cool 10
2:17:01 seconds two golden receivers podcasting
2:17:04 on a
2:17:16 Mountaintop that's cute doggies is cute
2:17:19 all right an instructional cooking
2:17:21 session for homemade nooki hosted by a
2:17:24 grandmother's social media influencer
2:17:26 set in a rustic Tuscan country kitchen
2:17:29 with cinematic
2:17:33 lighting all
2:17:35 right weirdness with the with the
2:17:38 potatoes but like all this is pretty
2:17:40 good it's good until she picks up the
2:17:42 spoon the spoon materializes in the
2:17:45 middle of her
2:17:47 hand still got the Jank people still got
2:17:51 the
2:17:52 Jank definite Jank going on here but
2:17:54 wait what's the what's the prompt here a
2:17:56 bicycle race on the ocean with different
2:17:59 animals as athletes riding the bicycles
2:18:01 with a drone camera
2:18:04 view yeah that's pretty
2:18:06 trippy not quite
2:18:10 there The Bling zooo what's the prompt
2:18:13 here Welcome to The Bling Zoo this is a
2:18:16 single video generated by Sor oh this
2:18:18 was Sam didn't do this or maybe he
2:18:24 did Welcome to the Blingo a single video
2:18:27 generated by Sora shot changes in
2:18:29 all so 17-second video all of these
2:18:33 edits are from a single
2:18:38 generation Turtle I like
2:18:41 [Laughter]
2:18:47 turtles
2:18:50 anyway all right
2:18:54 [Music]
2:18:56 people you know what maybe maybe here's
2:18:59 the deal
2:19:00 about this
2:19:03 channel maybe this
2:19:06 channel is a place where you can park
2:19:09 your cynicism for a couple
2:19:12 hours just set it at the
2:19:14 doorstep and I'll keep it there
2:19:18 listen my favorite comedians are bitter
2:19:23 cynical like Bill
2:19:27 Burr [ __ ] you know Doug Stan
2:19:31 hope you know just bitter [ __ ] smarmy
2:19:37 like I get it I love it
2:19:43 and being cynical about this AI stuff is
2:19:48 not the play right now in my opinion so
2:19:52 maybe come hang out on these lives I do
2:19:55 these every night I generally start
2:19:57 around 8:00 p.m. mountain like that's
2:20:00 too late I doesn't matter it's what I
2:20:02 do and I do it seven nights a
2:20:08 week that's stupid I
2:20:11 know my wife agrees my kids think I'm
2:20:15 [ __ ] insane I am
2:20:19 I do it seven nights a week what if even
2:20:23 if you hate this [ __ ] even if you're
2:20:25 like it's
2:20:26 awful what if you came here for like an
2:20:29 hour 30 minutes you just park your
2:20:32 cynicism at the door and while you
2:20:34 listen to me ramble you just
2:20:37 play if you get stuck on something you
2:20:40 ask a question in here and one of the
2:20:41 Irregulars will answer you or I'll put
2:20:43 you point you over to the AI salon and
2:20:44 you'll go maybe join there God forbid
2:20:47 you can connect with Community that's
2:20:49 trying to figure the [ __ ] out
2:20:52 too and just kind of play with it like
2:20:54 that that could be a
2:20:57 thing that could be
2:20:59 a good
2:21:03 thing well what if I enjoy it I know
2:21:06 it's going to [ __ ] up your cynicism
2:21:08 isn't it but you just just do it for
2:21:09 like a half hour and then you pick the
2:21:11 cynicism back up when you go walk out
2:21:14 the door it's right it's in the umbrella
2:21:16 stand right by the door
2:21:19 just pick it up and you just walk out
2:21:21 you know whack an old lady in the head
2:21:22 with the umbrella you just you just go
2:21:24 right back to being
2:21:27 a a bitter
2:21:34 prick you can judge people make fun of
2:21:36 them you can yell at them for just
2:21:39 existing but maybe just you know for
2:21:41 like a little time come in here and have
2:21:44 fun I'm trying to master it so I can
2:21:46 help avoid all the AI Doom stories
2:21:48 shortcuts are not
2:21:50 okay yeah like the the that's the thing
2:21:53 the AI Doom stories are so easy to find
2:21:56 do you think the media is going to talk
2:21:58 about how people are Reinventing
2:22:00 themselves with
2:22:02 AI remember when John krinski did the
2:22:05 the good news channel at the beginning
2:22:07 of Co and it was like people were really
2:22:09 into it for like two weeks and then they
2:22:12 were like this is boring could you just
2:22:14 talk about murders and [ __ ] again that
2:22:16 would be good
2:22:27 but this isn't boring and this isn't
2:22:30 it's not all good news like sometimes
2:22:32 this [ __ ]
2:22:34 sucks sometimes put people put in a lot
2:22:36 of time to build something and then the
2:22:38 tool itself just [ __ ] falls apart
2:22:41 it's really
2:22:46 frustrating all right all right how much
2:22:49 longer do we have to wait until AI can
2:22:51 create a video game based on a prompt
2:22:53 vocal sync well it can sort of do it now
2:22:55 it depends on your definition of video
2:22:57 game but if you're talking about what I
2:22:58 think you're talking about which is how
2:23:01 long until I can I can prompt a world
2:23:04 into existence that's got game mechanics
2:23:06 in it and I can just enter the world and
2:23:08 start playing I'm guessing three years
2:23:11 and I think I think we'll start to see
2:23:13 this
2:23:14 year I'd be really surprised if real
2:23:17 engine doesn't do something with AI this
2:23:21 year where you'll be able to go in and
2:23:24 speak an
2:23:26 unreal five world into
2:23:30 existence and I I would also
2:23:36 probably
2:23:38 unity's Unity could happen this year but
2:23:40 maybe it's next year I don't know enough
2:23:42 about the development world for
2:23:46 Unity but you're going to have game
2:23:48 development environments are going to
2:23:50 add AI layers on top of
2:23:53 them and so I think we'll we'll we'll be
2:23:58 generating 3D worlds that we can enter
2:24:00 and interact with probably within a
2:24:04 year and then I think over the next year
2:24:08 there will be rudimentary game mechanics
2:24:10 added to those and and you know maybe
2:24:13 there's even rudimentary game mechanics
2:24:14 added you know this year
2:24:18 and I think within three
2:24:21 years it might not be a AAA game but it
2:24:25 might be a double a game that you just
2:24:27 speak it into existence you say I wanted
2:24:30 to have these game mechanics I wanted to
2:24:31 have this kind of world environment I
2:24:33 wanted to have car racing with realistic
2:24:38 physics and I want it to be raining or
2:24:40 not raining or sunny
2:24:42 or and then you'll be racing a car Kyle
2:24:45 what are your thoughts on VR and AI um
2:24:49 well they they they right now they are
2:24:51 so if you go look at what's happening
2:24:54 inside the Vision Pro goggles there's a
2:24:57 [ __ ] ton of machine learning going on in
2:24:59 there and Ai and realtime processing
2:25:03 data processing the R1
2:25:08 chip so so they are very tightly coupled
2:25:16 um
2:25:18 and I I think that
2:25:21 um here's here's what I'm excited about
2:25:24 for for AI with with the VR
2:25:29 stuff where we've been and what
2:25:32 Zuckerberg
2:25:35 spent a lot of money on and a lot of
2:25:38 years
2:25:39 on was paying artists to build
2:25:46 worlds and the problem with that is
2:25:51 that to the extent that the artists
2:25:53 build a world that you don't
2:25:56 like what are you going to
2:26:00 do so they built this world at meta that
2:26:03 looked like second
2:26:05 life and I think most people looked at
2:26:07 it were like uh didn't we have this in
2:26:11 like 1999 is Is this different than
2:26:16 that
2:26:19 but if if I figure so if you
2:26:21 figure Apple Apple has product cycles of
2:26:25 approximately 18 months right so we'll
2:26:28 see we'll see a V2 of the Vision Pro in
2:26:31 in 18
2:26:32 months and then 18 months after that
2:26:34 we'll see the
2:26:37 V3 the V3 of the Vision
2:26:40 Pro are going to be pretty [ __ ] slick
2:26:43 right they're going to be small they're
2:26:45 going to be
2:26:46 light they're going to be nastily
2:26:49 capable and that'll be right around the
2:26:51 time where the thing I just mentioned
2:26:53 where you can just speak worlds into
2:26:55 existence and enter them I think you're
2:26:58 going to see these two worlds come
2:26:59 together very very
2:27:03 quickly very very
2:27:08 quickly um but but then it's it's all
2:27:11 the other things that you can do with AI
2:27:14 right
2:27:16 like this year we're going to have
2:27:18 agents and we're going to have
2:27:21 um these things are going to get good at
2:27:24 reasoning like when when the agents show
2:27:26 up I right now we don't know we don't
2:27:30 know what what's about to happen but
2:27:33 what's about to happen
2:27:35 is you're not going to have to go to an
2:27:38 AI and tell it to do
2:27:42 something you're going to say hey I
2:27:45 would like this accompl ACC lished in my
2:27:48 life and some agents will go off and do
2:27:50 [ __ ] for
2:27:52 you and they'll come back and go okay
2:27:54 boss it's done and then you look at it
2:27:56 and you go you did
2:27:58 shitty be like okay you want me to go do
2:28:01 it again yes and then go off it come
2:28:04 back okay boss I did
2:28:06 it that's pretty good could I have it in
2:28:09 yellow sure off they go come back here
2:28:11 it is in yellow that's
2:28:15 good can I have it here tomor tomorrow
2:28:17 sure Lear enough and get that 3D printed
2:28:19 for you have it on your doorstep
2:28:21 tomorrow what is Unity Unity is an
2:28:24 open-source game development environment
2:28:26 that is incredibly powerful and
2:28:30 Incredibly
2:28:32 complicated there's like there's
2:28:35 like 17 or 18 different careers in each
2:28:40 of the
2:28:41 functional the feature functional areas
2:28:44 of unity right this one's a modeler this
2:28:48 one's a rigger this one's a a texture
2:28:51 person this one's a game mechanics
2:28:54 person this one's a particle physics
2:28:56 person this one's it's just like like it
2:28:59 does like all of the [ __ ] that you need
2:29:02 to do a
2:29:04 game except you have to have 17
2:29:08 degrees to be able to to do
2:29:12 anything or you need to start a company
2:29:15 with like five people that each know
2:29:17 three of those functional areas and then
2:29:19 you start your little company and you go
2:29:20 make a game
2:29:22 and no one plays it and you get
2:29:25 depressed cuz it took you five years of
2:29:27 your life and it's okay but not great
2:29:30 and you launch it well imagine
2:29:33 if you could just say hey I'd like a
2:29:37 game that has these kind of shapes in it
2:29:39 these kind of characters in it these
2:29:41 kind of sounds in it these kinds of
2:29:43 lighting in it these kinds of Worlds
2:29:44 these kind of game mechanics and the AI
2:29:47 just figures out all the T tactical
2:29:51 [ __ ] that's the kind of [ __ ] that we'll
2:29:53 start to see emerge over the next 18
2:29:55 months to three
2:30:09 years oh
2:30:14 wow so cam cam K can cam cat came from
2:30:18 Cleveland who's one of the Irregulars
2:30:20 and she's a mod she just wrote this in
2:30:21 the mod
2:30:23 Channel Kyle I'm legally blind and I'm
2:30:25 using VR accessibility features to bring
2:30:27 images close to my eyes and see things
2:30:30 that I can't see otherwise I visit the
2:30:33 world through Google Earth I'm so
2:30:35 excited to see what accessibility
2:30:37 options are with Vision Pro 2 or three
2:30:39 and what AI can
2:30:41 do th I'm telling you the I think it's
2:30:44 the thing I'm most excited about with AI
2:30:47 it is that what it what
2:30:49 it what it ultimately is is it's an
2:30:52 accessibility
2:30:55 engine and it's like accessibility to
2:30:57 knowledge accessibility
2:30:59 to um like that vision and and
2:31:05 um like being able to just connect
2:31:08 people and things and ideas
2:31:12 instantly people that couldn't
2:31:15 speak for forign languages now have we
2:31:19 now have Babel
2:31:21 Fish did you know this have you played
2:31:24 with this
2:31:27 [ __ ] oh you're welcome Kim thank you for
2:31:29 sharing
2:31:30 that so this is this is just an iPad
2:31:33 have you played with pi or the the the
2:31:36 chat
2:31:37 GPT app on on iOS the mobile
2:31:41 app so here's chat
2:31:43 GPT so I'm just going to click the
2:31:45 little headphones down here little
2:31:52 headphones how are you doing
2:31:59 tonight I'm just a computer program so I
2:32:02 don't have feelings but thanks for
2:32:04 asking how can I assist you
2:32:06 tonight uh I think I'd like to speak
2:32:09 German so what I'm going to do is I'd
2:32:10 like to speak an English sentence and
2:32:12 then I'd like you to translate it in
2:32:14 German without repeating the English
2:32:15 back just speak the German
2:32:20 okay sure I can do that go ahead with
2:32:23 the English sentence you'd like
2:32:24 translated to
2:32:26 German I'm thirsty and my Apple watch
2:32:29 battery died where can I get a
2:32:44 beer a friend just came and they only
2:32:46 speak Japanese could you repeat that in
2:32:48 Japanese
2:33:05 please
2:33:09 accessibility you know we were talking
2:33:11 about Alzheimer's
2:33:12 before imagine being able to
2:33:15 create
2:33:18 to take things that you've written or
2:33:21 blog
2:33:23 posts and if your loved
2:33:28 one is maybe in a home and they're
2:33:30 lonely and you can't spend 24/7 with
2:33:33 them cuz you have your life to
2:33:37 live imagine being able to create an
2:33:39 avatar of yourself with your
2:33:43 voice and your knowledge of
2:33:47 the
2:33:48 world and you could be there for your
2:33:51 loved one
2:33:53 247 with infinite
2:33:56 patience always willing to tell the
2:33:58 story
2:34:00 again always willing to explain who they
2:34:02 are
2:34:04 24/7 for that person with
2:34:07 Alzheimer's imagine
2:34:09 that that's not like five years away
2:34:12 that's like this
2:34:15 year
2:34:25 imagine if all your life all you wanted
2:34:28 to
2:34:30 do was tell
2:34:34 stories but you [ __ ] around in high
2:34:37 school and you skipped class and you
2:34:40 didn't get good at the
2:34:43 words but you got great ideas for
2:34:46 stories but the minute you sit down with
2:34:48 the word
2:34:50 machine the brain [ __ ] seizes
2:34:56 up what if
2:34:58 overnight you were good with
2:35:03 words
2:35:05 oh I can just do that
2:35:12 now what if when you were 17 you had an
2:35:14 idea for a kids book
2:35:17 about kids that felt the way you
2:35:20 felt and now you're 54 years old and you
2:35:24 still think about that [ __ ] kids book
2:35:26 and you've never written
2:35:28 it cuz you don't have the artistic
2:35:31 talent and you don't know artistic
2:35:33 people to make the images and you don't
2:35:34 have money to go find them and you just
2:35:37 you feel trapped and but you still have
2:35:39 this idea you're like but I don't know
2:35:41 how to write a kids book what if this
2:35:45 weekend you could write that book
2:35:47 illustrate it and publish it on Kindle
2:35:54 publishing
2:36:00 accessibility like that's [ __ ]
2:36:02 miracle level
2:36:04 [ __ ] that the tools are there right now
2:36:08 that people can go do like those things
2:36:10 that I was talking about right
2:36:13 now not like five years out
2:36:17 and that's what's happening in the in
2:36:19 the AI salon and in this community night
2:36:21 after night after night day after day
2:36:24 hey Kyle I just want to let you know
2:36:26 Rudy from South Africa the other day he
2:36:30 calls me Mr Kyle hey Mr Kyle he's so
2:36:34 sweet he's a [ __ ] nutcase he from
2:36:37 South Africa he lives in a small town in
2:36:39 South Africa hey Mr Kyle he's been
2:36:42 making Tik Tok videos and he make these
2:36:44 he makes these elaborate [ __ ] crazy
2:36:46 crazy ass editing Tik Tok video I don't
2:36:48 know how the [ __ ] he does it they have
2:36:51 like all this energy he's like I'm
2:36:53 excited about this
2:36:55 [ __ ] hey Mr Kyle I I I've been making
2:36:59 these videos and someone wanted me to
2:37:01 come over and tell them about this AI
2:37:03 stuff oh they they wanted me to tell
2:37:07 them how much I would charge oh Mr Kyle
2:37:10 I just got the
2:37:13 job this is a kid in a small town in
2:37:16 South
2:37:18 Africa that doesn't have
2:37:22 access that all of a sudden has access
2:37:25 is all of a sudden getting paid to
2:37:28 do stuff that he's learning on the
2:37:31 Fly accessibility holy
2:37:34 [ __ ] imagine thinking of a device and
2:37:37 having AI create a prototype for you
2:37:39 imagine the ideas
2:37:41 yeah
2:37:45 imagine
2:37:48 imagine one of my
2:37:50 like from my 20s one of my
2:37:54 um hero
2:37:57 companies that I thought oh my God I
2:37:59 want to meet those guys I want to be
2:38:01 part of that is Ideo right they they
2:38:03 they solve these problems the they
2:38:05 create all these amazing designs they
2:38:07 these industrial designers
2:38:10 right
2:38:13 imagine all your life you've looked up
2:38:15 to idio
2:38:18 and you want to create products and but
2:38:19 you don't have the skills and you don't
2:38:21 know how to do Drafting and you don't
2:38:23 know how to do industrial design but you
2:38:24 like but you've had this
2:38:27 aspiration what if you could just go hey
2:38:30 Dolly 3 I have an idea for a product how
2:38:33 would I actually manufacture that or or
2:38:36 chat GPT and it tells you and well what
2:38:39 would what would be some interesting
2:38:41 Design Concepts for that and it tells
2:38:43 you and you're like could I see one of
2:38:44 those and then it shows you
2:38:47 like oh that's really
2:38:49 good what if over a two- week period you
2:38:52 do a deep dive on that product idea of
2:38:55 yours and you come out that two weeks
2:38:58 with a plan for how to sell
2:39:01 it a pitch Deck with
2:39:05 prototypes working
2:39:12 code oh PS that Brandon just said PS it
2:39:15 will never talk down to you you know one
2:39:17 of the things that I'm most excited
2:39:19 about about
2:39:20 AI it's like having a golden retriever
2:39:23 you know how you know how like why do we
2:39:25 love dogs we love dogs because when you
2:39:28 get home they're like oh God where you
2:39:30 been it's so good to see you this you're
2:39:33 the
2:39:34 best that's what chat GPT is like you're
2:39:37 like ah I kind of have this shitty idea
2:39:39 it's like oh dude such a good idea you
2:39:41 want to do some more you you want me to
2:39:43 tell you more about you want 10 more
2:39:44 like that
2:39:50 and people are sitting on the sidelines
2:39:52 don't sit on the [ __ ]
2:39:54 Sidelines jump
2:39:56 in all right I gotta go it's late people
2:39:58 here it's so [ __ ]
2:40:02 late it levels All Humans anyone can do
2:40:06 anything you're our golden retriever
2:40:08 Kyle I do have I do have golden
2:40:10 retriever energy don't I I've been told
2:40:12 that a couple of times in my life what
2:40:15 are you going to do take it it's it's
2:40:18 much better than a rubber duck debugger
2:40:20 it's good Kyle the golden retriever
2:40:22 thank you Kyle thank you phah great I
2:40:25 can get ready for work now all right Dr
2:40:27 Bud your excused go get ready for
2:40:32 work be not afraid I can't do a 30
2:40:34 minute goodbye tonight Danielle I can't
2:40:36 do it I can't do it I gotta I gotta
2:40:39 leave rough rough thank you um okay so
2:40:44 here's the
2:40:45 deal
2:40:48 if you're like that sounds good I'd like
2:40:50 to do that where do I start how do I
2:40:52 find people like that go to the salon.
2:40:55 click on the first link that's the AI
2:40:57 Salon Community read about what we're up
2:40:59 to read about our
2:41:02 values if you resonate with the values
2:41:04 join the salon and introduce yourself go
2:41:07 to the community section introduce
2:41:09 yourself and say
2:41:13 hi do that if you like this channel if
2:41:17 you like my little rant or even if you
2:41:18 didn't if you just think like you know
2:41:20 what I wouldn't mind showing up here and
2:41:22 parking my cynicism at the door for half
2:41:24 an
2:41:25 hour follow me follow the channel I
2:41:28 don't know how you do that it's Tik Tok
2:41:29 figure it out you're younger than
2:41:33 me or you're older it doesn't [ __ ]
2:41:35 matter follow this channel if you want
2:41:38 to support it you can subscribe to these
2:41:41 lives you can buy in one of these
2:41:43 Corners is um a video series series that
2:41:46 I put together on why I think AI is
2:41:48 actually underhyped not overhyped if you
2:41:51 can believe
2:41:53 that you can buy one of those who won
2:41:55 the Valentine's Day Challenge I do not
2:41:57 know I haven't seen it
2:41:59 yet I think Leo was heading that
2:42:02 up um she took that
2:42:04 on
2:42:07 um yeah and like let's see how many we
2:42:10 got 27,000 likes tonight we didn't do we
2:42:13 didn't do a lot of like bombing so
2:42:14 27,000 not not bad for a rant and for
2:42:18 showing a bunch of stuff Empower thank
2:42:20 you so much that's very generous of you
2:42:22 I appreciate that Friday night double
2:42:23 date night with you and Leah exactly
2:42:25 yeah so tomorrow night okay so here's
2:42:27 the deal tomorrow night at um 6 PM
2:42:33 Mountain um Leah and Jane's event
2:42:37 happens at 8:00 pm will be this or I
2:42:41 might go at 9: I don't know what's going
2:42:43 on tomorrow night depending on if I'm
2:42:45 going out G hopping or whatever it might
2:42:47 be 9 but it'll probably be eight and
2:42:50 then Saturday at noon Pacific so 1 pm
2:42:55 mountain is an Murphy's GPT jams so
2:42:59 Friday night date night and Saturday
2:43:01 hangover you can have your Bloody Mary
2:43:03 while you're at an Murphy's GPT jams all
2:43:07 right I did at least I didn't turn into
2:43:10 an orange tonight thank you Rachel
2:43:11 Sparkles that was pretty funny I brought
2:43:13 someone up on stage and popped an orange
2:43:15 on my head and I couldn't figure out how
2:43:17 to turn it off it was like that dude
2:43:18 that showed up for court and he had a
2:43:20 cat hat on he couldn't he didn't know
2:43:22 how to turn it off and the judge was
2:43:24 just increasingly pissed
2:43:33 off oh
2:43:40 man all right so and then the last thing
2:43:43 I like to do here cuz I just like to do
2:43:44 it remember ROM ER room those of you gen
2:43:47 xers out there so Romper Room was the
2:43:51 show at the at the end of Romper Room
2:43:53 they they'd pull out the magic mirror
2:43:55 and they'd look in the magic mirror and
2:43:56 they'd say I see Jimmy and I see Sally
2:43:59 and I see Bart and I see and you'd
2:44:01 always just be waiting what are they
2:44:02 going to say Kyle they'd never say
2:44:04 [ __ ] Kyle Kyle didn't
2:44:06 exist and so it just always made me
2:44:10 smile so I like doing that show some
2:44:11 love before we go oh thank you very much
2:44:13 yeah Danielle thank you so much Tobias
2:44:15 great to see you AO Kim katkim from
2:44:18 Cleveland cuno Elizabeth Jim let's see
2:44:22 who else we got in here tonight Brian
2:44:24 Whitney good to see you Joker Source
2:44:26 Camp
2:44:29 Tams Rachel
2:44:31 Sparkles Brian Whitney laughing
2:44:37 Thunder Dr J good to see you James
2:44:41 Cal Source Camp Empower pixels thank
2:44:45 thank you as always vocal sync glad to
2:44:47 have found you welcome vocal sync I'm
2:44:49 glad you found us
2:44:52 too
2:44:55 uh okay so oh by the way I know we
2:44:59 played with the with the that open AI
2:45:01 video [ __ ]
2:45:03 earlier a human being good night Kyle
2:45:06 another educational and exciting evening
2:45:08 thank you um that's that open AI video
2:45:11 shit's not available right now it's in
2:45:14 what's called red team testing so
2:45:15 they're testing it so it's probably
2:45:17 going to be weeks just forget about it
2:45:19 it'll come back when it comes back we we
2:45:22 will play with the [ __ ] out of that
2:45:24 stuff um Tams I love this group can't
2:45:27 wait to learn more in the AI Salon yep
2:45:29 you'll love it you'll love it it's good
2:45:31 just jump in participate share [ __ ]
2:45:34 learn [ __ ] break [ __ ] just be just be
2:45:38 nice just be engaged go follow the
2:45:40 values and uh it's a remarkable
2:45:42 Community all right everybody um thank
2:45:45 you thank you thank you Kyle matters
2:45:48 matters and turtles Kyle and turtles
2:45:51 matter I agree all right uh I will see
2:45:54 you tomorrow night for date night um 600
2:45:57 PM Mountain Leah and Jane 8:00 P PM
2:46:01 Mountain this channel Saturday noon
2:46:04 Pacific and Murphy's GPT jams it's a
2:46:08 jam-packed weekend people be there or
2:46:13 don't bye