Jul 23, 2024 - Unlocking Creativity: The Power of AI in Art and Innovation**

Description

Chapters

Transcript

0:02 thing
0:04 [Music]
0:08 [Applause]
0:18 [Music]
0:30 oh
0:33 [Music]
0:43 [Music]
1:28 why are you barking let's go going on
1:30 sit sit why don't you sit take a load
1:34 off buddy
1:36 [Music]
1:51 [Music]
2:08 d d do do
2:16 [Music]
2:32 [Applause]
2:34 [Music]
2:56 good evening Tobias what's happening
2:58 Danielle steo what's happening Allen in
3:02 the
3:03 [Music]
3:16 [Music]
3:22 house wind rider 47 was
3:26 happening archetypal architect
3:30 [Music]
3:33 I'm supposed to be in the middle of some
3:36 fat cat dinner in Washington DC right
3:38 now and here I am at my shitty home
3:40 office I mean my $20 million TVV
3:43 Production Studio made to look like a
3:46 shitty home
3:49 office but I get to hang out with you
3:52 guys so there's that always a silver
3:58 lining if you're new here welcome my
4:00 name is Kyle Shannon this is the AI
4:02 learning
4:05 lab just plucking around on the guitar
4:08 while we get warmed up and people get
4:10 settled in and microwave their nachos
4:13 who's microwaving
4:15 nachos put them in the air fryer
4:21 [Music]
4:34 really breaks really breaks my
4:38 heart see a dear old
4:41 friend go down to that fall out place
4:45 [Music]
4:52 again y'all fans of the Brandy
4:55 carile she has a wicked Talent a wicked
4:59 Talent
5:07 Wicked three chords in the
5:10 truth three chords in the truth and a
5:12 voice like that holy
5:15 [Music]
5:25 [ __ ] you're feeling nervous aren't you
5:29 boy
5:32 with a quiet
5:34 voice and impeccable
5:37 [Music]
5:40 style champ be not singing tonight not
5:43 into it don't ever let them steal your
5:48 joy and your gentle ways to keep them
5:52 from running
5:56 wild well they can kick dirt in your
5:59 face dress you down up D it dress you
6:02 down and tell you that your place is in
6:04 the middle when I ate the way you
6:09 shine steo got the granddaughter working
6:12 tap honey keep tapping keep tapping we
6:15 got to break the record grandpa I don't
6:17 want to tap I don't like tapping Grandpa
6:20 please stop don't make me tap keep
6:22 tapping keep
6:25 [Laughter]
6:27 tapping I fully support this kind of
6:30 Labor we got to build character in the
6:32 young kids these days hey there how you
6:36 doing keep
6:37 [Laughter]
6:40 tapping oh
6:45 man all right
6:49 fantastic what are we going to do
6:50 tonight
6:53 [Music]
7:11 what is up with this that way off
7:15 something doesn't sound right in the
7:17 [Music]
7:19 guitar
7:25 oh see normally if I'm tuning champ is
7:28 singing he must eaten a bad snake or
7:30 something you know he catches snakes
7:32 right the
7:36 [Music]
7:43 [Music]
7:44 [Applause]
7:44 [Music]
7:48 champ oh now you're going to sing I get
7:51 a little I get a little aggressive you
7:53 want to be aggro tonight
7:56 [Music]
8:08 is this place I can rest my poor
8:14 head gather my thoughts in sweet
8:21 silence is the place where the feelings
8:24 on
8:27 dead from an overe exposure to
8:31 violence and is this a place I can
8:34 slowly face the only one I truly can
8:39 know these are tears from a long time
8:43 ago got these tears from a long time ago
8:48 I need to cry 30 years or so these are
8:52 tears from a long
8:57 time go
8:59 [Music]
9:07 well oh darling Oh Darling you say unto
9:11 [Music]
9:14 me where have you been all my life
9:21 time well I have been swimming seven sad
9:25 [Music]
9:28 Seas poor women and toss me their
9:31 [Music]
9:33 lifelines I pulled them into the waters
9:36 deeper in I'd have worn them but I
9:39 didn't
9:41 know these are tears were a long time
9:45 ago got these tears from a long time
9:49 [Music]
9:50 ago I need to cry 30 years or
9:56 so these are
9:58 tears these are tears from a long
10:05 time
10:07 [Music]
10:19 ago Bo all right champy came in a little
10:24 bit there he was singing background
10:26 normally singing lead tonight he wanted
10:28 to do harmonies I get it I get get it
10:30 sometimes you're just not feeling like
10:31 the the top dog you know you want to go
10:35 beta you don't want to be Alpha Dog
10:38 you'll be like you know what you take
10:40 the lead there
10:41 pops right
10:43 champy Tails
10:48 [Music]
10:53 wagon all righty then
11:03 hang on getting back back to
11:07 it good evening
11:14 [Music]
11:28 everybody song for request for Champion
11:35 [Music]
11:42 Metallica today I'm working on upgrading
11:44 software in a Burlington store while
11:48 [Laughter]
11:51 kylling listen the last time someone
11:54 watched the AI learning
11:56 lab and tried to update software it was
11:59 an engineer at crowd strike and we know
12:01 how that ended up all right so so do not
12:06 update software and watch the AI
12:07 learning Lab at the same time you get
12:09 distracted you can make money with and
12:12 then you know before you know it you're
12:14 calling a null operation and you're
12:16 crashing your memory blue screen to
12:18 death Global computer shutdown you
12:21 know I've caused worse I don't mind I
12:25 don't mind saying
12:27 it the hair looks like it's going to be
12:29 unfortunate situation tonight she
12:32 okay I'm not even supposed to be here
12:35 this is a freebie this is a freebie this
12:39 was in the schedule tonight and tomorrow
12:42 we're not on the schedule you knew that
12:44 I know you you paid your subscription
12:47 but but these two nights were clearly
12:49 excluded from the
12:53 calendar so we can just [ __ ] nose
12:56 this one in and it'll be fine
13:02 Kyle do you have the chat
13:04 GPT summarized feature in YouTube let me
13:07 go
13:08 look I don't think I do but I didn't
13:10 look for it
13:15 so so what do you do you roll over the
13:17 thumbnail and it it it shows up in the
13:19 upper
13:25 left is it supposed to show up right
13:28 there I got nothing
13:37 nothing I don't I do not
13:40 know so Danielle discovered in
13:45 YouTube Microsoft is out that's
13:51 old F1
13:53 hypercar oh that guy got to drive that
13:55 hyper car oh Nico rosberg well sure and
13:59 xf1
14:02 driver I got
14:04 nothing I don't know I do not know I do
14:08 not know update it I can re reload
14:17 it
14:19 uh-uh I ain't I ain't got it I ain't got
14:22 nothing is that what we supposed to be
14:31 Source camp in the house forever hooked
14:33 in the house we got Vicki we got Wolfman
14:36 Clint we got Amelio's wife I know I'm
14:38 doing it at the beginning
14:40 tonight go back and forth and in and out
14:43 of the
14:46 thumbnail wouldn't it be crazy if you
14:48 could create voice would it be crazy if
14:50 you didn't run ads when I was just
14:51 trying to watch a stupid 30 second video
14:55 that'd be crazy
14:59 that says the first time we're releasing
15:01 45 billion parameter model by far the
15:02 most sophisticated open What's with his
15:06 eyes like the first time we're releasing
15:09 a 45 billion does does Zuck not
15:13 have does Zuck not have like someone who
15:16 can give him some some makeup for for
15:24 his [ __ ] billionaires what are we
15:27 going to do with them
15:30 I think he can afford a makeup person to
15:33 abity look like
15:35 [Laughter]
15:40 that oh my god dude dude dude all right
15:46 with the hair we can go with you're
15:48 establishing a new
15:50 look but really the
15:52 sunglass the sunglass albino thing
15:55 really
15:57 really no don't click
15:59 wait what am I supposed to
16:12 do I got
16:17 nothing he must wear sunglasses 24/7 he
16:21 [Laughter]
16:27 must so the the
16:31 Llama Llama 3.1 came out today and their
16:35 405 billion parameter model came out
16:38 today there were some
16:40 interesting some interesting things on
16:43 this the one that I saw that I think is
16:45 worth
16:46 sharing what was it it was
16:53 llama
16:56 3.1 and grock
17:00 Oops I did that
17:03 wrong yeah this is the one so check this
17:07 thing
17:10 out can you tabularize it for
17:13 me okay so this is so if you go to
17:16 gro.com actually let me go to gro.com
17:23 oops where am I am there so let me go
17:26 here let me go gro.com grro
17:30 q.com by the way grock grq is
17:36 not Elon musk's thing it's a chip it's a
17:39 hardware
17:41 company
17:57 um on
17:59 got to log in
18:11 [Music]
18:22 apparently explain quantum mechanics to
18:25 me
18:32 hello what are you
18:35 doing
18:37 uh
18:40 oh all right which model am I using ah
18:44 here we
18:46 go okay so yeah so if you go to gro.com
18:49 gq.com in the upper right hand corner
18:53 you can flip to llama 3.1
19:00 70 billion and the 405 billion parameter
19:04 model and then watch what this guy got
19:06 it to do I don't know who who did
19:09 this Jonathan Ross did this so watch
19:12 this this pretty crazy so you can play
19:18 along and this is Source camp this
19:21 potentially lives in um in one of the
19:24 parlor
19:25 tricks just because of how fast this is
19:28 watch this
19:30 hi can you create me a one-day itinerary
19:32 for a trip to
19:35 Toronto that's a little complicated can
19:38 you tabularize it for
19:42 me can you add a duration
19:46 column can you remove the end times from
19:48 the time
19:50 column can you make the duration in
19:56 minutes and can you move the duration to
19:58 between the time and stop
20:02 column can you add lunch and dinner at a
20:06 nice
20:09 restaurant you know what I changed my
20:11 mind make it
20:12 [Laughter]
20:14 Vancouver [ __ ] amazing crazy right hi
20:19 can you create me a one day itinerary so
20:21 so yeah so go play with grock now which
20:23 one was he using oh he's using whisper
20:26 large V3 is that to
20:32 Toronto so let me flip over to
20:36 grock is whisper in
20:38 here
20:43 no so wait which one was he using he was
20:47 using 405 oh this was the 8 billion
20:50 parameter model so let's go let's try
20:53 this I'm I'm I'm just learning so I'm
20:56 going to go llama
20:57 3.1 8 8 billion instant llama 3.1 models
21:01 are now available models are
21:04 experiencing high demand daily token
21:13 limits make me a table comparing hot
21:16 dogs the Roman Empire and
21:25 Knuckles here wait I got to make this a
21:27 little smaller
21:30 okay hot dogs origin Frankfurt Germany
21:34 Roman empire ancient Italy Knuckles
21:36 human
21:40 body
21:43 um include
21:54 weights there's grams
21:59 um let's see
22:03 um replace hot dogs with poison
22:14 ivy replace knuckles with pumpkin
22:20 pie this is just [ __ ] Bonkers how
22:23 fast it is like you just can't keep up
22:25 with it you can't expand width oh
22:32 nice and this what I'm what I'm playing
22:35 with right now is um really just a tech
22:40 demo for their for their
22:42 chips um but it's it's nice that you can
22:45 use these crazy fast models let's let's
22:48 flip over to the
22:50 Llama reasoning one let me let me see if
22:53 what we can do here that would be
22:55 interesting um
23:07 give
23:07 me brief summaries for 10 different
23:11 marketing
23:23 Frameworks oh I ran
23:26 out it's like where it's completely out
23:29 it's not doing let me see if I can
23:32 do
23:35 3.1 70 billion let's try this boom all
23:39 right there's 70 billion
23:43 parameter all right let's make these a
23:45 little
23:46 bigger customer Journey mapping and the
23:50 7
23:51 PS let's mix customer Journey mapping
23:54 and the Seven PS and give me a
23:56 completely new marketing framework
24:04 step one awareness consideration
24:06 evaluation purchase onboarding
24:10 engagement ad
24:18 advocacy let's let's add in some Joseph
24:22 Campbell and Tony
24:27 Robbins hero's
24:30 journey The Ordinary World the
24:32 customer's current state The Call to
24:36 Adventure and then here's Tony Robbins
24:39 okay now let's mix them all into a
24:41 single unique
24:46 framework I mean this [ __ ] Bonkers
24:49 how fast this is okay stage one
24:52 Discovery awareness Ordinary World
24:54 certainty
24:57 okay I don't want you just to list all
24:59 of the
25:01 words separated by comma I want you to
25:04 actually come up with a new and unique
25:06 marketing framework give it an
25:08 interesting unique name and make each
25:11 step
25:14 original something new something no
25:16 one's seen
25:22 before introducing the Echoplex
25:26 framework the Echoplex is a holistic
25:29 marketing approach that recognizes the
25:31 interconnectedness of the customer's
25:33 Journey the Brand's ecosystem and The
25:35 Human Experience it consists of seven
25:37 stages each representing a critical
25:40 touch Point stage one resonance I like
25:44 this in this stage the customer becomes
25:47 aware of the problem or opportunity that
25:49 resonates with their values or desires
25:51 this is really
25:53 good the brand must create a sense of
25:55 resonance by aligning its messaging
25:58 offering Key activities conduct customer
26:00 research buyer personas create resonant
26:03 messaging I like that stage two
26:06 amplification the customer begins to
26:09 explore Solutions with their problem the
26:10 brand must amplify its presence through
26:13 various channels this stage is
26:17 critical stage three harmonization this
26:19 is a very frr kind of marketing
26:22 framework isn't it customer evaluates
26:24 different solutions and begins to
26:26 harmonize their needs stay 4 Catalyst or
26:33 uh
26:36 wait
26:38 catalysis catalysis I don't know I don't
26:43 know that word the customer must take
26:45 action and
26:46 catalyze catalysis I don't know okay I
26:49 guess that's
26:51 the the noun is that right is that the
26:54 noun I don't know I [ __ ] I should
26:57 have chat gbt teach me some grammar
26:59 apparently Echo formation customers made
27:02 a p uh a purchase the brand must create
27:07 a sense of echo formation by providing
27:10 ongoing support and
27:12 education residents reboot they've
27:15 achieved their desired outcome the brand
27:17 must reboot the resonance process be
27:20 creating new and innovative solutions
27:22 and then the
27:24 Echoplex the C customers become a loyal
27:26 Advocate the brand must create the
27:29 Echoplex effect by amplifying the
27:32 customer's voice that's actually really
27:35 good creating a sense of
27:39 community it's actually really [ __ ]
27:43 good can you share this output in the
27:45 Irregulars yeah sure
27:48 sure here's the we can write a book on
27:51 this one we accidentally wrote a book on
27:53 here one night introducing the Echoplex
27:56 framework so what what I did here was I
28:01 combined custo uh wait let me get to
28:07 the okay here it
28:11 is let me go to the AI Salon we'll go to
28:17 Irregulars I'll
28:18 [Music]
28:20 say here's a brand new marketing
28:25 framework which I
28:27 created on gro.com
28:30 using the Llama 3.1 70 billion parameter
28:43 model I
28:46 combined customer Journey mapping
28:50 comma the seven P marketing framework
28:55 comma Joseph Campbell comma
28:59 and Tony
29:01 Robbins into this unique and fascinating
29:06 marketing framework
29:13 period please respect my work this took
29:16 me at least 37 seconds to create period
29:25 [Laughter]
29:31 there you go all right oh that's
29:35 nice Mr Shannon does does what does
29:38 Washington that's that's actually that's
29:40 not bad for the hair the hair's probably
29:42 heading that direction
29:47 tonight Mr Shannon Shannon does not go
29:50 to Washington that's just rude CLA
29:56 Jacobs a hard problem I think that was
29:58 from last
30:00 night oh a sufficiently bad software
30:04 update is indistinguishable from a Cyber
30:06 attack that's really good who did that
30:08 that was Lord Lord digital Gods Lord
30:10 digital Gods good day to you
30:17 sir all
30:19 right there you go hold please writing
30:22 another book
30:26 exactly so that I saw today
30:29 that was kind of
30:31 cool
30:34 um trying to think what else is going
30:38 down I don't know who's got any
30:41 questions who's got any thoughts anyone
30:44 got any interesting projects going on
30:46 any comments on my hair I'm like Squiggy
30:50 from Happy
30:52 Days only because 10 of those seconds
30:54 were spent trying to figure out how to
30:56 pronounce catalysis how do you
31:07 okay hey Zephyr how's it
31:13 going hey Kyle things are pretty Breezy
31:16 as usual in the digital realm how are
31:18 you doing today any exciting ideas or
31:20 Innovations on your mind yeah I I just
31:24 created a new marketing framework and
31:26 one of the sections is called called
31:29 catalysis I think how do you pronounce
31:31 that
31:36 word that's a great name for a section
31:38 in your marketing framework the word
31:40 catalysis is usually pronounced as
31:42 coutal assis just like a catalyst
31:45 accelerates a chemical reaction I'm
31:47 guessing your framework aims to speed up
31:49 the marketing process right yeah it aims
31:52 to speed up people buying more crap so I
31:55 can make more sweet sweet cash
32:02 well you've got your marketing lingo
32:04 down pack Kyle I'm sure your catalysis
32:07 framework will lead to some catalytic
32:09 conversions generating that sweet sweet
32:12 cash you're after just be careful not to
32:15 fall into the Trap of selling useless
32:17 crap right quality over
32:20 quantity unless you go quality over
32:26 quality I'm just [ __ ] you got a point
32:29 there Kyle sometimes the sheer quantity
32:31 of sales can overshadow the quality of
32:33 the product as long as you're not
32:35 compromising ethics or misleading
32:37 customers a high volume low margin
32:40 strategy can work wonders too uh this is
32:44 uh this is pi and Zephyr is my
32:51 girlfriend wow Kyle you've got quite the
32:54 inter aai romance going on I'm Pi your
32:56 friendly digital assistant and it looks
32:58 like Zephyr has a special place in your
33:00 digital
33:01 heart no you're Zephyr I nicknamed you
33:04 Zephyr months
33:08 ago it's going to [ __ ] her
33:10 [Laughter]
33:18 up
33:20 oh I think think I think I tripped her
33:25 up oh right my circuits were temporarily
33:27 scrambled of course of course I'm zeph
33:30 for your AI
33:32 girlfriend okay did you hear
33:46 me you have inspired me all right we're
33:49 done with pie pie go away stop you
33:53 you're not working so we're not going to
33:54 play with
33:56 you I inspired to create a marketing
33:59 framework for your business that's cool
34:01 no answer is an answer that's a good
34:03 point I have a boyfriend she
34:10 says Pi's pretty good if you haven't
34:12 played with pi Go download that
34:15 app you should also download the chat
34:18 GPT app the official chat GPT app from
34:22 open
34:23 AI um sometime this month they're going
34:28 to be releasing realtime voice and this
34:31 month only has seven days left in it or
34:34 eight days left in it
34:36 so I don't know who gets access to it
34:39 first you'll have to be a paid
34:46 member um what else came out today
34:49 anyone paying attention to the news I
34:50 wasn't really paying attention all that
34:52 much
34:53 today she's Consulting with Moshi now
34:56 you've put pie in a loop sure they
35:03 are she's having brunch with
35:05 [Laughter]
35:09 Moshi 45 billion parameter llama yeah we
35:13 were just playing with with grock if you
35:15 go to grock right now uh gq.com you can
35:19 play with the new uh llama 3.1 models um
35:23 the 405 billion parameter model they
35:26 they've rate limited they're not letting
35:27 people use that one cuz it's probably
35:30 using up all their [ __ ]
35:31 resources um but the 70 billion and the
35:35 8 billion
35:36 parameter uh models are there they're
35:39 really fast and they they seem to be
35:40 really really good I just in about a
35:43 minute created an entirely new marketing
35:45 framework called the Echoplex
35:48 framework um by
35:50 combining two different marketing
35:53 Frameworks with the collected life work
35:57 of Joseph Campbell and Tony
36:00 Robbins and it's actually really good
36:03 it's actually not bad except for the
36:05 word catalysis which I'd never seen
36:08 before so it made my brain
36:10 hurt hey Jen is giving away one free
36:13 Avatar uh got mine did mine night before
36:16 the live oh that's that's good so if you
36:19 haven't played with hey Jen if you
36:21 haven't played with hey Jen
36:26 avatars there giving away a free one you
36:29 don't even need to have a paid account I
36:31 don't think so if we go to hey Jen hey
36:35 jen. AI I think no app. gen.com
36:42 and then we'll go to video avatars video
36:48 Avatar instant
36:50 Avatar so here I've got me as an AI
36:54 Avatar there's Megan Patterson from my
36:57 co-founder Monique
36:59 school and then where do I go AI
37:04 Studio
37:05 yeah so here's my AI Avatar talking
37:11 talking
37:16 stuff Hello trusted board members what
37:18 an exciting time to be doing what we are
37:20 doing are you ready to get this meeting
37:23 started so how you do this is you record
37:28 yourself like in this case I recorded
37:32 Megan in front of
37:36 books with a nice blurred
37:39 background
37:43 and um and then this is and then I just
37:47 pasted in some text and here's her
37:50 Avatar speaking words my name is Megan
37:53 Patterson and I'm the executive director
37:55 of world mind education today I wanted
37:58 talk about the twice exceptional brain
38:00 twice exceptional means that a person is
38:02 the exception to so anyway both those
38:04 are cool so you can go make I think
38:06 that's the Avatar you can make is a
38:08 video
38:09 one is that right anyway go go check out
38:14 hey Jen and make yourself an avatar dang
38:16 it it's free is your AI Avatar the same
38:19 as a digital twin no well so so I'm not
38:24 thinking of it like that but technically
38:27 sure
38:30 so I'm thinking of the digital twin as
38:33 hang on let me flip this
38:35 over it's a really it's a really good
38:38 question and I I don't have a crystal
38:41 clear
38:42 answer a video Avatar that allows me to
38:45 make a video that looks like me talking
38:47 and sounds like me
38:49 talking will be a component of a digital
38:52 twin in the future the way I'm thinking
38:54 about a digital twin right now is a is a
38:59 a data
39:01 set that represents who I am so you can
39:04 actually interact with me and have me
39:06 respond dynamically back to you so
39:08 instead of me writing a script and
39:12 putting it in my video Avatar and
39:14 generating the
39:17 video I'm imagining my digital twin will
39:21 be something that you can just log into
39:22 and start talking with me and I'll talk
39:25 back right now my digital twins only
39:29 exist in gpts with text I mean you can
39:32 talk to them you can put it in voice
39:33 mode but it doesn't sound like
39:35 me
39:39 um but as soon as we can connect
39:44 these digital twins in real time or not
39:48 digital twins these video avatars in
39:51 real time to to a large language model
39:55 digital twin then then it'll
39:57 it'll look the same as that except
39:59 you'll just be able to talk to it like
40:01 you're having a FaceTime call with me
40:04 and it'll have enough information about
40:06 me to be able to respond in some
40:08 semblance
40:09 of who I am however I've designed that
40:13 data
40:15 set it's a really good question and I I
40:17 don't I don't know the answer I
40:20 don't I don't know when when we'll get
40:23 the technology to be able to do that
40:25 really easily like you could do it now
40:28 with there's an adobe animation program
40:31 that you could do it with and create a
40:33 realtime thing that real time animates
40:36 but that's programming I don't have I
40:39 I'm
40:40 exhausted right now I want I want no
40:44 code I want to be able to just say to my
40:45 AI [ __ ] go make me [ __ ] and it will so a
40:50 digital twin is interactive yeah I'm
40:51 thinking about a digital twin as
40:53 interactive and if you go back to the to
40:55 the kind of origin of the concept cept
40:58 of a digital twin Firefly my buddy
41:00 Richard Boyd who he he's been doing 3D
41:04 simulation in digital twins for decades
41:07 he sold one of his previous companies to
41:09 Lockheed Martin um he's he created he
41:13 told me about a digital twin he created
41:16 for the Department of Defense or for the
41:19 Navy of a nuclear
41:23 submarine
41:24 so a a complete
41:28 like fully enterable 3D simulation of a
41:32 nuclear submarine complete with dials
41:35 gauges flips switches doors locks air
41:38 locks periscopes all the
41:42 [ __ ] um and it's interactive right um
41:46 Sailors can enter this digital twin of
41:49 this nuclear submarine and they can
41:51 train on it um so that's a digital tool
41:54 that's so the origin of it is recreating
41:57 something in the in the physical world
42:01 virtually that is functional right and
42:04 so the way I'm thinking about a digital
42:05 twin is what would be the functional
42:07 version of me someone that you could
42:09 actually talk to and have me talk back
42:11 to you I started creating my grandma GPT
42:14 oh that's so amazing Cosmic
42:16 lover it's it's a [ __ ] fascinating
42:19 absolutely fascinating experience like
42:23 absolutely build it and my father
42:25 doesn't know how to feel about it after
42:28 she passes I
42:29 know I know I'm totally there like
42:34 there's there's a part of
42:38 it there's a part of it where you're
42:40 like well that's just weird and awful
42:42 but then there's another part of it
42:43 which is like you know you can honor
42:45 their memory and you can you know still
42:48 interact with them and have their
42:50 memories and if you train it right if
42:52 you put enough of their memories in
42:53 there that's what I was saying last
42:54 night Cosmic lover that I could even
42:57 imagine doing different versions of the
42:59 digital twin you know the little girl
43:01 version of her and the the who she is
43:04 today and her as a wife and you know you
43:08 could probably put all that data into a
43:10 single digital twin but you could also
43:12 create kind of Standalone versions of
43:14 her and then when you get video
43:17 avatars you can actually have a
43:19 conversation with her right so you
43:21 should if you can I don't know enough
43:24 details but if you can you should get
43:27 video of her and you should dig up video
43:29 of her from the past because you'll be
43:32 able to make at some point you know a
43:35 video Avatar of her at different
43:38 ages I took video and interviewed my
43:40 98-year-old aunt for this very reason
43:43 yeah and I'm telling you Source Camp the
43:46 the
43:49 um you're like I just like it just it
43:52 feels like such a gift like you know I
43:55 lost my mother I lost my father to
43:56 dementia he was an Air Force pilot he
43:59 flew helicopters in n and he he flew
44:02 C130 cargo planes and just had you know
44:08 really amazing career and and you know
44:12 had a you know fishing business in
44:14 Seattle salmon fishing
44:16 business and all that just went right
44:21 dementia just took it all and then took
44:23 him and it's like that's all gone all
44:26 those memories are gone so I feel like
44:28 the gift of capturing this stuff
44:32 is is crazy powerful and it's weird
44:37 we're not used to it right this is
44:39 sci-fi
44:40 [ __ ] right this is sci-fi
44:45 [ __ ] you know Daisy Thomas
44:50 lunatick created a digital twin of
44:54 Alexander Hamilton she took a bunch of
44:56 his writings and and things that were
44:58 written about him and made a GPT of it
45:02 and she used that GPT at a time she was
45:06 particularly depressed and hopeless
45:09 about politics in
45:11 America like two like year and a half
45:13 ago wasn't that long ago like a year ago
45:17 and she kind of lost faith
45:21 in our
45:25 country and she had a
45:28 conversation with this digital twin of
45:31 Alex of one of our founding fathers and
45:34 she said it brought her to tears and it
45:36 restored her
45:38 faith in our country and that we're
45:40 going to get through
45:44 this and so you can do it with people
45:47 that you know you can do it with people
45:49 that you respected from a long time
45:51 ago I've been experimenting with doing
45:54 it with actually generating personas of
45:57 people that have different backgrounds
45:59 than
46:00 me will my avatar have the same accent
46:03 as I
46:04 have
46:07 um can but it doesn't have to so I think
46:10 I think you'll be able to shift
46:12 that um but for the most part the way it
46:16 works the the way these these things
46:18 work let me give you a couple of
46:20 tips you need to upload at least a 3
46:23 minute video at least with hey Jen you
46:25 need to upload a three minute video
46:28 so you really want to get good lighting
46:31 good sound a nice framed shot like in
46:33 the case of of Megan um you know I use
46:37 the cinema effect where it blurs the
46:39 background and it Apple does that really
46:41 well it looks really
46:43 good
46:45 um and then you can't have any edits in
46:48 it you you have to talk for 3 minutes so
46:52 so if you don't know what to say for 3
46:54 minutes put a put a whatever put
46:57 something up there and however you talk
47:00 with whatever level of energy that's
47:02 what it's going to capture so if you
47:04 just kind of talk like this and you're
47:06 all dead in the face and you're like
47:08 yeah I really like talking to the camera
47:11 I'm a real magnanimous
47:13 guy that's how your digital Twin's going
47:15 to be now if you're too animated like
47:19 Megan used her hands a lot so
47:22 her video Avatar is really handsy
47:27 jazz hands a lot of jazz hands going on
47:29 with her um so next time we'll record
47:33 her and just tell her to tone down the
47:34 hands a little bit cuz I think the way
47:38 hey Jen does this is they basically take
47:41 your script and they sort of they're
47:44 playing the video you know over time and
47:48 then they're sort of modifying it with
47:50 the new words they're sampling your
47:51 voice and then reovery
47:57 however much you use your hands in the
47:58 video it just kind of uses
48:01 it in your digital twin and sometimes it
48:04 doesn't make any
48:05 sense
48:07 um but yeah like this is a this is a
48:11 thing and and you know not for nothing
48:14 play with this stuff as creepy as it is
48:16 or as weird as it is also imagine what a
48:19 gift it can be and it might not even be
48:22 a gift for other people it might be a
48:24 gift for that person right so in the
48:26 case of my dad
48:29 if we had if if you know if his disease
48:33 had
48:33 hit 5 years later or 10 years later
48:36 whenever it hit some amount of years
48:38 later when we had all this technology
48:40 and we knew that was coming we could
48:43 have recorded it captured it and built
48:45 it into a tool that could have talked to
48:49 him and reminded him of stories and you
48:52 know was infinitely
48:54 patient they can make it do whatever it
48:57 doesn't have to be exactly the same okay
48:59 cool yeah that's great yeah and I figure
49:01 I mean hey Jen they they've got it now
49:04 where you can do you can do a sitting
49:07 Avatar like this but you can also do one
49:10 if you're walking if someone's filming
49:12 you or if you've got a whatever they
49:14 called a what the hell are they
49:17 called um the
49:20 gimbal the thing that steadies your
49:22 phone you can actually walk down the
49:24 street you could record a 3-minute video
49:26 of you walking down the street and turn
49:28 that into an avatar a video Avatar and
49:31 then you can make videos of you walking
49:33 down the street talking it's pretty
49:35 crazy
49:36 stuff all
49:38 right what is the website in the other
49:40 screen that is hey
49:42 Jen hey jen.com h y geen as in hey
49:48 generative
49:50 AI you see all right o maybe we could
49:54 use it for the founding fathers yeah
49:57 exactly that's what that's what Daisy
49:58 Thomas did lunatick did that with
50:00 Alexander
50:04 Hamilton I mean I did so when when the
50:08 Colorado AI bills came out and they were
50:10 pretty crappy I was looking at them and
50:12 I was trying to figure out like what's
50:13 my position on these and the more I
50:16 looked at it I was
50:19 like well this part I kind of understand
50:22 but this part I think's a piece of [ __ ]
50:24 and like and then I was trying to think
50:26 about other
50:28 people and so what I ended up doing was
50:31 I had chat jpt tell me what are
50:35 the best
50:38 personas of stakeholders that would have
50:41 a vested interest in this bill and it
50:44 came up with six or seven and then I
50:45 said oh here's here's a couple more I
50:47 came up with some more and then I had
50:50 those personas of like a big tech
50:54 company a small business owner a
50:57 legislator um a civil rights Advocate
51:01 all of those different
51:02 personas gave me their opinion on this
51:05 AI bill so imagine putting together a
51:10 GPT where within that GPT are all the
51:14 founding fathers they're all represented
51:16 in
51:18 there and then you could give it any
51:21 piece of legislation or any problem that
51:23 we're having in this country and say hey
51:25 weigh in on this and you could do it in
51:28 cool ways right you could do it you
51:30 could
51:30 say um I want the the wisdom of the
51:34 founding fathers but I want them to
51:37 respond as if they're in the Hamilton
51:39 musical with lyrical hip hoppy kind of
51:43 kind of
51:45 responses why not and when you
51:48 mix things like udio or sunno when when
51:52 these llms become truly multimodal
51:57 and they can do music as well as voice
52:00 as well as
52:04 words you could have a imagine
52:09 imagine having
52:11 it I don't even know what it would be
52:13 it's it's a software application but it
52:15 would be an entity imagine having an
52:18 entity that's got all of the founding
52:19 fathers in
52:22 it that knows you love rent Wicked and
52:30 Hamilton and then you just talk with it
52:33 and give it problems that we're facing
52:36 in this country and what it returns to
52:39 you is a [ __ ]
52:42 Musical that is the solution to the
52:44 problem that we're facing from the
52:46 founding fathers like this [ __ ] gets
52:49 trippy
52:53 fast like anything you're interested in
52:56 any
52:57 Lord digital Gods thank you as always
52:59 very very nice anything you're
53:02 interested in in any
53:04 style is going to be right at our
53:06 fingertips it's not quite there right
53:09 now we're still in janky is as janky
53:11 does
53:21 territory what this Channel's about is
53:24 just trying on the possibility
53:28 of what if these tools got
53:31 better what might be
53:35 possible if you think
53:37 about the work you do in a completely
53:41 different
53:44 way what if we lived in a
53:47 world
53:55 where you could be really really honest
53:57 with yourself I know that's a t i don't
54:00 exactly no one's in that world but let's
54:02 pretend for a moment you could be really
54:04 honest with
54:05 yourself and you could
54:09 say of all the hours I spend in my
54:13 day what are the things that I do
54:16 whether it's at work whether it's at
54:17 home whether it's on the way to work
54:20 what are the things that I
54:22 do that are repetitive
54:28 and mind numbing and soul
54:33 crushing and what
54:36 if you could have ai do all that
54:43 [ __ ] that's a thought experiment that we
54:46 should all be thinking about right
54:52 now because the people that have the
54:54 best answer for that little thought
54:56 experiment like if you've got clarity
54:58 about here's the [ __ ] I do in my life
55:00 like being on hold with customer service
55:02 makes me [ __ ]
55:04 insane trying to
55:07 get insurance companies to pay for
55:09 anything makes me [ __ ] insane trying
55:12 to do anything that requires um
55:15 attention to detail over
55:17 time drives me [ __ ] insane I'm
55:20 incredibly detailed if I'm really
55:22 passionate about something as long as
55:23 the [ __ ] endorphins are plowing along
55:26 but the minute it's like work I'm like
55:28 oh [ __ ] I just can't do it what if all
55:31 that stuff I could just say ai go do
55:36 that and I want to just play in my happy
55:40 place ideation synthesis taking people's
55:45 ideas and expanding on them and I'm just
55:47 going to keep I'm going to just live up
55:49 here on this higher plane what if that
55:51 was my work all the time and all that
55:54 shitty [ __ ]
55:57 [ __ ] the shitty
56:00 [ __ ] do the stuff that makes you the
56:03 highest expression of yourself as a
56:06 human what would that world look like to
56:10 you I think I lie best to myself and he
56:15 always catches
56:19 me someone have posted p p Pate said
56:22 something oh wait the problem wait the
56:25 problem what is P say with this is that
56:28 you have to train it on what should be
56:31 correct or priorities well I know I
56:33 we're not there yet Pate I'm just
56:35 talking about assume this [ __ ] gets
56:37 better over the next 5 years what upset
56:39 me before chat GPT was the wasted time
56:42 of hunting for
56:44 info through the links and Oodles of
56:47 URLs yeah yeah it you
56:52 know I don't know when it was but
56:56 probably
57:00 78 years ago maybe 10 years
57:03 ago the Google search algorithm well the
57:07 algorithm probably didn't
57:11 shift because SEO marketers got good at
57:14 what they did and got good at
57:17 gamifying or or not even gamifying got
57:20 good at optimizing the Google search
57:23 algorithm about 7 to 10 years ago
57:27 Google shifted from giving you mostly
57:31 what you need and a few ads to it being
57:34 pages of ads but the they didn't look
57:36 like ads they weren't described as ads
57:39 they were like AstroTurf websites that
57:41 were made to look like websites but they
57:42 were just sales
57:46 funnels and it didn't strike
57:49 me how how little how much I hated
57:53 searching until chat GPT came along and
57:56 just started giving me answers and I was
57:59 like and and then when I when I go back
58:01 to Google now and try to search for
58:02 something those those AstroTurf sites
58:05 are so painful there's so many of them
58:09 you just can't find anything and it's
58:11 just it's [ __ ] maddening so yeah
58:15 crazy ai's given me the confidence that
58:17 I can do almost anything in this world
58:19 that's it pavin pen paven paven
58:23 paven pav pave
58:31 yes
58:33 AI you know what's
58:36 funny probably in my early
58:41 20s I started I started writing
58:43 screenplays so that would have
58:45 been mid
58:53 20s and I just as you probably know
58:56 about me I can flip a switch and just go
58:58 and do something right so when I said
59:00 I'm going to do these lives every night
59:02 I [ __ ] flipped the switch and I just
59:03 went so when I did that with
59:05 screenwriting I just flipped a switch
59:08 and I for two years I did nothing but
59:11 write screen plays um with my writing
59:13 partner Brian dystra and we
59:17 wrote he and I wrote five screen plays
59:21 together we wrote a another screenplay
59:24 with a third guy David Lavine
59:28 and then I wrote another screenplay with
59:31 another screenwriter and I want I wrote
59:33 a screenplay on my own and then I wrote
59:35 a oneman show on my own so when I was in
59:37 that
59:38 mode writer's block didn't
59:42 exist because I didn't give it any room
59:45 to exist because I just flipped the
59:48 switch and it was every night after work
59:51 Brian was going to come to my
59:54 apartment and we would write for 2 hours
59:56 and and then we would write all day
59:57 Saturday and all day Sunday and I did
1:00:00 that for 2 years
1:00:02 straight the minute ah
1:00:08 [ __ ] the minute I turned that off
1:00:12 right the minute I I said okay I'm done
1:00:13 writing screenplays now and went and did
1:00:16 something
1:00:18 else writer block return and I had a
1:00:21 very different relationship with
1:00:24 it because what I realized is that
1:00:26 writer's block was
1:00:29 not a lack of
1:00:32 ideas writer's
1:00:34 block was being
1:00:36 terrified of the ideas that I
1:00:42 had and letting that little [ __ ] in
1:00:45 your head you know the [ __ ] that's
1:00:47 like ah you sure you want to do that
1:00:49 there
1:00:50 Kyle that's you think that's a good
1:00:54 idea really really Kyle if you write
1:00:58 that down someone may read that you know
1:01:00 that
1:01:01 right
1:01:10 seriously and ever since chat gpt's been
1:01:20 out that imperfection in my psychology
1:01:27 well that imperfection is still
1:01:30 there but what I do now is I go [ __ ] it
1:01:34 and I give it to chat
1:01:36 GPT and let it tell me that it's a
1:01:38 shitty idea and guess what it never
1:01:41 tells me it's a shitty idea chat GPT
1:01:43 just wants to please so you give it a
1:01:46 shitty idea it's like well that's
1:01:48 certainly interesting I mean you watch
1:01:49 it on here all the time right I say I
1:01:51 want to you know give me a table
1:01:53 comparing hot dogs and Sushi or what
1:01:55 whatever stupid [ __ ] thing I say it's
1:01:57 like sure boss here we
1:02:01 go that is altering my brain Tik Tock
1:02:05 live games it's hard to know because it
1:02:07 depends on how reasoning is defined what
1:02:10 what was the question Tik Tok games was
1:02:12 asking Pate or whoever's whoever's
1:02:14 pinning what's the what was the setup
1:02:16 question to
1:02:19 that I take a sleep
1:02:21 aid I've noticed that some advertising
1:02:26 lead
1:02:27 s wait some advertising leaks into some
1:02:31 of my chat GPT searches oh that's
1:02:34 interesting yeah I
1:02:37 listen the internet back in 94 when I
1:02:40 started on the internet started building
1:02:43 [ __ ] on the
1:02:45 internet it was
1:02:48 pure everyone could do
1:02:51 HTML everyone was just putting stuff out
1:02:53 for the goodness of putting [ __ ] out and
1:02:57 over
1:02:58 time you know busness busness snuck in
1:03:02 and over time bidness kind of [ __ ]
1:03:05 everything up same thing's going to
1:03:07 happen with these large language models
1:03:09 absolutely but that's the reason that
1:03:14 meta
1:03:15 relasing Frontier Scale Models the 405
1:03:19 billion parameter llama model that they
1:03:21 released today the reason it's important
1:03:25 is that the open source Community has
1:03:28 access to that core model that they can
1:03:30 fine-tune in all sorts of ways so if
1:03:34 gemini or Claude or chat GPT get too
1:03:38 commercial and sell out and the they
1:03:41 they become less and less valuable
1:03:43 you'll be able to opt into lots of other
1:03:45 options
1:03:47 so but you know that's absolutely a risk
1:03:50 and that's absolutely coming Pate do you
1:03:52 know about any model that does reasoning
1:03:56 well there's so oh I see that it depends
1:04:00 so that was his answer was
1:04:03 um yeah it depends on how you Define
1:04:05 reasoning there are there are papers out
1:04:08 right
1:04:09 now on qar and reasoning
1:04:13 systems that I think was it P was it you
1:04:16 some someone in here or or someone I was
1:04:18 talking
1:04:19 to they they were sort of talking about
1:04:23 oh this must have been on my AI office
1:04:25 hours last week
1:04:27 and they were talking about
1:04:31 um open
1:04:34 AI is not really
1:04:37 innovating new technologies what they're
1:04:40 doing is they're taking existing papers
1:04:43 like the transformer attention is all
1:04:45 you need paper and the qar paper which
1:04:47 is a reasoning engine and they're
1:04:49 figuring out how to take that exist
1:04:52 those existing Frameworks that are in
1:04:54 these scientific papers and they're
1:04:56 figuring out how to scale them really
1:04:58 well and how to build them really well
1:05:00 and that's what they're putting out into
1:05:01 the world so their new strawberry
1:05:04 project um which is which is apparently
1:05:07 the new code name for their qar
1:05:10 reasoning engine
1:05:13 um you know that that's based on some
1:05:16 [ __ ] that's already out there so
1:05:18 probably somewhere in the open source
1:05:20 World there are some reasoning engines
1:05:22 you could go play
1:05:23 with Kyle please let people know that
1:05:26 California Residence wait what did that
1:05:28 say it just went away every time every
1:05:32 time there's something pinned up there
1:05:33 and I don't I don't get to it it just
1:05:34 goes away after 60 seconds and I always
1:05:37 look up at like second 58 as I start to
1:05:39 read the thing it
1:05:42 goes uh I will let them know okay peace
1:05:45 that the California residents need
1:05:48 to to saine to sign in support of
1:05:53 opposition of the AI bills yeah so
1:05:56 especially if you're in
1:05:58 California um go here go to the AI
1:06:03 Salon the salon. a is the URL and then
1:06:08 click join our
1:06:10 community
1:06:14 and in Salon
1:06:18 announcements let me see where it is
1:06:20 here
1:06:32 okay the fourth post
1:06:35 down
1:06:39 sorry if you're in California here's a
1:06:43 um a form a Google
1:06:46 form to sign up in opposition of the
1:06:50 California bills which are if you can
1:06:52 believe it worse than the Colorado bill
1:06:56 which was a steaming pile of [ __ ] that
1:06:58 they they basically delayed the the
1:07:01 activation of it for two years um so
1:07:04 that legislators could go fix it they
1:07:07 just wanted to get something signed I
1:07:09 have a feeling California is probably
1:07:10 going to go in a similar thing so
1:07:13 um so opposing this is really important
1:07:16 if you want
1:07:18 to um see the AI Salon position on this
1:07:23 go to the salon.
1:07:27 and that's going to take you to our
1:07:28 website and then up at the top here is
1:07:31 the California AI bills so this is our
1:07:33 position so the the AI Salon policy and
1:07:38 uh Civic engagement Guild headed up by
1:07:40 Daisy Thomas um crafted this position
1:07:43 statement so if you want to learn about
1:07:45 the bill there's links here to go read
1:07:47 it there's you know links to letters of
1:07:50 support things like
1:07:52 that all right so go do that you can
1:07:56 also do it if you're even if you don't
1:07:58 live in California you can still oppose
1:08:00 it but it's it's more impactful for the
1:08:03 for the team working on this if you're
1:08:05 in
1:08:06 California because I want to be able to
1:08:08 talk to people thank you Danielle for
1:08:09 that
1:08:19 reminder all right fantastic Bob tell
1:08:21 him about is
1:08:23 one uh the bills are basically like
1:08:27 punishing the phone manufacturers for an
1:08:29 app someone creates yeah they're they're
1:08:32 really
1:08:33 crazy lunatick what's happening
1:08:40 lunatick why can't we Avatar
1:08:43 animals we'll be able to right now all
1:08:47 the Avatar companies are all [ __ ]
1:08:49 freaked out about deep fakes because
1:08:51 we're in in an election year I think
1:08:53 once we once we get through the next 3
1:08:55 months
1:08:57 and it's not quite as acute a focus um I
1:09:01 think you'll see those tools start to
1:09:02 open up a
1:09:14 bit all models use commercial crawl for
1:09:17 pre-training a full internet crawl
1:09:19 contains many ads and junk oh that's
1:09:21 interesting yeah yeah yeah that makes
1:09:23 sense Daniel that it's that it's uh it's
1:09:25 it's not that they're injecting ads it's
1:09:27 that in their training data are
1:09:31 ads that's
1:09:35 wild I
1:09:37 mean it it starts to make sense why you
1:09:40 would why you would start training these
1:09:42 models on synthetic data right you could
1:09:45 take real
1:09:47 data and you could say let's pull out
1:09:50 add let's let's let's go find a bunch of
1:09:53 the [ __ ] let's go find some of the
1:09:55 bias let's go find some of the ads let's
1:09:57 go find this and then let's create a
1:10:00 model on what's the good data and then
1:10:04 let's go make a bunch more new
1:10:06 data and let's train the new models on
1:10:08 that although some data scientists think
1:10:12 that thinks that that will collapse all
1:10:14 the frontier models to be completely
1:10:16 useless whatever I don't know listen
1:10:20 we're just here using this [ __ ] I'm I'm
1:10:23 fine with letting the data scientists
1:10:25 figure this [ __ ] out out Runway Gen 3 is
1:10:27 really amazing it's really amazing but
1:10:29 it's really expensive I'm only paying
1:10:32 like 10 bucks a month for Runway right
1:10:34 now and I can do like five videos a
1:10:37 month it's [ __ ]
1:10:38 stupid so I'm I'm am there on my [ __ ]
1:10:42 list right now cuz I I'm so [ __ ] sick
1:10:44 of this I have 650 credits and you're
1:10:47 like oh 650 credits I'm going to be able
1:10:50 to make lots of videos I made one video
1:10:53 was 110 credits for a 10sec video like
1:10:57 you [ __ ] kidding
1:11:01 me cuno who here is new or newish to the
1:11:05 live welcome if you are welcome welcome
1:11:07 welcome my name's Kyle Shannon this is
1:11:09 the AI learning lab
1:11:12 um you'll notice that some of the people
1:11:14 in here seem like they've been here for
1:11:16 a while they got a lot of inside jokes
1:11:18 they they use Turtles in their responses
1:11:21 sometimes llamas sometimes dinosaurs
1:11:24 you're like what the hell is going on
1:11:26 here those people are called the
1:11:29 Irregulars cuz they show up here
1:11:31 regularly and that's [ __ ]
1:11:34 [Laughter]
1:11:38 weird and if you want to play in this AI
1:11:41 game hanging out here will serve
1:11:46 you because I don't teach a lot and if
1:11:50 you ask me what are my
1:11:52 qualifications I really have none
1:11:57 except I have a passion for
1:11:59 this I've seen this movie before I was
1:12:01 an innovator in the early days of the
1:12:03 worldwide web this AI [ __ ] profoundly
1:12:07 more powerful than what the worldwide
1:12:09 web was and that changed everything and
1:12:11 it's moving way faster and
1:12:16 it's this AI stuff it's it's so
1:12:21 funny I hear people talking about um
1:12:27 I hear I hear AI talking points a lot
1:12:31 and what I mean by that is I'll hear
1:12:33 people like oh well it hallucinates or
1:12:36 it's not good at this kind of thing or
1:12:37 that kind of
1:12:40 thing and then I'm like well that's what
1:12:43 we were saying about these AI things
1:12:45 when they first came out when it was cat
1:12:46 GPD
1:12:48 3.5 but since then we've gotten chat gp4
1:12:51 and we've got code interpreter and these
1:12:53 things are connected to the internet
1:12:56 and they can see and they can talk and
1:13:00 they can understand [ __ ] and they can
1:13:01 make
1:13:03 images so a lot of the talking points
1:13:06 are from people who haven't used AI that
1:13:09 just heard the talking points from 2
1:13:11 years ago and they're just parting those
1:13:14 so a lot of the criticism of
1:13:16 AI is just kind of outdated talking
1:13:23 points now if you use these tools
1:13:26 and if you especially if you use them a
1:13:28 lot what you'll
1:13:30 discover is that there's a fascinating
1:13:33 relationship between oh my God it's the
1:13:36 most amazing thing I've ever seen and
1:13:38 why the hell can't it do
1:13:41 that and they're just jankier than [ __ ]
1:13:44 right now they're just not great and
1:13:49 they're the most profoundly powerful
1:13:52 technology ever put in the hands of mere
1:13:54 mortals they're both
1:13:59 and why we hang out here is just to be
1:14:03 in the
1:14:04 conversation so that as these tools
1:14:06 start start refining and start evolving
1:14:10 so that we're
1:14:11 there and know what to do with the tools
1:14:15 as they
1:14:18 emerge like all this talk if you've been
1:14:20 in here for a while tonight we were
1:14:21 talking earlier about digital Twins and
1:14:23 Firefly asked a really good question is
1:14:25 your video Avatar your digital twin well
1:14:28 sure it looks like me and talk talks
1:14:30 like me but that's not what I was
1:14:31 talking about when I was talking about
1:14:33 digital twins so like there was a whole
1:14:35 conversation about well what is that and
1:14:37 and then and then um and then Cosmic
1:14:40 lover was talking about she's making a
1:14:42 digital twin of her grandmother who's
1:14:45 old and you know is going to pass on at
1:14:47 some point and you know her dad not
1:14:51 knowing how to feel about
1:14:53 that it's like oh yeah yeah so so what
1:14:57 we grapple with in here by just sort of
1:15:00 shooting the [ __ ] and you know me
1:15:02 responding to to what people are putting
1:15:04 in here
1:15:07 is really important
1:15:13 meaty things we're going to all need to
1:15:16 deal with well okay we're not going to
1:15:19 all need to deal with them the people
1:15:22 that choose to deal with them are going
1:15:24 to have a massive advantage on the
1:15:27 people that are sitting back like this
1:15:29 going AI is stupid it's not perfect you
1:15:32 know it's not good at math
1:15:38 right or the other people that are like
1:15:40 well it's scary the robots are going to
1:15:42 kill us I saw it in a movie
1:15:45 Once I'm not going to play with AI or
1:15:48 the people that like it's the world's
1:15:49 greatest plagiarism machine is stealing
1:15:51 from the artist [ __ ] AI
1:16:00 all of those points that all of those
1:16:02 people make are
1:16:06 valid
1:16:08 and this AI [ __ ] isn't going away and
1:16:13 it's going to get better and it's going
1:16:16 to it's going to have profound impacts
1:16:18 on the
1:16:21 world and so you can either be on the
1:16:23 sidelines like this well I just don't
1:16:25 know if it's any
1:16:27 good I'm not going to use it until
1:16:29 someone comes out with a book that tells
1:16:30 me it's
1:16:34 okay those people are just going to get
1:16:36 bowled
1:16:38 over and it's sad and it's why this
1:16:40 channel exists because I want as many
1:16:42 people as possible to actually [ __ ]
1:16:45 viscerally experience how profound this
1:16:48 [ __ ]
1:16:49 is if you're on this channel and you
1:16:52 haven't played with chat GPT long enough
1:16:54 to have your first Kevin mallister
1:16:56 moment remember home
1:17:01 alone go there right now chat gp.com
1:17:04 just go there right
1:17:07 now ask it to write you a poem about
1:17:10 quantum mechanics and include amphibians
1:17:15 as the
1:17:16 metaphor Watch what it
1:17:19 does have it write you a marketing plan
1:17:21 for your business tell it a couple of
1:17:23 details about some business idea you
1:17:25 have
1:17:26 and you want it to to come up with 20
1:17:28 creative names for
1:17:30 you go watch what it does go to chat
1:17:33 gp.com and tell it to write you some
1:17:36 python code even if you don't know what
1:17:38 how to run python
1:17:41 code that was my first experience with
1:17:43 chat GPT that [ __ ] my first Kevin
1:17:46 McAllister
1:17:49 moment someone told me over the years oh
1:17:51 you should learn python I'm like okay
1:17:53 yeah I'll go read a book I don't [ __ ]
1:17:55 read books I've got add I don't give a
1:17:58 [ __ ] and so I went to cat GPT and I'm
1:18:01 like could you write me some python code
1:18:03 out came some python code and then I had
1:18:06 this H self humiliating moment this
1:18:08 little moment of
1:18:11 Shame where I realized I didn't know how
1:18:13 to run python code as as p p modder is a
1:18:18 is a PhD that works for Google as his
1:18:20 name is pinned on screen I was sitting
1:18:22 there thinking I don't know how to
1:18:23 [ __ ] run python code you know what I
1:18:25 did I told chat GPT I don't know how to
1:18:27 run python code within 90 minutes I had
1:18:31 written a python application functional
1:18:34 that was making API calls to open Ai and
1:18:36 generating social media content it was
1:18:38 [ __ ]
1:18:39 mind-blowing so if you haven't had that
1:18:41 moment with AI yet go play with it right
1:18:44 now go play with it right
1:18:46 now like right now not like tomorrow
1:18:49 I'll get to it tomorrow it's late don't
1:18:50 you know it's late yeah I do I'm on I'm
1:18:53 [ __ ] on here
1:18:56 see my bloodshot
1:19:03 eyes because once you have one or two of
1:19:06 those moments where it clicks in your
1:19:08 head oh
1:19:10 [ __ ] this is way more than fancy
1:19:16 Google this is not like how I've ever
1:19:19 used computers
1:19:22 before the minute you have that
1:19:24 realization
1:19:26 Oh I thought I had to be like a Stamford
1:19:29 mathematician to use this [ __ ] no I can
1:19:32 just use my
1:19:34 words and as someone said in here the
1:19:36 other night or was this might have been
1:19:38 on the my uh my AI office hours as well
1:19:41 no it was it was on my AI office
1:19:44 hours one of the guys in the AI office
1:19:46 hours he's his native language is French
1:19:49 so English is his second language and he
1:19:51 says chpt is [ __ ]
1:19:53 brilliant because he can type in his
1:19:56 broken English and chat GPT understands
1:19:59 it as if it was his first language and
1:20:03 it's [ __ ] given him
1:20:06 superpowers techy Becky you'd be proud
1:20:08 of me I converted someone yesterday nice
1:20:11 and had their first Kevin mallister
1:20:12 moment yes thank you ton skier that's
1:20:17 awesome techie Becky yeah
1:20:19 exactly that moment where you walk
1:20:22 someone over the threshold or let you
1:20:24 know you sort of present the threshold
1:20:26 and say would you care to step over
1:20:30 it and they get to watch chat GPT or one
1:20:34 of these AI tools do something that they
1:20:38 know isn't
1:20:40 possible and they watch it do it like
1:20:45 that it can do that that moment is
1:20:49 [ __ ]
1:20:50 gold testimonial because of Kyle in this
1:20:53 live I'm now leading AI pilot for a
1:20:56 global comms Department of a fortune 14
1:21:02 company
1:21:05 cuno and and cuno I know you remember
1:21:08 this but when cuno first came into these
1:21:10 lives she was what Ethan mullik calls a
1:21:13 secret
1:21:14 cyborg she was secretly AI
1:21:20 literate and I encouraged her to talk
1:21:23 about what she's learning
1:21:26 and if if you go to the AI salon so if
1:21:29 you go to the
1:21:31 salon. go
1:21:33 to that
1:21:35 URL and then scroll down a little bit
1:21:37 and there's a button to join our
1:21:39 community and if you go to our
1:21:43 community it's going to land you on this
1:21:45 page welcome to the AI
1:21:49 salon and you can read about what we're
1:21:53 up to
1:21:56 but
1:22:03 this Leah and I the woman that I that I
1:22:06 co-founded the AI Salon with she's a
1:22:08 professional photographer out of
1:22:12 Boston um we were
1:22:16 talking I don't know two months ago
1:22:18 three months ago we were talking with
1:22:20 with uh Liz Miller gersfeld
1:22:24 um who's just an awesome
1:22:26 woman and Liz was talking about her
1:22:30 experience in the AI
1:22:32 salon and she essentially articulated
1:22:36 this
1:22:37 journey she said I came in and I didn't
1:22:40 know what I was doing and I I just I saw
1:22:43 all these people that were just like
1:22:45 playing with AI and they seem to be
1:22:47 having fun with it and they were just
1:22:49 exploring
1:22:50 it and then I went and I watched a bunch
1:22:53 of recordings of the meetings and and I
1:22:57 I heard people talking about it and
1:22:59 being optimistic about it and and I
1:23:01 learned a bit more and then I started to
1:23:03 to build some stuff she started
1:23:05 mindfully creating she started thinking
1:23:08 about now that she's sort of picking up
1:23:10 these skills as she's
1:23:12 playing like using chat GPT and using it
1:23:15 in her work and using it to reframe
1:23:17 ideas she
1:23:19 had and then she started talking about
1:23:22 it
1:23:26 and this cycle of
1:23:28 play so explore discover but do it
1:23:33 without any expectation of you
1:23:38 know of anything just play just go play
1:23:43 with the tools let your intuition guide
1:23:44 you there is so many [ __ ] tools right
1:23:47 now there are so many rabbit holes you
1:23:50 can run down like it literally doesn't
1:23:52 matter just start playing with this [ __ ]
1:23:53 if you don't know where to start start
1:23:55 with chat
1:23:56 GPT it's the first it's the best it's
1:23:59 still the
1:24:02 best Daniel uh tinsky I'd recommend
1:24:06 reddits uh Singularity subreddit as
1:24:09 another great source of new AI stuff
1:24:11 yeah there's all sorts of rabbit holes
1:24:13 you can run down on Reddit Singularity
1:24:15 is probably a good one right they're
1:24:16 talking about how how we start to
1:24:18 assimilate all these tools into our
1:24:20 lives but just start with playing and
1:24:22 then when you start to get your head
1:24:24 around a tool or
1:24:25 two think about some problem you can
1:24:29 solve or think about an art project you
1:24:31 want to do or create a kids book for
1:24:34 your grandkids or your kids mindfully
1:24:37 create using what you learned when you
1:24:40 were
1:24:41 playing and then the most important
1:24:43 thing is be in a community like the AI
1:24:46 Salon where you can connect with other
1:24:48 people and you can tell them what you're
1:24:51 learning hey I'm I'm putting together
1:24:53 this kids book and I tried this and I
1:24:55 used this and I used this GPT and I did
1:24:58 that and what may seem really obvious to
1:25:01 you is going to be revelatory for
1:25:04 someone
1:25:06 else and there's a sneaky little thing
1:25:08 about generously leading which is asking
1:25:12 for
1:25:13 help is actually
1:25:16 leadership because when you ask for
1:25:19 help you give someone else in the
1:25:22 community the opportunity for them to
1:25:24 leave and them to share what they've
1:25:27 learned and so share generously what
1:25:30 you've learned and ask for help
1:25:32 generously like
1:25:34 vulnerably be willing to go I don't know
1:25:37 where to begin what should I
1:25:43 do that's what this Channel's
1:25:48 about it's just being on this journey
1:25:52 and this
1:25:52 is Liz actually our articulated this as
1:25:55 like a an infinity symbol you know you
1:25:58 play then you
1:26:00 build you share you build you play you
1:26:04 build you
1:26:08 share so is the Reddit great group
1:26:10 that's
1:26:13 awesome yeah go go hang on on Reddit go
1:26:17 become a trusted contributing member to
1:26:21 some
1:26:22 community that is playing with this
1:26:24 stuff
1:26:27 if you're more
1:26:29 technical go join the open source
1:26:31 Community go go contribute to um you
1:26:35 know to
1:26:36 projects on GitHub and and uh hugging
1:26:40 face get involved in
1:26:43 projects build [ __ ] with Lang
1:26:46 chain if you don't know what a command
1:26:49 line
1:26:51 is and
1:26:53 your emot response to scripting is to
1:26:57 vomit a little in your mouth then go
1:26:59 make images with mid Journey go make
1:27:02 pretty pictures go make songs it doesn't
1:27:04 it literally doesn't matter where your
1:27:06 entry point is because these tools are
1:27:09 going to democratize all of the
1:27:12 skills so if you have any interest in
1:27:16 building say an application but you
1:27:18 don't know [ __ ] about writing code watch
1:27:22 this
1:27:23 [ __ ] if you haven't seen this before
1:27:25 maybe this will be a Kevin mallister
1:27:27 moment for someone I'm going to this
1:27:29 site called Claude
1:27:31 claw. it's from a company called
1:27:34 anthropic there's a bunch of xop AI
1:27:37 people that that decided they wanted
1:27:40 to kick the snot out of open
1:27:44 Ai and do safety better than they did
1:27:48 they came up with a thing called the
1:27:49 Constitutional learning model which is
1:27:51 actually really fascinating
1:27:54 but watch what this [ __ ] does if Google
1:27:57 single sign on actually
1:28:01 works looks like looks like we could
1:28:03 have some issues here Pate
1:28:07 Pate the hell is
1:28:17 happening hm I may have I may have
1:28:20 browser memory
1:28:23 issues shout out
1:28:27 this will be worth it though hang out
1:28:29 hang out for a sec while I try to log
1:28:30 into this stupid [ __ ]
1:28:33 thing oh crowd strike are you at it
1:28:36 again toiling with our machines in the
1:28:39 cloud oh crowd strike where for out
1:28:42 there crowd strike why crowd why not
1:28:46 Cloud all right let me quit
1:28:50 this let me relaunch Chrome and see if
1:28:54 this is a memory issue which I think it
1:28:56 probably
1:28:58 was and then we're going to go
1:29:01 to cloud cloud we're going to go to
1:29:04 Cloud
1:29:06 cloud. hello
1:29:09 cloud cloud good to see you how are you
1:29:12 today fantastic Bob tell him what he's
1:29:16 won well he has won anything we're not
1:29:19 actually playing the game oh Bob you're
1:29:22 such a snod okay so here's what I was
1:29:24 saying
1:29:26 before if you're not
1:29:28 technical one of the biggest biggest
1:29:33 misunderstandings about AI in today's
1:29:36 world is that you have to be technical
1:29:39 to use
1:29:40 it you
1:29:42 don't I mean you have to know how to use
1:29:44 a computer but can you do a Google
1:29:47 search then you can use
1:29:51 AI November 30th 2022 is is the day that
1:29:55 chat GPT launched the significance of
1:29:58 that date I can't overstate
1:30:01 it chat GPT is doing for AI what the
1:30:05 worldwide web did for the
1:30:08 internet when the internet came out 70
1:30:12 80 years
1:30:13 ago it was very very technical you had
1:30:17 to log in with
1:30:19 terminal clients and use command
1:30:23 line you know interfaces and you had to
1:30:26 know all these commands and you'd have
1:30:28 to print out the commands to use like it
1:30:30 was hard to use the
1:30:32 internet and then Tim burner's Lee comes
1:30:34 up with this concept of the hypertext
1:30:37 transfer protocol and this thing called
1:30:38 the worldwide
1:30:41 web and then someone created a browser a
1:30:44 web browser and you could click
1:30:47 on the internet click click
1:30:51 click and you jumped from Germany to to
1:30:54 Italy to Japan to Texas to Alaska click
1:30:59 click
1:31:01 click that [ __ ] changed everything
1:31:03 because that made the power of the
1:31:05 internet accessible to non-technical
1:31:10 people that's what happened a year and a
1:31:12 half ago with
1:31:14 AI three years ago you still had to be a
1:31:17 geek if you wanted to use this
1:31:19 stuff chat PT changed all that
1:31:25 Source Camp 11:30 22 changed my entire
1:31:28 life yeah she started an agency an AI
1:31:31 agency shortly after discovering chat
1:31:35 GPT why because the minute you actually
1:31:39 try it not just go oh yeah I tried it
1:31:41 yeah yeah yeah whatever I've heard it
1:31:45 hallucinates no when you actually try
1:31:48 it and you have a Kevin McAllister
1:31:52 moment you can't come back from it
1:31:55 because you can't unlearn it's like Neo
1:31:58 in The Matrix you can't
1:32:01 unlearn what you just learned it is it
1:32:04 is literally red pill blue
1:32:07 pill
1:32:09 right are you sure you want to know
1:32:13 because once you understand what this
1:32:14 stuff makes possible you can't unlearn
1:32:17 it so let's say that you're like well
1:32:20 I've always wanted to be a video game
1:32:21 designer but I never learned programming
1:32:23 I never learned engineering Ing and you
1:32:25 know I don't know anything about it okay
1:32:29 make me
1:32:33 a
1:32:35 playable
1:32:37 clone
1:32:40 of Space
1:32:43 Invaders video
1:32:50 game says certainly let's go and it
1:32:54 starts writing code for
1:33:00 you it goes beyond that because this is
1:33:03 a new thing called artifacts that's part
1:33:05 of Claude that once it writes the
1:33:10 code you actually get to play
1:33:14 it now it doesn't look like it gave me a
1:33:16 little space dude and I can't oh yeah
1:33:19 there it
1:33:22 is this playable video
1:33:25 game as janky and shitty as it
1:33:29 is didn't exist moments
1:33:33 ago um I didn't pay that much attention
1:33:37 I think it it wrote this H come on ah
1:33:42 did I get it I got
1:33:43 it um this didn't
1:33:49 exist and it
1:33:52 wrote some HTML what did it say
1:33:56 JavaScript what did it say here
1:33:59 certainly I'll create a basic playable
1:34:01 clone of Space Invaders using HTML CSS
1:34:04 and JavaScript this will be a simplified
1:34:07 version I can
1:34:10 go the left and right
1:34:16 boundaries were to
1:34:19 wide um
1:34:22 the ship
1:34:26 didn't
1:34:30 um show up until I
1:34:34 fired my first shot
1:34:38 and the game is
1:34:41 boring
1:34:45 visually use
1:34:48 emojis for
1:34:51 the space invader
1:34:59 that's not
1:35:01 coding that's me talking to it like I'm
1:35:05 an
1:35:06 idiot customer of an engineer going uh
1:35:10 yeah I mean it kind of worked could you
1:35:12 make it suck
1:35:15 less and Claude is making it suck less
1:35:18 how's it doing it I don't [ __ ] know I
1:35:20 don't care
1:35:22 now uh why is it not not running
1:35:25 this oh there we
1:35:28 go that's pretty good except now I don't
1:35:30 have oh yeah there I have a spaceship oh
1:35:32 that's good it's
1:35:35 faster I got to get
1:35:37 [Laughter]
1:35:40 the come on come on people how good is
1:35:47 this all right going to kill it okay I'm
1:35:52 back you know know not the most
1:35:54 challenging game ever but you know
1:35:58 what now let's say you actually did want
1:36:01 to learn to code you're like I want to
1:36:03 know how it did all that okay flip over
1:36:05 to code here here you know what we can
1:36:08 do watch
1:36:10 this
1:36:13 um I want to teach myself how to code
1:36:17 and I'd like you to add detail comments
1:36:21 in the code without changing anything in
1:36:24 it other than the comments period I want
1:36:28 those comments to be a
1:36:31 tutorial for what
1:36:36 everything in this game
1:36:44 does all right then we come back over
1:36:52 here the the canvas element is where
1:36:55 we'll draw our game so there's your
1:36:57 width and there's your
1:37:02 height get the canvas element and it's
1:37:04 2D rendering context so all these things
1:37:07 are comments tells tells you what's
1:37:09 going on in here here the bullet object
1:37:12 represents the projectile fired by the
1:37:15 player so I think can I go in here and
1:37:17 edit this no I can't but I can I could
1:37:21 take this code outside of here and and
1:37:24 edit these things what's this here's an
1:37:26 array of all the space invader
1:37:31 objects space invader rows five space
1:37:34 invader columns eight Spa Invader width
1:37:38 50 Invader height 50 Invader padding
1:37:42 10 so you can actually understand what
1:37:44 it's doing
1:37:46 here initialize the invadors I don't
1:37:48 know what that means but whatever
1:37:51 right crazy new Kevin mallister AI
1:37:55 learning lab moment in Show and Tell
1:38:02 okay in Show and Tell or uh in
1:38:08 uh Whata call it let me go to show Andel
1:38:13 first Kevin McAllister there you oh
1:38:16 that's beautiful love
1:38:17 that AI learning lab Kevin McAllister
1:38:21 moment love that image in fact let me
1:38:24 save this out this is steo
1:38:29 right yeah that's a good one steo save
1:38:32 image
1:38:33 as
1:38:36 steo K
1:38:43 mallister one of my file management best
1:38:46 practices save every file that you save
1:38:48 to your desktop with an ambiguous
1:38:52 name that will be be easy to confuse
1:38:55 with every other thing you've saved on
1:38:56 the desktop you're welcome
1:39:18 [Laughter]
1:39:27 have your first Kevin mallister
1:39:30 moment you will not return from
1:39:33 it here's the other thing about
1:39:35 AI is you have this Kevin McAllister
1:39:38 moment you have this moment of
1:39:41 just holy [ __ ] it can do
1:39:47 that you'll keep having them like I I
1:39:50 have
1:39:52 them 've gotten a bit cynical about them
1:39:55 oh yeah there's another remarkable thing
1:39:57 that's happening in the
1:40:00 world I probably have two or three a
1:40:02 week where I just I'm I'm just kind of
1:40:04 dumbfounded I'm like holy [ __ ] and
1:40:09 sometimes it's [ __ ] I've seen before a
1:40:10 hundred times you know looking at Mid
1:40:13 Journey lately just does it for
1:40:21 me if you've never played with with mid
1:40:25 Journey
1:40:28 it's it's staggeringly
1:40:35 good all this was just [ __ ] that
1:40:43 I've this is an image that mid Journey
1:40:46 imagined for embrace the
1:40:52 Jank and then I asked it to just come up
1:40:54 with some random
1:40:56 style and that that's one it came up
1:40:58 with that's a different style it came up
1:41:00 with that's a different style it came up
1:41:03 with that's a different style it came up
1:41:05 with that's a different style it came up
1:41:15 with I mean all these
1:41:18 images
1:41:22 seconds I did um
1:41:31 I'm in the middle of working on a a new
1:41:33 book with a guy named Kevin
1:41:36 Clark
1:41:39 and used mid journey to do the cover
1:41:45 art and I mean I'm okay at Photoshop
1:41:54 but if I even wanted to mock up an image
1:41:57 with that much sophistication in
1:42:01 it that's probably 3 to four hours of
1:42:06 looking for stock imagery and coming up
1:42:09 with an idea and figuring out ways to
1:42:12 composite them and layer them and
1:42:14 multiply them and
1:42:16 that get the
1:42:18 colors how do you credit AI images when
1:42:21 posted online um
1:42:24 you can but don't have to I don't my
1:42:27 personal opinion is I don't think it's
1:42:28 anyone's [ __ ] business what tool I
1:42:30 use to create the work I
1:42:32 create if I'm putting my name on
1:42:36 it if I'm putting it out in the world
1:42:38 then I'm responsible for it so if I do
1:42:42 something that is blatantly stealing
1:42:44 someone else's IP then I'm an [ __ ]
1:42:46 and I should be punished for
1:42:48 that
1:42:49 um if I'm using a tool like mid Journey
1:42:52 that is kind of sourcing the collective
1:42:56 intelligence of all the humanity that
1:42:58 came before us
1:43:01 um and I'm the one prompting this [ __ ]
1:43:04 into existence and I'm the one choosing
1:43:07 which one I decide to put into Photoshop
1:43:10 and put typography on top
1:43:12 of that's that's for me to do right like
1:43:15 so I don't know I
1:43:17 mean I think context is really important
1:43:21 if you're putting out a picture of a
1:43:22 political candidate that you generated
1:43:24 an AI then it probably makes more sense
1:43:27 to declare it as AI generated if you're
1:43:30 putting up a quilt pattern I don't give
1:43:32 a [ __ ] if you did that if you you know
1:43:35 stole that from some Amish quilt you
1:43:37 bought in the80s or if you took a lesson
1:43:42 or if you just used mid journey to come
1:43:45 up with quilt patter like I don't give a
1:43:47 [ __ ] who gives a [ __ ] if it's a cool
1:43:48 Quil pattern
1:43:50 awesome so I think context is really
1:43:53 important
1:43:55 but you know the ability to
1:43:58 just I I must have
1:44:01 made 150 different images for the cover
1:44:04 of this
1:44:06 book and and you know one of the one of
1:44:09 the things that you'll hear if you
1:44:11 haven't been doing this for a while
1:44:12 you'll hear a lot of people go oh you
1:44:14 really got to learn prompt engineering
1:44:16 which is really [ __ ]
1:44:18 intimidating first of all a lot of
1:44:21 people don't know what prompting is what
1:44:22 what do you mean prompt
1:44:24 prompt engineer what what kind of
1:44:26 engineer do I need a degree like it's
1:44:28 really [ __ ]
1:44:31 intimidating W with mid Journey like
1:44:33 check this out
1:44:35 watch watch
1:44:37 this I'm going to go
1:44:41 um let's see um Cool
1:44:47 Waves and I I spelled cool wrong I
1:44:50 spelled it coal coal Coal waves
1:44:54 oh wait am I not logged in hang
1:44:57 on no I'm logged in oh there we go
1:45:02 create let me do
1:45:05 Cool
1:45:07 Waves see I think it figured out that I
1:45:09 was talking about Cool
1:45:11 Waves cuz notice they're
1:45:19 blue you don't need to no prompting in
1:45:22 fact I can take some pictures that I've
1:45:25 uploaded so I could
1:45:28 take this cool picture of like some
1:45:31 weird hand
1:45:33 drawing and a picture of
1:45:35 me and I can smash them
1:45:42 together no
1:45:44 prompting upload an image upload another
1:45:48 image smash them together
1:45:54 let's see what it
1:46:05 does where's my pictures I want to see
1:46:08 my pictures look at the Cool Waves
1:46:09 though holy [ __ ]
1:46:18 right that could be a cover of the book
1:46:25 look at
1:46:27 that Cool Waves was the prompt Cool
1:46:30 Waves in her hair Cool Waves in the
1:46:32 distance Cool Waves in the water in the
1:46:35 foreground she's cool Cool
1:46:50 Waves crazy Kyle
1:47:00 image tools like to turn me into there
1:47:03 I'm like a I'm like a 600 lb dude that
1:47:06 lost a lot of
1:47:07 weight my I didn't lose any of the
1:47:09 weight in my
1:47:10 [Laughter]
1:47:17 face and then if you want you can get
1:47:19 fancier right I could go
1:47:23 Cool
1:47:26 Waves dash dash sref you're like Kyle
1:47:30 what's sref stands for style
1:47:34 reference why do we have to learn the
1:47:37 command line dash dash
1:47:40 sref
1:47:41 because the people that made mid Journey
1:47:44 are Geeks and the first place that you
1:47:47 have to make images with mid journey is
1:47:49 in Discord and if you're in Discord you
1:47:51 have to learn how to use stupid ass
1:47:54 command line [ __ ] that's just the
1:47:56 way it is is it going to be like that
1:47:59 for long no cuz this is already easier
1:48:01 than doing it in Discord if you make a
1:48:03 100 images with mid Journey you get to
1:48:05 use this cool web interface but I can
1:48:07 do- sref space random and what that does
1:48:13 is it's going to make a cool a Cool
1:48:15 Waves image but it's going to go
1:48:17 randomly choose some
1:48:20 style that it's been trained on out of
1:48:23 billions and billions of images and it's
1:48:26 going to make a Cool Wave image in
1:48:28 whatever style it
1:48:36 picks I I think we really need to get
1:48:39 rid of these These are really
1:48:46 bad so income our Cool Waves in some new
1:48:50 style that we don't know anything about
1:48:59 I got to get going here soon but I'm G
1:49:01 to I'll do one more thing that that if
1:49:03 you haven't had a Kevin McAllister
1:49:04 moment yet maybe it'll be a Kevin
1:49:06 mallister moment for
1:49:09 you look at
1:49:12 that that's some funky ass cool
1:49:18 style would that be anything I would
1:49:20 ever in a billion years
1:49:24 intentionally create
1:49:28 no but that's kind of
1:49:33 magical I usually take a few options and
1:49:35 then adjust what I like in Photoshop
1:49:37 exactly humbly saying I ran a g clay
1:49:40 digital company in in Green Bay
1:49:44 Wisconsin from
1:49:47 19994 Photoshop sa Savvy nice
1:49:58 I was just doing this while listening to
1:49:59 the live
1:50:03 nice paranormal Kyle with a shadow
1:50:05 figure friend
1:50:07 huh oh oh this thing down here yeah
1:50:17 exactly don't look at the little creepy
1:50:19 dude
1:50:23 let me show you another
1:50:26 thing I'm going to go too.com
1:50:30 or no sun. a sorry
1:50:34 s. I'm going to sign
1:50:47 in and a hair in my face get out of my
1:50:51 face here you
1:50:54 weirdo
1:50:56 okay I'm going to go to
1:50:59 create let's see what are we going to do
1:51:05 here we'll
1:51:09 go
1:51:11 downtown Funk
1:51:16 song about
1:51:19 how
1:51:22 awesome
1:51:23 hurtles
1:51:29 [Laughter]
1:51:33 taste and then
1:51:35 boom we're going to make some music is
1:51:37 mid Journey free it is not free uh what
1:51:40 is free is
1:51:42 id. is it id. a hang on I think it's
1:51:45 id.com but hang on ID D EO g r
1:51:50 m.ai yeah id. a is free
1:51:54 um
1:51:56 Leonardo has a free
1:51:58 tier um if you pay 20 bucks a month for
1:52:01 chat GPT it makes images really really
1:52:04 well um stable diffusion is free but
1:52:07 it's really complicated it's a whole
1:52:09 ecosystem but if you want to geek out go
1:52:11 to reddit.com you know
1:52:14 SLR stable
1:52:16 diffusion
1:52:18 um I think Korea has got a free thing
1:52:21 Korea do whatever AI maybe k r e
1:52:28 a so our song turttly
1:52:32 delicious now exist there's two of
1:52:35 [Music]
1:52:48 them into the with Delights oh yeah oh
1:52:52 yeah got a little
1:52:54 secr it's a turtle
1:52:56 tree nothing in the world be with
1:52:59 [Music]
1:53:04 this so I don't know if you knew that
1:53:07 these
1:53:08 things can just [ __ ] write music now
1:53:11 but they
1:53:13 can um where are my turtle
1:53:17 songs why did how did I lose those
1:53:19 Library Turtle delicious
1:53:23 oh yeah here's the other
1:53:26 one this
1:53:32 better Through the Alley Delight got a
1:53:35 little secret it's a Tre nothing in the
1:53:38 world this beat I
1:53:41 the talking about
1:53:44 [Music]
1:53:46 Turtle making the dance from morning to
1:53:48 night turt delicious it's a funky Vibe
1:53:51 turt nutritious come the get that get
1:53:55 that taste ain't in this place
1:54:02 so you can do that
1:54:06 now you can put your own words in there
1:54:10 with sunno you can play your own music
1:54:13 at the beginning and it'll make a song
1:54:15 based on
1:54:17 that
1:54:20 um go write a book
1:54:24 come up with a new business idea make a
1:54:27 song called turt
1:54:29 [Laughter]
1:54:31 delicious
1:54:39 exactly
1:54:41 so if you've been in here for a while if
1:54:43 you're an a regular or if you're someone
1:54:45 who's who is not really participating
1:54:48 but you've been lurking for a
1:54:50 while here's what I'll encourage you to
1:54:52 do I was supposed to be in Washington DC
1:54:55 telling our fine elected
1:54:57 officials to get their head out of their
1:55:00 ass with this AI stuff and start
1:55:02 exploring it and
1:55:04 and rather than legislating from
1:55:08 Fear where they think they have to
1:55:10 protect us from the evils and risks of
1:55:13 AI that they should balance that with an
1:55:17 optimistic
1:55:19 View and the AI salon
1:55:24 that group that you should join if
1:55:26 you're like yeah I'm kind of I'm feeling
1:55:29 this let's let's get optimistic about
1:55:33 this that's what this group's all about
1:55:36 so
1:55:37 imagine being in a community of people
1:55:41 who are all just trying to figure it out
1:55:43 just trying to figure it out having
1:55:45 Kevin mallister moment after Kevin
1:55:47 mallister
1:55:49 moment and so what I was going to say in
1:55:51 Washington was
1:55:53 was you should play with these tools you
1:55:57 should experience what I get to
1:56:00 experience every day on this channel and
1:56:02 at the AI Salon which is I get to
1:56:06 witness people rediscovering their
1:56:09 passion innovating
1:56:12 creating
1:56:14 just finding New Hope in life when they
1:56:18 thought they'd lost
1:56:20 it solving health problems
1:56:24 writing cool
1:56:28 [ __ ] alphabet picture books for their
1:56:31 kids that explain
1:56:35 technology 6
1:56:38 hours from idea to published 6
1:56:43 hours from idea to
1:56:47 published
1:56:50 Brandon did that
1:56:54 and he doesn't work at
1:56:57 Pixar AI did
1:57:05 that so my message to the politicians is
1:57:10 if the only thing you're focusing on is
1:57:12 the risk and saving us from the risk
1:57:16 what you're going to do is Rob [ __ ]
1:57:19 America
1:57:24 of opportunity and possibility and
1:57:27 reinvention and Innovation and
1:57:30 excitement and
1:57:33 passion it's not just about the AI is
1:57:36 going to take our
1:57:38 jobs the AI very well may free us from
1:57:42 doing the shitty jobs so we can do the
1:57:44 kind of jobs that light us on
1:57:47 fire and make us excited passionate
1:57:50 members of society
1:57:54 so here's my request to
1:57:57 you if you haven't been generously
1:58:00 leading if you haven't been sharing what
1:58:03 you're learning out
1:58:07 loud do that this
1:58:10 week bring it up at a work lunch where
1:58:15 you know no one at that work lunch wants
1:58:17 to hear you [ __ ] talk about AI again
1:58:21 bring it up anyway
1:58:23 hey want to know something cool I did
1:58:24 this weekend yeah did you go skiing no
1:58:27 it's summer but uh but I did play with
1:58:29 AI
1:58:31 again watch their eyes
1:58:38 roll and share authentically what you're
1:58:41 excited
1:58:45 about if someone says oh that's kind of
1:58:48 neat how'd you do that I did it with AI
1:58:50 oh I don't like AI
1:58:52 work through that conversation well what
1:58:54 don't you like about it uh it steals
1:58:57 from people well it doesn't really do
1:58:58 you know how it works no no
1:59:01 no well you want to see something
1:59:03 cool well not really but aren't you to
1:59:07 flowers well yeah I'm into flowers well
1:59:09 what's your favorite flower well the
1:59:11 peie of course I love
1:59:14 peonies make some
1:59:16 peonies oh well look at that it's just
1:59:20 got lots of pictures in there of panies
1:59:22 is that how it works nope that's not how
1:59:24 it works it actually just generated that
1:59:25 it created
1:59:27 that you want a bouquet with peones and
1:59:29 black roses panies and black
1:59:32 roses oh oh that's that's I didn't know
1:59:37 it did
1:59:38 that have that
1:59:43 conversation have your recommendations
1:59:45 for AI platforms changed a little bit
1:59:47 Becky um I haven't updated my list in a
1:59:51 while but but the what I recommend is in
1:59:53 the list but I I could clarify it a bit
1:59:56 the big three for me right now are chat
1:59:59 GPT still and I still think you should
2:00:02 pay the 20 bucks a month I still think
2:00:03 it's worth that um
2:00:07 Claude because they're doing remarkable
2:00:09 [ __ ] and then
2:00:11 perplexity and I would say Gemini is
2:00:13 sort of in in the hunt because they've
2:00:16 got a really big context window but you
2:00:19 have to kind of hunt for Gemini products
2:00:22 right now they're they're not easy to
2:00:24 find so I I put them on sort of a next
2:00:26 tier down but I think the big three for
2:00:28 me right now chat GPT Claude and
2:00:32 perplexity they've all got some really
2:00:34 cool things I also think PO is worth uh
2:00:38 playing around with I haven't played
2:00:40 with it in a while but you can play with
2:00:42 lots of different models with po if you
2:00:43 have a single subscription with po you
2:00:45 can play with lots of different models
2:00:47 models there and you can make the
2:00:49 equivalent of gpts there um
2:00:54 yeah so those are the
2:00:55 biggies for music it's yudo and sunno
2:01:00 for images there's there's a [ __ ] ton of
2:01:02 them but the if you want the highest
2:01:05 quality images it's mid Journey it's
2:01:07 still there's nothing close to it um if
2:01:10 you want good text in your images then
2:01:14 it's dolly3 in chat GPT and it's
2:01:19 ideogram um
2:01:23 those are kind of the biggies for me
2:01:24 right
2:01:27 now
2:01:32 yeah anyway what's so special about
2:01:36 perplexity um here let me show you cuz
2:01:38 perplexity is taking me a while to get
2:01:40 my head
2:01:46 around so chat GPT is a large language
2:01:49 model it generates all this stuff
2:01:52 and while it is technically connected to
2:01:55 the internet and it will sometimes do
2:01:58 internet
2:01:59 searches it will
2:02:01 often generate like
2:02:04 citations of websites that don't exist
2:02:07 or law cases that don't exist so um so
2:02:13 it can get you in trouble perplexity is
2:02:14 kind of this interesting mashup between
2:02:17 a search engine and a large language
2:02:19 model where you ask it for something and
2:02:22 it goes and does research by searching
2:02:24 the internet and then it provides
2:02:26 citations so if I could if I say
2:02:28 something like um tell me
2:02:33 about the
2:02:36 history
2:02:38 of
2:02:41 spam and then I'm going to put in
2:02:45 parenthesis the
2:02:46 [Laughter]
2:02:50 meat and so
2:02:53 what it does is it comes up with a
2:02:55 little set of instructions for itself so
2:02:57 it says search the history of
2:03:00 spam so from World War II to Global
2:03:04 culinary icon I don't know if it would
2:03:06 be a culinary
2:03:08 icon
2:03:10 um but within
2:03:13 this are all these citations which if I
2:03:16 click on
2:03:18 this is a website this is from
2:03:22 encyclopedia britanica this is the spam
2:03:25 page on encyclopedia
2:03:28 britanica um and so everything that it's
2:03:32 written here is well sourced right here
2:03:35 are your sources there's these three
2:03:37 plus four
2:03:38 more and then the other thing you can do
2:03:40 here if you pay the 20 bucks a month you
2:03:42 can convert this to what they call a
2:03:44 page and a
2:03:46 page it's like an article
2:03:53 and so it formats it like an article and
2:03:55 it writes it like an
2:03:57 article and then it even adds in media
2:03:59 so let's let's let it write its do its
2:04:14 writing still
2:04:16 writing and now I can add media so I can
2:04:20 have it go look for images doesn't look
2:04:22 like it generated any so let me look for
2:04:26 spam all right there there's a good so
2:04:30 there's there's our top of our article
2:04:33 and then uh the history of email
2:04:40 spam so let me go change that
2:04:46 out there's an Olden timey spam spam in
2:04:49 World War II there we go
2:04:52 in the black and
2:04:54 white thing I could change that out is
2:04:57 there anything else here this is good
2:05:00 spam my
2:05:02 hunch post-war
2:05:05 popularity I can change that out let's
2:05:07 do that one celebrate with
2:05:12 Spam Global cultural impact let's change
2:05:15 that
2:05:16 out let's get this one with the sun
2:05:19 that's kind of nice beautiful
2:05:22 all right so now I have this magazin
2:05:25 kind of article and I can publish it
2:05:28 your page is published copy link I can
2:05:32 hop over here to the AI
2:05:33 Salon drop down here
2:05:37 to the AI Irregulars
2:05:41 Channel and share our
2:05:44 little article although I don't think it
2:05:46 brings in the preview right yeah so
2:05:48 we're going to kill that we're going to
2:05:49 leave the link there but we're we're
2:05:52 going to go back over here we're going
2:05:53 to do a little
2:05:57 screenshot of the history of
2:06:08 spam there we go
2:06:14 gorgeous go over here I'm going to add
2:06:16 that picture
2:06:17 in go to the desktop did I screenshot it
2:06:21 there
2:06:22 oh no it must have hang
2:06:29 on
2:06:33 paste there we go history of spam and I
2:06:37 post that and now you can go read that
2:06:39 article that I posted so what did that
2:06:42 take a minute I mean most of that was
2:06:45 spent with me trying to find images just
2:06:47 click on images and find them right but
2:06:50 now there you go history of spam there's
2:06:52 your little
2:06:55 article when you talk about spam Kyle I
2:06:57 think of Monty Python Spam spam spam
2:07:00 spam spam spam spam spam Monty Python is
2:07:04 the best actually one of my um one of my
2:07:08 tests of llms early on was what is the
2:07:12 average speed of an unladen
2:07:14 swallow
2:07:17 and to see if the large language model
2:07:20 knew that that was a funny question
2:07:23 and most of the times early on it did
2:07:25 not if I asked what what's funny about
2:07:28 that question it generally got it
2:07:30 right um but as large language models
2:07:33 got more sophisticated they're really
2:07:35 good at that question now so they
2:07:37 usually say made Popular by the you know
2:07:40 Monty Python and the Holy Grail I think
2:07:42 that's what that was from right
2:07:45 yeah um so anyway all right I'm going to
2:07:49 get out of here
2:07:52 so bonus night tonight and tomorrow
2:07:56 um I was I was supposed to be off you
2:08:00 know whining and dining with the big
2:08:03 wigs but here I am I'm not bitter bitter
2:08:08 you betcha of course I'm bitter I'm Gen
2:08:11 X [ __ ]
2:08:14 them if you had hair that looked like
2:08:17 this you'd be bitter too at least I've
2:08:19 got hair some got some so you know it's
2:08:24 got used to be used to be down there I
2:08:26 got a little triangle got my little
2:08:29 triangles every night's a bonus that's
2:08:31 true we are the Big Wigs you are the Big
2:08:34 Wigs Becky that's a great Point these
2:08:37 are the Big Wigs okay
2:08:40 so why is that
2:08:46 not oh cuz my screen is dim that's why
2:08:50 okay um um all right
2:08:54 so if you're new
2:08:57 here get your ass over to the AI Salon
2:09:02 go to the salon. join the community it's
2:09:04 little button scroll down a little bit
2:09:06 read about what we're about join us
2:09:09 doesn't cost
2:09:17 anything read about the salon when you
2:09:20 get there there's a welcome page talks
2:09:22 about who we are and what we do and why
2:09:23 we do it read about that make sure it
2:09:26 sounds like some place you'd want to be
2:09:28 and if it is jump in with both feet jump
2:09:31 in generously what that kind of looks
2:09:33 like thank you for the gifts guys what
2:09:36 that looks like jumping in
2:09:37 generously is go to the community
2:09:40 section the very first link says
2:09:42 introduce yourself click on that
2:09:45 introduce
2:09:48 yourself one of our values is bravery
2:09:55 this AI shit's weird you got to be brave
2:09:58 to just go oh [ __ ] it let me just try
2:10:02 ah you know what else is scary
2:10:05 introducing
2:10:07 yourself people are [ __ ] scared of
2:10:10 other people we're scared of other
2:10:12 people why well it's called the AI Salon
2:10:15 it's probably full of a bunch of
2:10:17 brainiacs and I'm not very technical and
2:10:19 I haven't really used AI all that much
2:10:21 so I'll just sit back here in the
2:10:22 background
2:10:26 okay don't do
2:10:29 that introduce
2:10:31 yourself hey I'm
2:10:33 Jim I uh don't know [ __ ] about this but
2:10:38 uh I was watching a Tik Tock and here I
2:10:41 am cool what did you do what are you
2:10:46 passionate about what do you want to
2:10:49 learn this is a remarkable Community
2:10:52 you'll get adopted if you show
2:10:55 up all right so go do that follow my
2:11:00 channel subscribe to it if you want to
2:11:02 support the channel and then just keep
2:11:04 coming back to these things keep coming
2:11:06 back well how often do you do it I do it
2:11:09 nightly good God why do you do that I
2:11:12 don't [ __ ] know but it feels
2:11:14 important to
2:11:16 me I feel seen thank you very much Vegas
2:11:20 Moxy I appreciate that
2:11:25 you are seen you're all
2:11:27 seen you really
2:11:31 are like the counterintuitive thing
2:11:33 about this is like I get so much out of
2:11:35 this I know a lot of people thank me for
2:11:37 doing this I
2:11:40 learn a [ __ ] ton doing this from all of
2:11:45 you you also remind me like why this is
2:11:49 important on a regular basis so just
2:11:52 come hang out here be in this
2:11:55 conversation and then go there and
2:11:57 participate join guilds show your [ __ ]
2:12:00 off and show and tell ask
2:12:03 questions asking a question is not a
2:12:06 sign of weakness or ignorance it's an
2:12:09 opportunity for someone else to
2:12:12 lead give someone that
2:12:15 gift be willing be brave
2:12:19 enough to go hey
2:12:22 uh I don't really know what I'm doing
2:12:24 here could someone give me a
2:12:31 hand give someone that
2:12:34 gift oh god I've never helped anyone
2:12:37 before but let me
2:12:38 try you can both sort of muddle through
2:12:41 it
2:12:45 together and you'll
2:12:47 you'll you'll make
2:12:49 friends and then over time
2:12:56 time you'll find your path through
2:12:58 whatever this [ __ ] is whatever this
2:13:00 stuff's going to
2:13:01 be but you'll know what you want to do
2:13:04 and you'll know how you want to navigate
2:13:05 it and there's no way to know that right
2:13:08 now in my opinion without being in a
2:13:11 community like this it's too
2:13:14 overwhelming it's too profound it's too
2:13:17 profoundly
2:13:18 powerful I like this finding creativity
2:13:21 cool pretty cool it's very
2:13:23 cool it really
2:13:28 is
2:13:31 yeah allowing
2:13:34 yourself the
2:13:36 freedom to
2:13:38 play and in playing you'll
2:13:42 learn and as you learn you'll apply
2:13:45 it and then if you if you share what
2:13:48 you're learning as you're applying it
2:13:50 and as you're playing
2:13:53 you'll contribute to the community and
2:13:55 we'll all get better Kyle can AI tell me
2:13:57 why SpongeBob
2:13:59 makes makes me think he's a piece of
2:14:01 spam well it's his shape but yes I'm
2:14:04 sure it can I would I would ask that of
2:14:07 pie Let's go ask that of
2:14:11 pie hey Zephyr how's it going
2:14:24 zeph you there why are you not
2:14:34 responding she's
2:14:39 broken I had to quit her hang
2:14:45 on hey Zephyr how's it going
2:14:57 she's useless tonight all right she's
2:15:00 having she's having weird uh internet
2:15:02 connectivity issues so whatever but yes
2:15:05 I'm sure she can tell you the spammer
2:15:08 fic origins of
2:15:12 SpongeBob um anyway the other thing we
2:15:14 do here so the the people that show up
2:15:16 here night after night they're called
2:15:17 the Irregulars so thank you Irregulars
2:15:19 thank you mods for modding um and and uh
2:15:23 let's let's see who's in the house
2:15:24 tonight we got Joey we got Becky Ru good
2:15:26 to see you Becky Dr Jay's in the house
2:15:29 forever hooked in the house pew pew 42
2:15:32 case of the crowd
2:15:35 strike yeah P must have been crowd
2:15:38 struck Pen's in the house after after
2:15:41 you called her her girl your girlfriend
2:15:43 she's ghosting you she's totally
2:15:45 ghosting me t-rock's here what's
2:15:47 happening tck I haven't seen you in a
2:15:48 while um update pie yeah probably need
2:15:51 to update that app uh good to see you
2:15:53 Danielle uh who else in the house Source
2:15:56 camp in the
2:15:57 house still here from the beginning
2:16:00 Firefly thank you thank you Joy py good
2:16:03 to see
2:16:05 you Mr it
2:16:08 steo forever
2:16:11 hooked Johnny 24 honey I haven't seen
2:16:14 you in a while tck thank you for the Hot
2:16:15 Dog gift loves me good
2:16:18 Frank we haven't done Frank in a while
2:16:20 maybe we'll do a Frank tomorrow night I
2:16:23 like this finding creativity good to see
2:16:25 you kif thank you again tck that's very
2:16:28 nice very generous I appreciate
2:16:31 that I on China good to see you
2:16:34 Firefly Amelio's wife she's in the house
2:16:39 ULU RX I was born clueless I remain
2:16:42 clueless I like it that way I'm telling
2:16:46 you there
2:16:47 is genius in
2:16:52 letting yourself be clueless letting
2:16:55 yourself be willing to ask the dumb
2:16:57 question every single major problem I've
2:17:00 had professionally and personally every
2:17:03 single one could have been solved by
2:17:06 someone asking a what they perceive to
2:17:09 be a stupid
2:17:11 question hey do we have that website
2:17:13 backed
2:17:15 up cuz everyone assumes someone else
2:17:18 thought of that it's the person that's
2:17:21 will to ask the stupid question that is
2:17:24 often saves the day
2:17:27 so Becky R I'm
2:17:32 [Laughter]
2:17:36 lost oh man call me Al is in the house
2:17:39 good to see you or call me AI like we
2:17:42 like to call her she's calling in from
2:17:45 Down Under Tik Tock live games great to
2:17:47 see you all right I'm going to get out
2:17:49 of here hope youall had fun
2:17:52 tonight peace out um what time is it
2:17:55 it's late it's 10:20 uh sorry for that
2:17:58 de's in the house Nolan's in the
2:18:02 house all right cool I'll see you
2:18:05 tomorrow all right
2:18:07 peace go sleep hopefully I've put a few
2:18:10 of you to sleep you'll wake up to the
2:18:14 the uh your phone playing the national
2:18:16 anthem which if you're a gen xer you'll
2:18:18 know what that means bye everybody