Dec 3, 2023 - AI-Driven Creativity: From Music Composition to Digital Avatars

Description

Chapters

Transcript

0:03 [Music]
0:13 got a blue
0:15 telescope looking at the world tonight
0:19 through blue
0:21 telescope I wish I may wish I might not
0:25 say what I
0:26 seen she metal on sheets of
0:30 [Music]
0:33 has got the blue
0:38 telescope down a moon struck a road
0:43 [Music]
0:59 tonight
1:00 good evening good evening sing it champe
1:04 Champ's doing his thing a minor he loves
1:07 a
1:08 [Music]
1:16 [Applause]
1:16 [Music]
1:28 minor cracks me up every time
1:31 it's like the it's like a it's like the
1:33 I like I like turtles kid makes me laugh
1:37 every
1:38 time I want to be pissed at him
1:40 sometimes sometimes I just want to play
1:42 my
1:43 guitar it does doesn't happen in this
1:46 house
1:52 anymore oh Silver Fox I love this so
1:56 much it's good it's good good good uh
2:00 hey steo what's happening forever hooked
2:03 saw your post about the mask good
2:05 question did you ask the artists uh yeah
2:07 I post I tagged the artist in in the
2:11 video I forwarded him I I also wrote an
2:14 article on LinkedIn which didn't get
2:16 picked up today I guess posting articles
2:18 on Sunday is
2:21 stupid I got I got nothing I got no
2:23 action on it I thought it was a really
2:24 good article but um but I I uh I sent
2:28 him the article on Instagram so we'll
2:30 see we'll
2:32 see um yeah it's a good question you
2:35 know I I can talk about what I did I
2:37 mean it was it was just it was just one
2:39 of those things
2:41 where I was kind of marveling at his
2:44 craft right this is this is
2:47 clearly clearly someone who's done you
2:51 know Decades of work to get that good
2:56 and you know like Talent you know and
2:58 then and then work
3:00 ethic on top of that and then just you
3:03 know just you know just remarkable
3:05 craftsmanship I was just kind of
3:07 marveling at that and then I thought oh
3:10 I wonder how quickly I could get to
3:11 Something in the
3:14 neighborhood and I got there really
3:16 quick and so there was a part of me that
3:18 was like oh this is like fan generated
3:20 art fan art and then there was a part of
3:23 me it was like but is
3:25 it and I mean this is the thing that you
3:28 know that Grimes said to her fan base
3:32 you know synthesize me all you want make
3:34 all the songs you want just give me 50%
3:36 and you know other artists are going
3:38 I'll sue your ass and it's like I don't
3:41 I don't know that there's a right answer
3:43 I think I think some artists are going
3:44 to be like cool you know get the stuff
3:47 out there give me some attribution and
3:48 we're cool other artists are going to be
3:50 like [ __ ] off and die so anyway I was
3:53 equally excited and equally dejected for
3:56 the artists well so here's the thing
3:59 though the work that that artist does
4:02 that's not getting replaced with AI I
4:05 mean sure I suppose you could have ai
4:07 generate that thing print it on shrink
4:10 wrap plastic and then wrap a helmet um
4:14 but but you know he's creating works of
4:16 art I you know I I I actually feel that
4:20 with all this AI stuff work like what
4:22 that artist does actually goes up in
4:24 value right because it's handcrafted so
4:27 I think
4:28 authenticity goes up in value I think
4:30 handcrafted goes up in value I think
4:32 live performances go up in
4:35 value and then all the digital stuff I I
4:39 I I don't have a [ __ ] clue all the
4:42 digital stuff like all software all
4:46 software all video
4:48 games all digital art all of it all
4:52 website development is just you're going
4:54 to be able to instantly knock it
4:57 off or be inspired by it
5:00 right what is it
5:05 um wait good artist plagiar eyes great
5:08 artist steel right I think we're g to
5:11 move into a world of great art very soon
5:15 lots of great
5:16 [Laughter]
5:21 art but who was it someone I think it
5:25 was lunatick on the on the AI Salon made
5:29 a really good point about
5:34 um you know where I chose to take it is
5:39 me right where I chose to take it where
5:42 I chose to stop right the first things
5:44 that it
5:45 did it wasn't really anywhere well it
5:49 was in the neighborhood of what he was
5:50 doing but just it didn't look like what
5:52 he had done and I wanted to see if I
5:54 could get it close to what he had done
5:56 and but then where I ended up stopping
5:58 with the sort of glowing eyes and things
5:59 like like that was a really different
6:01 kind of thing I think having taste will
6:04 be very important with the AI Revolution
6:06 I think taste is everything I think
6:07 taste and point of view becomes
6:09 everything right so
6:12 so the the power will be in the curators
6:17 right
6:19 so maybe the skill moving forward is
6:22 curation over
6:24 creation and then and and what does that
6:26 skill look like I don't know I took a
6:29 picture someone had taken and improved
6:31 it with AI and he was
6:33 angry I fed the Beast with his art well
6:36 the Beast has already fed with his art
6:38 or it couldn't do what you did Vicki so
6:42 listen like I said I mean this this is
6:44 why empathy is one of the uh is one of
6:47 the um values of the AI Salon
6:52 because people are going to be angry and
6:55 and especially people that are not in
6:57 this and don't understand how it works
6:59 and just you know have whatever
7:01 preconceived notions they have yeah
7:03 that's going to be like that
7:04 contemplating if I should submit AI art
7:06 to my state fair absolutely someone did
7:09 it here in Colorado a year ago and won
7:13 it I don't oh if I want to open that can
7:15 of worms listen I listen I don't know I
7:18 mean it's it that's up that's up to you
7:20 it you know it that's largely up to your
7:24 comfort zone
7:26 I you know from where I sit
7:31 I I
7:32 don't I don't think it's anyone's
7:34 business quite frankly how I got to what
7:37 I got
7:39 to like do do do do do I need to
7:43 disclose that I used Photoshop and
7:45 Lightroom to edit my photographs I took
7:49 on my Sony digital
7:50 camera do I need to give the model
7:53 number and the lens and the settings I
7:57 put in Lightroom or that I used
7:59 Lightroom room that I used a digital
8:01 image manipulation thing no that's
8:04 ridiculous like everyone's like of
8:06 course not well when when Photoshop
8:10 first came out and digital cameras first
8:12 came
8:13 out that was what everyone said well you
8:16 should declare this a digital photograph
8:19 it's just not a real photograph right
8:21 there was all that [ __ ] so it's just
8:23 it's just the same it's the same
8:25 argument with different
8:26 technology um anyway um welcome
8:29 everybody happy Sunday night this is the
8:32 AI learning
8:33 lab we'll get going here
8:36 um I'm if you if you don't know what the
8:40 [ __ ] I'm talking about I um I I'll show
8:43 you
8:47 here um let me go
8:50 to where can I get to
8:55 that oh I know how uh let's see tick
9:00 Tik tock.com
9:01 [Music]
9:14 that's
9:18 it couldn't find this
9:23 account thought it was CBO airbrush is
9:26 that not right hang on
9:32 AI learning
9:41 lab it's Kyle Shannon with AI learning
9:43 lab I think we're entering an
9:47 a oh jbo not CBO okay all right so so I
9:52 went to I was just I was drinking my
9:55 morning coffee damn how long has this
9:58 show been on
10:00 um not long I I kind of started in I I I
10:05 started a conversation based on a
10:07 comment and didn't set up the
10:09 conversation so it seems like like we're
10:12 deep in the middle of the conversation
10:13 we just went live N9 minutes ago so you
10:16 we're just we're just
10:19 live but I'm explaining uh what the
10:22 conversation was about now okay so so
10:25 this morning I'm drinking my coffee and
10:27 this video pops into my my
10:31 live and so this is um a guy named
10:36 Jordan Boral or Baro I don't know if
10:40 it's baral or Bargo um I assume he is
10:43 the official Calgary Flames airbrush
10:47 artist um and so this is It's a five
10:51 minute you know process video it's one
10:52 of these process videos where you don't
10:54 quite know what he's doing when it
10:57 starts and then oh this is a different
10:59 video hang
11:01 on this is
11:05 here so you you see him masking and he's
11:09 painting the
11:10 [ __ ] and then he starts airbrushing the
11:13 stuff and like he's airbrushing like
11:15 these the like the Little Fibers of the
11:17 yarn he's turning this this pattern into
11:20 a
11:21 crocheted you know a knitted sweater
11:24 right very very
11:27 Canada and so I'm just what watching the
11:30 like normally with these process videos
11:32 I'll get to about you know a third of
11:34 them I'm like I'm good I'll either just
11:36 skip to the end or just move on and this
11:38 one I watched the whole thing cuz like
11:39 look at the look at just his precision
11:43 and like that he that he has figured out
11:46 how to do this in such a I don't know
11:49 like how he's making the letters look
11:51 like they're knitted in
11:53 and and all that sort of stuff I don't
11:55 know why it keeps jumping videos that's
11:58 weird um
12:00 let me
12:07 see I mean look at that God damn and
12:11 like like he's even sanding it like like
12:14 how could you sand that work you just
12:15 did like what if you rub through it
12:17 somewhere but like he's done you know
12:19 this he's clearly done this for decades
12:21 right like he just knows this [ __ ] cold
12:24 yeah it's gorgeous right it's so
12:27 beautiful and so
12:30 what hit me as I'm watching
12:34 this is I will never in my life be this
12:39 masterful at anything right it's one of
12:42 the one of the downsides of ADHD is
12:44 you're you're good at a lot of [ __ ]
12:47 you're masterful at almost nothing right
12:50 and that's just that's just part of the
12:52 part of the deal right so I was kind of
12:55 I was kind of mesmerized by this thing
12:58 and and you could do anything Kyle well
13:00 and that's kind of what happened so so I
13:03 thought well I wonder I wonder how well
13:05 chat GPT could describe what this did so
13:08 I went over to chat
13:10 GPT is this it
13:14 [Music]
13:18 no I went over to chat JP T are we not
13:22 going to switch again what what are we
13:25 doing here
13:27 what ah
13:30 Microsoft Edge is such a giant Pooh hole
13:33 of a
13:34 browser
13:36 okay so I went over to chat
13:39 GPT and I I grabbed you can upload so so
13:43 for those of you that don't know what
13:44 I'm what I'm showing you here is taking
13:48 advantage of two things within chat G
13:51 well three things within chat GPT one is
13:54 its ability to see so you can upload
13:56 images to it and it can describe them
13:58 and then it could also generate images
14:00 and then it's got its large language
14:01 model so you could talk to it so you can
14:04 upload up to four Images so I just went
14:07 and I took some screenshots of of the
14:09 Tik Tok live and then I uploaded those
14:11 so this is this is Jordan's
14:15 work and you know gorgeous go like the H
14:19 it's just it's it's just
14:21 mesmerizing and then I
14:23 said these are images from different
14:25 angles of a professional goalie hockey
14:27 mask please describe them in as much
14:29 detail as possible and so it
14:32 did and there's all the description and
14:35 I'm looking at it I'm I you know I've
14:36 got add so I'm not going to read all
14:38 that
14:40 [ __ ] so then I said create a 16x9 image
14:44 of a goalie wearing this mask although I
14:47 wrote wearing this make it it understood
14:50 though where we see a closeup of his
14:52 face in the mask from an angle we can
14:54 see the detail as well as his intense
14:57 eyes with the blurry background of the
14:58 net rink and Arena and so it generated
15:02 these and these are in the neighborhood
15:05 but they're like missing the Calgary
15:06 Flames you know logo and and color
15:10 scheme and and they're like really ropey
15:13 and big and it just it wasn't quite
15:15 right right so I was like eh okay is
15:17 Dolly the best image
15:23 generator there there there are four
15:26 that I think are the top ones they all
15:28 have some different advantages the big
15:30 advantage of
15:32 do is that you don't actually have to
15:35 learn how to prompt it all of Del image
15:38 generation is done in the context of a
15:40 chat GPT chat you can learn how to
15:43 prompt it but you don't have to and so
15:46 what I'm doing here is I started with it
15:49 looking at the images of his mask
15:52 describe it for me and then I had it use
15:55 chat GPT use that description to to
15:58 create images and so I for that reason
16:01 it's it's dolly is just a fundamentally
16:04 different tool because it's more like
16:06 you're working with an art director than
16:08 it is you're working with an image
16:10 generation tool even though that's
16:11 exactly what you're doing
16:13 okay so then I said um match the colors
16:18 to the original it's a Calgary Flames
16:20 color scheme with Reds and yellows
16:21 accenting the colors you have analyze
16:24 the uploaded image uh for color details
16:27 as well as texture and so if rewrote its
16:29 description and then I said perfect Now
16:32 update the images you created to
16:34 incorporate these new details and it
16:37 gave me these and the the thing that
16:40 really blew me away here was that's the
16:42 Calgary Flames logo and the safety guard
16:47 rails on these things like if you ask it
16:49 to to make art of a living
16:52 artist it it won't do it and and same
16:55 with sometimes if you use trademark
16:57 stuff but it did this so that's
16:58 surprised me a bit
17:01 um you know nice nice Calgary Flames
17:04 here and sort of the a bit of one there
17:06 a little janky one there um but they're
17:09 pretty cool and then I realiz I I
17:11 remembered oh uh the more thing we've
17:14 been talking about in Dolly 3 you can
17:16 say make it more whatever I could I
17:18 could have said make it more crocheted
17:20 or make it more you you know uh whatever
17:23 I said make it more
17:24 intense make make the design and eyes
17:27 more intense and so there's one of the
17:30 more intense versions I was like oh
17:32 that's cool and there's there's another
17:34 one and I and I was like okay make it
17:36 more
17:38 intense
17:40 and now it's got these kind of devil
17:42 eyes and they're kind of you know they
17:44 kind of burn and they match the Flames
17:46 but now now it lost the Calgary Flames
17:49 thing so I was like well let me go one
17:51 level more intense but I said more
17:54 intense with players clashing in the
17:56 distance because I noticed this one this
17:59 one had some players in the background I
18:00 thought oh that would be cool let's get
18:02 some players in the background so I said
18:04 make it more intense with players
18:05 clashing in the distance keep the
18:07 Calgary Flames branding active on the
18:11 goalie uniform and then it came up with
18:13 these
18:14 two and so this has got an NHL logo here
18:20 it's got a little bit of the Flames but
18:22 then there was this one and this one for
18:25 me it's got the the Calgary Flames logo
18:28 on the on the neck protector here but
18:30 it's got these Flames for eyes and it
18:32 created this new kind of unique thing
18:34 and if you zoom in on it it's it's um it
18:38 is crocheted it's like knitted but it's
18:41 it's much smaller it's not that big
18:42 coarse thing and then and then look at
18:45 the hockey players in the background
18:47 there it just it's like it's it's like
18:49 he just turned around quick you know
18:52 while the Puck's coming down the ice to
18:54 look at the camera and it and
18:56 it's it's this crazy easy ass thing and
19:01 so this all happened in the space of
19:04 about 10 minutes while I'm just sort of
19:05 sitting on the couch drinking
19:07 coffee
19:11 and I it put me in in a
19:14 weird it's not a dilemma I
19:17 mean this technology is here it's not
19:20 like this technology is coming this
19:22 technology is here so I can be flipping
19:25 through anything I could be surfing the
19:28 internet could be looking at a magazine
19:31 I could be looking at a billboard
19:32 outside I could be at a fair watching
19:35 someone paint something I could you know
19:38 find a piece of you know a street sign
19:43 inspiring and I can grab an image of
19:46 that thing and pull it into this tool
19:49 and start jamming on
19:52 it and so part of me is like I was so
19:56 inspired by Jordan's work that I was
19:59 like you know this is Fan Art and this
20:02 is really this is a really cool
20:03 extension kind of a collaboration of who
20:05 he is and what he does and there's
20:07 another part of me that's like listen
20:09 you sleazy
20:11 [ __ ] this is this is Talent
20:16 appropriation and and you know I I think
20:19 that you know reality is likely
20:21 somewhere in
20:22 between um and in the end like you know
20:25 I'm not going to I'm not going to go try
20:26 to get a job you know doing hockey masks
20:29 with do 3 like this literally like this
20:33 is something I I would have never done
20:35 before this was more like a thought
20:37 experiment when when I re when I got
20:39 depressed at how talented this guy was
20:41 realizing I could never do that and then
20:44 10 minutes later later to have this
20:47 image which is a clear
20:51 um
20:53 omage to to the work that he did but
20:57 it's it's clearly a different thing
20:58 thing so do gpts lose their short-term
21:02 memory despite the
21:04 documents in the knowledge base do they
21:08 lose their
21:11 short-term well they lose it if you're
21:14 talking about the context of the
21:16 specific conversation you're in if you
21:18 start a new conversation it will lose
21:21 the memory of that conversation it can't
21:24 maintain the thread of conversations
21:26 across conversations
21:29 within the same conversation it it won't
21:32 lose its memory and it won't lose the
21:33 memory of what's in the um documents but
21:36 it will lose the me memory of your
21:38 interaction if you change it to a new to
21:41 a new thing so anyway so that was the
21:43 that was the
21:45 um that was what I did this morning and
21:48 that was the thing that kind of blew my
21:50 mind and then I I wrote an article about
21:52 it for LinkedIn so if you go to
21:55 LinkedIn
21:57 um
21:59 [Music]
22:01 it's
22:04 called Zen and the Art of casual
22:09 Excellence um where where I basically
22:12 just I I put in article form what I just
22:14 showed you all I said you know here's
22:16 the guy here's this remarkable video I
22:18 said please go watch this video so you
22:21 know I'm not trying to pull anything
22:23 over on him um and then I just talked
22:26 about the process I said here are the
22:27 four screen shots I took I uploaded them
22:30 to DOI here's what it did and so it's
22:32 just an article that does that and I
22:34 basically end it you know with this
22:36 thing of like you know you know am am I
22:40 a you know am I a fanboy or or a thief
22:44 or both and is it good or bad it part of
22:49 what I try to do on this channel is not
22:51 make any
22:54 judgments on the technology or how
22:56 people are using it um
22:59 what I'm trying to explore here is
23:01 what's
23:03 possible this is now possible and it's
23:06 not only possible like this is now like
23:09 that [ __ ] image like that image
23:12 that's a stunning image to my eyes can
23:16 Dolly create a user interface design for
23:19 yes it can it can and I can show you
23:22 that in fact one of the things you can
23:24 do with with Dolly from an interface
23:26 design standpoint is 8:22 p.m. thanks
23:30 Siri you're awesome
23:34 um you could upload a whiteboard sketch
23:37 or a paper sketch of an interface um
23:40 have it describe the interface and then
23:41 tell it to design the interface for you
23:43 and it it'll do it now it's not
23:45 actionable it'll be more like a design
23:47 inspiration it's not going to create you
23:49 know a separated out figma design but
23:53 it's going to design something that will
23:55 probably give you some new
23:57 ideas
24:00 that is super cool Silver Fox wow that's
24:09 [Music]
24:16 beautiful wait what's going on
24:21 [Music]
24:27 here
24:30 that's a good idea
24:31 Daisy
24:34 lunatick yeah so yeah if I'm running
24:36 down a rabbit hole um I probably won't
24:39 see stuff that gets pinned so it's
24:41 probably that's that's a good idea all
24:43 right okay so there you have it that was
24:49 [Music]
24:57 that
24:59 all
25:01 right the goalie
25:04 moved can you explain how to best use
25:06 seed numbers for images and do I don't
25:09 really know cam
25:12 um I know you can ask it for a seed
25:15 image but they also call it something
25:16 else the the
25:19 reference I haven't PL I I haven't I've
25:23 played with it a little bit I haven't
25:25 had terrific luck with it I can explain
25:27 what it is so
25:29 basically when you make an image with a
25:33 generative image generation tool like
25:35 mid Journey or stable diffusion or
25:37 Dolly it's generated with what's called
25:40 a seed and it's it's just a unique
25:43 number that is kind that references the
25:46 the the specific amalgamation of you
25:50 know
25:52 pixelized data that gets denoised into
25:55 an image and so if you reuse use that
25:58 seed you tend to get
26:02 um consistent images across
26:05 prompts
26:08 um there's some people have claimed
26:10 they've gotten really good results with
26:11 it I've gotten kind of janky ass results
26:14 with it so I I I haven't explored it
26:16 enough to know oh gen ID number thanks
26:18 Papa JJ yeah so so what I would say what
26:21 I would say cam is is um generate an
26:25 image and then ask chat GPT to give you
26:27 the
26:29 gencore ID hash hashtag
26:33 number so
26:35 Jore ID hashtag or
26:39 hash symbol hashtag
26:42 symbol and then when it gives you that
26:46 the next time you do a prompt say use
26:48 this gen ID and give me whatever the new
26:52 part of the prompt is you seated some
26:56 images for a child's book I saw I saw
26:59 the video yesterday I know I did Silver
27:02 Fox but I just
27:04 it it I couldn't I couldn't get it to
27:08 maintain the style so like the the
27:17 um I think it was the Llama it was it
27:19 was some animal I had that had you know
27:23 patches of color and then the next image
27:25 it didn't have the patches of color so
27:31 I I just I'm not good at it I like I I
27:35 don't want to tell you how to do
27:36 something if I don't actually know how
27:38 to do it right so I don't actually know
27:40 how to do it right there are probably
27:42 some videos on it here's the deal I
27:45 guarantee you that there's someone that
27:46 has geeked the [ __ ] out on this and that
27:50 video probably exists um mine's not the
27:53 one to do on it um I've just named the
27:56 character like the Hedgehog demo and the
27:58 other characters have the same name of
28:00 it yeah that that was that was that's
28:03 another that's a good point Sherry D so
28:05 you can get the seed number the the
28:07 other thing you can do what they did in
28:09 the in the uh when they announced Dolly
28:12 3 was they named um the Hedgehog Larry
28:18 so one of the reasons hedgehogs come up
28:20 as our as our animals here as we were
28:23 talking about Larry the Hedgehog a lot
28:25 so you can name the you can name a
28:27 specific character in the scene and then
28:30 say hey you know what does Larry's house
28:33 look like or I want to see Larry in an
28:35 open field that can help as well can I
28:37 share my open in login with my family
28:41 I'm pretty sure you can um if too many
28:45 people use it you're going to run up
28:47 against rate limits though right so if
28:49 you get someone really excited about it
28:51 and they go knock out 50 images and it
28:54 rate limits for 3 hours and you're at
28:57 work and need to use it you're going to
28:59 be pissed off so um I don't think
29:03 they've got anything in there with
29:04 different sessions where where you can't
29:06 have a different session because they've
29:08 got it rate limited I think they'll just
29:10 shut you down so um Cindy [ __ ] who spoke
29:13 at the AI Salon her her daughter really
29:17 got into it and she ended up having to
29:20 um join a different uh create a
29:23 different subscription so Cindy [ __ ] is
29:26 a 2 GPT for subscription family uh for
29:29 that very reason uh can I join I assume
29:32 you mean that means the AI Salon yeah
29:34 sure hang on let me let me show you what
29:37 the AI Salon is
29:40 so I want my chat I want my chat
29:45 J okay so this is the AI Salon um the AI
29:49 Salon is a um group that I
29:53 started a year ago on Tuesday um so the
29:57 week after chat GPT came out I started
29:59 this group
30:03 and where am I going to go I will go
30:07 here so the URL for the it's it's it's
30:11 hosted on Mighty networks so the URL is
30:14 AIS salon. mn.
30:17 Co and if you go
30:20 there and decid you want to join my my
30:24 first request is go to the about the
30:28 Salon go to the about the salon um tab
30:32 or whatever this is thingy thingy the
30:35 thingy over here um and and read what
30:38 the salon's about um it it's it's really
30:40 short it's kind of you know like who we
30:42 are why we're here how we explore what
30:44 we seek um so super simple and then it
30:48 talks about our values of
30:50 curiosity generosity like be willing to
30:53 share what you're learning is is a big
30:55 one
30:56 exploration collaboration you know I
30:59 want people that are are you know
31:01 excited to engage with others inclusion
31:04 doesn't matter where they're from some
31:06 of the best artists in our community
31:08 right now have no art
31:10 background why because they've
31:11 discovered this tool where they've got a
31:13 voice and they're [ __ ] going nuts on
31:15 it and empathy um empathy is's a big one
31:19 there's a lot of people are scared right
31:22 now A lot of people
31:24 are angry
31:26 and you know things like that so thank
31:29 you very much
31:30 Wiston uh so if you if you resonate with
31:34 those values great you know but please
31:35 go read read about them and then uh once
31:38 you get into the
31:39 community um there's all sorts of cool
31:41 areas there's there's a bunch of you
31:43 know FAQs and how to navigate and Rule
31:45 Rules of Engagement things like that um
31:49 there's a calendar so the the AI Salon
31:52 we meet every other week um so this
31:54 coming Tuesday is our one-year
31:56 anniversary so we're bringing back as
31:58 many speakers from the past year that
32:01 can join us as possible and we're just
32:03 going to basically do a year in review
32:05 from people that have spoken so we had
32:07 John Knack from Adobe has spoken twice
32:09 at the salon talking about Ai and what
32:11 adobe's doing with AI we've had artists
32:14 we've had technologists we've had
32:16 business people you know we've had
32:18 ethicists um so there a lot of them are
32:21 coming back to talk about the year uh in
32:24 in review so so that's super cool so the
32:27 uh the calendars are there for that um
32:30 then there's a community section um the
32:33 community the the first link in the
32:35 community section is the Welcome Wagon
32:37 so if you join please introduce yourself
32:40 in the value of generosity the First Act
32:43 of generosity you can do in the
32:44 community is tell us who you are tell us
32:47 who you are what you're up to where you
32:48 are on this AI Journey um there's a
32:52 water cooler you know just a chat where
32:54 you can just talk about anything there's
32:55 a showand tell channel so one of the
32:57 things that happens in here a lot of the
32:59 people that are here regularly will as
33:01 I'm sort of talking about stuff will oh
33:04 yeah the the new KISS thing is pretty
33:06 cool what those guys just did I'll talk
33:08 about that in a bit um but you can you
33:10 can show images you can share your gpts
33:12 you can tag them um there's a news
33:14 channel there's a there's a resources
33:16 Channel there's an ask ask for help
33:17 Channel and then the thing that's really
33:20 exciting is there's these guilds and so
33:22 the guilds are sub communities within
33:24 the salon where you can sort of geek out
33:26 on a specific thing so there's a writing
33:28 Guild and an Art Guild and a beginners
33:30 Guild and a technologist Guild and all
33:33 that sort of happy horse [ __ ] so so
33:35 that's what the salon's all about so if
33:36 you want to join that go join that the
33:39 other thing I would say is my my request
33:42 here how you can support this channel is
33:45 share the live share this live out right
33:49 now
33:50 um and uh
33:53 follow follow so cuz you know I put up
33:56 videos occasionally I do these lives
33:58 nightly so if you want to do that um if
34:01 you want to subscribe to the lives that
34:03 supports the channel you don't have to
34:04 but if you do awesome um and there's
34:08 also in the lower leftand corner there's
34:10 a series I created a series called
34:12 underhyped where I talk about why I
34:14 think
34:17 um AI is underhyped why I think it's
34:20 significant um how it works how to how
34:23 to think about prompting how to not get
34:26 overwhelmed um
34:28 things like that so so that series will
34:29 also support support the channel um I
34:33 think that's about it oh yeah you can
34:34 tap the screen too feel free to tap the
34:36 screen just let people know that you're
34:38 having a good time here if you're having
34:39 a good time here if you're having a
34:40 shitty time here don't do [ __ ] just sit
34:43 back and
34:44 troll oh if you're going to troll at
34:46 least be creative we we have smart
34:49 people in here so if you do dumb dumb
34:50 trolls we'll make fun of you if you if
34:52 give a good troll the best trolls are
34:54 the ones that seem like a compliment on
34:56 the surface
34:59 and he like hey wait a
35:04 minute you look like Rosie o' donnal
35:08 [Laughter]
35:14 [Music]
35:15 ohuh all right join us we have llamas
35:17 and cookies exactly human being
35:22 exactly oh yeah um and then the other
35:25 thing we do here is if you have
35:26 questions about AI so there's mods in
35:28 here that are are are looking through
35:31 the the crazy amount of comments that
35:32 are coming through they're pinning
35:34 things so that I can see them and focus
35:36 on geeking out and things like that so
35:38 if you have questions ask them um I'll
35:40 also scroll through here and just
35:42 randomly answer things as well um ASU so
35:46 I can finally troll have at it have at
35:48 it ASU absolutely you're an irregular at
35:51 this point Irregulars have a different
35:54 um a different standard like Irregulars
35:57 can pick on each other pick on me
35:59 there's a whole there's a whole
36:01 engagement language that happens with
36:03 the Irregulars so the Irregulars are
36:05 people that show up here night after
36:07 night I do these things seven nights a
36:09 week two to three hours a night because
36:12 I make good life
36:14 decisions Mr K what's happening I've
36:17 been having issues uploading documents
36:19 I'm realizing there are limits to number
36:22 and size yeah I think it's 100 megabytes
36:25 size and then I think it's 20 but I I I
36:29 just remember seeing that somewhere so
36:32 it might not be that I've been having
36:35 like on the lives here I've been having
36:38 first of all I'm having my voice seems
36:39 to be
36:41 going I hear it's a little crunchy like
36:45 I gotta smoke his voice I'm like hey how
36:47 you doing Sally you're looking good
36:49 tonight you know what I'm saying oh
36:51 you're such a good Dame Sally I like
36:56 you
36:58 so that's how it goes sometimes um
37:02 thanks I wrote a city council Watcher
37:04 GPT and it still needs some better
37:06 instructions I
37:08 guess yeah so what I'm finding with the
37:12 gpts
37:14 is you can get them like like all of
37:16 this AI stuff you can get
37:20 them to some version of useful really
37:24 quickly and then if if you want to get
37:26 them really d in it's much
37:28 harder Kyle is one person please be
37:31 mindful about requesting info over over
37:33 and over um thank you you need warm
37:36 drinks I should go get a tea you know
37:38 what maybe I'll do
37:41 that all right yeah let me go let me go
37:45 uh that's a good idea I'm going to go
37:48 put on the water and then I'll be back
37:51 so just talk about yourselves I'll be
37:56 back
38:03 hey my
38:26 uh
38:35 all right I think there is
38:39 thr I think there is
38:56 thr
39:24 yep water's
39:25 on
39:34 all
39:36 right he needs to create an AI adverti
39:39 to entertain us when he takes breaks
39:42 from prompts C listen listen you can
39:44 make
39:45 money this is all the entertainment you
39:48 need this is all the entertainment you
39:51 need
39:53 um yeah uh I did do an avatar did you
39:56 see see my avatar the avatars are pretty
39:59 insane the avatars are pretty [ __ ]
40:03 nuts
40:06 [Music]
40:09 um here I will show you some if you have
40:12 not seen the avatars from hey Jen you'd
40:14 be like hey C that's crazy C I didn't
40:18 know they could do that can they do that
40:20 hey
40:22 c a learning
40:25 lab
40:28 so after my Calgary Flames
40:31 video are these two videos where I look
40:35 like I'm a little too close to the
40:36 camera and I'm looking super intense so
40:40 the first one is um so so what this is
40:44 hey Jen has this Avatar
40:46 feature no champy commercial break we
40:49 can do a champy commercial break we'll
40:51 see if he wants to sing sometimes he
40:52 wants to sing sometimes he doesn't let's
40:55 see champy
40:56 [Music]
41:25 oh
41:27 yes yeah that's a good singing
41:30 dog um have a few intermission videos
41:33 you could play I could do that that's
41:34 not a bad idea um anyway okay so ha Jen
41:38 has this this thing called The Avatar
41:40 maker so what you do is you upload like
41:44 a 3 minute video and it synthesizes the
41:47 whole video and um and then you can just
41:51 type text and it will generate video or
41:54 you can record audio and it will will
41:57 replace the video so the look and feel
41:58 of it is whatever you upload and then so
42:01 this one is just me typing text hey it's
42:04 Kyle Shannon with the AI learning lab
42:06 you know there are some days I just
42:08 don't want to turn on the camera days
42:10 where I don't even want to get out of
42:11 bed well good news we now live in a post
42:15 sci-fi world and I can just type up this
42:17 script and select my avatar of choice
42:20 this is one from a video I recorded
42:22 about a year ago you might be
42:24 wondering Kyle what is this magic
42:28 wand sorry Sherry D Sherry D wrote now
42:31 my dog is barking headphones headphones
42:35 on sorry sorry about that jrc love
42:38 champy champ champ is what this channel
42:40 is really about autotune champy
42:45 exactly um if you go to Instagram uh the
42:48 the Instagram account champ sings um you
42:52 can see lots of videos of me singing
42:53 poorly and him singing lovely
42:58 um okay so that video was just me typing
43:01 text and it synthesizing The Voice this
43:03 next one is me recording an MP3 just an
43:07 audio file and uploading it and it
43:09 generated this video hey it's Kyle
43:12 Shannon with the AI learning lab so the
43:15 previous video in my timeline is an
43:17 avatar generated with he Jen where I
43:20 didn't say the words I all the hand
43:22 movements that was from my video and it
43:25 somehow
43:28 to whatever I'm saying which is bizar
43:31 I'm just recording AUD my cat looks
43:34 worried it's matching this audio to the
43:38 meta okay meta's AI Chief doesn't think
43:41 Theon is okay meta's AI Chief doesn't
43:44 think AI super intelligence is coming
43:46 anytime soon meta's AI Chief has
43:49 consistently um put down open AI
43:54 achievements like time after After Time
43:57 After Time he usually makes a statement
44:00 like that shortly after it becomes clear
44:02 that that they've done something so I
44:06 would take his comments with a grain of
44:07 salt but Hey listen it's possible is ha
44:09 Jen free or is it credit based it's
44:11 credit based um there is not to my
44:16 knowledge um one of these things
44:19 commercially that is that is decently
44:21 cheap the the video generation tools are
44:24 pretty expensive and these Avatar tools
44:26 are pretty expensive um but this this is
44:31 the the ease with which you can do this
44:34 I mean it's literally upload a video and
44:37 then just type text and it will
44:39 synthesize it I mean I could go make a
44:41 new one right now if we if if we wanted
44:44 to um so yeah so that's that's that's
44:48 those things
44:51 crazy um the ones on YouTube don't work
44:55 instructions are never are clear getting
44:57 errors when trying to
45:00 download the ones on YouTube for what
45:04 for what are
45:07 unclear what was the Instagram channel
45:10 so I've got well I've got three but I
45:13 I've got champ
45:14 sings is the one where champ sings
45:18 That's the name and then there's Kyle
45:21 Shannon dreams which is a u stable
45:24 diffusion dream Booth kind of self
45:27 portrait um character art project I did
45:31 that was the thing that got me into
45:35 AI uh there's a gap between AI talk and
45:39 businesses actually using it
45:41 absolutely and as there should be
45:44 like so couple of things about where we
45:47 are with AI right now for one for one
45:49 thing is AI is not new ai's been around
45:52 for decades big businesses have been
45:54 using ai ai is in all the software we
45:57 use so ai's been around what's different
46:00 and what this channel is about is the
46:02 inflection point that started a year ago
46:05 Thursday November 30th 2022 chat GPT
46:09 launches chat
46:11 GPT does for AI and machine learning
46:15 what the worldwide web did for the
46:17 internet back in the mid 90s it made it
46:20 it's a simple interface on top of a
46:22 really powerful underlying technology
46:24 and so
46:27 no one right now knows what to do with
46:30 that how to how to use it how to use it
46:33 in
46:34 business all of these tools one of the
46:36 things that we talk about you know that
46:38 I joke about in here is all of these
46:40 tools right now all of them 100% of them
46:44 janky pieces of
46:45 [ __ ] they they are janky pieces of [ __ ]
46:49 it doesn't mean you shouldn't be
46:51 building things on them it doesn't mean
46:52 you shouldn't trying to be doing
46:55 business with them
46:57 but what it means is your expectations
47:00 should be low your expectations should
47:02 be let's assume that these things are
47:05 just horrible useless things that no
47:08 business can ever
47:10 use and then they will surprise you and
47:13 and and you'll you'll find things that
47:15 they can do well Corps can't stop
47:17 talking about art artificial
47:18 intelligence tools but few businesses
47:21 report using them there's a there was an
47:23 Ethan mullik study not too long ago that
47:26 that
47:28 um
47:31 the forget what it was it was a uh was
47:35 it Bost was it Boston Consulting Group I
47:37 I don't remember what it
47:38 was um but it basically said yeah
47:41 current businesses are not using them
47:43 and but almost all of them from the SE
47:47 Suite to the director level said we've
47:49 got to be up in our game what's your
47:51 third Instagram just Kyle Shannon um uh
47:55 Kyle Shannon
47:56 so there's Kyle Shannon which is my
47:58 photography
48:00 Instagram I also have Kyle shannon.com
48:02 which is my photography site and then
48:05 Kyle Shannon dreams is the art project
48:07 and champ sings is champy and they're
48:10 most they're they're mostly dormant I
48:12 don't really use Instagram but if you
48:14 want to go see some
48:15 archived some archived works you can do
48:19 that I haven't taken a photograph in
48:22 [ __ ]
48:25 ages
48:27 oh good thanks baby all right I have hot
48:31 beverages hot beverages have been
48:34 delivered thank
48:41 you all right it's definitely not all
48:45 the software oh it's definitely not in
48:47 all the
48:48 software what's definitely I don't I
48:51 don't understand that
48:53 comment oh oh oh AI yeah
48:56 well it's some version of AI right like
49:01 AI is in Siri and AI is in Alexa and AI
49:05 is in you know
49:08 um
49:10 Spotify and Netflix's streaming
49:12 recommendations like there's there's
49:14 lots of AI in what we
49:17 do the generative AI stuff the stuff
49:20 that we get to play with the stuff that
49:21 we get to touch that's that's the new
49:23 stuff right the the democratization of
49:26 access to these tools is the thing that
49:29 um that chat jpt has has really shifted
49:33 and the the core technology beneath chat
49:36 GPT the Transformer that Google invented
49:39 I will give Google credit for that they
49:43 invented the core technology that all of
49:45 the stuff that we're playing with right
49:46 now is built
49:47 on um and and it makes all of this stuff
49:52 possible whether it's music generation
49:54 or or
49:56 um uh word generation or images or
50:01 videos all this stuff is made possible
50:04 from from the same basic
50:07 architecture
50:09 um the other thing that we're starting
50:11 to see now
50:14 is the fruits of what happens
50:19 when a piece of software like chat
50:23 CHT gets to 100 million users and 6
50:26 weeks so so when chat GPT launched they
50:30 didn't think anyone would really give a
50:32 [ __ ] about it and uh people
50:36 did so they got to a million users in
50:39 five days they got to a 100 million
50:41 users in six weeks it was the fastest
50:43 adoption of technology in
50:47 history why well because all of a sudden
50:51 the thing that all of us thought you had
50:53 to have a mathematics degree to be able
50:56 to use you could just type in words and
50:59 it would spit out this remarkable
51:02 [ __ ] and and uh so it was all word of
51:07 mouth and and the impact of that was
51:10 that all tech companies all investors
51:14 all technologist all Engineers all open-
51:17 Source projects all of them
51:20 went and turned their focus to AI
51:23 remember we were into metaverse for a
51:25 while we're talk about metaverse then we
51:27 were talking about web 3 and nfts and
51:29 Defi and like we were all over the place
51:31 we were trying to we were trying to find
51:33 it and then chat GPT comes out and and
51:37 the Global Tech Community has
51:41 gone and focused on AI so um we're we
51:46 are right now starting to see the fruits
51:48 of that labor all of that investment so
51:51 the reason you're starting to see like
51:53 these video tools getting better and and
51:55 better and better is they've just been
51:58 hammering away out it and there's they
52:00 keep raising $50 million here $100
52:03 million there is anyone using co-pilot
52:05 I'm not using co-pilot because we don't
52:07 have a an Enterprise license for uh for
52:11 365 so and I and I haven't played with
52:14 the the co-pilot Studio because so
52:18 co-pilot if you don't know is
52:19 Microsoft's version of chat GPT it's
52:22 being Incorporated in everything
52:24 Microsoft which is why the firing of Sam
52:27 Altman
52:29 was one of the reasons it was so
52:32 dramatic is you've got you know one of
52:34 the biggest companies in the world has
52:36 bet their entire company or you know a
52:39 good chunk of their future on this one
52:42 company's technology and so when that
52:44 happened uh you know it got spooky
52:47 anyway co-pilot is their version of chat
52:50 GPT and then just like chat GPT has gpts
52:54 co-pilot has co-pilots and there's a
52:56 thing called co-pilot Studio where you
52:59 can make these sharable customizable
53:01 versions of your co-pilot um I haven't
53:05 played with it I don't have access to it
53:06 so anyway if anyone's played with it let
53:08 me
53:09 know
53:15 um co-pilot plus power bi equals monster
53:21 data analysis yeah I would think so
53:27 just yeah exactly um I don't know
53:29 powerbi power business intelligence but
53:32 I don't know powerbi but um yeah I would
53:35 think
53:37 so co-pilot's going to allow you to
53:41 easily integrate your data from across
53:45 the Microsoft ecosystem into these
53:48 custom uh co-pilots so yeah it's going
53:52 to be
53:53 amazing all right
53:56 all teachers and let's
54:02 see us compels Saudi to
54:05 fund Saudi fund to exit Alman backed AI
54:09 chip startup oh interesting that doesn't
54:13 surprise me they that got some bad press
54:16 quickly so that Sam mman was raising
54:20 money in Saudi Arabia for a chip company
54:23 but you know what what that says to me
54:24 is open AI is realizing they can't they
54:27 can't be beholden
54:29 to Nvidia and they can't be beholden to
54:32 Microsoft they they've got to have
54:34 they've got to be able to stand
54:36 alone um which was really interesting
54:39 that that when that thing was about to
54:41 blow up that it open AI was just going
54:44 to become Microsoft like it was just
54:46 going to become the entire employee base
54:49 was going to go over there um I've asked
54:51 a bunch I have a bottle of Blandon I'd
54:54 like to send a Kyle is a thank you holy
54:57 [ __ ]
55:01 deal I've got I've got that little
55:04 amount of blandon's just sitting there
55:05 and it's all dried out and it's all
55:07 evaporated and it's completely
55:09 undrinkable at this point um Blandon is
55:12 my jam and this shit's impossible to get
55:15 so if you've got one I want to send it I
55:18 will [ __ ] take it um just DM me you
55:22 can DM me here or my um my uh my contact
55:26 info is in my um is in the AI learning
55:30 lab in my profile I've got a link tree
55:32 there um but awesome thank you so much
55:36 I'll take it I actually if I had a
55:38 little blands in my hot tea right now it
55:40 probably help the help the uh vocal
55:42 cords chill out a little
55:45 bit but [ __ ] far out um yes um I do
55:51 accept uh I do accept support in the
55:54 form of Blandon
55:55 [Laughter]
55:58 I if you've never had
56:01 blantons it's it's awful you should
56:03 never buy it I'm trying to I'm trying
56:05 I'm trying to lower the uh lower the uh
56:10 demand um I will D DM you his address uh
56:14 but his real name isn't Kyle and he
56:16 lives in Kentucky
56:18 exactly that's good
56:22 Joker Joker is going to get my Blandon
56:25 bastard
56:29 you don't have to do that but hey man if
56:31 you're offering I'll take it and I'll
56:34 tell you what when it arrives I will
56:36 crack
56:37 it live on the on the Tik Tok the day it
56:43 arrives oh man that's funny uh did you
56:45 see CNN the whole story special on AI
56:48 tonight really well done I did not see
56:50 it actually no cool we'll work for
56:52 blantons absolutely okay there's there's
56:55 an Irregulars there's an Irregulars if
56:58 you want to make some images will work
56:59 for Blandon is a good one here's what
57:01 the bottle looks
57:04 like it's got a little horse on the top
57:07 this one is
57:11 the uh the O so see next to the Little
57:17 Foot there see the
57:19 O so there's um these cork tops are all
57:24 different um
57:26 different stages of the you know the
57:29 famous film of the horse running the the
57:31 first film of a horse
57:33 running they're all different stages of
57:35 a horse running and they all spell out
57:37 the word
57:39 Blandon um let me go look at uh show and
57:42 tell see if there's anything new and new
57:44 and fun in there so one of the things
57:47 that that uh people have been doing on
57:49 here some of the Irregulars is as I talk
57:52 about [ __ ] and I'll say something stupid
57:53 like will work for blantons
57:55 they run off to dolly or mid Journey or
57:57 wherever and they make funny images in
58:00 response to that and so um so um so then
58:07 they post them in Show and Tell on the
58:09 salon which is which is really fun um
58:12 there's there's nothing new in there
58:13 right now oh wait Don just oh no there
58:16 we
58:18 go very
58:21 cool this is such a such an amazing
58:23 space it really is
58:28 all
58:40 right Vicky's going to bed she's going
58:43 to bed she goes I'll be listening but no
58:45 glasses so I won't be able to pin any
58:46 comments sweet dreams
58:49 Vicky that's really funny I've collected
58:52 them all oh that's awesome yeah I've got
58:54 I've got a like a Ziploc baggie full of
58:57 them all I think I'm missing one right
58:58 now I forget which one I'm missing but
59:01 uh but I think what I might do is I
59:02 might three 3D print myself um the
59:05 holder for them
59:08 all that's awesome um very very cool
59:13 Clint Copperhead here at work Macy's we
59:16 have co-pilot on Microsoft Edge oh yeah
59:18 yeah tricky are you liking it yeah
59:21 anyone anyone who's uh who's using uh
59:24 co-pilot let me know what you're
59:27 thinking of it cuz I'm curious if it if
59:31 it is as good as it demoed CU it demoed
59:34 pretty damn
59:37 well uh can you explain what IBM's
59:41 Watson AI is um not in great detail but
59:45 but Watson AI
59:48 so so Watson AI as far as I understand
59:52 it is is is essentially an expert system
59:56 so it's it's an AI That's trained on a
59:58 very specific data set so Watson Health
1:00:02 was about taking Health Data and making
1:00:05 you know doing sort of remarkable stuff
1:00:08 with with AI in that particular thing so
1:00:10 I think I think the Watson product was a
1:00:13 methodology and a set of Technologies to
1:00:16 be able to take your data train Watson
1:00:19 on it and make it an expert system for
1:00:21 your
1:00:23 data Watson health failed I don't know
1:00:26 why it failed um but but you know
1:00:30 possibly because it didn't work right it
1:00:34 it could have been equality issue I
1:00:35 don't like I don't know um but it kind
1:00:39 of flamed out and and went away and then
1:00:41 in this past year one of
1:00:44 the every single tech company every
1:00:47 single tech company after chat GPT did
1:00:50 what it did came out of the woodwork and
1:00:52 said we got AI too IB M was no exception
1:00:56 so IBM came out and said hey we've got
1:00:58 we've had AI for years we've got Watson
1:01:01 AI which is their new revamped I assume
1:01:04 Transformer based you know generalized
1:01:08 model of of Watson but I but I honestly
1:01:13 don't know um it is it is an attempt at
1:01:17 a resurrection of that brand which is
1:01:20 it's not a bad thing to
1:01:22 do
1:01:23 but it like it didn't go down like it
1:01:28 didn't go away elegantly
1:01:30 so the brand has been damaged right you
1:01:33 know it's not quite the equivalent of
1:01:36 Microsoft bringing back
1:01:40 clippy but you know it's in the
1:01:46 hood oh
1:01:48 man powerbi is a Microsoft app used for
1:01:51 data visualization and dashboards oh
1:01:53 yeah I bet I bet bet then that that's
1:01:55 absolutely insane
1:01:58 Edson yeah where you can yeah where you
1:02:00 can just just talk to this tool and it's
1:02:02 calling sophisticated visualization and
1:02:05 dashboards and and assuming it's prog
1:02:08 you know it's it's handling all the
1:02:10 complexity of what that tool does just
1:02:12 with the large language model I bet it's
1:02:14 absolutely
1:02:15 insane when I looked it just said to
1:02:18 call I don't know that's a reference to
1:02:19 prompt add two or three Greek animal
1:02:22 names and tell Del to draw it as one
1:02:24 animal oh that's interesting that's I
1:02:27 bet that's fun yes walking blue walking
1:02:31 blue what's the drink called I don't
1:02:34 know what drink oh this drink throat
1:02:41 coat with a little bit of honey it's
1:02:43 actually
1:02:47 helping
1:02:49 ah ah very good
1:02:52 Luke Luke listen
1:02:56 I need you to run down to the
1:02:57 7-Eleven I need you to pick me up a pack
1:03:00 of smokes all right Luke Luke Luke I am
1:03:04 your
1:03:09 father now when now we know what to get
1:03:11 for Christmas I'll [ __ ] take it if I
1:03:14 if I see a a porch full of square boxes
1:03:18 be
1:03:23 like
1:03:29 uh can AI chapot be
1:03:32 convicted of an illegal wire
1:03:36 tap will work for
1:03:39 Blandon
1:03:48 um it's going to be really interesting
1:03:51 Thelonius to see where the
1:03:53 law where the law gets challenged with
1:03:59 so so
1:04:02 where where things like chat GPT are
1:04:04 quickly going is to a more autonomous
1:04:08 place which which basically means like
1:04:10 right now you have to prompt chat GPT
1:04:12 and you have to tell it to do [ __ ] and
1:04:14 it can only kind of do [ __ ] within its
1:04:16 own little
1:04:17 universe but with these gpts and and
1:04:20 with their API you can connect the power
1:04:23 of chat GPT to other services so you
1:04:27 could say to chat gbt in the not too
1:04:30 distant future and probably write right
1:04:33 now with with a clever programmer or two
1:04:37 hey um I want to know what Kevin Vasquez
1:04:42 is up to help me figure out how to tap
1:04:45 on into that now you you you won't be
1:04:48 able to do that with chat GPT but you'll
1:04:49 be able to do that with open source
1:04:51 stuff the the question is if those
1:04:54 autonomous
1:04:55 agents decide on their
1:04:58 own that they need to figure out how to
1:05:01 tap someone's phone line and then tap it
1:05:03 and get
1:05:04 that maybe all you did was you said hey
1:05:08 I want to get paid back for this
1:05:11 thing and the autonomous agent decided
1:05:14 on its own to tap someone's phone who's
1:05:17 liable for that that's really
1:05:19 interesting uh what is blandon's whiskey
1:05:21 yes blandon's it's BL NT wait b a n t o
1:05:26 n s and it is uh Kentucky bourbon yeah
1:05:32 and it's super smooth and it's just like
1:05:34 there's just not a lot like it it's just
1:05:36 it's it's it's amazing so what are your
1:05:39 thoughts on kiss going digital avatars
1:05:41 would you go to a concert I'd absolutely
1:05:43 go to a concert
1:05:45 um so there's so they've been doing this
1:05:47 for a while there's there's a Korean is
1:05:50 a Korean or a Japanese group that is
1:05:53 it's all um Holograms
1:05:57 basically and they've got the stage
1:05:59 technology decently developed that that
1:06:02 um it looks it looks pretty realistic
1:06:04 from what I've heard I like I've seen
1:06:06 video of it but you never know
1:06:10 um I think it's I think it's [ __ ]
1:06:13 brilliant I mean how those guys have got
1:06:15 to be I'm [ __ ]
1:06:17 58 kiss was Kiss was touring when I was
1:06:20 a kid so they're in their 70s
1:06:26 like you know it's got to get [ __ ]
1:06:29 old man putting on the makeup
1:06:33 and playing the same old [ __ ] tired
1:06:38 songs so yeah why
1:06:41 not why
1:06:44 not you know an actor you know I think
1:06:47 Bruce Willis has got some sort of brain
1:06:49 disease where he might not be able to
1:06:50 act
1:06:53 anymore
1:06:55 if he can be in new films why not why
1:06:59 not um it's weird now
1:07:05 because we we haven't seen we've only
1:07:07 seen a lot of what's a lot of The
1:07:09 Surreal [ __ ] that's going on right now
1:07:11 the thing that I started to chat with
1:07:13 with this you know I watched this video
1:07:15 of this guy doing this thing that I
1:07:16 couldn't do in a million years and then
1:07:19 10 minutes later I've generated
1:07:20 something that's [ __ ] in the
1:07:23 neighborhood
1:07:25 and we've never been able to do that
1:07:28 before and that's going to that's going
1:07:29 to happen in every skill set like like
1:07:33 someone who's just an administrative
1:07:36 assistant is going to go you know I
1:07:37 always wanted to create a really
1:07:39 sophisticated financial analysis app and
1:07:42 then they're just going to do
1:07:44 it with with no computer background and
1:07:49 and and and someone like that is going
1:07:50 to come up with something in a new way
1:07:54 that's never been done before that's
1:07:56 going to be novel and we're going to see
1:07:57 that we're going to see that kind of
1:07:58 innovation happening everywhere and so I
1:08:01 think the like the you know like the
1:08:05 holographic projections of avatars on
1:08:08 stage why
1:08:10 not because quite frankly not only not
1:08:14 only will they have the the images of
1:08:16 the kiss guys probably at different ages
1:08:18 right because they can it was industrial
1:08:20 Light and Magic that did the avatars so
1:08:23 I would be surprised if they didn't do
1:08:26 here's 70s Kiss ' 80s kiss '90s kiss
1:08:30 2000s
1:08:32 kiss the the grand
1:08:34 FW farewell you know 2023 tour kiss so
1:08:40 you can probably dial in the age you've
1:08:42 got all their music you've got however
1:08:44 many live performances you got you can
1:08:45 train an AI on all of
1:08:48 that and
1:08:50 then you could have kiss be writing new
1:08:53 songs
1:08:54 into the
1:08:57 future so maybe they roll out new songs
1:09:00 maybe some of them are
1:09:03 hits why
1:09:06 not they put in the [ __ ] work to you
1:09:10 know tore and turn that thing into a
1:09:13 [ __ ] Juggernaut why not leverage that
1:09:16 why
1:09:17 not they can just sit in the sit in the
1:09:20 old age
1:09:21 home smoking joints and [ __ ]
1:09:24 collecting
1:09:26 checks I'm all for it yeah the Rolling
1:09:31 Stones
1:09:34 exactly Kevin Vasquez they had an ABBA
1:09:37 concert that's right with holographic
1:09:39 avatars and had a wait and had a had
1:09:45 quite a big audience
1:09:47 yeah if it's a good show it's a good
1:09:50 show and these things will be good shows
1:09:53 you know why
1:09:54 why because we're all we're all
1:10:00 wearing
1:10:02 um things that measure our emotional
1:10:05 response so it's not going to be that
1:10:07 long that these things if they're not
1:10:12 already are measuring in real time how
1:10:15 the audience is digging it and they'll
1:10:18 change up the set dynamically to play
1:10:20 the audience like a [ __ ] fiddle go to
1:10:23 show tell
1:10:29 [Music]
1:10:37 okay oh yeah
1:10:40 baby kiss GPT oh that's
1:10:45 awesome that's amazing look it got the
1:10:47 label right it got the text right here
1:10:53 look
1:10:55 there's the
1:10:57 logo there's the bottle see the little
1:11:02 scallops look at that that's
1:11:05 amazing it made it a round bottle but it
1:11:07 got it got most of it
1:11:10 right that's so cool clippy freaking
1:11:14 clippy back from the
1:11:19 dead we will'll work for
1:11:23 Blandon
1:11:27 that's
1:11:34 awesome oh yeah that's what I'm talking
1:11:39 about that's a good one I miss New York
1:11:42 so that would totally be me in New
1:11:44 [Laughter]
1:11:49 York you are [ __ ]
1:11:53 awesome
1:11:55 oh it would be sad if it weren't
1:11:58 [Laughter]
1:12:03 true this is good he's homeless with
1:12:05 Jack
1:12:06 [Laughter]
1:12:10 Daniels you guys are
1:12:12 great oh man that's that's really good
1:12:16 you see you see you see what what kind
1:12:19 of world we're living in I can make a
1:12:22 stupid joke and 5 minutes later there's
1:12:26 like media created about
1:12:30 that like take that that little
1:12:34 interaction that we just had so over a
1:12:36 15minute
1:12:37 period someone offered to send me
1:12:39 Blandon I said [ __ ] far out someone
1:12:42 made the joke will work for Blandon I
1:12:45 laughed at this that said that's a good
1:12:46 one boom there's those
1:12:51 images that's increasingly going to be
1:12:54 our
1:12:55 experience with work product and with
1:12:58 media where we will will it to be and it
1:13:03 will be and it's going to get faster and
1:13:06 faster meme Heaven it's totally meme
1:13:10 heaven and it's sophisticated meme
1:13:13 Heaven you know it's not like the Llama
1:13:16 the llama on the street with the bottles
1:13:18 of plants and the cardboard sign like to
1:13:21 photoshop that you you had to spend like
1:13:23 a good 45 minutes this is like I want a
1:13:26 llama that got matted hair cardboard
1:13:30 sign will work for plantant Jack Daniels
1:13:33 bottles lay laying around it's good it's
1:13:36 really
1:13:40 good oh man it's crazy I use Joker's
1:13:45 GPT pepper ghost are there any decent
1:13:49 Music Creation AI apps okay so so I'm
1:13:52 gonna do two things here
1:13:54 um
1:13:56 Randy r h a n d y
1:14:00 Randy
1:14:02 Randy um I'm going to do two things here
1:14:06 I'm going to give you
1:14:07 all a series of screenshots to do the
1:14:11 Irregulars know what's coming here I'm
1:14:13 going to go pretty quick through this
1:14:14 and then I'm going to show you some of
1:14:15 the music generation
1:14:18 tools because there's some there's some
1:14:20 good ones okay um so that's chat GPT if
1:14:24 you've not played with chat GPT go there
1:14:27 right now but Kyle I'm not very
1:14:30 technical go there but Kyle I don't need
1:14:33 to go
1:14:34 there I'll get to it tomorrow just go
1:14:36 there right now do it do it do it go
1:14:40 there now if you have not tried it do it
1:14:44 but why it's just like fancy Google no
1:14:48 it's not it will melt your [ __ ] face
1:14:52 go there and play until you've had your
1:14:54 Kevin mallister
1:14:56 moment what's that
1:14:59 Kyle remember Home
1:15:02 Alone that
1:15:08 moment I don't know that I've met a
1:15:10 person that when they've really tried AI
1:15:13 hasn't had that moment something it will
1:15:16 do will melt your face and if you
1:15:19 haven't had that moment yet you owe it
1:15:21 to yourself to do it
1:15:26 okay so get ready to
1:15:28 screenshot so chat gbt is the biggie
1:15:32 right now no one's even close to it I
1:15:34 mean there's a couple of people that are
1:15:35 close but not
1:15:36 really
1:15:38 um these four are the big four grock's
1:15:41 not out yet so grock's coming but it'll
1:15:44 be a big one just because Elon will will
1:15:46 it into existence and the the thing that
1:15:50 I think a lot of people sometimes forget
1:15:52 is every single one of Elon companies
1:15:54 are AI companies SpaceX is an AI company
1:15:59 um the solar panel companies and AI like
1:16:02 they're all AI companies and so you know
1:16:06 the this is going to get better really
1:16:08 fast and then you got co-pilot was which
1:16:10 is Microsoft's Claude which is a big
1:16:12 competitor to chat gbt and you got Bard
1:16:14 which is Google bard is going to get
1:16:17 really good in the coming months um they
1:16:21 were supposed to launch a thing called
1:16:22 Gemini next week as a surprise launch
1:16:25 their CEO just killed it until January
1:16:28 or sometime um but they're really good
1:16:31 so these are the big four so screenshot
1:16:34 that screenshot this Pi perplexity and
1:16:37 po are kind of the next three down um
1:16:41 Pi's really good at conversational stuff
1:16:43 perplexity is really good at research
1:16:45 and po lets you play with lots of
1:16:46 different models and it's got something
1:16:48 very similar to goots which you can
1:16:51 develop for free and you can monetize
1:16:54 and you can add knowledge to so a direct
1:16:56 competitor to gpts fun fact Adam
1:17:00 D'Angelo the CEO of quar is on the board
1:17:04 of open AI he was one of the ones that
1:17:07 fired Sam Alman who's got a competitive
1:17:10 product but Kyle that seems like a
1:17:13 conflict of interest doesn't it g
1:17:18 yes I do not
1:17:20 know what that Dynamic is but there's a
1:17:23 Dynamic there okay that's that so here's
1:17:25 the music ones oh no this is the image
1:17:27 ones here's the image ones Dolly 3 mid
1:17:30 Journey Leonardo and firly I said before
1:17:32 there's kind of four flavors of these
1:17:34 image generation tools that all do
1:17:36 different
1:17:37 things Dolly 3 the real power of it is
1:17:40 you're generating images in the context
1:17:42 of a conversation in a in a chat GPT
1:17:45 conversation so you can treat it like
1:17:47 mid Journey or stable diffusion or
1:17:49 Firefly but you can also just be talking
1:17:52 about marketing
1:17:54 um tactics and strategies and just say
1:17:57 oh by the way could you just whip up an
1:17:59 image for that and when when you say
1:18:02 that it will know what that meant
1:18:04 because you're in the context of a
1:18:05 conversation it prompts for
1:18:08 you um Mid Journey if you're an artist
1:18:11 or if you want high-end high high high
1:18:14 quality art is the best Leonardo is
1:18:17 built on top of stable diffusion which
1:18:20 is like the cockpit of a 747 if you
1:18:23 really want to geek out on [ __ ] get into
1:18:25 the world of stable diffusion Leonardo's
1:18:27 kind of your your gateway drug there
1:18:29 it's a it's a web front end commercial
1:18:31 product on top of stable diffusion and
1:18:34 then Firefly from Adobe if you work in a
1:18:36 company it has not been trained on any
1:18:39 copywritten material so so that's what
1:18:42 that one's good for and it's getting
1:18:43 better all the time not great because of
1:18:45 its training set but getting better all
1:18:47 the time and uh generative fill within
1:18:50 Photoshop is actually pretty remarkable
1:18:53 okay video
1:18:55 creation Runway ml Moon Valley and P
1:18:58 labs are in Runway ml just released
1:19:01 motion BR brush which is face melting
1:19:04 interesting good P Labs has announced a
1:19:08 thing called P 1.o which looks like it's
1:19:11 going to kick the [ __ ] out of Runway ml
1:19:12 all of these things are getting really
1:19:14 good really fast um Visa will put
1:19:17 together sort of stock video corporate
1:19:19 videos really slickly um but it's it's
1:19:23 kind of you know just sort of assembling
1:19:24 stock video together but it does it well
1:19:27 and fast and then ha Jen and did are
1:19:31 Avatar makers and ha Jen is the one that
1:19:33 I showed where it's a synthesis of me
1:19:35 it's a deep fake of me um Without Really
1:19:38 Trying and then Runway TV is like a net
1:19:43 Netflix competitor three years from now
1:19:47 but it but it's three years early you're
1:19:49 three years early Runway TV right now
1:19:51 looks like a joke people are going to
1:19:54 make fun of it networks are going to
1:19:55 make fun of it Netflix is going to go
1:19:57 well that's not that's not real
1:20:02 entertainment and then three years from
1:20:04 now Runway TV or something like it is
1:20:06 going to obliterate one of the big
1:20:07 companies just watch watch watch I'm
1:20:10 telling you might not be that one but
1:20:11 it's going to be one like it where
1:20:13 everyone dismisses it dismisses it and
1:20:15 all of a sudden that's what everyone's
1:20:18 watching and who are the new star who
1:20:20 are the new Spielberg of generative
1:20:22 movie be making I don't think they're
1:20:24 coming from Hollywood they
1:20:27 might but not with Hollywood's current
1:20:29 attitude toward this
1:20:32 [ __ ] going to be some 17-year-old right
1:20:35 now that [ __ ] Geeks out on that stuff
1:20:38 and [ __ ] [ __ ] takes down a
1:20:42 studio all right music so you've got 11
1:20:47 Labs which is not music but that's voice
1:20:49 so uh Hey Jen's voice synthesis is 11
1:20:52 labs technology it's really good sunno
1:20:55 and riff Fusion are both similar kinds
1:20:57 of things sunno right now is is my music
1:21:00 maker of choice thank you for the tap
1:21:03 Taps and then sound raw is like a loop m
1:21:08 music composer kind of like Garage Band
1:21:10 where you you know do Loop composition
1:21:12 except it does it it uses AI to assemble
1:21:14 the loops really really fast with with
1:21:18 lots of control so I'm going to show you
1:21:20 those two I'm going to show you sunno
1:21:21 and sound draw and then what's I think
1:21:24 there's some 3D stuff so screenshot that
1:21:26 that's 3D
1:21:30 stuff and then here's people you should
1:21:33 follow if if if you're going to take my
1:21:36 advice and get curious about AI start
1:21:39 following these four people and there's
1:21:41 there's more one of the tricks you can
1:21:43 do is figure out who these people follow
1:21:45 and follow them it's a it's a good hack
1:21:48 for finding smart
1:21:52 people
1:21:54 all right so let's go back
1:21:57 to music and
1:22:00 audio music and
1:22:09 audio I thought I had sunno up and
1:22:11 running oh that's what this
1:22:17 is all right so the way this
1:22:22 works
1:22:23 this [ __ ]
1:22:25 staggering the way this works is you
1:22:28 say
1:22:30 um I
1:22:36 want oh come
1:22:38 on I
1:22:43 want an
1:22:47 Americana
1:22:48 singer
1:22:52 songwriter
1:22:53 meant
1:22:57 about being out of work and willing to
1:23:03 work for
1:23:05 blandon's
1:23:13 bourbon oh wait I ran out of wait so I'm
1:23:16 just going to go Americana singer
1:23:19 songwriter
1:23:21 Lament
1:23:23 and willing to
1:23:25 work for
1:23:28 blandon's
1:23:30 bourbon okay so um that's
1:23:36 it and so what it's going to do now is
1:23:39 it's going to write me lyrics first
1:23:40 thing it'll do is write lyrics and and I
1:23:43 assume they're using gp4 because they're
1:23:45 not bad and when it comes to things like
1:23:48 rhyming and and lyrics and stuff like
1:23:49 that gp4 just kind of Beats other [ __ ]
1:23:52 I've been playing playing my guitar
1:23:53 singing my songs but I find myself out
1:23:55 on the street where I don't belong times
1:23:58 are tough the gigs have dried up down on
1:24:01 my luck feeling like I'm stuck but I got
1:24:04 a little plan a dream I can chase I'll
1:24:07 work for blandon's bourbon I'll sing for
1:24:09 my
1:24:10 supper I'll strum my guitar make the
1:24:13 crowd
1:24:14 holler give me a shot I'll make it worth
1:24:17 your while all right so once it writes
1:24:19 the lyrics now it's going off and
1:24:20 writing two songs
1:24:23 with the chords the music the
1:24:26 instruments the
1:24:28 singer and then it mixes it all so and
1:24:32 it also makes an image representative of
1:24:35 of it I they're they're they're using
1:24:37 some crappy this must be Dolly too
1:24:38 they're using the images are not great
1:24:41 but there's a there's a dude singing for
1:24:44 his
1:24:45 Blandon and then here's the little play
1:24:47 button so here's the two versions of the
1:24:50 song so we'll see we'll see if they're
1:24:53 any
1:24:58 good
1:25:00 [Music]
1:25:03 gu
1:25:05 I
1:25:06 [Music]
1:25:09 Street
1:25:11 time get dried
1:25:15 up I'm down on my luck feeling like I'm
1:25:19 stuck but I got a little plan a dream I
1:25:21 can chase out
1:25:23 I bur I sing for my suer I strong my
1:25:31 guitar make the crowd
1:25:34 harder give me sh I'll make it work you
1:25:40 want just give me a chance and I'll make
1:25:44 you
1:25:46 smile I'll work for BL bir pour me a
1:25:51 glass
1:25:54 sing out
1:25:58 [Laughter]
1:26:00 good that's actually not bad that sounds
1:26:04 like a Blake Shelton song I'll work for
1:26:06 BL BL B I'll sing for my supper right
1:26:12 not bad all right let's hear the other
1:26:15 [Music]
1:26:18 one this good this is
1:26:21 sad
1:26:23 out on the street where I don't
1:26:29 belong times
1:26:33 are the gates have all dried up I'm down
1:26:38 on my Randy just said wow doing my Kevin
1:26:42 mallister face I'm telling you man it's
1:26:45 like night after night here it it just I
1:26:50 I experience it night after night like
1:26:52 this this thing still blows me away
1:26:58 because six not even 6 months ago four
1:27:01 months ago you could only make sunno
1:27:04 songs on
1:27:05 Discord and they sucked they were just
1:27:08 they they they it sort of was words they
1:27:10 were sort of okay they was sort of like
1:27:14 these are um uh Kevin who's in here is
1:27:18 is a professional music producer he's
1:27:20 using sunno right now as a writing part
1:27:23 partner this isn't my style but oh my
1:27:26 [ __ ] god tell me your style lunatick
1:27:28 I'll we'll we'll do one we we'll do one
1:27:31 uh I'll do the blandon's Bourbon One in
1:27:32 a different style tell me your style and
1:27:34 and we'll we'll make another
1:27:37 one that sounds like the bside of
1:27:39 Champy's Greatest Hits oh yeah I gotta
1:27:43 get hey champy champy come
1:27:49 here come
1:27:51 here you want to hear a song you want to
1:27:53 sing do you want to sing here
1:27:58 [Music]
1:28:02 listen you want to
1:28:09 [Music]
1:28:12 sing no no
1:28:17 rapping
1:28:21 let's
1:28:23 no you don't want to
1:28:26 [Music]
1:28:30 sing n he's not in the
1:28:33 mood oh epit Gerald okay
1:28:38 okay
1:28:41 um all right so we're going to
1:28:45 do
1:28:48 um Ella
1:28:51 fits
1:28:57 Gerald
1:29:00 style
1:29:08 um style
1:29:14 um yeah we'll just say that all right so
1:29:16 the same thing about being out of work
1:29:18 and willing to work for blandon's
1:29:21 bourbon
1:29:29 it's got to be an A Minor yeah I
1:29:31 know that's so sad what that champy
1:29:34 wouldn't sing I know listen man he's
1:29:36 he's like he's like he sings when he
1:29:38 needs to sing although it might just be
1:29:40 that didn't have enough guitar in it cuz
1:29:42 I
1:29:43 bet well that's not an A Minor here I
1:29:45 need a capo
1:29:47 [Music]
1:29:51 old
1:29:55 [Music]
1:30:07 you're so easy champy one little a minor
1:30:10 and he
1:30:12 goes the growing number of misleading
1:30:15 ads using AI to create fake celebrity
1:30:17 endorsements that is we're going to
1:30:19 start thonus like that is coming so
1:30:24 fast or so hard it's like yeah you see a
1:30:27 lot of them on Tik to talk right now
1:30:29 we're going to see more in more of them
1:30:30 all right here's the here's the elep
1:30:32 Fitzgerald one um I've been knocking on
1:30:36 doors but they won't let me in got my
1:30:38 resume in hand but they won't give me a
1:30:41 chance been down on my luck feeling so
1:30:44 low got a proposition something to show
1:30:48 see I'll work for blandon's bourbon pour
1:30:50 me a glass all right let's let's hear
1:30:52 how it
1:30:53 [Music]
1:31:02 is
1:31:04 mil but won't give me a
1:31:10 chance huh that one sounds like crap but
1:31:14 [Music]
1:31:21 interesting but they won't give me a
1:31:24 chance down on my love feel it alone but
1:31:29 I've got a proposition little something
1:31:33 to
1:31:34 show I'll swe those Flo I'll scub those
1:31:39 wall just give me a shot I'll do it
1:31:44 [Music]
1:31:50 all that's my boys in the
1:31:53 that's
1:31:54 [Music]
1:32:01 [Applause]
1:32:03 champy so anyway so that's sunno so
1:32:07 crazy right just crazy and then um sound
1:32:13 draw sound and so sound draw is an
1:32:18 interesting they this is whoever is in
1:32:21 their naming uh committee needs to be
1:32:24 [ __ ] fired it's sound
1:32:27 raw not sound draw which I thought it
1:32:32 was for for
1:32:33 months and but but they they use all
1:32:36 caps so you can't tell so so I
1:32:40 capitalized this because I'm considerate
1:32:43 of human beings
1:32:44 needs
1:32:46 um sounds like it belongs to the AI Mary
1:32:49 Tyler Mo show coming soon that's pretty
1:32:51 funny funny it does yeah it wasn't quite
1:32:54 Ella was it but but it was it was it was
1:32:56 at least in a hood that's what I'm
1:32:58 saying like all all these tools right
1:32:59 now they're not quite there but holy
1:33:01 [ __ ] are they going to be amazing in a
1:33:03 year um the the other thing about all of
1:33:06 this music stuff
1:33:08 is right now chat GPT is just multimodal
1:33:11 with pictures and it can do voice
1:33:13 synthesis talking and and hearing but
1:33:17 it's going to do music it's going to do
1:33:19 video it's going to do animations it's
1:33:20 going to do programming it's going to do
1:33:22 like all of that stuff is coming and for
1:33:25 all of them not just for for um open AI
1:33:29 um so sound draw is is another one let's
1:33:33 see it is sound draw I think it's doio
1:33:36 not aai hold
1:33:42 please
1:33:45 iio
1:33:47 yeah let me make sure that I had that
1:33:50 right over here yeah sound draw
1:33:53 .io will you be holding workshops or
1:33:56 classes on AI Salon
1:34:00 um eventually we will um me personally I
1:34:04 I might do workshops I haven't been I've
1:34:07 been doing these things in lie of
1:34:11 workshops but I might put something
1:34:12 together my my reticence in doing
1:34:14 workshops right now is everything's
1:34:16 changing so fast
1:34:18 that I don't know I just feel
1:34:24 I feel like the the the most important
1:34:27 thing I can teach anyone right now is
1:34:29 get curious and and get
1:34:31 um get hungry just get hungry to learn
1:34:35 and don't worry about the right tool the
1:34:38 right prompt any of that [ __ ] n none of
1:34:40 it none of it's going to matter in three
1:34:41 months because the technolog is all
1:34:43 changing so what what you actually need
1:34:45 to
1:34:46 learn right now is what this channel is
1:34:50 show up get curious play with [ __ ] if
1:34:54 I'm showing stuff log into that thing
1:34:56 and be making [ __ ] too if I'm making
1:34:58 jokes about will work for Blandon turn
1:35:00 it into an image and share it on Show
1:35:02 and Tell like
1:35:04 that that
1:35:06 nimbleness
1:35:08 in engagement with AI is the skill that
1:35:13 we're going to need for probably the
1:35:15 next three years it is the
1:35:18 skill the individual tactical how do I
1:35:21 do this how do I do that is essentially
1:35:27 irrelevant most important thing is
1:35:30 understand the problem you're trying to
1:35:31 solve and just understand what does the
1:35:34 current lay of the land look like enough
1:35:36 to go well I could probably try that
1:35:38 with uh with chat GPT oh it didn't work
1:35:41 there oh you know what I heard Bard was
1:35:43 better at this than that let me go try
1:35:45 Bard oh Bard seemed to sort of do
1:35:49 it so
1:35:53 but that
1:35:55 said there is an AI there so what I
1:35:57 would say is go to the AI Salon join the
1:36:00 AI 101
1:36:02 Guild that's Peter kaminsky's Guild he's
1:36:05 doing I think weekly meetings where he's
1:36:08 just picking a topic and teaching it um
1:36:11 and then the other thing I would do if I
1:36:13 were you is as you're learning this
1:36:17 stuff you should start giving workshops
1:36:19 you should start teaching start teaching
1:36:24 why do you think I do
1:36:27 this cuz I'm
1:36:31 learning in order to teach you have to
1:36:33 learn so anyway okay so this is sound
1:36:36 draw
1:36:37 so like I said this is so this is not
1:36:41 this is a totally different style of
1:36:44 comp music composition so these are
1:36:46 pre-recorded Loops that are just getting
1:36:48 stitched together but the way they did
1:36:50 this interface is pretty slick so let's
1:36:53 say I I'm doing a one minute Tik Tock
1:36:55 video and I need a one minute piece of
1:36:57 music I can pick the the speed of you
1:36:59 know the length of the music I can pick
1:37:02 the the pace of the music so I'm going
1:37:05 to do normal and
1:37:08 fast um you can select a genre so so
1:37:11 these these blocks are are kind of
1:37:14 templates or I can select mood or I can
1:37:17 select
1:37:19 theme so mood you know Epic Happ hopeful
1:37:22 laid-back so a lot of times you'll have
1:37:23 a client or you know someone at work
1:37:25 will go I need dreamy music right so you
1:37:29 can just come in here and click on
1:37:30 dreamy Restless funny hopeful whatever
1:37:33 um Angry um and then these are genres
1:37:37 you know hip-hop trap drill R&B ambient
1:37:40 beats techno house Orchestra Funk
1:37:45 Pop
1:37:47 um Christmas so let's go in here we'll
1:37:50 just do uh
1:37:55 let's
1:37:58 see uh we'll do euphoric so I'm going to
1:38:01 do a one minute
1:38:04 euphoric medium to fast
1:38:07 song and so boom there's all your songs
1:38:10 it did I think 20 of them or 15
1:38:13 something like that and you can all you
1:38:15 can just play them right
1:38:17 away and then you'll see here this
1:38:19 Soundwave you'll you'll see it gets
1:38:22 these funny little labels see it says
1:38:24 low medium very
1:38:27 [Music]
1:38:28 [Applause]
1:38:29 [Music]
1:38:30 [Applause]
1:38:35 high all right I hate
1:38:38 that so you can just come in and you
1:38:44 know
1:38:47 awful I think euphoric and fast is not a
1:38:50 good combo so I can now go back up here
1:38:53 and I can
1:38:55 say mood I'm going to get rid of
1:38:57 euphoric and I'm going to turn
1:39:00 on happy and uh
1:39:05 sexy and then under theme I'm going to
1:39:09 say motivational and
1:39:12 inspiring and under instruments I'm
1:39:15 going to turn off acoustic
1:39:19 guitar and ethnic and voice and
1:39:26 Brass yeah that's good and so as I'm
1:39:31 making these
1:39:32 selections it's rewriting all the
1:39:43 [Music]
1:39:50 songs
1:40:04 I can also go in and I can edit these I
1:40:06 can say where it's it's going sort of
1:40:08 low low low here I can go low
1:40:15 [Music]
1:40:19 high now back to low
1:40:24 and back to
1:40:27 high there we go
1:40:30 [Music]
1:40:31 and all right so you can download this
1:40:35 um they're all royaltyfree in Pro
1:40:39 mode you can actually turn on and off
1:40:42 different
1:40:45 tracks
1:40:47 so again like from a technical
1:40:50 standpoint this isn't doing anything
1:40:52 super remarkable except it's doing it so
1:40:56 [ __ ] fast right there's there's one
1:41:00 of the ways I like to think about um
1:41:03 content hang on let me let the Tik Tock
1:41:07 Gods catch up
1:41:11 here
1:41:14 [Music]
1:41:20 um oh this is a good one from Ted this
1:41:23 is a bunch of Civil War guys with a with
1:41:26 a MacBook Pro someone did some time
1:41:29 travel and there's a MacBook Pro with AI
1:41:32 learning labs in the Civil War picture
1:41:35 it's pretty good um one of the ways I
1:41:37 like to think about content and and when
1:41:40 when you see sites like
1:41:46 that musicians who haven't been playing
1:41:48 with this AI stuff will see sites like
1:41:50 that and like oh [ __ ] F I'm [ __ ] not
1:41:53 necessarily I mean
1:41:55 yes all all knowledge work all creative
1:41:59 work is being upended in real time as we
1:42:02 speak it's it's all going to get [ __ ]
1:42:04 up
1:42:06 but there are sort of different levels
1:42:09 of
1:42:11 content and I did a I did a video on
1:42:13 this so I've got a video that's got this
1:42:14 triangle on it and I talk about this at
1:42:18 the bottom of the
1:42:20 pyramid
1:42:21 is I need work that is good
1:42:26 enough so if I just need some background
1:42:28 music for a two-minute corporate video
1:42:33 that's just about the the VPS talking
1:42:36 about if we hit our numbers or
1:42:38 not that music doesn't need to be
1:42:41 awardwinning it just needs to be good
1:42:43 enough the images for AI learning lab
1:42:46 like that says AI
1:42:49 LSI G ring
1:42:52 La LS it doesn't say I AI learning laugh
1:42:56 it's it's it's not good but it's good
1:42:59 enough like we're just using I'm just
1:43:01 using it here like if it's misspelled
1:43:04 does it really [ __ ] matter no so at
1:43:06 the bottom of this pyramid is good
1:43:08 enough automate all of that who gives a
1:43:10 [ __ ] if that's automated who gives a
1:43:12 [ __ ] because I with with a with a tool
1:43:16 like sound draw I'll use Music where I
1:43:19 might not have used it before I'm not
1:43:20 going to it's not like a musician is
1:43:22 necessarily losing their job for that
1:43:24 kind of music because I'm going to use
1:43:26 it in different ways right now does this
1:43:30 impact things like Fiverr and sort of
1:43:32 you know bottom of the barrel to truly
1:43:35 commoditized marketplaces yeah I think
1:43:38 it I think it obliterates
1:43:39 them um in the middle of the of the
1:43:44 pyramid is like I need it to be good and
1:43:47 if I need it to be good I I'll likely
1:43:50 use some of these AI tools but likely in
1:43:53 conjunction with a human being that's
1:43:55 going to refine them polish them up I
1:43:57 like
1:43:58 turtles and then at the top of the
1:44:00 pyramid is I need it to be excellent and
1:44:03 if I need it to be
1:44:04 excellent it may or may not involve AI
1:44:07 it will likely involve lots of human
1:44:10 power and
1:44:13 energy and and that's where you
1:44:17 know craft is still going to matter at
1:44:20 the top of the
1:44:22 pyramid but the bottom of the pyramid
1:44:24 stuff that stuff is just going to like
1:44:26 the value of it's going to drop to zero
1:44:28 so it's going to it's going to compress
1:44:30 the top of the pyramid you're going to
1:44:31 have to be really good to rise above and
1:44:33 then what's going to happen
1:44:35 is
1:44:37 both existing practitioners so let's
1:44:40 let's say it's
1:44:41 musicians existing musicians that
1:44:43 Embrace AI are going to get really good
1:44:45 at AI music and non-musicians who get
1:44:49 really good at AI are going to accept
1:44:51 sell at music or yeah at at AI
1:44:55 music musicians that resist it are going
1:44:58 to get you know sort of squeezed out by
1:45:02 those two you know musicians who embrace
1:45:04 it and non-musicians who embrace
1:45:07 it um and they they're going to get
1:45:09 squeezed they're either going to have to
1:45:10 do top of the pyramid kind of work or
1:45:13 they'll get squeezed into doing you know
1:45:14 live performance things that can't be as
1:45:17 easily you know disrupted so um so
1:45:22 anyway yeah that stuff's coming I know
1:45:24 it's crazy isn't
1:45:25 it um AI or preset sounds framework so
1:45:29 so Pete the the last thing I showed
1:45:31 sound draw is is preset sounds it's it's
1:45:34 a loop composer so they're taking pre
1:45:37 pre-written loops and assembling them
1:45:39 with
1:45:39 AI the sunno and refusion are actually
1:45:44 generating um
1:45:48 music but I still think sound draw is is
1:45:51 worth having in Your Arsenal or at least
1:45:54 knowing it's there if you just want to
1:45:55 knock some [ __ ] out quick that you just
1:45:57 need background you know mood
1:45:59 establishing music it's fine it's
1:46:02 fine you know might you get inspired to
1:46:05 write something from that you might but
1:46:07 it's like you know it's you know [ __ ]
1:46:10 Loop music it's Loop elevator music but
1:46:13 but if I can if I can knock out 20 songs
1:46:17 in 5 seconds and then just sort of push
1:46:20 buttons and tweak them and then listen
1:46:23 to those 20 songs download one and do
1:46:25 that for an hour I'll have a library of
1:46:28 40 or 50 songs that I can use in my
1:46:31 videos without having to spend two or
1:46:33 three months working with people off of
1:46:36 you know um freelance
1:46:39 sites which I did last year I did it
1:46:42 last
1:46:45 year it was [ __ ]
1:46:47 painful musicians are flakes you know
1:46:50 creative people are
1:46:52 flakes some of them are good some of
1:46:54 them are not good some of them are
1:46:56 reliable some of them are not reliable
1:46:58 some of them are bad [ __ ]
1:47:02 crazy if I can get 40 songs in an hour
1:47:05 [ __ ] yes if it's just in that good needs
1:47:08 to be good enough layer of
1:47:11 content I'll do it all day long and so
1:47:14 will everyone
1:47:16 else and that's that's something that
1:47:18 we're going to have to confront cuz that
1:47:19 kind of work is is going to you know
1:47:21 it's just going to dry
1:47:28 up some of us were all of the above what
1:47:31 in the Triangle you mean yeah I think I
1:47:34 think most creative people
1:47:36 do you got to pay them you gota you got
1:47:39 to pay them bills you got to do the the
1:47:42 [ __ ] day job the day job uh version
1:47:45 of your craft
1:47:49 right
1:47:52 new activity let's see what's going on
1:47:54 over in Show and Tell
1:47:58 land what the hell is going on
1:48:02 here James James ziga he's got he's got
1:48:05 a turtle part where I'm afraid to show
1:48:08 it because I'll get i'll get kicked off
1:48:12 Tik Tok um that's a that's a
1:48:19 lovely
1:48:23 the Depression era this is
1:48:27 good yeah that's good that's solid oh
1:48:31 nice the gummy bear wow that the Blandon
1:48:34 bottle is really good here look at
1:48:36 that damn that's actually really
1:48:41 good Kyle you have to change shirts now
1:48:44 no you can see a little bit up here
1:48:47 we're we're far we're far past I'm I'm
1:48:50 let's see 948 I'm not I'm not here for
1:48:52 much longer my voice is dying oh that's
1:48:55 a good
1:49:00 one mcau caulin got his star scream from
1:49:04 Home
1:49:06 Alone from the scream that's pretty
1:49:11 good oh that's beautiful
1:49:14 lunatick with the uh classic holiday
1:49:17 Vibes there that's pretty beautiful nice
1:49:25 really
1:49:26 cool nice work you people you weirdos
1:49:29 you dang your
1:49:43 regulars which one is best for knocking
1:49:45 out songs well it it I think I just
1:49:49 talked about that it depends on the kind
1:49:50 of song if you want if you want songs
1:49:52 with words then sunno and if you just
1:49:56 want background music then sound draw we
1:49:58 can't unsee oh Cherry pasties oh that oh
1:50:02 that's why I need to change the shirt
1:50:04 it's a skull people it's not it's Johnny
1:50:08 Cupcakes you get your minds out of the
1:50:11 gutter you you're right I should
1:50:12 probably keep it a little lower I'm
1:50:14 going to get who's going to tell
1:50:16 him shush I'll keep him lower
1:50:24 it's Johnny Cupcakes people leave me
1:50:27 alone you kids get off my lawn you can
1:50:30 make money with you get out of here you
1:50:33 get out of here you damn kids pick it on
1:50:35 me with my booby
1:50:39 eyes AI is perfect for elevator music
1:50:42 Biz
1:50:43 specific so long music yeah exactly I
1:50:47 know but but Fon is quite quite honestly
1:50:50 think
1:50:51 about at every level of knowledge work
1:50:54 and at every level of creative output is
1:50:57 that shitty low-level work that that
1:51:00 work that's just
1:51:01 it it just needs to be good
1:51:05 enough and that's the that's the stuff
1:51:09 that is mind-numbing soul crushing
1:51:14 repetitive just [ __ ] like everyone
1:51:16 hates doing it so if we can get the
1:51:19 machines to do the shitty work and it
1:51:22 frees up our brains and our cycles and
1:51:24 our souls to do the higher level
1:51:29 work maybe you know quality of life
1:51:32 quality of work goes
1:51:35 up
1:51:36 possible I don't know but you know Kyle
1:51:39 what'sapp has a chat bot for meta
1:51:42 Facebook pretty cool personality
1:51:44 questions oh
1:51:46 interesting yeah I haven't played with
1:51:49 the The Meta I am so not on Facebook
1:51:51 anymore like I don't even I don't even
1:51:54 post content to Facebook anymore um I
1:51:58 haven't played with any of their AI
1:51:59 chatbots the the personality
1:52:03 ones but so I got nothing there who
1:52:06 pinned
1:52:08 [Music]
1:52:11 that now I can't unse see the Cherry
1:52:13 pasties I know dang
1:52:17 it it is it is a it's a
1:52:20 a Mayan
1:52:24 uh man Mandela pattern that's what it
1:52:30 is see now I can't unsee it now I'm
1:52:33 embarrassed by my little shirt
1:52:36 [Music]
1:52:41 boobies would be interesting to have it
1:52:44 cater the music to realtime
1:52:45 conversations in a room yeah I think
1:52:51 I think the idea so so there's you want
1:52:54 to see something let me see if I can
1:52:56 find it first hang on let me see if I
1:52:59 can find it then I'll flip the I'll flip
1:53:00 the camera
1:53:02 um how am I going to find this
1:53:19 uh
1:53:27 gptv
1:53:49 narration
1:54:16 [Music]
1:54:19 um
1:54:25 all right well we'll just do we'll do it
1:54:26 with the soccer one all right
1:54:29 so he's a one-man show ladies and gentle
1:54:33 like a magician on the field dodging one
1:54:35 2 three Unstoppable look at him go the
1:54:38 crowd is roaring can you believe this
1:54:41 he's taken on the whole defense he's a
1:54:42 one-man Show Ladies and Gentlemen he
1:54:44 shoots go messy messy
1:54:47 messy so that narration
1:54:52 is dynamically generated there's another
1:54:54 one there's another one with
1:54:58 um is it League of Legends they're
1:55:02 playing I think it's I think it's
1:55:05 League where it's doing realtime
1:55:08 narration of Esports so what these what
1:55:10 the what these developers are doing is
1:55:13 they're taking a
1:55:14 screenshot every in this case I think
1:55:17 it's every 15 frames the other guy did
1:55:20 it every 15 seconds they're taking a
1:55:23 screenshot using chat GPT Vision to
1:55:27 describe the scene and they're doing
1:55:30 that narration dynamically so it's
1:55:32 basically watching the video and
1:55:34 generating the narration in real time so
1:55:37 you could also do the same thing with
1:55:39 music right so as the scene changes you
1:55:43 could you know do that you could also
1:55:45 use other kinds of inputs moving forward
1:55:48 so the Google's um Vision Pro goggles
1:55:55 they've got sensors they've got I think
1:55:57 16 different sensors in those goggles
1:55:59 and they've got
1:56:00 realtime um data processing on device so
1:56:05 you could you could literally be
1:56:08 monitoring the micro movements in
1:56:10 people's pupils to understand how
1:56:13 they're feeling and
1:56:15 dynamically um compose
1:56:18 music to go with the visuals and to
1:56:21 match their
1:56:23 emotion so like not only is that
1:56:26 possible like the technology to do it is
1:56:30 already being played with like it
1:56:31 already
1:56:33 exists it's just you kind of have to
1:56:35 look around the corner a bit to see
1:56:37 what's coming but yeah this is this idea
1:56:40 of realtime Dynamic personalization
1:56:44 hyper
1:56:45 personalization is not that far
1:56:48 away 18
1:56:51 months probably we'll see hints of it
1:56:54 before
1:56:57 then so [ __ ] [ __ ] cuckoo isn't it
1:57:01 just [ __ ]
1:57:03 nuts what what what what is this world
1:57:07 we're living
1:57:12 in yes Auto adjusting music based on
1:57:15 conversations will likely leave you
1:57:17 without
1:57:18 guests
1:57:21 night n Nancy thank you for perplexity
1:57:23 thanks Nancy good
1:57:26 night what do you recommend for writers
1:57:29 chat gp4 unequivocally
1:57:33 unequivocally chat gp4 it just even chat
1:57:36 gbt 3.5 um some people like Claude if
1:57:39 you're doing uh writing like journalism
1:57:42 that needs research then perplexity AI
1:57:44 you want to lay that on top of
1:57:46 it but for quality of writing GPT 4 if
1:57:50 you don't have access to chat gp4 use
1:57:54 copilot
1:57:56 microsoft.com that's Microsoft's version
1:57:59 of chat GPT sitting on top of the GPT 4
1:58:02 underpinnings or use po the paid version
1:58:05 of
1:58:06 Po you can use the chat GPT for model AI
1:58:11 is going to revolutionize the compliance
1:58:13 and performance space
1:58:16 absolutely every knowledge work every
1:58:18 knowledge work
1:58:21 every
1:58:23 creative with the exception of analog
1:58:25 stuff in the real
1:58:27 world but even that is going to be
1:58:31 impacted how you write songs how you
1:58:35 produce 3D
1:58:37 objects everything's going to be
1:58:39 impacted anything that is that is in any
1:58:42 way mediated by
1:58:45 computer is going to be is going to be
1:58:48 disrupted by this stuff
1:58:54 book's already written just to
1:58:57 edit think about government and data
1:58:59 being uploaded and tracked and analyzed
1:59:01 by powerful AI that's been there for
1:59:03 that's been there
1:59:05 for a decade or two
1:59:08 anyway what what's maybe different here
1:59:10 is the ability to really
1:59:13 easily digest and visualize and exploit
1:59:17 that data that's the difference
1:59:20 so I think the collection of the data
1:59:23 that ship has already sailed and and in
1:59:25 fact that ship has sailed and we gave it
1:59:27 away willingly to Facebook and
1:59:31 Google so we've we've already given all
1:59:33 that [ __ ] away the difference now is the
1:59:36 ability to exploit that data has just
1:59:38 gotten a hell of a lot easier what do
1:59:40 you recommend resource for professors
1:59:42 with help with grading and rubrics um
1:59:46 join the AI Salon pronto
1:59:50 find Mr K who's in here Mr K also has um
1:59:54 a Tik Tok Channel he's a teacher he
1:59:58 actually built a GPT um for for doing
2:00:02 some of that he's done he's done a video
2:00:04 on on grading um rubrics or something
2:00:07 like that just chat gbt is the one cat
2:00:09 gbt will do all
2:00:11 that Bart is not bad either Bard's
2:00:14 getting better um it's it's not it's
2:00:16 it's sort of in the three chat GPT 3. 5
2:00:21 territory I want an AI to replace my
2:00:23 mother when she passes two model oh
2:00:26 actually no I I thought that was a joke
2:00:28 um no not too modeling at all in fact a
2:00:32 good friend of mine Richard Boyd um
2:00:35 before chat GPT came out he did this
2:00:38 three or four years before chat GPT came
2:00:40 out he had access to uh gpt2 to the
2:00:45 API and he used a different technology
2:00:48 than than gpt2 I think he used a
2:00:51 different kind of Technology he
2:00:53 transcribed all of his father's writings
2:00:57 for all of his life he found everything
2:00:59 that his father had ever written and he
2:01:01 created a virtual version of his father
2:01:05 and and um it
2:01:09 checks um like what books his father
2:01:12 would be reading what news stories his
2:01:14 father's paying attention to and he can
2:01:16 talk to him I think that idea of
2:01:18 Memorial memorializing people or doing
2:01:21 it the other way where maybe you've got
2:01:23 you know someone in in an old folks home
2:01:27 with
2:01:28 alimer um or or just you know they're
2:01:31 they're in an assisted living facility
2:01:34 and they want to be able to talk to you
2:01:36 know their
2:01:38 kids why not and especially if you can
2:01:41 you know with these avatars where it can
2:01:43 look like the kids sound like the kids
2:01:45 be infinitely
2:01:48 patient
2:01:50 I don't think it's modlin at
2:01:52 all I think it's
2:01:57 potentially I don't know healing and
2:02:00 empowering and
2:02:03 inspiring is it going to be a little
2:02:05 weird and odd
2:02:12 yeah for a while until it's
2:02:16 not till it's
2:02:18 not
2:02:24 gy you need to get a gold button you can
2:02:27 make
2:02:31 money we'll check in tomorrow dog just
2:02:34 got home from the vet oh okay Kim hope
2:02:36 the doggy's
2:02:38 okay everything will be tokenized I
2:02:40 agree with
2:02:41 that people have suggested politicians
2:02:44 wait all I've got is audio of my mom
2:02:48 yelling my name it's my ringtone for
2:02:52 when she calls oh
2:02:54 wow yeah I oh oh that's all you've got
2:02:56 in terms of in terms of uh of the Audio
2:03:00 I don't think you need I I think um try
2:03:04 11
2:03:05 Labs um I don't think you need audio
2:03:07 that's that long for 11 Labs I think
2:03:09 it's I think some of the some of the
2:03:11 audio synthesis things can the clip can
2:03:14 be as short as 3
2:03:18 seconds
2:03:20 that's very similar to a Black Mirror
2:03:22 episode that's so
2:03:24 beautiful actually I doubt we'll have a
2:03:26 choice I swear there was a movie about
2:03:28 that someone made a robot of a deceased
2:03:31 family
2:03:33 member I listen
2:03:35 I if it if if it's something that is
2:03:38 relevant to you the technolog is there
2:03:41 right now to be able to experiment with
2:03:44 this why not why not experiment with it
2:03:52 if you can interview them or if you've
2:03:54 got recorded interviews here's an
2:03:57 idea and it doesn't even need to be
2:03:59 voice synthesis where you hear them but
2:04:02 it could be it could be well if if
2:04:04 you've got an interview of them you
2:04:05 could obviously synthesize that for the
2:04:07 for the audio
2:04:09 piece
2:04:11 um take a transcript of them being
2:04:16 interviewed upload that transcript to a
2:04:19 GPT you could make a GPT where you can
2:04:22 interact with you know you could create
2:04:25 the GPT in such a way that it uses that
2:04:27 transcript for tone of
2:04:30 voice and then you could give it other
2:04:32 rules you know she never said these kind
2:04:34 of things she always said these kind of
2:04:37 things and then you could interact with
2:04:39 that and then you know you
2:04:43 could you could pop
2:04:46 open chat GPT app
2:04:49 and let's say you you had a uh a
2:04:55 GPT that represented you know your
2:04:58 mother or an uncle or something like
2:05:01 that
2:05:04 and let's
2:05:14 see so you could pop chat GPT into
2:05:18 conversation mode
2:05:27 hey I just wanted to tell you about my
2:05:32 day of course I'm here to listen tell me
2:05:36 about your day and feel free to share
2:05:38 any highlights or challenges you faced
2:05:42 right so you could you could have you
2:05:43 know stuff in there that says I want to
2:05:45 tell you about my day it could respond
2:05:47 like she responded I mean you can do
2:05:50 that right now lunatick my Alexander
2:05:53 Hamilton GPT made me cry in hopeful
2:05:55 tears today
2:05:58 wow amazing in what in in what way hope
2:06:02 hopeful for politics hopeful just for
2:06:06 for what these tools can do I'm just
2:06:08 curious what what made it so
2:06:11 inspiring yeah they say to interview and
2:06:13 make a database of them yeah why
2:06:16 not I see her in the spring I'll see
2:06:19 here in the spring and interviewer
2:06:20 thanks absolutely absolutely what I
2:06:22 would do oh so tonius if you're going to
2:06:25 see her in the
2:06:27 spring work with chat GPT like tell chat
2:06:31 GPT what you're doing say hey I'm going
2:06:33 to see my mother you know she's maybe
2:06:36 toward the end of her life or I don't
2:06:37 know if she's got cognitive stuff
2:06:39 whatever that might be tell it what's
2:06:40 going on and have it create an interview
2:06:44 framework for you have it create a set
2:06:45 of questions and then just you know
2:06:47 modify that and then what whatever you
2:06:49 can get asked ask
2:06:51 it because the ability now to process
2:06:56 this data see the problem before was
2:06:59 everyone was talking about doing these
2:07:00 interviews with loved ones and things
2:07:02 like that but in order to turn it into
2:07:04 something you actually had to program it
2:07:05 and program logic and manually break it
2:07:08 up into semantic groupings and and just
2:07:12 do all that manual [ __ ] large language
2:07:14 models now do all that and they do it
2:07:18 instantly
2:07:20 so just I would say get as much as you
2:07:22 can because your ability to process it
2:07:25 is is kind of infinite and instant I had
2:07:27 a discussion with it on citizens united
2:07:30 oh that's
2:07:31 amazing wow that's super cool yeah yeah
2:07:36 create a GPT of Alexander Hamilton and
2:07:38 say hey what the [ __ ] going on in our
2:07:40 country I'm sure wait ask it to tell me
2:07:43 how Alexander Hamilton would would
2:07:44 respond and deal with it beautiful
2:07:48 amazing
2:07:54 yeah see at this point it's it's it's
2:07:57 almost not about the technology anymore
2:07:58 at this point it's about us figuring out
2:08:03 what is a problem in the world I'm
2:08:05 passionate about that I want to
2:08:08 solve and then let me see if I can use
2:08:11 these tools to do that and what lunatick
2:08:14 just came up with was our politics are
2:08:18 [ __ ] she was in politics are politics
2:08:21 are [ __ ] can one of the founding
2:08:23 fathers help me through
2:08:25 this yes
2:08:28 actively as if they were there why
2:08:30 because Alexander Hamilton wrote so much
2:08:34 [ __ ] we we we've right gave me
2:08:38 legitimate
2:08:39 hope like from a political strategy for
2:08:42 real people I'm hopeful again
2:08:47 amazing
2:08:53 I mean we're just in a [ __ ]
2:08:54 remarkable time and and so few people
2:08:57 know it they don't know what this is and
2:09:00 you can't this is the thing that
2:09:02 frustrates
2:09:03 me you it is
2:09:06 impossible to explain to someone who's
2:09:09 not played with this [ __ ] what it
2:09:12 is that experience that lunatick had
2:09:16 she's like frustrated to the point of
2:09:18 like quitting politics pounding her head
2:09:20 against the wall but passionate about
2:09:24 it and just found a glimmer of
2:09:28 hope for the
2:09:31 future with with this tool that kind of
2:09:34 looks like fancy
2:09:36 Google there's no way she could explain
2:09:38 that to
2:09:41 someone well I put it in and I said how
2:09:44 would Alexander Hamilton answer this oh
2:09:46 yeah well that's that's no no no like
2:09:50 based on all of the data we've got from
2:09:53 Alexander Hamilton and it knows a lot
2:09:55 anyway cuz it's been trained on all of
2:09:57 the
2:09:58 internet thank you
2:10:05 Winston she had an active
2:10:08 living
2:10:10 conversation with one of our founding
2:10:12 fathers and it almost moved her to
2:10:17 tears that's not business as
2:10:23 usual imagine you work for
2:10:26 Ogie you take all the writings of D
2:10:29 David Ogie like if Ogie hasn't made a
2:10:31 GPT of David
2:10:34 Ogie they're missing the
2:10:37 boat anyone here work at
2:10:44 Ogie how do we manage through this
2:10:47 transition David
2:10:49 Mr
2:10:50 Ogie here's what i' do holy [ __ ] you got
2:10:55 a recording of him synthesize that
2:10:58 create an
2:11:00 avatar everyone that starts at
2:11:04 ogal maybe you make a requirement you
2:11:06 spend two hours talking with David
2:11:10 Ogie as part of your
2:11:17 onboarding you using this feels so
2:11:19 intuitive to me I feel like I've been
2:11:21 actually upgraded to this
2:11:24 new Earth to this new Earth
2:11:28 yeah I know me
2:11:34 too when I when I said I was I was doing
2:11:38 the 24 hours of chat GPT someone said to
2:11:41 me is isn't that more like what a
2:11:43 25-year-old would
2:11:47 do I feel like a
2:11:49 25-year-old I feel
2:11:53 like I don't know I can't I don't
2:11:58 know I can't spend enough time right now
2:12:02 getting my head around this and I and
2:12:04 and
2:12:05 it's I can't get my head around it like
2:12:09 the experience I had today where I watch
2:12:11 this video of this guy doing something
2:12:12 I'm totally not capable of and 10
2:12:15 minutes later I have something that
2:12:17 looks like I'm capable of of
2:12:22 it like that's
2:12:27 not that's not a world I expected to
2:12:30 live in it's not a world I thought was
2:12:39 possible I mean as remarkable as iPhones
2:12:43 are you know and and they're like Star
2:12:45 Trek technology it's like the
2:12:46 communicator well we we've now got a
2:12:48 communic Ator in our
2:12:50 hands that was kind of predictable right
2:12:53 that was like you know in our real world
2:12:55 we had phones that had wires that were
2:12:57 connected to the wall and if you were a
2:12:59 fancy person you had a 25t cord so you
2:13:02 could walk into the other
2:13:04 room and so you could kind of see what
2:13:06 the future would look like maybe someday
2:13:09 they can stick all of this technology in
2:13:11 something small I can carry around and
2:13:12 talk into it and that felt like
2:13:16 sci-fi and when we got that it was like
2:13:18 okay I've got that now that's that's
2:13:20 cool that's
2:13:21 powerful but it's it was kind of an
2:13:24 extension of what we had before this
2:13:26 stuff that lunar stick's talking about
2:13:28 that we're talking about
2:13:30 here we've never had this
2:13:33 opportunity when when you interact with
2:13:36 chat
2:13:37 GPT it does not
2:13:40 so these machine learning things are
2:13:43 getting a bad rap because they're
2:13:44 calling them Bots right and sci-fi has
2:13:48 told this that the Bots are going to
2:13:49 kill us and you're like well you're just
2:13:50 talking to a bot you're just talking to
2:13:52 a
2:13:53 bot no what you're actually talking to
2:13:57 is a large language model which is it's
2:14:00 been trained up on all of the [ __ ] that
2:14:03 we've said well whoever got [ __ ] on the
2:14:05 internet in the past four decades five
2:14:09 decades their voice their humanity is
2:14:13 what it's been trained on so when you
2:14:16 interact with it it's kind of like
2:14:17 you're interacting with the collective
2:14:20 people that came before that put [ __ ] on
2:14:23 the
2:14:25 internet because you
2:14:29 are and it's instant and it's not like I
2:14:33 can type in a Google Search and it shows
2:14:35 me here's some documents you can go read
2:14:37 no no it's like it it's it sort of
2:14:40 divines out of this word soup here's
2:14:44 what you people have
2:14:46 said in some Amalgamated way and if we
2:14:50 get our prompt
2:14:52 right that thing gets better and better
2:14:54 and
2:14:56 better that's that's not anything we've
2:15:00 ever experienced and it's I don't know
2:15:04 it's
2:15:05 it's it's a Prof it's a profound thing
2:15:09 it's profound it's like it's like holy
2:15:14 [ __ ] I feel every moment of my life with
2:15:16 AI now that I'm not working I have so
2:15:18 much energy I can't sleep I'm the same
2:15:21 way
2:15:23 Suzanne like why am I doing this two
2:15:27 three four hours a night I'm talking
2:15:30 into my stupid [ __ ]
2:15:33 phone because I come in here and people
2:15:37 are like lutter
2:15:39 stick I spun up Alexander Hamilton and
2:15:43 he's helping me have hope for the future
2:15:46 of our political system
2:15:50 [ __ ] people complain about
2:15:52 hallucinations making AI based on people
2:15:55 is going to do that a lot for now yeah
2:15:58 and hallucinations are not always bad
2:16:01 right hallucinations are bad when you
2:16:03 need facts and
2:16:06 predictability and
2:16:08 accuracy but if you're doing creative
2:16:10 stuff hallucinations are fine right so
2:16:13 so again it's about understanding the
2:16:15 tools understanding how they work today
2:16:18 where the boundaries are where the
2:16:19 limits are what they're good at what
2:16:20 they're not good at where they [ __ ]
2:16:26 up I have had very
2:16:30 few experiences with AI where it's done
2:16:33 exactly what I've wanted it to do I have
2:16:36 had thousands of experiences with AI
2:16:40 where it's gotten me to
2:16:42 80% that last 20% is a [ __ ]
2:16:46 [ __ ] that last 20
2:16:48 20% if you hear the term AGI that's what
2:16:52 AGI is Agi is closing the Gap from 80 to
2:16:57 to
2:16:59 100 because when they get that good when
2:17:02 they're better than us at most things we
2:17:06 do that's that 20% so the reason AI
2:17:10 right now can kind of get you to 80% and
2:17:13 not much above that it's just Where We
2:17:17 Are but you can't know that
2:17:21 without being in the
2:17:23 conversation without playing with these
2:17:25 toys without coming to conversations
2:17:28 like this where someone asks a question
2:17:31 and I go off on some wild tangent and a
2:17:33 bunch of people go and experiment with
2:17:34 [ __ ] and say hey Kyle I just found out
2:17:36 it it works here's how it works oh holy
2:17:38 [ __ ] we all find it out together like
2:17:40 that happens nightly
2:17:47 here I feel like I just went to [ __ ]
2:17:49 church or I'm I'm
2:17:58 pastoring who did this retr
2:18:03 Punk Jim Jim
2:18:06 [Laughter]
2:18:11 Cary that's [ __ ]
2:18:16 awesome Mike kn lovely combining the
2:18:19 gummy bear and the Llama memes here in
2:18:21 the AI learning lab I can't show the
2:18:24 turtle
2:18:25 picture I can wait I can show I can show
2:18:29 that amount of the turtle picture but
2:18:30 right below the the fold there we got
2:18:33 some is we got some turtle parts that
2:18:36 that are
2:18:37 [Laughter]
2:18:40 MPL oh
2:18:44 man all right
2:18:47 people let's see do me a favor um if you
2:18:53 would if you're if you'd be so kind
2:18:57 follow
2:18:58 me um because yeah because I go live
2:19:02 nightly and if you follow me then I'll
2:19:03 get uh you'll get
2:19:07 notified
2:19:10 [Music]
2:19:13 um oh yeah that was a good
2:19:16 one The Irregular Christmas look we have
2:19:20 an Irregulars ugly Christmas
2:19:26 sweater what there we
2:19:28 go why is that so dull there we go
2:19:33 beautiful love
2:19:36 it Lauren thank you for the bear all
2:19:39 right everybody listen my voice was
2:19:42 going oops got to got to keep the
2:19:44 pasties below the fold there check the
2:19:47 irregular section
2:19:56 [Music]
2:20:01 okay oh that's good that's the
2:20:04 uh who is it Eisen stat that did the
2:20:07 that did the horse like that that's the
2:20:10 Llama version of it oh good this is good
2:20:12 some more will work for
2:20:14 Blandon that's the the St Bernard llama
2:20:17 it's very
2:20:25 good there's I think Sherry has started
2:20:29 a uh a whole new trend in the in the AI
2:20:32 salon right now of these kind of 60s
2:20:36 snapshot style photos they're really
2:20:41 cool that's really
2:20:44 cool
2:20:46 huh
2:20:48 [ __ ] crazy man it's crazy I tell you
2:21:02 crazy
2:21:12 um all
2:21:14 right when I bring up AI people think
2:21:16 I'm crazy blank stairs
2:21:18 Sabrina keep talking about it when I
2:21:22 bring up AI people think I'm crazy blank
2:21:24 stairs keep talking about it let them
2:21:26 think you're a
2:21:28 weirdo cuz here's what's going to happen
2:21:30 I promise you I guarantee you cuz it's
2:21:32 happened it's it's been happening in
2:21:34 here pretty regularly just talk about
2:21:36 that you're doing this talk about that
2:21:37 you're interested in it talk about that
2:21:39 you're experimenting with it oh that's
2:21:41 weird let him do that just have
2:21:46 empathy that that's that's coming from a
2:21:49 place of fear that's that's all it is
2:21:53 and I don't even mean to dismiss it
2:21:55 that's not it's all it is that's what it
2:22:01 is
2:22:03 and just keep standing
2:22:07 in staying curious and staying
2:22:10 adventurous and figuring this stuff out
2:22:13 and keep talking about it because what's
2:22:14 going to happen at some point is
2:22:16 someone's going to hear you and they're
2:22:17 going to go oh I I need to figure that
2:22:20 stuff out hey can you come help me with
2:22:22 this and then that's going to lead to
2:22:24 something else and then that's going to
2:22:26 lead to something else and at some point
2:22:27 that friend that was like I don't know
2:22:28 why are you doing that AI stuff I'm
2:22:30 tired of hear about that AI stuff
2:22:32 they're going to come back to you and go
2:22:33 hey I kind of need to know about that AI
2:22:35 stuff could you help but you're not
2:22:38 going to tell anyone are
2:22:40 you the horse Galloping film would Ed
2:22:43 Edward mybridge I I knew that's what it
2:22:45 was I couldn't for the life of me
2:22:47 remember his name thank you myy
2:22:50 bridge Eisen stat Eisen stat did
2:23:01 uh I don't know what the film was but
2:23:04 anyway the first film the first film
2:23:06 that did
2:23:07 editing that was Eis and
2:23:10 Stat uh my my film history is is
2:23:14 escaping
2:23:15 me all right um all right I'm going to
2:23:18 get out of here people listen first of
2:23:21 all I want to say thank you to the mods
2:23:25 I also has anyone used AIS salon. MN
2:23:30 yeah I was just going to talk about that
2:23:32 so if you are looking if if you're one
2:23:35 of those people who that you know that
2:23:37 just said I keep talking about Ai and
2:23:39 everybody thinks I'm a weirdo so I
2:23:41 started this group called the AI
2:23:45 Salon um let me make this bigger so you
2:23:47 can see it down here can you see that
2:23:50 there so go to AIS salon.
2:23:54 mn. that's the um AI salons Mighty
2:23:59 networks that's our home so we have a
2:24:02 meeting coming up Tuesday night it's our
2:24:04 one-year anniversary I started this
2:24:06 group the week after chat GPT launched
2:24:09 because I lost my [ __ ] mind this is a
2:24:12 group of people that are curious about
2:24:14 AI adventurous about AI have empathy are
2:24:17 generous are really cool are up to
2:24:21 [ __ ] everyone in this group every single
2:24:24 person in this group is trying to figure
2:24:26 it out no one's got it figured out if
2:24:28 you're
2:24:29 like Curious and engaged and trying to
2:24:32 figure this stuff out this is the group
2:24:34 for you so join it come join it go to
2:24:37 the about the salon section and look at
2:24:39 our values make sure you resonate with
2:24:41 them if you don't that's cool it's not a
2:24:43 group for you but if you do great there
2:24:46 are things like curi iOS it and
2:24:48 generosity and empathy and inclusion and
2:24:52 exploration and
2:24:55 collaboration go look at
2:24:57 those and then join us and come do cool
2:25:02 [ __ ] and it's not paid it's this is like
2:25:05 this I just do this we're just doing
2:25:09 that repeat the salon website please
2:25:12 this is it here AIS salon. mn. Co
2:25:19 all
2:25:20 right I'm going to get out of here it's
2:25:22 Sunday night
2:25:26 tgim why because you get to [ __ ] play
2:25:29 with AI [ __ ] show up Monday like
2:25:31 let's figure some [ __ ] out I'll check
2:25:34 that out tomorrow thank you awesome Z
2:25:36 thank you thanks my friend shared with
2:25:39 me oh thanks everyone oh that's great
2:25:41 welcome welcome
2:25:42 welcome U he was on Tik Tok but he's
2:25:45 been quiet I have no idea
2:25:48 all right much fake accounts oh
2:25:52 okay yeah oh by the way the this is
2:25:56 maybe important to know if if you f
2:26:01 there I've got a scammer
2:26:03 account that has copied all of my
2:26:06 content here so this account is AI
2:26:09 learning lab and the scammer account is
2:26:11 AI LL learning lab make sure you're not
2:26:15 following that person that person is a
2:26:17 scammer trying to sell people crypto
2:26:20 [ __ ] I won't try to sell you crypto [ __ ]
2:26:24 so if someone reaches out to you that
2:26:26 sounds like this channel going hey want
2:26:28 to get involved in a thing that ain't me
2:26:32 so be careful
2:26:35 okay all right everybody um I think
2:26:39 you're all amazing thank you mods thank
2:26:41 you Irregulars thank you subscribers um
2:26:44 thank you anyone who just shows up here
2:26:45 a lot hair growth geek good to see you
2:26:48 Sher D Dr J Lauren random thoughts thank
2:26:52 you for being a subscriber chaos
2:26:54 generation good to see you prompts cart
2:26:58 I like that name good to see you thank
2:27:00 you Tik Tock dude dude Mia Tobias Kevin
2:27:06 Vasquez goal go oh yeah clink Copperhead
2:27:10 thank you as always theones thank you as
2:27:12 always I really appreciate you dropping
2:27:14 the news bombs and things like that that
2:27:16 is very cool lunatick awesome awesome
2:27:20 thank you for sharing the thing about
2:27:21 Alexander Hamilton that's really
2:27:22 inspiring Silver Fox good to see you jrc
2:27:26 a human being Sabrina oh my God it's
2:27:28 everyone's in here tonight madcap you
2:27:31 guys are great you're awesome
2:27:33 z um who else we got Drew learning AI
2:27:36 he's down under he's probably just been
2:27:38 mowing I think all he does is mow he
2:27:40 must live on like 40 acres and he just
2:27:43 has a little push mower it's all he does
2:27:45 he's got noise cancelling headphones he
2:27:46 listens to this and he mows I don't know
2:27:49 maybe he's got a lawn mowing business we
2:27:51 don't know we don't know and we don't
2:27:55 care all right everybody you have a good
2:27:57 night I'll see you tomorrow uh I would
2:27:59 assume we're going to see some some news
2:28:01 the P the next two weeks there's
2:28:03 probably going to be a fair amount of
2:28:04 [ __ ] launching is just my suspicion it's
2:28:07 my suspicion so be safe out there and
2:28:10 you can make
2:28:14 money have a good night everybody