12/30/2024 - Reflections on AI Festivus and Plans for Future Events in the AI Learning Community

Description

Chapters

Transcript

0:02 I don't want to have to recombobulate
0:04 you you know do you understand what I'm
0:07 saying this is not the can9
0:09 recombobulation
0:15 area all right
0:23 [Music]
0:36 there's been something baby I've been
0:38 trying to
0:39 [Music]
0:42 say don't
0:46 know the past and the future now
0:49 surrounding
0:50 [Music]
0:52 me surrender to give me
0:57 found there's been a little trouble
1:01 since you came to my
1:03 [Music]
1:06 rescue and if you like all of the rest I
1:10 would have quit you long ago but I
1:12 couldn't do
1:16 that oh tell me now women and I never
1:21 went to
1:23 well make him crazy and make him cold as
1:26 hell
1:30 let you wish me
1:32 well but it's SP you try still have to
1:38 find way
1:40 [Music]
1:54 through good evening good
1:58 people good evening good good people
2:02 here I'll put this back up for the for
2:04 the folks on the Tik Tock we are now
2:08 post festiv
2:09 us we did our AI festiv us AI for the
2:13 rest of us and good Lord it was a hit
2:16 people loved it they stayed for hours we
2:21 had 2300 people from literally all over
2:25 the world amazing
2:29 [Music]
2:37 so I'll talk about
2:39 that what we did what my what my
2:42 takeaways
2:43 [Music]
2:47 were we'll do a little play in tonight
2:49 we'll make a little creative stuff I saw
2:52 a really interesting
2:54 um comment from David Shapiro you guys
2:58 know I'm a fan of his
3:01 and hang on let me get this all
3:03 straightened up for y'all let me get
3:06 things straightened up
3:08 here I I ain't done this in a few days
3:12 Pap you understand what's happening
3:15 right uh yeah it looks like you don't
3:17 really know what you're doing
3:20 exactly oh some of us stayed way after
3:23 hours I I saw that being organized and I
3:27 was going to crash it and I thought if I
3:29 crash it I'll be there all night I can't
3:31 do it your head is lower oh yes it is
3:37 there my head is higher
3:38 [Laughter]
3:42 now oh man is there an AI FES of song
3:46 and drink yet no no but we could do we
3:49 could go do that tonight maybe we'll go
3:51 make an AI Festivus
3:55 song
3:57 um we are working on an AI Festivus
4:00 inspired project that was requested
4:03 during Festivus that I'm really excited
4:06 about
4:08 [Music]
4:15 [Applause]
4:16 [Music]
4:27 um is this place I can rest my poor
4:35 head gather my thoughts in sweet
4:42 silence bu this place where the feelings
4:46 [Music]
4:49 AR from an over exposure to
4:53 violence this is place I can slowly face
4:58 the only one I truly can
5:02 know these are tears from a long time
5:06 ago got these tears from a long time ago
5:11 I need to cry 30 years or
5:15 so these are tears of a long
5:21 time go
5:27 [Music]
5:35 30 years of Tears by John hi it he's one
5:39 of my
5:41 faves prolific songwriter oh by the way
5:44 not for nothing if you ain't seen
5:47 complete unknown with Timothy shamay as
5:49 Bob Dylan the [ __ ] are you doing with
5:52 your life that movie is it's it's one of
5:54 the better biopics I've ever seen and
5:57 it's you know what's amazing about it is
6:00 it's got It's essentially got no story
6:03 there's not a lot of story there it it
6:07 it honest to God just has this feeling
6:09 like you're you're sort of a fifth wheel
6:13 just sort of hanging out with with Bob
6:15 and the people that were around him at
6:18 that time in his life when he was sort
6:19 of discovering his musical
6:22 [Music]
6:26 identity but the acting's really good
6:29 the music is amazing it was all played
6:32 live um I ordered adan's book and it
6:35 came already oh that's amazing is it so
6:38 Cuna what's cool about that book is
6:41 those guys just met like a month ago
6:44 like a month or six six weeks ago they
6:47 came from two different communities that
6:49 that I'm participating in and they met
6:51 on one of my office hours and they hit
6:53 it off and they're like we should write
6:55 a book and you just got the book that
6:57 it's just amazing it's just am
7:00 amazing it's it's absolutely Bonkers and
7:03 you know if you were at AI Festivus
7:05 which I'm sure a lot of you were
7:09 um that event came together in 15 days
7:13 start to
7:14 finish like there was nothing there was
7:17 no ideas
7:19 nothing so everything from recruiting
7:22 the speakers to organizing the speakers
7:25 to creating the itinerary to building
7:27 the website to building the schedule to
7:30 figuring out what the hell we were going
7:31 to do to figuring out
7:35 technology to then realizing that this
7:38 event was a hell of a lot more popular
7:40 than we planned on and having to Pivot
7:42 all that happened in 15 days
7:47 [Music]
8:05 standing
8:07 between you and hard plac is
8:11 [Music]
8:13 insane standing too
8:18 near I'm slowly compiling a Kyle
8:20 inspired playlist up to nine songs so
8:23 far oh this one this one you can't add
8:25 this is an
8:26 [Music]
8:28 original uh
8:32 [Music]
8:37 um you're trouble to
8:41 [Music]
8:43 me real trouble can you
8:49 say leaning in
8:52 clothes smell of your perfume scares me
8:56 [Music]
8:58 most leaning
9:01 away feels stronger every
9:08 day you're trouble to
9:14 me real trouble can't you
9:22 say that is a Kyle sh had an original
9:25 called
9:26 trouble um is there an AI festive song
9:29 and drink yet no but we'll probably make
9:31 one tonight digital Gods I can't wait to
9:33 hear you present again it was so great
9:35 and oddly soothing Lord digital Gods had
9:38 the most comments and the most the most
9:41 people were like who is this guy this is
9:43 awesome they loved
9:47 you I think Lord digital Gods was the
9:50 first session where it was truly like oh
9:52 we got some Stars
9:54 here it was pretty awesome and then and
9:57 then we started looking for what we
9:58 loved about speakers after that was
10:00 pretty
10:02 good his uh his fan club spontaneously
10:05 erupted in the back
10:06 [Laughter]
10:14 channel was so good oh
10:18 man so if you don't know what the hell
10:20 I'm talking about well anyway if you
10:22 don't know who the hell I am my name's
10:23 Kyle Shannon this is AI learning lab
10:25 what we do here we learn about AI
10:28 there's listen David David Shapiro put
10:30 out a post today on X that basically
10:33 said the more he uses AI the more he's
10:36 convinced that he his skills are no
10:38 longer necessary in this world and like
10:41 he was a guy that you know when million
10:44 dooll an hour systems went down he was
10:46 the guy that fixed them and brought them
10:48 back up right he's a brilliant dude he's
10:51 written five books on AGI he's got I
10:54 don't know three or four different well
10:56 probably more than that probably more
10:58 like 10 or 20 open AI uh open source AI
11:03 projects and he basically just
11:05 stopped making content about
11:10 it but what he said today is just
11:13 basically he's just not really necessary
11:15 and no one's going to be necessary as we
11:18 move into more and more
11:22 um um as the as these agents start doing
11:27 our work for us and they start getting
11:28 smarter
11:30 [Music]
11:50 Freedom came away that night hang on a
11:53 sec let me do this let me do
11:56 this Lord digital Gods we are unworthy
12:00 [Music]
12:03 not to mention he has the most video
12:05 replay views too that's
12:12 awesome lead singer is overlaying on the
12:14 bed he sang One song he put in his Union
12:16 minimum and Tapped Out Kyle are you
12:20 taking New Year's off yeah I think I'm
12:21 gonna take New Year's off New Year's Eve
12:23 tomorrow night I'm going to take off but
12:25 I'll be here New Year's Day
12:30 New Year's Day night so tomorrow night
12:32 will be off Wednesday will be be
12:35 back and
12:37 uh got to figure out how to kick off
12:40 2025 what we want to focus
12:43 [Music]
12:57 on who was it that was making the the
13:01 Kyle Shannon
13:04 playlist was it side hustle
13:07 [Music]
13:15 Mimi oh DIY
13:17 [Music]
13:23 May no it wasn't you side hustle Mimi
13:26 Daisy May there it is slowly compiling a
13:30 Kyle inspired playlist nine songs so
13:33 far I got a bunch of Martin sexon
13:40 there's what's this one called every
13:43 time I see you now get that looking
13:48 M every time I see your
13:51 mouth hear that
13:54 smile that's Martin
13:57 Sexton there's Martin seon that
14:00 there's Black Sheep by Martin
14:05 Saxton and then there's the
14:09 um his cover of purple rain
14:15 [Music]
14:29 I didn't mean to cause you in a
14:32 [Music]
14:35 trouble didn't mean to cause you in a
14:43 pain only one one time to see you
14:47 [Music]
14:49 laughing only I want to see you
14:54 laugh Source Camp I think 2025 will be a
14:58 repeat of 2024 as the the world begins
15:00 to wake up to AI us a year ago I think
15:02 it's going to be
15:04 different I think it's going to be
15:08 different here's my Tak Source camp and
15:12 this you're not going to want to hear
15:13 this because I know you're already like
15:15 I can't take anymore I think 2024 was
15:19 our year to
15:20 rest I think
15:23 2025 I think 2025 shit's going to start
15:27 acceler like truly accelerating what's
15:29 accelerated so far is the intelligence
15:32 of the
15:33 models but that hasn't trickled down to
15:36 us yet right we don't have access to 03
15:39 we still don't really know how to use 01
15:41 we don't know how to prompt it we don't
15:43 know what to ask it once we have
15:45 autonomous agents they're going to know
15:48 how to use those tools and they're going
15:50 to start I forget I think it was Andre
15:55 karpathy this week basically said
15:59 I think was it him it was him or Demis
16:03 hassabis that basically said
16:06 that probably in 2025 we're going to get
16:09 to the point where the AI is going to
16:13 take over the research for the AI
16:19 researchers that basically Sam mman
16:21 could just sit there and go okay here's
16:23 a bank of servers here you know here's
16:26 100,000 gpus you know here's some tweaks
16:30 to to how we want to optimize [ __ ]
16:32 go and off they go so I have a feeling I
16:37 have a feeling things are going to
16:39 get see slug of Doom artist formerly
16:42 known as Kyle
16:44 [Music]
16:48 exactly if you are oh we don't have many
16:50 people on YouTube do we I not no we have
16:55 11 if you're on YouTube and you don't
16:57 subscribe Please Subscribe over there
16:59 [Music]
17:02 Kyle what happens to those people who
17:04 haven't tried it
17:05 yet
17:08 [Music]
17:23 um I don't
17:26 know if I compare
17:30 what's happening now with what happened
17:32 with the worldwide
17:50 web there's some percentage of people
17:52 that never really got into the worldwide
17:54 web right they were just always
17:55 consumers of it and so I think that's
17:57 going to be similar here
18:10 here I I I honestly don't know the weird
18:13 thing about AI is it's moving so fast
18:16 and changing so fast and it's it's
18:19 computers have never acted like this
18:21 like in the history of
18:24 computing computers were all these
18:26 deterministic machines that we
18:28 programmed logic into and they did what
18:31 we told them to
18:33 do and that's not how they're working
18:37 now it's just
18:40 not I know we want them to be
18:44 more we want them to be more predictable
18:46 we want them to hallucinate less
18:50 [Music]
19:05 adoption on our hands seems so much
19:07 faster than when we rolled out the web
19:09 in our hands yeah yeah exactly like
19:13 we've got access to stuff that is
19:15 dramatically more capable like the web
19:19 took a long time to materialize you had
19:21 to be a geek for a lot
19:23 longer to to really take advantage of it
19:29 here here's the the reason I'm so I'm so
19:32 I'm so flued right
19:35 now is that this is this GNA sound
19:38 [ __ ] up to say
19:40 but it's inspired partially by David
19:43 Shapiro's
19:47 post
19:50 [Music]
19:55 um I think we still think we matter
19:59 when it comes to
20:03 work I know that's G to sound really
20:05 [ __ ]
20:08 up and I think to a degree we
20:13 [Music]
20:15 do but I think it
20:17 becomes optional I think it becomes a
20:20 choice now is this going to happen in
20:23 2025 no it's not but in the next five
20:27 years probably
20:29 where the machines are just going to
20:31 take over doing most of the stuff we do
20:34 I just got an app Robert uh scoble
20:37 talked about this yesterday on Ax X I
20:41 just got an app it's called Simple or
20:43 something like that so will there be any
20:45 benefit to being an early adopter well
20:48 so yes I okay so here's the thing Source
20:51 Camp I think the early
20:54 adopters get lots of opportunity but the
20:57 opportunity is going to be shortlived
20:59 the opportunity is going to be Bridging
21:00 the Gap while these Technologies catch
21:03 up and start doing the work right and
21:06 and while companies ignore
21:09 Ai and all of a sudden realize their
21:13 business model is no longer relevant
21:15 there there's going to be a lot of money
21:17 to be made with those kind of companies
21:19 struggling to survive or figuring out
21:21 what a future looks like so I think I
21:23 think early adopters benefit in the
21:25 short
21:26 term and then I think we all end up in
21:29 this place where we can kind of choose
21:33 the stuff we want to do like there will
21:35 be people that choose to use computers
21:38 to do stuff but like the the need to
21:42 have to do that I think is the thing
21:44 that starts to go away and I don't I
21:47 don't I don't think we
21:51 have there's nothing in our history like
21:54 for Generations that prepares us for
21:57 that
21:58 [Music]
22:02 Redux of our conversation this morning
22:04 yeah Kevin Clark Kevin Clark's in the
22:06 house Hey Kevin what's happening yeah
22:09 exactly we the
22:12 like we are culturally not prepared for
22:16 what's
22:18 coming and I don't so I don't know
22:23 I like one of the things that drives me
22:25 [ __ ] baddy on this channel is when
22:30 I can't articulate where we are and I'm
22:32 in the middle of one of those right
22:35 now I think humans will struggle to turn
22:37 over the busy work yeah too tired to
22:40 well because so a lot of the busy work a
22:43 lot of the busy work in hindsight we
22:46 will look at as busy work but right now
22:49 we look at as
22:50 work and particularly if you're good at
22:53 your work you probably take a lot of
22:55 pride in that and that's everything from
22:58 creative work to knowledge work to admin
23:01 work to like all the stuff that we do
23:03 that is is support of the
23:07 machine almost all of that is going to
23:11 be um we won't need to do
23:15 it it will get done better without us we
23:19 may choose to do
23:21 it but I don't think I don't think
23:24 people are prepared for that um
23:28 [Music]
23:29 that
23:31 [Music]
23:36 reduction oh can you find simple in the
23:38 App Store what does it do hang on one
23:39 sec on Tik Tok let me go find the
23:54 app dang it
23:59 oh my
24:00 God s IIM simple AI
24:05 okay okay it's called Simple
24:08 Ai and what it does is you talk to it
24:13 like chat
24:15 GPT and then it will call up businesses
24:18 for you and have conversations on your
24:20 behalf and report back to
24:22 you so you could say um call up all the
24:26 hardware stores in a 20m radius and find
24:30 out if they have you know air
24:32 conditioners in
24:34 stock and then it'll say you know check
24:37 the ones you want to call and you hit a
24:38 button and it makes phone calls and then
24:41 it shows you the
24:43 transcriptions of the conversation it
24:45 had with people in those
24:47 stores do you know how many [ __ ]
24:50 Saturdays all of us have
24:53 spent [ __ ] dialing up lows when you
24:56 realized you're a slacker and you didn't
24:58 prep hair and now it's hot and all the
25:00 air conditioners are sold out and you're
25:02 a loser and your wife's not very happy
25:05 about
25:06 it not that I have personal experience
25:08 with that I'm I'm I'm saying that merely
25:10 as a
25:14 hypothetical well now you just go into
25:16 your app and 20 phone calls are made
25:19 simultaneously and conversations are had
25:21 with either human beings or Bots it
25:23 doesn't
25:25 matter because my bot will call up
25:29 something on the other end that's a
25:31 human or a bot will find out if they
25:33 have that inventory in stock and let me
25:37 know and I mean that's like like that's
25:40 that's just like a teeny little use case
25:43 but it's
25:44 huge it's like take take you know Google
25:48 searches or the Yellow Pages in the
25:51 olden timey
25:53 days and none of the stuff that we used
25:56 to have to do for that has to be done
26:00 anymore nobody has to print the Yellow
26:02 Pages nobody has to do the Google
26:06 searches side hustle Mimi you can also
26:09 use it to find which ones don't have a
26:11 website or an optimized business listing
26:14 oh wow yeah I like I haven't played with
26:16 simple AI at all I played with it once I
26:19 had it call up what did I have it do oh
26:23 I had to call up two restaurants
26:27 [Music]
26:30 oh side hustle Mimi is giving gifts on
26:32 Tik Tok tonight thank you side hustle
26:34 Mimi that's very sweet of you I
26:35 appreciate it um um I I also had um
26:40 simple Ai call up two restaurants I felt
26:43 bad I didn't want to call up a bunch
26:45 because I didn't in intend on buying
26:47 tacos but but I had to call up two
26:50 restaurants and ask if they had tacos
26:52 and one of them was a Mexican restaurant
26:54 the answer was yes and one of them was a
26:56 burger joint and they said not yet but
26:58 we're thinking about
27:04 [Music]
27:04 [Laughter]
27:08 [Music]
27:13 it how much does it cost $200 a month
27:17 what for simple AI it's free as far as I
27:20 can
27:20 tell I don't know how they make their
27:22 money maybe they're making their money
27:23 from small businesses paying to be
27:25 listed I don't I have no idea I have no
27:27 idea
27:31 what what what simple AI likely
27:35 is simple AI is
27:38 likely I mean you can probably find it
27:40 on TechCrunch if it's a if it's a real
27:42 funded company you can find it um my
27:46 guess my guess is that someone used the
27:49 real the the advanced voice API or the
27:53 Google API one of them with their
27:56 advanced
27:57 voice the open AI Google's API for
28:00 advanced voice and then they basically
28:02 just wrote an application that's got a
28:04 little chat GPT front
28:06 end that that is connected to the
28:09 internet that goes out and finds
28:10 businesses and then you know has the
28:13 call feature hooked up to it and they're
28:15 just using advanced voice so it's
28:17 probably something that was thrown
28:18 together over I don't know couple of
28:22 weeks maybe two
28:24 [Laughter]
28:26 months will it will it wait on on hold
28:28 for hours it might I don't know it might
28:32 people don't like AI nurse technology
28:34 but it's coming well I'll tell you what
28:36 fa Fab
28:41 Fabiana people don't like people don't
28:44 like AI nurse
28:47 technology that
28:50 sucks but I'll tell you what people
28:53 like hate hate more than AI nurses is
28:57 shitty nurs nures or no
29:00 nurses so if we go from having shitty
29:03 nurses or no nurses to AI nurses that
29:06 suck okay that's a toss
29:09 up to AI nurses that are decent great to
29:14 AI nurses that are
29:16 awesome that's coming if you've ever had
29:20 a personal conversation with pie or with
29:24 advanced
29:26 voice um it is a remarkably patient and
29:30 understanding interaction like a
29:32 remarkably patient and understanding
29:34 interaction so imagine a nurse an AI
29:37 nurse that you just talk to like a
29:39 regular
29:41 nurse they know every single thing about
29:43 your case about your medical history
29:45 they know all of it they know all the
29:47 answers they're programmed not to
29:49 hallucinate and give you shitty stuff
29:52 right they've got guard rails and
29:54 they're
29:57 compassionate now they're not in person
30:00 you might
30:01 say all right Reva James hi
30:07 kylen said goodbye to my best buddy dog
30:10 today oh no advice for using AI to
30:14 remember him with piics first of all ree
30:16 I'm so sorry about that doggos doggos
30:19 are the best and the the worst thing
30:21 about doggos is their lifespans don't
30:24 match ours so if you're if you're a dog
30:27 person actually if you're an any animal
30:29 person you're sort of signing up for for
30:32 losing that so I'm very sorry about that
30:35 um best way to do it I I mean one thing
30:39 one thing I would do is um go to chat
30:43 GPT Andor sunno probably chat GPT first
30:48 and just maybe even in advanced voice
30:50 just tell chat GPT all the things you
30:52 remember about your per and what you
30:55 loved about them and
30:59 maybe even have it write lyrics for you
31:01 have it write a poem and then take that
31:03 over to sunno and just describe the kind
31:05 of you know song that you think the dog
31:08 would like or that you know represents
31:10 the dog so you could write it you could
31:12 write a song really
31:14 quickly
31:20 um you know in terms of in terms
31:27 of us using AI I mean I I would say
31:30 something
31:32 like like like low-level AI is probably
31:35 more appropriate something like
31:37 iMovie where you can just dump a bunch
31:39 of pictures in there with some music so
31:40 maybe go to sunno and make a custom song
31:43 and then take a bunch of pictures into
31:44 iMovie and just have it you know it can
31:46 it can do those or I think even photos
31:48 if you're on a Mac photos can do like a
31:52 a photo album with music so you could go
31:54 write a custom song and then make a make
31:57 a movie out of that I think would be a
31:58 really fun way to go the other thing you
32:00 could do is you could take photos into
32:02 Runway ml or photos into Luma labs and
32:06 have it animate photos and sort of bring
32:09 them bring them to life might be kind of
32:11 interesting so oh but I feel so
32:14 bad so sorry about that
32:18 [Music]
32:36 back to the nursing
32:40 [Music]
32:43 thing I don't know if you've been if you
32:46 haven't been
32:49 watching it's [ __ ] crazy if you
32:52 haven't been
32:53 watching what's going on with the
32:55 humanoid robot thing
33:00 we're not that far away from humanoid
33:04 robots oh simple AI is 15 bucks
33:08 okay all right it's 15
33:14 bucks the nursing thing we're going to
33:17 have we're going to have humanoid robots
33:18 that are going to be able to do 24hour
33:23 you
33:24 know
33:26 good you know nursing care in in not in
33:30 not too much
33:31 time um so again it's it's kind of
33:36 hard Kevin and I were talking about this
33:39 this morning as
33:41 well we are we are in an exponential
33:44 curve have you seen the hang on let me
33:46 see if I can go find the the
33:49 chart um
34:16 [Music]
34:20 exponential 03
34:30 yeah this is the
34:38 one this chart right
34:41 here
34:43 um this is the ark
34:46 AGI
34:47 um test which The Arc AGI test to be
34:53 clear is not a test of AGI
34:58 it's a test of one aspect of it and they
35:00 named it that because they're good
35:02 [Laughter]
35:05 marketers but no one else has come even
35:08 close uh to to hitting that thing until
35:11 03 did this but look at these
35:14 numbers gpt2 didn't even
35:17 register on the on the thing so this is
35:21 gpt2 is
35:23 20 18 gpt3 Is 2020 right so this is when
35:30 when news stories first came out about
35:33 gpt3 being this thing's pretty good this
35:35 is where it was then you remember in
35:38 March of
35:40 2023 gp4 comes out and it was so much
35:43 better than gpt3 look how low it was on
35:46 the arc AGI
35:51 score it's like not even a blip then GPT
35:55 40 comes out that comes out
35:59 when was that a year ago right something
36:02 like a year
36:05 ago then this year in the summer we got
36:10 01 preview then we got 01 High 01 Pro is
36:14 now at 50% 01 Pro is at
36:18 50% and then 03 low is at
36:23 77% humans are at um 8
36:28 85% on the AR AGI test and 03 when they
36:34 [ __ ] turned it Loose it was like 1.1
36:36 million in compute for it to answer all
36:38 the questions so it's expensive but it
36:41 hit 87.5 it beat
36:45 humans this is in four years five
36:52 years this is in five years now has this
36:56 affected us yet no
36:58 no this is this is
37:01 the intelligence
37:04 technology that like we don't even have
37:06 access to this 03 [ __ ] only if you're
37:09 paying 200 bucks a month do you have
37:10 access to to that stuff but for 20 bucks
37:14 a month you've got access to this thing
37:16 that is
37:18 dramatically smarter than gp4 so if
37:21 you're still using gp4 and GPT
37:25 40 and you're like yeah that's pretty
37:27 smart like like we're we we're in a
37:29 completely different Ballpark and right
37:32 now it requires humans to understand the
37:35 problems that they're trying to solve to
37:37 use all this
37:38 intelligence I don't think it's going to
37:41 require humans to figure out how to use
37:43 these tools by the end of 2025 maybe
37:46 maybe sometime into
37:50 2026
37:52 um can you talk about deep seek
37:56 V3 I don't know a ton about it I haven't
37:58 used it but deep seek V3 is a model that
38:03 beats GPT 40 and Claud Sonet in a bunch
38:07 of tests including programming I don't
38:10 know about creative writing I don't know
38:11 a ton of details about it but it was
38:15 trained on I think $5.5 million in
38:19 compute 2,000
38:22 gpus and $5 million in compute 2,000
38:25 gpus over a two or three month period
38:28 and it's now better than the
38:30 state-of-the-art models that they spent
38:32 billions on and it's open source and we
38:36 have the
38:37 weights so what's going to happen Becky
38:39 is
38:40 this
38:42 um because of all this open source stuff
38:46 just just imagine you know this line
38:48 that at the very top of this right now
38:50 is
38:51 03 from from open Ai and only special
38:55 people have access to that and it's
38:56 going to be really [ __ ] expensive
38:58 expensive coming in behind that very
39:01 quickly probably in the six month to a
39:03 year is open- Source models whether they
39:06 come from meta or whether X starts open-
39:09 sourcing things or whether they come
39:10 from
39:11 China are are going to all start being
39:13 released and and what the what the um
39:17 what the research Community is going to
39:19 do is they're going to take those models
39:21 and they're going to optimize them and
39:22 make them smaller and smaller and better
39:25 and better and our devices are phones
39:28 are going to get better chips so the
39:30 phones are going to get more processing
39:34 power and then these tools are going to
39:36 get smaller and they're going to get
39:38 small to the point that you can run them
39:40 on a what they call an edge device on
39:42 the
39:43 phone so locally on your
39:48 phone soon like three
39:53 years you're going to have like a
39:56 thousand
39:59 genius level intelligences like way
40:03 smarter than
40:04 Einstein that will do anything you want
40:06 them to
40:07 do locally on your
40:10 phone you don't need a connection to the
40:14 internet but that's coming really
40:20 soon I I like I what does that mean I
40:23 have no [ __ ] clue I I mean I got I
40:25 got no clue I got nothing I got nothing
40:27 for
40:30 you I I thought he knew what he was
40:32 doing this is supposed to be called the
40:33 AI learning lab isn't it is he some sort
40:35 of Professor what kind of professor
40:37 doesn't know what's
40:42 happening
40:45 D from
40:47 Ocean hi Kyle congrats
40:50 on AI Fest of this it was great my
40:53 question how did you come up with the
40:56 idea and if you about the outcome and
40:59 did you ever think it would be this
41:00 successful great question um okay so if
41:05 you don't
41:06 know um we did this thing called AI
41:08 Festivus if you if you go to the website
41:11 it's still there you can go to AI festiv
41:17 us.com and we did this thing that was
41:21 two days on Friday and Saturday
41:29 how did you convene the speakers to
41:31 attend that's these are great questions
41:36 okay let me show you something cool I
41:39 went over to my friend Claude And I
41:43 uploaded I uploaded the um
41:49 [Music]
42:03 um I uploaded the zoom attendee
42:06 statistics into um
42:09 Claude And I said build me a dashboard
42:13 of statistics for AI Festivus so in 15
42:18 days AI Festivus went from being a not
42:21 an idea to the event that happened this
42:24 past weekend we had 3250
42:28 or so people register of those we had
42:32 2335 attend you know some session or
42:35 other which I think is astounding from
42:38 87 countries um with an average
42:41 concurrent attendee rate of 546 people
42:45 so over two 12-hour days on a holiday
42:49 weekend we averaged 546 people watching
42:53 those sessions it's it's absolutely
42:57 mind-blowing so to your question about
42:59 how did we come up with it so last
43:03 year right before the holidays open AI
43:07 launched
43:09 um custom
43:11 gpts and Sam Alman said here's custom
43:14 gpts and what's coming is the GPT store
43:17 where you're going to be able to
43:18 monetize these things and it's going to
43:20 be great and everybody's going to love
43:21 you and you're going to love it and
43:23 everything's going to be good with the
43:24 world they never launched the GPT store
43:27 they launched what they call a store but
43:29 it's really just a bad directory and it
43:33 hasn't gotten better um so so that ship
43:36 is long sale but the custom gpts are
43:38 still very
43:39 valuable so last year at on this same
43:43 weekend um everyone was really excited
43:46 about these so I just kind of had it in
43:48 my I I just kind of had the brainstorm
43:50 what if we did a
43:52 24hour um GPT buildon kind of like a
43:56 hackathon but for gpts and I've done you
43:59 know website building hackathons over
44:01 the years and we always built websites
44:02 for nonprofits so I said why don't we
44:04 build these for nonprofit so I called up
44:06 an Murphy because I didn't want to do 24
44:09 hours on my own I couldn't sustain it
44:12 and I said hey how about you and I
44:14 splitting 24 hours and you know you
44:18 invite your community I'll invite my
44:20 community and we'll just teach people to
44:22 make gpts for 24 hours and I did mine on
44:25 Tik Tok she did hers on zoom and
44:27 everything 3 hours we swapped and in 24
44:30 hours we made 150 custom gpts for
44:33 nonprofits super inspiring the the
44:37 communities were you know people said
44:39 they were going to come and make a GPT
44:41 or two and they stayed for the whole 24
44:43 hours and made 20 I mean it was just it
44:45 was just a cool event so about a month
44:49 ago um an and I connected and and we
44:52 said well are we going to do anything
44:54 we're like ah we could do something and
44:56 we were just sort of kicking ideas back
44:58 and forth and what we came up with was
45:00 wouldn't it be cool if rather than you
45:02 know she and I just teaching a
45:04 tool that we've got all these remarkable
45:07 people in our communities so what if we
45:10 just reach out to the people in our
45:12 communities so she's got she leads AI
45:15 she's part of Rachel Woods's AI exchange
45:18 she does some stuff with Paul rater's
45:20 group the marketing AI Institute I've
45:21 got the AI salon so we've got these
45:24 different communities that that we're in
45:27 and we literally just called up the
45:28 people we know that are doing cool [ __ ]
45:31 like I knew Cindy [ __ ] was doing
45:32 something cool and I knew that Nate St
45:34 Pierre had this cool Hollywood thing and
45:36 I knew that CJ was making good images
45:38 and Kelly Bosch makes these inspiring
45:41 videos like you know you know the people
45:43 that you like in your communities if
45:44 you're if you're playing around in this
45:46 stuff and so we just reached out to
45:50 people and with the exception of two
45:52 people that couldn't do it because they
45:54 were traveling Source Camp was one of
45:56 them Kelly camp I'd love to you know
45:58 still do something with Kelly um with
46:02 the exception of two people who just
46:03 couldn't do it for for Logistics reasons
46:05 every single person we asked said
46:08 yes so we had 24 slots right we had we
46:12 had two 12h hour days in our head and we
46:16 ended up with 34 speakers that said yes
46:18 so then we combined some of them into
46:20 panels and things like that because we
46:21 didn't want to say no to anyone like
46:23 these are all people willing to come
46:25 talk so Annex actually did the curation
46:28 and put together the the um the Run of
46:31 show basically who was going to speak in
46:33 which
46:34 slot um the uh a woman named Marisol
46:38 from from she lead's AI built the
46:41 website um we figured out uh you know
46:45 what technology to use and how to use it
46:47 and all that sort of stuff Vicky and
46:48 Brandon from the AI Salon helped with
46:50 some of the backend stuff Vicki ended up
46:53 you know basically hurting all the cats
46:55 in the AI salon we built within the AI
46:58 salon so if you don't know the AI
47:02 Salon we built a whole custom area for
47:06 AI festiv us that had its own
47:09 chat and so there was just this ongoing
47:12 chat so the back channel for the event
47:15 still exists in the Festivus chat there
47:17 was a showand tell area where people
47:19 could show off work they were working
47:20 off uh you know work that they were
47:23 building based on what they were
47:24 learning oh there's a cute dogger cute
47:27 PPP
47:29 well that's enough internet for the
47:33 day and then we have a resource area and
47:38 the resource area if you haven't been
47:40 there has all of the
47:45 recordings now you can't find these on
47:49 YouTube the only place you can find
47:51 these is in the AI salon so you have to
47:53 be a member of the AI Salon to get these
47:55 and we're going to do some really we we
47:57 we've got an idea to to to a really
48:00 interesting way to repurpose this
48:01 content that I'm really excited about so
48:03 I can't talk about it yet because it's
48:05 not
48:06 baked but that's how it came together
48:09 and then we just basically put it
48:12 together in two weeks and it was just it
48:14 was just kind of tireless work of a
48:16 bunch of volunteers quite frankly that
48:19 made it happen and then some sponsors
48:21 came on that helped us offset a lot of
48:23 the cost because um at one point Sabrina
48:26 Roman
48:27 from LinkedIn and from YouTube Fame
48:30 who's got a really big Community
48:32 globally she said this is a great event
48:35 you should all go attend and so I think
48:37 a huge amount of the Boost came from her
48:43 community I'm starting to call the AI
48:46 haters analog
48:50 artists curiously close to the total
48:53 number of billionaires in the world
48:55 what's that the uh the the uh the the
48:58 number of people that showed up at AI
49:00 festiv us I like it I'll take it yeah
49:03 Vicky was the chief cat hurder she was
49:06 definitely the the chief cat
49:09 [Laughter]
49:13 herter oh
49:16 man so yeah so that was that whatever
49:20 what other question did you have have
49:22 you thought about the outcome oh the
49:24 outcome um
49:29 well like one of the outcomes was when
49:31 we ended the
49:32 event the attendees did not accept that
49:38 reality so so they created their own
49:41 Afterparty after 36 hours of hanging out
49:44 with the likes of us they were like
49:46 we're not going to let this drop we're
49:48 gon to go hang out and so they went and
49:50 hung out for I guess the rest of the
49:51 night I don't [ __ ] know I was going
49:54 to drop in but I knew if I dropped in I
49:55 wouldn't get out um Kyle you've got two
49:58 people waiting to present NE wanting to
50:00 present next year you know what's you
50:02 know what's amazing Mimi is we have we
50:05 have way more than two a bunch of the
50:08 speakers said I want to reserve slots
50:10 for next year so I this this is a model
50:14 so let me let me tell you the thing that
50:15 I'm happiest about with
50:19 this when we first created the salon one
50:22 of the things that immediately started
50:24 happening it's just this happens in
50:26 every Community right right you got a
50:27 bunch of people that sign up and never
50:30 do anything you got a bunch of people
50:31 that sign up and do stuff for a little
50:33 bit and then fade off and then you end
50:35 up with some some amount of people some
50:37 core amount of people in any Community
50:40 IM emerge As Leaders right they they
50:43 just take over and they they welcome
50:45 people in and they're nice and they're
50:46 helpful and they're good and that of
50:49 course started happening within the AI
50:51 salon and one of the reasons we created
50:53 guilds within the salon was we didn't
50:56 have a place for that for that
50:58 leadership to be channeled we didn't
50:59 have we didn't have an opportunity for
51:01 people
51:03 to you know put that energy somewhere
51:06 and so that's why we we created guilds
51:08 what I love about AI festivas is it
51:11 gives members of the community and
51:14 opportunity to put themselves out there
51:16 like one of the things we talk about if
51:18 you haven't seen the AI Salon oh and I
51:20 want to talk about this too an and I
51:22 talked about this today if when you
51:24 first go to the salon you land on this
51:26 welcome page
51:28 and the welcome page is a video from
51:30 Leah faston and I who started the salon
51:33 and we talked about I used we used to
51:36 call this the journey to AI literacy and
51:38 I've I've changed that today I actually
51:40 changed this today um after a
51:43 conversation with
51:45 an I don't think you can be AI literate
51:48 right
51:49 now because things are moving too fast
51:51 right I talk in here all the time about
51:53 I'm not an expert and no one's an expert
51:55 anyone that tells you they're an expert
51:57 is full of
51:58 [ __ ] because things are moving too fast
52:01 but you can be ready you can be ready
52:05 for what's coming right so the journey
52:07 to AI Readiness is is what this is all
52:11 about but in this in this journey it's
52:14 it's a three-step
52:15 journey um quick chatty DD okay my
52:18 crisis has been where are all the
52:20 medical
52:21 breakthroughs we we've got music images
52:25 podcast Etc but where's the medical
52:27 so ton SK the medical stuff is going to
52:29 take
52:31 longer anywhere where you've got
52:33 someone's life on the line is going to
52:36 take longer right driving right
52:39 autonomous driving is you know we
52:42 thought this would be here five years
52:43 ago we're a decade or 15 years into it
52:47 and it's still not quite there it's
52:48 still close um autonomous flying of
52:52 commercial airliners not happening right
52:55 now um
52:57 we do have a bit of robotic surgery but
53:00 it is all you know teleportation kind of
53:03 [ __ ] rather than autonomous robots doing
53:06 the surgery
53:09 um if your medical issue is with janky
53:12 hands we've got you covered um and and
53:15 every you know things have to go through
53:16 the FDA right you can't make claims that
53:18 you've cured something without It Go
53:21 without it going through testing so the
53:23 medical stuff is just going to take
53:25 longer it's a much more conservative
53:28 industry and quite frankly here's the
53:32 deal medical malpractice and incompetent
53:35 doctors kill some large number of people
53:38 every year right let's say it's a
53:41 100,000 shitty drivers kill a 100,000
53:46 people a year or 200 or 500 whatever it
53:48 is medical staff oh medical stuff is
53:51 happening my friend developed an MRI
53:53 test for infants or oh an MRI vest for
53:56 infants oh okay cool yeah the again when
53:59 we're in this exponential curve the
54:01 problem solving is going to start
54:03 getting it's going to start coming much
54:04 more rapidly so I'd say probably in the
54:06 next three years we're going to start
54:09 seeing really dramatic stuff it's just
54:11 going to be
54:12 delayed um the the problem with it is if
54:16 you have one Tesla you know kill someone
54:19 because self-driving was bad that makes
54:22 the headlines a 100,000 people you know
54:24 killed by just bad drivers no it's just
54:27 completely ignored same with human
54:29 doctors just performing bad medical care
54:33 right but the first you know robot
54:36 doctor that causes someone's death
54:39 that's going to make the headlines right
54:40 so so it's just it's just a natural
54:43 transition that we're in the middle of
54:45 is why that will come later okay back to
54:47 this thing with with the salon the
54:50 journey to AI Readiness play first just
54:52 play play with the tools right this
54:55 entire weekend was about playing with
54:57 the tools mindfully Crea is once you
54:59 learn a little bit now start to build
55:01 something build something for a friend
55:03 build something for yourself um oh CJ FL
55:06 Fletcher's in the house CJ Fletcher
55:07 what's happening
55:09 CJ
55:10 um oh CJ's getting some love in the
55:13 comments I love that I had fun talking
55:15 you did great by the way by the way I
55:17 don't know if you saw CJ but I did I had
55:20 to do a spontaneous um session because
55:23 Daisy Thomas got sick I hope you're
55:24 feeling better Daisy if you're watching
55:26 um
55:28 and uh and I used I used a lot of the
55:30 stuff that you talked about and Nate St
55:32 Pierre talked about to to start with a
55:34 really simple shitty prompt and then
55:36 refine it forward it was pretty cool um
55:38 radiator of light thank you for doing AI
55:41 festiv us I attended from start to
55:43 finish I've been
55:45 utilizing what was taught
55:47 fantastic fantastic like so so here's
55:50 here's what we got
55:52 to you're not AI literate now but you're
55:55 probably AI ready you're least ready to
55:57 know how to think about it how to
55:59 experiment with it and so this journey
56:01 is play first mindfully create and then
56:03 generously lead this generously lead
56:05 thing is super important because because
56:09 things are moving so fast no single
56:11 person can learn enough about anything
56:16 to actually have any kind of Mastery
56:18 it's just it's
56:20 it's it is functionally impossible
56:24 there's too much too fast
56:27 when you're in a community CJ Fletcher
56:30 figures out how to do kind of what I
56:32 would call Deep prompting and and Kelly
56:36 Camp learns how to apply things to
56:39 business like Jim Ross figures out how
56:41 to apply things to his business like
56:43 everyone's got these different lenses
56:45 that they look at this stuff
56:47 through what the AI Festivus event did
56:51 was it it gave what ended up being a
56:53 really big stage to the smartest people
56:56 the natural leaders that were already in
56:58 the community already being generous
57:01 what blew me away more than anything is
57:04 is it was to a person we said hey would
57:06 you you know would you want to you know
57:09 participate yep I'm in just tell me
57:12 where to be like here's my here's my
57:14 schedule I've got you know these six
57:16 family events I've got to attend but
57:18 I've got a 4-Hour block here and a
57:19 two-hour block here if you can squeeze
57:21 me in I'll do it not a
57:24 question and that ended up being a
57:26 really powerful thing for all the for
57:28 all the speakers right and it ended up
57:30 being an incredibly powerful thing for
57:32 the for the
57:33 attendees because they they just got
57:35 this schoras board and for the people
57:38 that sat there through the two days
57:39 through 24 hours of this just think
57:43 about like all of what washed over them
57:46 and all the perspective shifts and at
57:48 some point they're checking out and
57:49 they're making nachos and their you know
57:52 Uncle Bob's coming in and talking about
57:54 politics and then they go back to the
57:57 festiv US and they learn a little bit
57:58 more and then they go back and it's all
58:00 just kind of seeping
58:02 in those people that sat through that 24
58:05 hours and kind of went through the
58:06 marathon with us I guarantee you they're
58:09 in a very different place right now than
58:10 they were at the beginning of
58:12 it promise you I like we've got you
58:15 should see the testimonials that we've
58:17 got like this changed people's
58:20 lives it changed people's
58:24 lives cuno
58:27 did we get any new members of the AI
58:29 Salon we got a lot I I don't know if
58:32 you're in here Vicky but do you know do
58:34 you know how many how many we got it was
58:37 a lot because normally when we get when
58:40 in the introduce yourself section this
58:43 section I normally have like one or two
58:45 or three a week right I tell people to
58:48 go go here all the time but see all
58:50 these ones where I I I don't have the
58:52 hearts checked that these are all new
58:54 people so like there's so so just even
58:57 in the past day there's been so I would
59:01 I'm guessing
59:04 3400 but probably more than
59:06 that the other thing that was really
59:09 cool is let me go back to this thing um
59:15 if I go to
59:17 Regional these were we had 950 attendees
59:22 from the United States we had 650 from
59:26 Europe 450 from Asia 150 from Africa 855
59:33 from South America and then um 50 from
59:36 Australia New Zealand
59:40 oceanana like that's a [ __ ] Global
59:42 event like that's a global event right
59:45 you can call it Global if two people
59:47 from London show up we we had 650 people
59:52 from
59:53 Europe 450 people from from Asia 150
59:58 from Africa I mean holy
1:00:02 [ __ ] Vicky on
1:00:05 YouTube about 700 new members of the
1:00:08 salon to put that in context the salon
1:00:13 we were
1:00:15 at 1,200 when when we switched from
1:00:18 Discord to Mighty networks we pretty
1:00:21 quickly got to about a thousand and then
1:00:24 we we kind of bumped up to 200 and then
1:00:26 over the next year year and a half we
1:00:29 just broke
1:00:31 2000 maybe a month ago not even three
1:00:34 weeks
1:00:36 ago so we just added 700 people to the
1:00:40 AI
1:00:41 Salon that's [ __ ]
1:00:43 huge that's [ __ ]
1:00:46 huge like just percentage
1:00:49 wise so to say that the the AI Festivus
1:00:53 was a success like it was a it was a
1:00:55 success
1:00:57 beyond what we could have imagined and
1:00:59 for an event that literally was a vapor
1:01:02 of an idea 15 days before it
1:01:05 launched to what that team ended up
1:01:09 pulling
1:01:11 off
1:01:14 staggering Cosmic lover holy [ __ ] yeah
1:01:17 exactly
1:01:23 exactly does this qualify me to say that
1:01:26 I did did a TED talk you you know it's
1:01:27 funny CJ I mean we we've talked about
1:01:30 this in the salon for a while like one
1:01:32 of our goals as Salon
1:01:36 organizers is that if someone says hey
1:01:39 I'm a member of the AI Salon we want
1:01:42 that to mean something right we want it
1:01:45 to be like it was a TED talk right we
1:01:48 want it to be oh if you're if you're
1:01:50 with the AI Salon you know you're in the
1:01:54 no so
1:01:57 um I don't think we're quite there yet
1:01:59 but I think this
1:02:01 event is very much in the right
1:02:03 direction what I love about this event
1:02:06 is it wasn't just the AI
1:02:08 salon
1:02:09 right Ann's got I think 1,700 members of
1:02:13 she leads AI or something like that
1:02:14 she's got a lot of that and we now have
1:02:18 2700 people we had 2,000 people in AI
1:02:21 salon and then we've got those other
1:02:23 communities that we had speakers come
1:02:24 from and then Sabrina Romanov brought a
1:02:26 bunch of people from her community all
1:02:28 these communities overlapping is a very
1:02:31 good thing you might think oh that's
1:02:34 competitive and those are your people no
1:02:36 no no no no no no we are in an
1:02:39 expansive an exponentially expansive
1:02:43 phase of AI development right now so the
1:02:47 more people we can get involved anywhere
1:02:49 like it doesn't matter where they go if
1:02:51 they're involved and in the conversation
1:02:53 we all
1:02:55 win um the Woodstock of AI it was kind
1:02:58 of a Woodstock of AI I mean it was I
1:03:01 mean look at the look at those stats the
1:03:04 global stats and look at the engagement
1:03:06 I mean the engagement was Bonkers too
1:03:09 let's see here are the sessions right so
1:03:13 I don't they called this engagement
1:03:15 percentage but here's all the different
1:03:17 day one these are this is day one these
1:03:19 are labeled wrong trying to get Claude
1:03:21 to actually refine this little dashboard
1:03:24 has been stupid uh
1:03:27 but but you know the average was 500
1:03:30 what was it
1:03:32 536
1:03:34 546 average concurrent people in
1:03:38 sessions and I mean you know there's day
1:03:41 one there's day
1:03:43 two and then here's all the you
1:03:47 know 588 viewers for the for the
1:03:51 individual there's Suzanne Welker 623
1:03:54 viewers mer El
1:03:58 489 right Bonkers Glenn Osman
1:04:02 [Music]
1:04:03 478 all the engagement numbers in the
1:04:06 80s uh if someone pinned something I
1:04:08 think I missed
1:04:14 it all right what are we going to do
1:04:18 so if you want to see the videos if you
1:04:22 missed AI
1:04:24 Festivus um you can join the salon so if
1:04:28 you go to AIS salon. mn. Co or you can
1:04:33 go to the salon. Pate wanted to know how
1:04:36 many watched his please yep let's go
1:04:38 look and Pate just to be clear um this
1:04:42 dashboard was created by um Claude so I
1:04:47 have no idea if the math is correct I
1:04:50 did not check the math as a math as a as
1:04:53 a doctor in in mathematics I will defer
1:04:57 to you to do actual math on this but
1:05:00 according to the claw dashboard that I
1:05:02 created out of my butt um P modder had
1:05:05 423 viewers with 83%
1:05:08 engagement I mean every like like no
1:05:12 like even at the very end like we had
1:05:14 cyber security you know cyber security
1:05:17 was the was the last session at the end
1:05:19 of 36 hours of programming still had 432
1:05:24 viewers
1:05:26 it's just
1:05:30 amazing actually P's wife on him though
1:05:32 that's awesome oh let's see what CJ got
1:05:35 CJ I think yours was pretty good
1:05:41 too CJ flesher yeah 434 oh yeah that's
1:05:43 what you're responding to yeah exactly I
1:05:46 mean come
1:05:51 on amazing just amazing
1:05:58 I mean if
1:06:25 you
1:06:55 e e
1:07:52 hey guys if you're just tuning into the
1:07:54 AI learning lab we're having some
1:07:56 technical difficulties with the stream
1:07:58 and Kyle's Wi-Fi you can join on Tik
1:08:01 Tock or we'll be back very soon thanks
1:08:04 for your
1:08:24 patience
1:08:54 e
1:09:24 e
1:09:54 e
1:10:24 e
1:10:54 e
1:11:24 e
1:11:54 e
1:12:24 e
1:12:54 e
1:13:24 e
1:13:54 e
1:14:24 e
1:14:54 e e
1:15:40 let's see am
1:15:43 I
1:15:47 back it looks like I might be
1:15:52 back all right let me quit tick tock
1:16:06 how about now you're back on yes
1:16:11 [Laughter]
1:16:18 sir never a dull moment with the
1:16:24 Internet thanks thanks
1:16:26 Xfinity Comcast is the
1:16:30 best oh sorry I wasn't wearing my
1:16:35 uniform I love
1:16:37 Xfinity it's the fastest gigabit
1:16:40 internet in all of
1:16:46 Denver
1:16:48 um say what say it's good it it it's the
1:16:53 fastest internet I ever had did
1:16:58 use thanks
1:17:08 [Laughter]
1:17:19 Xfinity we still got 47 people on Tik
1:17:24 Tock good
1:17:26 Lord oh
1:17:33 Xfinity I did want to do a cool thing or
1:17:35 two tonight all right let's go do a cool
1:17:38 thing or two all right so here's what
1:17:40 we're going to do am I back I'm good
1:17:43 right is this a sponsored
1:17:47 post does the Creator earn commission
1:17:50 the the Creator should [ __ ] earn
1:17:52 commission for all of the positive uh
1:17:56 reviews I give about Xfinity
1:17:58 [Laughter]
1:18:06 internet all right hang on I gotta
1:18:08 scratch the bugs out of my
1:18:13 scalp I'm sure it's just a rash
1:18:17 okay let me present my
1:18:20 screen share my screen screen
1:18:25 SC what the hell happened
1:18:28 there why oh because that's not up on
1:18:30 stage no there we go all right we're
1:18:34 good all right so here's what we're
1:18:36 gonna
1:18:38 do
1:18:41 [Music]
1:18:45 um okay I am gonna go
1:18:56 wonder if there's an easy
1:18:59 way I don't think there's an easy way to
1:19:02 do this yeah there is I'm G to go here
1:19:05 all
1:19:07 right okay so here's what we're going to
1:19:09 do we're going to go copy get sharable
1:19:12 link okay so here's what I'm
1:19:15 doing so if you were at AI
1:19:18 Festivus and you're a member of AI
1:19:21 Salon you can go see recordings we're
1:19:24 not going to put these out into the
1:19:26 public because right now [ __ ]
1:19:30 them if they if they didn't abandon
1:19:34 their family for 36 hours to hang out
1:19:37 and learn about AI then they can't they
1:19:40 can't come watch it they got to come
1:19:41 where the where the video is
1:19:43 [Laughter]
1:19:47 okay so that's that's the way we're
1:19:49 going to do this we can be a little
1:19:51 Dicky we can be a little Dicky sometimes
1:19:54 and if if you didn't spend 36 hours with
1:19:56 us then you have to join the salon to
1:19:58 see these
1:19:59 videos stat omatic
1:20:01 TM
1:20:07 Lord adding unnecessary trademarks to
1:20:10 the end of things
1:20:14 um and coming up with stupid
1:20:17 names
1:20:21 stomatic it just gives me great joy okay
1:20:28 ah all
1:20:31 right radiator radiator of light was
1:20:34 there John Hess is going to watch it
1:20:37 great listen if you're if you come into
1:20:39 the salon that's where that's where all
1:20:40 the replays are but here's what I'm
1:20:41 gonna do I'm gonna go
1:20:44 backstage because I have that power so
1:20:46 this is our this is our backstage
1:20:49 YouTube area and I'm going to go get all
1:20:52 of the URLs for these and we're we're
1:20:55 going to go
1:20:56 to not there notebook LM we're going to
1:21:00 create a new
1:21:02 notebook and we're going to do YouTube
1:21:05 link paste your YouTube url bang
1:21:09 insert then we're going to go back why
1:21:12 ises that say it's
1:21:15 red oh the transcripts aren't available
1:21:18 damn
1:21:22 it all right well at some point they
1:21:25 will
1:21:29 be so here's what I was gonna do I was
1:21:32 gonna go here and do
1:21:35 um notebook
1:21:39 LM to be able to interact with all the
1:21:42 sessions so and and make podcasts of
1:21:44 like day one and day two stuff like that
1:21:47 so we'll do that some other night we'll
1:21:48 do that Wednesday once we have it t it
1:21:50 takes like 24 hours for YouTube to get
1:21:53 the things there
1:21:55 they will someone someone asked for them
1:21:57 today yeah they'll they'll they'll get
1:21:59 their YouTube just takes a while they
1:22:01 they just take a while to get your
1:22:02 transcripts okay so we're not doing that
1:22:05 now okay so here's what we are GNA do
1:22:07 well what are we gonna do now Kyle you
1:22:09 said you were gonna show us something
1:22:12 cool you can make money
1:22:15 withing I can send you the transcript
1:22:17 files really wait are they not
1:22:23 there do you have them all
1:22:26 Vicki you want to drop them in a Google
1:22:28 drive
1:22:35 folder you think we did pretty
1:22:37 spectacularly already holy [ __ ] yes by
1:22:40 the way not for
1:22:43 nothing the
1:22:46 single most
1:22:49 Relentless question that was
1:22:52 asked 4,000 722 and 62 times an hour was
1:22:59 is there going to be a recording of This
1:23:01 Is there going to be a recording of This
1:23:03 Is there going to be a recording of This
1:23:05 Vicky fielded every one of those
1:23:07 questions for 36 hours so and she got
1:23:13 all of the videos up on the AI Salon in
1:23:17 um like record time so so um first of
1:23:21 all thank you Vicky and you all should
1:23:24 be doing very nice things to Vicki if
1:23:26 you see Vicky in the in the salon you
1:23:29 should say very nice things to her for
1:23:31 one thing she has the power to boot your
1:23:33 ass if you're a
1:23:35 jerk and she'll do it but she's actually
1:23:38 very fair she will not boot people just
1:23:41 because they had a bad day but she will
1:23:43 boot your ass if it needs needs to be
1:23:45 booted um she dealt with every single
1:23:49 person was like is there going to be a
1:23:50 recording is there going to we're doing
1:23:52 an event for 2,000 people you think
1:23:53 we're going to not record it of course
1:23:55 there's going to be a recording do you
1:23:57 get access to it not if you ask like
1:24:00 that but she she got all the videos up
1:24:05 within 24 hours of that event
1:24:08 happening when we do our AI Salon events
1:24:11 until I had Vicky to help me do that it
1:24:13 would sometimes be four months until I
1:24:15 got videos up from our sessions so the
1:24:18 fact that she got she split them all
1:24:20 into individual panels and got them up
1:24:22 in the resources area come on on all
1:24:27 right is excessive chatting a bootable
1:24:30 offense no we love excessive chatting we
1:24:33 we want participation um excessive
1:24:36 stupid chatting you know that thing
1:24:38 there are no stupid questions that's not
1:24:43 true so many people are on the naughty
1:24:46 list exactly the fesus naughty list
1:24:55 oh
1:24:56 man oh sending bitra and Band-Aids to
1:25:01 Vicki from her bloody typing fingers
1:25:03 seriously well she did she did get smart
1:25:06 Vicki's pretty clever she had she had a
1:25:08 bunch of copy and
1:25:10 paste copy and paste responses to stupid
1:25:13 questions excess excessive chatting is
1:25:15 encouraged excessive stupid chatting is
1:25:18 is what will get you banned
1:25:26 oh man all right so let's go do
1:25:27 something fun I'm gonna change my
1:25:30 sharing
1:25:32 style here for a second I'm gonna
1:25:34 present I gotta make the tab first hang
1:25:40 on I'm so glad our internet is back that
1:25:43 that was we haven't had a a [ __ ] just
1:25:46 fullon Wi-Fi meltdown in
1:25:50 ages um Luma labs
1:26:01 okay so if you don't know Lia
1:26:06 Labs it's cool it's a very simple you
1:26:09 can just type in um let's
1:26:13 see
1:26:16 um
1:26:19 um
1:26:21 horse
1:26:23 running a
1:26:25 cross a
1:26:27 river in the style of anel
1:26:33 Adams and then I'm G to make it 16 by9
1:26:36 I'm gonna make it a video and then I'm
1:26:37 gonna make a video no screen share oh
1:26:40 did you want me to share
1:26:43 this ah y'all are
1:26:53 cute so so feel free to play along if
1:26:56 you've got a Luma Labs uh subscription I
1:27:00 think they may have
1:27:02 free free
1:27:09 gens and so here's the horse running
1:27:12 across the river in a style of anel
1:27:14 Adams why is it not playing I guess it
1:27:17 made the video it made the image
1:27:20 first so I guess it'll make the video
1:27:22 after this
1:27:33 [Music]
1:27:37 [Laughter]
1:27:40 physics physics not the strong suit here
1:27:44 okay Vicki you are the best this is a
1:27:46 reminder to all if your router is two
1:27:48 plus years old restart or replace it if
1:27:51 it doesn't help oh that's actually good
1:27:53 advice
1:27:55 mine is I think less than a year so I
1:27:57 think I'm
1:28:02 good Queen Vicky yep she's the best she
1:28:05 really
1:28:11 is
1:28:13 okay all right so one of the things that
1:28:16 Kelly Bosch talked about last night was
1:28:19 was Luma labs and I figured what we'd do
1:28:22 is we' do a little we do a a little
1:28:25 action we do a little something and I
1:28:27 want to show you so in Luma laabs
1:28:28 there's two modes there's just kind of
1:28:31 this video
1:28:32 mode
1:28:35 where how do I get out of
1:28:39 here click there where you know you've
1:28:41 got this is all the crap that we've
1:28:44 created and then some of them are videos
1:28:46 and some of them are
1:28:50 not and then they've got this new thing
1:28:52 called boards and boards are
1:28:56 weird they're a weird interface but
1:28:59 they're kind of cool so like here's a
1:29:01 civil war board that I did I made this
1:29:03 Civil War video called Shadows of gray
1:29:06 and
1:29:07 blue and let me scroll all the way to
1:29:10 the top here what I started out
1:29:14 making was
1:29:16 images and then I turned those images
1:29:18 into videos and the way it works is you
1:29:21 scroll left and right like if you want
1:29:23 to do iterations of a prompt it'll give
1:29:27 you like these Mad Libs pull Downs so
1:29:30 this was like scared soldiers but I
1:29:31 could
1:29:32 say frightened civilians or determined
1:29:35 generals and now it'll make images of
1:29:37 determined generals and so you can just
1:29:40 make images left and right and then any
1:29:42 one of these you can turn into a
1:29:44 movie and so it's just this long
1:29:46 scrolling list of you know some
1:29:50 horizontal digging deep on something and
1:29:52 some are just video things and then if
1:29:54 you want like there a generation failed
1:29:57 I can just say show more and it'll make
1:29:59 new videos you know based on that
1:30:02 prompt all right so that's what a board
1:30:07 is and
1:30:17 then and then the other thing that that
1:30:21 they
1:30:22 added into this this is I
1:30:27 can let's go from so I can click on I
1:30:31 like let's say I like this kangaroo
1:30:33 right and I can say make this into a
1:30:35 video now you'll notice in the lower
1:30:37 left here it added that as the opening
1:30:41 frame and I could say let's go from that
1:30:44 kangaroo to this
1:30:47 kangaroo I'll say make video and that'll
1:30:50 add that as the second frame so now it's
1:30:51 going to go from this flat art to this
1:30:54 more 3D art and I can just with no
1:30:56 prompt I can just say make a
1:30:58 video and so it's going to make a video
1:31:01 of
1:31:02 that and so this is the thing that Kelly
1:31:05 Bosch said she was completely enamored
1:31:07 with so what I thought might be fun was
1:31:09 to do a little we'll go over to chat
1:31:14 GPT oh that's so cool those images
1:31:16 turned out nice yeah they were cool
1:31:17 right and and it was I just came up with
1:31:20 a a silly little prompt and it came up
1:31:22 with something cool and then you can
1:31:24 just iterate on those so I figured what
1:31:26 we'd do is we'd go over to chat GPT and
1:31:28 get some
1:31:30 inspiration all right let's look at our
1:31:32 videos let's see how they came
1:31:34 out that's kind of
1:31:38 cool not
1:31:41 really not really morphing from one to
1:31:44 the other that one
1:31:48 is it's kind of janky kind of janky but
1:31:51 kind of
1:31:52 cool right like like that one kangaroo
1:31:55 was hanging on the wall and the other
1:31:56 kangaroo is just sitting is the
1:32:03 artist oh man Kyle she was talking about
1:32:07 making a
1:32:08 still of the last gram of videos to use
1:32:11 as the first that's exactly what I'm GNA
1:32:13 do so that's what exactly what I'm going
1:32:14 to do Joy so so what I did here so let
1:32:18 me let me go back out to my
1:32:21 board okay so what I did
1:32:25 was I started out with bosot so I said
1:32:29 what did I say a bosot style
1:32:32 illustration of a kangaroo I did a
1:32:34 stupid thing just a stupid simple prompt
1:32:38 and then basically what I did was I
1:32:40 picked a start frame that I liked and
1:32:43 then a second frame that I
1:32:46 liked and then I made a movie so here
1:32:49 was the first movie I made
1:32:55 and then you
1:32:56 take the end frame and you use that as
1:32:59 the beginning frame of the next
1:33:03 movie right so this is the end frame of
1:33:07 that last
1:33:08 movie oops this is not being viewable by
1:33:12 you
1:33:14 folks this is shitty interface for this
1:33:18 this is how I made my sunshine keeper
1:33:20 vid oh cool
1:33:22 nice there's nothing stupid about
1:33:25 kangaroos uhoh the austral the
1:33:27 Australians are getting defensive we we
1:33:30 got to be careful people when the when
1:33:32 when the Aussies start getting defensive
1:33:34 nothing good can come of it nothing good
1:33:36 can come of it Just Praise
1:33:39 marsupial praise marsupial we don't want
1:33:42 steo on our bad side just call meow's
1:33:45 gonna come in throw throw a fit it's
1:33:48 it's going to be a whole thing Drew Drew
1:33:51 is going to have a contion there's going
1:33:53 to be discom
1:33:55 ulation all right so let's let's do this
1:33:58 from scratch so rather than showing you
1:34:00 what I did before let's do it from
1:34:03 scratch we'll do it live we'll do it
1:34:08 [Music]
1:34:22 live we're going to go to the CH
1:34:25 J and we're going to say hey chat
1:34:32 J let me make this a little bigger so
1:34:34 you can see it can you see that yeah
1:34:36 that looks
1:34:37 good hey chat
1:34:40 je should we get um should we talk to
1:34:44 Quinn let's talk to Quinn hang
1:34:52 on so what I'm about to show you is a is
1:34:55 a prompt engineering
1:34:58 technique I by the way I hate the term
1:35:00 prompt engineering I think it's
1:35:02 stupid um prompt crafting I'm fine with
1:35:07 just like talk [ __ ] into existence is
1:35:10 better so let's just talk [ __ ] into
1:35:13 existence
1:35:18 here hey Quinn how's it
1:35:21 going hey Kyle I'm doing well thanks how
1:35:24 about you I'm doing good I want some
1:35:26 help I want to come up with some really
1:35:27 interesting images and I want to use a
1:35:31 new technique I learned called the
1:35:32 double twist technique have you ever
1:35:34 heard of
1:35:35 it the double twist technique sounds
1:35:38 intriguing could you tell me a bit more
1:35:40 about it yeah so I'm gonna have you help
1:35:41 me come up with some ideas for images
1:35:46 and then we're g to do we're g to pick
1:35:47 one and then we're going to do a twist
1:35:49 on that idea and then we're going to do
1:35:52 a twist on the twist that's the double
1:35:54 twist get
1:35:57 it got it that sounds like a fun
1:36:00 challenge let's start with a base
1:36:02 concept any particular themes or
1:36:04 subjects you have in mind well I want
1:36:06 something that's super artistic and
1:36:08 something that's surreal and something
1:36:10 that's going to be really good for
1:36:11 generative AI to make where accuracy
1:36:14 doesn't matter as much so I'd like you
1:36:16 to help me come up with interesting
1:36:19 subject matters and interesting Artistic
1:36:22 Styles and it should be
1:36:27 photographic all right let's brainstorm
1:36:30 how about starting with something like a
1:36:31 cityscape where the buildings are made
1:36:33 of glass and filled with flowing objects
1:36:36 we could explore Styles like surrealism
1:36:38 photo montage or even a dreamlike
1:36:40 quality what do you think boring can we
1:36:44 go less boring that like that feels very
1:36:47 expected I want I want to win the Nobel
1:36:51 Peace Prize because these things are so
1:36:53 good
1:36:55 all right let's take it up a notch how
1:36:57 about a scene where a forest is made
1:37:00 entirely of musical instruments and the
1:37:02 trees play themselves in the wind we can
1:37:04 explore Styles like photorealistic
1:37:06 surrealism or even a painterly approach
1:37:08 with elements of abstraction does that
1:37:11 spark any inspiration that's good why
1:37:13 don't rather than coming up with a
1:37:14 single concept why don't you give me 10
1:37:17 just give me 10 little short ones and
1:37:19 then I'll pick one that I
1:37:21 like sure thing here are 10 Concepts one
1:37:25 a library where the books are made of
1:37:27 clouds drifting off the shelves two a
1:37:30 cityscape at night where the buildings
1:37:32 are giant glowing crystals three a
1:37:34 desert landscape where the sand dunes
1:37:36 are made of folded paper four a mountain
1:37:39 range with Peaks made of floating let's
1:37:41 get let's let's do these as people I
1:37:43 want I want Humanity in this right now
1:37:46 we just got the cold 3D world and it's
1:37:49 it's making me go to sleep I want
1:37:52 interesting characters that are we weird
1:37:54 and funky and Unforgettable and will
1:37:56 make me cry when I see
1:38:01 them got it here are some
1:38:03 character-driven Concepts one a street
1:38:05 performer whose Shadow has a life of its
1:38:07 own dancing differently from him two a
1:38:10 wise old woman with a cloak made of
1:38:12 living fluttering butterflies three a
1:38:14 child whose hair turns into
1:38:16 constellations when they dream okay
1:38:17 let's do the old woman with the coat of
1:38:21 butterflies great choice now for the
1:38:24 first twist how about the butterflies on
1:38:27 her cloak write write a really detailed
1:38:30 image prompt of that woman and be very
1:38:33 detailed in her look her setting her
1:38:37 coat the kinds of butterflies everything
1:38:39 be incredibly
1:38:42 detailed certainly here's a detailed
1:38:45 image prompt imagine an old woman with
1:38:47 deep Wise Eyes and a gentle smile okay
1:38:50 that's good her great I'll trust that
1:38:52 what you came up with is good Now give
1:38:54 me five ideas for a
1:38:57 Twist sure here are five twist ideas one
1:39:01 as the fireflies emerge from her cloak
1:39:04 they illuminate hidden mystical symbols
1:39:06 on the ground two the butterflies
1:39:08 transform into tiny glowing spirits that
1:39:10 whisper Forgotten Tales three the old
1:39:13 woman begins to float as the fireflies
1:39:15 lift her gently into the air wait four
1:39:18 Quinn you're giving me you're giving me
1:39:20 animation ideas I just need I want to
1:39:22 start with a single image so just give
1:39:24 me twists on the image
1:39:29 itself got it here are five twists on
1:39:32 the image one the butterflies on her
1:39:34 cloak are made of stained glass
1:39:36 reflecting colorful light around her
1:39:38 that's good I love that one now give me
1:39:40 another
1:39:41 twist how about the old woman's cloak is
1:39:44 adorned with butterflies that are
1:39:46 actually tiny intricately designed
1:39:48 Clockwork creatures each with delicate
1:39:50 gears and cogs well but let's do it so
1:39:53 that it's building on the previous twist
1:39:55 right the previous twist was butterflies
1:39:58 now made of stained glass so give me
1:40:02 five ideas for a Twist that builds on
1:40:04 that rather than replaces
1:40:06 it got it here are five twists building
1:40:10 on the stained glass butterfly concept
1:40:12 one the stained glass butterflies
1:40:14 reflect intricate Vivid scenes from the
1:40:16 woman's past telling her life story love
1:40:19 that keep
1:40:21 going sure here here's the next twist
1:40:25 three the stained glass butterflies
1:40:27 change colors with the shifting light
1:40:29 casting Prismatic patterns on everything
1:40:31 around her
1:40:34 nope keep
1:40:37 going four each stained glass butterfly
1:40:40 has a unique pattern reflecting
1:40:42 different emotions and memories no
1:40:45 what's the fifth
1:40:47 one five the stained glass butterflies
1:40:50 gradually detach and form a vibrant
1:40:52 glowing Mosaic IM the air around
1:40:55 her let's go with the first one where
1:40:58 the the stained glass reflects images
1:41:03 from her past okay so what I want you to
1:41:04 do now is write me an incredibly
1:41:07 detailed
1:41:09 prompt certainly here's a detailed
1:41:11 prompt imagine an old woman with deep
1:41:15 Wise Eyes and a gentle smile standing in
1:41:17 a lush Enchanted Forest okay great now
1:41:21 rewrite that prompt so that it's less
1:41:23 than
1:41:24 300
1:41:26 words certainly imagine an old woman
1:41:29 with Wise Eyes on a gentle okay and so
1:41:32 now I'm gonna stop and what's cool about
1:41:35 this oh I think I might have needed
1:41:41 to I think I cut her off so it's so i'
1:41:44 I've got to let her say it all
1:41:48 right sorry say the last prompt prompt
1:41:51 again I need a prompt that's 300 words
1:41:53 or left Les of the the new woman
1:41:58 prompt of course imagine an old woman
1:42:02 with deep Wise Eyes and a gentle smile
1:42:04 standing in an enchanted forest her face
1:42:07 is lined with the wisdom of years and
1:42:09 she wears a cloak adorned with stained
1:42:11 glass butterflies each butterfly
1:42:13 reflects vibrant intricate scenes from
1:42:16 her past casting colorful patterns of
1:42:18 light around her the forest is filled
1:42:20 with a soft magical glow enhancing the
1:42:23 serene and mystical
1:42:26 atmosphere okay that's
1:42:28 good I'm glad you liked it
1:42:32 okay so there's our prompt okay
1:42:36 so let me do this for a
1:42:40 second so one of the reasons I love
1:42:43 doing that is you don't have to be a
1:42:46 prompt engineer to do this stuff like
1:42:48 this that conversation I had with her
1:42:51 while I was I was sort of forcing the
1:42:53 issue and moving really fast because you
1:42:55 know I don't want it to be [ __ ]
1:42:57 boring that's not dissimilar to the kind
1:43:00 of conversation you would have with an
1:43:02 art director if you worked in an agency
1:43:04 or with a designer or with yourself if
1:43:07 you're thinking about coming up with
1:43:09 ideas for something so when when it
1:43:11 comes time to use generative AI don't
1:43:15 feel like you have to be so [ __ ]
1:43:18 formal it's not programming you can have
1:43:20 fun with it you can you can experiment
1:43:23 and push boundaries and tell her no that
1:43:26 sucked like like here's the
1:43:28 thing what generative AI is giving us
1:43:31 all the ability to be I just wrote an
1:43:33 article on this is we get to be like
1:43:34 Rick Rubin or we get to be like the
1:43:36 creative director at an agency or the
1:43:38 producer of a movie we get to just sit
1:43:41 back and and direct it and just say I
1:43:44 want it to be more in this direction or
1:43:46 more in that direction or I want the
1:43:47 poem to be like this or I want it to
1:43:49 rhyme less or I want to create a novel
1:43:51 or a screenplay or whatever we get to
1:43:54 direct
1:43:55 that that's incredibly powerful and it's
1:43:58 incredibly creative but you have to give
1:44:01 yourself permission to go there and you
1:44:03 have to give yourself permission to not
1:44:05 feel like you have to get it
1:44:11 right if it's a creative process right
1:44:14 if you're trying to do math like I'm
1:44:17 trying to make that dashboard of
1:44:19 Statistics it would be best if those
1:44:21 statistics were right right
1:44:24 so different applications have different
1:44:26 things but all right here we are we have
1:44:28 a
1:44:29 prompt now let's so go see if we can
1:44:32 turn this prompt into something
1:44:35 interesting so I'm going to go back and
1:44:38 I'm going to
1:44:39 share Luma Labs oh Luma Labs where are
1:44:45 thou Luma Labs I do not see you oh there
1:44:48 you are kangaroo dreams okay so here we
1:44:51 are so we're g to go back out
1:44:54 we're going to create a new
1:44:57 board and we're going to start with
1:44:59 images 16 by9
1:45:02 images and we're just going to paste
1:45:04 that prompt in
1:45:09 okay
1:45:12 boom and then what Luma laabs should do
1:45:15 is it it should um whatever you call it
1:45:19 Mad Lib eyes this
1:45:21 thing see it's identified all the these
1:45:26 different oh cool it's got a silver fox
1:45:30 vibe to
1:45:34 it they're not exactly on her coat there
1:45:37 they are there's two of
1:45:41 her let's see let's make her how can we
1:45:45 make her more interesting
1:45:51 [Music]
1:45:57 oh there's a brainstorm button here
1:45:59 let's hit the brainstorm
1:46:04 button brainstorm is pretty cool
1:46:08 too so brainstorm is going to come up
1:46:10 with different things all right stained
1:46:12 glass
1:46:13 inspired lighting casting Kaleidoscope
1:46:16 patterns intricate butterfly patterns
1:46:18 reflecting art Nuvo all right let's try
1:46:22 that one
1:46:24 and then let's go back to brainstorm and
1:46:26 there's climp we'll do some
1:46:29 climp oh wait oh this is clamped okay so
1:46:33 let's go back to
1:46:37 brainstorm we'll do this
1:46:43 one black bar there you go all right now
1:46:48 let's let's just play with our so I'm
1:46:51 going to go show more so if I say show
1:46:55 more show more show more show more it's
1:46:58 going to take this exact
1:47:00 prompt and just make a bunch more and
1:47:05 then what I'm going to do is I'm going
1:47:06 to say instead of an old woman we're GNA
1:47:08 say an ancient spirit it's going to do
1:47:11 that and now we're going to say a
1:47:12 mystical Shaman and now we're going to
1:47:15 say a Celestial Elder and now we're
1:47:17 going to say an ethereal
1:47:19 sage and then we're going to go back to
1:47:22 an old woman and instead of an enchanted
1:47:25 forest we're going to
1:47:27 do a Mystic Grove and then we're going
1:47:30 to do an Arcane Woodland and then we're
1:47:32 going to do a [ __ ] Glenn I don't know
1:47:35 what that is and then we're going to do
1:47:36 a sorcerer's Thicket and then instead of
1:47:39 stained glass butterflies we're going to
1:47:41 do crystal
1:47:43 moths opalene
1:47:46 beetles jeweled
1:47:49 dragonflies and iridescent cicas and
1:47:52 it's going to be soft now it's going to
1:47:53 have a magical glow now it's going to
1:47:56 have ethereal luminescence now it's
1:47:57 going to have so remember before when
1:48:01 people told you you had to become prompt
1:48:03 Engineers no what if you just had to be
1:48:06 like Mad Libs
1:48:10 Masters right why
1:48:12 not why not just be [ __ ]
1:48:18 inspired and then it just like [ __ ]
1:48:21 just BL BL BL plant PL plant just blasts
1:48:25 out
1:48:26 images all right so here's the gusta
1:48:30 climp so here's wood block from Japan
1:48:33 here's Aztec geometric things here's
1:48:36 Celtic not
1:48:38 workor there's mauy tattoo art here's
1:48:41 Enchanted Forest here's a haunted
1:48:43 Marshland that's good let's go down to
1:48:45 this one instead of butterfly wings
1:48:47 let's do leaves in the wind and
1:48:49 dragonfly wings and Cosmic Galaxy and
1:48:53 old maps and instead of art newvo let's
1:48:56 do Cubist and let's do
1:48:59 Roco and let's do this and let's do that
1:49:03 now it is very possible that all the
1:49:05 [ __ ] I'm creating I'm gonna hate all of
1:49:07 it does it matter
1:49:11 no but I'm kind of learning I'm kind of
1:49:14 learning some cool stuff
1:49:19 here uh oh did I crash I think I crashed
1:49:25 what's going on oh my mouse could be
1:49:28 dying I think my mouse is dying wait or
1:49:31 am I crashing hang on Mouse is dying
1:49:35 good Lord the technical difficulties
1:49:42 tonight oh no I crashed page
1:49:47 unresponsive wait okay I broke it
1:49:54 here we go oh these are cool look at
1:50:05 these you digging this all
1:50:12 right Silver Fox hanging in by a thread
1:50:15 but not
1:50:18 sleeping oh F Glenn equals a fairy Glenn
1:50:21 cool I need to go to
1:50:25 bed those mad libs are moving at the
1:50:28 speed of your brain yeah exactly if you
1:50:30 have ADHD this is a tool that you're
1:50:32 going to dig I
1:50:34 love the beetle these Beetle things
1:50:56 all right
1:50:58 let's all right how we going to do this
1:51:00 okay okay so here's what we're gonna
1:51:03 do let's let's go is this the beginning
1:51:07 oh it is okay so do do we like anything
1:51:14 here that one's kind of interesting
1:51:27 all right so let's go let's start with
1:51:29 her so we're going to say make a movie
1:51:32 so notice she becomes the first
1:51:35 frame and then let's have her transform
1:51:48 into this one oops
1:51:53 make a movie so that becomes our second
1:51:55 frame right so now I'm going to hit
1:51:58 generate
1:52:00 movie so that's going to do its
1:52:03 thing and so now I
1:52:06 say let me pick the second frame that
1:52:09 now becomes our first
1:52:11 frame right and so let's go
1:52:14 find another version of her here she is
1:52:17 with a
1:52:19 beard let's go with this one because
1:52:21 it's got lots of butterflies in the air
1:52:23 all right we'll do that all right and
1:52:27 then we'll go grab her with the
1:52:30 butterflies in the
1:52:39 air they're insisting on putting a nice
1:52:42 frothy beard on her look look this kind
1:52:45 of looks like Gemini bear so we'll get
1:52:47 like Gemini L Gemini bear
1:52:50 lady all right there she is
1:52:55 boom we got to get to the
1:53:02 uh to the beetle the beetle ones are so
1:53:10 cool yeah here let's let's
1:53:13 do all right she's
1:53:16 praying let's go to this one
1:53:24 come
1:53:25 on you can do it you can do it Chrome
1:53:29 don't crash on me
1:53:30 [Laughter]
1:53:34 yet waiting your
1:53:36 [Music]
1:53:45 turn that's a cool one
1:53:53 so which one did we do this top one so
1:53:55 we'll do that as the first
1:53:58 frame and then we'll
1:54:02 go this
1:54:05 one all right so let's let it catch up
1:54:09 now any questions this making
1:54:14 sense and then basically what this is
1:54:17 going to allow us to
1:54:18 do
1:54:20 oops hang on right
1:54:25 what this is going to allow us to
1:54:27 do is Stitch all these videos back to
1:54:34 back it's going to allow us to stitch
1:54:37 all these videos back to back and it'll
1:54:39 just be like one
1:54:41 flowing transform into another Isn't
1:54:44 that
1:54:50 cool all right
1:54:53 bees for the people make sense great
1:54:56 what's the blat best platform to watch
1:54:59 you on the one that doesn't crash I'm
1:55:03 guessing YouTube probably but I would
1:55:06 say Tik tok's more
1:55:09 intimate YouTube you can see [ __ ]
1:55:11 better um I don't know how X is and I
1:55:14 don't know how LinkedIn
1:55:19 is all right
1:55:23 let me go back and share a little a
1:55:25 little more
1:55:26 pictures all right there's some of our
1:55:30 videos so let's see how bad they came
1:55:32 [Laughter]
1:55:37 out oh that one
1:55:41 beautiful
1:55:51 nice I feel very intimate on Tik
1:56:06 Tok yeah this is sort of cool and now we
1:56:10 could you
1:56:11 know these are kind of working oh love
1:56:14 that
1:56:15 one love that one
1:56:33 very mystical very
1:56:38 magical I think that's it okay so now
1:56:42 you could
1:56:45 go I think you know what we could
1:56:47 do let's grab one of these images
1:56:54 that we
1:57:04 used this one so let's download that
1:57:07 image and then let's go to
1:57:10 sunno and I think in sunno you can
1:57:13 actually
1:57:14 upload an image can't
1:57:18 you
1:57:21 custom upload
1:57:23 audio I thought there was a way that you
1:57:25 could use
1:57:27 images with
1:57:34 sunno let's see create
1:57:36 explore soes anybody know how to do
1:57:39 images in sunno oh I'm not sharing sunno
1:57:41 hang on
1:58:05 custom
1:58:08 Persona
1:58:13 anybody is isn't there an image upload
1:58:17 thing where soon will just write a song
1:58:19 based on your image
1:58:27 am I forgetting that or like am I
1:58:28 hallucinating
1:58:31 that what did I ask no clue I thought
1:58:34 there was a thing in sunno that you
1:58:36 could you could upload an image hang
1:58:40 on
1:58:44 um
1:58:45 sunno image
1:58:49 upload create music with an image take a
1:58:52 picture use a photo to create a 30-
1:58:55 second song with sunos
1:58:58 scenes here's how to use sunos
1:59:02 scenes click on create custom is on
1:59:04 select
1:59:08 camera to get your oh it's only
1:59:13 in okay it's only in the
1:59:17 app got it all right now it's in the app
1:59:21 well it might be 3.5 hang on that's a
1:59:25 good thought Brandon maybe 3.5 go back
1:59:28 to
1:59:31 create
1:59:33 3.5 upload
1:59:38 audio Yeah it's not here it's only in
1:59:40 the
1:59:42 app all right whatever let's do we're
1:59:44 going to go
1:59:46 instrumental um
1:59:50 ethereal fairy
1:59:56 music from an enchanted
2:00:00 forest let's just see how it
2:00:04 does oh I'm doing this and I'm sharing
2:00:06 the wrong tab of course because I'm a
2:00:09 lose a baby why don't you sue
2:00:12 me oh yeah he's an
2:00:16 [ __ ] suo create is this it this is it
2:00:23 [Music]
2:00:30 so um let's see
2:00:33 strings
2:00:35 and mystical
2:00:45 synths who forgot to put the shot collar
2:00:48 on him shut
2:00:49 [Music]
2:00:50 up this well welcome to Chad
2:00:53 [Laughter]
2:00:58 [Music]
2:01:07 add let's see um let's let's call it
2:01:12 sweeping
2:01:15 sweeping ethereal music
2:01:23 okay while that's doing that well we we
2:01:25 want some music let's just get something
2:01:27 that's decent shall we oh I'm in Ah
2:01:30 that's what it is I'm in 3.0 3.5 let's
2:01:33 create a couple in 4.0 these will be
2:01:37 better uh archetypal architect I'm high
2:01:40 enough to be following along it's a
2:01:42 great place to
2:01:42 [Laughter]
2:01:48 be the the AI learning lab we being
2:01:51 stoned is encouraged and an
2:01:54 [Laughter]
2:01:58 [Music]
2:02:07 advantage me
2:02:15 [Music]
2:02:42 being these are pretty but why am I not
2:02:45 getting strings let's see
2:02:48 orchestral
2:02:51 strings
2:02:53 for an
2:02:56 enchanted
2:03:05 forest long strings something strings
2:03:08 this keeps giving me keyboard keeps
2:03:11 giving me pner I don't want
2:03:13 pner I want Fairy music
2:03:17 [Music]
2:03:23 what the
2:03:29 [ __ ] um let's do
2:03:32 custom and I think I isn't there a uh
2:03:35 let's
2:03:36 see trap Funk pop ambient
2:03:41 IDM and traditional
2:03:45 folk ENT
2:03:47 IDM
2:03:50 sarangi um
2:03:57 orchestral
2:04:00 strings
2:04:05 and let's see
2:04:07 and
2:04:09 minimalist violins
2:04:23 exclude Styles
2:04:26 piano let's try that
2:04:31 boom
2:04:32 Pio here we go
2:04:37 [Music]
2:04:52 [Music]
2:05:15 pting Coral voices I'm going to say
2:05:18 style
2:05:21 of songs from liquid
2:05:26 days it seems to be okay with song names
2:05:30 it it's not okay with artist names but
2:05:32 it seems to be okay with album names
2:05:35 we'll see
2:05:37 [Music]
2:05:58 [Music]
2:06:39 this is fine let's grab this
2:06:42 one um download MP3 video okay so we
2:06:45 have a
2:06:46 song now we'll go back over here
2:06:59 [Music]
2:07:04 all right let's
2:07:10 see like that one download
2:07:28 think that one was good Tik Tock pin
2:07:32 when you get a new phone try all the
2:07:34 ring tones okay are they cool that's
2:07:40 grovy that one's kind of beautiful that
2:07:43 one's more beautiful I like that one all
2:07:45 right download
2:07:49 that uh that one's a
2:07:55 mess but she emerges from that butterfly
2:07:58 so that's
2:08:01 good
2:08:06 no that was kind of cool that's kind of
2:08:14 spooky all right that's it that's all we
2:08:16 got so
2:08:23 oh you're not I'm you're not seeing all
2:08:24 this that doesn't matter hang on let
2:08:26 me let me share my whole
2:08:30 screen
2:08:38 yeah oh trying ringtones me with sunno
2:08:41 right now yeah that's a good point okay
2:08:44 uh let's see uh do I have it open no so
2:08:47 if I go like
2:08:48 this and I go iMovie
2:08:59 black bar bring up the black bar Mr
2:09:04 Shannon fantastic Bob tell him what he's
2:09:07 won only hasn't won anything we're not
2:09:09 playing a
2:09:12 game okay we're going to do a new
2:09:15 project we're going to add media time
2:09:17 check 1010 I know we're late but now I
2:09:20 got to make a movie it's just has to
2:09:32 happen are these in order let's see
2:09:34 we'll drop this in first bom
2:09:38 boom no we didn't okay now we
2:09:41 did boom there's
2:09:44 that then we'll drop in
2:09:48 that right
2:09:57 wait why is she not
2:10:02 moving oh because it's a
2:10:05 picture is this where we
2:10:16 start I'm so confused hang on a sec I
2:10:19 got to DEC confuse myself all right
2:10:21 right so the first one is okay it's that
2:10:25 one so it's that one okay so this can go
2:10:31 away this can go away and then we
2:10:35 go this
2:10:37 one and then this
2:10:40 one right and then
2:11:08 y all right that's it okay so then we go
2:11:13 rant and then we grab this and we go all
2:11:16 right let me play it
2:11:23 [Music]
2:11:30 that's
2:11:32 [Music]
2:11:42 horrible it's
2:11:46 awful the music's horrible the
2:11:49 animations are horrible the is janky
2:11:53 [Music]
2:11:57 but [ __ ] it let's put it out in the
2:12:02 world um fairy lady
2:12:05 fairy
2:12:15 lady all right let's go to the Twitter
2:12:22 let's go to close this let's go to home
2:12:26 and
2:12:30 then okay so tomorrow there's no live
2:12:33 I'm not going to be going live tomorrow
2:12:34 have a happy New Year what else um New
2:12:38 Year's resolution we should think about
2:12:41 what we want to do let me stop sharing
2:12:43 that let me start sharing X so you can
2:12:46 hear it and then we're going to go um
2:12:52 while this is janky and
2:12:58 aad boring it
2:13:01 was made
2:13:04 live
2:13:08 tonight with a high dose of
2:13:15 ADHD yay AI
2:13:22 there's Mr Beast all right let's uh
2:13:25 let's go get our little video little
2:13:27 VI fairy
2:13:31 lady all right here she is bang boom
2:13:36 bang and then I'm going to go uh hashtag
2:13:40 we'll do
2:13:41 at um
2:13:45 Luma Luma
2:13:47 AI at uh
2:13:51 let's see
2:13:53 used chat GPT
2:13:58 Luma um what are those things called
2:14:02 boards and
2:14:04 sunno yay AI
2:14:12 suoo soono music and then at chat GPT
2:14:22 I think that's it yeah and then we'll
2:14:24 tag Ki who's the head of the creative
2:14:27 program over there Bang all
2:14:31 right let's go 45 okay we're
2:14:35 uploading okay last time this year black
2:14:38 bar all right there's your black bar
2:14:40 people so you can see
2:14:42 everything um
2:14:53 an Murphy's in the house hey Murphy with
2:15:00 shaking we were going to do uh we were
2:15:03 going to do some notebook LM on all of
2:15:04 our cool sessions and but the the
2:15:07 transcripts weren't up on YouTube
2:15:10 yet all right
2:15:13 post your post has been posted all right
2:15:16 so go go find my
2:15:19 Twitter Kyle Shannon and then give this
2:15:22 some love share it up share it up here
2:15:26 you want to watch it
2:15:28 [Music]
2:15:56 [Music]
2:16:09 all right let me get rid of this all
2:16:12 right do me a favor everybody subscribe
2:16:13 to You YouTube if you're on Tik Tok say
2:16:17 hi to Ann Murphy she's in the house
2:16:21 [Music]
2:16:24 thank her for AI festiv
2:16:27 us AI for the rest of us so like what we
2:16:31 did tonight some of those
2:16:33 techniques I didn't really have th that
2:16:37 combination of techniques in my head to
2:16:39 be able to do that that clearly but that
2:16:41 was because of CJ and Kelly Bosch and a
2:16:45 couple other people that did things that
2:16:46 I was like oh let me do that and that's
2:16:49 how we could do that so anyway all right
2:16:53 I'm out of here guys Happy New Year see
2:16:54 you on the other side all right
2:16:57 everybody thank you Happy New Year
2:16:59 everyone thanks for everything you do
2:17:01 thanks for hanging out here and I will
2:17:03 see you uh in the new year so tomorrow
2:17:07 nights off we'll be back New Year's Day
2:17:09 at 800 PM mountain time all right
2:17:12 subscribe to
2:17:14 YouTube thank you an Murphy
2:17:19 beautiful all right everybody
2:17:21 um peace out let me get rid of the live
2:17:25 stream on YouTube bye everyone happy New
2:17:29 Year