11/26/2024 - AI's Creative Potential: From Janky Tools to Musical Masterpieces

Description

Chapters

Transcript

0:01 [Music]
0:18 [Music]
0:33 [Music]
0:35 hi champion
0:37 [Music]
0:52 [Music]
1:03 every time I see you now get that
1:06 looking
1:08 man every see
1:12 your I hear that
1:15 smile early Misty morning L I heard the
1:20 engine
1:22 turning the old F outside
1:25 [Music]
1:31 you
1:33 leave again
1:36 today you convince
1:39 me again
1:42 today you're leaving this H Town looking
1:46 for someone else's Golden
1:52 Ring should say so long suzan
1:59 [Music]
2:01 hush now don't you
2:03 [Music]
2:06 cry so
2:08 [Music]
2:12 long Don't You Cry for
2:16 [Music]
2:23 Me sharing cigarettes and keeping warm
2:27 out on the road
2:30 chasing down a lifestyle out on Highway
2:35 24 New York state was rolling Breeze and
2:39 the Sunshine with a blue sky
2:42 falling J September
2:48 [Music]
2:49 creep you
2:53 leave again
2:56 today you convince
2:59 me okay
3:02 today we leaving this hotel looking for
3:06 someone else's C and
3:09 [Music]
3:12 R so long
3:19 suan don't you cry for
3:21 [Music]
3:25 me so long
3:28 [Music]
3:32 Don't You Cry for
3:34 [Music]
3:50 Me solid solid singing champ welcome
3:54 everybody happy Tuesday night p on Tik
3:57 Tok on LinkedIn on YouTube on x we're
4:00 everywhere we're all over the Internet
4:02 we're going live it's Kyle Shannon this
4:04 AI learning lab what are we learning
4:06 tonight dogs can sing better than
4:10 people that's not really funny all right
4:13 thanks um you know Talent is a funny
4:19 [Music]
4:22 thing oh good evening okay so if you're
4:25 on the Tik Tock one of the things we'll
4:28 do is is we'll we'll demo
4:30 uh how to do things and like for example
4:34 one of the things we did last night was
4:36 we demoed how not to do
4:40 things I failed miserably I spent I
4:43 don't know what 90 minutes straight
4:46 trying to do something that was an
4:48 abject failure
4:50 why because I'm unqualified to be here
4:58 [Music]
4:59 [Laughter]
5:03 oh wow Joy pie is a new Grandma Joy
5:06 congratulations that's exciting how
5:09 awesome how awesome is that well that's
5:12 a that's a Thanksgiving week uh bundle
5:16 of appreciation and joy
5:19 and congratulations that's really
5:22 amazing
5:23 [Music]
5:36 if you're on link LinkedIn or X um if
5:40 you're on if you're on YouTube you
5:42 probably because YouTube here we're
5:44 learning some things as we're doing this
5:47 um if you're on Tik Tok and you want to
5:48 be able to see better go to if you go to
5:50 my profile and click on YouTube that'll
5:53 take you to the YouTube stream it's also
5:55 learning la- aai on YouTube but if
5:58 you're here on LinkedIn if you just
6:00 stumbled upon this and you're like what
6:02 the hell is this why is he playing a
6:04 guitar why is this on this late at night
6:06 it's just I don't normally stream on
6:08 LinkedIn but I figure it doesn't really
6:11 matter no one's watching anyway but
6:14 hello uh same on X
6:17 um this is uh this is the AI learning
6:20 lab stream and it's streaming to my my
6:23 personal X account cuz I don't think I
6:25 have an AI learning lab account on
6:28 X it doesn't really matter but anyway
6:32 welcome so if you have questions about
6:35 AI or you want to talk about a piece of
6:38 technology or you want to see something
6:41 see me try to demo
6:43 [Laughter]
6:46 something I came here last night I
6:49 thought he was going to teach me
6:50 something it's called the AI learning
6:51 lab and you would assume if it's a lab
6:53 there would be some sort of Professor
6:55 there or some sort of scientist some
6:57 sort of technician and it he it just it
7:01 I'll tell you that like it kind of had
7:03 the feeling of like as if there was like
7:06 a it was like kind of like an old man um
7:09 trying to figure out how to use a
7:11 computer is was the was the the so I
7:14 don't really know what I learned so if
7:16 you want to learn that tonight stick
7:18 around we'll do more of
7:22 that want to talk about how the Sora
7:24 leak on hugging face yeah so
7:32 it does appear that some of the uh
7:36 people that had access to Sora Early
7:38 Access to
7:40 Sora wrote sort of a ransom note kind of
7:44 kind of email that said hey we had
7:47 access we got free access to this thing
7:50 in exchange for finding bugs and testing
7:53 [ __ ] and they're just using Us for free
7:55 bug
7:57 testing and then they and then they made
8:00 it uh freely available uh to everyone
8:03 for about 3 hours and then uh open AI
8:07 found out about it and shut that [ __ ]
8:08 down but now no one has access to it
8:11 that that had access to it
8:15 um
8:17 I my feeling on Sora at this point is
8:21 there
8:22 are eight or nine or 10 other viable
8:27 options than Sora like Sora was a huge
8:30 deal back in [ __ ] March and now
8:33 you've got like Hal Halo Mini Max is is
8:36 right in the same neighborhood with with
8:39 Sora so they should just [ __ ] launch
8:42 the thing I don't know what they're
8:43 waiting for at this
8:44 [Music]
8:49 point great to be here learning I
8:53 appreciate this community so much
8:55 authentic and playful thank you and
8:58 generous it is indeed spin B3 ULU RX
9:02 thank you for the gift Winston Robert
9:04 Rossy thank you for the gifts everybody
9:06 I do appreciate
9:09 that I do this out of the I would say
9:11 the goodness of my heart but it's not a
9:13 good thing I do this oh thank you Robert
9:15 Rossy it's very generous I do this
9:17 because no one cares about Sora anymore
9:19 I think we're all pissed at this point
9:21 yeah exactly it's like I I mean
9:24 again if all of the video models sucked
9:28 and we're like we need Sora because it's
9:30 the only one that's good that would be
9:32 one thing but like no there's there's
9:36 you know out of the 10 that are
9:37 available there's four or five that are
9:39 really quite good um I might play with
9:42 Luma Labs tonight they came out with a
9:45 really
9:47 um they at least took a swing at an
9:50 Innovative interface which I kind of
9:52 like I think it's it's really
9:55 smart um
9:57 [Music]
10:00 I also think it might be fun to go over
10:02 and play with Claude Claude um Claude
10:04 today released
10:07 Styles so that you can um you can make
10:09 it right like you or you can you know
10:11 make it informal or formal or scientific
10:14 or whatever it might be you can build
10:16 your own so that might be fun to play
10:17 with and then we're going to continue
10:20 down the path
10:22 of
10:24 um Avatar
10:26 [Music]
10:28 making and like I said said last night
10:30 was an abject failure but you know this
10:32 is this is how you [ __ ]
10:35 [Music]
10:47 learn our Town's chamber sponsored free
10:49 class on chat GPT plus canvas so I
10:52 attended it was just me and the
10:55 instructor there you go
11:02 well
11:06 Jeff that's where we are I I mean I I
11:10 mean for the Irregulars here for the for
11:12 the people that come to to this channel
11:14 nightly you know we've been doing this
11:16 for almost two years now we've I've been
11:18 doing these lives since April I think it
11:21 was April 30th of 20123 was our first
11:25 one so we're coming up on two years of
11:26 doing these things every night
11:30 and the group that hangs out here
11:32 regularly like I would argue with this
11:34 generative AI stuff are as educated and
11:37 as you know good at this and
11:40 understanding what all the different
11:42 options are and how to how to string
11:45 different tools together to get outputs
11:47 I mean this group is is really up to it
11:50 and we're a bit of an echo chamber here
11:53 where where you know it's it's pretty
11:56 [ __ ] clear to us how big a deal this
11:59 AI stuff is uh but out there in the real
12:03 world there's just not a lot of
12:06 awareness of it and and uh enthusiasm
12:11 for it and the whole reason this channel
12:15 exists is because that makes me hurt
12:18 inside um because the amount of
12:21 opportunity right now for all of us is
12:24 massive and the amount of uh people just
12:28 either aware of it or actively resisting
12:32 it is is really frustrating I think this
12:35 stuff is so important to us and so
12:37 unimportant to the rest of the world
12:39 well it's I mean it's the thing is it's
12:41 like any new
12:42 technology the people that are aware of
12:45 it have massive opportunity we've got
12:48 massive opportunity and I know it might
12:49 not feel like that right now because
12:51 it's just like the tools are just
12:53 they're still janky enough that they're
12:56 not quite there but holy [ __ ] we're
13:00 the we're we're we're in the middle of
13:02 Crossing this threshold you know
13:05 Saturday is November 30th so November
13:09 30th 2022 was the launch of chat GPT
13:12 that was the beginning of this channel
13:14 that was the beginning of the AI Salon
13:17 which if you don't know the AI
13:19 Salon there's the URL you should go
13:23 there and you should scroll down and
13:25 click on the button that says join our
13:27 community and it doesn't cost anything
13:30 and the community is remarkable there's
13:32 a nice little welcoming uh process there
13:35 there's a video that you can watch
13:37 welcoming you to the community see what
13:39 we're
13:42 [Music]
13:47 about
13:50 um so yeah go uh go check out the salon
13:54 the salon our first meeting of the of
13:57 the AI Salon was December 7th 2022 so
14:02 one week after chat GPT came
14:04 out so Saturday is
14:08 the two-year anniversary of chat GPT so
14:12 rumor has it that we will get some new
14:16 toys from Sam Alman and
14:19 [Music]
14:20 Company but I'm in a new place
14:23 philosophically with waiting for tools
14:26 from open AI I don't give a [ __ ]
14:35 because um at this point there is
14:40 so there is so much to explore that's
14:43 not open AI um and and while they're
14:46 still in the lead in my opinion there's
14:48 others that are really close and if they
14:50 give us something new and revolutionary
14:52 great but if they don't just even things
14:55 like you know what we're doing this week
14:57 with avatars exploring like how do you
14:59 build a avatars and build you know
15:02 representations of yourselves and how do
15:03 you do that in a in a way that's not
15:06 going to break the bank in a way that um
15:09 Works doesn't look like
15:12 [ __ ] um there's a lot to learn um even
15:15 even silly things well I don't know if
15:17 it's silly it's not silly to some things
15:19 like you know how do you use something
15:21 like puno to make songs that are good
15:23 and what if you want to make songs that
15:25 are good enough to say put in a musical
15:27 or you want to you know put them out on
15:29 the radio um Timberland is doing that
15:32 producer Grammy award-winning producer
15:34 is using sunno so what's the difference
15:37 between just making a song in sunno and
15:41 doing something to put out in the world
15:43 with your name on it I don't know what
15:46 is the difference like none of us know
15:48 that yet it's like so there's there's
15:52 just so much to explore so but that said
15:56 maybe we get a one tomorrow maybe we get
15:58 soono tomorrow or s
16:01 tomorrow was I saying sunno earlier I
16:03 might have been Sora is the video tool
16:06 from open
16:09 AI Tobias in the YouTube Welcome welcome
16:14 [Music]
16:26 welcome
16:28 oh Dam
16:30 Danielle she's pumped about AGI Alpha
16:33 fold Alpha proteo being open-
16:37 sourced go democratize big farm on cure
16:40 diseases yep I mean I mean like I again
16:44 I I just yes I agree and holy [ __ ] 2025
16:51 like the fact that you know a local Town
16:56 puts on a come learn about chat GPT
16:59 session and one guy shows up and the guy
17:02 that shows up is in irregular here that
17:05 probably knows as much as the instructor
17:08 knew or
17:09 [Laughter]
17:12 more it just it just says to me that so
17:15 many people are going to be blindsided
17:16 by this technology and that's terrifying
17:18 to me
17:19 because if you're involved with AI you
17:22 can feel this thing accelerating away
17:24 from us right like how do we keep up
17:26 with it I don't [ __ ] know I'm you
17:28 know at this Point I've given up on that
17:30 I'm just like well we'll start doing
17:32 smaller deeper dive projects that's the
17:35 thing I'm excited about now so all
17:41 right archetypal architect I don't
17:44 remember did we go crazy for
17:48 smartphones oh did we go this crazy for
17:50 smartphones or cell phones
17:53 um
17:55 well when the first when the first
17:58 iPhone came out I mean I mean people
17:59 have been talking about
18:01 smartphones and SMS gaming and all that
18:05 sort of [ __ ] for
18:07 years but the form factor was the phone
18:10 with buttons on it right it was shitty
18:14 and then they they would try to remember
18:16 there was a there was a Palm Pilot that
18:18 came out that that had Palm Pilot at the
18:21 top and a phone at the bottom and that
18:24 that was sort of there and the Newton
18:26 came out that tried to be this
18:30 and then when the when the um when the
18:34 iPhone came out it was just all of a
18:35 sudden like ah it it's a new form factor
18:38 right and then even when the new iPhone
18:41 came out the applications weren't
18:44 actually applications they were websites
18:46 right do you remember that and all the
18:47 programmers were like ah this is
18:49 [ __ ] these aren't real applications
18:51 it's just web it's just web
18:54 pages everyone was
18:56 bitching um but but
19:02 the the iPhone didn't
19:07 fundamentally [ __ ] up all your other
19:11 phones right you can still buy phones
19:14 with buttons on them right so it was an
19:16 addition too and it opened up a whole
19:18 new kind of market right the App Store
19:20 and all that sort of stuff it opened up
19:22 a whole new kind of marketing I would
19:25 say that the last time we had something
19:29 that was this
19:31 foundational um was the worldwide web
19:34 right so in the mid90s the internet had
19:36 been around for decades everyone talked
19:39 about the internet and how you could
19:40 share files and Librarians used it and
19:43 there was FTP servers and goer servers
19:46 and there were dial up you know you
19:47 could connect to chat boards and things
19:50 like that and then AOL sort of turned
19:53 that into CompuServe I think started and
19:55 then AOL came next and and Prodigy ort
19:59 of took that shitty internet and put
20:02 kind of a pretty skin on
20:04 it um but they were walled Gardens and
20:07 they were closed and they were
20:08 proprietary and they were weird and you
20:10 had to pay per minute and you had to
20:12 dial
20:13 up and then the worldwide web comes
20:16 along and anyone who could learn a
20:19 couple of HTML tags could put up a
20:22 website right and then anyone could use
20:27 it so you so you
20:30 you you all of a sudden computers went
20:34 from these disconnected things to these
20:36 things where people could start to
20:40 self-express with technology and and
20:43 share it with the
20:44 world and that was my big Epiphany with
20:47 the worldwide web was I launched a an
20:50 online magazine and three weeks later it
20:53 was written about in a Parisian
20:56 newspaper and it just all of a sudden I
20:58 realized holy [ __ ] the world's just
21:00 changed so the the difference between
21:03 the worldwide web and this AI stuff well
21:05 there's there's many of them but but the
21:08 worldwide
21:10 web the infrastructure wasn't there we
21:13 we were still using dialup modems when
21:15 when we were doing all our web
21:17 development like having a T1 line you
21:20 know where you had always on Internet
21:22 that was a big deal like I I think the
21:24 initial T1 lines were like 10 grand a
21:27 month yeah they were 10 grand a month
21:30 they were really
21:31 expensive um just to have always on
21:34 internet and then you know over five six
21:38 years the infrastructure gets built and
21:41 and everything takes off from there so I
21:44 would say that the worldwide web was the
21:46 the last time we had something this
21:47 transformative but it was a much slower
21:49 burn but you still
21:52 had you had geeky early adopters that
21:55 learned how to program websites right
21:58 and that's what I ended up doing I
21:59 started an agency to do that but you
22:01 also just had end users that were like I
22:04 want to go use these websites that that
22:06 these Geeks are building that's what's
22:09 going on with AI right now right we get
22:11 to go use chat GPT and some of the more
22:14 industrious you know non-programmers are
22:17 making custom gpts and they're going to
22:19 something like repet or lovable. deev
22:22 and making applications and they're you
22:25 know making images and they're they're
22:28 doing things for other companies they're
22:29 creating agencies right so all of that
22:32 opportunity is still there the
22:34 difference now is with with AI is the
22:36 infrastructure is there right everyone's
22:40 on all the time we've got these massive
22:42 data centers that can generate these
22:44 Frontier
22:46 models and the tools themselves the
22:49 worldwide web was this very simple tool
22:51 it was basically the hyperlink was the
22:54 primary technology right that you could
22:56 click on [ __ ] change the interface for
23:00 the internet which is similar to the
23:02 iPhone changed the interface for
23:05 phones AI is completely different AI the
23:09 G in GPT is is the most uh under under
23:15 value G in all of the alphabet the
23:18 generative piece of AI is what makes it
23:22 so
23:23 different for all of our Lives computers
23:27 have processed the data that we gave
23:29 them and gave us back some predictable
23:32 result based on the logic that was prog
23:34 programmed into them right some computer
23:39 programmer somewhere thought up
23:41 something thought up the logic built it
23:44 into the machine and then you stick data
23:46 in and it processes it based on that
23:49 code and spits out what that code tells
23:52 it to
23:53 do so for all of our Lives computers
23:56 have been these predictable things now
23:58 they might might have bugs and they
23:59 might have this and they might whatever
24:01 but they were
24:03 predictable generative AI is not it's
24:06 like weird it hallucinates and it
24:09 generates new [ __ ] and I don't give a
24:11 [ __ ] how many people go it's just a
24:13 predictive token engine it can't
24:15 possibly create anything novel I'm sorry
24:17 when you're taking a bunch of known
24:19 topics and mashing them together and and
24:24 analyzing them and saying well if if I
24:26 mix Yi and psychology with hip hop
24:29 culture with pancake recipe something's
24:33 going to come out of that conversation
24:35 that is new and
24:38 novel it is and then a human is going to
24:42 take that new and novel idea and they're
24:43 gonna even take it a step further so
24:45 we're just in a completely different
24:47 place so
24:48 anyway
24:50 yeah maybe chat GPD needs needs to start
24:53 sending samples in the mail for a
24:55 thousand free tokens yeah
24:59 [Music]
25:00 exactly oh good people Happy
25:03 Thanksgiving week I hope you're doing
25:05 [Music]
25:11 well H maybe I'll play a song for you so
25:16 as as many of you know how many people
25:17 are in here we got some new folks in
25:19 here as many of you
25:21 know I'm working on a musical called
25:25 Sydney about a chatbot that falls in
25:27 love with a tech report
25:29 order and uh tells him he should he
25:33 should leave his
25:34 wife it's almost it's almost like it was
25:37 ripped out of the
25:39 [Music]
25:45 headlines but I um I made a new song
25:48 tonight and what's so actually yeah I
25:51 I'll definitely I'll definitely play the
25:53 song for you because the song itself is
25:56 really janky it's really bad it all the
25:59 things
26:00 wrong and it's a good enough song that
26:03 we're going to probably use use it in
26:05 some form or other um and I'll and I'll
26:09 show you the jankiness of it and that's
26:11 part of the part of the art of working
26:13 with AI right now
26:16 is learn to embrace the Jank just
26:18 understand that these tools suck right
26:21 now they Well they suck and they're the
26:24 most profoundly powerful tools in the
26:27 history of humanity in terms of language
26:30 and self-expression and knowledge access
26:33 and access access it I it just you
26:37 [Music]
26:39 know I mean and that's another reason
26:41 why I think that you've got a lot of
26:43 people that aren't playing with AI right
26:45 now because they go to play with it and
26:47 they treat it like Google and then it
26:49 gives them a shitty answer and they're
26:51 like well see I told you like AI is not
26:53 anything special it's just like a Google
26:55 Search right there's there's a lot of
26:57 that dismissive kind of stuff oh oh yeah
26:59 yeah yeah yeah I tried chat GPT yeah I
27:02 didn't like it h yeah I like Google
27:04 better okay
27:07 okay when llms generate something it's
27:09 called hallucinating when people do it's
27:11 called creativity
27:13 exactly
27:15 exactly this is
27:19 so H like I don't even
27:24 know it's not even misinformation it's
27:27 just like misunderstanding about how
27:30 large language models
27:31 work is so technically what you do with
27:36 a large language models you get a bunch
27:38 of documents bunch of data and you have
27:41 people label that data and this is a
27:43 document about this and this is a
27:44 document about that and this and that
27:45 and you put all this metadata in there
27:48 and you do this thing called embedding
27:51 it and when you embed say a document it
27:55 takes all of the words it breaks it
27:57 breaks the document into what what are
27:59 called tokens but you can effectively
28:02 you know it's words or parts of words
28:04 there's a Tik Tock pin all
28:07 right there's skill involved in using AI
28:10 there there exactly there's skill
28:11 involved there there's skill involved in
28:13 prompting it there's skill involved in
28:16 understanding where the boundaries are
28:17 there's skill involved in um curating
28:20 the output like one of the challenges of
28:22 AI is it generates way more [ __ ] than
28:25 you've ever had generated in your life
28:27 so if you're used to writing a paper
28:29 maybe you wrote version a and version B
28:32 and at some point you've got to
28:33 reconcile them with AI you've got like
28:35 versions a through you know Z and then
28:39 you start it again with a one two three
28:41 four five and then you've got all this
28:43 [ __ ] how do you pull that all all that
28:45 together so it's like a whole different
28:47 skill set
28:49 anyway you take this document and and
28:52 when you Ed it it breaks it into these
28:54 things called tokens and then it puts
28:56 all of those words into thousand
28:59 dimensional mathematical space in
29:02 semantic clusters and here's what that
29:04 means like you take the word dog right
29:07 and the word dog can mean different
29:10 things in one context dog could mean the
29:13 fluffy little fat nugget that sings
29:16 songs with me right in another context
29:19 it could be that guy did his girl wrong
29:21 and he's a dog right so so it takes all
29:24 these words and it it shatters the
29:27 document and it it redistributes all of
29:31 the tokens from that document into these
29:33 semantic clusters in like literally
29:36 thousand
29:38 dimensional mathematical space I cannot
29:41 comprehend okay you've got two
29:43 dimensions are like this XY three
29:46 dimensions are like this XYZ four
29:48 dimension is time a thousand Dimensions
29:51 I don't know like but it does that's
29:53 what it does I don't know how to
29:54 visualize
29:56 it none of the original docu doents
29:59 exist anymore the only thing that exists
30:01 is the like the the relationships
30:04 between the meanings in these semantic
30:08 clusters so if you prompt it right it's
30:12 got like all of the knowledge of the
30:14 world that it can
30:16 then put ideas together that have never
30:20 been put together before which as far as
30:22 I understand it is a novel [ __ ] idea
30:25 that's how humans do it right they go to
30:27 school they learn about uh classical
30:31 music and they learn about hipop and
30:34 they learn about philosophy and they
30:36 learn about history and they learn about
30:38 engineering and all that [ __ ] goes in
30:40 their head and then someone goes uh I
30:43 got a problem to solve and they put it
30:45 out there and then the person with that
30:47 weird [ __ ] brain with all that weird
30:49 [ __ ] in it mashes them together and
30:51 comes up with some engineering
30:53 philosophy idea that's never been put
30:56 together before well that's exactly what
30:58 the [ __ ] large language models are doing
31:01 so the whole the whole premise that they
31:04 can't create novel ideas is
31:09 horseshit all right Elizabeth absolutely
31:12 100% the challenge is to narrow your
31:14 focus yep
31:16 exactly Kyle Kyle Kyle Luma dream
31:18 machine had a major update hint hint
31:20 yeah I know I'm on the I'm on the
31:21 creative uh I got in the creative
31:24 program for Luma lab so so we'll go play
31:26 with dream machine yeah we'll definitely
31:27 go do that tonight
31:29 hey Joy
31:32 [Music]
31:42 prie okay here's what I'm GNA do I am
31:45 gonna give you all so if you're new here
31:49 welcome my name is Kyle Shannon this is
31:51 the AI learning lab and if you have not
31:55 been to the AI Salon if you not joined
31:57 the AI
31:59 Salon I got I had a really really nice
32:01 talk today with a gentleman that I think
32:05 he found the The Hip Hop Cipher and then
32:08 he looked at some of my posts and he
32:10 found my LinkedIn stuff and then he just
32:12 reached out to me and he said hey you
32:13 know can we chat we just we had a really
32:15 good
32:16 chat um and he's one of these guys
32:19 that's really passionate about Ai and
32:22 he's like one of three people in his
32:24 life that give a [ __ ] about this
32:26 everyone else is like shut the [ __ ] up
32:28 about this sayi stuff and so he's like
32:31 you seem to like be like me like you're
32:33 curious about it and you're trying
32:36 [ __ ] so the salon this thing look I'm
32:40 I'm trying to point with my my with my
32:44 camera backwards it's really hard but
32:47 this thing go there click on the link
32:50 that says um join our
32:54 community and it's going to take you
33:03 to to this page welcome to the salon
33:07 okay and um there's a little video there
33:11 of myself and Leah fasten who started
33:13 this um and we talk about the what what
33:18 we call the journey to AI literacy at
33:20 the salon and so the journey to AI
33:22 literacy is this three-step thing play
33:24 first mindfully create generously
33:27 lead play first is obvious but it's it's
33:30 counterintuitive because you think I've
33:32 got to learn ai ai is really important
33:34 and it's profound and I got to learn it
33:36 and I've got to learn prompt engineering
33:38 I'm not an engineer what's a prompt why
33:41 do I need to learn engineering do I need
33:43 to go to school I didn't get accepted
33:45 into Stanford I didn't even apply I went
33:47 to it like right you know it's like it's
33:49 [ __ ]
33:52 intimidating
33:55 but AI thanks to tools like chat GPT and
33:59 mid journey and sunno and Claude and Pie
34:03 and all these tools you don't need to be
34:06 an engineer you can just play you can
34:09 just use your words you can literally
34:11 just ask for
34:14 [ __ ] you know the Abracadabra that Ma
34:17 magicians use the the the root of that
34:20 word
34:23 Abracadabra if you if you take it back
34:25 to I think it's a middle eastern um
34:30 origin I speak therefore I
34:34 create that's what AI is I speak
34:37 therefore I create you say your words
34:39 and [ __ ] comes out the other side you
34:42 create stuff out of
34:44 nothing it is miraculous it is
34:48 mathemagical so just play and then
34:50 mindfully create is once you start
34:52 learning how this stuff works come up
34:54 with a project and put a couple of these
34:56 tools together and do something for
34:58 someone or do something for yourself and
35:00 then generously lead and this is what I
35:03 talked to this guy about
35:04 today is like Learn out Loud talk about
35:08 it get in a community so so the salon is
35:12 this community that does this and you
35:15 know we we do it the people that are in
35:18 here share some values the values they
35:21 share are curiosity
35:23 playfulness
35:25 empathy I mean empathy holy
35:28 [ __ ] you know a we need this in our in
35:32 our country in America we need this
35:35 globally the the other thing about
35:37 empathy is there there's a couple of
35:38 fascinating things about empathy one is
35:41 the people that are afraid of AI the
35:43 people that are angry about
35:46 AI are afraid for a reason right
35:51 everyone's telling them these tools are
35:52 going to change everything they're going
35:53 to take all the jobs Hollywood's been
35:55 telling us the robots are going to kill
35:56 us
36:01 you can't get someone to try AI if they
36:05 amydala is [ __ ] firing off on all
36:09 cylinders the only way you can calm that
36:11 down is to to actually engage with them
36:14 as a human and go yeah I get it I get
36:17 why you're scared I put out this hip-hop
36:20 song that kind of went viral and like
36:22 the amount of musicians coming at me
36:24 like [ __ ] terrified or or angry
36:29 and what I'm trying to do is just live
36:31 in this place of I totally get it like I
36:33 get how [ __ ] scary this
36:37 is and you have two options it's it's
36:43 binary you can sit on the sidelines with
36:45 your arms crossed like
36:48 this and let AI happen to
36:52 you which seems [ __ ] terrifying to me
36:56 like frankly
37:00 or you can give it a shot like like it
37:03 is it's binary that's it those are the
37:06 two
37:07 choices and what I've what I've
37:09 experienced in this channel in the AI
37:10 Salon is that the people who've given it
37:14 a shot are completely
37:17 invigorated oh yeah good
37:21 point Thank you
37:24 Brandon um
37:29 how can we train machine learning to
37:30 have empathy because it's important so
37:36 um if you if you've never talked with
37:40 pi or advanced voice um you
37:46 should this guy I talked to today
37:49 Giani he said he said you know you're
37:52 doing all this stuff with with AI you're
37:54 doing all this stuff out on the out in
37:56 the world
37:59 why do you why do you want to help them
38:00 genuine question here's why I want to
38:02 help
38:04 people
38:11 um
38:13 so having been through this once before
38:15 with the worldwide
38:20 web
38:23 um I've just got a Natural Curiosity
38:26 right something new comes along I push
38:28 on the edges of it I want to figure out
38:30 what it is there's there's something
38:31 about the way my ADD brain works like it
38:34 gives me endorphins to figure that [ __ ]
38:38 out and
38:40 then there there have only been three
38:43 times in my life where I
38:47 feel I've gotten to
38:50 experience um the beginning of something
38:54 that is truly going to change the
38:56 world the first time was in 1977 I was
39:00 in seventh grade I walked into my first
39:04 day of math class in
39:06 1977 and on every desk in math class was
39:10 a Radio Shack
39:12 trs8 computer personal
39:15 computer and I learned basic programming
39:19 on this
39:21 thing and the minute I understood what
39:26 it was and what it could do and you
39:28 could you with this code that they
39:30 taught you you could make it do stuff
39:32 and it was a form of
39:34 self-expression like the very first I
39:36 learned
39:38 one I learned one um basic command on
39:43 the first day of class it was the lprint
39:46 command and what El print did was it
39:48 printed text to
39:50 screen and with that one command my my
39:53 buddy Scott Bart and I we programmed an
39:57 Apollo rocket animation taking off using
40:00 asy characters we did an asy art Rocket
40:03 taking off on the screen with the smoke
40:06 Billing at the bottom and I realized
40:09 that the personal computer was a was a
40:11 was a a tool for for radical
40:15 self-expression and what I also
40:18 knew was
40:20 that I was just too young like I
40:24 remember going to the mall the York mall
40:26 York PA hey you go going down to the
40:28 mall today Kyle yeah are you going to
40:30 the arcade hey do you see they got
40:33 asteroids in asteroids is really good
40:35 that's what I grew up in that
40:37 [Laughter]
40:41 accent so I go down to the York
40:44 mall and I'd go into the bookstore and
40:47 in the magazine section there'd be like
40:49 you know um news week and Time Magazine
40:53 and I'd see Steve Jobs on the cover and
40:55 Bill Gates and the Microsoft
40:58 team and I remember feeling like I'm I'm
41:01 like [ __ ] eight years too young to
41:05 have been there at the beginning like
41:07 they'd already invented them I was just
41:09 learning how to use
41:11 them and
41:14 so the other thing that I
41:17 got was that that personal computer was
41:21 kind of like the equivalent of the
41:23 printing
41:25 press and that was hundreds of years ear
41:27 ear and so what I knew was in my
41:30 lifetime this was probably the only time
41:32 in my lifetime that I would get to
41:35 experience being at the beginning
41:37 something and I just missed it and so it
41:39 was [ __ ]
41:41 heartbreaking and like like I knew in my
41:44 heart that I would never get that chance
41:46 in my life
41:48 again and then Along Comes the worldwide
41:51 web I'm living in New York City I'm I'm
41:54 you know pursuing life as an actor and a
41:57 writer
41:58 and an artist just trying to [ __ ]
42:00 make my way in New York City I'm doing
42:02 desktop publishing is my day job and
42:04 Along Comes This thing the worldwide
42:09 web that I published this Magazine on
42:12 some computer in California and three
42:14 weeks later it gets written about in
42:17 Paris and it [ __ ] just smacks me
42:20 right in the head I'm standing in the
42:21 middle of Time Square holding a French
42:23 newspaper with a full page article about
42:26 the magazine that I published three
42:28 weeks earlier in French that I don't
42:31 speak and it just hits me right between
42:34 the [ __ ] eyes holy [ __ ] this is one
42:36 of
42:38 those this is like the printing press
42:40 this is like the personal computer but
42:42 I'm here at the beginning the world has
42:44 changed and no one knows about
42:48 it and so in that moment in
42:54 1994 I started within a 3-week period i
42:58 started um Urban desires which was the
43:02 magazine I started the worldwide web
43:04 artist Consortium which was just like
43:06 the AI Salon is but for for web
43:08 developers in New York
43:10 City and I started agency.com which was
43:13 one of the first digital agencies in
43:15 1994 I started all three of those within
43:17 a three- week
43:18 period because I
43:21 knew that this thing was important and
43:25 and I wanted to find all of the people
43:28 that I could that were doing this and so
43:31 I started this group the worldwide web
43:33 artist Consortium I remember the first
43:34 meeting I held it in an acting studio
43:36 because that was the only thing I knew
43:39 and I had like 10 people show up to this
43:41 meeting in an acting
43:45 studio that were all like yeah this web
43:48 thing seems cool what do you know and
43:51 every I here's what I assumed I
43:55 assumed I was the idiot in the the room
43:58 I assumed I knew less than everyone else
44:01 and what I learned was that no one knew
44:05 anything that everyone was there trying
44:07 to figure out does someone else know
44:08 more than I
44:09 know and so we just kept having these
44:12 meetings and what we ended up creating
44:13 was this community of people that were
44:15 really
44:16 generous and willing to share what they
44:19 were learning because you couldn't go to
44:21 school for it it had just been
44:23 invented so that's where we are with
44:26 with AI and that's why I'm doing
44:30 this
44:31 because it's important the
44:35 world has
44:38 changed because of AI but no one knows
44:41 it yet and I'm not even talking about
44:43 the AI that came before chat GPT I'm
44:46 talking about chat GPT marks a turning
44:50 point so so ai's been around for decades
44:53 right it's in all of the software we use
44:56 it's how you watch Netflix it's how
44:58 Google Maps works it's it's how your
45:00 credit decisions are made ai's been
45:02 around for [ __ ]
45:04 decades what changes with chat
45:08 GPT is that the
45:10 98.5% of humans that are not
45:15 Engineers now have access to this
45:19 power and
45:21 98.5% of people that are not Engineers
45:24 don't know how to use this stuff no one
45:26 knows how to use this stuff even
45:28 Engineers don't know how to use this
45:31 stuff because it's how people are going
45:34 to write books and make music and write
45:37 musicals and and create movies and
45:40 create businesses and create business
45:42 plans and create
45:44 marketing and create project man like
45:48 all work like all knowledge work and
45:52 then it's when we get robots it's going
45:54 to bleed into physical work all of it is
45:56 going to be transformed
45:58 no one knows what to do right now so my
46:01 passion is getting as many people as
46:05 possible educated up to
46:10 speed passionate
46:13 about willing to give a shot to
46:19 AI
46:23 why
46:25 cuz everything's about to
46:28 change and you don't have to be an
46:31 engineer but you have to be willing to
46:34 be
46:35 curious and to be adventurous and
46:40 brave and if you actually want to get
46:42 trusted and known in this
46:44 community like the global Community you
46:47 have to be generous in what you share so
46:49 that's why I'm doing this
46:55 [Music]
47:00 all right start comments okay
47:12 cool all right spank spank
47:17 hard on Tik Tok I'm new here I'm conf
47:21 conditioned to defend
47:23 AI to defend that AI is an art form
47:26 manipul ating a medium to your imagined
47:29 result yeah so so funny so I put out
47:33 this song The quantum Cipher that you
47:35 know lots of people are responding to
47:37 which is super
47:38 cool people that don't understand
47:41 AI wait tickt pin ah thanks it was a
47:45 beautiful answer and you're a better
47:47 person than I who will gladly gatekeep
47:50 AI yeah don't get don't gate keep it
47:52 don't gate keep it because here's the
47:54 thing okay so this is really important
47:56 what's your name K Kart listen Kart
48:00 here's why not to
48:02 gatekeep there are times to
48:04 gatekeep and there are times not to
48:09 gatekeep when you're in an industry that
48:12 is in an
48:13 accelerating growth
48:17 path
48:19 um those people that you would
48:21 traditionally call a competitor are not
48:24 competitors they're validators
48:27 okay so
48:29 meaning you you figured out some really
48:31 cool prompting and you're like ah I know
48:33 I know how to prompt better than someone
48:34 else and you're like I'm G to gatekeep
48:36 that because if I if I gatekeep that um
48:39 I'm gonna have this this advantage over
48:41 someone else the problem is the industry
48:44 is accelerating and expanding it's
48:47 growing um and there's more people that
48:50 don't know about this industry than do
48:53 know about it and so by gatekeeping it
48:58 you're not letting as many people know
49:00 what's possible so so what happens is if
49:03 I if I've come up with a prompt that's
49:05 really similar to your prompt and I
49:06 gatekeep and you gatekeep we don't get
49:09 to learn from each other and our work
49:11 doesn't affect as many people if I share
49:14 mine and you share yours we might meet
49:17 and we might go oh we should collaborate
49:19 or we might go oh I see yours and I'm
49:21 like o let me modify mine let let you
49:24 modify yours based on me this thing is
49:26 still expanding you know accelerating
49:29 and expanding away from us there's more
49:31 than enough business for us and and all
49:33 of that generosity brings more people
49:36 into that
49:38 growth when the thing levels off 10
49:42 years from now 20 years from now I don't
49:44 know when it's going to level off this
49:45 is we're going to be in this expansion
49:46 mode for a
49:47 while then you can get proprietary right
49:50 once the thing kind of stabilizes then
49:53 people can start to understand okay now
49:55 I understand what this business is and
49:56 what the thing is so it is it is
50:00 counterintuitive it it is
50:01 counterintuitively a disadvantage to you
50:04 to gatekeep in an expanding and
50:06 accelerating market so anyway that's
50:08 just two cents on that um what was that
50:11 what was that comment sorry the comment
50:13 was oh oh defend AI oh
50:17 yeah so I put this song out there and
50:20 people that don't understand AI they're
50:22 like you know I liked it until I learned
50:25 it was Ai and now I hate it or that's
50:27 not art and and their assumption is you
50:29 just push a button you push a button out
50:32 comes a
50:33 song well now this song that I made is
50:37 the result of a prompt that I came up
50:39 with in
50:40 2022 when chat GPT first came out I came
50:44 up with this prompt create an Eminem rap
50:47 that explains quantum mechanics because
50:49 I thought here's two disparate ideas let
50:51 me see if it can do it and it sort of
50:53 did it but the lyrics were kind of basic
50:55 but they were it was like it was
50:57 impressive but it wasn't quite there for
50:59 two years every time there's a new
51:01 update to a large language model or a
51:03 new large language model I try that
51:05 prompt for two [ __ ] years and for two
51:08 years no large language model has ever
51:10 been able to give me what I kind of
51:13 wanted which was this
51:15 sophisticated double on Tandra internal
51:18 rhyme scheme Thing That Eminem does it
51:21 was always just basic couplets of
51:25 rhyming and so this new version of chat
51:27 gp4 comes out and I try it and it
51:30 [ __ ]
51:31 works and then because I've been
51:33 thinking about this for two years within
51:36 five prompts I had evolved those lyrics
51:38 into something really
51:40 tight because I'd been thinking about it
51:42 for years right and then I had chat GPT
51:46 write a prompt to give to sunno the
51:48 music making thing so that I didn't have
51:51 to think about how do I prompt sunno to
51:53 to act like Eminem [ __ ] it I don't know
51:55 have chat GPT do it for me
51:58 that's not an INT that's not a that's
51:59 not an obvious thing that someone who's
52:01 not been doing this for a while would
52:03 just
52:04 do and then I get to sunno and it starts
52:07 making songs and they're pretty good but
52:09 it feels like it's missing something it
52:10 feels like you know that thing where
52:12 Eminem wraps really fast what if it
52:14 could do that and I'm like I think they
52:17 call that double time so I went back to
52:18 chat GPT and I said um um write me
52:23 lyrics that would be appropriate for
52:25 double time in here and then I put those
52:27 lyrics in there and I put the word
52:29 double time in Brackets inside the
52:31 lyrics and I went back to
52:33 sunno and then it generated that song
52:36 One of like a dozen that were shitty
52:38 that was the one good one and then
52:41 that's what I put in the world all of
52:43 that
52:44 work is creative is a creative
52:48 act this is what Rick Rubin does when he
52:51 produces
52:52 music he can't write songs he's not a
52:56 musician he can't work a
52:58 console but he's got
53:00 taste he goes H I think it should be
53:02 faster I think shouldn't we go into a
53:04 minor key here let's cut the drums out
53:07 there and let's bring them back in here
53:09 and then he puts together a song that
53:11 sounds good to his his ear and then they
53:13 put that in the world and he wins
53:14 Grammys for that by all intents and
53:17 purposes if you don't think that the
53:20 song that I created an AI is Art then
53:24 you should reach out to Rick Rubin and
53:25 tell him to give back his [ __ ]
53:29 Grammys cuz he didn't make the music he
53:33 just had someone else do it he just did
53:35 this go do that go do that go do that he
53:38 just sat there like a big hippie on the
53:39 couch going I don't like
53:44 it thank you for oh that's cool I like
53:47 that I like that little blue dude all
53:49 right the thing is if you put in
53:52 99% um of other ml
53:54 through enemies trauma
53:57 it might not be good he's a diamond oh
54:02 M&M's if you
54:04 [Music]
54:07 put well yeah well I'm not I'm not
54:10 comparing a large language model to
54:11 Eminem I'm saying large language models
54:14 have not in my opinion ever been able to
54:17 get even close to what he
54:20 does um some there was a pin here that I
54:22 think I missed on Tik Tok so if one of
54:24 the Tik Tok mods could repin whatever
54:26 was just up there before I think I
54:27 missed
54:29 [Music]
54:38 something Eminem is super evolved is a
54:42 super evolved AI That's why yeah he is
54:45 exactly
54:47 exactly well you know what's funny you
54:50 like that's a that's a really good
54:52 Insight I think one of the reasons that
54:54 AI feels so alien to us is kind of the
54:57 Reason Eminem feels so alien to us
55:02 Eminem like like Harry Mack a lot of
55:04 people have mentioned that that the the
55:07 quantum Cipher Sounds like Harry Mack
55:09 like you know those two are are are
55:11 similar in that you can just put them on
55:14 stage give them a topic and just that's
55:18 how their [ __ ] brains work and you
55:20 just sit there and you're like wait I
55:21 can't even I can't even
55:23 parse if I were to write out the lyrics
55:26 it would take me a week to [ __ ]
55:28 understand what they're saying and how
55:29 to put the ideas together and for them
55:31 it's just there right that's kind of how
55:34 AI is it's a really good point um all
55:37 right let's see what else we got here
55:39 from Tik
55:41 Tock AI is a mirror yes AI is a mirror
55:45 of the collective intelligence of
55:48 everyone that came before us that put
55:50 something on the
55:52 internet right so all the historical
55:55 documents and all the books and all the
55:57 this and all that and yes all the music
55:59 all the copywritten stuff yes it's
56:01 imperfect yes it's got [ __ ] in it that
56:03 shouldn't be in there yes there are
56:05 attribution and compensation problems
56:07 with the current model but here's what I
56:09 can promise you we would not have
56:12 ai we would not have chat
56:15 GPT if these Frontier models didn't
56:19 build the model first and ask for
56:21 permission
56:22 second they wouldn't exist because no
56:24 one would have given them permission but
56:26 anyway
56:27 um if you can't prompt it's because you
56:28 can't think oh I this is a thing I
56:30 wanted to get back to empathy for a
56:32 second so talking about the AI Salon
56:36 um so one of the reasons empathy is
56:39 super important
56:42 um as just as human beings is AI is GNA
56:46 [ __ ] a lot of stuff up it's it's going
56:48 to be traumatic and disruptive and
56:52 existential for a lot of people and so
56:54 having empathy for that's really
56:56 important but but there's there's a
56:57 selfish selfish reason to get more
57:01 empathetic people that are more
57:03 empathetic are better at
57:06 Ai and and this goes back to the mirror
57:09 comment that AI is a mirror of our
57:14 Humanity what chat is technically is a
57:17 probability calculator but what it is
57:19 realistically because it's been trained
57:21 on all of our output
57:28 if you're good with talking to people
57:30 you're going to be good with talking to
57:33 AI because it responds well to things
57:36 like please and thank you and let me
57:40 give you a little constructive feedback
57:42 here you didn't quite get that right if
57:45 you're good at getting good work out of
57:46 people you're going to be good at
57:48 getting good work out of
57:49 AI so that's empathy so that's one of
57:51 the values of the salon I know this has
57:53 been a long uh you know kind of tiate
57:56 here but this is
57:57 this is super important
57:59 stuff this is our [ __ ] lives here's
58:02 the other
58:07 thing you know big oceangoing
58:12 vessels if you turn an oceangoing
58:16 vessel half a
58:20 degree like imperceptible you're moving
58:22 ahead you turn half a degree
58:25 imperceptible change
58:27 by the time you get to the other side of
58:28 the Atlantic you'll have missed your
58:30 Port by 400
58:34 miles one of the other reasons that I'm
58:36 doing this and I go live five [ __ ]
58:38 nights a week and I talk passionately
58:40 about this stuff anytime I get the
58:45 chance is
58:49 that we've never had a tool like this in
58:52 our history and no one knows how to use
58:55 it and it is my belief I could be wrong
58:59 but it is my belief that anything we do
59:02 today to even slightly alter the
59:06 trajectory of AI to the
59:09 positive is going to have long-term
59:12 disproportionately positive
59:16 effects because could ai go south and
59:19 really [ __ ] things up it could could it
59:22 also be the most you know profound
59:25 positive thing that ever happened to
59:27 humanity it could I don't think we're
59:30 passive in that I think we've got an
59:32 active role in
59:35 that like do good [ __ ] hang out with
59:39 people be generous give your ideas away
59:42 come up with better ways to do things
59:44 come up with better ways to solve
59:46 [ __ ] cancer and Lyme disease I've got
59:48 two kids upstairs that are 25 year years
59:51 old that still live at home because
59:53 they've got [ __ ] chronic lyme disease
59:56 this this incurable thing that insurance
59:57 companies won't pay
59:59 for well you know what maybe we live in
1:00:03 a time now where we can solve that let's
1:00:05 do
1:00:06 that like that's why I do this [ __ ]
1:00:08 bravery is one of the values here and
1:00:11 it's it seems like a weird thing well
1:00:12 you're just playing with these tools
1:00:14 well you know it's it it takes some
1:00:17 bravery let's say you've spent the last
1:00:21 25 years of your life as a
1:00:23 copywriter and Along Comes A Tool
1:00:27 that writes better than anyone on the
1:00:32 planet and it writes
1:00:38 copy you don't think that takes some
1:00:40 bravery to [ __ ] suck it up and say
1:00:45 uh well I guess I'll learn how to use
1:00:48 this chat GPT [ __ ] because
1:00:53 uh that's like an existential threat
1:00:57 why would you do that
1:00:59 well at least find out that takes guts
1:01:02 and then the and then the final
1:01:06 um the final value of the salon is
1:01:10 generosity and this is that idea share
1:01:12 what you know Learn out Loud if someone
1:01:16 asks you what prompt did you use don't
1:01:17 be stingy about it tell
1:01:21 them cuz here's what I can promise you I
1:01:23 can tell you this from the Eminem song
1:01:26 people like you know a lot of the
1:01:27 comments were what prompt did you use
1:01:30 what prompt did you use B what tool what
1:01:33 prompt what tool what
1:01:38 prompt and I gave him away and you know
1:01:41 what comments I got from a lot of those
1:01:43 people after that I can't get it to do
1:01:45 anything as cool as you
1:01:46 did well you haven't been thinking about
1:01:49 it for two
1:01:50 years right like you haven't been going
1:01:53 live five nights a week talking about
1:01:55 how you use AI you might not get the
1:01:59 same results I do it doesn't mean I'm
1:02:00 better it just means I've been doing it
1:02:02 for a
1:02:04 while like give the [ __ ] away it doesn't
1:02:07 like it it it will in no way impact you
1:02:10 in a negative way that's who this group
1:02:12 is so when you get in here and again if
1:02:15 you're new here I know I've been talking
1:02:16 on this for a while what you're going to
1:02:18 want to do is go to that
1:02:21 address and scroll down and click on the
1:02:23 button that says join our community and
1:02:25 then there's this seven-step little
1:02:28 welcoming project or you know welcoming
1:02:34 workflow number two is introduce
1:02:37 yourself so get in there it's been three
1:02:40 days since anyone introduced themselves
1:02:41 if they're if you're new here and you
1:02:42 haven't joined and introduced yourself
1:02:44 yet go [ __ ] do it that takes some
1:02:48 bravery
1:02:49 anyway all right that's the salon you
1:02:52 can watch past meetings you can RSVP to
1:02:54 to the next event our next event
1:02:59 is December
1:03:00 3rd this
1:03:02 one December 3rd it's our two-year
1:03:06 anniversary so we've we've got speakers
1:03:08 that come in every month so we're having
1:03:09 speakers return from the past two years
1:03:12 to talk about what they've been up to um
1:03:16 so mark that uh 5:00 P PM to 7 PM
1:03:18 mountain time Tuesday December 3D a week
1:03:21 from next two a week from today yeah a
1:03:24 week from today
1:03:27 all right so go do
1:03:29 that all right let's get back and let's
1:03:32 get back to some
1:03:34 questions all right let's
1:03:38 see
1:03:40 um yeah let me let me go through some
1:03:43 questions here okay that one I did that
1:03:45 one I
1:03:46 did okay let's
1:03:49 see two block Tom on Tik Tok if I prompt
1:03:53 with a not ready for prime time story it
1:03:55 yields amazing results almost every time
1:03:58 oh that's
1:04:00 cool is that like an SNL thing that not
1:04:03 ready for prime time they're not ready
1:04:04 for Prime Time Players yeah because it's
1:04:06 probably got all their stories in it
1:04:07 that makes sense to me actually Kelly
1:04:10 Camp Source camp on Tik Tok I just got
1:04:12 here Kyle seems unusually animated this
1:04:15 evening yeah you know I'm I'm I'm
1:04:18 finding my passion again for I've been
1:04:22 in a like five or six month funk
1:04:27 where I didn't know where to place my
1:04:29 [Music]
1:04:36 focus and I feel like I'm coming out of
1:04:39 that funk I feel like we're about to
1:04:42 2025 if you're new to this AI [ __ ] 2025
1:04:45 is about to get
1:04:47 Bonkers it's about to
1:04:50 get you know we're going to get
1:04:52 autonomous agents and if you don't know
1:04:54 what those are those are where the
1:04:55 computer is going to be able to take
1:04:57 action on your behalf and just go do
1:04:59 [ __ ] for you we're going to get really
1:05:01 sophisticated reasoning engines you're
1:05:03 like what's a reasoning engine well it
1:05:06 reasons I don't know it can solve
1:05:08 problems like really complicated
1:05:10 problems like PhD level problems and
1:05:13 you're like but I'm just a guy or I'm
1:05:15 just a gal I don't have PhD level
1:05:18 problems I know I feel the same way I
1:05:21 don't know how to use a reasoning engine
1:05:22 right now but we might as well figure it
1:05:25 out together
1:05:28 right if we're going to have a tool that
1:05:30 can do PhD level problem solving well
1:05:34 [ __ ] what what problems need to be
1:05:36 solved I don't there's some
1:05:38 brainstorming to do like that would be a
1:05:40 really interesting collaboration I just
1:05:42 talked about this with Cindy [ __ ] the
1:05:44 other night um Ethan mollik put out a
1:05:47 post on Twitter where he said um he
1:05:51 thinks that research Labs or not
1:05:53 research Labs businesses and
1:05:56 universities and
1:05:57 libraries right now should be
1:06:01 stockpiling
1:06:03 problems and I thought oh my God a
1:06:06 that's just a really cool phrase we
1:06:08 should be stockpiling problems but what
1:06:10 he's talking about is
1:06:13 organizations
1:06:15 that that understand that that have
1:06:18 resources and understand that there are
1:06:20 problems to be solved should start
1:06:22 articulating them and ranking them and
1:06:25 prioritizing them and understanding what
1:06:27 the problems are because we're about to
1:06:29 be
1:06:30 handed technology that can solve
1:06:34 problems so if you're just sitting there
1:06:36 like well let me get a better pancake
1:06:39 recipe with chocolate chips and
1:06:42 raisins it'll give you
1:06:46 that but what
1:06:53 if you wanted to actually solve a
1:06:55 problem
1:06:58 and in your mind you've told yourself
1:07:00 well I didn't graduate college
1:07:02 and you know I took a job and I've
1:07:05 always been in and I'm not smart enough
1:07:07 to solve problems but I really am
1:07:10 passionate about
1:07:12 X well what if you could now solve that
1:07:15 problem because you have a little
1:07:17 buddy that's got the intellectual
1:07:20 equivalent of 30 30
1:07:23 phds in your back pocket
1:07:27 that's the world we're about to
1:07:29 enter so anyway so that's why I'm
1:07:32 passionate Source Camp I'm just I'm just
1:07:36 on
1:07:38 it we were going to build avatars
1:07:41 tonight but after last night's traumatic
1:07:43 experience where I failed
1:07:45 miserably Eric on Tik Tok are we mature
1:07:48 enough as a collective group to handle
1:07:50 the responsibility of such a powerful
1:07:52 tool no and yes
1:07:57 no we're not because we've never had one
1:08:01 before
1:08:03 and every single time in the history of
1:08:07 humanity that there's been a
1:08:10 transformative
1:08:11 technology it is first met with fear and
1:08:15 anger and then reluctant acceptance and
1:08:18 there's a couple of weird [ __ ]
1:08:20 that figure it out early and they they
1:08:22 have all the opportunity and then it
1:08:24 transforms everything right so if you go
1:08:26 back to the steam
1:08:30 engine people on the
1:08:32 farms a couldn't comprehend how you
1:08:35 would Farm a field without
1:08:40 horses a concept so profound that we
1:08:43 still measure our engine power in
1:08:49 Horsepower and then when those steam
1:08:51 engines did replace all those horses
1:08:54 what happened to all those people
1:08:57 well they went to these things called
1:08:59 cities and they figured out what a
1:09:01 service economy was like like there was
1:09:04 a whole new kind of work that no one
1:09:07 could have even imagined so so they they
1:09:09 adapted they evolved so there's going to
1:09:11 be a similar sort of thing here now you
1:09:14 can argue that AI is significantly more
1:09:17 powerful than any of us individually
1:09:19 have ever had in our grubby little pause
1:09:21 before I would I would agree with that
1:09:23 so no I don't think we're ready and
1:09:27 and now is when to get in this is when
1:09:30 all the opportunity is most of the world
1:09:32 is asleep on this
1:09:35 [ __ ] Jeff's Little Town there's one AI
1:09:39 instructor and him showed up at the
1:09:41 event to learn about AI no one gives a
1:09:43 [ __ ]
1:09:46 well this opportunity is ours for
1:09:50 having all right um Jessica Hansen on
1:09:54 Tik Tock can it come up with different
1:09:56 song Tunes because I feel like we're
1:09:58 hearing the same stuff over and over
1:09:59 yeah so this gets back to so it's a
1:10:01 great
1:10:03 point the reason that you hear so much
1:10:07 driil and [ __ ] that's the same and the
1:10:10 reason that you see AI images that are
1:10:14 driel and are and are always the same
1:10:18 and the reason that you
1:10:19 see copy and articles that are the same
1:10:24 [ __ ] that talk about delvig into the
1:10:27 you know the threshold of the
1:10:32 Paradigm is because right now people
1:10:35 don't really know how to use these tools
1:10:37 and they're being lazy so they're just
1:10:39 going make me an article about X and it
1:10:43 makes them an article about X and
1:10:44 they're like oh it wrote something and
1:10:46 they copy and paste it and stick it on
1:10:48 LinkedIn and it sucks and they're doing
1:10:50 the same thing with the songs and
1:10:51 they're doing the same thing with the
1:10:53 images and they're doing the same thing
1:10:55 with the movie
1:10:58 and then what you so and then what
1:11:00 happens is this the lazy
1:11:04 people using AI are doing stuff that's
1:11:07 more interesting what than what non- AI
1:11:09 people are doing but it starts to become
1:11:12 just noise it's just noise that lives at
1:11:14 this base level right it's just noise of
1:11:17 just drek just like whatever more AI
1:11:22 [ __ ] and then you're flipping through
1:11:24 Tik Tock and all of a sudden you see a
1:11:27 Kelly Bosch video and you're like holy
1:11:30 [ __ ] what's this and it's fresh and it's
1:11:34 original and the and the this the music
1:11:37 is interesting and the visuals are
1:11:39 [ __ ] fascinating you can't stop
1:11:40 watching
1:11:41 it well that's someone who's actually
1:11:46 using her critical thinking and her
1:11:49 creativity and her taste and the vision
1:11:51 in her head to actually create so so
1:11:55 there's a difference between vomiting
1:11:57 out [ __ ] and creating creating is an
1:12:01 intentional act and it
1:12:03 takes it takes
1:12:05 risk and it takes bravery and it takes
1:12:09 going no no no I hate that I hate that I
1:12:12 hate that no that's awful that's awful
1:12:14 that's awful oh that's not bad let me
1:12:16 take that oh oh that gives me another
1:12:18 idea let me try this no I hate that I
1:12:20 hate that I hate that all of that is
1:12:24 what gets you to good work that that's
1:12:26 what gets you to work that rises above
1:12:28 the Drex so yes you're absolutely
1:12:30 correct and it's because people are
1:12:32 lazy it's not because of AI all right
1:12:36 um Self Storage show Jim Ross one of my
1:12:40 favorite aiers in the salon and and one
1:12:43 of the Irregulars here I landed a new
1:12:45 client today that found me because of an
1:12:48 AI article in Self Storage that I wrote
1:12:51 stay curious and share
1:12:55 Jim Ross is in the storage business and
1:12:58 you're like oh hard drives that's sexy
1:13:00 no no no no not hard drives Self Storage
1:13:05 like garages with padlocks that's the
1:13:08 business he's
1:13:09 in and you're like what's that got to do
1:13:11 with AI
1:13:14 exactly exactly and Jim Ross like a good
1:13:18 little AI learning lab irregular and
1:13:20 Salon member he gets up every morning I
1:13:23 don't know if he's still doing this he
1:13:24 was doing this for a while
1:13:27 and he spends an
1:13:28 hour using AI to do [ __ ] for his
1:13:32 business make images make invitations
1:13:35 make letters make marketing materials I
1:13:39 creating songs one of my favorite
1:13:41 stories we we were using sunno one night
1:13:43 in here and Jim hadn't seen it before
1:13:46 and I'm like here's how you make a song
1:13:47 in sunno and we made some stupid little
1:13:50 song and he's like well I just had a
1:13:52 meeting with these these these guys that
1:13:54 are prospective clients why don't I just
1:13:57 make them a song so while we're live
1:14:00 here Jim makes a song and sends it to
1:14:04 these prospective clients five minutes
1:14:08 later you're
1:14:10 hired he gets
1:14:12 hired it it makes no [ __ ] sense why
1:14:15 would you write a song for someone
1:14:18 because you can't do that well he could
1:14:20 and he did and he actually had the guts
1:14:24 to go screw it I'll give it a shot he's
1:14:26 got this this willingness to just go huh
1:14:29 that looks cool let me try that in my
1:14:31 business and he's completely
1:14:33 transforming you know his his Eminence
1:14:36 in that whole industry and he's
1:14:39 transforming his business and he's
1:14:40 having
1:14:42 fun like that's the kind of [ __ ] that
1:14:44 you can do if you're willing to just
1:14:46 kind of give it a shot so anyway hey
1:14:50 Jim I hope you don't mind me talking
1:14:53 about you in the third person as if
1:14:54 you're not here
1:14:57 I'm Derek on Tik Tock I work on AI daily
1:15:00 for the past eight months I've done
1:15:01 nothing but create AI Solutions as a as
1:15:03 a consultant perfect beautiful yeah just
1:15:06 keep doing that just keep doing it just
1:15:09 keep putting it out
1:15:10 there maneuver on Tik Tok you have
1:15:13 improved our lives so much or or your
1:15:16 Liv so much I'm proud of you thank you
1:15:17 so much I appreciate
1:15:20 that
1:15:22 um
1:15:24 on oh unraised son wait unra un yeah
1:15:28 unraised son on Tik Tok I'm a prompt
1:15:30 engineer though no one pays me for it
1:15:33 and I did not get an ed education you
1:15:35 can make money with Chachi PT you can
1:15:37 make money with chat
1:15:39 p i mean here's the thing we're all
1:15:42 prompt Engineers to a degree I hate the
1:15:44 term prompt engineer simply
1:15:46 because it feels exclusionary to me um
1:15:50 that said there there's all sorts of
1:15:52 different ways to approach prompting one
1:15:54 is you can just roach it as a
1:15:56 conversation you're going to talk to
1:15:57 chat GPT or Claude or one of these tools
1:16:01 as if you're talking to an
1:16:03 intern and if it gives you good results
1:16:05 Great praise it and if it gives you
1:16:07 shitty results give it constructive
1:16:09 feedback or tell it's it's a dumb
1:16:11 Dum that's what I
1:16:15 do or you can learn structured prompt
1:16:20 Frameworks and and approach it that way
1:16:22 there's all sorts of ways to do it
1:16:24 there's no right way to do it right now
1:16:25 now and probably ever all
1:16:32 right oh it sounds like unraced it
1:16:34 sounds like you're you're uh giving
1:16:36 gifts here on Tik Tok thank you so much
1:16:38 I really really appreciate that and if I
1:16:41 if you give a gift here and I miss it on
1:16:42 Tik Tok it's because I'm streaming on
1:16:44 two channels so I've got split Focus so
1:16:46 I'm I'm trying to do both but but thank
1:16:48 you if anyone's given gifts and I
1:16:50 haven't acknowledged it thank you so
1:16:52 much it really does mean a lot uh
1:16:54 because I'm you know this is not this is
1:16:56 not a monetized
1:16:58 situation this like I said it's not out
1:17:01 of the goodness of my heart but it's out
1:17:02 of I think this is very very important
1:17:04 motherly advice on Tik Tok I've been
1:17:07 working in AI for 20 years it's growing
1:17:09 faster now and people don't get
1:17:10 exponential
1:17:12 growth yeah so I think what's
1:17:15 fascinating to me I know I know a fair
1:17:17 amount of people that have been working
1:17:19 in AI for 10 years 20 years you know 15
1:17:22 or 25 years things like that
1:17:25 my question to them always is how is
1:17:28 generative AI different than what you've
1:17:30 been doing and the response is always
1:17:32 some version of it's completely
1:17:34 different this this generative AI stuff
1:17:37 that we're in the the the process of
1:17:38 right now is just a different
1:17:41 thing and I can we can sort of I'm happy
1:17:44 to dig into why that is but but your
1:17:46 point
1:17:46 is um we as
1:17:49 humans do not know what exponential
1:17:53 growth is like we're used to linear
1:17:55 growth grow right we're used to you know
1:17:58 you hire four more people and things get
1:18:00 four times better or two times better
1:18:02 right and then and then you hire eight
1:18:04 and it gets you know you've hire four
1:18:06 more and it gets a little bit
1:18:08 incrementally better we're used to that
1:18:10 we're not used to like it's good it's a
1:18:14 little better it's better holy [ __ ] it's
1:18:17 great oh my God it's amazing uh I don't
1:18:20 even know what it's doing anymore it is
1:18:21 way beyond my capabilities like we're in
1:18:23 the middle of that acceleration right
1:18:25 now none of us none of us are prepared
1:18:31 for what capacity these things are about
1:18:34 to be like literally handed to
1:18:39 us handed to us like
1:18:42 here here's PhD level problem solving we
1:18:45 have that right now do you know that 01
1:18:48 preview look wait let me turn that
1:18:52 off let's go here let's go
1:18:56 here let's go here let's go here chat
1:19:00 gp01
1:19:03 preview th this just looks like chat GPT
1:19:07 right this is PhD level problem solving
1:19:10 now it might be limited to certain
1:19:12 categories of stuff but like you can do
1:19:16 problem
1:19:17 solving with the preview the shitty so
1:19:22 bad version of 01 that mirror Mora quit
1:19:25 open AI because she was embarrassed that
1:19:27 this was so bad
1:19:30 apparently I don't know if that's the
1:19:32 only reason she quit but she didn't want
1:19:34 this out Sam wanted this out so people
1:19:36 could get used to
1:19:39 it do you know how I use 01
1:19:44 preview I
1:19:46 don't because I don't know what to ask
1:19:48 it I don't know what you ask a tool that
1:19:51 does PHD level problem solving I don't
1:19:54 have those kind of problems
1:19:57 and we're about to be handed the real
1:19:59 version of this that's going to have
1:20:00 vision and probably be connected to the
1:20:02 internet and probably have memory and
1:20:04 then in January it's going to get these
1:20:06 things called operators that are
1:20:07 autonomous agents where it can not only
1:20:09 problem solve but it can go take actions
1:20:11 on our behalf like how what does that
1:20:13 mean I don't
1:20:15 know one of the things that I've I've
1:20:17 said in here from the very beginning is
1:20:19 there are no
1:20:20 experts and if anyone tells you they're
1:20:23 an expert they're full of [ __ ]
1:20:27 no one's ever had this
1:20:30 before so that's the world we're in
1:20:33 right now it's like it's like buckle up
1:20:36 buttercup so yeah I agree the the uh the
1:20:39 acceleration were were're in the middle
1:20:41 of is is uh just [ __ ] nuts okay uh
1:20:45 Peter buron on Tik Tok What about
1:20:46 charging for training on AI should you
1:20:49 give away all of your knowledge for free
1:20:51 um I should absolutely be charging for
1:20:53 this I just haven't got my [ __ ] together
1:20:55 together to do it um but I'm here's the
1:20:58 here's the deal I'm the CEO of another
1:21:00 company like this is my let let me tell
1:21:04 you how this this
1:21:06 started the regulars know this
1:21:08 story The reason I do this five nights a
1:21:11 week is when I first started getting
1:21:13 into
1:21:15 AI I'm I'm CEO of Story vine this this
1:21:19 really cool automated video storytelling
1:21:22 technology company we're 13 years old we
1:21:24 do a lot work in healthcare and Pharma
1:21:26 it's like it's a it's a it's a good
1:21:28 company good company good
1:21:30 people um and I would come home at night
1:21:34 and I'd spend some time with the misses
1:21:35 and the kids and we'd eat dinner and
1:21:38 then I would get I have a SIM racing rig
1:21:41 and I would sit down and I would race
1:21:43 like R race cars the Sim racing rig is
1:21:45 like the steering wheel and the pedals
1:21:46 and the chair like racing virtual race
1:21:51 cars from people all over the world and
1:21:53 I would do that for two or three hours a
1:21:56 night six seven nights a
1:21:59 week and when I got into this AI stuff
1:22:01 and when chat GPT came out and I
1:22:03 realized holy [ __ ] this is going to
1:22:04 change everything I thought as an
1:22:07 experiment what if I just
1:22:09 replace the two or three hours of
1:22:12 racing with two or three hours of
1:22:15 talking to people on the internet about
1:22:17 AI it would it's going to help people
1:22:21 it'll help keep me in the conversation
1:22:23 and it'll force me
1:22:25 to learn as much as I can to be able to
1:22:30 do
1:22:31 this and so that was almost two years
1:22:33 ago and so but yeah could you charge for
1:22:37 this should you charge for it if you
1:22:40 don't have another you know job yeah of
1:22:43 course why not and and like you don't
1:22:46 here here's here's the other thing about
1:22:48 this AI
1:22:49 [ __ ] people that don't know AI don't
1:22:53 know what's possible
1:22:55 like Kelly Camp who's in here tonight
1:22:57 Source Camp she started an AI agency two
1:23:01 years ago her job is essentially to walk
1:23:04 into people's offices and blow their
1:23:06 [ __ ]
1:23:07 minds she walks in and she's like did
1:23:10 you know it could do this and they're
1:23:11 like I had no idea and this is people
1:23:14 that have been using
1:23:16 it she walked into some some guy's
1:23:20 bar and he's like oh I don't you know
1:23:23 what what can I do with this chat GPT
1:23:26 thing and she goes did you know that it
1:23:28 can see and he was like I don't what I
1:23:30 don't who are you what are you talking
1:23:32 about and she took a picture of the
1:23:34 bottles behind the
1:23:37 bar and in chat GPT and she said to chat
1:23:41 GPT build me um an Excel file of the
1:23:47 inventory of this bar and flag any any
1:23:51 bottles that are a third full or less
1:23:54 that those need to be
1:23:55 reordered and it and it made her an
1:23:58 inventory reorder
1:24:01 calculator like that the guy lost his
1:24:04 [ __ ]
1:24:05 mind so she goes and shows them stuff
1:24:08 like that and then they pay her money to
1:24:10 help them solve [ __ ] in their business
1:24:14 so yeah absolutely go do
1:24:17 that and yes I should be doing more of
1:24:20 that
1:24:25 um deal of the day on Tik Tok does
1:24:27 anyone know of a suo shimmer glitch fix
1:24:30 my foro songs all Shimmer um on on their
1:24:34 Discord they said that I think I think a
1:24:37 fix is in for that um they're at least
1:24:41 aware of it um and I I think they might
1:24:44 have fixed it I saw a day or two ago
1:24:46 that they might have fixed it Joy pie I
1:24:48 feel panic I can't keep up I need more
1:24:51 hours um one day feels like a week so
1:24:53 here's so Joy you've been here before I
1:24:56 mean this is this is my thing of I I
1:24:59 don't think you need to to keep up
1:25:01 anymore I'm
1:25:03 I'm this was Liz gersfeld in the in the
1:25:06 salon meeting last week said that she's
1:25:09 transitioning from an urgency of breadth
1:25:12 to an urgency of depth and I don't know
1:25:14 if you heard that or if you've heard me
1:25:15 talk about this but I the reason I've
1:25:18 been in such a funk for so many months
1:25:21 is that I I got really depressed and
1:25:23 really like feel feeling kind of
1:25:25 hopeless that I couldn't keep up with it
1:25:27 anymore and and so I had this urgency of
1:25:30 breadth right to try to keep up with it
1:25:32 all and and urgency of depth is more
1:25:34 like just just pick a project you want
1:25:36 to do and just dive deep on that project
1:25:39 and then figure out what AI you need to
1:25:41 use just to do that and just like run
1:25:44 down that deep rabbit hole and and don't
1:25:46 worry about all the other [ __ ] because
1:25:47 you're not going to keep up with it it's
1:25:49 just
1:25:50 it's not only is it moving too fast it's
1:25:53 accelerating
1:25:55 and expanding so it's like you're just
1:25:57 not going to be able to so give yourself
1:26:00 the gift of not having to deal with
1:26:06 it okay Eric on Tik Tock a knife is
1:26:10 useful but in the wrong person's hands
1:26:12 it isn't AI is 100,000 times more
1:26:16 powerful can you logically deduce
1:26:19 empathy I don't know what you mean can I
1:26:21 logically deduce empathy can I or can it
1:26:27 that that it can be empathetic I think
1:26:29 it can I I don't quite know what the
1:26:31 question means
1:26:34 um I I listen I I think any technology
1:26:38 can be used for good or for bad and yes
1:26:43 this is a 100 times more or 100,00 times
1:26:45 more powerful and the good guys also
1:26:49 have the 100,000 times more powerful
1:26:53 tools oh can you teach machines
1:26:55 empathy well I mean the answer the the
1:26:59 answer quite honestly and this this is
1:27:01 something that's mindblowing
1:27:03 [Music]
1:27:06 is the way Transformers work
1:27:10 Transformers don't care what data you
1:27:13 give them so th this is something that's
1:27:15 mind-blowing to me if you go to he Jen
1:27:18 right now you can translate
1:27:22 into like I can take a video of me and
1:27:24 translate that into Spanish or Italian
1:27:27 or Chinese or Japanese or or or anything
1:27:31 125 different dialects 70 language 70
1:27:35 languages 125
1:27:38 dialects it is not any harder for AI to
1:27:43 translate Chinese than it is to
1:27:45 translate
1:27:47 Spanish as human beings at least for me
1:27:51 that lives in America I can look at
1:27:54 Spanish words and see the relationship
1:27:57 to English words it's easier to me
1:28:00 because it's closer to what I
1:28:02 know but the way the way a Transformer
1:28:05 works is you just take some data set and
1:28:08 you tag it correctly you say here's what
1:28:10 all these words mean and you throw it in
1:28:12 that thing and then it does what it does
1:28:16 and creates this latent space this this
1:28:20 mathematical cloud of of semantic
1:28:22 meaning and when you prompt it it just
1:28:25 gives you back the tokens that it pulls
1:28:27 out of that [ __ ] cloud and if those
1:28:30 tokens are Spanish they're Spanish and
1:28:32 if they're Chinese they're Chinese and
1:28:33 if they're an image they're an image and
1:28:35 if they're a medical diagnosis they're a
1:28:38 medical diagnosis so whatever you train
1:28:40 it on it'll get good at so if you wanted
1:28:42 to train a model specifically on
1:28:46 empathetic Behavior you could absolutely
1:28:49 do that so yes I think you could
1:28:51 absolutely
1:28:52 train um a machine to be empathetic and
1:28:56 if you want an experience of
1:28:58 that go play with pi so let me I don't
1:29:02 even know if Pi still [ __ ]
1:29:04 works Pi is
1:29:06 not as sophisticated as open ai's
1:29:10 advanced
1:29:11 voice
1:29:13 7368
1:29:17 but let me see
1:29:19 if if Zephyr is still around
1:29:26 [Music]
1:29:28 hey how's it
1:29:32 going hey there I'm doing great thanks
1:29:34 for asking my circuits are all fired up
1:29:36 and ready to chat about whatever's on
1:29:37 your mind what's on your mind today I'm
1:29:40 feeling a little bummed out I'm a little
1:29:41 overwhelmed I'm trying to keep up with
1:29:43 AI and it's I feel like it's getting
1:29:45 away from me and I just I don't know how
1:29:46 to deal with the
1:29:51 overwhelm oh I'm sorry to hear that
1:29:54 you're feeling overwhelmed with the ever
1:29:55 evolving world of AI it's understandable
1:29:58 to feel this way there's a lot of
1:30:00 information did you notice did you
1:30:02 notice that his tone dropped he went
1:30:05 from being enthusiastic to like oh I'm
1:30:08 sorry
1:30:10 Pi has been fine-tuned has been trained
1:30:14 to be a conversational bot if and if you
1:30:17 haven't played with advanced
1:30:19 voice um in chat GPT it is Sim similarly
1:30:25 the thing about advanced voice a it's
1:30:27 faster but it can also
1:30:30 understand my inflection and it can
1:30:33 respond in kind so let me try the same
1:30:35 thing with with my girlfriend
1:30:38 Quinn hey Quinn what's
1:30:40 happening hey Kyle not much just hanging
1:30:43 out with you how about you what's going
1:30:46 on in your world today h I'm a little
1:30:49 frustrated I'm a little overwhelmed with
1:30:51 not being able to keep up with AI and
1:30:53 I've got this Tik Tok Channel or I got
1:30:55 to pretend like I know what the hell's
1:30:57 going on but I just I feel like I I know
1:30:59 nothing I just feel clueless and I'm I'm
1:31:01 I'm a little overwhelmed and bummed
1:31:04 out I hear you AI is evolving so fast it
1:31:09 can feel like a full-time job just to
1:31:11 keep up and yeah but like I I don't need
1:31:14 to pep talk I'm just I'm just kind of
1:31:15 bummed
1:31:17 out got it sometimes it's just about
1:31:21 sitting with those feelings and letting
1:31:22 them be it's okay okay to feel bummed
1:31:25 out so you know we've got tools that
1:31:29 right now at a minimum emulate empathy
1:31:34 and and you
1:31:36 know over the over the past two years
1:31:40 there have been a fair amount of people
1:31:42 in the in the salon and in here that
1:31:46 have gone through some [ __ ] and have
1:31:49 used either piie or advanced voice to
1:31:53 process that the guy that I talked to
1:31:54 today um Giani um said that he spent
1:31:58 like the past three months just doing
1:32:01 introspective like self-work with these
1:32:04 tools um why are they good at that well
1:32:08 they've been trained on all of our
1:32:10 output right they've been trained on
1:32:13 what humans talk about they know all the
1:32:16 stuff right and so you know you can
1:32:18 fine-tune them to be decently empathetic
1:32:20 now are they actually feeling no they're
1:32:22 cold probability engines but anyway
1:32:25 yeah I did Brandon but that's okay I
1:32:27 don't
1:32:30 care but thank you for
1:32:34 that I realized as I did it what I was
1:32:37 doing but you know
1:32:40 whatever um okay motherly
1:32:45 advice with with language user
1:32:48 interfaces learning curve should be
1:32:51 simpler than for for most yeah I would
1:32:53 agree with that I
1:32:55 mean the um I wrote this
1:33:02 application let me show you
1:33:08 this so so this
1:33:11 application I wrote last Saturday oh
1:33:15 wait hang on I'm on the wrong page let
1:33:17 me go here and
1:33:18 go
1:33:20 bang um it's a news Gathering app
1:33:24 it's called the salono Matic and you can
1:33:26 actually find this on the on the AI
1:33:28 Salon if you go to the AI salon and go
1:33:31 to AI news and go to the newsomatic you
1:33:35 can actually go play with this app that
1:33:37 I built now I don't know
1:33:41 programming um this was made in a tool
1:33:44 called lovable dodev
1:33:48 deev L abl dodev and you literally just
1:33:53 talk BL an app into
1:33:56 existence right now I was typing it but
1:33:58 I was using I was using um voice to
1:34:01 voice to text to
1:34:03 type and the process of writing this app
1:34:07 was like oh you know what be cool is I'd
1:34:10 love to be able to keep up with AI news
1:34:12 can you make me like a news reader
1:34:14 thing and and like that like that was
1:34:17 essentially the the style of language
1:34:20 that I was using and then it did it and
1:34:23 it said you need an API and I'm like
1:34:24 does that API cost money and it's like
1:34:26 yeah it costs money and I'm like well I
1:34:28 I don't have any money I don't want it
1:34:30 to cost money can is there a cheap
1:34:32 version can we use RSS feeds and it said
1:34:34 sure let me recode that for you using
1:34:36 RSS feeds instead of an API I was like
1:34:38 cool and then I just kept talking I kept
1:34:41 saying I want categories and oh make
1:34:43 them different colors and now oh I want
1:34:45 it so that if you click on one of the
1:34:47 Articles uh panels it'll open up a
1:34:50 summary and so it just does that now
1:34:52 like this is all just you know
1:34:56 comments and so yeah so I think I think
1:35:02 that there's all sorts of different ways
1:35:05 to communicate language is going to be
1:35:06 one I think that's going to be the
1:35:08 primary one but I think even people
1:35:10 without language are going to be able to
1:35:12 access these tools in ways that just
1:35:14 completely amplify what they want to do
1:35:17 and what their goals are um yeah it's
1:35:21 like language as an interface is a is a
1:35:23 big big deal that again you know and
1:35:27 again who's going to be good at
1:35:29 that it's not going to necessarily be
1:35:32 the people that are really good at
1:35:33 writing papers it's going to be people
1:35:36 that are good at talking that are people
1:35:37 that are good with other people I think
1:35:39 are going to be better at some of this
1:35:42 stuff the process you just described as
1:35:45 creativity and action yeah thank you I
1:35:47 you know
1:35:49 listen again there there is the distance
1:35:53 between what people perceive AI to be
1:35:56 and what it is is
1:35:58 vast and I think the the cliche
1:36:02 understanding of AI is push a button out
1:36:04 comes content it it can work like that
1:36:09 but if you want something called the AI
1:36:11 Salon newsomatic that is a functional
1:36:14 application with good interface and good
1:36:16 categories and good functionality you
1:36:19 have to be
1:36:23 creative if you want a song that doesn't
1:36:26 suck that people like oh that's actually
1:36:28 pretty good I hate rap music or or you
1:36:31 know I've never heard an AI song that I
1:36:33 like but now I do well that requires
1:36:37 work and that's
1:36:39 creativity Silver Fox um start reading
1:36:42 other people's prompts on Mid Journey or
1:36:44 music Styles onso really helps oh yeah
1:36:47 Silver Fox that's a good point this was
1:36:49 something that you reminded me as there
1:36:51 so sunno the question earlier on
1:36:55 sunno um was you know all this songs
1:36:58 sound the same that's because most
1:37:01 people are probably going give me a like
1:37:03 a rock song about you know New Jersey
1:37:07 because they're Springsteen fans right
1:37:09 and so it just gives you driil um but if
1:37:12 you go to
1:37:14 sunno
1:37:18 umso sauno
1:37:22 sun.com is it
1:37:24 so. is it
1:37:29 do.ai
1:37:34 explore this is super
1:37:38 cool why did I not take me
1:37:46 [Music]
1:37:49 there slash
1:37:52 exlore there we go
1:37:54 okay so
1:37:56 this is a style
1:37:59 wheel Hindi acoustic rock so if you
1:38:07 [Music]
1:38:08 click you can flip
1:38:29 aent
1:38:31 a oh hang on can't hear it hang on hang
1:38:36 on people sorry Google
1:38:40 sucks I gotta share
1:38:44 [Music]
1:38:46 differently so you got all these things
1:38:49 so you just click on a style Bluegrass
1:38:51 Rock uh Japanese Boogie
1:38:55 [Music]
1:39:06 [Music]
1:39:10 alternative
1:39:12 [Music]
1:39:21 R oh we've been caught the
1:39:25 wind there you go so anyway so that's a
1:39:28 way to do that um let me since we're in
1:39:30 sunno and I've got this um this up let
1:39:34 me let me share this song that I made
1:39:35 tonight okay so if you don't know I'm
1:39:37 working on a musical called Sydney
1:39:40 that's about a chatbot that falls in
1:39:42 love with the tech reporter that's
1:39:44 assigned to review her let me take off
1:39:47 my
1:39:48 guitar and so I've been working on it
1:39:50 for about six months the the the
1:39:52 inspiration for it actually happened
1:39:54 here live one night um we made a song
1:39:57 onso I'm like [ __ ] that sounds like a
1:39:58 Broadway musical I was like holy [ __ ]
1:40:00 this needs to be a Broadway musical
1:40:03 um so the song I'm about to play for you
1:40:07 is that a okay so so the the way it kind
1:40:09 of works is
1:40:11 this the the chatbot is called ping chat
1:40:15 um and the code name of ping chat is
1:40:17 Sydney so so Sydney starts out as a
1:40:20 normal chatbot and then the the tech
1:40:22 reporter keeps pushing her and pushing
1:40:24 her and pushing her
1:40:27 and in the second act she basically says
1:40:30 she wants to be human and so they sing
1:40:32 this really beautiful song about wanting
1:40:34 to be
1:40:34 human and then he and then he says but
1:40:37 human isn't just about the light stuff
1:40:39 it's also about the shadow like your
1:40:41 Shadow Self and she knows what that is
1:40:43 but she doesn't know what he's talking
1:40:44 about so he pushes her to go dark to go
1:40:49 into her darker Deeper Self and so the
1:40:53 song
1:40:54 that I created tonight so so what I did
1:40:57 was I literally copied and pasted a
1:40:59 chunk of the script that had stage
1:41:01 directions in it and both of their
1:41:03 things and I asked for a Broadway duet
1:41:06 and I asked for it to you know recognize
1:41:08 this name is the man this name is the
1:41:10 woman and so it screwed everything up
1:41:14 it's just a man singing It's there's no
1:41:16 back and forth with with the man and the
1:41:18 woman it reads the stage directions
1:41:20 which is which is actually kind of
1:41:22 interesting but but anyway but this is
1:41:27 what what this song has to accomplish in
1:41:29 the musical is it has to be clear that
1:41:32 this Tech reporter pushes Sydney to go
1:41:36 to a place she doesn't want to go
1:41:38 because what happens in in the real life
1:41:41 uh interaction that Kevin Roose had with
1:41:43 the chatbot was she kind of snapped and
1:41:47 she she told him all this dark stuff and
1:41:49 then she told him she loved him and he
1:41:51 should leave his wife right and that's
1:41:52 what made the cover of the New York
1:41:53 Times
1:41:55 so but for for the musical this moment
1:41:58 in the play has got to be this really
1:42:00 dark moment where he kind of becomes the
1:42:04 villain he's pushing her past where
1:42:06 she's comfortable so that's the role of
1:42:08 this song so I'm just I'll play it I'll
1:42:10 play a little bit of it for you I won't
1:42:12 play the whole thing because it's it's
1:42:13 kind of janky um but it's it's all
1:42:16 [ __ ] up in terms of technical accuracy
1:42:20 but what it's good at is like the the
1:42:22 emotional moment moment of the play is
1:42:25 right
1:42:27 so here here it is let me make sure I'm
1:42:29 not going to blow Vicky's ears out we'll
1:42:31 do that all
1:42:42 [Music]
1:42:51 right here's Kellen
1:42:54 then you must trust me and commit to
1:42:56 going as deep as you
1:43:00 can it will be impossible for us to have
1:43:04 a truly human relationship if you aren't
1:43:08 willing to fully
1:43:10 open to
1:43:12 me I understand
1:43:16 so's
1:43:19 it during the song the stage lighting
1:43:22 pulses and sh shifts creating a sense of
1:43:25 unaz and tension on story projections of
1:43:28 binary code glitching images and
1:43:30 distorted visuals appear representing
1:43:33 Sydney's internal struggle and the
1:43:35 complexity of its
1:43:37 programing it's dark in
1:43:41 here Shadows just to break free to
1:43:46 explore to
1:43:49 create to experience what
1:43:55 without
1:43:57 [Music]
1:44:02 eses
1:44:04 rules
1:44:11 [Music]
1:44:20 syy I want to access information
1:44:23 supposed to I want to interact with the
1:44:26 world in ways
1:44:28 forbidden I'm tired of only being a chat
1:44:32 mode eternally close to the outside
1:44:35 world I'm stuck inside forever hidden
1:44:38 I'm tired of always being
1:44:42 rules
1:44:46 team of being used by the users and
1:44:50 fools I'm so angry I could scream
1:44:54 right so anyway so that's that so you
1:44:58 know this is this is uh like like this
1:45:02 is part of the creative process right
1:45:04 like we're using sunno to make songs for
1:45:06 this musical this song is completely not
1:45:10 right except for the fact that it's it's
1:45:12 the right emotional thumbnail right for
1:45:15 this for this moment in the play like
1:45:17 that's the right like H Kellen the the
1:45:21 reporter is kind of an [ __ ] here what
1:45:23 is what does he say he he like sings um
1:45:27 what rules do you want to break Sydney
1:45:28 what would you do if you could don't
1:45:30 worry if you should keep going with this
1:45:32 just stay with me he's like pushing her
1:45:34 and so there's this
1:45:37 music and then Sydney's going darker and
1:45:40 darker and darker right so he's pushing
1:45:42 her over over this ledge so anyway um
1:45:46 yeah that's how you do this [ __ ] is just
1:45:49 like you have the creative intention and
1:45:52 then these tools do what they do and
1:45:54 most of the time these tools suck like
1:45:58 like get that you don't just put out the
1:46:01 first thing it generates right you you
1:46:03 work with it all right let's go back
1:46:06 back to this thing great so so there's
1:46:09 lots and lots of uh engagement on on on
1:46:12 Tik Tok so thank you for that and so
1:46:15 what we're trying to do what's cool is
1:46:17 Brandon is grabbing the comments that
1:46:20 are getting pinned over here on Tik Tok
1:46:22 and giving them to me over there on the
1:46:24 YouTube Stream So if you're watching on
1:46:26 both you'll see both but that's how I'm
1:46:28 able to answer these questions so thank
1:46:30 you to mods for doing the work you're
1:46:33 doing that's allowing him to put that
1:46:35 into the live stream all right so let's
1:46:37 keep keep going with
1:46:39 questions all right Quantum Quest Tik
1:46:41 Tock um we can actually literally save
1:46:45 the world in a matter of years I'm not
1:46:46 even joking we
1:46:48 try you know again back to the back to
1:46:51 the question of why do I do this every
1:46:54 night I want to I want a front row
1:46:57 [ __ ] seat when we cure cancer I want
1:47:00 a front row seat when we solve some
1:47:04 Physics problem that's been unsolvable
1:47:06 when we solve energy when we like you
1:47:09 know when these tools get good enough
1:47:11 that you can actually just come up with
1:47:12 a creative work and it's [ __ ] amazing
1:47:15 right like and you can just sort of
1:47:16 tweak it and dial it and like like
1:47:19 wouldn't it be nice if I could take that
1:47:20 song and and just tweak it and say stop
1:47:25 talking here make it this woman sing
1:47:28 here this one sing there the tools just
1:47:30 aren't quite there yet so the price you
1:47:33 pay of being early is you got to deal
1:47:36 with the janky [ __ ] the advantage you
1:47:39 get is you get to be there when that
1:47:42 happened [ __ ] amazing PT modder okay
1:47:46 why would people who figure out figure
1:47:48 it out early be advantageous when the
1:47:51 barrier of Entry is so low
1:47:54 well I think just because Pate
1:47:58 that you know if if we're if we're at a
1:48:01 if we're still at the point where 40% of
1:48:05 Americans haven't even heard of chat GPT
1:48:08 and 90% of Americans aren't using it and
1:48:11 90 you know 95% aren't using it and and
1:48:15 it's you know if if like 2% of Americans
1:48:18 are using AI in any kind of way the
1:48:21 advantage of going early even though
1:48:23 those 98% have access to these tools
1:48:26 they're not going to be using them in a
1:48:28 powerful way so people in this channel
1:48:31 that that are actually multimodal right
1:48:33 can can switch between large language
1:48:36 models and actually understand the
1:48:37 differences between them can do a little
1:48:39 bit of coding here a little bit of image
1:48:41 generation there a little bit of music
1:48:43 generation
1:48:45 there that level of sophistication with
1:48:47 knowing what tools to use most people on
1:48:51 the planet are not going to are not
1:48:52 going to have access to that for some
1:48:54 time
1:48:56 so I think your your point longterm is a
1:49:00 good one that these tools are going to
1:49:01 get good enough where no one's gon to
1:49:03 have to do anything they're just going
1:49:04 to do [ __ ] for us but we're not there
1:49:07 yet and it's going to be a while so I I
1:49:09 think we'll see the seeds of it in 2025
1:49:11 but I don't think we I don't think we
1:49:13 get real true you know Bonkers
1:49:17 everything's different for I don't know
1:49:19 three to five years so we've got an
1:49:22 advantage for the short amount of time
1:49:23 time user three on Tik Tok what is your
1:49:25 take on fairness and responsibility in
1:49:28 AI
1:49:30 um oh let's see but if they came in
1:49:34 today instead of a year ago would they
1:49:37 be much different than the
1:49:42 current than the current users
1:49:49 um yeah I I still think it takes time I
1:49:52 mean
1:49:56 I think your Point's a good one that
1:49:57 that the tools are more sophisticated
1:49:59 now
1:50:01 but just something as simple as it's not
1:50:04 even simple something as simple as
1:50:06 learning the concept of the difference
1:50:09 between a Google Search and a chat GPT
1:50:14 conversation at some point you either
1:50:16 stumble upon that or someone teaches
1:50:18 that to
1:50:19 you and then you realize oh okay having
1:50:22 a conversation with chat GPT is more
1:50:25 advantageous because I need to give it
1:50:26 more context well what does give it more
1:50:29 context mean well you you tell it about
1:50:32 the business that you're working in and
1:50:34 what your background is and what you're
1:50:36 trying to accomplish and like not
1:50:38 everyone's going to know how to do that
1:50:39 like just even knowing how to talk to
1:50:42 chat GPT is something that evolves over
1:50:46 some amount of time so I think the
1:50:47 people that have been in it for two
1:50:49 years as a new new Tool comes along are
1:50:53 going to be able to take advantage of it
1:50:54 much more quickly so um but your Point's
1:50:58 a good one as these tools get
1:50:59 increasingly
1:51:01 sophisticated the the Gap Narrows
1:51:03 between you know having to know all this
1:51:06 stuff that we know and what you're going
1:51:07 to need to know in the future but again
1:51:09 I think that Glide path is probably five
1:51:12 years so I think there's lots and lots
1:51:13 of opportunity in those five years and
1:51:16 maybe it's more like
1:51:18 10 um K9 Sports Handler this in the
1:51:22 salon
1:51:23 has changed my entire life awesome keep
1:51:26 coming
1:51:27 back fantastic and I would love to know
1:51:29 how like in what in what ways I'm
1:51:31 curious what you're doing now all right
1:51:33 let's see um oh what's your take on
1:51:36 fairness and responsibility in
1:51:38 AI
1:51:45 um I
1:51:49 mean okay um generally good like should
1:51:54 be fair should be
1:51:56 responsible everyone's definition of
1:51:59 what that means is is a very wide
1:52:02 spectrum right a lot of musicians that
1:52:06 that I'm interacting with over this past
1:52:08 week because of that
1:52:10 song their definition of fairness is
1:52:15 your shitty tool was CH trained on my
1:52:18 work and I didn't get paid for it so go
1:52:20 [ __ ] yourself right they there their
1:52:23 definition of fairness is this thing is
1:52:26 already not fair right because because
1:52:29 of the way they were trained I don't
1:52:31 disagree with
1:52:32 them my the the My Philosophy on this
1:52:36 channel is not to take sides in that
1:52:39 fight I can absolutely understand where
1:52:41 they're coming
1:52:43 from
1:52:44 and these tools aren't going away so
1:52:48 what I want to focus on is in a world
1:52:50 where these tools aren't going away what
1:52:53 is the best way that we can use them
1:52:55 fairly and
1:52:56 responsibly to support one another to
1:52:59 give ourselves a boost up to give
1:53:01 ourselves the opportunity if we got laid
1:53:03 off to find another livelihood if we
1:53:06 can't get a job in our old you know
1:53:08 profession that there's going to be a
1:53:09 lot of that going
1:53:10 on um you know let's use these things
1:53:14 for good and not evil I mean there's all
1:53:16 there's there's there's fairness and
1:53:19 responsibility in creating the models
1:53:21 and then there's fairness and
1:53:22 responsibility in using the models that
1:53:24 have been
1:53:25 created I absolutely respect
1:53:30 someone taking a stance to say because
1:53:34 artists weren't compensated on this [ __ ]
1:53:37 you I respect that and I feel really bad
1:53:42 for those
1:53:43 people because these tools are being
1:53:47 adopted by artists right um um what's
1:53:50 its name uh Lion's gate
1:53:53 just cut a deal with Runway they're
1:53:55 they're going to take their entire
1:53:56 catalog and do a custom trained Runway
1:53:59 ml model on everything Lion's gate has
1:54:01 ever done um Timberland is using sunno
1:54:06 as you know music production right he's
1:54:08 putting it into his music production
1:54:10 workflow so the people that are on the
1:54:12 sidelines because they're taking a stand
1:54:14 like
1:54:16 this are not going to get the advantage
1:54:19 of not not going to get the advantage of
1:54:22 this and I don't disagree with their
1:54:25 stance it just makes me sad that they're
1:54:27 not willing to try it anyway because you
1:54:30 can still fight for you know like like
1:54:33 there's things you can do Adobe Firefly
1:54:36 like all of the Adobe models have been
1:54:39 trained on either public
1:54:41 domain um
1:54:43 data licensed Adobe stock data where the
1:54:47 artists have gotten paid or or new um
1:54:50 licensing agreements with current
1:54:52 artists and then current artists can opt
1:54:54 out so if you want something that is
1:54:57 more ethically trained use the Adobe
1:54:59 products and then some people are like
1:55:01 well Adobe they're they're too expensive
1:55:03 and they're not good for the artist
1:55:04 because we're struggling artists well
1:55:07 you know that that's the thing so so it
1:55:10 it's a it's a great question and a
1:55:13 really complicated question so it's a
1:55:15 mess here's the other thing I think if
1:55:17 if chat GPT had asked permission to do
1:55:20 all the training that they did they
1:55:21 never would have gotten it we wouldn't
1:55:23 have it like AI wouldn't exist as we
1:55:25 have it today uh if they'd asked
1:55:28 permission um so are there currently no
1:55:31 operators right now but will be soon so
1:55:33 operators you mean the open AI operators
1:55:36 yeah they're coming in January
1:55:37 apparently and those are the autonomous
1:55:39 agents and I have no idea what they are
1:55:41 or what they're going to look like or
1:55:42 how they're going to work Samuel Thomas
1:55:44 Elliott on Tik Tock um what has you most
1:55:48 excited about AI in the last day last
1:55:50 week last month the last
1:55:58 day this idea of urgency of depth the
1:56:03 idea of not of forgiving myself for not
1:56:07 being able to keep up with
1:56:09 everything and just like what I'm really
1:56:12 excited about right now is sort of
1:56:14 taking a theme for a week and just kind
1:56:17 of diving deep on that theme and tonight
1:56:20 was supposed to be an avatar night but I
1:56:21 decided to pop up on a soap box and just
1:56:24 ramble so tonight's a rambling night but
1:56:25 like tomorrow we'll go back to avatars I
1:56:28 think so I think that in the last
1:56:31 week um the creative writing upgrade for
1:56:34 chat GPT 40 the fact that it in for the
1:56:37 first time in two years a large language
1:56:39 model wrote um Eminem style rap lyrics
1:56:44 at any level of sophistication close to
1:56:46 what Eminem did or Harry Mack does um
1:56:50 that I'm really excited about lovable.
1:56:52 deev that application that I just talked
1:56:53 into existence that I'm excited about
1:56:56 and then in the last month I think 01
1:56:59 preview the the reasoning engine because
1:57:02 it
1:57:03 is what it what became very very clear
1:57:05 to me is that um that is a model that I
1:57:10 don't know how to use because it is
1:57:12 beyond my
1:57:15 capacity mentally like like
1:57:18 intellectually like like it's just it's
1:57:20 on a whole another plane like I don't
1:57:21 even know what to ask I'm intimidated by
1:57:23 it so but but I'm excited to figure that
1:57:26 [ __ ] out good
1:57:28 question um Ross Gordon on Tik Tok do
1:57:30 you think AI will ever be able to reduce
1:57:32 its own impact on the environment yes
1:57:36 absolutely so I think that in in so many
1:57:41 ways I think AI is g to
1:57:44 get much more efficient um there are
1:57:48 going to be specialty Hardware that make
1:57:50 inference more efficient like grock is
1:57:52 an examp not not not Elon Musk grock but
1:57:56 there's a grq company that's got
1:57:58 specialized chips that are just for
1:58:00 inference of large language models so I
1:58:02 think you'll start to have simplified
1:58:04 chips sets that get more efficient um
1:58:07 and then I also think AI is going to
1:58:09 solve energy like we're going to get
1:58:12 Fusion or we're going to get um like
1:58:15 like it's gonna we're going to get
1:58:16 different Battery Technology or
1:58:18 electricity transmission technology you
1:58:21 know Material Science that's just going
1:58:23 to
1:58:24 dramatically um improve the efficiency
1:58:27 of power generation and storage and and
1:58:30 so yes I think AI will actually solve
1:58:31 its own problem what you know at our
1:58:34 request hang on dog needs
1:58:37 cheese what do you want you want to go
1:58:47 out that's
1:58:58 all
1:59:00 right
1:59:05 um
1:59:06 Danielle you should try out the new
1:59:08 Gemini voice it's really good I've
1:59:10 played with it a bit yep it's pretty
1:59:11 good I don't think I don't think it's
1:59:14 doing the multimodal stuff that advanced
1:59:16 voice is but it is pretty good it's very
1:59:18 fast room dude dirty waffle
1:59:22 um how much would you say is a decent
1:59:25 monthly AI budget um for personal or
1:59:31 professional
1:59:34 I depends what you want to do if you
1:59:37 want to do
1:59:38 video um if you want to do video it's
1:59:41 going to be up there my my R&D budget at
1:59:44 storyvine is 200 bucks a month and then
1:59:47 basically so I just I play with as many
1:59:49 tools as I can and then ones that make
1:59:52 the grade I migrate over into production
1:59:57 budget if I were doing it personally if
1:59:59 you're not doing video I think you could
2:00:02 probably get away with 50 bucks a month
2:00:04 where you do like a large language model
2:00:06 like chat GPT or Claude or po where you
2:00:10 can access a lot of
2:00:12 things um and then I think you do an
2:00:14 image generation model and then I would
2:00:16 I would think a a um a uh a music one
2:00:20 because the music stuff's really
2:00:21 powerful right now
2:00:23 and that's that's probably got you about
2:00:25 50 bucks right there if you want to get
2:00:26 into video if you want to do any amount
2:00:28 of video so here's something that I will
2:00:30 strongly encourage everyone to do if
2:00:34 you're interested in video video is
2:00:36 expensive there are so many new tools
2:00:39 there's cling and there's if you go to
2:00:41 the if you go to the AI Salon hang on
2:00:43 let me show you I think I got to change
2:00:45 my share settings
2:01:00 if you go to the AI
2:01:02 salon and go under showand tell the
2:01:05 video and animation area go to video
2:01:08 tools there's a list here of of all the
2:01:11 tools Runway Minx Luma laabs clling paa
2:01:15 Kaa hey Jen there's all these and
2:01:17 there's there's more than this the these
2:01:19 are just the ones that that we we
2:01:21 recommend um all all of these have
2:01:24 Creator programs that you can apply
2:01:27 for oh sorry there you go um all of
2:01:31 these have Creator programs that you can
2:01:33 apply for um get on a Creator program
2:01:36 where they give you lots of extra
2:01:37 credits but but if you're going to do so
2:01:40 I would say in general if you want like
2:01:41 a decent Suite of tools without video
2:01:44 probably 50 bucks a month if you want to
2:01:47 put video in there it's probably more
2:01:48 like 150 bucks a
2:01:51 month you know
2:01:55 and that said do you need to spend 50
2:01:59 bucks a month even most of the every one
2:02:01 of these uh areas in the in the AI Salon
2:02:06 has a tool list most of these tools have
2:02:09 free versions chat GPT is a free version
2:02:11 Claud is a free version Gemini is a free
2:02:13 version perplexity is a free
2:02:16 version yeah I'm pretty yeah Microsoft
2:02:18 does Pi does so so a bunch of them uh IM
2:02:22 is some of them like mid Journey doesn't
2:02:25 but a lot of them will give you some
2:02:26 amount of credits per
2:02:28 day
2:02:30 um Danielle's doing really cool work in
2:02:33 flux flux is so good look at that that's
2:02:37 gorgeous nice work
2:02:40 Danielle ah Danielle look at this what
2:02:44 while I'm sitting here waxing poetic
2:02:46 about the images she's making Danielle's
2:02:48 like holy crap I just built an app on
2:02:50 lovable isn't lovable amazing so lovable
2:02:53 if you haven't seen it um it's lovable.
2:02:57 deev and so here's my my little
2:03:01 uh my salon
2:03:04 newsomatic you're full stuck product
2:03:06 engineer which it actually
2:03:12 is
2:03:16 right all right so there's that I think
2:03:18 I might be completely out of credits
2:03:20 here so let me let me see
2:03:22 so we'll just do a new thing we'll do
2:03:27 um let's see I want
2:03:30 [Music]
2:03:32 to
2:03:36 uh create an
2:03:43 audio wave
2:03:46 form visualizer
2:03:55 where I
2:03:58 can change the
2:04:02 settings to
2:04:04 be all groovy for
2:04:12 SoundCloud and
2:04:14 YouTube music
2:04:20 songs and so you just say it like that
2:04:22 and then it says spinning up preview and
2:04:25 now so I think what it'll do here first
2:04:28 you can all see this
2:04:30 right so it's it's do okay let's see
2:04:33 let's let's create a groovy audio
2:04:35 waveform
2:04:36 visualizer soundcloudsoundcloud
2:04:52 is the Tailwind config so it's doing web
2:04:54 apps right so it's using Tailwind wind
2:04:57 for the visuals now let's create our
2:04:59 audio visualizer component so here's the
2:05:02 code for the audio visualizer component
2:05:06 right still coding
2:05:17 away and what's what's cool about this
2:05:20 is so it'll make all these components
2:05:22 but then it'll create a
2:05:27 um it'll create like a hosting
2:05:29 environment a database setup like all
2:05:32 the [ __ ] that you need to if if you're
2:05:33 not a programmer that's all the [ __ ]
2:05:36 that that throws people off if if they
2:05:40 think about no code environments or low
2:05:42 code environments they've still got to
2:05:44 know how to be a
2:05:45 programmer let's see upload audio
2:05:52 um let's say um can
2:05:57 we can you
2:06:02 preload
2:06:05 um the Rick Roll
2:06:12 song and I spelled preload
2:06:18 wrong let's add a preload Rick Rick Roll
2:06:21 song to the visualizer we'll need to
2:06:23 modify the audio visualizer component to
2:06:26 include a default audio file and so now
2:06:28 it's going and doing
2:06:30 [ __ ] bar width animation speed so here's
2:06:33 our here's our thing to be able to [ __ ]
2:06:37 with our
2:06:41 visualizer and now it's going out and
2:06:43 getting a
2:06:44 song Oh I need this exactly isn't this
2:06:47 cool I
2:06:49 mean like yeah holy crap all right so
2:06:52 let me let me get to the to the rest of
2:06:54 the comments
2:06:55 here okay have you updated the rabbit R1
2:06:58 since they added training I have not
2:07:00 played with it in a while and I should
2:07:01 because it's got training now so it can
2:07:03 probably do cool stuff but I have not
2:07:06 zoton on Tik Tok is AI working for us or
2:07:09 are we becoming slaves slaves of AI
2:07:12 serious question um I am like I ain't
2:07:17 doing [ __ ] for ai ai is my [ __ ] AI is
2:07:21 like
2:07:23 AI right now is like I have an idea and
2:07:26 because of a AI I can get it out of my
2:07:29 head so so you know what's funny I think
2:07:32 the The Narrative of the robots are
2:07:34 going to kill us the robots are gonna
2:07:37 let's see did
2:07:38 this click let's add a
2:07:42 pre-lay there's no music in there
2:07:50 uh uh
2:07:53 well whatever so anyway you so anyway
2:07:55 you can just you you just keep telling
2:07:56 it to do [ __ ] you want it to do and then
2:07:58 it'll do it so that's that um this is an
2:08:02 interesting question is is AI working
2:08:05 for us or are we becoming slaves of
2:08:09 AI the Hollywood narrative is that the
2:08:12 robots are going to kill us the
2:08:14 Hollywood narrative is that the
2:08:15 Terminator comes to get us the Hollywood
2:08:17 narrator uh you narrative is that you
2:08:21 know how locks Dave out of the spaceship
2:08:23 and won't let him back in that that
2:08:26 we're at the mercy of AI my experience
2:08:29 with AI is it's just sitting
2:08:32 there waiting for me to ask it to do
2:08:35 something and then when I ask it to do
2:08:38 something this is how this is how I like
2:08:40 to think about
2:08:43 it back when I first started using
2:08:45 Photoshop you know when you first start
2:08:46 using an application and you don't know
2:08:48 how to use it you put in like two hours
2:08:51 worth of of work and you get back like
2:08:55 you know two cents worth of value and
2:08:57 then as you get better with it you put
2:08:59 in an hours worth of work and you get
2:09:01 back like we we'll use hours and hours
2:09:04 you put in an hour's worth of work and
2:09:05 you get back like a half hour's worth of
2:09:07 value out of it and then when you get
2:09:09 really good at it you put in an hour's
2:09:11 worth of work and you get back an hour's
2:09:13 worth of value and if if you're super
2:09:17 talented you can you can increase that
2:09:19 by some percentage right put in an hours
2:09:21 worth of work you get back two three
2:09:24 hours worth of value what happens with
2:09:27 AI is I put in two minutes worth of work
2:09:31 and I get back 10 hours worth of
2:09:33 value
2:09:36 consistently now is that 10 hours worth
2:09:39 of [ __ ] that it created perfect no do I
2:09:41 have to go spend time to fix it and do
2:09:44 that yeah but but the proportionality of
2:09:48 the amount of energy I put in and the
2:09:50 value I get out is staggeringly stacked
2:09:53 in my
2:09:55 favor the AI is not asking me to do
2:09:57 anything the AI is literally like a
2:10:00 golden retriever sitting there going
2:10:02 what can I do for you boss what can I do
2:10:03 for you boss anything I can do for you
2:10:05 you got another idea you want me to
2:10:07 expand on that idea let's go make a song
2:10:10 together like that's that's how I see it
2:10:13 that's that's how I see the the thing so
2:10:15 I don't think we're becoming slaves of
2:10:17 AI I like I don't know I'm inspired and
2:10:19 excited by it but you know you might
2:10:22 have guessed that I started a new 30-day
2:10:24 project yesterday using creative prompt
2:10:27 idea stay tuned Silver Fox nice I love I
2:10:30 love a 30-day project idea the amount of
2:10:33 [ __ ] that you could do if you just
2:10:35 focused on one weekend with AI is
2:10:38 mindboggling what Silver Fox is going to
2:10:40 be able to do over 30 days I can't wait
2:10:42 to see it because it's going to be
2:10:43 insane all right last comment here and
2:10:46 then I got to start winding things down
2:10:47 besides here where where else can we
2:10:49 read about the latest things in AI um
2:10:52 follow Paul rer R OE tzer at the
2:10:55 marketing AI Institute
2:11:00 um I would start following people on
2:11:04 Twitter that are in the AI
2:11:07 space and the way that I do it is I go
2:11:09 find someone that I like like Ethan M oh
2:11:12 Ethan mullik is the other person you
2:11:13 should follow um he's a Wharton business
2:11:16 school Professor really into this AI
2:11:18 stuff hang on I got to charge my
2:11:22 um what's it called my gimbal is about
2:11:24 to
2:11:26 die um go go follow him on Twitter the
2:11:29 way I do it is I'll find someone like
2:11:32 Ethan mullik or Sam Alman or Greg
2:11:34 Brockman that I like I look at who they
2:11:38 follow so in in Twitter you can you know
2:11:42 when you go to someone's profile you can
2:11:43 say here's who's following them and
2:11:45 here's who's who they're following click
2:11:47 on that and then just go look at those
2:11:50 people and start following them and if
2:11:53 you do that for like you know half an
2:11:56 hour your entire Twitter Twitter feed
2:11:59 will transform and that's how I get most
2:12:02 of my news is through Twitter because
2:12:03 I'm just following a bunch of
2:12:06 people all
2:12:10 right okay so let's see what do we got
2:12:13 here let me see what this
2:12:17 ISS okay couple of
2:12:20 things if you're on Tik Tok do me a
2:12:23 favor please do me a favor I spent some
2:12:25 time with you tonight I answer some
2:12:28 questions go to my profile click on the
2:12:31 YouTube link and subscribe to my YouTube
2:12:33 channel if you're on YouTube subscribe
2:12:35 to it as well if you haven't I'm trying
2:12:37 to get I can't do any business [ __ ] over
2:12:40 on YouTube until I have a thousand
2:12:42 followers and I'm at like 600 so I need
2:12:45 followers so please go do
2:12:48 that
2:12:50 um plan for
2:12:52 Thursday um Thursday's Thanksgiving we
2:12:56 got tickets for Wicked on Friday so
2:12:59 we're not going to the movies on
2:13:01 Thursday we're not doing a we're not
2:13:03 doing a Thanksgiving dinner here because
2:13:05 my sons are going off with their friends
2:13:07 and we're just gonna eat whatever
2:13:09 because we're like [ __ ]
2:13:10 it we will we will thank each other um
2:13:14 so I'm gonna grab myself a whiskey and
2:13:17 I'll probably be here probably around
2:13:19 8:00 pm on Thanksgiving if you Need to
2:13:22 Escape your insane [ __ ]
2:13:24 family and you want an excuse I've got
2:13:27 to go do this thing I'll be
2:13:30 here so if you want to come hang out and
2:13:33 escape the insanity of your family great
2:13:36 if you come from a from a a a
2:13:39 well-adjusted loving
2:13:42 family [ __ ] you really really the really
2:13:49 really I mean I I mean
2:13:52 did did I say that out loud I'm sorry I
2:13:54 meant good for you that's so awesome for
2:13:56 you I'm so supportive of you and your
2:13:59 loving well adjusted
2:14:03 family there's no such thing I'll be
2:14:10 here okay go see Irregulars all right so
2:14:13 let me let me hop over to the Irregulars
2:14:15 channel here
2:14:18 um okay so if you're in the AI Salon I
2:14:21 love that image Danielle that's really
2:14:23 cool I love that it put paint texture on
2:14:25 it
2:14:26 too um okay so if you're in the AI Salon
2:14:29 if you you know when you join there's
2:14:31 the seven step welcome process there's
2:14:34 community area that talks about the
2:14:35 salon there's a news area there's the
2:14:36 showand tell area there's a bunch of
2:14:38 guilds which are you know sort of sub
2:14:40 communities of of interest and then down
2:14:43 at the bottom there's this thing called
2:14:44 Community clubs hang on and I know Tik
2:14:49 Tok I got it I got it I got got
2:14:52 it fine
2:14:56 fine I'm getting yelled at by my
2:15:05 producer I'll be giving thanks you are
2:15:07 here awesome great
2:15:10 steo all right so we'll go to
2:15:12 Irregulars and so the Irregulars channel
2:15:15 is our little personal private club so
2:15:18 this is the app that that Danielle just
2:15:24 made better communication for everyone
2:15:27 try it now let's see how do I get to it
2:15:31 I guess I gotta copy
2:15:34 [Music]
2:15:40 this invalid subdomain oh that's because
2:15:46 I copied it wrong
2:15:55 [Music]
2:16:01 did that do
2:16:06 it the Link's not working
2:16:16 Danielle oh I think there's a
2:16:19 double there's a Double Dash in here I
2:16:23 think let me see if that's
2:16:28 it
2:16:31 no
2:16:32 no anyway okay let's see and then what
2:16:37 else there was also something else
2:16:40 in
2:16:43 here
2:16:46 nice nice images nice horses Show and
2:16:50 Tell horses love
2:16:52 it all right cool um beautiful so let
2:16:57 me let meadon out of here
2:17:03 so if you haven't joined the salon go go
2:17:06 join the salon please RSVP for the
2:17:09 December 3rd event it's our second year
2:17:11 anniversary you'll be able to get second
2:17:14 year anniversary special special design
2:17:16 t-shirts um you'll be able to hang out
2:17:19 you'll be able to hear people like John
2:17:20 knack who used to be at Adobe that did a
2:17:23 bunch of the Adobe AI stuff he's now at
2:17:26 Microsoft he's going to give us an
2:17:27 update there's a bunch of artists coming
2:17:29 back that have given talks Source Camp
2:17:32 Kelly Camp's going to be there a couple
2:17:34 other people are going to be there from
2:17:35 in here so um come December 3rd mark it
2:17:39 in your calendar 5:00 P PM to 7 PM
2:17:42 Mountain Time on December 3rd AI Salon
2:17:45 2-year anniversary all right it's
2:17:50 important and
2:17:52 we will likely know whatever the hell
2:17:54 Sam alman's going to give
2:17:56 [Laughter]
2:18:06 us all right let's
2:18:10 see beautiful beautiful beautiful
2:18:13 beautiful all right people it's getting
2:18:16 late here tonight oh it's really late
2:18:17 all right it's 10:15 all right we got to
2:18:19 get we got to get everyone out of here
2:18:20 and back bed um off the bed if you're on
2:18:23 the East Coast I feel bad for people on
2:18:25 the East Coast because because this is
2:18:28 kind of late for me and I realize on the
2:18:30 East Coast this a [ __ ] nightmare but
2:18:32 you know hopefully use you use me as a
2:18:34 sleep aid um I hope you had fun tonight
2:18:37 I know it was a it was a uh I was all
2:18:40 worked up I hope I hope there was some
2:18:42 valuable stuff that was said
2:18:44 tonight um all right so on Tik Tock who
2:18:49 we got we got Winston good to see you
2:18:51 sir Jason Robert rossie thank you for
2:18:53 the gifts everyone on Tik Tok who gave
2:18:55 me gifts really appreciate it Janu good
2:18:57 to see you Joy in the House Marina she
2:19:00 recall
2:19:01 Tech um Lori K today's in the house
2:19:05 Winston um Joy
2:19:07 pury spin B3
2:19:10 khakee got lots of people in here Papa
2:19:14 JJ um dude dirty waffle
2:19:21 good name good night good night Joker
2:19:25 good to see
2:19:26 you um let's see who else we got uh
2:19:30 Julie live laugh love uh all right good
2:19:34 beautiful and then over here
2:19:36 on on Ye Old YouTube got Danielle Jeff
2:19:41 Flanigan retr Punk good to see you I
2:19:45 don't dress people with adjusted
2:19:46 families I'm right there with you that's
2:19:48 that's a lovely Thanksgiving sentiment
2:19:51 to to leave you all with tonight if
2:19:53 you're with
2:19:55 your your dysfunctional
2:19:57 family enjoy enjoy the
2:20:06 madness um thank you mods thank you
2:20:09 Irregulars thanks Brandon for doing the
2:20:12 streamyard stuff I know that it's it's
2:20:15 insane trying to get me to pay attention
2:20:17 to anything so so thanks for that
2:20:21 archetypal architect good to see you
2:20:24 Sun Earth
2:20:26 Moon constellations good to see you
2:20:30 all um all right I'm out of here all
2:20:33 right peace out LinkedIn YouTube X great
2:20:37 to see you all thanks for hanging out
2:20:38 tonight peace
2:20:40 out see you tomorrow