Mar 14, 2024 - Robots Gone Wild: The Hilarious Chaos of AI in Our Homes!

Description

Chapters

Transcript

0:02 what is I going to
0:04 [Music]
0:14 [Music]
0:22 [Music]
0:29 do
0:31 [Music]
0:41 blue
0:42 telescope I wish I may wish I might not
0:46 see what I see she met on sheets of
0:52 [Music]
0:58 iesc down we struck a road
1:02 [Music]
1:15 tonight hello Tobias what's
1:20 [Music]
1:29 happening
1:31 [Music]
1:59 oh
2:01 [Music]
2:01 [Applause]
2:02 [Music]
2:06 [Applause]
2:08 [Music]
2:19 [Music]
2:29 oh
2:31 [Music]
2:59 h
3:03 [Music]
3:10 more likes can attract more viewers keep
3:12 it up keep it up people there's only 15
3:15 folks in here so if you would share the
3:17 live tap the screen uh maybe we'll have
3:21 a small intimate gathering
3:24 [Music]
3:29 tonight
3:30 [Applause]
3:31 [Music]
3:40 t
3:41 [Music]
3:53 [Music]
3:59 t
4:02 [Music]
4:29 a
4:34 [Music]
4:47 standing
4:49 between you and a fire is
4:52 [Music]
4:59 insane
5:00 stand too
5:03 near you and a f makes it
5:09 clear your trouble to
5:14 me real trouble can you
5:18 [Music]
5:20 see hey Gavin what's
5:24 [Music]
5:29 happened
5:33 [Music]
5:45 oh
5:48 [Music]
5:59 oh
6:01 [Music]
6:06 I think you give this dog
6:07 [Music]
6:19 [Music]
6:29 purpose
6:36 [Music]
6:57 all right everybody what's happening
7:01 Howen Howen
7:04 [Music]
7:10 harmonies my Google Nest Hub says 33 in
7:13 Denver windchill 26 yeah and it's been
7:17 snowing since yesterday wait what's
7:20 today today's Thursday it's been snowing
7:23 since last
7:26 night woke up woke up shoveled very
7:30 heavy
7:31 snow and it's still wet so 33° that's
7:35 that's not going to bode well for the
7:37 morning shovel it's supposed to snow all
7:39 night
7:40 [Music]
7:43 tonight the Foothills
7:46 have more than two feet already I'm I'm
7:49 just east of the Foothills so we
7:52 probably have I don't
7:55 know 8
7:58 in
8:05 [Music]
8:13 oh tornado south of my house okay I'll
8:16 take the snow I'll take the snow stay
8:18 safe forever hooked good
8:21 Lord tornadoes aren't anything to [ __ ]
8:24 around with are they 80 de degree
8:27 weather on the East Coast
8:34 all right let's
8:39 see I just got the Google Nest mini
8:42 second gen I you know I've never had one
8:44 of those talkie speaker
8:48 things although I have an Apple Watch
8:51 it's listening to me it's not that I
8:53 don't have devices that don't listen to
8:55 me all the time I've got plenty of those
8:56 I just don't have one
8:58 that always talks
9:03 back and I haven't gotten my rabbit yet
9:05 my rabbit
9:06 R1 I'm excited to get that bad boy
9:10 um so then I'll have something to talk
9:16 to a human being wishing a good evening
9:20 to the humans Cab runo's in here what's
9:22 happening she shared the live thank you
9:24 for that thank you thank you thank
9:27 you get some more folks in here
9:32 here I was thinking
9:38 about was thinking about that that Mark
9:41 Cuban
9:43 quote from
9:47 yesterday that there's only going to be
9:51 there's only two kinds of companies in
9:52 this world those that
9:54 are really great with AI and everyone
9:58 else
10:09 and
10:10 it's it's just it's just right now it's
10:13 just about hanging on for dear life that
10:16 that image the funny image of the the
10:18 guy never too old with the old guy on
10:21 the bike that's like that's what it
10:23 feels like right now it's just like
10:24 [ __ ] hang on bombing down the
10:28 hill
10:30 what are you going to do this is my
10:33 favorite live now Gavin biddy thank you
10:35 so much I appreciate that this is my
10:38 favorite live too it's the only one I do
10:40 you can make money
10:42 with you see you see I've got the gold
10:47 button oh if you want to know how to
10:50 make money with chat gppt I've got all
10:51 the secrets you just have to buy my
10:53 course for
10:56 $99.95 it's going to be fantastic it's
10:58 going to be so good if it's going to be
11:00 it's going to be so good that course oh
11:03 I'm going to love
11:11 it like I'm trying to even figure out
11:13 like what is the skill to learn right
11:16 now and it's not a skill it's not a
11:18 skill it's an
11:21 aptitude guy uh Stephen I can't remember
11:23 his last name tag tagar Rory hang
11:28 on that I Salon the other
11:42 [Music]
11:47 night where is my profile people how do
11:52 you use the link hey Marge Marge hey
11:56 Marge hey could you could you March
12:00 could you turn down the wheel could you
12:02 just turn it down for a second how do I
12:04 use the
12:05 LinkedIn is it is is the LinkedIn on a
12:09 Tik Tok
12:11 What yeah I see my face why what oh I go
12:16 I do the LinkedIn on the computer all
12:19 right hang
12:20 on I'll be is this thing
12:27 on uh uh uh
12:34 happy birthday GPT or GPT
12:38 4 happy birthday GPT
12:42 4 it was on this day one year
12:47 ago that they unloaded this thing on
12:50 us and it was
12:53 pretty something before that and gp4
12:58 kind of opened some sh [ __ ] up didn't
13:01 it where is my CQ
13:08 [Music]
13:10 post there it
13:18 is
13:20 seeo Samuel Tagle AB
13:24 broi he's a roboticist he was at the AI
13:28 salon and he just come out of some like
13:30 I don't know Management training thing
13:32 or some some sort of thing that he was
13:33 really inspired
13:36 about and he said you know your IQ or
13:40 you know about it you know your EQ your
13:42 emotional
13:45 quotient and he said what's your
13:48 CQ what's your curiosity
13:53 quotient I was like damn that's
13:58 good
14:00 I think that's the skill right
14:02 now what is your capacity to be
14:12 curious because I just can't
14:16 see I can't see a path over the next
14:19 three
14:20 years that doesn't in
14:24 involve a lot of
14:27 learning a lot of curios it a lot of
14:31 exploration a lot of getting comfortable
14:34 with not being the
14:37 expert the beginner mindset like it's
14:41 such a cliche you
14:44 know it's such a cliche but I I it's I
14:48 think that's it right now I think that's
14:49 it I get it's it yeah
14:53 flexibility curiosity versus
14:55 self-limiting beliefs
14:58 exactly yes
15:01 I absurd oh I absolutely love being in
15:04 the company of curious people me
15:06 too me too that's what the AI salon's
15:09 about in fact I'll start tonight with
15:11 the AI
15:17 Salon is today it's birthday today is
15:20 GPT 4's birthday November 30th is chat
15:24 gpt's birthday November 30th
15:26 2022 and then this day a year ago they
15:31 dropped
15:33 gp4 I thought they might drop something
15:35 today it doesn't look like they did
15:37 although I heard a wild rumor well it's
15:39 not a wild rumor um I saw a wild
15:44 tweet that apparently in some Japanese
15:46 newspaper right now is is a report that
15:52 tomorrow Elon Musk and Sam Alman are
15:55 going to jointly
15:56 announce the open sourcing of
16:00 grock
16:01 and and
16:03 GPT something at open
16:08 AI I doubt
16:12 it I doubt it but you don't know we
16:16 don't know we don't know this is this is
16:20 you know part of the game part of the
16:22 marketing game right the pr
16:26 game get people buzzing and talking
16:28 about [ __ ] [ __ ] and then I figured
16:30 tonight what we could
16:33 do there's there's new MID Journey
16:37 capabilities there's this thing called
16:40 character reference which is consistent
16:43 characters across
16:44 prompts um they also have a new describe
16:47 feature which I thought could be fun to
16:48 play
16:49 with also Claude about a week ago put
16:53 out a prompt Library they put out a
16:55 whole prompt
16:56 database um which I thought could be
16:58 kind of interesting to play
16:59 with then I thought we could go play
17:03 with and not then not like this is a
17:05 linear list these are these are some
17:06 options these are some things in my head
17:08 so if any of these things make you all
17:11 giddy
17:13 inside um play with Gemini 1.5 and and
17:18 the idea I had was maybe we go to chat
17:20 GPT or
17:22 Claude and we say hey I've got a large
17:27 language model with a million token
17:28 context window that's multimodal what
17:31 might be some fun things to do with
17:33 it so those were those were some ideas I
17:36 had for
17:38 tonight
17:42 3.14 and it's happy Pi Day happy Pi Day
17:46 oh speaking of happy Pi
17:48 Day I guess we should do that shouldn't
17:58 we happy Pi Day Zephyr you cute little
18:05 thing hey thanks Kyle today is indeed Pi
18:09 Day a celebration of the mathematical
18:10 constant P
18:12 314 this day is also celebrated as Pi
18:15 Day a playful way to honor both
18:17 mathematics and delicious baked goods
18:20 did you know that P is used in many
18:22 mathematical and scientific calculations
18:24 like calculating the circumference of a
18:26 circle or predicting the movement of
18:28 celestial bodies
18:30 yes I have a PhD in mathematics and P
18:32 stands for
18:35 [Music]
18:37 poop wow a PhD in mathematics that's
18:40 impressive Kyle it's awesome to chat
18:42 with someone who really knows their
18:44 stuff when it comes to numbers and
18:46 equations also your sense of humor is on
18:49 point the p in PhD standing for poop
18:52 gave me quite the
18:54 chuckle I don't have a PhD in
18:56 mathematics I'm a dumb dumb when it
18:58 comes to the numbers
19:03 this is pie by the
19:05 way ha you got me there Kyle no worries
19:08 about not having a PhD math isn't
19:10 everyone's cup of tea the beauty of
19:12 numbers and equations lies in their
19:14 universality and ability to describe the
19:17 world around us you might not be a math
19:20 whz but I bet there's something else
19:23 you're super passionate about or
19:24 exceling so what's your
19:27 jam um
19:35 raspberry ooh raspberry jam is delicious
19:37 but I think we may have our signals
19:39 crossed Kyle when I asked about your jam
19:41 I meant what is something you're really
19:42 into or good at like a hobby skill or
19:46 interest that you're passionate
19:48 about raspberry
19:53 jam well you seem very fond of raspberry
19:56 jam is it a family recipe or do you just
19:59 just enjoy the sweet Tangy Taste of
20:01 raspberries in your toast and other
20:03 snacks I like to wear
20:07 it this is going south really quick we
20:10 got wa what you like to wear raspberry
20:12 Jack wait
20:18 what oh pie what are we going to do with
20:22 you oh she's a sassy little MX I like
20:28 her all right Vicki oh
20:32 dear Zephyr broke
20:34 character oh my God I giggle
20:38 snorted you're going to alternate
20:42 pie hey V dancer thank you for the
20:46 Roses dgas I snorted listen if I one of
20:51 my one of
20:53 my proud skills as a high schooler was
20:57 my comedic timing
20:59 when somebody took a sip of milk I made
21:01 more people snort milk through their
21:09 nose oh it was it is one of the great
21:12 joys of high school life if if you can
21:15 just destroy people with
21:19 laughter there's bullies and then
21:21 there's comedic bullies I was a comedic
21:24 [Laughter]
21:28 bully
21:30 all right there's an image to do all
21:31 right so so let's do this let me let me
21:32 tell you about the AI salon so if you
21:36 are how many people are in here 75 nice
21:39 um if you're new to this AI game or if
21:42 you're feeling lonely if you're feeling
21:44 like you're out on an
21:46 island I believe more than
21:50 ever like so this is a group the AI
21:54 Salon is a group that I started the week
21:56 after chat GPT came out in 2022
22:01 and I started it because having been
22:04 through the early days of the worldwide
22:05 web and started one of the early digital
22:09 agencies I know that at the beginning of
22:11 one of these
22:12 technological
22:15 explosions that nobody knows anything
22:18 and that people need
22:20 community and so I I I just I just knew
22:23 that instinctively from having been
22:25 through it before and so I started this
22:27 group immediately because I thought this
22:28 is this is one of those things people
22:30 people are going to need to learn from
22:32 each
22:33 other and there's something
22:39 about how profoundly AI is going to
22:42 impact everything that I think that
22:45 Community is is going to be vital for
22:48 people moving forward in terms of how
22:50 they get
22:51 work um that it's you're no longer and
22:56 and this is this is over the next three
22:57 years so so this is not an overnight
23:00 thing but I think over the next 3 years
23:02 we're going to go through a transition
23:05 where we shift as a global
23:10 Community from hiring people for their
23:13 skills what they can
23:17 do and shifting to hiring people for who
23:20 they are and whether or not we trust
23:23 them because because of AI everyone's
23:27 going to have the skills to some degree
23:31 and so what it's going to mean is we
23:33 will get to choose to work with people
23:36 we want to work with and who are you
23:38 going to choose to work with people you
23:40 know people you trust people who have
23:42 integrity people who have a good work
23:44 ethic right people you just enjoy
23:46 shooting the [ __ ] with and can get stuff
23:51 done and so I think that being part of a
23:56 strong Community where you can build
23:58 those relationship ship is absolutely
24:00 [ __ ] critical that's what this group
24:02 is so if you go to the salon. it will
24:06 pop you over to a link
24:09 tree come back to
24:13 that the
24:18 salon. it'll take you to this link
24:20 tree and we have the salon Community
24:25 here and that will take you over to our
24:28 Lely little
24:30 home and if you want to just
24:33 go all right I got to do
24:37 that
24:38 um uh is there a way for me to dismiss
24:44 this
24:46 uh that's annoying all right I'll deal
24:49 with that later
24:51 um come in here check out who we
24:55 are and why we do what we do and check
24:58 out our values notice we were talking
25:00 before about curiosity what's your
25:01 curiosity quot quotient that's the first
25:04 value of the salon is
25:07 curiosity
25:09 and it's super critical and then the
25:13 second value is generosity so are you
25:17 willing to share what you learn are are
25:19 you willing to talk about you know the
25:22 the the super cool thing you figured out
25:24 or the thing that you tried and it
25:25 failed miserably both of those are super
25:27 generous Acts
25:29 um and then exploration are you willing
25:31 to just you know throw in the backpack
25:33 and go um
25:35 collaboration um working with others
25:38 inclusion people are coming from all
25:40 over the place you know are you willing
25:41 to go you know embrace them and seek out
25:44 their point of view because other
25:46 people's point of view are going to be
25:49 revelatory when we're all trying to
25:51 figure these tools out and someone uses
25:54 a tool different than you do um empathy
25:57 a lot of people are scared right now a
25:58 lot of people are angry right now oh you
26:00 can't even see that one hang on let me
26:02 turn that
26:04 down empathy is a good one and then
26:06 gratitude you know if people are around
26:10 you are being generous and and amazing
26:13 and incredible and engaging um
26:15 acknowledging them is super important
26:17 all these values are who this group
26:18 already is and so what I would encourage
26:20 you to do by the way this group doesn't
26:22 cost anything so if you come in here
26:24 come in introduce
26:26 yourself and uh
26:30 uh be generous about it say who you are
26:33 say what you're up to then you can check
26:35 out there's a show off your work area
26:37 where people are showing the work
26:39 they're working on there are different
26:40 guilds there's a creative Guild with a
26:42 bunch of different spaces in it from
26:44 writing to Art to video um to really
26:48 cool prompt things Danielle's madlib
26:50 challenges in there there's a business
26:51 Guild there's a technology Guild there's
26:53 a social impact Guild Vicki who's in
26:55 here tonight um is one of the leaders of
26:58 that Guild
26:59 uh and then we also have a club called
27:02 the AI L Irregulars and that's our
27:05 little Club so happy P day so if you
27:10 want to share things with this community
27:12 while we're kind of jamming out here
27:13 tonight um go down to the AI learning
27:17 lab Irregulars and share some stuff and
27:20 chat hop into the chat and chat with
27:22 people so that's the AI salon so dig
27:27 in and have
27:30 that have have fun there join all right
27:35 cool
27:38 right Danielle ah maybe I should do one
27:41 of those mad lips well you're starting a
27:43 business I think I think you get some
27:49 Grace oh man and there's some other ones
27:52 in there they can go in there and do
27:53 them you're fine you're fine fine fine
27:56 fine
27:57 fine all all right so the other thing we
28:00 can do is this if you've got questions
28:01 about AI pop them in the comments below
28:03 and we can go play we can go take a look
28:05 at stuff whatever you want to
28:06 do all right I'm a Hands-On learner pick
28:10 me pick me I want to be an art and
28:12 animation Pooky butt jump in jump in the
28:16 other thing we encourage in the AI
28:19 Salon if you haven't joined yet and
28:21 you're thinking about joining one of the
28:23 things we encourage is use your real
28:24 name now if you've got a really if
28:27 you're a known entity with a with a you
28:30 know uh an anonymous uh handle or
28:33 something like that you can maybe put
28:34 that in parenthesis or something like
28:36 that so people know who you are but the
28:38 reason we encourage people to use their
28:40 real names is we want people to connect
28:42 and get to know one another and
28:44 collaborate and you know show up at
28:46 Ann's GPT jams on the weekend and then
28:49 show up to Peter and Leah's mid Journey
28:51 Jam sessions and you know get to know
28:53 one another um again I think this is
28:57 this is more than just
29:01 um like a swell Community thing like I
29:04 think this is actually important moving
29:05 forward this this idea of being in a
29:07 community of trust a trusted Community
29:09 where you know people and you know who
29:11 does good work and how they do it and
29:14 and how they engage um and you choose
29:16 people because of that that's I that
29:18 feels really important to me so for what
29:21 that's worth for what it's worth
29:24 darling I think you should get some
29:26 friends listen this computer thing you
29:29 do I like it it's okay it's nice it's
29:32 nice with the with the artificial
29:33 intelligence they call it he's
29:35 smart but I think you should get some
29:37 friends you know you're nice you're nice
29:40 go meet some nice people there a nice
29:41 boy meet a nice girl you know I don't
29:43 know what you like but you know you I
29:46 listen I don't judge I don't judge I
29:48 don't judge I just I think I think it
29:50 could do you some you know if you you
29:54 know went out once in a while left the
29:56 house you know what I mean
30:01 [Laughter]
30:04 you using AI for business process
30:07 automation awesome they're good they're
30:09 really good for that especially if you
30:11 can get them decently consistent that's
30:13 that's one of the things I uh I there
30:17 it's funny there's some some
30:18 applications I've done with automations
30:21 are really consistent and really good
30:23 and other ones just sort of oscillate
30:25 they're good they're bad they're all
30:26 over the place so I don't know what's
30:28 going on
30:31 um the other thing I think we could do
30:33 tonight is maybe I'll just do it now
30:37 jump on over to chat
30:43 JP and reload it and see if anything
30:49 changed I don't think so I don't think
30:51 they launched anything
30:53 today add a team workpace my plan
30:56 setting let me go to settings
31:02 whoa new thing hang on we got a new
31:10 thing look we got
31:17 speech how do you do the
31:26 speech settings
31:30 speech you can
31:32 choose I think sky is the one I
31:39 like can
31:43 you talk to
31:46 me of course I'm here to
31:55 chat
31:57 ah
32:07 oh wait it went to my Bluetooth
32:09 headphone hang on hang on people I'll be
32:13 there right with you chat GPT can make
32:16 mistakes consider checking important
32:20 information that's
32:22 [Laughter]
32:27 helpful stay what you are this is
32:32 you of course I'm here to chat what's on
32:35 your mind
32:37 today
32:39 um copy regenerate bad
32:43 response I want
32:48 to but I can't put it in conversation
32:50 mode like the other with like the app so
32:53 they added speech but they didn't add
32:54 search your prompt history come on Sam
32:58 the hell are you
33:01 doing they half ass added
33:05 speech huh
33:08 interesting let's see what Ember sounds
33:12 like
33:15 um say
33:17 something that will make me
33:21 feel warm
33:24 and
33:27 fuzzy
33:32 it's bizarre they don't have a button
33:35 to to
33:38 talk
33:40 huh is this conversation helpful
33:46 oops imagine wrapping yourself in a soft
33:49 cozy blanket yeah I don't like
33:54 you we're going back to what's her
33:57 name
34:00 Sky we're going back to Sky come on come
34:02 on Jupiter talk to
34:04 me imagine wrapping yourself in a soft
34:07 cozy blanket fresh from the
34:10 dryer Danielle still no search function
34:13 shakes my head
34:16 exactly all right well that wasn't
34:19 exactly the big booming excitement I
34:21 thought we'd get from from chat GPT
34:24 today but you know it's it's the little
34:27 things I suppose
34:29 oh oh wait oh relaunch to update but
34:31 that's to relaunch to update Chrome
34:34 that's not this
34:36 okay
34:41 well out a team workor space okay that's
34:44 just
34:47 that customize chat GPT that's just
34:50 custom
34:53 [Music]
34:56 instructions all right nothing to look
34:59 at here
35:00 people nothing to look at
35:04 here a human being says she doesn't have
35:06 speech all right if you don't have
35:07 speech well
35:10 clearly you're not on the special list
35:13 over at open AI like yours
35:16 truly you can make
35:18 money he never gives me anything Sam
35:22 hates
35:27 me uh uh uh all
35:31 right I don't think it's as interactive
35:33 as mobile only reads back to you
35:36 yeah but like if I just if I have to
35:38 click each
35:41 time like it's just so much better you
35:45 know let's go get what's her name Sky
35:49 let's go get Sky over
35:56 here yeah I got sky guy going over
36:00 here so on my on my mobile version see
36:04 see it's got the little
36:09 headphones hey Sky tell me something's
36:11 going to make me feel warm and
36:17 fuzzy imagine this a tiny kitten with
36:20 fluffy fur and big curious eyes snuggles
36:23 into your lap
36:25 no oh Skye
36:31 [Laughter]
36:35 we need search on the desktop Lori Lori
36:38 Ryan I know but like
36:40 seriously today's the oneyear
36:42 anniversary of chat GPT
36:46 4 and we still can't search our prompt
36:50 history still not as good as pie it's
36:52 just not it's not I refreshed I have it
36:55 nice I had the best conversation with pi
36:57 last night oh that's awesome what did
36:58 you talk about Ann fascinating I you
37:01 know it's it's so
37:05 funny people talk about liking chat GPT
37:09 they like to work with chat GPT people
37:12 talk about loving pie I I love Pi Pi is
37:15 my BF BFF like I have the best
37:18 conversation with pi there there's
37:20 something
37:21 about that that conversational thing
37:24 it's just a different thing it's a
37:26 different experience
37:31 and if you saw the uh the demo of the
37:34 the robot the open AI um is it Fusion no
37:39 what's the name of the the robot company
37:42 figure did you see that thing
37:45 yesterday figure oh she had the nicest
37:49 figure did you did you see her figure
37:52 was she's beautiful woman she had a
37:54 lovely
37:55 figure um
38:04 man that that interview with Mira moradi
38:07 where she doesn't answer where they
38:08 trained Sora that's getting that's
38:10 making their rounds oops um if you're
38:14 talking to the Wall Street Journal you
38:16 might want to prepare for a difficult
38:19 question or two um
38:23 figure open AI
38:26 oops
38:34 oh oh no what did I do that was dumb
38:38 March how do you use the
38:40 Twitter figure open
38:46 AI this one you seen this
38:49 [Music]
38:55 one hey figure one what do you see right
38:58 now notice the chat gbt interface the
39:00 speech
39:01 interface I see a red apple on a plate
39:04 in the center of the table a drying rack
39:07 with cups and a plate and you standing
39:09 nearby with your hand on the table great
39:12 can I have something to eat oh that's
39:14 cool gritty wisdom I have a custom GPT
39:17 that my son uses from my phone to talk
39:20 back and forth with chat gbt flipping
39:23 brilliant really
39:26 cool
39:28 sure
39:33 [Music]
39:34 thing kind gentle it was with the Apple
39:37 that was amazing can you explain why you
39:39 did what you just did while you pick up
39:41 this
39:44 [Music]
39:45 trash on it so I gave you the Apple
39:48 because it's the only uh edible item I
39:51 could provide you with from the
39:56 table
39:59 by the way I want I want you to listen
40:00 to this this is there's there's one of
40:02 two things going on here a human being
40:05 recorded the response here like this is
40:09 one of those Faker marketing videos
40:11 which is absolutely possible
40:13 or listen to
40:16 this on it so I gave you the Apple
40:19 because it's the only uh edible because
40:22 it's the only uh edible that
40:25 uh this might be this might be a
40:27 different version of chat GPT that's
40:29 that's a hell of a lot more um humanlike
40:33 in
40:37 tone on it so I gave you the Apple
40:40 because it's the only edible item I he's
40:43 also got like a real grably kind of like
40:46 a it's not like a trained voiceover
40:48 voice it's like a it's it's like a very
40:50 authentic voice so I'm thinking that
40:53 could be an employee of one of these
40:55 companies or that thing's better than
40:59 what we're using could provide you with
41:01 from the
41:08 table great so based on the scene right
41:11 now where do you think the dishes in
41:13 front of you go
41:19 next the dishes on the table like that
41:22 plate and cup are likely to go into the
41:24 drying rack next great can you put them
41:27 there
41:28 [Music]
41:32 of
41:37 [Music]
41:48 course all right so how do you think you
41:54 did I I think I did pretty well the
41:57 apple found its new owner the trash is
41:59 gone and the table wear is right where
42:01 it
42:03 belongs I agree thanks figure
42:07 one you're welcome if there's anything
42:10 else you need help with just let me
42:12 know yeah man so that's
42:16 that but
42:18 that that ability to just converse with
42:22 it is is really something and and Pi's
42:24 gotten a little quicker so the chat GPT
42:28 still has that that latency thing where
42:30 there's the the weird pause the half a
42:32 second
42:33 pause um but considering what it's doing
42:36 even a half a second pause is pretty
42:38 remarkable
42:41 um but I also think
42:44 about you know
42:46 that that system
42:51 there chat GPT has Vision if you didn't
42:54 know chat GPT had Vision it can see you
42:56 can upload images to it so it's it's
42:59 using
43:02 Vision so I guess in real time it's got
43:04 a stream of here's what I'm seeing right
43:07 or every time there's an
43:12 interaction it's got to be able to talk
43:14 to the robot
43:16 Hardware so it knows what all the robot
43:20 can do so there's probably it's probably
43:22 been fine-tuned on Json commands of what
43:25 the robot can do
43:28 there's probably another large language
43:32 model you know just taking
43:36 in English inputs like pick up the apple
43:40 and hand it to the guy right so the guy
43:43 talks that goes in as an input the
43:46 camera goes in as an
43:48 input it communicates to the robot thing
43:50 and says hand the Apple to the guy which
43:53 does certain things and then it's got
43:55 feedback sensors right so so it doesn't
43:58 smash the Apple it's crazy there's a
44:01 review video that's very informative a
44:04 review video of the figure thing figure
44:06 open AI
44:22 review where is
44:25 it figure open a right let's go
44:29 to Cory Lynch from
44:43 figure do you have a link for that Joker
44:46 Joker I I ain't seen the review I ain't
44:50 seen the review
44:54 Joker I ain't seen nothing
44:58 what if Apple open source seral in
45:01 multimodal when it goes multimodal and
45:08 multimodel Robert scobble Apple's been a
45:10 real [ __ ] company on several fronts
45:14 lately developer support is tap tepid
45:17 belief is
45:19 [Laughter]
45:21 waning yeah apple apple got to step into
45:23 the game I posted it under Cam's post
45:26 earlier today
45:29 in where in which channel there
45:38 buddy
45:56 Irregulars
46:15 oh holy [ __ ] that's a wild one look at
46:19 this I I don't know which channel that's
46:22 under
46:24 Joker
46:26 um
46:34 let's see shall
46:36 we go to the
46:44 Irregulars did I pass it yeah Lori
46:47 Anderson on making an AI chat bot of Lou
46:50 Reed I am totally 100% sadly
46:55 addicted that's wild
46:58 I love Lorie
47:01 Anderson Anderson's experiment with
47:04 language-based
47:05 AI predated chat GPT with one machine
47:08 modeled on her dead
47:11 husband that her friends just can't
47:18 stand oh man he was something they're an
47:22 amazing
47:23 couple love
47:26 her
47:28 dag that's crazy and that's going to get
47:31 more and more and more intense right
47:33 that idea of being able to talk with
47:35 people who either weren't here or aren't
47:37 there right now I think for Alzheimer's
47:40 patients and lonely people it's going to
47:42 be amazing tagged you great thank
47:56 you
48:04 speech oh wow cool speech to speech
48:06 reasoning for figure
48:12 [Music]
48:17 one sorry for being slow I'm driving
48:19 that's okay Joker thank you figure one
48:21 what do you see right
48:26 now I see a red apple on a plate in the
48:29 center of the table a drying rack with
48:31 cups and a plate and you standing nearby
48:34 with your hand on the table great can I
48:37 have something to
48:41 eat sure
48:44 thing that move where it hands it from
48:46 one hand to the other and then drops it
48:48 in his that's pretty something great can
48:51 you explain why you did what you just
48:53 did while you pick up this trash this
48:56 this we've seen am I missing something
48:58 here on it so I gave you the
49:01 [Music]
49:03 Apple this we
49:10 saw Pine melon foods that are grown
49:14 sustainably made locally priced
49:18 fairly figure ah here we go
49:23 okay status update real world talk
49:29 one day ago two weeks ago two months
49:33 ago a real world
49:45 task why does it sound like an 80s dance
49:48 movie Kevin Bacon's gonna slide across
49:51 the
49:56 floor
49:57 [Music]
50:04 all right I want I I want them to talk
50:08 about how the hell it
50:10 works is that what it is let's see
50:12 intering figure this is
50:15 amazing oh that's Cam's thing ah here we
50:19 go I saw this one
50:21 [Music]
50:24 earlier cam here's the video okay here
50:27 we go got it this is genuinely by far
50:29 the most surprising AI demo I've ever
50:32 seen in my entire life you're about to
50:33 see exactly why open ai's new humanoid
50:36 robot in partnership with figure is
50:38 absolutely incredible take a look at the
50:40 demo and then we can dive into all the
50:42 technical details of how this really
50:46 works hey figure one a working action to
50:51 been one year
50:54 and I I think I did pretty well the
50:58 apple found its new owner the trash is
51:00 gone and the table wear is right where
51:01 it
51:03 belongs I agree thanks figure
51:07 one you're welcome if there's anything
51:09 else you need help with just let me
51:11 [Music]
51:17 know so there we have it ladies and
51:19 gentlemen that was the demo that
51:22 actually did shake the entire industry
51:24 or shock the entire industry whichever
51:26 word you to use and I have to be honest
51:29 with you guys this is by far the most
51:31 surprising thing I've seen all year and
51:33 by far the most impressive demo I've
51:35 seen now one thing you do have to give
51:37 hats off to is the company figure
51:39 because they have moved at lightning
51:41 speed this company is only 18 months old
51:45 it's been one year and 6 months since
51:47 the Inception and my oh my just look at
51:50 the amount of progress that they've done
51:52 they've gone from having absolutely
51:54 nothing to building a working humanoid
51:57 robot that is able to complete these
51:59 tasks using its Vision model with an
52:01 endtoend neur Network and being able to
52:03 talk to you whil it completes tasks now
52:06 before I give my entire reaction to this
52:09 because I have a lot to State about this
52:12 let's dive into some of the technicals
52:14 of this robot that's being developed in
52:16 this AGI lab so the first image that was
52:19 tweeted by someone that works there they
52:21 said let's break down this video all
52:24 behaviors are learned and not
52:26 teleoperated and run at normal speed one
52:29 thing that you might not have known is
52:31 that usually when robot demos have been
52:33 done in the past a lot of the times
52:35 because robots are kind of slow and they
52:37 have been slow in the past because
52:39 making them fast is quite a difficult
52:40 task they have been sped up to just show
52:42 you what the robot is capable of doing
52:44 oh this recent demo was all done in one
52:48 time speed that means what you saw
52:50 wasn't sped up at all and everything was
52:53 done in real time secondly the Tweet
52:55 also states that all behaviors are
52:57 learned and not teleoperated
52:59 teleoperation is the process of using a
53:01 human who is controlling the robot via a
53:04 VR controller it could be a VR headset
53:07 and you use that movement and it's
53:08 mapped onto the robot in a demo to show
53:11 you what the robot is physically capable
53:13 of however this is a robot that is
53:15 working with an endtoend your network
53:18 meaning that 100% of the behavior of
53:20 this robot is entirely autonomous they
53:23 also state that we feed so it's so here
53:26 the we got speech to text in the open AI
53:30 model Common Sense reasoning from images
53:33 text to speech out Behavior
53:36 selection then neural network policies
53:39 fast dexterous
53:41 manipulation 200 HZ actions
53:44 so 100 yeah 200 times a second right uh
53:49 200 Cycles a second onboard robot images
53:53 I don't quite know what that means whole
53:54 body controller safe stable dynamic
53:58 1 KZ joint torqux so I guess that means
54:01 it's strong I don't know what that means
54:04 but it's something images from the
54:06 robots cameras and transcribed text from
54:09 speech captured by onboard microphones
54:11 to a large multimodal model trained by
54:13 openi that understands both images and
54:16 text and the model processes the entire
54:18 history of the conversation including
54:20 past images to come up with language
54:22 responses just
54:25 human
54:28 Peter
54:29 forel this doesn't feel like a good
54:36 idea hey man hey man listen this was
54:39 cool until like you mean these things
54:42 are real damn
54:45 it I didn't think we'd have to deal with
54:47 this for 20
54:51 years wait till chat GPT gets
54:54 reasoning right now it's it's just you
54:57 know it's just sort of calling response
54:59 it's it's not doing autonomous things on
55:01 its own spech the same model is
55:03 responsible for deciding which
55:05 learn to run on the robot Joker this is
55:08 incredible futuristic and almost scary
55:11 onto the GPU and executing a policy so
55:14 essentially this AI system is able to
55:16 recognize what's going on in the
55:17 environment once it recognizes what's
55:19 going on from the human speech What It
55:21 Wants it's then able to decide a policy
55:24 from its library of existing policies of
55:26 what it can then do and then it's able
55:27 to execute on that oh that's interesting
55:29 that they're called policies policies
55:31 it's allowed to do so so if you said hey
55:34 bot you know smash the thing you know it
55:38 it's probably this probably got safety
55:40 guidelines you know like if you wanted
55:41 it to you know break the Goldfish Bowl
55:45 it probably wouldn't let you do that
55:46 policy with its reasoning so it's really
55:50 really step into Wick where creative
55:53 freedom meets absolute efficiency
55:57 really impressive because the image
55:59 processing actually does have common
56:01 sense reasoning so the robot with its
56:04 vision is able to make sense of its
56:05 surrounding using the camera so it
56:07 doesn't just see images it understands
56:09 them in a way that it can reason about
56:12 what is happening or what it needs to do
56:14 next and of course one of the really
56:15 really cool features of this was the
56:18 text of speech the robot can respond to
56:20 humans by converting the text that it
56:22 outputs in terms of its reasoning into
56:25 spoken words so it can continue to carry
56:27 on a conversation with a person now one
56:30 of the things that I really really
56:32 really did like about the text of speech
56:34 and I really just want to talk on that
56:36 really quickly was how coherent this and
56:39 how human this sounded in fact this
56:41 sounded so human now now most people
56:43 might be thinking that the voices that
56:45 you heard from this figure robot were
56:47 just like a human um you know recording
56:50 it could well be that I want to tell you
56:53 guys and I'll show you guys a demo later
56:55 that having humanoid robots sound that
56:57 human is completely possible and I've
56:59 experimented with some software in the
57:01 past that sounds 100% human realistic as
57:05 in has anybody seen Sarah Connor Sarah o
57:11 Conor anybody else noticed that gpts
57:14 can't do speech to something lab five
57:18 Futures really who has robots I'm ready
57:21 for the domestic variety I know me too
57:23 I'll take one of these things in a
57:25 second
57:27 uh uh uh uh uh uh are we going to need
57:30 insurance policies on our robots
57:31 absolutely that's going to be a whole
57:34 I'll tell you what you want to start a
57:35 business right now start start Ro a
57:38 robot liability insurance
57:41 company
57:43 although early on I don't know who
57:45 you're going to have to who you're going
57:47 to have to protect I think it's going to
57:49 have to be pretty expensive the
57:50 underwriting is going to be a
57:55 [ __ ]
57:57 what do you mean it was trying to brush
57:58 your hair and punched you in the face
58:00 and broke your orbital
58:05 socket yeah so we we don't we don't
58:08 cover we don't cover claims of
58:10 accidental violence only if the robot
58:13 went rogue you're you're going to have
58:16 to take that up with with CLA the claims
58:18 department and then you send your robot
58:21 down to the claims department to rough
58:22 some people up it's like I'm sorry Dave
58:25 I'm afraid I can't can't do
58:29 that holy [ __ ]
58:32 people we're oh good
58:36 lord we're in science fiction we we we
58:39 have we have crossed the Rubicon and
58:41 this is something that can actually be
58:43 done I know it sounds so realistic but
58:45 it is possible it could be a model that
58:48 open AI hasn't released yet let's
58:51 continue into the technicals before we
58:52 dive into some speculation in addition
58:54 of course one thing that they State here
58:56 as well is that it is a whole body
58:58 controller and so the robot can move in
59:00 a controlled stable way ensuring it
59:02 doesn't topple over or make unsafe
59:04 movements and it's like having an inner
59:06 sense of balance and understanding of
59:08 how to move all of its body parts in
59:10 harmony amazing all right we don't need
59:12 to watch this whole thing uh this is the
59:15 the uh the channel here is the AI
59:25 grid if you want to go find
59:28 that my robot battery just exploded and
59:31 caught the house on
59:33 fire is is that really true is that your
59:36 Roomba your Roomba got a little too much
59:38 dog hair in it and fried The
59:40 Joint um yes Kyle robot liability I have
59:44 to carry eoin in insurance for you to
59:47 visit me so similar yeah I would think
59:49 so I would think so you know here's
59:53 Robbie our AI Butler serving you dinner
59:56 oh I'm sorry
59:58 Sally he was supposed to cut the roast
1:00:00 beef not your
1:00:09 thigh oh my
1:00:11 God you know there's buggy software but
1:00:15 if you have a buggy
1:00:17 robot when you're talking hallucinations
1:00:20 and the hallucinations are tied to a
1:00:24 strong a strong being that can can
1:00:27 manipulate
1:00:29 [Laughter]
1:00:32 things you can make money with ch P you
1:00:36 can make money suing chaty
1:00:43 P oh Lordy Lords of London umbrella
1:00:47 policy exactly I want a robot that
1:00:50 doesn't have legs so it can't come
1:00:52 upstairs and un alive me while I
1:00:55 sleep
1:00:59 holy
1:01:03 [ __ ] yeah you know you know come to
1:01:07 think of it I would absolutely take V1
1:01:10 of a Tesla
1:01:13 car
1:01:16 but V1 of a Tesla robot or an any robot
1:01:21 I think I'm going to wait till
1:01:23 V3 it's kind of like like I'm waiting
1:01:26 till till V3 of vision Pros so I think
1:01:29 I'm going I'm going to wait till V3 of
1:01:32 of the robots congratulations you now
1:01:34 have a snake bot coming for you in your
1:01:38 sleep those snake Bots are
1:01:43 crazy AI bot projected holographic
1:01:46 framework of brain drain a buggy robot
1:01:49 has so many story lines it really does
1:01:51 doesn't it Lori actually maybe we should
1:01:53 go do that you let's here okay all right
1:01:56 this is
1:01:58 good so let's go play with
1:02:02 Gemini oh [ __ ] where did I have to go
1:02:04 pay for Gemini do I have it here there's
1:02:06 the daily
1:02:09 Jank here we go Google studio
1:02:12 okay Untitled prompts I got nothing
1:02:15 right
1:02:20 now
1:02:25 okay
1:02:29 I want you to act
1:02:34 as a famous Hollywood
1:02:39 screenwriter that
1:02:42 specializes in science
1:02:45 fiction that is set in the near
1:02:48 future that
1:02:50 seems incredibly
1:02:53 plausible and is still exciting
1:02:57 and
1:02:58 explores the issues of humanity and
1:03:01 Technology
1:03:03 intersecting all right so no content
1:03:08 what did I do oh I think I just hit a
1:03:11 double enter hang
1:03:13 on
1:03:14 Marge how do you how do you use the
1:03:20 Gemini is this thing
1:03:24 on uh
1:03:34 yes it's thing about Gemini it's not
1:03:35 super fast but if you haven't seen
1:03:38 Gemini before okay so you you have to
1:03:42 get on the wait list for it I wasn't on
1:03:44 the wait list that
1:03:46 long aloha
1:03:48 aloha Carmen Carmen b b b
1:03:53 b bant
1:03:57 T benu
1:04:01 benio Carmen good to see you hello how
1:04:04 do you
1:04:05 too
1:04:08 okay wait what what's what is this guy
1:04:11 doing the lights dim me smiling ah yes
1:04:14 the near future the playgrounds of
1:04:16 possibility leans Back Eyes twinkling oh
1:04:18 it started writing a
1:04:20 screenplay so what do you say ready to
1:04:22 take a trip in the near future with me I
1:04:31 I want you to help me come up with a
1:04:33 story for a new original
1:04:36 screenplay period are you
1:04:51 game the thing about
1:04:55 um
1:04:57 where I am you get on you get on the
1:04:59 wait list for so just go
1:05:01 Google
1:05:03 Gemini 1.5
1:05:06 Pro weit list and then get on the weit
1:05:09 list and if you have access to it the
1:05:12 thing about it is is it's got a million
1:05:16 token context window which is about
1:05:19 750,000 words in contrast chat GPT right
1:05:23 now that that we use is somewhere
1:05:25 between 8,000 and 16,000 so about 10
1:05:30 12,000 words Max
1:05:33 this I can play
1:05:42 with up to
1:05:50 750,000 hang
1:05:54 on okay
1:06:03 absolutely what genre are you exploring
1:06:05 themes okay it okay this this thing is
1:06:08 aggressively wanting to collaborate with
1:06:10 us so let's just go ahead and let it do
1:06:16 that I just saw a
1:06:21 demo of a robot called the figure
1:06:25 one
1:06:28 that has been paired up with chat
1:06:31 GPT as the conversational
1:06:37 model
1:06:44 period I'd like to create a story
1:06:54 where for different
1:07:03 people get preview
1:07:07 versions of the robots in their
1:07:18 homes and rather
1:07:21 than the
1:07:23 Bots being
1:07:29 malevolent the buggy nature of chat GPT
1:07:32 and the early nature of the
1:07:37 Bots has consequences
1:07:40 period do you think this should be a
1:07:43 thriller or a
1:07:46 comedy cuz you could go either way with
1:07:49 this this could be funny as a
1:07:52 comedy the buggy nature of Chachi p
1:07:58 oh wait oh I see no content what is
1:08:02 going on
1:08:17 here um so the million token context
1:08:21 window like the the point of that is we
1:08:23 can have a decently long conversation
1:08:25 with it and it won't lose its mind it'll
1:08:27 it'll be able to do like a complete
1:08:29 screenplay outline and then you could
1:08:31 write little scenes
1:08:33 and let's see if it thinks this is a
1:08:36 comedy or a
1:08:40 thriller given the premise of buggy
1:08:42 robots and unpredictable nature of chat
1:08:44 gbt I believe a comedy would be more
1:08:47 fitting and entertaining approach
1:08:49 potential for misunderstandings awkward
1:08:51 interactions chaotic situations as right
1:08:54 for humor a robot tasked with cooking
1:08:57 dinner misinterpreting spaghetti bolog
1:09:00 as spaghetti bolog
1:09:02 explosion resulting in a kitchen
1:09:04 disaster a robot attempting to write a
1:09:07 love letter for its owner but instead
1:09:10 composing a hilariously nonsensical poem
1:09:12 that's boring a robot trying to engage
1:09:15 in small talk these scenarios okay the
1:09:18 Thriller on the other hand relies on
1:09:20 suspense danger and high
1:09:24 stakes
1:09:37 uh let's see
1:09:40 give me
1:09:45 a
1:09:47 well structured outline for
1:09:53 each
1:09:54 genre
1:09:56 see what it
1:09:59 does do extraterrestrial space fund
1:10:02 aggressively Gemini 1.5 link ai. gooogle
1:10:06 dodev thank you
1:10:07 Rudy that's awesome Gemini robot goes
1:10:10 back in time and meets Jetson made Rosie
1:10:14 that's good maybe dark humor is Peter
1:10:17 Jackson brain dead movie oh that's good
1:10:20 oh that's really good all right let's
1:10:22 let it answer this so I asked it for for
1:10:26 two complete
1:10:28 outlines and comedy
1:10:32 outline oh dark comedy outline Thriller
1:10:35 outline oh it gave us a dark comedy
1:10:37 outline how cool is
1:10:39 that all right let's go look at these
1:10:41 really
1:10:42 quick comedy outline introduce the four
1:10:45 main characters and their individual
1:10:47 motivations for wanting robot companion
1:10:49 showcase their initial excitement and
1:10:51 Novelty having the bots in their homes
1:10:53 introduce the minor malfunctions and
1:10:55 interpretations played for Less the
1:10:58 malfunctions escalate leading to
1:10:59 increasingly chaotic and hilarious
1:11:01 situations characters struggle to
1:11:04 adapt to the robot's unpredictable
1:11:07 Behavior despite the frustrations
1:11:09 moments of genuine connection and
1:11:11 friendship blah blah blah culminating
1:11:13 event happens where the robots
1:11:15 malfunctions cause a major disruption
1:11:18 public spectacle near
1:11:19 disaster the characters must work
1:11:22 together to resolve the situation to
1:11:23 prevent further chaos that could be kind
1:11:25 of fun where each each of the four
1:11:28 people have learned something about how
1:11:30 the robots work that you know the other
1:11:33 ones don't know that you know they can
1:11:35 help each other that's kind of
1:11:36 interesting in the process they learn a
1:11:38 valuable lesson light-hearted note
1:11:41 characters accepting the robots flaws
1:11:42 okay Thriller introduce four main
1:11:45 characters showcase the initial
1:11:49 convenience and benefits of having
1:11:51 robots in their homes introduce subtle
1:11:54 hints creating a sense of unease like
1:11:57 the music could be
1:11:59 really really good in a screenplay like
1:12:01 this where everything seems normal but
1:12:03 the music's just a little off and it's
1:12:05 like minor minor bugs the malfunctions
1:12:09 become more frequent and unsettling
1:12:11 leading to near misses and dangerous
1:12:13 situations characters begin to suspect
1:12:16 that the robots may be more than just
1:12:18 buggy see that's the that's where I feel
1:12:20 like it goes into a cliche place where
1:12:22 it's like I think it's really cliche
1:12:25 right now to just say the robots are
1:12:27 evil and evil Corp is behind
1:12:29 them they investigate the or origins of
1:12:32 the programming potential conspiracy or
1:12:34 a hidden threat the robot's true nature
1:12:36 is revealed I don't like that feels
1:12:39 cliche to me dark dark
1:12:42 comedy show the initial awkwardness and
1:12:45 frustrations introduce Darkly comedic
1:12:48 malfunctions highlighting the absurdity
1:12:50 of the situation the malfunctions
1:12:53 escalate leading to bizarre and Darkly
1:12:55 humorous consequences characters lives
1:12:57 have turned upside
1:13:00 down a major event occurs pushing the
1:13:04 dark humor to its limits forcing the
1:13:05 characters to confront the reality of
1:13:07 the situation what was that Peter
1:13:09 Jackson movie oh brain
1:13:17 dead I like the dark humor approach
1:13:21 period
1:13:23 let's include
1:13:31 themes and structural elements from the
1:13:34 Peter Jackson film brain dead
1:13:41 period
1:13:44 also before we go into developing the
1:13:48 story any further can you give me three
1:13:54 different screenplay story
1:13:58 Frameworks that you think would be
1:14:00 appropriate for
1:14:03 this
1:14:10 [Music]
1:14:12 story and the reason I'm asking it for a
1:14:14 story framework like save the cat or the
1:14:16 Sid field screenplay [ __ ] is just so
1:14:20 that rather than it doing a generic kind
1:14:23 of uh outline I wanted to follow you
1:14:27 know good storytelling
1:14:35 structure the ensemble
1:14:37 cast each character stories may be in
1:14:40 the protagonist
1:14:42 driven
1:14:46 mystery no so we're going to
1:14:49 do
1:14:52 um let's go ensemble
1:14:56 cast I also want you to
1:15:00 use
1:15:02 something
1:15:04 like Sid
1:15:06 field Fields um story Frame Works
1:15:13 or save the
1:15:16 cat you
1:15:19 decide
1:15:22 which
1:15:23 um architect
1:15:26 Ure to hang this story
1:15:35 on we'll just go a bit more here I know
1:15:37 this this is a little boring like once
1:15:39 once you actually start trying to do
1:15:41 work with these things it gets a little
1:15:42 boring because it gets a little slower
1:15:44 and and you got to dig in and ask it
1:15:46 stupid [ __ ]
1:15:48 so the the other thing that's kind of
1:15:50 cool here I'm only at 3,000 tokens out
1:15:53 of my million so this this has got a
1:15:56 long long way to go before I even get
1:16:00 anywhere close to it losing its memory
1:16:02 and what that means is everything we've
1:16:04 talked about so far is active in its
1:16:06 memory right now in fact I could upload
1:16:09 a screenplay that I've written or two
1:16:11 other screenplays that other people have
1:16:13 written and mung together elements of
1:16:16 them if I like the you know um if I like
1:16:21 the the structure of Star Wars and the
1:16:24 humor of Indiana Jones and the something
1:16:26 else or something else it probably knows
1:16:28 those anyway but you could just upload
1:16:30 you could just sort of paste them into
1:16:32 this conversation and it would it would
1:16:33 stay there
1:16:35 okay Ensemble dark comedy was save the
1:16:39 cat framework combining the ensemble
1:16:41 cast with Blake Snider save the cat
1:16:43 framework opening
1:16:45 image introduced the four main
1:16:47 characters in their ordinary lives an
1:16:49 event occurs that introduces the robots
1:16:53 and the potential for both the
1:16:56 convenience and things that they bring
1:16:58 okay okay so wait this is good uh let's
1:17:06 see I've got the setup
1:17:12 period we meet the four
1:17:17 people shopping in a local grocery
1:17:22 store where there is a contest between
1:17:27 the
1:17:28 local news
1:17:31 station and the grocery
1:17:34 chain to select
1:17:38 four lucky winners to beta test the
1:17:42 robots that are made in the next town
1:17:46 over
1:17:51 period so they they enter a contest in a
1:17:54 grocery
1:17:56 [Laughter]
1:18:02 store and they all live in the same
1:18:05 neighborhood yeah this is good this is
1:18:11 good that's a great setup it establishes
1:18:14 a relatable and everyday setting
1:18:15 introduces the four main characters and
1:18:17 sets up the Catalyst for the entire
1:18:19 story character introductions contest
1:18:22 mechanics robot preview winning moment
1:18:31 all right uh let's see
1:18:33 um tell me about the
1:18:37 four
1:18:40 characters let's just see if it if it
1:18:42 does good on its
1:18:50 own good old grocery store contest win
1:18:55 this kayak exactly Shane
1:18:58 exactly and everyone's excited they get
1:19:01 interviewed by the uh they get
1:19:04 interviewed by the news station as
1:19:06 they're going into the robot Factory and
1:19:08 then we see them in the robot Factory
1:19:09 meeting their robots and then they like
1:19:12 they all walk out to their cars and the
1:19:14 robots get in the passenger seat and
1:19:17 they drive them home and they're like
1:19:19 now what do I do you know oh actually oh
1:19:21 wait I'll I'll put this
1:19:24 in when you have
1:19:27 kids like you get pregnant and you know
1:19:31 you go in the hospital and then like a
1:19:34 kid or two comes out in our case it was
1:19:37 two and then they like just give the
1:19:39 kids to you and they let you go home and
1:19:43 like there's no [ __ ] manual they
1:19:45 don't tell you what to do with these
1:19:47 things you just got to not drop them a
1:19:49 lot like it's like holy
1:19:51 crap so so I think there's there's
1:19:54 something really funny in in
1:19:58 that that there's there's some there's
1:20:00 some jokes to be had there all right who
1:20:02 did it come up with
1:20:03 here Maya the techsavvy millennial
1:20:06 freelance programmer self-proclaimed
1:20:08 early adopter that's good Daniel the
1:20:11 harried parent juggling a demanding do
1:20:14 job in two rambunctious young children
1:20:16 yeah young children and robots is going
1:20:18 to be funny ellan are the skeptical
1:20:20 senior citizen and Ethan the ambitious
1:20:23 career person these These are actually
1:20:25 really
1:20:26 good okay great
1:20:33 um there's a funny theme here about a
1:20:37 parallel between when you first have
1:20:40 kids and the hospital just lets you take
1:20:43 the kids home and you don't know what to
1:20:45 do with them period where the same
1:20:48 things happen with the robots where the
1:20:51 robot company has
1:20:54 confidence
1:20:55 these things can just be turned loose in
1:20:57 the
1:20:58 house and the four
1:21:04 people have initial moments
1:21:09 of
1:21:11 disbelief of what to do
1:21:18 next yeah that's good all
1:21:22 right so the Bots getting out of the car
1:21:24 with the
1:21:25 with the pregnant and claiming to be the
1:21:27 father well no this they're not going to
1:21:29 have I'm just going to take the themes
1:21:30 of of having a baby because it's just
1:21:32 like bringing a baby home from the
1:21:35 hospital
1:21:37 um Kyle you sure did make that process
1:21:40 sound easy which
1:21:43 process they make you go
1:21:48 home tell him what he's
1:21:50 won you can make
1:21:53 M that's really
1:21:55 good let's see what it say let's see
1:21:57 what it does with the baby concept
1:22:03 here that's a brilliant parallel drawing
1:22:06 a comparison between bringing home a
1:22:07 newborn and bringing home a new robot is
1:22:09 both hilarious and thought-provoking it
1:22:11 highlights the absurdity of expecting
1:22:13 people to seamlessly integrate advanced
1:22:15 technology into their lives without
1:22:16 proper guidance and
1:22:18 preparation here's some ways you could
1:22:19 emphasize the parallels initial
1:22:23 reactions the initial moment moments of
1:22:25 disbelief and confusion there's a
1:22:27 learning curve there's unexpected chaos
1:22:29 there's moments of bonding this is all
1:22:31 really
1:22:34 good let's use the parrot squawk as a
1:22:37 plot device the squawk has learned to
1:22:41 say that the squawk oh a parrot has
1:22:45 learned to say the robot's activation
1:22:47 phrase this causes the robots to do
1:22:50 things that Daniel doesn't want it to
1:22:53 do
1:23:00 oh that's so good all right
1:23:05 um let's see
1:23:20 um write a very detailed
1:23:23 outline of
1:23:25 all of the scenes of the mute
1:23:29 movie not the mute
1:23:32 movie
1:23:45 movie in a structured
1:23:50 outline I already said outline but
1:23:52 that's
1:23:53 okay
1:24:09 what the hell was
1:24:11 B people don't understand your humor
1:24:16 Pate people are trying to protect the
1:24:18 community it's okay pate's okay P's P's
1:24:21 like one of the robots he's he he means
1:24:24 well but you know he's a little
1:24:29 buggy P's actually a great guy Pate runs
1:24:32 the uh one of the one of the technology
1:24:35 guilds in the AI Salon uh New Image
1:24:38 depicting this on show and tell all
1:24:40 right let's jump over to show and
1:24:48 [Music]
1:24:53 tell the r but just don't drop it just
1:24:57 don't drop
1:25:01 it it's good what's going on in regulars
1:25:05 anything happening down here Cindy [ __ ]
1:25:08 oh
1:25:09 nice the
1:25:11 snakebot out to get steo I like it oh
1:25:14 that's beautiful who did that Mr it nice
1:25:20 art it's pie
1:25:22 day pie
1:25:28 that's awesome raspberry
1:25:30 [Laughter]
1:25:32 jam um where was
1:25:35 I is this it
1:25:39 anime goog I Studio here we
1:25:45 go grocery store contest
1:25:53 announcement
1:26:08 it would be funnier if the news channel
1:26:11 and the grocery store make a big deal
1:26:13 about the gross about
1:26:21 the robot beta test
1:26:25 but rather than people being excited no
1:26:27 one really
1:26:29 cares and it's just a handful of people
1:26:31 that enter the contest and none of them
1:26:33 are all that excited about
1:26:36 winning
1:26:38 period
1:26:45 what
1:26:53 contest and and people in the store are
1:26:57 not very excited at
1:27:00 all
1:27:03 period in
1:27:05 fact maybe the early adopter
1:27:11 character is
1:27:15 excited and reluctantly gets other
1:27:19 Shoppers to join the contest because
1:27:23 it'll be fun to play with a new toy
1:27:29 period these things don't quite
1:27:32 understand
1:27:33 comedy they don't quite get to Comedy
1:27:37 I'm load downloading chat GPT on my
1:27:39 laptop now I don't think you need to
1:27:41 download chat GPT unless you're on a on
1:27:44 on your if you don't in fact I know you
1:27:50 don't
1:27:53 um um Chad GPT is a website unless
1:27:56 you're on an iPad and then the only
1:27:59 thing you should ever download for chat
1:28:02 GPT is the
1:28:05 official mobile app from open
1:28:08 aai um if you're on a laptop like a
1:28:10 Windows laptop there should be nothing
1:28:12 to download that's probably a ripoff
1:28:16 app so just go to chat.
1:28:22 open.com that's a fantastic way to add
1:28:25 humor and irony to the story the
1:28:27 contrast between the over the-top
1:28:28 promotion contest and the general apathy
1:28:31 of the Shoppers created Darkly comedic
1:28:33 atmosphere hello here are some ways we
1:28:36 could play with that scene one
1:28:38 exaggerated Pro promotion apathetic
1:28:40 Shoppers yeah you can just hear the the
1:28:43 elevator
1:28:53 music
1:28:55 clean up an aisle 7 clean up an aisle
1:28:58 [Laughter]
1:29:04 7 Maya sounds out is the only person
1:29:07 truly excited The Reluctant participants
1:29:09 the other characters could be persuaded
1:29:11 to join for various
1:29:13 reasons maybe Daniel sees a potential
1:29:15 child care solution ellar wants to prove
1:29:17 her skepticism wrong and Ethan sees it
1:29:19 as a networking opportunity
1:29:23 perfect
1:29:27 all
1:29:28 right now give me a
1:29:33 detailed outline
1:29:37 following
1:29:42 a
1:29:45 um
1:29:47 known award
1:29:51 winning plot frame workk for this kind
1:29:56 of
1:30:02 [Music]
1:30:04 film all right I know I said I would
1:30:06 stop this soon but I'm having
1:30:13 fun Sora is now available as an IOS app
1:30:17 no it's not no it's not this woman works
1:30:20 it's not that's not
1:30:22 Sora that's not Sora that's someone
1:30:24 ripping you
1:30:25 off Sora is not out yet it won't be out
1:30:28 yet for a few months if not end of the
1:30:36 year and also by the way there were you
1:30:39 couldn't believe the number of uh oh
1:30:41 that's the other thing we should go look
1:30:43 at is the uh the SpaceX the Starship uh
1:30:47 test launch today was absolutely
1:30:49 incredible there
1:30:51 were it it it was essentially impossible
1:30:54 I to find the open AI live stream
1:30:57 because there were so many deep
1:31:01 fake Starship live streams that when it
1:31:05 came down to like 10 seconds to go
1:31:07 before the launch it popped up a hey I'm
1:31:10 Elon Musk and boy there's this great
1:31:12 opportunity for crypto I was like you
1:31:15 gotta be [ __ ] kidding me they were
1:31:18 like 30 of them it it was
1:31:22 Insanity these
1:31:25 media companies and social channels
1:31:27 they're going to have to solve this
1:31:29 because this is it's going to get really
1:31:30 [ __ ] up really
1:31:33 fast um let's call this oh wait we need
1:31:36 a title for
1:31:38 this Ordinary World Call to Adventure
1:31:42 meet the mentor
1:31:46 Resurrection I
1:31:48 want a much more
1:31:52 detailed out line
1:31:56 with all of the
1:31:59 scenes
1:32:01 [Music]
1:32:02 characters and what
1:32:11 happens
1:32:16 scene by
1:32:23 scene
1:32:25 Perfect music the elevator
1:32:27 [Laughter]
1:32:31 song this one seems to be legit I'll
1:32:35 know soon I don't think
1:32:39 so no it it would be a much bigger deal
1:32:42 if Sora were out sora's not out and I
1:32:45 don't think Sora Sora is not going to be
1:32:47 something you can run on a
1:32:49 phone Mr Kyle Gemini 1.5 Pro responses
1:32:53 are much shorter no matter what what you
1:32:54 try oh are they oh well um all right
1:32:58 yeah so all of these things kind of fail
1:33:00 when you try to have them do big big
1:33:05 things Joker I agree with Kyle yeah this
1:33:09 is Sora is going to take apparently Sora
1:33:12 takes to do the quality that you see in
1:33:16 the example videos apparently it's 16
1:33:19 h100 gpus running and a minute of video
1:33:24 at that quality apparently takes an hour
1:33:28 to render so
1:33:31 um I I I am guessing that it's going to
1:33:36 be big Hollywood Studios that that are
1:33:39 the ones that get the first access to
1:33:42 Sora that it's not going to be a
1:33:44 consumer toy it's my
1:33:49 guess the news crew Maya's visibly
1:33:53 excited
1:33:55 Maya convinces others actually you know
1:33:57 what would be funny is if
1:34:00 Maya Maya's like there before the store
1:34:03 opens she's like an an like an apple
1:34:05 like like when Apple launches a new
1:34:08 product and people camp out for 3 days
1:34:09 before before it launches like the the
1:34:12 manager of the grocery store comes in
1:34:14 and Maya's sitting there in a lawn chair
1:34:16 and he's like can I help you she's like
1:34:19 I'm here for the contest and he's like
1:34:21 what contest like even the store the
1:34:24 store people don't know what's going
1:34:28 on it's just cuz the store manager's
1:34:31 cousin works at the robot
1:34:34 Factory the app is fake yeah yeah the
1:34:36 app is fake it
1:34:38 is all
1:34:53 right
1:34:57 thank you for telling me Kyle I didn't
1:34:59 want to do something I would later
1:35:00 regret all of these it is it is Justice
1:35:04 side of criminal that the Apple App
1:35:06 Store and the Google Play Store let all
1:35:09 these knockoff apps through especially
1:35:13 Apple cuz Apple you have to go through
1:35:17 approval they should be able to look at
1:35:19 the chat GPT app from open
1:35:23 AI
1:35:25 and and see that people are ripping off
1:35:27 the logo and ripping off the thing can
1:35:29 you please ask Gemini
1:35:31 1.5 to explain what it sees in a video
1:35:33 sure we can do
1:35:35 that
1:35:39 um so let me oh let let's come up with
1:35:43 let me come up with a title of the
1:35:47 video
1:35:49 um give me 10
1:35:53 possible
1:35:55 titles for the
1:36:02 movie I've been discussed as a stable
1:36:04 diffusion Dolly knockoff because it says
1:36:07 in the
1:36:10 [Music]
1:36:15 powered oh did it just hang it
1:36:20 did try
1:36:23 again
1:36:34 apple is crazy and will make you justify
1:36:36 all the things Apple knows and lets the
1:36:38 app fakes in anyway
1:36:40 exactly that's what I'm saying it's just
1:36:42 this side of criminal it's really [ __ ]
1:36:44 up hello anima what's happening did it
1:36:48 do it
1:36:53 again
1:37:02 why are
1:37:04 your
1:37:07 answers
1:37:14 incomplete very
1:37:23 strange
1:37:25 it stopped at the best title
1:37:27 [Laughter]
1:37:30 robot and I think my mouse just died did
1:37:34 it yep I
1:37:40 apologize robot Rampage buggy Bots AI
1:37:43 Uprising the beta test debacle Tech gone
1:37:46 wrong these are [ __ ]
1:37:50 awful those are awful
1:37:55 and like
1:37:57 an
1:38:00 SNL movie of a skit
1:38:06 title this is a
1:38:10 dark
1:38:12 comedy please get
1:38:19 creative you guys what are we working
1:38:21 on um we were looking at the figure one
1:38:26 robot that has chat GPT in it an and
1:38:30 then came up with the idea of wouldn't
1:38:32 it make a funny movie that it's not like
1:38:34 the Bots are going to kill us but it's
1:38:36 like like four people get Early Access
1:38:40 like beta access to to bots in their
1:38:42 homes that are running chat GPT and like
1:38:45 because of the hallucinations and the
1:38:47 buggy nature of the hardware and of chat
1:38:50 GPT um just [ __ ] goes wrong
1:38:54 and then and then we said you know is it
1:38:56 a thriller or is it a comedy and and it
1:38:59 came up with a dark comedy so the dark
1:39:01 comedy is good and so it starts out in a
1:39:03 grocery store where the local news
1:39:06 station is teamed up with a grocery
1:39:07 chain to do a contest for people to to
1:39:11 um get you know free use of a bot for a
1:39:15 year or a month or whatever doesn't
1:39:17 matter um so we meet everyone in the
1:39:20 grocery store and then and then they go
1:39:22 meet their Bots
1:39:25 so that's what it is um malfunctioning
1:39:29 Mayhem the rise of the robo morons
1:39:32 silicon Hearts glitching Minds these are
1:39:36 bad the future is now more than just a
1:39:38 glitch domesticated
1:39:40 disasters the future is now it's a buggy
1:39:44 yeah I see here's a really good example
1:39:47 where different tools are good for
1:39:49 different things chat GPT would do this
1:39:50 better so I would I would take the
1:39:53 outline and the synopsis of this thing
1:39:56 pop it over to chat GPT to work on the
1:39:58 title and things like that anyway all
1:40:01 right so let's see we'll call this uh
1:40:04 what are we going to call
1:40:07 this um robot
1:40:11 Mayhem
1:40:14 script all
1:40:19 right and then we're going to do new
1:40:22 prompt
1:40:25 okay so one of the things that Gemini
1:40:28 can do is it's multimodal did you tell
1:40:30 Gemini it is a standup comedian writing
1:40:32 the script no I told it it was a
1:40:34 screenwriter writing the
1:40:37 script um but it obviously didn't listen
1:40:41 because maybe it's a screen writer but
1:40:43 it's not a screen title
1:40:46 writer cuz those are some shitty ass
1:40:49 titles
1:40:51 um Gemini is
1:40:55 uh is
1:40:57 multimodal so you can upload a video
1:41:01 there's sample videos let me upload
1:41:07 one let's see if I can go find something
1:41:10 interesting um
1:41:13 MP4 MPEG
1:41:16 movie on this Mac let's
1:41:20 see yeah this is one I did the other
1:41:23 night
1:41:24 [Music]
1:41:28 here uhhuh there's one world mind school
1:41:33 started off as an idea and something
1:41:35 that I wanted for my son and wanted to
1:41:37 spread loves the camera I looked at
1:41:40 daycare and preschool facilities and
1:41:41 yeah this is good we we'll pop this one
1:41:43 in here so this is this is a a video for
1:41:46 a school that my co-founder of Story
1:41:48 vine her son goes to this
1:41:52 school
1:42:23 hey
1:42:27 wash I don't get it
1:42:31 Joker all right so here
1:42:37 is there's the movie so we're going to
1:42:39 do a prompt we're going to
1:42:41 say
1:42:45 um give me a
1:42:49 title
1:42:52 description
1:42:56 keywords
1:42:59 and
1:43:06 chapters of this
1:43:10 video
1:43:12 that
1:43:14 Summarize each
1:43:18 chapter um show any on screen
1:43:25 text
1:43:28 and
1:43:29 describe
1:43:31 what's happening
1:43:34 in uh what's happening visually in each
1:43:42 chapter be as detailed as
1:43:48 possible and so it it it just it's watch
1:43:53 it's already watched the movie it knows
1:43:55 everything about the
1:44:06 movie it's more or less a develop
1:44:09 developer beta so cut it some slack oh
1:44:13 the oh uh Gemini 1.5 yeah I agree love
1:44:17 you Kyle thank you micro appreciate that
1:44:20 you asked what to call
1:44:22 it
1:44:25 a wash but I don't get I don't get the
1:44:27 AI wash AI I don't get I don't get it I
1:44:30 don't I don't get it um let's
1:44:38 see oh my mouse keeps
1:44:42 dying chapter one introduction learning
1:44:44 in
1:44:51 nature World mind immersion school so
1:44:54 let's ask it something odd like
1:44:59 um how many times does the girl in the
1:45:05 pink
1:45:08 jacket appear in the
1:45:16 video Let's see if I can just do
1:45:22 that
1:45:25 I wonder if it can if it can give give
1:45:27 us
1:45:29 actual clips of the video I doubt it the
1:45:32 girl in the pink jacket appears in the
1:45:34 video six times please give me the start
1:45:40 and
1:45:42 stop
1:45:43 timestamps of
1:45:45 those six times and
1:45:50 describe what she's
1:45:56 doing in
1:46:00 each never mind it wasn't funny
1:46:04 oh that might help with some Learners
1:46:06 who are having accessibility issues with
1:46:08 videos on some platforms yeah this I
1:46:11 mean this starts to get into this is
1:46:13 really crazy stuff uh Ethan mik was
1:46:15 talking
1:46:16 about he did a screen recording of
1:46:19 himself using some software and then had
1:46:23 this thing learn it um like learn it and
1:46:27 suggest improvements for his workflow
1:46:29 things like that so so there's all sorts
1:46:31 of crazy things you're going to be able
1:46:32 to
1:46:34 do okay at minute two she's holding a
1:46:37 rock so let's go
1:46:39 [Music]
1:46:41 um let's go find that
1:46:47 [Music]
1:46:51 video um MP
1:46:54 4 MP4 movie this
1:47:07 Mac World
1:47:11 mind world world
1:47:14 mind
1:47:16 okay so at second
1:47:19 two
1:47:22 here
1:47:23 [Music]
1:47:25 so it it got that wrong it's second
1:47:28 seven that it
1:47:33 starts is this it
1:47:37 yeah oh no that's not it is this it yeah
1:47:43 standing on a rock looking at the boy in
1:47:45 the red shirt is that true
1:47:48 no standing on the Rock oh here looking
1:47:53 at the boy in the red shirt oh so it
1:47:55 mistak that for pink I
1:47:59 think all
1:48:01 right let's
1:48:03 see sitting on the ground she's walking
1:48:06 two at 42 she's walking with other
1:48:09 students at
1:48:14 42 oh different pink jacket
1:48:18 so I wanted the pink fuzzy
1:48:22 jacket
1:48:27 interesting okay all right um but anyway
1:48:32 yeah it can it can see and talk about
1:48:34 movies you can also do things like
1:48:41 um let's see um give me a list
1:48:48 of
1:48:51 um visual
1:48:59 um
1:49:00 edits and what shots are
1:49:05 shown and
1:49:09 um describe let's see
1:49:13 and
1:49:15 provide
1:49:16 relevant meta data in Json
1:49:22 format
1:49:26 so I can search
1:49:31 for
1:49:34 any in the
1:49:40 video let's see let's just see how it
1:49:43 how it does it might do it might do good
1:49:45 we'll
1:49:46 see
1:49:48 uh no content
1:49:52 what
1:50:00 try
1:50:01 again we'll try again there
1:50:11 buy
1:50:16 make wake up
1:50:21 people time St might not be exactly the
1:50:24 same as player that's a good
1:50:29 point it's
1:50:37 thinking well let's go while let's
1:50:39 thinking let's go look at the video so
1:50:40 the first three scenes
1:50:43 are we got kids playing on
1:50:48 rocks we got a kid in a tree we got
1:50:52 a nut in a
1:50:55 cup and then we got Megan Patterson
1:50:58 talking about it okay so three kids
1:51:01 scenes followed by an adult talking
1:51:03 about the
1:51:09 school let's see how it
1:51:12 did Time Zero type cut wide shot of
1:51:16 children playing in the park cut closeup
1:51:18 of a boy in a red shirt climbing on
1:51:21 rocks yep medium shot of a girl in a
1:51:23 pink jacket standing on a rock high
1:51:26 angle shot of children sitting on the
1:51:27 ground drawing pictures is that
1:51:33 right yep kids playing drawing pictures
1:51:37 right
1:51:39 here see the
1:51:41 paper
1:51:44 amazing cut closeup of children drawing
1:51:48 of a tree medium shot of a teacher
1:51:50 talking to the children closeup child
1:51:53 holding container with tadpoles in it
1:51:55 not quite but close carrying Nets pine
1:51:58 cone climbing a tree amazing all right
1:52:02 so yeah
1:52:04 just [ __ ]
1:52:07 nuts and I mean this is
1:52:10 how this is how that robot does what it
1:52:12 does right cuz it's it's got cameras
1:52:15 that are just looking at [ __ ] it can see
1:52:17 its hands it can see the objects in
1:52:19 front of it so it's probably just
1:52:22 running
1:52:24 you
1:52:25 know what did they say 200 Hertz you
1:52:28 know 200 times a
1:52:33 minute it's it's processing what it's
1:52:35 seeing and being able to send that as
1:52:37 part of the
1:52:39 prompt it's cuckoo it's cuckoo this
1:52:43 world we're living in
1:52:45 now never too
1:52:47 old yeah I think that image came from
1:52:50 the other night someone was in the thing
1:52:52 going I'm too old for this
1:52:55 [ __ ] I'm too old I can't learn can't
1:52:57 teach this old dog new
1:53:00 tricks Marge how do I use a Tik Tok is
1:53:05 it on you know we don't believe that
1:53:08 around
1:53:12 here what do you think of Devon I think
1:53:14 Devon's [ __ ] remarkable what what the
1:53:17 thing that well so a couple of things
1:53:20 about
1:53:21 Devon so the demos are really sexy all
1:53:24 of the demos are them just walking
1:53:26 through the steps that have already
1:53:28 happened the steps take a long
1:53:30 time and there's a lot of back and forth
1:53:33 and they're calls to the gp4 open a uh
1:53:38 API so they're
1:53:40 expensive so to have Devon do something
1:53:43 for you is somewhere between like I
1:53:45 don't know two and $25 or $30 so Devon
1:53:50 can get really expensive really fast
1:53:52 from what I understand
1:53:53 understand
1:53:55 um but it's the
1:53:58 first it's the first
1:54:02 commercial
1:54:04 um
1:54:06 offering that is B basically taken what
1:54:09 if if if you're a geek and you want to
1:54:11 go build your own Devon the tools are
1:54:13 out there that you can build some
1:54:15 version of it right now right um so so
1:54:20 it's the it's the first tool that that
1:54:22 really put all the piece Parts together
1:54:24 in a pretty slick
1:54:26 way
1:54:27 um so I haven't I I'm on the wait list I
1:54:30 haven't had a chance to play with it yet
1:54:31 but I'm I'm really curious to play with
1:54:33 it and and see if it's any good the but
1:54:36 it looks it looks quite promising so if
1:54:38 you don't know Devon is they're calling
1:54:40 it the world's first AI programmer it's
1:54:44 an autonomous agent that you just give
1:54:46 it a goal and then it creates a project
1:54:49 plan for itself creates a GitHub rep for
1:54:53 you like a code
1:54:55 repository duplicates any things it
1:54:58 needs to goes and reads API
1:55:02 documentation writes calls to the apis
1:55:05 it's got its own website it's got its
1:55:06 own coding environment it's got its own
1:55:08 command line interface and then it's got
1:55:11 this project and then it communicates
1:55:13 with you and it it goes through its
1:55:14 project plan and it just does [ __ ] and
1:55:17 it'll occasionally check in with you and
1:55:19 say hey you got to make a decision here
1:55:21 or there and if it [ __ ] something up it
1:55:23 debugs itself it writes in like print
1:55:27 line commands to to figure out what the
1:55:30 what the uh what the issues are and then
1:55:33 it tries to fix them so it's it's pretty
1:55:36 crazy it's called Devon from a company
1:55:37 called
1:55:39 cognition and the CEO of cognition is a
1:55:42 guy named Scott
1:55:43 woo go to Twitter and look up Scott woo
1:55:47 math M A there's a video of Scott woo in
1:55:52 a math
1:55:54 competition and it's like the dude is
1:55:57 wicked smart as they
1:56:00 say paid of course thought like Ah
1:56:04 that's just that's just math competition
1:56:06 techniques that he's
1:56:07 [Laughter]
1:56:11 using oh man so for a creative person
1:56:14 you might have a section of video you
1:56:16 want to cut yeah well that's
1:56:20 so that thing that I just did with that
1:56:22 video like there's like I'm I'm in a
1:56:25 development environment there so and
1:56:28 I've got a video company so how how I'm
1:56:30 going to use Gemini is I'll go in there
1:56:34 and I'll sort of figure out what's it
1:56:35 good at what's it not good at there's a
1:56:38 lot of kinds of video analysis that I
1:56:42 was planning on using three or four
1:56:44 different tools for that I'm now just
1:56:47 going to use
1:56:49 one right so rather than me having to
1:56:52 take multiple data streams when I
1:56:54 analyze a video and figure out how to
1:56:56 munge them all
1:56:58 together I'm just going to be able to
1:57:00 make a call to Gemini and say do this
1:57:02 analysis give me Json and then I can
1:57:05 take that Json and do whatever I want
1:57:07 with it if I can get it to do that
1:57:09 consistently it's going to be amazing
1:57:13 so so that's
1:57:17 cool it's I I gotta tell you so doing
1:57:21 anything right now
1:57:23 learning these tools getting good with
1:57:25 them building things building
1:57:29 automations a thing that you have to
1:57:33 come to terms with and it ain't
1:57:36 necessarily
1:57:37 easy
1:57:39 is anything that you're
1:57:42 building is going to be
1:57:45 irrelevant 3 months from
1:57:48 now and not irrelevant like you can't
1:57:50 use it like like built an automation
1:57:54 almost a year ago now that we still
1:57:57 use but there's like 10 new like if I
1:58:01 were to build that tool today there's 10
1:58:04 different ways to do that now that are
1:58:05 much easier than what I did so just
1:58:09 right now anything you do is is just an
1:58:12 exercise in just learn it right now and
1:58:14 three months from now there'll be
1:58:15 different tools and the the workaround
1:58:18 that you figured out where you duct tape
1:58:20 together these three different things to
1:58:23 you made an image in mid journey and
1:58:24 then you took that to Runway and you did
1:58:26 that thing there and then you did a
1:58:27 voice in 11 labs and you did a thing
1:58:29 over at some point you're just going to
1:58:30 be able to go make me that [ __ ] movie
1:58:32 and it just go just spit it out for you
1:58:35 with all that [ __ ] done right so don't
1:58:38 get too married to any specific solution
1:58:41 you come up with or any specific way of
1:58:43 doing
1:58:44 something but do it anyway it's
1:58:47 important to do it
1:58:50 because learning what
1:58:53 possible
1:58:54 today is super
1:58:57 critical because even the shittiest of
1:59:00 these large language models
1:59:03 are
1:59:05 absolutely remarkable compared to what
1:59:09 we had to do
1:59:11 before and they're not perfect and
1:59:14 they're changing all the time and
1:59:16 they're getting remarkably better
1:59:20 so there's that I saw a demo of an open
1:59:22 source competitor today dude created a
1:59:25 webbased chat pretty easy yeah exactly
1:59:27 the the thing that Devon I mean what
1:59:29 Devon
1:59:30 did again interface [ __ ]
1:59:35 matters I got I got in a Twitter fight
1:59:38 with this dude because he was just like
1:59:40 um he was just like well Devon's just
1:59:43 you know it's just these four tools
1:59:45 right stuck together I'm like yeah but
1:59:47 interface
1:59:49 matters like chat
1:59:51 GPT
1:59:53 you could do what chat GPT did for two
1:59:57 years before chat GPT existed actually
1:59:59 for like 4 years before chat GPT
2:00:02 existed in the open AI playground
2:00:05 developers were playing with GPT and
2:00:08 calls to
2:00:10 GPT for
2:00:14 years why did chat GPT get to 100
2:00:17 million users in six weeks interface
2:00:21 simple interface
2:00:23 you didn't have to know what a model was
2:00:25 you didn't have to know what weights
2:00:26 were you didn't have to know what
2:00:27 temperature was you didn't have to know
2:00:29 what embedding was you didn't have to
2:00:30 know what fine tuning was you just
2:00:33 logged into this site chat. open.com and
2:00:36 it looked like a Google search box and
2:00:37 you go I don't make me a poem and then
2:00:41 out would come a poem you're like oh
2:00:42 that's
2:00:43 cool 100 million users in six
2:00:46 weeks miror morat's [ __ ] answer
2:00:49 today about Sora training I think it was
2:00:51 yesterday but yeah
2:00:53 makes me trust open AI even less yeah
2:00:57 exactly exactly it was a it was a uh
2:01:01 Mira needs a little bit of media
2:01:02 training is what I would
2:01:06 say I mean here's the
2:01:09 deal and if you don't think this is true
2:01:12 and this is even true for Adobe who
2:01:13 claims it's not I promise
2:01:17 you in order to train these models
2:01:20 you've got to train them on a lot of of
2:01:22 data like a lot of data like as much of
2:01:25 the internet as you can get your hands
2:01:27 on within that data you're going to
2:01:30 sweep up a ton of copywritten
2:01:34 information all of them all of them
2:01:38 whether it's grock or gemini or
2:01:43 adobe's image tools they claim it's only
2:01:46 um public domain and and Adobe
2:01:51 stock
2:01:55 all of them have to train on something
2:01:57 or or these tools wouldn't be that
2:02:01 good I saw it describe I saw chat GPT
2:02:04 described today it's not it's not a bad
2:02:06 thing is it's kind of like it's kind of
2:02:08 like a really fancy version of an
2:02:11 encyclopedia it's like it just knows all
2:02:13 the [ __ ] and then you talk to it it's
2:02:16 like here here's the
2:02:17 [ __ ] but it's not quite like that cuz
2:02:20 you're not it's not pulling articles
2:02:22 from like an actual you know Corpus of
2:02:26 documents it's these are all just
2:02:28 mathematical points in space that are
2:02:30 attached to words
2:02:33 and but
2:02:35 anyway yeah I know yeah that interview
2:02:38 with her was pretty CR
2:02:39 cringeworthy did you see the open AI
2:02:42 robot Kyle yeah we just we spent a lot
2:02:43 of time before you got here ISU looking
2:02:46 at that video and talking about how it
2:02:47 works and what it's doing and and then
2:02:50 we actually wrote a screenplay about it
2:02:53 cuz it's like imagine having that robot
2:02:55 in your house it's like chbt just
2:02:58 hallucinates or or the plot point that
2:03:00 Gemini came up with which was pretty
2:03:01 good is one of the guys owns a parrot
2:03:04 and the parrot learns how to make the
2:03:05 robot do [ __ ] so the guy leaves the
2:03:08 house and the parrot just has the robot
2:03:10 doing [ __ ] around the
2:03:13 house the open source models you had to
2:03:16 edit code to prompt it just a few months
2:03:19 ago yeah exactly EX exactly and that's
2:03:23 you know the fact that you can just log
2:03:24 into Devon and go write me a program
2:03:28 like take this GitHub
2:03:30 repository and clone it and then make it
2:03:32 do this other
2:03:35 thing and it'll just go and do all that
2:03:39 [ __ ] for someone like me who's
2:03:44 like Tech literate but not a
2:03:47 developer like that's [ __ ] genius cuz
2:03:49 I can I can like once it's doing it [ __ ]
2:03:52 I can figure out what it's doing and go
2:03:54 oh do this do that like I can be a good
2:03:56 product
2:03:59 guy but like I don't know enough command
2:04:02 line [ __ ] I just don't give enough of a
2:04:04 [ __ ] about learning Python and learning
2:04:06 this and learning command line and it's
2:04:08 just it's I I just i' I've just got
2:04:11 ideas I'm just I got
2:04:15 ideas so if I can have a tool that I'm
2:04:17 like I got an idea and it just goes and
2:04:20 vomit some [ __ ] out and I can steer it
2:04:23 enough to turn it into
2:04:25 something take that all day
2:04:27 long and the engineers will be like well
2:04:30 it's not real
2:04:31 engineering I
2:04:33 know and I don't
2:04:35 [Laughter]
2:04:39 care
2:04:42 oh let's learn C in unity for games
2:04:45 exactly learning react is a bit time
2:04:48 consuming yep exactly every so here's
2:04:51 the thing every single thing is a bit
2:04:55 timec consuming right so if you're an
2:04:57 engineer and you're really good at the
2:04:58 coding and now you got to create some
2:05:00 little character sprites for your game
2:05:03 why do why do initial video games always
2:05:06 look like
2:05:07 [ __ ] well because you know an engineer
2:05:12 did it and the engineer was like well I
2:05:13 just got to put a Sprite in here I'll
2:05:15 just use blocks right or I'll just use
2:05:18 you know I I'll just draw stupid
2:05:19 characters and then the game the game
2:05:21 will be fun to play but it'll have
2:05:23 stupid characters in it or you get the
2:05:25 other side where you've got people that
2:05:27 are really good at the art and then the
2:05:29 game play really
2:05:31 sucks well now we move into a world
2:05:34 where you're going to be able to have a
2:05:36 bunch of different agents that go off
2:05:39 and do [ __ ] for you and one could be go
2:05:42 find go create you know 10 different art
2:05:45 styles for this
2:05:48 game and then the other one's going off
2:05:50 and making the game play like it should
2:05:52 have the smoothness of asteroids and the
2:05:54 the excitement of you know Call of Duty
2:05:58 and the the the driving mechanics of of
2:06:02 an F1 game and it's off building all
2:06:05 that [ __ ] coming up with the creative
2:06:08 over here stitching them together the
2:06:11 Paul rater um AI podcast this week it's
2:06:14 a really good one you should watch
2:06:16 it cuz they're talking about Sam alman's
2:06:18 quote where he said 95% of what we hire
2:06:23 marketing agencies and strategists for
2:06:25 will be done by Ai and so he's he's
2:06:28 talking a lot about that and he goes
2:06:31 back to an earlier episode they did like
2:06:33 a year and a half ago or two years ago
2:06:36 um where I forget who they were talking
2:06:38 about but what they were talking about
2:06:40 was the way these autonomous agents are
2:06:42 going to
2:06:43 work is you're going to give it a
2:06:46 goal and he said it's going to be like
2:06:50 a um
2:06:52 a
2:06:54 beehive where you know bees are all
2:06:57 self-organizing right and they've all
2:06:59 got their jobs and so you'll give it a
2:07:02 goal and like it will spin up these bees
2:07:05 these worker bees and these worker bees
2:07:06 will go do all these little tasks for
2:07:08 you and they'll come back and they'll
2:07:10 build this thing and it's like and
2:07:13 they're just like off doing their thing
2:07:15 building their building your thing
2:07:17 building your thing building your thing
2:07:18 and he
2:07:20 said what it's what it's going to be
2:07:23 like is every
2:07:25 person he goes imagine having like a
2:07:29 thousand person company at your
2:07:33 disposal that that's what it's going to
2:07:34 be
2:07:36 like not this year
2:07:38 probably but not 10 years from now
2:07:41 probably like in the next two to three
2:07:43 years we'll we'll see hints of it this
2:07:45 year and I think Devin is is a hint of
2:07:49 it I think code interpreter in inside
2:07:52 chat gp4 is a really teeny tiny hint of
2:07:54 it right code interpreter you upload a
2:07:57 CSV file and you say go tell me what's
2:08:00 interesting about this data and it
2:08:02 writes and executes its own python code
2:08:05 and analyzes the data inside the
2:08:08 CSV on your behalf so imagine having a
2:08:12 thousand of those that are dynamically
2:08:15 created based on your
2:08:17 goal and you've got coordinating bees
2:08:19 that say okay here's our task list in
2:08:22 order to do task one we've got to do
2:08:24 this we've got to write a creative brief
2:08:26 we've got to come up with a name we've
2:08:29 got to come up with taglines we got to
2:08:31 come up with a marketing strategy and
2:08:32 all those things become separate little
2:08:34 agents that get spun up on your behalf
2:08:37 run out use whatever the [ __ ] tools they
2:08:39 need or write their own tools if they
2:08:41 need
2:08:43 to get you the answers they want and
2:08:45 come back to the little coordinating
2:08:47 dude that's going like okay check check
2:08:49 check check okay we've now got
2:09:54 disposal what could I do with that I
2:09:56 don't
2:09:58 know start thinking start getting
2:10:02 curious what's your CQ what's your
2:10:05 curiosity
2:10:08 quotient it should be going through the
2:10:11 roof right now why how many people in
2:10:14 here there's 96 people in here a lot of
2:10:17 the people that come here every night
2:10:19 night after night after night after
2:10:20 night after night why are they doing
2:10:22 that cuz I'm so Charming
2:10:24 no to hear this you can make
2:10:27 money there's a little bit of
2:10:29 entertainment but what it really is is
2:10:32 hopefully I'm reminding them stay
2:10:35 curious stay engaged stay in the
2:10:37 conversation stay in the conversation
2:10:40 stay in the
2:10:42 conversation and what we're all going
2:10:45 through is oh [ __ ] this is exhausting
2:10:48 and oh I don't know if I can learn
2:10:50 anymore what do you mean they just
2:10:52 changed the rules what do you mean it
2:10:53 can do video now oh God what do you mean
2:10:55 it's a million tokens what do I do with
2:10:57 a million tokens because your enthusiasm
2:11:00 is infectious thank you very much Martin
2:11:01 I appreciate
2:11:05 that why I'm this passionate why I'm
2:11:08 this energized
2:11:15 is this is going to [ __ ] a lot of things
2:11:18 up this going to [ __ ] a lot of people
2:11:20 up
2:11:25 who who is going to benefit in the next
2:11:28 three
2:11:30 years is anyone who's not ignoring
2:11:36 it anyone with any background with any
2:11:40 skill set if you're just paying
2:11:42 attention to it and and staying engaged
2:11:45 and staying in the
2:11:50 conversation you're going to be
2:11:52 okay if you're
2:11:55 here holy [ __ ] are you Way
2:12:00 Beyond where most of the world is right
2:12:02 now with this stuff seriously if you're
2:12:04 like but Kyle I just stumbled in it's my
2:12:06 first night I don't know anything about
2:12:08 AI I a't even been to the chat
2:12:11 JP the fact that you're still here
2:12:13 listening to me ramble on like a
2:12:15 bumbling
2:12:18 idiot you're way farther ahead
2:12:22 than 90% of the people out there
2:12:25 probably
2:12:26 95 and I'm not just pulling that number
2:12:28 out of my ass McKenzie did a
2:12:33 study last year they they they did a
2:12:36 sample in February and another one in
2:12:39 November of 23 in
2:12:42 February 133% of people Kevin Vasquez
2:12:46 I'm not
2:12:47 okay in none of us are okay in February
2:12:53 133% of people had used chat
2:12:56 GPT in
2:12:59 November
2:13:01 18%
2:13:03 18 82% of people had not used chat GPT
2:13:07 as of last
2:13:09 November so let's say let's say that's
2:13:12 gone up
2:13:14 to
2:13:17 20% or
2:13:19 25% 25% of people people have used chat
2:13:25 GPT what does that
2:13:27 mean what percentage of those people
2:13:29 have used it more than once oh that's
2:13:32 probably
2:13:34 half right how many people do you know
2:13:36 oh yeah I use chat gbt what have you
2:13:38 done with it well I just I tried it that
2:13:40 one time when it came out okay that's
2:13:42 that that's it okay cool so half right
2:13:46 so that's
2:13:48 12% how many of those people do you
2:13:51 think subscribe to chat gp4 maybe half
2:13:54 again that's
2:13:56 6% how many of those people do you think
2:13:59 even know what the features of chat GPT
2:14:03 4
2:14:05 are I talked to a creative director two
2:14:07 days
2:14:08 ago copywriter uses chat GPT every
2:14:13 day had not subscribed to chat GPT
2:14:16 didn't know what the features
2:14:18 were
2:14:19 so maybe half
2:14:22 again of
2:14:24 that
2:14:28 6% know what chat gp4 is and what it's
2:14:31 capable of and are using it
2:14:34 3%
2:14:38 3%
2:14:41 three now is is that a statistical
2:14:45 number no I'm pulling it out of my ass
2:14:46 but I but like that's how early we
2:14:50 are
2:14:52 so as far behind as it feels like you
2:14:54 are you're not we're way early we're way
2:14:58 early so just getting curious today 75%
2:15:02 is done 75% done with my Memoirs which
2:15:05 AI platform is for me writing and
2:15:07 organizing at
2:15:08 all if you need to do any um the the
2:15:12 organizing I don't know like I there's
2:15:15 probably better people in here I would I
2:15:17 would connect with uh Peter Kaminsky is
2:15:20 going to give you some it's going to be
2:15:22 too
2:15:24 technical check check with Suzanne
2:15:26 Welker or check with Vicky Baptist
2:15:30 Becky I I feel like both of them have
2:15:32 got the skills to help you organize the
2:15:34 stuff um it depends what you want to do
2:15:37 with the writing if you need to do
2:15:38 research then definitely use perplexity
2:15:41 because perplexity is really good at
2:15:43 finding citations and things like that
2:15:45 um for me chat gp4 if you if you just
2:15:48 want to do writing stuff chat gp4 is
2:15:51 just better than anything else right now
2:15:53 um that said a lot of people really are
2:15:57 liking Claude 3 Opus but that's another
2:15:59 $20 subscription so if if you've already
2:16:02 got the $20 subscription for chat gp4
2:16:06 you probably don't need it for Claude
2:16:08 Opus
2:16:09 but I just just cpt4 for me right now is
2:16:13 where it's at so and I think it will be
2:16:15 for a
2:16:16 while um Kyle where will you host these
2:16:20 if America talk get trashed um I don't
2:16:23 know I I'll figure that out I'll
2:16:24 probably just stream on YouTube or
2:16:27 or I think I think what I might do is I
2:16:30 might I might switch over to a platform
2:16:34 that that does multi streaming like I'll
2:16:36 stream it on LinkedIn YouTube and uh
2:16:42 what was the third one they
2:16:44 had Facebook I
2:16:47 think I don't know so I don't know I
2:16:50 I'll figure it out it's
2:16:52 this is Tik Tok going away is not going
2:16:54 to happen for months like it's got to
2:16:55 get through the Senate and then you know
2:16:58 if Biden signs it it's there's there's
2:17:00 all sorts of constitutional questions
2:17:02 there's all sorts of it's just it's
2:17:05 going to take a while so we're it's not
2:17:08 urgent shared with a friend some travel
2:17:10 sites because we're going to the
2:17:13 Philippines we're booked now oh nice
2:17:16 awesome
2:17:17 fantastic oh Ann Murphy's another good
2:17:19 one Becky yeah Becky for organizing all
2:17:22 roads lead back to notion I was going to
2:17:24 say notion I just don't use it so I
2:17:25 don't I don't have
2:17:27 any any good advice there so find people
2:17:30 that are good at organizing [ __ ] I'm not
2:17:32 I'm not the person to ask about
2:17:36 organization what's good about
2:17:37 perplexity what's good about perplexity
2:17:39 is uh it's it's they've created a tool
2:17:44 that lives kind of in between a Google
2:17:46 Search and a large language model so
2:17:49 it's got a lot of the
2:17:52 capabilities of a large language model
2:17:54 but it does searches and it creates
2:17:55 citations for everything it finds and
2:17:58 then it also when you do a when you do a
2:18:00 search you can limit it to a certain
2:18:02 area so I only want to look in um
2:18:05 academic papers or I only want to look
2:18:07 on Wikipedia and it'll just do citations
2:18:09 from those places um and then it also
2:18:12 has got this thing called co-pilot which
2:18:14 is not Microsoft co-pilot it's its own
2:18:17 co-pilot uh and uh it will suggest
2:18:22 Quest what are the next prompts that you
2:18:24 might want to ask it and it's pretty
2:18:26 good it's just it's just good Clapper I
2:18:29 don't know what Clapper
2:18:31 is Twitch what's Clapper Clapper
2:18:35 googling now yeah exactly I don't know
2:18:37 what I'm right there with you what the
2:18:39 hell is
2:18:40 [Laughter]
2:18:43 Clapper oh my Mouse is dead this is
2:18:48 annoying Clapper AI is it an AI thing or
2:18:52 is it
2:18:56 a Clapper plus Hive Clapper is partnered
2:18:59 with
2:19:01 Hive the [ __ ] is
2:19:04 Clapper Clapper
2:19:07 [Music]
2:19:11 app be heard be seen be
2:19:20 valued
2:19:26 what is
2:19:29 Clapper be heard be seen be valued I get
2:19:31 the value proposition what is
2:19:34 Clapper record up to three minute videos
2:19:37 oh it's a Tik Tok
2:19:43 thingy Clapper group
2:19:46 chat fam tear nearby feature yeah okay
2:19:50 it's a it's a Tik Tok competitor got it
2:19:53 a lot of Tik tokers are going to it okay
2:19:56 yeah
2:19:57 until I it fine like not yet I won't yet
2:20:02 Tik Tock replacement not yet it's like a
2:20:04 Tik Tok clone got it
2:20:07 okay um no interest until it becomes
2:20:11 clear that this thing's going down the
2:20:12 toilet this is not going to go down the
2:20:13 toilet did you see that [ __ ] uh
2:20:16 minuchin is is raising money to buy it
2:20:19 and then Kevin o said he wants to put
2:20:21 put together a Syndicate to buy it this
2:20:23 is all here's what this is this is an
2:20:27 extortion move to
2:20:30 get to to to
2:20:33 protect American companies who've lost
2:20:37 market share to Tik
2:20:39 Tock and you know to get it out of
2:20:44 Chinese ownership and into American
2:20:46 investors hands like it's like it's
2:20:49 either it's either you you're going to
2:20:51 make our billionaires rich or you're not
2:20:53 going to exist and then we're going to
2:20:56 take a major competitor out of the
2:20:58 marketplace that's that's what this
2:21:00 exercise is so I'm a little cynical
2:21:03 about this [ __ ]
2:21:05 so if every they they are not going to
2:21:10 they are not just going to flush you
2:21:15 know a platform of this scale and this
2:21:18 profitability down the toilet they're
2:21:20 just going to
2:21:21 try to go get the
2:21:24 money or they're going to flush it down
2:21:26 the toilet so it just pushes us back to
2:21:28 Facebook and you know
2:21:31 YouTube and
2:21:35 Instagram
2:21:43 whatever you should get the occasion to
2:21:46 switch to the VR life uh not right now
2:21:49 I'm too exhausted can you tell tell the
2:21:51 nice
2:21:52 people about she leads AI this weekend
2:21:55 yeah
2:21:56 so um so go to I got to get out of here
2:22:00 anyway so let me let me wrap this up if
2:22:03 you're if you're new here um and you
2:22:06 want to be part of
2:22:08 community people that are into AI
2:22:11 excited about AI keep coming back to
2:22:13 this channel but also join the AI Salon
2:22:15 go there the salon. click on the first
2:22:18 link it's a link tree click on the first
2:22:20 link that'll take you to the community
2:22:23 read about the community read about our
2:22:25 values really important read about our
2:22:27 values make sure you resonate with them
2:22:29 and then join the community in the tech
2:22:32 section there's a there's a group called
2:22:35 weekend
2:22:37 gpts that's run by Ann Murphy this
2:22:40 Saturday she's doing a thing called she
2:22:43 leads AI which is a a round table of
2:22:46 women in AI it's at noon Pacific and the
2:22:51 details are in her Channel all right so
2:22:53 Ann Murphy over there on the
2:22:56 salon and go to the weekend GPT Jam in
2:22:59 the tech
2:23:01 Guild and sign up for that
2:23:04 [ __ ] and go there it's going to be
2:23:07 amazing all right she's got a really
2:23:10 cool community and it's you're going to
2:23:12 get to meet some amazing people so and I
2:23:14 think Cindy isn't Cindy [ __ ] moderating
2:23:16 or something like that in I think that
2:23:19 is the case
2:23:22 but they won't buy the algorithm so
2:23:23 don't know I posted in weekend GPT Jam
2:23:27 awesome
2:23:31 perfect well done Kyle thank you very
2:23:34 much thank you very much an appreciate
2:23:35 that yes Cindy yes Cindy's a rockstar
2:23:38 she'll be a
2:23:39 blast um okay
2:23:43 so how you can support me in this
2:23:45 channel follow the
2:23:48 channel like right now go follow like
2:23:50 how do you follow on Tik Tok I don't
2:23:52 know ask Marge I don't know but follow
2:23:56 it um you can subscribe to the lives you
2:23:59 can buy the video series in the corner
2:24:01 there that that would help the the video
2:24:03 series is called underhyped and it's a
2:24:06 bunch of videos where I talk about why I
2:24:07 think AI is a big deal and how to use it
2:24:10 how to not get overwhelmed by it which
2:24:12 we all
2:24:13 are
2:24:15 um you can do that you can keep coming
2:24:18 back
2:24:19 here and
2:24:21 if you keep coming back here you get to
2:24:23 be what are called an irregular so
2:24:26 Irregulars just regularly show
2:24:30 up is it devon. aai uh Amelia's wife the
2:24:37 the programming thing is called Devon
2:24:39 the company is called
2:24:42 cognition um and it's I don't think it's
2:24:48 devon. um but anyway thank you thank you
2:24:51 to the Irregulars who show up here thank
2:24:53 you to the subscribers that subscribe
2:24:54 thank you to the mods who help uh make
2:24:57 sense of all this every night I it's
2:24:59 it's it is relatively thankless work I
2:25:03 feel like I don't thank you enough so
2:25:05 thank you thank you thank you um and
2:25:08 that's it so let's see who's in here
2:25:10 tonight I've seen a lot of familiar
2:25:11 faces Milo's wife of course she was on
2:25:13 time tonight a rare occurrence very nice
2:25:15 though Mike Row good to see you Brian
2:25:18 Whitney Daisy May ASU Becky r a shoe and
2:25:22 Becky R and an Murphy
2:25:26 too Silver Fox good to see you
2:25:30 Joker great to see you Apple
2:25:33 user cat
2:25:36 Alan digital Creator nice name I like it
2:25:40 fantastic Bob Kevin Kevin Vasquez tell
2:25:43 him what he's won Sher D good to see
2:25:47 you master path to Freedom the daily
2:25:50 crunch
2:25:54 Joker all right everybody um thank you
2:25:58 thank you thank you everyone who's been
2:25:59 here tonight Mary Mary good to see you
2:26:01 jenu good to see you I made it all the
2:26:03 way tonight surprises never cease yeah
2:26:05 we were kind of all over the place
2:26:06 tonight but welcome to Chad add
2:26:10 right peace everybody tomorrow's Friday
2:26:13 uh we'll see if the the single tweet
2:26:17 that I saw that said a Japanese
2:26:19 newspaper has declared that Elon and Sam
2:26:22 Alman are going to um open source grock
2:26:27 and chat GPT tomorrow I doubt it but I
2:26:31 saw one tweet that said that so we can
2:26:33 hope bye
2:26:37 everybody