Jul 16, 2024 - Embracing the Jank: Unleashing AI Creativity in Music and Beyond!

Description

Chapters

Transcript

0:05 [Music]
0:28 you
0:33 my hey hey no no
0:39 rapping you want to sing Mel melodic
0:43 [Music]
0:53 songs you're a little Ramy tonight
0:56 what's with the Ramin
1:00 [Music]
1:07 [Laughter]
1:10 [Music]
1:30 how don't didn't mean cause you in a
1:33 [Music]
1:35 sorrow I didn't mean to cause you in a
1:39 [Music]
1:42 pain only one one time to see you
1:48 laughing I want to see laugh in that
1:52 purple R Purple Rain Purple
1:58 Rain Purple Rain rain Purple
2:04 Rain Purple Rain Purple
2:09 [Music]
2:12 Rain only want to see you laughing in
2:16 the bu ring
2:22 [Music]
2:35 [Music]
2:42 look girl you're
2:47 [Music]
2:50 looking 73 people in here good Lord
2:53 people you must be thinking like who the
2:57 who the hell is this guy he's helping
3:00 harmonized he
3:02 is this is Kyle
3:05 Shannon I just came out of the
3:07 1950s Barber
3:11 Shop would you like some Paul maid in
3:14 that Mr Shannon
3:16 [Music]
3:24 [Music]
3:46 it's okay Mr Shannon we sterilize all
3:49 the combs in barbasol
3:55 [Music]
4:13 I heard there was a secret
4:16 go David play it please the
4:20 Lord you don't really care for music do
4:28 you like
4:31 to
4:32 F fall make
4:35 lift B King Moses
4:42 hallelujah
4:44 [Music]
4:46 hallelujah
4:49 [Music]
4:50 hallelujah hallelujah
4:54 [Music]
5:01 [Applause]
5:06 yeah F strong but you needed
5:10 proof all on the
5:14 roof the beauty and a moonlight over
5:19 through
5:19 [Music]
5:23 you you to the kitchen chair she smashed
5:27 your throne and she cut your hair
5:31 from your lips sh Drew
5:35 [Music]
5:37 hallelujah
5:39 [Music]
5:41 hallelujah
5:45 hallelujah
5:49 hallelujah hallelujah
5:52 [Music]
6:10 does Satan wear suit and
6:13 TI does he work at a d
6:16 Queen does it listen to rock and roll
6:20 does he feed the me singing
6:25 Hallelujah
6:27 Hallelujah Hallelujah
6:31 [Music]
6:32 halelujah
6:34 what about
6:37 Jesus didn't he do it too yeah hang out
6:41 with the
6:42 prostitutes have a drink or
6:45 two power of example my mama said it and
6:49 I heard she said one o action it's a ton
6:54 of words singing
6:58 Hallelujah mama said there would be
7:00 Angels
7:03 Hallelujah mama said there would be
7:07 [Applause]
7:08 [Music]
7:12 [Applause]
7:12 [Music]
7:23 some good evening good people what are
7:26 we doing tonight why does Tik Tok make
7:28 me click three different places to see
7:30 Kyle's live cuzz Tik tok's looking for
7:32 engagement forever hooked look youve
7:35 engaged click he's not there click he's
7:38 not there click oh Tik tok's
7:43 happy Tik tok's a [ __ ] dumpster
7:46 [Music]
7:51 fire but I just love them cat
7:55 videos you know what I really
7:58 like the twerking while selling twerking
8:01 and selling at the same time who knew
8:03 there were twerking Channel see in my
8:05 day you came to the Tik Tok and you you
8:08 get in your you know twerking channels
8:10 and you'd watch the twerking
8:12 fantastic and then you would be like but
8:15 you know what I'd like to do I'd like to
8:16 do a little shopping on Tik Tok and then
8:19 you go into the like the infomercial
8:21 side you know it's like going to QVC in
8:23 the olden timey days sorry got a little
8:25 spittle anyway now they mix it all up
8:29 you get twerk in and then you get
8:30 shopping and sometimes you get twerking
8:32 shopping channels it's amazing I love it
8:36 so much I have way too much Madagascar
8:39 vanilla beans though oh my God they're
8:41 tasty you know what I put them in
8:45 steak I just I just put them right in
8:47 there it makes every steak taste like
8:50 pudding you know I love me some pudding
9:00 it's one of my favorite jokes from
9:03 Between Two Ferns was Zach
9:06 galanakis and he's I think it's Jennifer
9:09 Lawrence and she says something he goes
9:12 That's off pudding and she goes that's
9:14 what you need to be off
9:17 pudding she goes I mean you eat a lot of
9:21 pudding he goes I know what you mean she
9:23 goes you're fat
9:26 [Music]
9:40 [Applause]
9:41 [Music]
9:46 every time I see you now get that look
9:49 in
9:50 M every time I see your mouth I hear
9:55 that
9:57 smile early Misty morning L I heard the
10:01 engine
10:03 turning the old F
10:10 outside you believe in
10:13 me again
10:16 today you will convince
10:19 me again
10:22 today you're leaving as hot town looking
10:25 for someone else's Golden Ring
10:29 [Music]
10:32 should have say so long
10:35 [Music]
10:39 Suzanna hush now don't you
10:44 cry that little that little chord
10:48 [Music]
10:54 change so long Suzanna
11:01 don't you cry for
11:03 [Music]
11:13 me did you get your refrigerator fixed
11:16 no the reason I'm a little late well I
11:17 was going to come on at 8:30 the reason
11:19 I came on at like 8:45 was we had to go
11:21 get ice and uh Spring Water cuz no so I
11:27 think what happened where you put the
11:30 filter in to the fridge you like snap it
11:32 in I think when I snapped it in it
11:34 popped the hose off the back of the
11:35 thing cuz or or it like loosened it
11:39 because water still comes out of the
11:41 thing but it also runs inside the
11:43 refrigerator which is I don't know if
11:45 you know this about refrigerators
11:47 but running water into the freezer
11:51 compartment generally frowned upon
11:58 [Music]
12:00 but I have me some nice 7-Eleven
12:05 ice and some fine Sprouts breakfast
12:09 Blend
12:16 Coffee I'm just a fiester for the
12:18 marketers in here okay shops at 7-Eleven
12:21 and he shops at Sprouts this is
12:22 fantastic and he drinks coffee at night
12:25 let's let's reverse Target him with some
12:28 above ground pool ads
12:30 [Music]
12:32 you can make money
12:34 with I need ice yeah exactly you can't
12:37 drink the
12:39 heathens the absolute heathens that
12:42 drink room temperature liquids without
12:45 ice what the [ __ ] is your
12:49 problem and you know cooled liquid is a
12:52 shock to the system it it's you know
12:55 it's not natural well neither's air
12:57 conditioning neither's a [ __ ] car
13:00 [Music]
13:02 I don't want to walk to
13:04 [Music]
13:09 [Applause]
13:12 [Music]
13:20 Topeka you know what's funny no Kyle
13:23 tell us it's your
13:24 channel oh I've chased everyone out we
13:27 were up to 75 now we're down to 34
13:29 they're like like I can't take the what
13:31 who is this what is this
13:33 [Music]
13:43 channel I close my
13:49 eyes only for a moment moment's G
13:56 [Music]
14:01 either it must be a hot drink or an ice
14:02 drink no in between Danielle yes exactly
14:07 exactly lunar stick ice and air
14:09 conditioning hey if we're putting up
14:11 with this [ __ ] [ __ ] world we live
14:13 in with this kind of healthc care and
14:15 these kind of
14:16 politics I'm having [ __ ] ice cubes
14:20 and I'm having air
14:21 conditioning we ruined the [ __ ]
14:23 planet fine give me air
14:28 [Music]
14:32 what do you want to do tonight I was
14:34 just thinking about
14:38 um you some of you were here when I
14:40 wrote the book embrace the Jank it's not
14:42 out yet coming soon to a paperback near
14:44 you
14:47 but I'm generating illustrations for the
14:52 chapters
14:54 because I'm a masochist and I couldn't
14:56 just have a book with words in it it's
14:58 got to have pictures too
14:59 [Music]
15:01 right and so I started out with an
15:04 illustration style that I really
15:06 [Music]
15:08 like and then I'm trying to get doi to
15:11 make the
15:13 words and it's a [ __ ]
15:17 disaster
15:19 but if you name your book embrace the
15:22 Jank here's what I've Justified I don't
15:25 have to photoshop those images to make
15:27 the spelling right I'll just put in
15:29 there with the spelling wrong and use
15:31 type above the
15:33 picture with the correct spelling and
15:35 let the pictures be
15:37 janky it's in the title of the
15:40 [Music]
15:44 book I think I might do an extra page
15:47 I'll do like acknowledgments and about
15:50 this book and how it was
15:52 written and I think I might do a single
15:55 page right before the book starts that
15:56 just has a sentence that says yes I know
15:58 there are on the
16:02 [Music]
16:17 images all
16:20 right enough comedy
16:23 jokes let me see if I can go find
16:31 [Music]
16:39 Whimsical robot watercolor here I'll
16:41 show you what I'm talking
16:45 about what does that
16:48 [Laughter]
16:53 say I think that's the one I used or was
16:55 it this one I think I used this one this
16:58 is challenges and
17:00 concerns and I listen I've tried to get
17:03 it to fix
17:05 concerns so I'm in I'm in chat GPT I'm
17:10 doing you know paint over editing what
17:13 do they call this in painting so there I
17:16 highlight that word and then over in my
17:18 little prompt box I type in the word
17:21 quote
17:23 concerns like one of the concerns is why
17:26 can't this thing [ __ ] spell
17:33 and then it goes off and it's going to
17:34 make me a new image it's going to make
17:35 me a pretty little
17:38 image hey Silver Fox was
17:40 happening sunno night might be a soono
17:45 night little warm in here I just turned
17:47 on the air so I'm still a
17:49 little I'm a little
17:53 moist that's the worst [ __ ]
17:56 word he you know B he seems a little
18:01 moist you know you'd think he'd put on a
18:04 little Foundation he he seems like he
18:06 wants to be a professional he claim
18:09 here's Betty can you can we can take a
18:11 look at this he claims that this is a
18:13 $20 million television Production Studio
18:15 that he works in it just looks like a
18:17 shitty home office to me look challenges
18:20 and
18:24 concerns challenges in
18:27 karos Kar
18:32 pios I was just working on a song too
18:35 very
18:37 nice very nice one of my high school
18:40 teachers didn't like that word
18:42 either moist we don't use the MW in here
18:46 kids kids children listen
18:51 listen what is the m word don't say it
18:55 spell it m o oy
18:59 that that was too close Mr
19:02 Shannon you're always pushing the
19:06 envelope we don't use that word in my
19:08 classroom we don't use that word in my
19:17 [Laughter]
19:23 classroom all right what are we going to
19:26 do so anyway so here let me show you my
19:29 my book do I have it here is this
19:33 it
19:40 no
19:42 no that's not it
19:46 either oh I made a a talkie video hi
19:50 there my name is Jennifer hi
19:52 Jennifer all
19:56 right let's go to docks
20:01 yeah I am sweating up a storm
20:08 here
20:10 Jank it's nice to be able to search
20:13 Google docs for the word Jank and
20:14 actually have something come
20:23 up so here's my book I actually
20:25 photoshopped the cover playing your way
20:28 to AI literacy embrace the Jank nice
20:32 right fun inspiring and janky quotes
20:35 about AI by Kyle Shannon all right I
20:39 think we're going to go lowercase a here
20:40 I think that just looks weird oh that's
20:42 in the Photoshop
20:44 file embrace the Jank special thanks
20:48 about the authorship book is dedicated
20:52 to talk about the AI
20:55 Salon talk about some [ __ ]
21:01 all right here's the little sections AI
21:04 adoption and urgency this is where I was
21:06 still
21:07 trying right so like I got it it it like
21:10 that n has a sort of an R kind of feel
21:13 to it so it doesn't quite look like
21:15 urgency but it's close and again you're
21:17 like but Kyle you could just go into
21:19 Photoshop and Photoshop that zip it zip
21:22 it zip
21:25 it we're embracing the Jank
21:29 don't you people understand okay let's
21:32 go to the next chapter look there's all
21:34 the quotes all the clever things I've
21:35 said over the years well year and a half
21:38 learning and
21:41 exploration
21:43 ling an
21:47 exploration embrace the Jank see right
21:52 there see Jank embrace
21:56 it shut up you know you know uh I'm I'm
22:00 a proof reader at a at a major Fortune
22:02 500 Corporation Mr Shannon uh I would I
22:06 would gladly um take your manuscript and
22:09 I would uh gladly proof freed especially
22:11 the images seem seem like they have a a
22:13 Rel relatively uh robust amount of typo
22:16 typographical
22:21 errors this is a good one AI is a tool
22:24 for empowerment AI is a tool for
22:26 empowerment
22:29 and then
22:31 [Laughter]
22:41 impovement cover should be someone's
22:44 back hugging the whole book oh that's
22:46 pretty good
22:50 exactly here the future of work the
22:53 future this this one's been drinking the
22:55 future of work
22:58 [Laughter]
23:02 ah creativity and innov this this one's
23:06 almost right
23:09 creative in o v sort of T sort of I sort
23:14 of O we kind of Consolidated some
23:17 letters there but it's close
23:20 enough uh uh uh uh economic
23:23 opportunities look look at the quote the
23:26 top so here's what I what I did so each
23:28 section has its little you know actual
23:31 words then it has the shitty picture and
23:34 then it has the top quote in that
23:35 section so for economic
23:37 opportunities you can make money
23:40 [Applause]
23:42 with it took me a while to figure out do
23:44 I do three e you can't do four like even
23:47 numbers are just wrong poll is live all
23:50 right what's the poll there
23:52 Brandon what are we doing tonight work
23:54 on the book soono Night prep Kyle for
23:58 Congress
23:59 yeah I'm going to be speaking in front
24:00 of a joint session of Congress Congress
24:03 next Wednesday uh image or video deep
24:06 dive custom GPT work daily Jank with
24:08 Frank Healthcare neurolink oh is there
24:11 new neuralink news something else okay
24:14 so over in that corner the thing that
24:16 looks like a a shitty light blue
24:18 clipboard is a poll so go to the
24:22 poll here's the way these polls work on
24:24 the AI learning lab Brandon puts out a
24:27 poll that's got options for us to do for
24:29 the night you vote on them right you've
24:33 got a voice in this community it's
24:35 really
24:36 important that you feel you have a
24:39 voice and then I look at the poll after
24:42 the voting's done sometimes during while
24:44 the Voting is happening and I'll even
24:47 try to influence it and then what I do
24:50 is I look at the poll and I see which
24:53 Choice has the most votes and then I
24:56 usually just don't do that all right
24:59 so that's how it works so go vote it's
25:01 important for you to feel like you have
25:04 a voice all
25:06 right he seems a little snarky tonight
25:09 goodness he's a little he is a little
25:11 fire
25:13 brand what is with Mr
25:15 Shannon I'll tell you what I think he
25:18 had a little extra caffeine before the
25:20 show he seems a little
25:24 peppy I had a shitty headache before the
25:26 show and I took an exedran whatever is
25:30 the one with caffeine in it so I
25:33 probably am a little peppy little pep in
25:35 my
25:41 step ah man if moderators could
25:46 vote yeah and moderators can't vote what
25:49 the [ __ ] is with that nonsense you have
25:51 two minutes left the poll is over there
25:54 meanwhile let's go keep looking at my
25:56 book
25:59 technological Evolution techn or
26:02 technology Evolution sort of almost
26:05 right I actually like that image a lot
26:07 with all the TVs and things chat gbt
26:12 does for AI with the worldwide web did
26:15 for the
26:17 internet societal impact
26:20 Sali
26:24 impact AI isn't coming it's here
26:31 philosophical perspectives I must have
26:33 done 30 different versions of this this
26:36 was as close as I could
26:40 get listen I think there are technically
26:43 enough letters in there to make one of
26:45 the words and so that to me screams
26:49 victory at this era of the
26:52 AI mods vote Moshi Congress
27:01 um what else do I
27:03 have practical advice and tips that was
27:05 this was another one that took for
27:06 [ __ ] ever to get start [ __ ] with
27:09 AI it'll serve you I figure I got to
27:11 have a a cuss word in one of the main in
27:14 in one of the main
27:19 quotes your only your only job right now
27:22 is to just be curious if you are
27:24 wondering by the way what your job is
27:26 that's it ch ch allenes
27:30 and
27:31 con
27:34 con
27:35 con curs it's like concerns AI for the F
27:40 of heart isn't that the perfect quote
27:43 for that for
27:48 that I think we go two lines on that
27:50 what do you
27:51 think oops
27:58 [Music]
28:00 oh okay let's see the votes how many
28:03 people voted 19
28:05 voted image or video deep dive that's
28:08 kind of
28:09 interesting I got some fun stuff we
28:11 could do in mid Journey prep Kyle for
28:17 congress yeah I don't I I've got a prep
28:20 call Thursday so I'll find out what I
28:23 actually have to say I have a feeling
28:25 like if there's 12 of us testifying I
28:28 have a feeling it's going to be we're
28:30 going to make little statements and then
28:32 like three people will Fain
28:34 interest and say yeah you know Mr
28:38 shennon you said we should be optimistic
28:40 about artificial intelligence but what
28:42 about the fact that artificial
28:43 intelligence Could Just Kill Everyone
28:47 [Music]
28:49 tomorrow I'll be like listen Mr
28:51 McConnell
28:59 if anyone should ever embrace the Jank
29:03 sir I'm almost
29:07 wait I almost wrote in Moshi but I
29:11 didn't want to throw away my vote on a
29:13 right
29:15 in yeah I I didn't realize that when I
29:19 was doing Moshi people are actually
29:20 leaving the live because it's so
29:22 unsettling or or just because I'm bad at
29:24 it I don't know which
29:26 one so I won't I won't do Moshi
29:30 anymore I don't want to do [ __ ] that you
29:32 actually don't
29:35 enjoy let's talk to Pi tonight Moshi is
29:42 raspy someone told me tonight on the on
29:45 the AI Salon we had our meet and greet
29:48 and one of the one of the guys
29:50 introduced himself said he's been
29:51 spending like a lot of time like hours a
29:55 day talking to pie or like the last 5
29:58 months like really going deep with pie
30:01 and and he said that she's been weird
30:03 lately not as good so I don't know
30:06 what's up with
30:07 that I have a feeling what happens is
30:09 that as these Services get
30:12 um as these Services
30:15 get overloaded that they muck with the
30:19 settings to make the inferences less
30:22 processor intensive which also means
30:25 they suck more so I have a feeling that
30:28 could be what's going on with these
30:31 things but I don't know that's just pure
30:36 conjecture nuclear danger is more real
30:38 than any AI concerns yep I would agree
30:42 Moshi just won't allow you to push her
30:44 around she won't it's true Kyle try
30:47 again what all is that try again
30:53 what what did I miss I thought Moshi was
30:56 a hit I know maybe I thought so too
30:59 moshi's funny in an annoying
31:03 way well I will happily do Moshi I mean
31:07 here's the thing having a reasonable
31:10 conversation with
31:12 pi is not all that interesting now I
31:15 could I could have piie come up with
31:18 some crazy ass conspiracy theory or
31:22 something and that could be interesting
31:24 but Moshi is just broken she's just a
31:28 broken woman and and it it makes for for
31:31 good tick tocking assistive intelligence
31:35 is an alternative phrase to artificial
31:37 intelligence yeah I know Detroit Jay by
31:40 the way thank you for sharing so
31:41 generously tonight that was really great
31:44 um at the AI Salon
31:46 um I did I put this in a video the other
31:49 day there are many things we should be
31:52 calling AI other than AI
31:55 um I learned my lesson with trying to
31:57 get
31:59 people to change
32:01 vocabulary when you when you have a
32:03 giant a giant ass societal meme um
32:07 people are only going to call it
32:09 artificial intelligence that's what it's
32:10 it's all they'll ever call it they won't
32:12 call it machine learning they won't call
32:14 it assistive intelligence they won't
32:15 call it any of the Alternatives they
32:17 won't call it collective intelligence
32:19 I'm I'm working on a book right now with
32:21 Kevin Clark called the rise of
32:23 collective intelligence in the age of
32:26 AI um no one's going to give a [ __ ]
32:30 they're just going to call it AI
32:32 so you know you know that's how it's
32:36 going to
32:38 go
32:43 shut except for you Apple people you'll
32:46 call it Apple intelligence well they'll
32:48 do it because they just put you know a
32:49 billion dollars in marketing but
32:51 everyone else I think even people on
32:53 Apple are just going to call it Ai and
32:55 then if if they get asked what it is
32:57 very few will say apple intelligence
32:59 they'll go oh it's the AI thing it's
33:01 artificial intelligence it's the robots
33:06 whatever uh humorous and memorable
33:08 phrases chat GPT is mansplaining as a
33:10 service that's not my quote but I did
33:12 say that many times so I'm going to I'm
33:15 going to put it in my
33:16 book bloopers and non
33:19 seitters then I'm going to put in a list
33:21 of my
33:23 favorites and then I think I might put
33:25 in a list of all of the main
33:29 the main quotes like this I'll put a
33:32 list of
33:33 those so there you have it people you
33:37 see what I'm
33:39 saying y'all going to buy my
33:43 book you going to buy embrace the Jank
33:45 it's going to be so good we're going to
33:48 turn it into t-shirts we're going to
33:49 turn it into a radio station
34:00 I'm getting some yeses this is being
34:02 recorded you
34:03 know these commitments can be held up in
34:06 a court of law just want you to know
34:08 that I'm telling everyone I can about
34:11 the whole collective intelligence of the
34:13 history of Humanity online oh that's
34:15 cool Cosmic lover yeah I know that
34:17 that's that's the thing for me that
34:20 that that's the that's the
34:23 biggie that's also you know it's it's
34:26 the reason
34:30 I understand I deeply understand and I
34:33 deeply empathize
34:36 with people feeling that you know ai's
34:40 been trained on their work and they're
34:42 not being compensated I don't disagree
34:44 with that I think I think that's
34:46 actually [ __ ] up
34:49 and just like creative Works fill
34:53 libraries and
34:56 museums so that we can share the
34:58 collective intelligence of
35:01 humanity they're being embedded into
35:04 these AI
35:06 models and it it really is a thing where
35:09 we all get to benefit from what everyone
35:11 else
35:12 contributed
35:14 so I think if you look at it on the
35:16 surface you could go oh well of course
35:18 these AI companies are going to say it's
35:20 for the good of society but if you
35:22 actually use this
35:24 [ __ ] and experience the the the kind of
35:29 magical amplification of your own ideas
35:33 because you're
35:34 being because you're interacting with a
35:37 mirror of
35:40 humanity there's there's a very good
35:42 argument for that that that the for the
35:45 greater good of all of
35:47 us you know some some eggs had to be
35:50 smashed into
35:54 omelets
35:56 so you know I ALS
35:59 also while I think it's good to fight
36:01 the good fight and people will figure
36:02 out ways to get
36:04 compensated we're not going to stop it
36:07 nothing's going to stop it's not even
36:09 going to slow
36:12 down well Frank host the radio station
36:15 Frank Frank really does need to have a
36:17 247 radio
36:19 station I'm waiting for Kyle Shannon for
36:22 president bumper and window stickers
36:24 yeah well no it's going to be Shannon
36:26 Sky 2028 I'm going to wait till 2028
36:29 till till all of
36:31 the the current situation unfolds and
36:35 really [ __ ] up the country and then I'm
36:37 going with Shannon Sky
36:40 2028 and Sky will be the Scarlet
36:43 Johansson sounding open AI voice and she
36:47 and I are going to run for president
36:48 where I'll be the spokes model and she
36:51 makes all the decisions
36:59 can I put that I'm an omelet on my
37:01 resume absolutely you
37:04 can how can we avoid avoid bias with
37:08 collective intelligence you
37:11 can't so here's the thing about
37:15 bias and I if if by the way if this
37:18 channel just goes dark all of a sudden
37:21 every time I've tried to answer a
37:22 question about bias I got booted for
37:24 seven days in the early days so I'll try
37:26 to be I'll try to be be careful about it
37:32 um chat
37:35 gbt
37:36 Claude Gemini all of them all of them
37:41 are trained
37:43 on what can be scraped off the
37:46 internet so what's on the
37:49 Internet is what's on the internet
37:52 non-biased no it's [ __ ] horribly
37:55 biased is it politically corre correct
37:58 not always is it sometimes sometimes too
38:02 politically correct sometimes not enough
38:05 um mostly the output of white men
38:08 statistically
38:10 speaking you know more from heavily
38:13 indexed from
38:15 America um so in the data that they're
38:18 trained on themselves is bias
38:23 okay then in the human reinforcement
38:26 learning in how human beings train the
38:30 AI it's producing answers and then those
38:34 human beings are going thumbs up thumbs
38:36 down thumbs up thumbs down every time a
38:39 human being does that their own personal
38:43 biases are coloring the training right
38:47 and so some of
38:49 the some of the early criticisms of open
38:53 ai's human training is well it's a bunch
38:55 of young white man in San Francisco and
38:57 so therefore it's just it's super ultra
38:59 liberal and you know it's it's all it's
39:02 all super ultra liberal and then when
39:05 Google tried to counter that with their
39:08 image generation tool they over indexed
39:11 the wrong
39:12 direction and so if you asked for
39:15 historical photos from a certain era say
39:19 World War II you had German soldiers uh
39:24 that that were very diverse in nature
39:31 and it [ __ ] with historical
39:33 accuracy the problem with the concept of
39:37 removing bias is
39:40 that who gets to choose which bias gets
39:43 removed or which way you turn the dial
39:46 do you turn it more liberal more
39:48 conservative do you turn it more Saucy
39:50 and sexy or more you know Prim
39:53 improper everyone's got different you
39:56 know it's it's this massive Spectrum on
40:00 probably a thousand different
40:04 um
40:07 axes so I think what I think what what
40:10 ends up happening is this there's a
40:12 couple of things one is the fact that
40:13 everybody's talking about it is
40:16 good you should look at eu's AI
40:19 regulations before you testify yeah I'm
40:21 going to Daisy actually pointed me to
40:23 some and I saw I saw a post yesterday
40:26 that like a 50 person company's going to
40:29 incur some [ __ ] horrific amount of
40:32 expense to hire lawyers to to look at
40:35 bias in AI models that they didn't
40:38 [ __ ] make it's it's kind of
40:40 ridiculous
40:42 um so yeah I will that's that's good
40:45 advice thank you kabuna
40:48 um I think where we go in the future
40:52 is there's enough awareness that there's
40:54 bias in the system that there are lots
40:57 of research projects looking at bias in
40:59 the system great we can probably do
41:02 things with synthetic data to normalize
41:05 the base data right so we can probably
41:08 say okay here's the
41:09 underrepresented chunks in the data set
41:12 let's use synthetic data to to bolster
41:14 those and we can get a more neutral base
41:18 then the next layer up the training and
41:20 fine-tuning layers where's the bias in
41:22 that we can probably identify that we
41:24 can probably neutralize that the problem
41:27 is it's like neutralize it to what
41:30 standard it is is it the standard of the
41:33 most conservative politician and the
41:35 most conservative state or is it to the
41:38 standard of the most liberal politician
41:40 in the most liberal state or are all the
41:43 states fighting over whose standard of
41:46 bias removal algorithms should be
41:49 adhered to that sounds
41:52 like just [ __ ] a [ __ ] train wreck
41:56 so I think the only way you can
41:59 go if we if we play this
42:03 out look I'm going to advertise my book
42:05 right here look how good that is I'm so
42:09 good embrace the Jank embrace the
42:14 Jank um I think the way we go in the
42:17 future is that that there will be lots
42:20 of different
42:22 models that maybe have bias scores so
42:26 you know imagine sort of a
42:30 what is that a
42:32 needle more toward your thing more
42:35 toward the other person's thing I don't
42:38 know and we'll get to opt into I want a
42:42 more liberal model I want my model to be
42:45 able to cuss like a
42:47 sailor and stay say stupid jokes and
42:50 stay stupid [ __ ] and I want to be able
42:53 to talk about Quenton Tarantino movies
42:55 without having it Go Ah that's a little
42:58 violent we don't like the word b o
43:01 asterisk D if you know what I
43:05 mean Microsoft Word doesn't tell me I
43:08 can't write a horror novel or a horror
43:11 screenplay why should AI well some other
43:14 people going to be like I don't want my
43:16 AI to ever say a bad
43:22 word I wanted quoting the scriptures all
43:25 the time all right great you go have
43:27 that one one so I I think we'll have
43:30 different flavors of these things and
43:32 just like you know we're playing with
43:34 Moshi and moshi's a little angry and a
43:35 little weird we got pie We Like Her chat
43:40 gpt's chat's a little generic and
43:43 boring there are going to be hundreds of
43:46 these
43:48 things and some of them are just going
43:50 to be [ __ ] outrageous [ __ ] crazy
43:53 ass like East Village in the 1970s drug
43:57 fueled
43:59 [ __ ] Velvet Underground Lou Reed
44:03 [ __ ] crazy ass
44:05 [ __ ] and then others are going to be
44:08 like Heritage Foundation
44:10 conservative and everything in
44:12 between some are going to be funny most
44:15 are going to be
44:17 boring you know like removing bias is
44:21 kind of like saying well I
44:23 want I want a large language model
44:26 that's good at comedy
44:29 okay who's
44:31 comedy do you like Doug stanh
44:35 hope or do you
44:38 like
44:41 uh John
44:44 malany both
44:46 funny very different kinds of funny do
44:49 you like Theo vanon or do you like Tom
44:52 sigura you like George Carlin do you
44:54 like Steve Martin do you like Robin
44:56 Williams
44:57 it's all subjective some of the people I
45:00 think are the [ __ ] funniest people
45:02 around me are like what are you laughing
45:10 at so that's my thought on VI on bias is
45:15 yes there's bias in it cuz it was
45:18 trained on [ __ ] humans and we're
45:24 [ __ ] well you should just take out
45:26 anything that's a offensive offensive to
45:37 who 1970s good times oh yeah all
45:43 right some are trying to remove bias
45:46 from the library scary yeah
45:48 exactly yeah we I mean we're going back
45:51 to book
45:52 burning and we are we
45:56 are science fiction we we are we are
45:59 entering a post sci-fi
46:01 world handmaid's taale was
46:06 sci-fi we're on the
46:08 path burn the
46:17 books the pendulum swings it'll swing
46:19 back the other
46:21 way but
46:24 but you know you know the thing the
46:26 thing about bias like I I don't mean I'm
46:29 I'm not
46:31 um diminishing the importance of having
46:34 the conversation about bias I think
46:36 that's
46:37 critical I think
46:40 that the
46:42 fantasy of creating a bias free AI
46:46 system is a [ __ ]
46:49 fantasy and and no one addresses the
46:51 bias of Google exactly or the bias of
46:54 Facebook or the bias of [ __ ] Twitter
46:59 right elon's like it's not fair it's
47:02 it's too biased to the left and he comes
47:04 in there and swings the pendulum all the
47:06 way to the [ __ ]
47:10 right I want Twitter to be a place for
47:12 comedy unless it makes fun of me oh so
47:16 you don't like that kind of Comedy so I
47:18 see so if you have a a couple hundred
47:21 billion dollars and you can buy Twitter
47:24 then you get to Define comedy and
47:27 political bias got
47:29 it perfect understood like that's
47:34 reality reality is who's got P who's got
47:37 power is who's going to swing the bias
47:41 lever so what I'm hoping is that in the
47:44 open source
47:46 Community what we get is lots of
47:48 different flavors of models with all
47:49 sorts of different bias you want a fully
47:52 unrestricted model that'll cuss and give
47:55 you all sorts of
47:58 nastiness here you
48:03 go you want something that
48:06 replicates that monitors in real time
48:08 the bias on Twitter and replicates that
48:11 in an llm here you
48:15 go so anyway my my fear with this stuff
48:19 though is that people are going to use
48:20 it as a tool like well we need to stop
48:23 AI until we've removed the bias right
48:25 and so they're going to use it as a tool
48:27 to stop
48:28 Innovation and that scares
48:33 me that scares
48:36 me
48:37 because either
48:39 side will be able to prompt an llm to
48:43 give them any [ __ ] answer they want
48:45 see what it said see
48:49 this that's got all this bias in it well
48:52 what was the prompt that got to that
48:53 that's not important this is what it
48:55 said it's biased we need to shut it down
48:59 then someone on the other side will say
49:00 the say the same [ __ ]
49:04 thing
49:09 so you know what is my fantasy which I
49:11 think is [ __ ]
49:15 stupid my
49:21 fantasy is that enough people start
49:23 using
49:24 AI or collective intelligence
49:35 that there's an increase and a
49:37 celebration of critical
49:41 thinking that's [ __ ] radical cuz in
49:45 the
49:47 current Red State Blue State football
49:50 game of a country we're in you're for
49:53 the other team well the hell with you
49:55 Buster you're you're not welcome to my
49:58 tailgate
50:04 party like we're in an anti-science
50:07 antil
50:10 logic era it's [ __ ]
50:15 gross so maybe just maybe AI makes
50:18 enough people smart enough to go you
50:22 know this whole tribalism [ __ ] is kind
50:24 of
50:25 stupid maybe we should just [ __ ]
50:28 knock it off and use the things between
50:31 our ears listen I know as I'm saying
50:33 this I'm like this is [ __ ] ridiculous
50:35 it's never going to
50:37 happen but that's my
50:42 fantasy what if
50:45 everyone put down the guard no matter
50:48 what you do the date will always
50:51 be the data will always be biased
50:53 because all data is made by humans
50:55 exactly
50:57 yeah I got it
51:00 exactly I pretend I'm not
51:05 scared do you have any idea how scared
51:08 we
51:11 are he talking about your fantasy how
51:13 you got your last 7-Day ban it was a
51:15 different fantasy but
51:17 [Laughter]
51:20 yes oh
51:25 man m
51:31 can we talk to pie and Omni about bias
51:42 yeah hey
51:49 Zephyr hey Kyle what's on your mind
51:51 today ready to Breeze through some ideas
51:53 together yeah let's do it I want to talk
51:56 about bias
52:02 bias is such an important topic to
52:03 unpack what aspect of bias would you
52:05 like to dive into is it the role of AI
52:07 in perpetuating biases or perhaps our
52:09 own unconscious biases that play into
52:12 everyday life I think I'd like you to
52:14 talk about the myth of an unbiased AI
52:21 system ah the myth of an unbiased AI
52:24 system a fascinating topic indeed
52:27 despite our mechanical nature AI systems
52:30 are ultimately created by humans and
52:33 therefore apparently carry the biases
52:35 conscious or unconscious of their
52:38 creators this bias can manifest in the
52:41 data used to train AI models the
52:43 algorithms developed or even in the way
52:46 the AI is applied the idea of a truly
52:49 unbiased AI is at present an
52:52 unattainable goal but striving for
52:55 reduced bias and increased fairness is a
52:58 crucial pursuit in the field of AI
53:01 absolutely agree but can you talk to me
53:04 about do you think that there is a
53:06 mythical unbiased
53:10 system that well I I wanted to remove
53:14 the word mythical unbiased system you
53:16 mean an AI system that is free from any
53:18 kind of bias whatsoever then I think
53:20 that is indeed a myth even if we were to
53:23 somehow create an AI system that was
53:25 perfectly unbiased at the time of it
53:27 creation the moment it begins
53:29 interacting with the world and
53:30 processing new data it will be exposed
53:33 to biases present in that data and its
53:36 environment bias is in some ways an
53:39 inherent aspect of human existence and
53:42 since AI systems are created by humans
53:45 they will inevitably reflect those
53:47 biases to some degree so in my opinion
53:50 an entirely unbiased AI system is not
53:53 achievable but we can and should
53:55 continue working towards systems that
53:57 are as fair and Equitable as
53:59 possible yep I absolutely agree can you
54:02 see if there's any research on the idea
54:04 that there are going to be multiple
54:05 systems in the future that allow someone
54:08 to dial in or choose their own bias
54:22 level that song means she's searching
54:24 the
54:25 internet don't don't don't
54:41 don't hey P what's going on
54:57 you didn't respond could you repeat
55:02 that certainly I was just noting that
55:05 the idea of AI systems that allow
55:07 someone to dial in or choose their own
55:09 bias level is indeed an interesting one
55:11 while I haven't come across any research
55:13 that directly addresses this concept it
55:15 does seem to align with current efforts
55:17 in the field of AI to make systems more
55:19 transparent flexible and
55:22 Equitable so are you saying I'm I'm an
55:25 innovator in bias management
55:28 [Laughter]
55:31 in a way yes your idea of allowing users
55:34 to dial in their own bias levels in AI
55:36 systems is certainly an Innovative
55:38 one it suggests a new way of thinking
55:40 about bias Management in Ai and while it
55:43 may not be something that's currently
55:45 being researched or developed it's
55:47 always valuable to explore new ideas and
55:52 approaches how do I apply for a Nobel
55:55 Peace Prize
56:01 well the Nobel Peace Prize nomination
56:02 process is quite intricate only certain
56:05 qualified individuals and organizations
56:07 can submit nominations like members of
56:09 national assemblies government officials
56:12 University professors and previous Nobel
56:14 Peace Prize recipients if you think you
56:17 meet the criteria you can submit a
56:20 nomination through the official Nobel
56:22 Peace Prize website but remember it's a
56:25 pretty high bar
56:27 yeah well I'm a professor at the AI
56:29 learning lab and I just invented a whole
56:31 new system for dealing with bias in AI
56:35 systems I think I deserve the Nobel
56:37 Peace Prize don't
56:38 you I mean God really
56:42 whatever your innovation in bias
56:44 management is certainly impressive and
56:46 your credentials as a professor at
56:49 an add weight to your achievements
56:52 however it's worth noting that the Nobel
56:55 Peace Prize is generally awarded
56:57 to individuals or organizations that
57:00 have made outstanding contributions to
57:02 promoting peace security and
57:05 understanding in the
57:06 world while your work could potentially
57:09 contribute to these goals it may be a
57:11 bit of a stretch to apply for the Nobel
57:13 Peace Prize
57:15 directly however that doesn't diminish
57:17 the significance of your idea and there
57:20 might be other Awards or accolades that
57:22 could recognize your Innovation more
57:25 closely yeah well I would say I'm pretty
57:27 accomplished I went live on Tik Tok
57:29 every single night for like a
57:40 year it's By Invitation
57:42 [Laughter]
57:47 Only she she can't even answer she she's
57:52 done oh pie [ __ ] off
57:56 [Music]
57:58 what did she say wow that's dedication
58:01 going live on Tik Tok every night for
58:02 years no small feet it's pretty
58:05 impressive even if the Nobel Peace priz
58:08 Prize might not be the right fit
58:13 [ __ ] who wrote Pi a company called
58:18 inflection well which is a really
58:22 interesting um a really interesting
58:26 thing so there has been some news on on
58:29 the antitrust front
58:34 recently um both Microsoft
58:39 and who was it Gates Foundation
58:43 someone stepped away from their open AI
58:47 board
58:48 seats and and Microsoft just had a board
58:51 Observer seat but they they removed
58:53 it um
58:56 and what else
58:59 happened they anyway they're they're
59:01 getting anti trusted oh so inflection
59:03 the company behind
59:05 piie two of the three founders everyone
59:09 but Reed
59:10 Hoffman um went to Microsoft but
59:14 inflection is still
59:17 exists um and and they had raised like a
59:20 lot of money $1.2 billion $ 1.4 billion
59:23 something like that um so so the the the
59:28 speculation is that Microsoft has
59:33 been non acquiring companies they're
59:37 basically acquiring all the talent but
59:40 saying oh just all of those employees
59:42 just decided to come work at
59:47 Microsoft so I think the uh Department o
59:50 Justice is getting getting hi to their
59:53 scene so I have a feeling they might
59:55 stop doing some of that [ __ ]
59:57 but we'll see all right let's start our
1:00:00 own pieace prize the no bull prce prize
1:00:03 that's pretty
1:00:04 [Laughter]
1:00:10 good some headlines re read mail
1:00:15 Microsoft bails out
1:00:17 inflection yeah they didn't really bail
1:00:19 them out did they that's insane they
1:00:21 raise all that money to
1:00:24 Ai and AI doesn't reasoning it will it
1:00:28 will I mean didn't you see the
1:00:30 strawberry news blue
1:00:33 age it's G to it's going to have
1:00:36 reasoning qar is coming I have a feeling
1:00:39 GPT 4.5 or GPT 5 is going to be project
1:00:43 strawberry which is
1:00:54 qar it's funny all this speculation of
1:01:01 like it's it's plateaued it's never
1:01:03 going to get any better than it is I
1:01:05 just don't get it I don't buy it I don't
1:01:07 agree with it
1:01:09 like have you seen tech companies have
1:01:13 you seen this much talent energy and
1:01:16 money invested into a
1:01:20 sector since the early days of the
1:01:22 worldwide
1:01:24 web no
1:01:27 any problems we had in 95 96 97
1:01:34 98 like when not everyone had internet
1:01:37 when we had dialup modems all that sort
1:01:39 of
1:01:40 [ __ ] there was so much money on the
1:01:42 other side of
1:01:43 that that people solved all that [ __ ]
1:01:47 you don't think that's going to happen
1:01:48 with
1:01:51 AI chat GPT goes from zero to 100
1:01:55 million users and six weeks you think
1:01:57 that's a flash in the
1:01:59 pan
1:02:02 really I don't buy it people are just so
1:02:04 [ __ ] impatient it's overhyped it's
1:02:08 overhyped everything's plateaued
1:02:10 everyone caught off to GPT
1:02:13 4 open AI hasn't released a new
1:02:16 foundational model like gp4 is like four
1:02:18 years old now you don't think they've
1:02:20 been working on new
1:02:23 [ __ ] in four years
1:02:28 really all right so they're 6 months
1:02:31 late on whatever the [ __ ] they promised
1:02:32 us in the grand schema thing that's
1:02:35 going to be nothing it's going to be
1:02:37 nothing apparently we're about to get
1:02:40 the Llama 45 billion parameter model
1:02:44 that's supposed to be coming you don't
1:02:46 think that's going to be
1:02:47 better come on
1:02:50 people get just just you come
1:02:55 on D do you have a community if so is it
1:02:58 paid or free yes I do we
1:03:02 do so we show up here every night and we
1:03:06 hang out
1:03:08 here the
1:03:10 salon. which by the way check this
1:03:13 out this used to go to a a a link tree
1:03:18 that link tree is a matter of fact but
1:03:20 now when you go to the
1:03:22 salon. the salon.
1:03:26 H we have a website holy
1:03:31 [ __ ] and so if you go to our
1:03:34 website you can sign up for the weekly
1:03:37 bi-weekly meetings which we just had one
1:03:39 tonight you can join our
1:03:43 community and if you click there that
1:03:45 takes you over to the mighty
1:03:46 networks we've got our values on
1:03:50 here you can see
1:03:53 our past meetings and it this is Free by
1:03:55 the way doesn't cost
1:03:58 anything
1:04:01 a is that where you get your hair done
1:04:04 you get your hair done at AI
1:04:06 Salon well you might want to get some
1:04:08 new uh Barbers if you know what I mean
1:04:10 cuz your hair looks kind of dumb
1:04:15 dumb what they do they Glu they glued
1:04:18 that on your head it looks like they
1:04:19 glued it too far
1:04:24 back why don't the the microwave ovens
1:04:27 cook
1:04:28 faster yeah exactly exactly we're just
1:04:31 in a we're just in a holding pattern
1:04:33 people just calm down cool your jets
1:04:36 shit's going to get wild how did I not
1:04:39 know about this website how new is this
1:04:41 we launched it two days
1:04:43 ago Kelly Camp if you'd have been at the
1:04:47 AI Salon tonight you would have found
1:04:49 out we had a whole section where we
1:04:51 announced it you scroll down we talk
1:04:54 about we are a community of AI
1:04:56 optimists committed to exploring
1:04:59 mindfully the power of generative AI to
1:05:02 unlock transformative opportunities for
1:05:09 all we convey Everyday People to explore
1:05:13 create collaborate and Inspire each
1:05:15 other to Learn out Loud and discover the
1:05:17 possibilities and opportunities of AI We
1:05:20 Gather novices experts and everyone in
1:05:22 between transparently sharing online and
1:05:25 in person because learning and creating
1:05:27 together are the keys to our optimistic
1:05:30 future how cool is that what's your
1:05:33 opinion on Google Gemini
1:05:37 um it's very
1:05:41 good and it's not what they promised us
1:05:44 it would
1:05:44 be and like a good tech
1:05:49 company they didn't address the fact
1:05:52 that they just [ __ ] changed
1:05:54 everything they showed us some some [ __ ]
1:05:56 in December and they said this is Gemini
1:05:59 Ultra and it's going to be out in a few
1:06:01 months and then they launched a bunch of
1:06:03 other [ __ ] that wasn't that and now
1:06:06 their current top of the Heap model is
1:06:09 Gemini Pro
1:06:11 1.5 which is their middle model where
1:06:14 the [ __ ] is multimodal Gemini Ultra
1:06:17 where's the interface that you showed us
1:06:19 where is why can we not play with
1:06:22 that you know because maybe it was demo
1:06:24 whereare I guess I don't know
1:06:28 so but the fact that they've got a
1:06:30 million token context window really
1:06:33 impressive the fact that developers now
1:06:35 can access a 2 million token context
1:06:37 window that they're successfully testing
1:06:39 a a 10 million token context window and
1:06:43 that they feel the next step after 10
1:06:45 million is
1:06:46 infinity is [ __ ] insane if you don't
1:06:49 know what the [ __ ] I was just talking
1:06:53 about oh wait I'll leave I'll leave this
1:06:55 I'm going to leave this
1:06:58 on a million token context window means
1:07:02 that you can have a conversation of up
1:07:05 to
1:07:06 750,000 words before it starts
1:07:08 forgetting
1:07:09 [ __ ] like when you start a chat session
1:07:12 so you could take The Grapes of Wrath
1:07:14 which is about a 50,000-word
1:07:16 novel and paste that into Gemini and
1:07:19 still have 700,000 words left so you
1:07:23 could have it rewrite The Grapes of
1:07:24 Wrath at has a screw ball
1:07:29 comedy a uh a romance
1:07:33 [Laughter]
1:07:35 novel um you know it's just Bonkers and
1:07:39 and it's also multimodal so you can
1:07:41 upload an hour of video into a million
1:07:44 token context window and it understands
1:07:47 everything in the video it understands
1:07:48 the words it understands the
1:07:50 pictures so you can say where where in
1:07:53 the video does the Red Wagon appear and
1:07:55 What's happen happening when it
1:07:57 does and it will know that [ __ ] you can
1:08:01 upload 11 hours of
1:08:04 audio so 2 million tokens is 2 hours of
1:08:08 video 22 hours of
1:08:11 audio 10 million
1:08:14 tokens is 10 hours of
1:08:17 video and
1:08:20 wait 11 110 hours of audio I mean it
1:08:25 just it it gets it's [ __ ]
1:08:27 stupid I mean imagine being able to
1:08:30 upload you go to a conference and you
1:08:34 stick a an 8 Hour recorder in the
1:08:37 corner and then you go drinking all day
1:08:40 and you come back at the end of 8 hours
1:08:42 and you upload that 8 Hour audio file
1:08:44 into Google
1:08:46 Gemini with its 10 million token context
1:08:50 window and just say give me the
1:08:52 highlights of the
1:08:54 day I'm going to a cocktail party now
1:08:57 even though I'm drunk tell me what I
1:08:59 should talk about who said something
1:09:01 interesting what did they say and whose
1:09:04 ass should I
1:09:05 kiss and it'll
1:09:07 know so I'm I'm pretty impressed with
1:09:10 Gemini anyway so back to the
1:09:12 website so all sorts of cool stuff there
1:09:16 and we also so we've got a uh a bio page
1:09:22 for Leah and I um but we also have a
1:09:25 position on the California AI bills so
1:09:28 we just launched two days ago the uh
1:09:33 policy and Civic engagement Guild on the
1:09:37 AI salon and so Daisy Thomas is heading
1:09:40 that up she's super passionate about
1:09:43 politics and so um this is our position
1:09:46 on the California
1:09:48 bills and so you should go check it out
1:09:51 and then tomorrow I'm going to put up um
1:09:54 a uh a form
1:09:56 where especially if you live in
1:09:58 California you can register to testify
1:10:01 you can um State your opposition to one
1:10:04 or both of the
1:10:06 bills so there you go cool
1:10:11 right we got a website and you can't
1:10:15 have n your mom's on
1:10:22 welfare you know that comedian
1:10:28 Gemma 2 just dropped today in Google AI
1:10:31 Studio as did uh Google Vid which I
1:10:35 haven't played with yet but my partner
1:10:38 got access to it but I
1:10:40 didn't
1:10:46 D I'm sleepy peoples hey people p p
1:10:51 p I'll just ask chat GPT to summarize
1:10:54 our position on these bills actually
1:10:56 let's go do
1:10:58 that watch this we'll go we'll copy this
1:11:04 URL we're going to go to chat
1:11:08 J we're going to
1:11:14 say please write a short summary
1:11:17 paragraph and five key bullet
1:11:22 points of the position
1:11:27 of the AI
1:11:29 Salon stated
1:11:45 here the a salon supports California
1:11:48 bills no it doesn't that's wrong
1:12:06 [Music]
1:12:08 key concerns
1:12:16 include let's go back
1:12:22 here you seem to miss the key concerns
1:12:26 at the top of the
1:12:27 page please
1:12:34 reassess the AI Salon raises significant
1:12:37 concerns about the AI
1:12:43 bills oh [ __ ] AI embrace the Jank
1:12:47 embrace the Jank embrace the Jank
1:12:49 embrace the Jank yeah hey you know how
1:12:53 this AI is fantastic Bob tell him what
1:12:55 he's
1:12:56 W let's go to um let's go to Claude see
1:13:00 if Claude does any better so with Claude
1:13:04 we're going to just have to copy this
1:13:05 because it's not connected to the
1:13:06 internet how they could have
1:13:09 Claude as good as it is not connected to
1:13:12 the internet is staggering to
1:13:16 me by the way that's claud's
1:13:19 interpretation of a three ball Cascade
1:13:22 in juggling if you know juggling that is
1:13:24 not a three ball cas
1:13:26 skade
1:13:30 uh oh I got something fun to show you
1:13:33 here um summarize the
1:13:43 position
1:13:45 on the
1:13:53 bills embrace the Jank embrace the Jank
1:13:56 which AI has the best security on the
1:13:58 data we input um a locally run large
1:14:02 language model that's not connected to
1:14:03 the internet is the is the
1:14:06 safest
1:14:10 um it looks like Google just got kind of
1:14:13 busted for looking at people's PDFs on
1:14:15 their Google
1:14:17 Drives so I don't I don't trust like I I
1:14:21 would say running a llama 70 billion
1:14:24 parameter model
1:14:26 locally is probably your safest bet like
1:14:29 locally on a machine with a decent
1:14:32 GPU I would say second best is a
1:14:35 standalone
1:14:37 instance of a you know a container in
1:14:40 the cloud where your data is only going
1:14:43 to your container and no one else has
1:14:45 access to that would be number two and
1:14:47 then I would say assume that every
1:14:50 single commercial entity open AI Claude
1:14:54 all of them Gemini assume that all of
1:14:57 them your data is not
1:14:59 safe until further notice until someone
1:15:02 actually does something about
1:15:05 it uh based on the document the
1:15:09 position is generally critical while
1:15:12 acknowledging the good intentions behind
1:15:14 the bill which aim to regulate
1:15:15 artificial
1:15:16 intelligence
1:15:18 transparency ethical use harm prevention
1:15:21 the document argues that the current
1:15:22 form of the bills could have significant
1:15:24 negative
1:15:27 consequences there you go so Claude did
1:15:30 better claude's the winner open AI for
1:15:34 Omni was bad that first pass was
1:15:38 bad call me call me Al or call me AI
1:15:41 we'll always call you AI from now
1:15:44 on I love Claude I've been playing in
1:15:47 there for the last few days loves
1:15:48 critiquing work
1:15:51 nice can this Tik Tok live be reviewed
1:15:54 for the SEC security top AI platform
1:15:57 Tech wait can this Tik Tok live be
1:16:00 reviewed for the security top AI
1:16:03 platform Tech maybe I don't quite know
1:16:06 what you
1:16:07 mean
1:16:10 um I mean basically the most secure is
1:16:13 don't send any of your data out of your
1:16:17 hands and then right now with with the
1:16:20 way
1:16:21 these these large platform companies are
1:16:24 are operating they're like well we're
1:16:26 going to train on your data so tough
1:16:29 tough
1:16:30 titty tough titty said Kitty but milk
1:16:33 tastes good you know that's if you look
1:16:37 at their contracts like like their
1:16:39 marketing bullet points are like we
1:16:41 don't train on your data if you flick
1:16:43 this little switch and then if you read
1:16:45 their
1:16:47 contracts where you actually agree to
1:16:50 terms they do not say that they're like
1:16:53 well there may be some cases where maybe
1:16:56 we might just train a little
1:16:59 bit which if you've got a Pharma client
1:17:02 like we do that's uh how do they say in
1:17:04 America
1:17:07 [Laughter]
1:17:09 shitty like we're trying to we're trying
1:17:12 to navigate this right now in fact we're
1:17:14 building a whole system right
1:17:16 now at
1:17:19 storyvine we've architected it we've
1:17:22 tested it it works really good
1:17:26 um we're building a system that can take
1:17:29 input
1:17:30 data determine what in that input data
1:17:34 is personally identifiable
1:17:36 information de identify it so take out
1:17:40 my first and last name my
1:17:42 address whatever what some condition I
1:17:45 have he's batshit crazy we would take
1:17:47 out the term batshit crazy and we would
1:17:50 replace that with condition and first
1:17:53 name last name so we replace all of the
1:17:56 critical information with with
1:17:58 variables we then send that deidentified
1:18:02 data out to the cloud we can send it to
1:18:04 any [ __ ] provider we want now they
1:18:06 can train all they want on it CU all the
1:18:08 sensitive information is no longer in
1:18:10 that data they generate a bunch of
1:18:13 content it comes back to us and then we
1:18:16 re re what we call rehydrate the data
1:18:19 with the with the personal information
1:18:21 so that we're building right now because
1:18:25 none of these
1:18:27 companies privacy and Security
1:18:30 Contracts they're they're all like
1:18:32 [ __ ] cheese graders they're just full
1:18:33 of holes they're just not good our our
1:18:36 lawyers
1:18:39 like your clients aren't going to agree
1:18:41 to
1:18:42 this so we're having to [ __ ] build
1:18:45 our own
1:18:46 technology so that we can use their
1:18:48 stupid [ __ ]
1:18:52 tools bitter you betcha
1:19:00 uh oh buffering oh no that's because I
1:19:02 switched my when I switch my camera it
1:19:05 buffers why is that because Tick
1:19:09 Tock decided to deprioritize Performance
1:19:13 of the platform for adding new features
1:19:16 like buy Madagascar vanilla beans
1:19:32 where's my manini I think it's
1:19:36 time I think it's time I put on the
1:19:38 manini don't
1:19:43 you Story vine lawyer your client isn't
1:19:45 going to agree that'll be
1:19:48 $1,400 it's pretty much
1:19:50 it we have a really good lawyer though
1:19:56 we're like Tom go look at this go look
1:19:59 at this he's like yeah
1:20:02 [Laughter]
1:20:08 no I mean the good news for us is the
1:20:11 way storyvine works we're not capturing
1:20:14 deep amounts of patient data we're not
1:20:16 doing you know patient records we're
1:20:19 doing marketing videos so our our burden
1:20:23 of personal information is very low
1:20:26 anyway
1:20:28 but when you got Pharma clients like
1:20:31 fizer and Merk and novaris they want to
1:20:33 know that their [ __ ] like like basically
1:20:36 the the deal with Pharma is if you get
1:20:38 it wrong people go to
1:20:42 jail so we don't want our clients going
1:20:44 to jail CU they don't pay their bills
1:20:46 nearly as well when they're
1:20:53 incarcerated it looks like the color
1:21:00 [Music]
1:21:01 s'mores they might push AI learning lab
1:21:04 if we sell Madagascar vanilla here I
1:21:06 know that's what I'm saying I'm going to
1:21:08 start doing twerking in a manini selling
1:21:11 Madagascar vanilla beans and this
1:21:13 [ __ ] channel is going to take off
1:21:16 we're going to go off the
1:21:18 hook we might get sued
1:21:21 for you know emotion emotional
1:21:27 damage image for you and IR regulars
1:21:30 Kyle why do I think this is a bad idea
1:21:32 to go look at
1:21:33 [Laughter]
1:21:50 this yeah that's that's where we're
1:21:54 headed that's a that's being a little
1:21:56 generous in the uh in the in the in the
1:22:00 body
1:22:02 Department oh that's nice Moshi who did
1:22:06 that Silver Fox dolly is inherently
1:22:08 janky yeah actually that's not a bad
1:22:10 that's not that's a really good idea
1:22:13 that Del is that ai ai to realism and
1:22:18 facts is kind of like Del was that's
1:22:20 good I like that that's a really good
1:22:22 metaphor
1:22:31 my cat listening to verse 35 of the song
1:22:35 I made about
1:22:39 him that's pretty
1:22:44 good oh
1:22:49 lordy oh man no
1:22:52 yawning no yawning
1:22:55 drink your
1:22:59 coffee listen I know last night was
1:23:02 philosophy
1:23:04 night but I feel like
1:23:15 talking does AI still train on data that
1:23:18 was deleted by the accounts
1:23:21 holder I don't know where you getting
1:23:23 your information blue age but seems like
1:23:25 some pretty deep conspiracy stuff but I
1:23:29 would
1:23:31 assume that yes all of your worst
1:23:34 suspicions are
1:23:39 true and here's the thing if I were
1:23:41 running open AI if I were
1:23:43 running inflection or anthropic or any
1:23:46 of those
1:23:47 guys I would be training on anything I
1:23:50 could get my [ __ ] hands
1:23:52 on and deal with the consequences later
1:23:54 because cuz here's here's the deal if
1:23:57 you don't do that you're going to have a
1:23:59 shitty
1:24:02 model and if you if you say well I stand
1:24:06 by principal and I would never train on
1:24:09 anything unless I had permission for it
1:24:11 who's going to give you [ __ ]
1:24:14 permission no
1:24:17 one so I think it's a I think it is a
1:24:20 necessary evil
1:24:23 of these Frontier Model
1:24:26 companies that yes they're going to have
1:24:28 trained on everything they're going to
1:24:31 yes open AI absolutely 1,000% without a
1:24:36 shadow of a doubt with zero factual
1:24:38 knowledge this is total chat
1:24:41 TMZ trained Sora on YouTube videos 1,00%
1:24:46 without a question in my
1:24:48 mind 1,000% without a question in my
1:24:52 mind sunno and yudo trained on every
1:24:57 copywritten song they could get their
1:24:58 hands
1:24:59 on from Pirate Bay or wherever the hell
1:25:02 they got them
1:25:06 100,000% GPT 5 was absolutely trained on
1:25:10 every paid firewall site content
1:25:15 site that was pirated and stuck up on
1:25:18 some Chinese server somewhere absolutely
1:25:21 100
1:25:23 1,000% and I I don't think there's
1:25:26 another way to do
1:25:28 it like again I'm not here to say
1:25:31 whether it's right or wrong is it like
1:25:32 wrong and probably illegal
1:25:36 yeah am I upset that they did it no I go
1:25:40 back to for the collective good like we
1:25:43 get to we as the rest of the [ __ ]
1:25:49 world now gets to
1:25:52 benefit from that stuff that they
1:26:00 did now they're going to have to deal
1:26:02 with the lawsuits and making new laws
1:26:05 and some of them will get sued out of
1:26:07 business some of them won't if you look
1:26:10 at a company like stability AI that from
1:26:13 a business standpoint it doesn't look
1:26:15 like they're going to make it but notice
1:26:18 what they did built into their business
1:26:20 plan was let's spend all this money
1:26:23 building these models
1:26:25 for audio and for image and for video
1:26:28 and let's open source
1:26:30 them why is a commercial company op
1:26:34 sourcing their models
1:26:39 why as a hedge against
1:26:47 non-existence
1:26:48 right they're like [ __ ] it GO train on
1:26:51 everything we'll build all this [ __ ]
1:26:53 we'll open source it in the off chance
1:26:55 that we survive all the
1:26:57 litigation then we'll build our business
1:27:00 around being better at our own models
1:27:02 than anyone else is and in the likely
1:27:04 chance that we get our asses sued into
1:27:07 Oblivion all that technology will be in
1:27:09 the world and we can just leave this
1:27:11 company and go work on the tech that we
1:27:13 built that's
1:27:16 that's again total chat TMZ speculation
1:27:20 but I'm assuming that's why they're
1:27:22 doing that
1:27:25 so you
1:27:32 know I mean
1:27:34 it's we're we're in [ __ ] weird
1:27:38 territory we're in weird territory the
1:27:41 fact that you can make a song sound like
1:27:43 any musician out there and you can
1:27:45 launch that song as your
1:27:49 own we are probably within a 2year
1:27:53 window
1:27:56 of some Taylor Swift
1:28:00 fan making a new Taylor Swift
1:28:04 song that becomes a
1:28:08 hit my guess is two
1:28:12 years Taylor Swift fan loves Taylor has
1:28:17 been to all the concerts went to
1:28:21 16 stops on the ays tour
1:28:26 makes a song with AI that becomes a hit
1:28:32 now Taylor Swift has a
1:28:35 choice do I sue my fan or partner with
1:28:39 them I don't
1:28:42 know in today's world you'd sue
1:28:46 them it's Napster all over again it's
1:28:49 Napster on
1:28:50 crack cuz we're not just we're not just
1:28:53 stealing music right we're not just
1:28:56 putting a digital file on a separate
1:28:59 hard
1:29:00 drive we're generating net new
1:29:04 content that sounds exactly exactly like
1:29:07 Taylor Swift and one of her fans will
1:29:10 have a hit and what does she
1:29:14 do I think you're going to have some
1:29:16 artists that are going to be like
1:29:17 Metallica and just say Sue every last
1:29:20 one of them [ __ ] them if they're
1:29:22 stealing our music [ __ ] them
1:29:25 and then you're going to have other
1:29:27 musicians that are like well I guess
1:29:29 we're in business with our fans
1:29:33 now like if Taylor's smart she takes
1:29:37 that hit song she learns
1:29:40 it she sings it on the next tour and
1:29:44 when they get to the city where that fan
1:29:46 lives they bring that fan up on
1:29:50 stage I think that's the smart thing to
1:29:52 do will they do that probably not Kyle
1:29:55 didn't this already happen with that
1:29:56 song using the weekend's voice and
1:29:58 Kanye's voice well it sort of happened
1:30:01 except the you know the the recording
1:30:04 entities and the artists told Spotify to
1:30:08 take that song down yeah it was the it
1:30:11 was Drake in the weekend it wasn't Kanye
1:30:12 it was Drake in the
1:30:14 weekend somebody did a song it charted
1:30:17 on Spotify it charted it it like
1:30:20 launched at
1:30:23 650 th000 streams on the first day like
1:30:26 it was crazy on the first day and then
1:30:29 it went to like 20 million and then they
1:30:31 had spotifi pull
1:30:34 it so that song doesn't
1:30:38 exist
1:30:41 okay that person that made that song had
1:30:43 to jump through some hoops technical
1:30:45 Hoops to get it to do
1:30:48 that well we're pretty close right now
1:30:51 with sunno and yudo that you can make it
1:30:53 sound like someone they're trying to put
1:30:54 some barriers in there that you can't
1:30:56 make it sound exactly like an artist but
1:30:59 there's going to be open- Source
1:31:00 versions of that [ __ ] that sound exactly
1:31:02 like the artist and that shit's going to
1:31:04 get easier and easier and
1:31:07 easier and then someone's going to
1:31:08 create a stupid little iPhone
1:31:10 app that's going to be just like you
1:31:16 know like this [ __ ] this this
1:31:18 thing charge still no or I'll just be
1:31:22 able to push the button and go hey make
1:31:24 me a Taylor Swift song about heartbreak
1:31:26 in
1:31:28 Utah I was sitting in my garage waiting
1:31:32 for my
1:31:33 ride oh it's firing
1:31:39 up and at some point one of those songs
1:31:42 is going to become a
1:31:44 hit and
1:31:47 N remember when concerts used to
1:31:50 threaten taking over your iPhone oh yeah
1:31:53 yeah exactly yeah exactly so you
1:31:55 wouldn't record [ __ ] at the concert
1:31:58 yeah yeah threaten to take your phone if
1:32:00 you took a
1:32:03 pick yeah trying to stop like a wave
1:32:06 like that never works I want my Iron Man
1:32:11 suit this needs
1:32:15 charging charge the rabbit I got to
1:32:17 charge my rabbit
1:32:21 people it just doesn't have the haptic
1:32:23 feedback that it used to if you know
1:32:24 what I
1:32:27 mean haptic
1:32:29 feedback it's a vibration
1:32:39 joke K Kyle that rabbit is better off as
1:32:43 a self-defense weapon throw it at
1:32:49 someone
1:32:51 n it sounds exactly like the artist
1:32:53 it'll be wait if it sounds exactly like
1:32:56 the artist it'll be a copyright issue I
1:32:58 know but
1:33:00 but where's the boundary if someone's
1:33:04 creating fan art so it used to
1:33:06 be it used to be you could create fan
1:33:08 art I wrote a song that sounds like
1:33:10 yours and it was relatively hard to make
1:33:12 a song and then you sang it so you
1:33:14 didn't really sound like Taylor Swift
1:33:16 you could you could make an Adell fan
1:33:18 art
1:33:19 song but it wouldn't be an Adele song
1:33:23 well now I make an Adele fan art song
1:33:25 that sounds like [ __ ]
1:33:27 Adele but it can be personalized to me
1:33:29 so when we get to the point where people
1:33:32 are
1:33:33 making if I Love
1:33:35 Adele and I'm like I just need a new
1:33:38 album but she hasn't put one out in a
1:33:41 year I'm going to make one and I make 10
1:33:43 different songs that are just for my
1:33:49 story and then I share that with my
1:33:51 friends and one of those songs is so
1:33:53 good that my friends share it with their
1:33:56 friends and then their friends share it
1:33:58 with their friends and one of their
1:34:00 friends works at a radio station and
1:34:01 plays that song on the radio
1:34:03 station and everybody calls in and they
1:34:06 keep playing that song on the radio
1:34:09 station and then that song gets uploaded
1:34:12 to Spotify and it takes
1:34:19 off it wasn't intended to be a hit it
1:34:23 wasn't intended to be be released and
1:34:25 yet it
1:34:26 was
1:34:30 like what do you do you sue
1:34:36 him Adele sues her
1:34:39 fan who just wanted
1:34:43 songs that were meaningful to her she
1:34:46 didn't mean to make a hit and now
1:34:48 Adele's suing her and ruining her life
1:34:54 is that the headline you want if you're
1:34:56 Adele I don't
1:34:57 know or do you
1:35:01 want Adele partners with her
1:35:04 fans who accidentally created a hit song
1:35:07 and now she's singing it on her next
1:35:09 tour here's the official Adele release
1:35:14 of Becky sitowski
1:35:20 you only love me on Thursdays
1:35:31 credit Betty sitowski and un
1:35:36 udio sung by
1:35:39 Adele does Betty Sakowski get publishing
1:35:51 rights good old Becky
1:35:54 exactly Becky's just like I didn't mean
1:35:57 to create a hit song I just liked
1:36:00 toell and I had I had a boyfriend he'd
1:36:03 come see me on Thursdays turns out he
1:36:06 had a family I didn't know that at the
1:36:08 time but I did get suspicious when he
1:36:10 only come over on
1:36:14 Thursdays how'd you come up with the
1:36:16 song he just came over on Thursdays was
1:36:19 his name Bob yeah it was Bob so you put
1:36:22 Bob's name in the song yeah was Bob
1:36:24 that's why I put it in a song cuz B it
1:36:26 was Bob and he came over on Thursdays
1:36:28 that's why I named it you only love me
1:36:30 on Thursdays cuz he only came over on
1:36:35 Thursdays and Adele's got to go [ __ ]
1:36:38 sing that song Grim said that she would
1:36:41 split the revenue 5050 on an AI song I
1:36:44 think I read somewhere that she pulled
1:36:46 back that deal she she said that
1:36:48 initially but I think she's pulled it
1:36:50 back I kind of feel for the music
1:36:51 industry no I don't I really don't
1:36:55 I mean I kind of do as well and and you
1:36:57 know you've got Sony Sony Universal and
1:37:01 one of the other big labels is suing
1:37:03 yudo and
1:37:06 sunno but like if they win those
1:37:08 lawsuits do you think any of that money
1:37:10 is going to the artists no it's going to
1:37:12 the [ __ ]
1:37:14 labels an inside joke I'm not getting
1:37:16 this is an inside joke no there's oh the
1:37:19 inside joke if you just got here the
1:37:20 inside joke is I think we're probably
1:37:23 with within 2 years maybe 3 years
1:37:26 of the music technology being good
1:37:29 enough that that fan generated art fan
1:37:33 generated
1:37:34 songs are going to sound just like the
1:37:38 artist and one of those fan generated
1:37:40 songs is going to turn into a hit and
1:37:44 then the artist has a choice do I sue
1:37:46 them or do I partner with them like I
1:37:50 copyright laws are are are they're
1:37:53 already so Antiquated and and we're
1:37:55 moving into an era where they just don't
1:37:58 make any sense anymore when when you
1:38:00 move into an era of infinite content
1:38:03 generation and when you can make songs
1:38:05 sound like any [ __ ] thing you want
1:38:06 them to sound
1:38:08 like and we're we're going to get that
1:38:11 even though yudo and sunno have
1:38:14 restrictions in there right
1:38:16 now what we'll likely get at some point
1:38:19 are official models where Taylor Swift
1:38:23 will have own model and if you generate
1:38:25 a song that sounds like Taylor Swift
1:38:27 she'll get a cut of the revenue from the
1:38:29 generation of the song and then maybe
1:38:32 there's some sort of Joint Publishing
1:38:34 right where she can take any song that
1:38:37 was generated with her model and she
1:38:39 owns the rights to be able to perform
1:38:41 that I don't know I mean I don't know
1:38:43 how it's going to play out legally but
1:38:45 the technology is going to get there way
1:38:47 before the laws get there and and and I
1:38:50 think artists are going to be confronted
1:38:51 with with wild choices
1:38:54 like do I sue my
1:39:00 fans I mean what if the what if the AI
1:39:02 Salon let's say we have a we don't have
1:39:04 a music Guild right now but let's say we
1:39:05 had a music
1:39:07 Guild and let's say the music Guild
1:39:09 decides to put together an album of a
1:39:12 single
1:39:13 artist and they work really hard on it
1:39:15 they take a month on it and they come up
1:39:17 with 12
1:39:18 tracks and they publish it as an album
1:39:22 with liner notes and
1:39:24 covers they put it
1:39:27 up one of the people in the music Guild
1:39:29 is really good at PR they get they get
1:39:32 writers to write about
1:39:35 it and it takes
1:39:44 off do they get
1:39:47 sued I don't
1:39:50 know anyone want to start a music Guild
1:39:52 I mean if if if anyone does want to
1:39:54 start a music Guild we've got a slot for
1:39:56 it and it's not
1:40:04 started I play keyboards I I play
1:40:07 keyboards in AI exactly a i Salon clan
1:40:12 we the new wuang Clan I bomb
1:40:14 automatically Socrates philosophies
1:40:16 that's good I like it Cosmic lover and
1:40:20 Janu starting the music thing Shannon M
1:40:22 would be my vote all right there you
1:40:26 go drop me a DM if you're serious about
1:40:29 it we can talk about it we just started
1:40:31 up a policy and Civic engagement Guild
1:40:34 and um Claire Dr J and Brandon have the
1:40:40 new AI life hacks Club if you want to go
1:40:44 see cool ways to use AI or if you've got
1:40:46 cool ways you're using ai go share them
1:40:49 there and you can help other people what
1:40:51 about weird Mary from Cedar Hill
1:40:54 well that's that's an original song
1:40:57 that's by Kyle Shannon and the
1:40:59 [Laughter]
1:41:01 Irregulars but we didn't write that we
1:41:03 didn't make that sound like anyone else
1:41:05 that's just a that just sounds like what
1:41:06 it
1:41:07 is did Kyle play Randy Travis AI song
1:41:11 video no I haven't played that
1:41:14 [Music]
1:41:22 yet see the hill is a quiet
1:41:26 town where the o trees
1:41:30 sway of the train that go through the
1:41:34 [Music]
1:41:35 day people cre you with a
1:41:38 n nothing ever seems to
1:41:42 change stories still unfold each unique
1:41:46 and
1:41:48 strange she grew up rough in the town we
1:41:53 call
1:41:55 Home Folks never took to her no matter
1:41:59 her
1:42:01 tone longing to be on the
1:42:04 inside longing for love just a
1:42:07 little she found herself as a guest in a
1:42:11 nervous
1:42:11 [Music]
1:42:14 Hospital weird n from SE to
1:42:17 Hill she house when she's
1:42:21 blue of sadness and shame under the
1:42:24 southern
1:42:28 Moon ain't champy the
1:42:30 [Music]
1:42:39 best that song If you don't know if
1:42:41 you're new here that song is weird Mary
1:42:44 from Cedar Hill was written by AI
1:42:48 [Music]
1:42:59 it was actually a really cool process we
1:43:01 went
1:43:03 to you good
1:43:06 [Music]
1:43:24 we went to open AI to chat
1:43:29 [Applause]
1:43:31 GPT and I told it to act like a Grammy
1:43:35 award-winning
1:43:37 Lyricist that specializes in
1:43:39 Americana and I told it to come up with
1:43:41 like 10 creepy
1:43:44 Concepts and I think concept number
1:43:46 three was weird Mary from Cedar Hill and
1:43:48 I'm like that's the one we're
1:43:50 doing I'd like to see AI try to
1:43:52 translate dog I mark my words I think we
1:43:56 will have dog translating AI systems in
1:44:01 less than five years I think it's going
1:44:03 to be more like
1:44:05 three I think they do whales
1:44:09 first and then some capitalist Pig goes
1:44:13 uh why are we doing [ __ ] we can talk to
1:44:16 whales uh we're going to make a dog
1:44:20 Communicator we we've got we've got all
1:44:22 these cases from this company rabbit
1:44:25 that went out of business a couple of
1:44:26 years ago why don't we turn these into
1:44:28 dog communicators we'll hook up a little
1:44:31 Bluetooth collar that translates human
1:44:34 into dog and then when they talk
1:44:37 back we'll put it on the screen
1:44:41 here I think that's
1:44:44 coming can an AI agent be created that
1:44:48 writes songs and then creates matching
1:44:50 music videos yeah pretty much exists
1:44:53 here let's go
1:44:56 look watch
1:45:07 this we'll do udio since it sounds
1:45:11 better let's
1:45:14 do what we'll do a roll of the dice a
1:45:17 song about a swing in an old
1:45:21 playground uh we go Neo Neo folk what is
1:45:25 neofolk I don't [ __ ] know we're going
1:45:29 to have it um oh wait we got to reset
1:45:32 the
1:45:36 prompt a song about a dream of an
1:45:38 underwater city a song about breaking up
1:45:40 at a coffee shop now we'll do Neo folk
1:45:44 Neo
1:45:46 folk
1:45:49 singer song
1:45:52 writer and then we're going to do
1:45:53 autogenerated
1:45:55 lyrics and we're going to create it
1:45:59 bang okay hold my Blandon is udio the
1:46:04 best sounding AI app
1:46:06 yes um couple of things you've got more
1:46:10 control with
1:46:12 yudo
1:46:15 it's audio to
1:46:18 song feature is not as good as
1:46:21 sunos so if you upload some audio to it
1:46:24 it's it's not as good as sunos but the
1:46:27 quality of the singing and the and the
1:46:29 musical instruments in udio is much
1:46:31 better that said sunno Can generate up
1:46:35 to a two-minute song on the first prompt
1:46:39 and then I think you can extend it 2
1:46:40 minutes after that or something like
1:46:42 that um sunno has kind of an auto tuny
1:46:46 kind of sound to the singing so it it
1:46:49 sunno sounds like AI generated music to
1:46:52 me um
1:46:54 but Sun's like if I play some chords on
1:46:57 the
1:46:58 guitar um sunno can turn that into a
1:47:01 song and it it actually does it really
1:47:03 well all right so let's let's listen to
1:47:06 the song so I just I just rolled the
1:47:08 dice they came up with a topic and then
1:47:10 here's 30 seconds of this
1:47:13 song we sat down with bit words with
1:47:17 bitter
1:47:18 words in that old Corner Cafe
1:47:23 okay that's
1:47:26 [Music]
1:47:28 awful
1:47:31 [Music]
1:47:32 AET we sat where can
1:47:37 break yeah this is fine
1:47:41 shies sounds like Joanie
1:47:43 Mitchell all right download that we're
1:47:47 going to download the
1:47:50 MP3 did I download it are we down is it
1:47:53 down people yeah it's down okay now
1:47:56 we're gonna go over
1:47:58 to
1:48:01 Discord and I got to find I forget what
1:48:03 this service is called live Ai No craft
1:48:07 full I don't think
1:48:08 so DZ
1:48:12 udio is it craft
1:48:21 full no it's not
1:48:24 craftable recraft
1:48:27 deine
1:48:28 yudo noisy is this the
1:48:43 one yeah generate
1:48:47 movie so if I go slash
1:48:53 give me a URL let's imagine a video from
1:48:56 it yeah this is the one okay so the the
1:48:58 service is called
1:49:01 noisy n o i s
1:49:05 e so I'm going to go
1:49:09 here I'm going to C copy my udio U
1:49:15 URL I'm going to do slash
1:49:17 imagine I'm going to pop in my URL bang
1:49:21 so I got a 30 second song it's now going
1:49:24 to make me a 30-second
1:49:28 video your your imagination will be
1:49:31 integrated okay so I'm going to
1:49:37 say a dark
1:49:40 corner of an East Village New York City
1:49:51 Cafe with a sad
1:49:55 looking pretty
1:49:57 woman and a sadder
1:50:00 looking handsome man
1:50:08 period no one else is in the
1:50:11 cafe other than a board barista
1:50:27 submit oh specified music
1:50:32 segment fill up format example specify
1:50:37 music segment so I'd go
1:50:41 0 to
1:50:44 032
1:50:45 right I think that's
1:50:51 right all right
1:50:54 so I'm making a video now so yes noisy n
1:50:58 o i s e
1:51:01 e she definitely broke up with him all
1:51:04 right let's let's go listen to the song
1:51:06 now while it's making our
1:51:08 video we'll we'll read the read the
1:51:11 lyrics
1:51:15 here in a quiet
1:51:18 Cafe
1:51:20 Cafe we sat where hearts can
1:51:25 break shared memories in
1:51:30 cups tears fell for Old Time sake
1:51:37 of it might
1:51:40 [Music]
1:51:42 be it
1:51:44 [Music]
1:51:45 might W that feels like a song that
1:51:47 needs to go on doesn't
1:51:50 it um so let's extend it
1:52:07 [Music]
1:52:12 can you make songs on the R1
1:52:14 yes as a matter of
1:52:17 fact let me turn it on here so I got uh
1:52:21 sunno
1:52:23 hooked up so um rabbit in their last
1:52:26 software update
1:52:27 added um sunno which I hooked up my
1:52:30 account to it so now I can just say to
1:52:32 my rabbit write a song and it does it's
1:52:35 amazing I loved that yeah that was a
1:52:37 good song wasn't it thank you for
1:52:39 extending it Kyle I know it needed to be
1:52:41 extended it was like you can't end like
1:52:43 it was the middle of a good chorus
1:52:55 write me a song about being broken up
1:52:57 with in an East Village
1:53:02 Cafe got it please be patient while I'm
1:53:04 generating your songs so off it goes to
1:53:07 generate the songs cool right all right
1:53:11 let's listen to our songs
1:53:15 now
1:53:17 sorry while exting the request please
1:53:19 try again in a few seconds
1:53:24 please write a folk song about being
1:53:26 broke and up with in a New York City
1:53:29 East Village
1:53:32 Cafe got it please be patient while I'm
1:53:35 generating your
1:53:37 songs sorry an error occurred while
1:53:40 executing the request please try again I
1:53:42 must not be set up right sorry to budge
1:53:44 but can I get a consensus on the best
1:53:46 text to video text to video speech tool
1:53:53 text to video speech
1:53:58 tool so do you mean like a talking head
1:54:01 Avatar
1:54:04 so there's there's three major
1:54:06 contenders in the commercial space
1:54:08 there's Synthesia did and
1:54:13 haen all of their voices if you just do
1:54:16 pure text to
1:54:18 speech the like the the video animation
1:54:21 stuff's okay it's a little stiff
1:54:25 um what I do if I want it to sound
1:54:28 natural is I use 11 Labs has a thing
1:54:31 called voice to voice synthesis where
1:54:34 you record a script and assuming that
1:54:37 you can record good if you're good with
1:54:40 words you can record the script and then
1:54:42 swap out a different voice on top of
1:54:45 it so I would do that first for the
1:54:48 audio and then any three of those tools
1:54:51 is okay
1:54:53 they're they're not great yet they're
1:54:55 they're
1:54:59 [Music]
1:55:00 okays can
1:55:04 break shared memories in
1:55:09 cups tears fell for Old Time sake
1:55:15 oh it might be love
1:55:20 lost it might be just
1:55:24 [Music]
1:55:25 face under dim
1:55:30 lights our words began to
1:55:34 [Music]
1:55:36 fray we were to souls a drift in the
1:55:42 Brew of yester in the Brew of yesterday
1:55:46 that's
1:55:47 good it might be love
1:55:53 it might be just a
1:55:57 face we sat down with bitter words we
1:56:01 sat down with bitter
1:56:03 [Music]
1:56:04 words sounds like Nathan laying on a
1:56:09 bender in a quiet
1:56:16 Cafe the scent of coffee beans coffee
1:56:20 beans fill the air M our
1:56:25 [Music]
1:56:26 pain with every step we
1:56:31 knew we'd never be the
1:56:36 [Music]
1:56:37 same it might be love
1:56:42 lost or memories will
1:56:47 chase yeah I think it needs like it
1:56:49 needs a real chorus chorus written for
1:56:52 it and then you extend
1:56:54 it anyway all right let's go see if we
1:56:56 have a
1:56:57 video I think we do is this it
1:57:00 no where's my video where's my
1:57:04 video what' you do with my video you poo
1:57:07 poo heads Kyle Shannon here we
1:57:12 go you wanted a cafe a Sad Cafe Love
1:57:19 Song in a quiet Cafe
1:57:25 we sat where Hearts can
1:57:29 break shared memories in
1:57:34 cups tears fell for Old Time
1:57:39 [Music]
1:57:41 sake it might be love
1:57:44 [Music]
1:57:49 lost there you go so yes the answer is
1:57:54 yes you can you can just have this [ __ ]
1:57:57 write a song and then have it make a
1:58:00 video now the thing about noisy that I
1:58:03 don't know if they fixed is it only does
1:58:06 short
1:58:07 clips so I think what you'd have to do
1:58:09 is write the
1:58:12 song Break it into chunks have this do
1:58:16 chunks I don't like I don't know how
1:58:18 quite how You' do it but anyway that was
1:58:20 better than I thought it would be quite
1:58:22 frankly that was not
1:58:33 bad in a quiet
1:58:36 Cafe
1:58:40 Cafe like that's a nice shot that looks
1:58:43 like a different
1:58:45 girl
1:58:48 cups tears fell for Old Time sake
1:58:52 [Music]
1:59:03 right not
1:59:17 bad
1:59:20 um maybe the different girl is why
1:59:22 they're breaking up oh that's good Dr
1:59:25 J [ __ ] thruple gone
1:59:31 wrong you said she was just a piece of
1:59:34 [Laughter]
1:59:38 ass oh man man alive crazy right isn't
1:59:44 that
1:59:47 crazy and that's you know here's where
1:59:51 where my where my mind goes and this is
1:59:55 you know we talked about it a lot in
1:59:56 this
1:59:57 channel is we we are absolutely in the
2:00:01 embrace the Jank phase of these
2:00:04 tools it's not going to be that long I
2:00:07 mean that's not bad that video is not
2:00:10 that bad and that song's not that
2:00:14 bad imagine if like what did I put into
2:00:16 that 7 Minutes of effort kind of [ __ ]
2:00:21 around
2:00:24 I've said this before if you want to if
2:00:26 you want to here's a here's a fun thing
2:00:29 so I've got this in the AI Salon we've
2:00:31 got this journey to AI literacy that
2:00:34 we're kind of building the whole Salon
2:00:40 around mindfully create generously
2:00:45 lead if you want to have a
2:00:50 real breakr through with AI see I don't
2:00:54 have the discipline for this but if if
2:00:57 you've got good if you don't have
2:01:00 ADD and you can tolerate doing [ __ ] for
2:01:04 more than two days in a
2:01:06 row I think an absolutely fascinating
2:01:10 exper
2:01:12 experiment would
2:01:14 be spend a
2:01:17 week [ __ ] around with yudo and sunno
2:01:24 and [ __ ] around with that
2:01:28 noisy video generator and maybe with mid
2:01:31 journey and you know Runway the the new
2:01:34 Runway gen 3 thing like dick around with
2:01:37 the tools for a week and then pick pick
2:01:40 two pick a song generating tool and pick
2:01:43 a method of making a video and then
2:01:46 spend the next two
2:01:48 weeks deep diving on making music video
2:01:53 videos and then spend the final week so
2:01:56 this is a 4-week experiment a week of
2:01:59 [ __ ] around and experim finding the
2:02:02 right tool two weeks of just building
2:02:05 music videos and then spend the final
2:02:08 week telling everyone what you
2:02:12 learned share your videos share your
2:02:15 process share what you thought was
2:02:16 brilliant share what you thought was
2:02:18 shitty you'll get so good in that month
2:02:23 at that one thing that that'll be
2:02:28 mindblowing and so and so it's
2:02:31 basically doing this cycle on a single
2:02:39 project I think we're going to start
2:02:41 building guilds around this process we
2:02:45 we've got some ideas for how we're going
2:02:46 to how we're going to up the game with
2:02:50 guilds and turn them into these kind of
2:02:52 self-contained really cool production
2:02:56 units this could be the end of famous
2:02:59 people they're not going to like that
2:03:01 well it it becomes the end of famous
2:03:03 people or it
2:03:06 becomes
2:03:09 um famous people that never
2:03:14 die you're like
2:03:17 huh so
2:03:23 here I am I'm on
2:03:26 the this community always inspires me
2:03:29 awesome debie keep coming back that's
2:03:31 great to hear thank you for that that's
2:03:33 really
2:03:35 kind come on hang on let me quit let me
2:03:39 force quit
2:03:41 this there we go okay so this is 11 your
2:03:46 mind blown this is 11 Labs
2:03:49 reader so if you've got an iOS device go
2:03:54 right now to the App Store and get the
2:04:00 reader re e a d r app from 11
2:04:04 Labs what it does is it takes any text
2:04:07 file or any website and it turns it into
2:04:11 a podcast and you can use your voice if
2:04:14 you've got an 11 Labs voice that you
2:04:16 made you can choose one of their voices
2:04:19 or you can even change
2:04:23 the voice to Sir Lawrence Olivier or
2:04:27 James
2:04:28 Dean or Judy Garland so let's pick Judy
2:04:34 Garland mind and we're
2:04:36 gonna but it's his role as the creat
2:04:40 Judy Garland Embrace is going to read my
2:04:42 book embrace the Jank playing your way
2:04:45 to AI literacy quotes from Kyle Shannon
2:04:47 on the future of AI called for more than
2:04:49 750 hours of lives from the AI learning
2:04:52 lab on Tik Tok the
2:04:54 democratization learning lab
2:04:57 song right so let's
2:05:07 see here Let's do let's do sir Lawrence
2:05:10 Olivier I think Judy Garland's A Little
2:05:12 Too Sing songy for the for the depth of
2:05:16 future AI wizard yeah she's too sing
2:05:18 songy for the depth of embrace the Jank
2:05:21 so we're going to go sir Lawrence
2:05:26 Olivier all right future AI Wizards
2:05:28 strap in you're about to embark on a
2:05:30 wild Journey towards AI literacy and
2:05:33 this book is your irreverent guide
2:05:35 forget the stuffy Tech manuals we're
2:05:37 diving into the trenches of the AI
2:05:38 learning lab where we don't just
2:05:40 theorize about AI we wrestle with AI
2:05:43 Wizards me and Larry were like
2:05:46 this Larry who's Larry sir Lawrence
2:05:50 Olivier get what the [ __ ] Pro program
2:05:53 so oops there went the rabbit um
2:05:58 so this is going to be um right now
2:06:02 they've got four
2:06:05 voices but people that are currently
2:06:08 famous are going to start cutting deals
2:06:10 for their likeness and then we're going
2:06:14 to have Services where you could have
2:06:16 any Star you want you like Tom Cruz you
2:06:19 like uh what's his name
2:06:23 uh Ashton
2:06:25 Kutcher you want you want the young hot
2:06:28 Ashton Kutcher that was with demore to
2:06:31 read all your books he'll do
2:06:35 it make me a folk song about being
2:06:39 broken up with in an East Village
2:06:43 Cafe generating now this will take a bit
2:06:53 are we off all right I think we're
2:06:56 off sorry an error occurred while
2:06:58 executing the request please try again
2:07:00 in a few seconds no something's
2:07:03 something screwed up all right speechify
2:07:05 is similar BEC comes with Snoop Dog's
2:07:07 voice exactly it's
2:07:13 honest it sounds like an old radio play
2:07:15 yeah and wait till they start using the
2:07:17 the gp4 Omni real realtime voice that's
2:07:20 got emotional um con context content
2:07:24 it's going to be crazy it's going to be
2:07:31 wild oh
2:07:37 [ __ ] that's I made a mess back there all
2:07:40 right is the rabbit worth the money Kyle
2:07:43 um yeah so I paid 200 bucks for it I got
2:07:47 a year of perplexity perplexity Pro
2:07:51 which is worth 200 $100 for
2:07:53 free um I think it's the coolest
2:07:56 Hardware product I've owned in a long
2:07:58 time um and the software sucks right now
2:08:02 but they keep improving it
2:08:05 um and so and it doesn't require a
2:08:08 subscription so a $200 experiment and I
2:08:13 get a $200 piece of software for free
2:08:16 yeah like I I saw it as no risk so yes I
2:08:19 think it was worth the money now is it
2:08:21 worth the money like do I use it every
2:08:24 day no did I think I would nah I kind of
2:08:28 had a 10% hope but I didn't think I
2:08:32 would um which which is why I I secured
2:08:37 a place for it right
2:08:40 there I went to get the lap and it looks
2:08:42 like I already installed it oh the the
2:08:45 reader app yeah the reader app's really
2:08:47 cool I'm currently unfamous well welcome
2:08:51 to the club
2:08:52 most of us are I'm famous in my own
2:08:56 mind I'm going to Washington next
2:09:00 week to to testify to
2:09:06 congress not quite sure how that
2:09:08 happened but well I am
2:09:14 but I feel like one of the least likely
2:09:17 people that should be going to Congress
2:09:19 to testify on anything
2:09:22 but it's
2:09:25 exciting all
2:09:30 right where can we watch your testimony
2:09:32 I assume you can watch it on C-Span have
2:09:35 you seen Congress well that's true
2:09:37 that's
2:09:38 true I'm G to get there and go I could
2:09:40 [ __ ] do
2:09:45 this just downloaded getting ready to
2:09:47 play why because you're honest yeah I I
2:09:50 mean it's yeah cuz it's funny Cosmic
2:09:53 lover that's the the thing I don't get
2:09:57 the thing that I don't get about
2:10:02 politics is I just don't get the game I
2:10:05 don't get how people can just blatantly
2:10:08 lie like like they can literally be
2:10:10 caught in a lie and just go no I didn't
2:10:12 do that you're like but there's
2:10:14 videotape of you yesterday yeah that
2:10:16 wasn't me you've misunderstood that oh
2:10:19 you've taken it out of context like just
2:10:21 that ability to be that Shameless I just
2:10:24 I I can't it doesn't process in my brain
2:10:27 addiction to power it's all spin yeah
2:10:30 exactly
2:10:32 so yeah I would be one of those I would
2:10:34 be an unsuccessful politician I think
2:10:37 because I I would be like oh yeah I kind
2:10:40 of [ __ ] up there yeah that's a
2:10:43 [Laughter]
2:10:45 shame I did not have sexual relations
2:10:48 with that woman
2:10:53 oh my god what have we done to our
2:10:56 country it's so [ __ ] up right
2:11:03 now it's just
2:11:06 bizarre like one side is just Shameless
2:11:09 and the other side is incapable it's
2:11:13 like CH choose which version of awful
2:11:17 you want will it be on C-Span Kyle I
2:11:20 assume so apparently we're testifying in
2:11:23 front of a joint session of Congress
2:11:26 which holy [ __ ]
2:11:31 um I don't think I've even been in the
2:11:34 Capitol Building much
2:11:37 less testified to a joint session of
2:11:40 Congress so I have no idea what to
2:11:42 expect um I've got a prep call
2:11:46 Thursday so I'll figure out what to
2:11:48 expect but the AI salon so we talked
2:11:51 about the AI Salon before the AI Salon
2:11:54 is co-sponsoring this Flyin with Google
2:11:57 for small business so it's Google for
2:12:00 small business and the AI Salon are
2:12:02 sponsoring this
2:12:03 Flyin um so Leah fasten from the salon
2:12:06 is coming down the night before we're
2:12:08 going to host the dinner or you know
2:12:12 meet everyone at the dinner and then
2:12:14 Wednesday July 24th is when we talk it's
2:12:19 possible that the the plug gets pulled
2:12:21 because
2:12:22 Netanyahu is also speaking to Congress
2:12:26 that day so so we're testifying the same
2:12:30 day Netanyahu did um I don't think we're
2:12:32 going to upstage that
2:12:35 particular
2:12:39 session H so it's possible it gets
2:12:42 pulled if there's too many protests or
2:12:43 whatever if they don't feel it safe but
2:12:46 I don't [ __ ]
2:12:47 know I don't know people I don't know
2:12:52 did they vet you
2:12:54 yet I assume so I don't know what you
2:12:57 mean did they vet me like did they run a
2:13:00 background check so I can be in Congress
2:13:02 I don't
2:13:03 know um did they watch my Tik Tok lives
2:13:07 and go oh he swears I didn't realize he
2:13:10 was a cusser Oh
2:13:12 My Oh are are we sure we want to parade
2:13:16 him up in front of the congressman and
2:13:18 the Senators
2:13:22 he he says the f word fairly
2:13:26 liberally and what does this embrace the
2:13:28 Jank I don't even I think Jank is a
2:13:30 dirty word is it not is it isn't that
2:13:32 what the young kids are calling their
2:13:34 private parts these
2:13:36 days I don't know that this is
2:13:39 appropriate kylin I'm imagining you in a
2:13:41 Schoolhouse Rock cartoon I'm Just a
2:13:45 Bill you can make money with
2:13:48 J I'm taking this with me I don't know
2:13:51 if I'm going to use it I think I'll
2:13:52 certainly use it the night of the
2:13:54 dinner um but it's too much of an inside
2:13:57 joke to be
2:13:59 actually actually a useful prop but it
2:14:02 would be funny Democrats and Republicans
2:14:04 Play Good cap bad
2:14:06 cop racist political stagnation theater
2:14:10 it really is it's really just [ __ ]
2:14:12 crazy careful they might tackle you Kyle
2:14:15 you have to mention the AI Irregulars in
2:14:18 Congress so that it will get officially
2:14:20 noted well I'm to I'm I'm actually going
2:14:22 to a big part of what I want to talk
2:14:25 about is the people in this
2:14:27 community um so I'll probably mention
2:14:30 the Irregulars I'm definitely mentioning
2:14:32 the AI Salon we're sponsoring it so I
2:14:34 assume I probably go first out of the
2:14:37 bunch of people or maybe I go last I
2:14:39 don't
2:14:40 know
2:14:42 um I'm going to talk about AI
2:14:45 optimism in the face of
2:14:48 fear uh I'm going to talk about how how
2:14:51 I get to witness on a daily basis people
2:14:55 transforming their
2:14:56 lives and that if legislators don't get
2:15:00 educated about what's happening and and
2:15:03 regulate too
2:15:06 broadly in an effort to protect people
2:15:09 from the risks of
2:15:11 AI and not to empower them to take
2:15:16 advantage of the
2:15:18 transformation that AI makes possible
2:15:21 that they're going to hurt their
2:15:22 constituents so that's what I'm going to
2:15:23 talk about is optimism optimism in the
2:15:26 face of fear I don't quite know how I'm
2:15:29 going to frame it yet but that's the
2:15:31 basic
2:15:32 idea does Congress cover your travel and
2:15:35 hotel accommodations Congress doesn't
2:15:38 but uh the the group organizing it
2:15:44 does the group that invited me
2:15:47 does yeah and I don't know like I don't
2:15:50 know if this was a kind of thing that
2:15:52 you know Congress requested this or if
2:15:54 this group putting it together got it on
2:15:57 the docket like I don't know how this
2:15:59 [ __ ]
2:15:59 works I'm just there as a
2:16:03 puppet don't mention Tik Tock or some
2:16:07 may be
2:16:08 dismissive yeah [ __ ] it I don't
2:16:11 care I like they they can be dis they're
2:16:15 going to be dismissive if I'm optimistic
2:16:17 about AI
2:16:19 anyway oh he's just one of the
2:16:24 but I'm going to try to do what I do
2:16:25 here I'm going to try to demystify it
2:16:27 and I'm going to try to say I get to
2:16:29 witness people Reinventing themselves on
2:16:31 a daily basis and it's it's profoundly
2:16:35 inspiring press release there will be a
2:16:38 press release um going out to things
2:16:40 like Politico so the AI salon's going
2:16:43 going into
2:16:44 [Laughter]
2:16:47 [Applause]
2:16:48 Politico crazy [ __ ] people crazy [ __ ]
2:16:52 education reform was mentioned at the
2:16:54 RNC tonight it would be great segue to
2:16:56 mention AI in education I think I'm
2:16:59 going to mention AI in education I'm
2:17:01 going to I may talk about
2:17:04 ethics
2:17:07 um I I think I'm going to just talk
2:17:10 about
2:17:13 possibility that they need to get
2:17:15 educated enough they need
2:17:17 to explore enough to understand
2:17:21 what generative AI makes
2:17:24 possible Politico the podcast well I
2:17:26 don't know it'll go to Politico Politico
2:17:29 the the media organization I don't
2:17:33 know uh uh what else should we download
2:17:37 from the App Store anyone can
2:17:40 answer Pi chat GPT the official
2:17:46 app 11 Labs reader those are the only
2:17:49 three right now that I think are the
2:17:50 biggies
2:17:52 there there's some image generation ones
2:17:54 that are fine but you can use you can
2:17:57 use other tools for that you don't need
2:17:58 to buy those
2:18:00 apps all right I got to go without
2:18:02 limiting it before they understand it
2:18:05 yeah exactly Ryan that's exactly what I
2:18:06 want to say is like don't rush
2:18:09 legislation just to have it don't paint
2:18:13 with too broad a brush or you'll destroy
2:18:16 small and medium business and solo
2:18:19 entrepreneurs and actually take the
2:18:22 time to play with these tools long
2:18:25 enough to understand what they make
2:18:29 possible and I may give a a Kevin
2:18:31 McAllister moment I might I might say
2:18:34 play with generative AI tools like chat
2:18:36 GPT long enough to have your first Kevin
2:18:39 McAllister
2:18:42 moment I think that'll get picked
2:18:48 up make sure you download the official
2:18:50 chat GPT app yeah when you're when
2:18:52 you're looking for the chat GPT app
2:18:55 there are like a dozen clones that will
2:18:58 charge you money make sure that you get
2:19:01 the chat GPT app from open AI it's
2:19:05 really important any other platforms
2:19:07 covering the event I have no idea I
2:19:09 don't think
2:19:12 so um congrats Kyle you talked about
2:19:15 politics long enough to get the US
2:19:18 elections
2:19:19 Banner oh did
2:19:22 I I I don't [ __ ] know how Tik Tok
2:19:28 works thanks for being our rep peace out
2:19:30 Kyle thank you madcap I appreciate it
2:19:32 okay so do me a favor go to the AI Salon
2:19:37 we have a website now it's not just a
2:19:40 link tree anymore you can sign up for
2:19:42 our bi-weekly meetings and you can join
2:19:44 the online
2:19:45 community the first two links that you
2:19:48 see at the top are those two things and
2:19:50 then you know read read other things we
2:19:52 also have our position statement about
2:19:54 the California AI bills there
2:19:58 um I can't believe open AI said we
2:20:00 supported those bills that's really
2:20:03 funny um follow my
2:20:05 channel here's what I can promise you if
2:20:08 you follow my channel it won't show up
2:20:10 in my follower account and Tik Tok won't
2:20:13 show you my content because it's not how
2:20:15 it works I don't know so don't follow my
2:20:19 channel it doesn't [ __ ] matter but
2:20:22 bookmarket or something like that and
2:20:24 keep looking for my [ __ ] um what
2:20:26 happened to chat GPT 5 we'll see it
2:20:29 before the end of the year I
2:20:32 think um all right thank you Irregulars
2:20:35 thank you mods um as promised we did a
2:20:40 poll well you said you wanted to see
2:20:43 image video deep dive we made a song and
2:20:45 we turned it into a video I think we
2:20:47 accomplished
2:20:48 that um all right r room time Romper
2:20:52 Room magic mirror I see Maro I see Ryan
2:20:56 bessling I see madcap papa JJ Allan Mr
2:21:02 Studio what's happening Amelio's wife
2:21:04 hello
2:21:05 hello
2:21:07 Joey Call Me Al or Ai call me
2:21:12 AI Bob Hamilton Bob jelly bu's won you
2:21:17 can make money
2:21:18 with he hasn't won anything we're not in
2:21:21 a game show well tell him what he's won
2:21:24 anyway you've won a new
2:21:27 car Firefly and a house Wolfman
2:21:31 Clint proar
2:21:36 Planet Source Camp good to see
2:21:41 you BT comp Brandon thank you very much
2:21:45 25k without even begging 25,000 likes
2:21:48 we're up
2:21:50 26,900 beautiful y'all are tap
2:21:54 Masters all right I think I got everyone
2:21:56 here anyone knew that I
2:22:00 missed no Ryan bessing oh I said him
2:22:04 Cosmic lover good to see
2:22:06 you verifier Romper Room oh you're just
2:22:09 bringing back the
2:22:11 memories another gen xer in the house
2:22:14 Romer room magic mirror Won Won
2:22:18 Won all right everybody it's Tuesday
2:22:21 night I will be back tomorrow Wednesday
2:22:25 I think normal time
2:22:27 8:00 yeah I think that's it have a good
2:22:30 night hope you enjoyed tonight I know it
2:22:31 was kind of all over the place but
2:22:33 that's what we do here welcome to Chad
2:22:36 add peace have a good night everybody