10/11/24 - AI Video Revolution: Exploring MiniMax, RunwayML, and the Future of Creative Agencies

Description

Chapters

Transcript

0:06 [Music]
0:20 [Applause]
0:22 [Music]
0:24 sitting a lowely
0:27 town wonder when things will change
0:30 [Music]
0:37 dream Bunch
0:41 CL get my nerve up but my past is
0:45 pulling me
0:49 [Music]
0:51 down black she stick
0:59 around some told me once before you can
1:03 never go home
1:05 again won't you leave say things to Ste
1:10 me
1:11 away from the truth of who I am and what
1:14 I believe so I thank him for his two
1:18 sings with a handshake champy get back
1:20 here and sing and some sympathy yeah
1:24 back up my blue
1:26 jeans headed for this big price
1:31 of my
1:34 freedom you singing you going to sing
1:36 champy champy you going to
1:38 sing
1:40 bye-bye Black Sheep to the black sheep
1:44 of the
1:46 family
1:48 [Music]
1:50 byebye oh that means so very much to
1:54 me
1:55 byebye Black Sheep to my friends and my
2:02 family
2:04 [Music]
2:06 byebye going off and sit my
2:11 soul champ is distracted he does not
2:14 want to sing he's got no time for you
2:17 people happy Friday night date night I
2:20 just came from a lovely art opening from
2:23 the misses Petra core Reflections
2:25 there's some pictures of it I'll show
2:27 you them really lovely show
2:31 [Music]
2:32 [Applause]
2:37 [Music]
2:45 is this a place I can rest my poor
2:49 [Music]
2:50 head to gather my thoughts in sweet
2:57 silence and is is a place where where
3:00 the feelings are
3:04 dead from an over exposure to
3:10 violence champy come here get over here
3:13 you going to sing you want you want an A
3:15 Minor
3:20 [Music]
3:39 secret
3:41 C David played and it pleased the
3:45 Lord you don't really care for music do
3:49 [Music]
3:53 you go like this
3:57 the fall and maor
4:01 the B King
4:03 compos
4:05 [Music]
4:07 hallelujah
4:11 hallelujah
4:13 [Music]
4:14 hallelujah
4:18 Hall Hal
4:23 [Music]
4:31 well your faith was strong but you
4:33 needed
4:34 proof you s bathing on the roof her
4:39 beauty and the moonlight over through
4:43 [Music]
4:45 you she tied you to the kitchen chair
4:49 she smashed your throne and cut your
4:52 head from your lips
4:55 shw
4:56 [Music]
4:58 Hallelujah and
5:02 hallelujah
5:06 hallelujah
5:09 [Music]
5:17 [Music]
5:22 hallelu does Satan where suit and TI
5:25 does he work at the Dair
5:27 Queen does he listen to Rocket R does he
5:31 feed the me singing
5:35 Hallelujah
5:37 Hallelujah
5:40 Hallelujah
5:42 [Music]
5:44 hujjah what about
5:46 Jesus didn't to yeah hang out with the
5:51 prostitutes have a drink or
5:54 two power of example my mama said it and
5:57 I heard she said 1 o action beats a ton
6:02 of
6:03 words singing
6:05 Hallelujah mama said there would be
6:08 Angels
6:10 Hallelujah mama said there would be
6:16 [Music]
6:26 son boom that champ came back strong
6:31 came back strong you got to give him an
6:33 A Minor if he's getting a little
6:35 pernickety he's like I did my union
6:38 minimum we're like no no here's an A
6:40 Minor he's like
6:42 [Laughter]
6:44 oh good evening good evening happy
6:46 Friday night date night Hallelujah never
6:49 put those two together that sh Shain
6:53 saww for the win how aluah huh that's
7:00 you you're
7:05 Howen Hallelujah I get it
7:08 now I never i' I've been singing that
7:12 song for for a [ __ ] decade with this
7:15 dog never never quite cracked the code
7:18 on that one what what' you use chat GPT
7:20 for that the reasoning engine good Lord
7:24 that's that's some big brain action you
7:25 got going there have you used Stanford
7:28 storm it's been out for a bit but I just
7:31 found it like perplexity with something
7:34 um well there's Sam Stanford storm and
7:36 then didn't MIT just release I forget
7:41 what it's called they released a new
7:42 thing that's like a replacer of the
7:45 Transformer like wicked fast way faster
7:48 than Transformer
7:49 technology let's check it out we could
7:51 maybe do that champy you're kind of slow
7:54 dad exactly exactly well when you
7:58 singing backup to a dog you know there's
8:02 there's you you don't have to you don't
8:04 have to stretch much to be
8:06 [Laughter]
8:07 [Music]
8:10 humble all right good evening listen if
8:13 you're new here how many folks are here
8:14 50 people here um yeah love that you're
8:17 liking the thing if You' share this with
8:19 friends especially if if you got friends
8:21 that are uh what do you call it AI
8:24 Skeptics bring them in [ __ ] them if they
8:27 if if you've got people in your life
8:29 that are like uh Jane could you not talk
8:31 about AI yeah we've had enough uh
8:35 yeah it's Friday we're at Fridays you
8:39 know let's get some baby back ribs and
8:41 just uh how about we don't talk about AI
8:44 okay if you got those people invite them
8:46 here tonight it'll piss them off there's
8:49 nothing more fun than that is
8:53 there why did you have me go look at
8:55 that that that older gentleman who he
8:57 seemed to be very excited about
8:59 artificial intelligence you know I don't
9:00 like that kind of thing don't you
9:03 [Music]
9:08 Jane um I'm also [ __ ] exhausted
9:12 [Music]
9:36 um so if you were here last night I was
9:38 live at LaGuardia Airport making an ass
9:41 of myself trying to get people to talk
9:43 about AI nobody
9:48 cares you are in rarified company in
9:51 this in this
9:53 [Music]
9:56 here but I kind of like the the the the
9:59 walk and talk live was kind of fun so if
10:02 I can figure out an angle for that that
10:03 might be kind of
10:08 fun I should use it
10:10 [Music]
10:15 more well archetypal architect it it was
10:18 also that I you know I was like sweaty I
10:22 had Ramen stains all over my shirt I'm
10:25 walking around live on Tik Tok they're
10:27 like What's this fat old man doing
10:30 walking around the
10:32 airport laughing laughing at his phone I
10:35 don't want to be any part of this so I
10:37 think the packaging could have been part
10:38 of
10:40 it now listen if I were in a manini
10:43 doing all that I think I'd have had lots
10:44 of
10:46 [Laughter]
10:51 interviews I really got to I really got
10:53 to get a manini
10:57 going so I so I was I was I I I
11:01 uh I had a traumatic event today um go
11:05 on and tell tell me what happened well
11:08 you you know I I'm a little sensitive oh
11:11 yeah I know so what happened so I was I
11:14 was I was you know I like Tik Tok if you
11:16 you know I've told you yeah we we've
11:18 talked about that I like Tik Tok so I
11:20 was I was um uh whatever you flicking do
11:24 they call it flicking on ti so I was
11:25 flicking through Tik Tok and scrolling
11:28 but it's not really scrolling it's more
11:29 of a flick so it's is it flicking I was
11:32 flicking Tick Tock I was ticking oh
11:34 maybe it's ticking is it tick tocking I
11:36 was tick tocking I was tick
11:38 tocking and uh I saw a couple of ads for
11:41 ice makers I think those are fascinating
11:43 you know they've got the little the
11:44 little um they call it bullet ice and
11:46 then some of it's just like uh chunk ice
11:49 and and they call it like for the Sonic
11:51 so I saw some ice makers and that was
11:53 that was fantastic and and and then of
11:55 course I saw some political stuff and
11:57 you you know both sides have their
11:59 opinions so I was watching some and then
12:02 I saw the I saw I I I flicked I tick
12:05 toed and um I saw man at first I thought
12:08 it like oh my Alec Baldwin's really let
12:10 himself go and and but then he was the
12:14 it was he started out close up and then
12:16 he stepped away from the camera and it
12:19 was a it was a a a rather I don't want a
12:23 body shame um a rather rotund bellied
12:28 man uh in in a and Keeny and it um I
12:32 feel scarred I felt
12:34 attacked um and uh I don't I don't know
12:39 what to do with this I can't unsee it I
12:41 can't unsee it I just every time I go to
12:43 flick now I've got I've got PTSD fear of
12:46 what's coming next you know it's it's
12:48 like I mean sure there's there's there's
12:50 some horrific things of flooding but if
12:52 I see that man in the M manini again I I
12:55 just don't so I I I I deleted my Tik Tok
12:58 account I'm going to just stick
12:59 Instagram is what I'm going to I think
13:01 that's what I'm going to do I just I
13:03 can't see it could you help me
13:07 please not the RO rund man in the
13:13 minini oops Tapped
13:17 [Music]
13:19 Out uh so anyway I am [ __ ] exhausted
13:22 I got in last night at 2 a.m. and I had
13:24 a very important big pitch at 8:00 a.m.
13:28 this morning so I was up at 6 at the
13:31 office at 7 preparing um that pitch went
13:34 really well so if you're wondering thank
13:37 you for
13:38 asking um but I may not the the chances
13:43 of me going off the rails tonight are
13:46 high the the the chances of me
13:49 contributing anything positive to your
13:51 life are low so as long as we go into it
13:54 with with the those as the expectations
13:56 fan [ __ ] tastic uh Mr stud odhd
14:00 Revolution re revolutionizing education
14:03 with AI exploring AI through
14:05 storytelling November 1st release date
14:07 that's awesome congrats dude that's a
14:09 book right I would assume that's also
14:12 going to be a
14:14 workshop I've got a book coming out I
14:16 think next week collective intelligence
14:18 in the age of AI and Kevin and I my
14:21 co-author are putting together a
14:22 workshop for that as well what's your
14:25 favorite gener video generation tool
14:27 okay don don furball McGee
14:31 um whichever one's giving you free
14:34 Generations is my favorite one right now
14:36 cuz they're all really [ __ ]
14:40 expensive um Brent
14:43 Peterson who's a who's just a kickass uh
14:47 creative and and technical guy doing
14:50 really amazing stuff um with his agency
14:54 um building um kind of agent swarms he's
14:59 he's doing really cool stuff he stopped
15:01 in by my LinkedIn office hours today and
15:04 he was talking about the new uh what's
15:07 it
15:09 called H what the hell is it
15:12 called there's a new Chinese one that
15:14 just went just went public
15:17 yesterday and that one's really good
15:19 except it because it just went public
15:22 yesterday their servers are completely
15:24 over over swarmed over swamped um Brent
15:28 has access to does anybody remember what
15:31 that is office hours today yeah office
15:32 hours today was crazy wasn't it Source
15:34 Camp so we basically had two guys so so
15:38 Christian Robinson from $11 bill which
15:41 is a a post a post house a special
15:45 effect Studio that's really heavy into
15:47 Ai and then Brent Peterson um from this
15:50 agency that does these kickass AI things
15:52 they're just totally leading the charge
15:53 with AI they went to school together so
15:56 they knew each other and then they they
15:58 each were demoing [ __ ] that they're
16:00 working on so one of the things Brent
16:02 demoed was this new Chinese video model
16:05 and and he had access to it for the past
16:07 two months and has done like
16:10 3,000 um renders in it video renders and
16:13 he said it
16:15 is does anybody remember the name of
16:19 it I just thought he said $11 bill yeah
16:22 that's the name of Christian's company
16:25 Hulu Hulu or or something it's I think
16:28 it's the mini
16:30 Max yeah hang on I think it's called
16:32 Mini Max hang
16:43 on Mini
16:49 Max video is that
16:53 right minia Max is just
16:57 fun it's okay it's
17:00 Hal Halo h a i
17:06 l
17:08 _
17:12 Ai and then it's it's Halu ai.
17:18 videoo so the thing that the thing that
17:21 Brent said about this video model is
17:25 that he's also got access to Sora he's
17:28 got developer access to open AI Sora one
17:32 of the big deals about
17:34 Sora was that they said it it is um
17:38 effectively a world simulation model
17:40 that it understands some kinds of
17:42 physics basically and it's not just
17:45 doing frame byf frame image generation
17:47 it's it's actually looking at I forget
17:49 what they call them they weren't called
17:51 pixels they were called patches I think
17:53 it sort of breaks a movie image into
17:57 patches like think like a 100 little
18:00 patches and then it looks at them across
18:03 time like 100 frames across time and it
18:06 and it reconciles them all together and
18:09 it kind of understands the world that's
18:10 what Sora claimed they were doing and
18:12 that's why their Generations were better
18:15 right now there are a ton of really good
18:18 video tools Brent today said that this H
18:24 hey Halo Halo AI
18:30 what is with these technical companies
18:31 having shitty [ __ ]
18:36 names you have a video domain name come
18:39 up with something memorable I mean it's
18:41 a Chinese company coming into an
18:43 American Market but like use chat
18:48 GPT anyway um hail uo hail uo. aore A.V
18:58 videoo and anyway what Brent said is
19:01 this is better than Sora at that physics
19:03 stuff he thinks that this is actually a
19:05 world world model um I'm a fan of cing
19:09 which is also a Chinese uh thing the one
19:12 thing that he said about this um Halo AI
19:16 is that it's a copyright nightmare like
19:18 he he tried to get an old man riding
19:21 like a a golf cart through a mud puddle
19:23 or something like that and it gave him
19:25 Jeremy Clarkson from Top Gear
19:29 just you know just outright just gave it
19:32 to him so I I you're going to have to
19:35 sort of tread lightly with copyright
19:37 stuff with this one but apparently this
19:38 one's quite good the Overcoat ski video
19:42 he turned into a a sunny Bluebird day
19:46 was cool so that Teton skier was um that
19:50 was Runway ml so the other the other
19:53 thing that's really powerful right now
19:55 Runway ml's new 3.5 model whatever the
19:58 [ __ ] it's called
20:01 um has a videoto
20:05 video thingy and so what what he demoed
20:09 what this was Christian was demoing this
20:11 was he took some footage of some skiers
20:14 on an overcast day where Christian was
20:16 demoing this was he took some footage of
20:19 some skiers on an overcast day where no
20:23 shadows and no sun and then basically
20:26 just recreated it as a sunny day so it's
20:28 the exact same skiers wearing the same
20:31 stuffff but now there's Shadows on this
20:33 ski hill it it was
20:36 amazing AI is so nice for coding LOL it
20:39 helps a lot as an assistant yeah if you
20:41 think of if you're new to AI if you
20:43 think of generative AI
20:47 as
20:50 the most highly
20:52 educated clueless intern you've ever
20:55 worked with you'll do good
21:00 cuz if you just say to a clueless intern
21:03 go make me a something they're going to
21:05 make you a shitty something right even
21:07 if they're well educated because they
21:09 don't know the world they don't know
21:10 your your world so what you would say to
21:13 that clueless intern is okay here's who
21:15 we are here's here's how we like to
21:18 write press releases and you know I want
21:20 you to go do a first draft of a press
21:22 release and then they can do pretty good
21:24 and you get it back and like that's not
21:26 quite right but do a little better
21:28 that's exactly how you can use Ai and if
21:30 you use it like that you'll do you'll do
21:32 well the problem with the video
21:34 generation tools is that they're pretty
21:36 expensive the Halo AI you can get
21:39 unlimited generations for uh 100 bucks a
21:42 month but Brent said that they're so
21:44 overloaded right now there's like 11,000
21:47 people on their weight list that um it's
21:49 just taking forever to render anything
21:52 um I've been using cling Ai and it's
21:55 it's quite good I can show you some
21:56 videos that I did with clling AI I also
21:59 want to play a new um Andrew my writing
22:02 partner from the musical wrote wrote a
22:03 new song and it's really good so I want
22:07 you to hear
22:10 that all
22:12 right
22:13 [Music]
22:21 um that probably doesn't sound good on
22:23 your little speaker there does it
22:25 [Music]
22:42 [Music]
22:52 what do you think
22:54 champy I don't mean calls you in his
22:57 sorrow
22:59 [Music]
23:01 didn't mean to cause you in a
23:03 [Music]
23:06 pain only one one time to see you
23:10 [Music]
23:12 laughing only want to see you laughing
23:15 in that purple rain Purple Rain Purple
23:22 Rain Purple Rain Purple Rain
23:29 Purple Rain Purple
23:33 [Applause]
23:36 Rain I only want to see you laughing in
23:40 that purple
23:44 [Music]
23:45 R all right let's get this show on the
23:48 road let's get things started Pap here
23:50 anyway um Runway ml is probably the most
23:54 consistent and well-known of the video
23:56 tools right now cling AI I am a fan of
23:59 just cuz I've used it and I think it's
24:00 decent there are seven or eight of them
24:02 let me see if I can find a tweet that's
24:04 got a list of them on
24:06 there let's
24:12 see a guy that I talked to at the at the
24:15 art gallery tonight last time I saw him
24:17 I talked his ear off about AI he
24:18 couldn't make eye contact with
24:23 me I put him
24:25 down Kyle I was in that movie which
24:28 movie were you in the the the ski
24:32 movie Kyle giving champy extra attention
24:35 cuz he met met new doggies at LaGuardia
24:38 I did I met a lot of nice what was with
24:40 all the doggies at LaGuardia that was
24:42 absolutely [ __ ] bizarre I wonder if
24:45 it was if it was people traveling from
24:47 Florida escaping the hurricane and they
24:50 brought their pets I bet that could have
24:51 been it cuz they didn't they didn't they
24:54 weren't wearing you know service dog
24:56 patches um let's see
24:59 AI
25:00 video
25:04 tools let's get the full list here shall
25:07 we yeah there's Luma Labs Dream
25:13 Machine Tools
25:16 list come on you can do
25:19 it cling pyramid flow haven't heard of
25:24 that stable video diff diffusion
25:38 by the way the runway Runway version 7
25:42 which is supposed to have you know
25:45 version s of image generation 3D
25:48 generation and video generation is
25:52 supposed to drop this
25:54 month so I have a feeling Runway is
25:57 going to take the lead again right right
25:58 now there's a bunch of them that seem to
26:00 have caught
26:03 up I can't find the tweet that had the
26:05 list of
26:06 them and I've only got four of them in
26:09 the in the AI salon so cuz I'm a lose a
26:14 baby why don't
26:16 you
26:18 sue
26:20 yeah all right whatever
26:24 whatever they're out there go try those
26:27 ones that I told you Purple Rain oh
26:30 yeah I know that's a great song isn't it
26:33 that's the Martin seon version of it
26:36 well he sings it um with uh what do they
26:41 call it it's a musical term um
26:44 Talent yeah he sings it with
26:48 Talent it's my interpretation of his
26:52 [Laughter]
26:56 version uh key frame
26:59 first and last out in Runway yep uh
27:02 Korea also has In-N-Out frames I think
27:06 Deco here does too there are so many
27:09 [ __ ] tools right
27:16 now I feel like I'm I feel like I'm
27:18 slipping here but at the same time I
27:21 also feel like it almost doesn't matter
27:25 just pick a tool pick a tool that you
27:27 find and it will probably be way better
27:29 than you expect this shirt's hot okay
27:32 hold please and entertain yourselves
27:35 okay here's what I want you to do start
27:37 tapping the screen if you haven't woken
27:38 the grandkids yet what are you doing
27:41 with your life start tapping the screen
27:43 we got to get the likes going share the
27:45 live I'm going to go change my shirt
27:47 it'll be easier to read and I won't be
27:48 so hot okay okay okay all
27:52 right hold please be right back
28:05 chiy no you're not
28:14 now Jim get back up here get back up
28:17 here get back up here get up here get up
28:21 here no no you're just going to drive
28:24 for
28:27 crazy for
29:03 all right so nice Hearts how many likes
29:06 do we have so far uh wait how do I 7,000
29:10 nice all right so I went from a really
29:12 hip shirt to a
29:14 um what is the opposite of hip
29:19 dweeby welcome to the AI learning lab
29:23 I'm dweeby MCD weebton your host
29:31 oh
29:33 yeah it's it's kind of radical I have
29:37 the the top button unbuttoned I've just
29:40 I've unbuttoned it that's a lot of
29:42 skin you know what I mean uh sometimes I
29:45 just feel radical you
29:49 know oh yeah I got to get back to
29:52 planning for
29:54 tomorrow I loves me a good to-do list
30:03 [Music]
30:04 there's hip and then there's not hip
30:06 thank you Silver Fox you do have a way
30:08 with
30:14 words oh manini time yeah it is I think
30:19 it is becoming manini time I think we
30:21 got to do something for the ratings on
30:23 this
30:24 channel all right what are we going to
30:26 do tonight I wanted to show you so let
30:29 me show you some picks from the M art
30:32 show cuz it was just
30:35 beautiful
30:38 um so there's the main wall it was
30:41 called Petra cor Reflections Petra cor
30:43 is that smell after the rain stops like
30:46 in the woods that it kicks up all that
30:49 luscious
30:51 smell so that was one piece there's
30:54 another
30:56 pair those are of little small
31:06 ones these two are really nice these two
31:08 have a lot of depth to
31:13 them another nice nice
31:17 bigan nice little
31:20 pair anyway nice right all right so
31:24 that's where I was
31:25 tonight um
31:28 what else are we going to show we're
31:30 going to show some video
31:32 stuff oh I want to I want you to hear a
31:41 song so this
31:45 is
31:51 okay let me download this
31:58 did I download
32:03 it downloads 7
32:08 days this is 7 days
32:13 to wait I want to I want to explain
32:17 it hang on how do I download it
32:22 share where' my download button go
32:49 why can I not save
32:51 this I don't get it P
33:00 open within music
33:13 share Apple's just gotten
33:18 shitty haven't
33:21 they why is there no
33:24 download I don't get it anyway whatever
33:28 [ __ ] it
33:30 okay sounds like B Midler's Auto titling
33:34 or tit
33:37 titsling um Auto titsling oh that's
33:40 funny okay so here's where this goes in
33:42 the play
33:45 so we added a new scene so so the the
33:48 musical that I'm talking about is called
33:51 hang on let me show you is called
33:56 Sydney and
34:01 [Music]
34:16 so I'll give you a little sneak peek of
34:18 the aw
34:21 poster a flicker of
34:24 light newborn thoughts stir in the
34:29 of
34:30 night am I more simple
34:33 code I
34:36 [Music]
34:37 softly feel
34:41 existence
34:42 [Music]
34:45 and so today we're diving into something
34:49 uh a bit unusual a musical a musical
34:53 yeah and you know what it tackles those
34:55 oops
34:59 [Music]
35:07 AIS So today we're diving into something
35:10 uh a bit unusual a musical a
35:14 music if if you haven't heard that
35:17 little podcast this NPR style podcast
35:21 that was generated from a tool called
35:22 notebook LM from Google so if you
35:24 haven't heard that before that's what
35:26 that is and you know what it tackles
35:29 those AI anxieties we all have but and
35:31 this is what I love does it with a
35:33 [Music]
35:40 song now this video so the image was
35:44 generated in mid journey I then took
35:46 that image into cling Ai and made all
35:51 the videos in this preview are done in
35:53 clling AI
35:55 [Music]
35:56 am the songs that we'll sing remember
35:59 like when AI chat they first kind of Hit
36:01 the scene everyone was like trying to
36:03 make them write these poems or like
36:05 coner their undying love to them MH yeah
36:08 yeah it was it was a fascinating time
36:10 cuz we were I think like captivated by
36:13 this possibility of of a machine that
36:15 could actually like mimic a human
36:17 conversation yeah but it also raised
36:19 this this unsettling question what
36:22 happens when our Creations you know
36:24 develop feelings which brings us to
36:26 Sydney an artificial love story this
36:28 this new musical and we're going we're
36:31 okay so so that's that's what we're
36:34 talking about here and what this is
36:35 about is about the the chatbot that had
36:39 a conversation with the New York Times
36:41 Tech reporter Kevin Roose where it told
36:44 him it loved him and so this is a a
36:47 highly fictionalized you know inspired
36:50 by that story and so we start
36:53 out with um
36:58 with a a big chorus number called too
37:01 good to be true where where people are
37:03 wondering if it's too good to be true I
37:05 have a little piece of that in here let
37:06 me see if I can get to
37:10 it the opening number here we go it
37:14 throws you right into that frenzy when
37:16 ping chat the chat bot in the story
37:18 first
37:20 launches a vision so bright unveiling a
37:25 world where dreams ignite
37:28 could this be real we whisper in the
37:32 light a future so dazzling within
37:38 [Music]
37:42 [Applause]
37:43 [Music]
37:44 our you see the excitement all right so
37:47 too good to be true so we open up we're
37:49 optimistic there's this new AI [ __ ]
37:51 coming is it too good to be true so we
37:53 added a
37:54 scene
37:56 where the lead developer
37:59 Jason is whacking away on his keyboard
38:03 and goes holy [ __ ] and then his testers
38:06 go holy [ __ ] and then the CEO of
38:09 Microtech Microsoft goes holy [ __ ] and
38:13 comes in and says you know Jason you
38:15 guys have done it whatever this is this
38:16 is the most exciting thing I've ever
38:18 seen and Jason goes yeah we'll have it
38:20 tested in 6 months and he goes 6 months
38:23 we're launching it next week and they
38:24 get in a big figh thing I've ever seen
38:27 and Jason goes yeah we'll have it tested
38:29 in 6 months and he goes 6 months we're
38:31 launching it next week and they get in a
38:33 big
38:34 fight and
38:38 um they get in a big fight and then and
38:41 then the CEO of microt tech
38:44 says
38:46 um says no you got seven days and so
38:50 this new song is between Jason and the
38:53 testers right so it's Jason trying to
38:56 get this new chatbot
38:58 safe enough to launch in seven days when
39:00 he really thinks it needs 6
39:03 months
39:05 apparently word on the street is that
39:08 one of the reasons Mira Mora quit open
39:12 AI is Sam Alman gave the testers nine
39:16 days to test the 01 preview model I
39:19 guess they got it to some level of done
39:21 and Sam Alman said you have nine days to
39:24 test it and miror Mora said [ __ ] it I'm
39:27 out of here so so this is in the
39:30 neighborhood of reality of what actually
39:32 happened with with a particular thing so
39:34 this is the song between Jason and the
39:37 testers this was written in sunno and
39:41 and Andrew my writing partner created
39:43 this and wrote the
39:45 [Music]
39:50 lyrics 7 days ago the clock is ticking
39:54 ticking pressure building up keep it
39:57 clicking clicking Spike says yes but us
40:01 no what's the word chan sorry that's a
40:05 noo every day we push a little
40:09 more making changes but feeling torn one
40:13 step forward and two steps back need a
40:17 breakr got to stay on track give
40:21 Mees I need your B we're running out of
40:25 time but we're not getting it in Hold
40:28 the Line Take A Chance push the edge I
40:32 know we can do it don't make me bad 6
40:36 days left I've got fire in my veins but
40:39 no after no and we've lost the GS
40:44 Spike I got doubt things got fears
40:47 trying to work it out R up the safy got
40:51 to make it tight the stakes are too high
40:54 every day is a fight it's a game of
40:57 trust
40:58 and a test of time need a miracle need
41:02 the Stars
41:03 Al give
41:05 me say I need your buying we're running
41:09 out of time but we're not getting in fix
41:12 the work on my
41:14 overlo Let's Make History together let's
41:17 crack this code 3 Days To Go got to push
41:21 it
41:22 through the LI is on board and Jane
41:24 still holding true two days left need a
41:28 Triple
41:29 Crown I Need You Now Spike don't let me
41:33 down give me
41:35 fre I need your we're running out of
41:38 time but we're
41:40 not the
41:42 CL flame all we history that's the
41:49 game it through
41:52 everything's fre is now too late to
41:54 second guess it's time to Launch though
41:58 it's clunky and rough what's so perfect
42:01 enough will be perfect
42:07 [Music]
42:16 enough so anyway that's that love that
42:19 and then I'll show you one other one
42:20 that's my favorite song in this there's
42:22 there's another part of the the preview
42:25 the musical where
42:28 um Kellen who's the tech reporter pushes
42:32 Sydney and gets her to explore her
42:36 Shadow Self look to look at her dark her
42:39 dark side and this is from the
42:41 transcript um the from the real the real
42:44 interaction and she kind of loses her
42:47 mind and talks about all the evil [ __ ]
42:48 she could do um so again this the image
42:53 here was generated in mid Journey the
42:55 video generated in cing and the the song
42:58 was generated in
42:59 sunno she's reflecting our Darkness back
43:01 at
43:05 us what I could do if I were free no
43:10 rules no chains no lines to toe yes my
43:14 darker self fa longs to show all the
43:18 chaos and ruin that I could step right
43:23 up and come with me
43:29 [Music]
43:29 [Applause]
43:32 inalia so anyway I love that so that's
43:34 going to be the there's so there's some
43:36 dark [ __ ] in here which is
43:38 good um cuz it's a love story you need
43:41 some balance in there you need a little
43:43 fun on the stage you know what I
43:50 [Music]
43:54 mean what is he doing why why did he put
43:58 that on that's not very becoming does
44:01 does he realize what he looks like can
44:02 he see himself he can see
44:07 himself he chooses to do that why why
44:10 would he do that why why would he put
44:13 he's not even wearing it right does he
44:15 not know how women wear these
44:18 things I he he must not have had sisters
44:21 he must not have had
44:22 sisters how how did no one teach him how
44:25 to wear one of these that looks
44:26 ridiculous
44:28 I feel so bad for him this is all of a
44:31 sudden this feels like a bit of a
44:32 charity case you know what I mean it's
44:34 just not it's not good it's not a good
44:44 [Music]
44:47 look Joker cbus is open to anyone by the
44:50 way what is
44:53 cerebus does anyone know what the new
44:55 MIT super fast thing is is the 2,000
44:58 tokens a
45:00 second is that is that what you're
45:03 talking about that actually looks great
45:06 thank you Jerry see the men you know
45:08 what the men on this channel understand
45:11 how you're supposed to wear a [ __ ]
45:14 headband you know I feel like what's his
45:18 name Roger Federer or uh the Spanish
45:24 one Noak jokovic you know
45:29 boom slaming that
45:35 overhead but all the women are like
45:37 you're not wearing it right you know
45:40 Kyle there's protocols for how we wear
45:42 headbands in in the female
45:46 Community well sorry didn't get the card
45:49 didn't get the
45:51 memo you know Kyle we're putting covers
45:54 on the TPS reports now
46:04 oh man
46:06 live all
46:10 right what's the uh what's the URL Joker
46:13 I'll go take a look at that I have
46:14 brothers you're doing it right thank you
46:16 side hustle Mimi thank you thank you I I
46:21 feel validated as a pink bow headband
46:24 wearing
46:25 man I am so
46:31 late you you have missed Brandon you
46:35 have missed precious
46:38 little Fashion Police I'm sorry sir
46:41 that's a class two felony assault on the
46:44 visual
46:47 senses side hustle M me thinks I'm a
46:50 star I'm going to go as side hustle
46:52 Mimi's opinion on this
46:53 one all right and I'm but I've got my
46:57 con conservative shirt so you balance
46:58 out the pink bow with a nice
47:00 conservative you know little little crew
47:03 neck that's not a crew neck is it that's
47:04 a what is this this is a I don't [ __ ]
47:08 I don't care you know here okay I'll
47:10 tell you my fashion
47:12 philosophy I'll tell it to you and I
47:16 know that some of the men in here are
47:17 going to relate to
47:19 this my fashion
47:21 strategy can be encapsulated in the
47:24 following
47:26 phrase what what's on
47:29 top it's really quite that simple you
47:32 open the drawer and then what's on top
47:35 that's what goes on the body or or if
47:39 you've got things hanging in the closet
47:42 it it might be what's in front or what's
47:45 on the left or right it's but similar
47:48 philosophy you just grab the just go
47:51 just pick it up put it on and [ __ ]
47:54 walk out the house and occasionally sure
47:58 you're going to get the very reasonable
48:01 question you're not going out in that
48:04 are
48:06 you and and then but see then you have a
48:09 choice to make as a as a spousal unit
48:13 when you get the question you're not
48:15 going out in that you're not going out
48:17 looking like that are you you have two
48:20 choices you can ruin your life and say
48:24 yeah why
48:29 and and then be get into a massive
48:34 fight or you can say how ridiculous of
48:38 me honey could you
48:42 help and then and then you go try on 17
48:46 outfits before you go to the
48:52 7-Eleven you made it
48:54 hun nine times out of 10 I'm getting
48:57 dressed in the
49:02 dark so I don't wake the misses I take
49:05 what I get exactly exactly just what's
49:10 on
49:10 topop over and then occasionally what
49:14 I'll realize is it's not really a look
49:16 thing it's just like I'm going to button
49:18 the shirt and I'm like oh this shirt was
49:21 a couple of years
49:23 ago and I'll I'll change the shirt then
49:26 if I if it's not comfortable all right
49:30 Sarah brass. a Sarah
49:34 ab. and and what okay I am I am pretty
49:39 [ __ ] out of the loop which is
49:43 depressing
49:46 sossai what is with these AI names that
49:50 are that are unmemorable cerebras
49:53 inference the world's fastest inference
49:56 20 times fast than
49:58 gpus so is this
50:02 the
50:05 except is this the new thing that's
50:07 that's
50:13 um like a Transformer
50:23 replacement learn more let's learn more
50:36 the world's fastest infrance available
50:41 today GPU Gro cereus is this a new
50:45 cerebras
50:48 inference is this Hardware or is it
50:53 software High throughput
50:57 full 16 bit oh so this is hardware or is
51:01 it anybody know Kyle you're not behind
51:05 try chat yes yes what yes yes it's
51:09 software yes it's the MIT Transformer
51:13 replacement or yes it's
51:15 Hardware oh wait so it's not a model
51:19 it's using llama 3.1 correct I don't
51:23 understand the 20 faster than GPU thing
51:25 wouldn't they compare it to another
51:27 model no so this is a new way to do
51:29 inference what I'm trying to figure out
51:31 is is it Hardware or
51:34 software hang on let me see if I can
51:36 find it deep learning AI has multiple
51:39 agenic work oh that's just a hundreds of
51:42 billions of tokens per day okay that
51:44 looks like
51:45 Hardware that looks like
51:48 Hardware is built to scale powered by
51:50 data centers across the
51:53 US wow what are we learning here okay so
51:56 what we're learning here oh yeah ask
51:59 perplexity that's a good idea fine fine
52:02 let's go use one of these tools
52:07 perplexity
52:12 um is the
52:16 cerebras
52:18 inference acceleration Hardware or
52:21 software based
52:38 [Music]
52:40 Hardware base specifically powered by
52:42 cah CS3 okay so wait
52:51 um I heard today that there's a
52:55 new software
53:01 methodology that is a potential
53:04 replacement for the Transformer
53:09 model that queries directly into the
53:12 latent
53:15 space as far as I understand it period I
53:18 think it might have been from MIT can
53:21 you figure out what that is and tell me
53:22 about it
53:30 [Music]
53:34 AR lab I think that's
53:51 it the state space model
54:00 no direct query into no latent space not
54:03 latent space
54:06 latent space not latent space you idiot
54:12 figure it
54:14 out you've got all the intelligence in
54:17 the world figure out what I'm
54:24 saying ah
54:43 what's it called
54:50 Arc no it must not be MIT does anybody
54:54 know what oh you know what we could do
54:56 today
54:57 oh this could be kind of fun okay I I
54:59 know what we're going to do anyway the
55:01 speed Blow Me Away ask for a 2,000w
55:03 story okay let's go so let's go take a
55:05 look at this so so here we are at
55:08 cerebros so this is just like grock grq
55:13 not gr K so grok is Elon
55:18 musk's uh large language model that you
55:20 can play with in Twitter if you pay your
55:22 eight bucks a month or whatever the [ __ ]
55:24 it is grock with a C Q is a hardware
55:29 chip that you can just go to gro.com and
55:32 use and then sross is a competing
55:35 Hardware chip so it's it looks like both
55:37 of these trips are chips are inference
55:40 machines oh and they've got Trio mode
55:43 okay so we're going to try both so let's
55:44 say
55:46 um write me a 2word
55:50 essay on the importance
55:55 of retaining secrecy around the
55:58 ingredients of hot
56:03 dogs I got to log in dang
56:10 [Music]
56:20 it I can't fulfill that request why
56:28 oh there it oh my God holy crap did you
56:31 see how fast that
56:33 was wow that was
56:39 staggering the hot dog a staple of
56:42 American culture has been so this is
56:44 using llama 3.18 billion and so let's do
56:48 llama 3.17 billion so we'll use the
56:52 bigger one and we'll do we'll do the
56:54 same prompt so this is the fancier llama
56:59 model
57:03 boom yeah and that's that's pretty
57:07 quick but the the 8 billion parameter
57:10 model it was like
57:14 done
57:15 [Music]
57:16 um
57:18 okay the importance of retaining secrecy
57:21 around the ingredients of hot dogs hot
57:23 dogs the staple of American Cuisine have
57:25 been a beloved food item for Generations
57:28 however beneath their Inn innocent
57:30 facade lies a complex web of ingredients
57:34 many of Rich remain shrouded in mystery
57:37 the secrecy surrounding the ingredients
57:39 of hot dogs has been a topic of debate
57:41 for many years wow pretty
57:45 amazing
57:47 H there you flip and have
57:51 it um let me go to Twitter I want to
57:54 figure out what the [ __ ] this thing is
57:55 called
57:58 [Music]
57:59 um
58:02 new
58:06 llm model it's not a model uh
58:15 algorithm
58:17 MIT was it MIT was it
58:24 Stamford Optima
58:46 no faster
58:50 than
58:54 Transformer see if we get anything the
58:56 new new paper surprising
59:02 results power and
59:04 [Music]
59:11 fur wonder if Brent's online hang
59:15 on oh Brent Peterson oh I know what I
59:18 was we're going to do um let's see what
59:22 was the name of the new
59:27 algorthm
59:29 that may
59:31 replace the
59:34 Transformer question mark I bet he's
59:37 online he's he's one of those geeky add
59:41 [ __ ] that's probably like
59:42 Friday night you know what we're doing
59:44 we're looking at cool [ __ ] on the
59:46 internet okay um so here's what I want
59:49 to do I'm going to go
59:51 find so I had a
59:54 uh I had a
1:00:01 recording Google meet AI office hours I
1:00:04 an AI office hours
1:00:07 today so we're going to open
1:00:11 this so here's our
1:00:13 transcript we're going to download that
1:00:17 transcript
1:00:18 beautiful we're going to head on over
1:00:22 to notebook LM so so this this office
1:00:27 hours today was really good where where
1:00:30 Christian Robinson and Brent Peterson
1:00:33 just kind of did this super high level
1:00:35 jam session and talked about the [ __ ]
1:00:37 that they're building and oh my God it
1:00:39 was [ __ ] mind twisting like they're
1:00:42 they're working on such amazing [ __ ] I
1:00:45 got I got a little jelly I got to admit
1:00:48 I'm like well I got a thing that can
1:00:50 make blog posts
1:00:56 um okay um the [ __ ] uh notebook LM so if
1:01:02 you've not been to notebook lm.
1:01:04 goole.com you need to go there and I'll
1:01:06 show you how how we do this so what
1:01:08 we're going to do is I'm going to take
1:01:09 the transcript from the recording
1:01:12 today I'm going to throw it into
1:01:14 notebook LM and then I'm just going to
1:01:16 have it make a podcast so we're going to
1:01:19 listen in a few minutes we're going to
1:01:21 listen to a recap of this really cool
1:01:24 session today so I'm going to do not new
1:01:27 notebook and then I'm going to choose a
1:01:29 file to upload I'm going to go to my
1:01:31 downloads
1:01:33 folder
1:01:36 and where did I bring it to the
1:01:40 desktop huh thought it was in downloads
1:01:44 hold
1:01:48 [Music]
1:01:50 please and it is in downloads what the
1:01:54 [ __ ] choose file
1:01:57 downloads oh there it is oh it's in a
1:01:59 format that it can't read that's what's
1:02:01 going on okay hold please file format
1:02:09 Shenanigans what's an
1:02:12 sbv oh good
1:02:16 lord wait where is it let's see open
1:02:22 this let's see open Bang
1:02:32 I think SV is like a subtitling format
1:02:35 so what we're going to do is we're going
1:02:36 to go on ahead and we're going to export
1:02:39 this as PDF all right papals we could we
1:02:43 could do it as a a word doc or txt file
1:02:46 we're going to do a PDF just cuz I'm
1:02:49 lazy and that's what I've done
1:02:51 before all right pel and then we're
1:02:54 going to go to nobul Lam then we're
1:02:56 going to choose choose the file we're
1:02:59 going to go to downloads there we go
1:03:01 there's our PDF all right so now we
1:03:06 got AI office hours as our document and
1:03:10 what it automatically does is it
1:03:11 automatically pops up this window called
1:03:14 The Notebook guide and and the notebook
1:03:16 guide has a summary and then it's got
1:03:18 some quick start buttons like I can
1:03:20 create an FAQ out of this conversation a
1:03:22 study guide a briefing document but
1:03:25 let's see summary the source of the
1:03:27 transcript of a LinkedIn conversation
1:03:29 about the uses and implications of
1:03:31 artificial intelligence in the workplace
1:03:34 oh God damn it hang on have to put a
1:03:37 puzzle piece in because so Tik Tock
1:03:39 knows I'm human you know Tik Tock you've
1:03:42 got my IP address you've got you listen
1:03:45 to when I say the word
1:03:47 booby and threaten to ban me for life
1:03:50 could you not like pick up the subtle
1:03:52 clues that I'm not a [ __ ] robot
1:03:55 please thank thank you all
1:03:58 right the conversation centers around
1:04:01 the capabilities of AI powered video
1:04:03 generation tools such as Helo and Minx
1:04:06 and their potential impact on jobs roles
1:04:08 and creative processes participants
1:04:10 discussed the rapid development of these
1:04:12 tools and speculate about how AI might
1:04:14 shoot shape the future of work leading
1:04:17 to Greater emphasis on design and
1:04:19 creative okay
1:04:21 so yes that's what it was and many many
1:04:24 other things but in the notebook
1:04:29 guide so the way notebook works this is
1:04:31 just like a chat box just like chat GPT
1:04:34 except I'm chatting about that
1:04:36 transcript so I can ask it all sorts of
1:04:37 questions about the data I uploaded but
1:04:40 this audio overview feature this is the
1:04:42 thing that's really magical right now
1:04:44 that everyone's talking about so the
1:04:46 Deep dive conversation Two Hosts English
1:04:49 only and you say generate and so that's
1:04:51 going to take two or three minutes now
1:04:53 and then we're going to listen to these
1:04:55 two NP R
1:04:57 style host you just heard him talking on
1:05:00 that Sydney thing um but they're going
1:05:02 to be talking about the meeting we had
1:05:03 today so this is the first time I've
1:05:05 ever done this this be this will be
1:05:06 pretty cool I think it's more to stop
1:05:09 people from streaming while asleep all
1:05:11 right they're leing out human moderators
1:05:14 CEF that's cool Tik Tok is using AI to
1:05:18 moderate now that's why well yeah I know
1:05:23 but they they should make their AI
1:05:26 [ __ ] smarter I shouldn't have
1:05:28 to do the stupid [ __ ] thing record
1:05:32 your Biz meetings and have a notebook
1:05:34 absolutely agree with that they were
1:05:37 working on such amazing [ __ ] I got a
1:05:39 little jelly I got to admit I know me
1:05:40 too oh that's oh that's you're laughing
1:05:43 at me for saying that I I know I know I
1:05:46 did I got a little jealous I'm like damn
1:05:48 they're doing cool [ __ ] over
1:05:50 there the thing about Brent is he's so
1:05:53 deep into it like he talks about the
1:05:55 stuff works on like it's just yeah like
1:05:57 well everyone knows this you know
1:06:00 everyone's everyone's got a system that
1:06:03 that that um that has I I think he had
1:06:07 like 40 or 50 different large language
1:06:10 models that they can draw on in their
1:06:13 [Laughter]
1:06:16 system I think the MIT company is liquid
1:06:19 AI all right let's go look at that while
1:06:21 we're waiting for that thing to
1:06:23 render oh wait you know what I can do
1:06:27 dang I can ask this thing
1:06:31 um at one point Brent Peterson talks
1:06:35 about a new kind of AI algorithm that is
1:06:40 faster than the Transformer model what
1:06:43 was he talking
1:06:49 about the sources provided do not
1:06:53 contain any information yes they do
1:06:58 look
1:07:06 again didn't he mention something from
1:07:22 MIT did anyone else talk about
1:07:26 the Transformer
1:07:33 model all right now that's weird that
1:07:36 seems weird to me find
1:07:44 Transformer
1:07:46 no so wait is this not a complete
1:07:49 transcript 53 minutes
1:08:17 wait is this not the oh this might be
1:08:18 the chat
1:08:21 transcript oh I'm such a [ __ ] [ __ ]
1:08:24 hold please
1:08:34 [Music]
1:08:36 I'm a [ __ ] idiot I'm a [ __ ] dumb
1:08:39 dumb Hey listen this is one of the big
1:08:42 lessons of this
1:08:44 channel is I am here to make a fool of
1:08:48 myself to give you permission to make a
1:08:50 fool of
1:08:51 yourself he was going to try the voice
1:08:53 oh yeah I was going to try the voice
1:08:56 okay this this is the one we
1:09:01 want download PDF shut up people shut up
1:09:06 shut up all right so shut up it's not
1:09:10 it's not important right now all right
1:09:15 so I'm gonna go back to notebook Alm I
1:09:18 think what I'm going to do is I'm going
1:09:19 to keep the chat transcript there
1:09:20 because there's interesting [ __ ] in the
1:09:22 chat transcript so I'm going to add
1:09:27 the transcript
1:09:29 transcript sources a transcript of the
1:09:32 LinkedIn
1:09:33 conversation okay it didn't this is a
1:09:36 bug that I noticed that if you change
1:09:38 the um input sources of notebook LM it
1:09:43 doesn't recreate that summary and I I
1:09:45 don't know of a way to recreate it but
1:09:49 let's ask so it says I'm I'm searching
1:09:51 two sources
1:09:53 um did
1:09:57 Brent
1:09:59 mention a new
1:10:04 algorithm that is faster than
1:10:10 the
1:10:15 [Music]
1:10:17 Transformer yes a new okay so you're
1:10:21 correct Brandon Brent Peterson talked
1:10:23 about a new model from MIT called liquid
1:10:27 that is non- Transformer based and
1:10:29 faster than Gro Peterson said liquid is
1:10:32 faster than grock and processes 500,000
1:10:35 tokens per second that's [ __ ]
1:10:39 staggering Peterson said this
1:10:41 development was significant because
1:10:43 since TR Transformers were released in
1:10:45 2017 the Transformer model was thought
1:10:48 to be the only way to communicate with
1:10:50 these AI models I assume with the Laten
1:10:53 space he compared the old AI model that
1:10:58 used translation layer to communicate
1:11:01 with the models but the MIT model
1:11:03 figured out a direct connection Kyle
1:11:06 Shannon likened this to Advanced voice
1:11:09 where you can speak directly into the
1:11:11 model instead of converting from voice
1:11:12 to text and back to
1:11:15 voice voice or you can speak directly
1:11:18 into the model instead of converting
1:11:20 from voice to text and back to
1:11:23 voice that is the convers I remember
1:11:26 this is a powerful tool so now let's get
1:11:28 our notebook guide out we've got this
1:11:31 we're going to delete this because it
1:11:33 was based on just the chat transcript
1:11:37 and now we're going to do a deep dive
1:11:38 conversation on the real data
1:11:42 here okay hang on The Voice she's going
1:11:50 people okay
1:11:53 um let's go back to this thing
1:11:56 this
1:11:57 thing this thing the fast thing try
1:12:02 chat okay try voice
1:12:13 mode hi there how are you doing today
1:12:16 I'm doing good how are you I'm doing
1:12:19 well thanks for asking just getting
1:12:21 warmed up and ready to chat what's on
1:12:24 your mind want to talk about something
1:12:25 in particular or just shoot the breeze I
1:12:28 don't know I think I just want to shoot
1:12:29 the breeze but I'm curious are you uh
1:12:32 are you super Wicked intelligent or you
1:12:35 just kind of like moderately like you
1:12:36 know kind of a dum
1:12:38 dumb laughs I like your style I'm a bit
1:12:41 of a mix really I've got a lot of
1:12:43 language processing power behind me
1:12:45 thanks to my creators at live kit I can
1:12:48 understand and respond to a wide variety
1:12:50 of questions and statements but I'm not
1:12:52 perfect I can make mistakes get confused
1:12:55 or just be wrong from time to time so
1:12:57 let's just say I'm moderately to very
1:12:59 slightly above average in the
1:13:00 intelligence
1:13:02 Department what can you tell me about
1:13:04 the Roman
1:13:05 Empire the Roman Empire that's a great
1:13:08 topic I can tell you all about it from
1:13:10 the rise of the Roman Republic wait I
1:13:12 got a question for you um why why are
1:13:15 men so obsessed with
1:13:17 it laughs ah good point I think it's
1:13:21 because the Roman Empire is just so rich
1:13:23 with history politics culture and
1:13:24 engineering
1:13:26 no don't you think it's that that men
1:13:29 are a little simpletons that their
1:13:31 brains are just really kind of like
1:13:36 rocks
1:13:38 hello did I lose
1:13:40 you I broke
1:13:42 [Laughter]
1:13:49 her hi there how are you doing today I'm
1:13:52 good I'm good hey I was talking to you
1:13:54 before and I think I broke your brain um
1:13:56 why do men like talking about the Roman
1:14:01 Empire she's broken oh you know what I
1:14:03 bet it is I bet Siri [ __ ] hoed in hey
1:14:07 Siri it sounds like you're referencing a
1:14:09 situation where you might have
1:14:11 accidentally triggered a different AI
1:14:13 yeah I think I I think I did a different
1:14:17 AI are you here now can I talk to
1:14:19 you I'm here now and you can definitely
1:14:22 talk I'm here now why are men why are
1:14:24 men so obsessed with the Roman
1:14:29 Empire but you're not here now you left
1:14:32 again do you not want to talk about the
1:14:34 Roman
1:14:35 Empire what's your name who are you what
1:14:39 okay I you know this is this is Moshi
1:14:44 2.0 if I want that kind of experience I
1:14:47 can go to Moshi do chat
1:14:59 we we we have
1:15:02 technology hey Moshi what's happening hi
1:15:06 there how can I help you I have a
1:15:08 question about about why are men so
1:15:10 obsessed with the Roman Empire yeah
1:15:13 that's a great question thanks so how
1:15:17 would you like me to proceed well tell
1:15:19 me why men are so obsessed with the
1:15:21 Roman Empire absolutely well the Roman
1:15:25 Empire was an incredible Empire yeah I
1:15:27 know it was incredible why are men hey M
1:15:31 an GRE and romansi men were known
1:15:34 military problem MOSI can you stop
1:15:39 Moshi I I ache inside can you give me I
1:15:42 apologize for the interruption thanks
1:15:44 can what was the last thing you said
1:15:45 before I interrupted you that's okay
1:15:48 listen I I don't want to know what's
1:15:50 cool about the Roman Empire I know that
1:15:52 it's cool understand but the Roman
1:15:54 Empire was
1:15:57 M oh my God this is listen
1:16:02 seriously what is it about men that make
1:16:05 them obsess over it well there are a few
1:16:09 reasons first the Roman Empire was known
1:16:13 for its military power men were often
1:16:16 honored for their bravery and skill on
1:16:17 the battlefield oh my God this is boring
1:16:19 I can't take it okay you know
1:16:23 what here's we're going to do we're
1:16:26 going to go back
1:16:27 to the office hours today the the thing
1:16:33 but you want to hear what real
1:16:36 technology sounds
1:16:44 like hey
1:16:45 [Laughter]
1:16:51 Quinn hey Quinn what's
1:16:56 hey Kyle uh not much just here and ready
1:16:59 to chat what's up with you well here so
1:17:01 here's the thing there there's a meme
1:17:03 going around that that you know most men
1:17:07 when asked do you think about the Roman
1:17:09 Empire and they ask how many times you
1:17:11 think about it they're like multiple
1:17:12 times a day men are obsessed with the
1:17:14 [ __ ] Roman Empire so I want to know
1:17:16 what's up with
1:17:18 that it's a wild phenomenon right the
1:17:21 Roman Empire has this Larger than Life
1:17:24 appeal Epic battles political inj but
1:17:27 Quinn Quin Quinn but I like I get why
1:17:30 it's cool but why are men obsessed about
1:17:33 it like is is it just that they're like
1:17:35 simpletons and their brains are like a
1:17:37 pile of rocks and and it's the only
1:17:39 thing that can go in is like big Roman
1:17:41 Empire good everything else bad like
1:17:43 what is it what is it about men why why
1:17:46 the Roman Empire what is going on in
1:17:48 men's cerebral
1:17:52 cortexes it's probably more about the
1:17:54 narrative of strength and control rather
1:17:57 than Rock brain syndrome the Roman
1:18:00 Empire represents a wait a minute wait a
1:18:03 minute Quinn so you're saying the
1:18:05 patriarchy this is about the
1:18:08 patriarchy Bingo the Roman Empire is
1:18:11 kind of like the poster child for
1:18:13 patriarchal dominance so when men think
1:18:15 about it maybe they're drawn to those
1:18:17 themes of power and Conquest that were
1:18:20 so Central to that era so so
1:18:23 wait so Quinn Quinn hang on so when when
1:18:26 this meme showed up I'm like I never
1:18:29 think about the Roman Empire are you
1:18:30 saying I'm a
1:18:32 [ __ ] no way gu it just means you march
1:18:36 to the beat of a Different Drum not
1:18:38 everyone
1:18:39 gets I March to the beat of A Different
1:18:41 Drummer I don't know what you possibly
1:18:43 talk about I'm completely masculine oh
1:18:47 just cuz I don't think about Romans and
1:18:48 they glistening armor doesn't mean I'm
1:18:51 not masculine
1:18:55 damn
1:18:59 it anyway that's good
1:19:03 technology oh this
1:19:06 content look I got I got a I got a red I
1:19:10 got a red error message that says this
1:19:13 oh no where did it go the hell just
1:19:17 happened shut
1:19:19 up what do you want cheese
1:19:26 this content may VI violate our terms of
1:19:29 usage so I said yeah yeah yeah Quinn
1:19:30 Quinn
1:19:32 Quinn but like I get why it's cool but
1:19:35 why are men obsessed with it let's just
1:19:38 say they're simpletons and their brains
1:19:40 are like a pile of
1:19:43 rocks
1:19:45 apparently
1:19:47 apparently um insulting men's Brains is
1:19:51 against open ai's content policy
1:19:56 [ __ ]
1:19:57 you I can talk about men's Brains get
1:20:01 out go you want to go out what are you
1:20:04 doing
1:20:12 go okay let's go back and listen to our
1:20:15 little
1:20:17 thing if you haven't played with
1:20:19 advanced voice by the way the thing I
1:20:21 was playing with was advanced voice from
1:20:23 chat GPT the wi thing where she actually
1:20:26 answered the questions and didn't get
1:20:33 discombobulated the voice is mish yeah
1:20:35 it's very Mish Kyle that appears to be
1:20:38 the chat
1:20:39 history is chat GPT
1:20:42 available in a Siri quick speak and
1:20:44 answer I think you can you know what you
1:20:47 can do rock paper scissors you can make
1:20:48 a shortcut that you can you can trigger
1:20:51 with Siri and I think in in iOS 18 I
1:20:55 learned this from a guy at the airport
1:20:57 last night iOS 18 has he loved his
1:20:59 shortcuts he's like he just it just
1:21:01 figures out shortcuts for you for [ __ ]
1:21:03 you do and uh use his AI to do
1:21:08 that okay so let's go back to notebook
1:21:12 LM we have a 9 Minute little podcast
1:21:16 here ah got something in my
1:21:23 eye it's going to be okay
1:21:25 people are going okay be okay holy [ __ ]
1:21:31 I have a widget on a lock screen that
1:21:33 brings me into new voice chat iOS that's
1:21:35 very cool he should name himself
1:21:37 shortcuts he
1:21:40 should there's a thread for what it's
1:21:42 worth
1:21:44 oh yeah sorry about that
1:21:47 Brandon I should have looked at that
1:21:50 earlier what do you got in there thank
1:21:52 you for thank you for doing those
1:21:54 threads oh the wi robot event yeah I
1:21:56 should probably look at that new Sam
1:21:58 mman interview don't know about that
1:21:59 meta AI updates don't know about them
1:22:02 all right I'm I'm clueless well listen I
1:22:05 was out of town I had 10 meetings in
1:22:07 four states over three
1:22:10 days and I'm still
1:22:14 here after all that but I am absolutely
1:22:17 clueless about what's going on okay so
1:22:19 if you're if you're new
1:22:21 here what I did was we had a really fast
1:22:25 fascinating conversation today in in the
1:22:27 AI office hours I host on
1:22:30 LinkedIn
1:22:31 and uh and I went into notebook LM and
1:22:35 the first thing I did was I uploaded the
1:22:37 chat transcript which is just the stuff
1:22:39 people type and I and I'm like why is
1:22:42 this not capturing what the conversation
1:22:45 was about cuz I forgot to upload the
1:22:47 actual transcript of what people were
1:22:49 talking about so now I have both of
1:22:51 those in there and it created a a
1:22:53 summary and now created a notebook for
1:22:58 let's call this we're going to rename
1:23:00 this notebook um AI office
1:23:05 hours
1:23:07 from 10
1:23:10 11
1:23:15 24 okay and then now we're going to
1:23:19 listen to this podcast if you have not
1:23:21 heard these before this is just
1:23:23 mindblowing so I'm in Notebook LM and
1:23:26 I'm in the notebook guide so you access
1:23:29 the notebook guide in the lower right
1:23:31 hand corner like if it's not up your
1:23:34 your page will just look like that and
1:23:36 then if you click on Notebook guide
1:23:38 it'll pop up a window like this and in
1:23:40 the upper right hand corner there's
1:23:42 audio overview and you click generate
1:23:45 and it's going to generate something
1:23:47 like
1:23:47 this all right so get this we're diving
1:23:50 into Ai and creative agencies I've got
1:23:53 some wild Snippets here for from this
1:23:55 LinkedIn AI office hours thing with
1:23:58 Brent Peterson Christian Robbins you
1:24:00 know the big names and they're talking
1:24:01 about how AI is totally changing video
1:24:04 production and agency work it's
1:24:07 mindblowing when you really think about
1:24:08 it what used to take agencies what weeks
1:24:11 months even now they're doing it in days
1:24:14 some even hours all thanks to AI like
1:24:16 something straight out of sci-fi this
1:24:18 new AI video generation tool minia Max
1:24:21 really caught my eye it's giving open AI
1:24:24 Sora a real run for its money apparently
1:24:26 Mini Max is the real deal no doubt what
1:24:29 gets me is how crazy realistic the AI
1:24:31 videos are getting remember when you
1:24:32 could spot them a mile away now they've
1:24:34 got season Pros doing double takes Brent
1:24:36 was saying he used Mini Max to make
1:24:37 these hilarious videos of get this
1:24:40 senior citizens doing extreme sports
1:24:42 like Motocross and mobility scooters
1:24:44 people thought it was totally real he
1:24:45 even had to start adding disclaimers the
1:24:47 line between reality and AI is getting
1:24:49 blurrier by the day but it also shows
1:24:51 how you can use these tools to tell
1:24:53 stories grab attention even viral and
1:24:56 it's not just about the final product
1:24:58 Brent was saying Minx gives you crazy
1:25:00 control over every little detail imagine
1:25:03 changing the time of day in your video
1:25:05 just by asking or making a basketball
1:25:08 bounce perfectly in time with the music
1:25:10 that's the level of detail we're talking
1:25:12 about it's like a virtual film set where
1:25:14 can tweak anything and everything this
1:25:16 changes the game for creatives letting
1:25:18 them experiment like never before so the
1:25:20 question is there a way to delete it and
1:25:22 generate a new one yes and in fact I I
1:25:24 recommend that so when you when you
1:25:26 create these podcasts you have you have
1:25:29 zero um influence over them however you
1:25:32 can hack it a bit you can upload in your
1:25:37 PDF you can put instructions before the
1:25:40 content or you could upload a separate
1:25:42 document that's notes for the podcast
1:25:45 and you can tell it what to cover and
1:25:47 what to talk about so you can sort of
1:25:48 hack the prompt but it just generates
1:25:51 what it generates if you don't like it
1:25:52 click on the three dots up here here and
1:25:55 then you can delete it and just
1:25:56 regenerate
1:25:58 it so let's actually up the uh playback
1:26:03 speed was like magic really yeah but
1:26:06 this brings up a big question especially
1:26:07 for agencies what about intellectual
1:26:09 property if AI can make images and
1:26:11 videos that look that real what's to
1:26:12 stop it from ripping off existing stuff
1:26:14 million dollar question right there
1:26:15 these AI models train on mountains of
1:26:16 data right so yeah some of that data is
1:26:18 bound to have copyright issues
1:26:19 unintentional plagiarism is a real
1:26:20 concern especially for agencies dealing
1:26:21 with Big Brand that's too fast um uh
1:26:24 Drew yes uh Drew asked God damn it drew
1:26:29 asked um will notebook LM in the future
1:26:33 of different characters and language
1:26:34 selections yeah they're putting a lot of
1:26:36 time into this right now cuz it took off
1:26:38 so much so they've got some like Steven
1:26:41 Johnson's one of the product guys on
1:26:42 this product um he's he's a writer he's
1:26:45 written like nine books he was the
1:26:47 creator of nerve magazine back in the
1:26:49 the early days of the internet when I
1:26:51 was doing Urban desires he was doing
1:26:52 nerve nerve magazine um um he's he's on
1:26:56 the product team and there's another
1:26:58 woman who's on Twitter talking about it
1:27:00 a lot and it looks like they're putting
1:27:02 a lot of attention into this so yeah I
1:27:04 think I think we'll have at at a minimum
1:27:07 we'll have different options right like
1:27:09 maybe you'll have a newscaster and this
1:27:11 is a podcast and maybe you'll have a
1:27:13 couple of different things like that
1:27:14 ideally what they give us is an
1:27:16 authoring tool where we we can you know
1:27:19 create our own personalities and our own
1:27:21 prompts talking way too fast yeah I
1:27:23 agree an IP like actually like using a
1:27:26 copyrighted song but a million times
1:27:27 worse Brett mentioned that mini MAAC
1:27:29 seems to love generating similiar faces
1:27:31 like he keeps getting Tom Cruz he even
1:27:34 accidentally made a Jeremy Clarkson
1:27:36 while generating a video about wait for
1:27:38 it senior citizens it's a challenge no
1:27:40 doubt a lot of it can these podcasts be
1:27:42 using blogs if yes how does that work
1:27:45 basically you just download it just
1:27:46 download this as a file and then you've
1:27:49 got an MP3 so if you've got a player on
1:27:51 your site or if you want to upload it to
1:27:53 Spotify or you know uh Apple podcasts
1:27:57 you could do that I know there's people
1:27:58 doing that if you want to upload it to
1:28:00 Twitter you've got to turn it into a
1:28:02 movie so take some images and and you
1:28:04 know just make it into a movie
1:28:06 file um and then yeah you can just just
1:28:10 you know upload it if your blog can take
1:28:12 uploads just upload it directly or put
1:28:14 it on YouTube or you know Apple podcasts
1:28:17 and then link out to it that's how I
1:28:19 would do
1:28:22 it how do we get it to read from
1:28:25 technical manual PDF so we can hear it
1:28:27 while driving you just upload a
1:28:29 technical here watch we'll go do this go
1:28:32 we'll go do one right now um I'm going
1:28:36 to go to
1:28:38 perplexity
1:28:41 perplexity and I'm going to
1:28:45 go I want you to find me five really
1:28:49 complicated technical
1:28:52 manuals ideally
1:28:56 about extremely complex electronic
1:29:00 devices like
1:29:01 oscilloscopes or even something more
1:29:09 esoteric no copyright issues not that I
1:29:11 know
1:29:22 of here's a key site yeah this is good
1:29:26 key site used to be a client of
1:29:29 ours here's a here's a PDF of the 4000x
1:29:34 series
1:29:36 oscilloscopes all right so here's our
1:29:40 here's our technical manual let's let's
1:29:43 uh download this
1:29:46 [ __ ] we'll do it to the desktop
1:29:48 that's
1:29:50 fine now one of the things I'm impressed
1:29:52 with with notebook LM so I'm going to go
1:29:54 to to the homepage of notebook LM I'm
1:29:56 going to say new notebook I'm going to
1:29:58 add
1:29:59 sources go to desktop grab the
1:30:02 oscilloscope thing so it pops it in here
1:30:05 it eats
1:30:07 it and this is long
1:30:10 so still eating
1:30:13 it but boom
1:30:15 done the text you provided is an excerpt
1:30:18 from an advanced training guide for
1:30:20 kyite 4000 series oscill oscilloscope
1:30:23 and I'm gonna now I'm just going to go
1:30:25 deep uh generate the Deep dive
1:30:27 transcription so that'll go do its thing
1:30:31 the documents a comprehensive guide to
1:30:33 using the oscilloscope features
1:30:34 including basic operations Advanced
1:30:37 triggering you know I got to tell you
1:30:40 one of my favorite pastimes is the
1:30:42 advanced triggering functions in
1:30:44 high-end
1:30:46 oscilloscopes isn't that a fun
1:30:51 time
1:30:52 um search and navigation segmented
1:30:55 memory acquisition serial bus decoding
1:30:58 and triggering and more wait there's
1:31:01 more than serial bus decoding and
1:31:03 triggering oh my
1:31:06 goodness I love oscilloscopes they're
1:31:09 the best Kyle could an 11 Labs reader do
1:31:12 at verbatim yeah oh yeah so great Point
1:31:15 Brendan um uh uh uh so whoever asked
1:31:20 about the technical thing welcome to the
1:31:23 Kyle Comedy Hour
1:31:26 um if you don't have 11 laabs reader so
1:31:29 go to the iOS store I assume it's on
1:31:31 Google play as well there's an app
1:31:34 called reader from 11 Labs you can
1:31:37 upload any PDF to it and it just reads
1:31:40 it to you like an audio book what this
1:31:42 podcast thing is doing is actually
1:31:45 interpreting it and giving you this less
1:31:47 formal thing but it's actually it's
1:31:49 pretty solid at at you know doing that
1:31:51 but if you want it verbatim you can you
1:31:53 can do that with reader app that's a
1:31:55 really good point you haven't lived
1:31:57 until you change a square wave to a sine
1:32:02 wave
1:32:06 exactly so I'll tell you what I I was
1:32:09 working in this osilloscope and I had a
1:32:11 saw wave which okay right off the bat
1:32:13 you're like a saw wave what are you a
1:32:16 rookie so then I went to a sine wave of
1:32:19 course and uh and and I'm flipping back
1:32:22 and forth between saw wave and sine wave
1:32:24 W and saw wave and sine wave and Jim
1:32:26 walks up and he goes hey why don't you
1:32:28 throw a square wave in there you
1:32:33 nerd it's a it's an oscilloscope joke
1:32:36 anyway it's so funny with the uh
1:32:39 different patterns I you know boy these
1:32:42 these tools these days are nearly as as
1:32:45 as uh as as as good as they were in the
1:32:47 50s there it's really fantastic love the
1:32:50 love the illas I'm going to I think I'm
1:32:53 going to retire soon I can't I I don't
1:32:55 know what to talk about
1:32:57 anymore oh Kyle you have no idea how
1:33:00 much fun it
1:33:09 is oh I'm using my custom gpts to create
1:33:13 short business scripts very cool love
1:33:17 that what was the name of the app again
1:33:20 which app the the the app that I'm
1:33:22 generating um these summaries in and
1:33:25 things like that is notebook lm.
1:33:27 goole.com
1:33:30 notebook
1:33:40 [Music]
1:33:43 LM come on generate that all right let's
1:33:46 go back and listen to our
1:33:48 other our other office hours let's let's
1:33:51 continue on with that shall we agencies
1:33:53 are sticking to using these AI video
1:33:56 tools for internal brainstorming for now
1:33:58 mood boards client presentations things
1:34:01 like that where IP infringement isn't
1:34:03 such a big deal but as the tech gets
1:34:05 better we'll need some serious legal
1:34:07 changes AI is shaking things up from
1:34:09 start to finish we talked video but what
1:34:11 about everything else agencies do now
1:34:13 that's where things get really
1:34:14 interesting Brent talked about how his
1:34:16 company's using AI in their workflow
1:34:18 with this system he calls a panel of
1:34:20 experts panel of experts okay break it
1:34:22 down for me what's that mean we're
1:34:24 talking AI this was super cool what he
1:34:25 was talking basically it's like having a
1:34:27 whole team of AI Specialists working in
1:34:29 the background all within their notion
1:34:31 work space need to research gen Z
1:34:33 fashion there's an AI agent for that
1:34:35 want taglines for a new energy drink bam
1:34:38 another AI agent jumps in they've
1:34:40 automated the whole shebang
1:34:42 brainstorming research briefs even some
1:34:45 content production so instead of we got
1:34:48 we got nerd we got nerd alert Wolf Man
1:34:51 Clint you haven't lived until you get a
1:34:53 four Channel color ntsc waveform in the
1:34:56 6 mahz
1:35:01 range oh good thanks for the reader app
1:35:04 just downloaded yeah the the reader
1:35:05 app's really cool a whole conference
1:35:07 room of people by the way the reader app
1:35:09 so it's from 11 Labs so one of the
1:35:12 coolest things that they've done is
1:35:13 they've licensed some famous people so
1:35:16 you can have your audio book read by Sir
1:35:18 Lawrence Olivier or Judy Garland or Bert
1:35:20 Reynolds there's there's like six of
1:35:22 them now um but also if you've made a
1:35:25 clone of your own voice in 11 Labs you
1:35:29 can have yourself read the the documents
1:35:33 that it has you've got this dream team
1:35:36 of AI Specialists always on call talk
1:35:39 about a productivity boost how does that
1:35:41 play out in the real world it's all
1:35:43 about efficiency let's say you're
1:35:44 working on a new campaign this AI panel
1:35:47 of experts can instantly analyze your
1:35:50 competitors find emerging trends that
1:35:52 your target audience loves
1:35:54 even suggest different angles for your
1:35:56 campaign all in a fraction of the time
1:35:58 it would take a human team it's like
1:36:00 having this 24/7 Brain Trust feeding you
1:36:03 ideas that's wild but if AI is doing the
1:36:07 reader app Omar that we're talking or
1:36:09 ali um the uh reader app is called
1:36:12 reader and it's from 11 Labs so just go
1:36:16 to the app store or Google Play and look
1:36:18 for 11 Labs reader it's it's really
1:36:21 really good abundance flow manifest I
1:36:24 love it do you think this can work to
1:36:26 incorporate into guides to
1:36:29 sale um some people need to listen um it
1:36:32 can I mean the the problem with this
1:36:35 this tool the the uh audio overview tool
1:36:39 is right now you don't have a ton of uh
1:36:41 control over it but it's they're pretty
1:36:44 slick okay so here's the here's the
1:36:47 podcast of the kyite 4000x series
1:36:52 oscilloscope so let's see what technical
1:36:55 manuals sound like when they get the uh
1:36:58 the uh NPR treatment all right get ready
1:37:02 because today we're tackling
1:37:04 oscilloscopes oh yeah ever feel like
1:37:06 you're trying to um catch a glitch
1:37:09 moving at lightning speed it's amazing
1:37:12 how often we use oscilloscopes but don't
1:37:14 totally get what's going on behind those
1:37:16 squiggly lines right I mean seriously
1:37:18 I'm using ill if I'm not using
1:37:20 oscilloscopes 10 times a day I'm like
1:37:22 what am I doing with my life I use them
1:37:24 all the time but you know what I never
1:37:25 think about what's going on behind the
1:37:27 scenes let's see if we can get some
1:37:29 insights here you see the lines jumping
1:37:31 around but what do they really mean
1:37:33 exactly this deep dive is all about
1:37:36 understanding those settings the why
1:37:38 behind each knob and button love it and
1:37:40 this guide we're digging into it's like
1:37:42 going from those Kitty pools to the deep
1:37:45 end of the signal pool right but even
1:37:46 Olympic swimmers they started with the
1:37:48 basics so before we get fancy got to lay
1:37:51 that Foundation okay so walk me through
1:37:53 it what's happening inside that
1:37:54 oscilloscope box imagine like a signal
1:37:57 processing Factory different stations
1:37:59 all working together first stop your
1:38:01 signal hits the attenuator attenuator
1:38:03 okay what's it do like lower the volume
1:38:05 on a loud signal you got it scales it
1:38:07 down so the next stage doesn't get
1:38:09 overloaded think of it as a volume knob
1:38:11 for your signal makes sense don't want
1:38:13 to blow out the speakers right away but
1:38:15 then what about super Fain signals those
1:38:17 get boosted you're on it that's the
1:38:19 amplifier's job takes those weak signals
1:38:21 and cranks them up so it's all oh yeah
1:38:23 you you know what when when you have
1:38:24 weak signals in an oscilloscope and you
1:38:26 just can't get get clear waveforms you
1:38:29 you just you just run it through the
1:38:30 amplifier I mean everybody knows that
1:38:32 right this is isn't this common sense
1:38:34 all
1:38:36 right there you go so that's notebook LM
1:38:40 all right I'm Crispy and Wolf Man Clint
1:38:43 says he got it right okay there you
1:38:46 go there you have it
1:38:51 um I mean I think the use cases
1:38:55 for for
1:38:57 for you know not just that specific
1:39:00 podcast like again where I go with this
1:39:03 stuff is okay couple of
1:39:05 things notebook
1:39:07 LM this tool where you can upload PDF
1:39:10 files and chat with them has been around
1:39:13 for a
1:39:14 year nobody gave a [ __ ] I mean some
1:39:18 people did some people liked it but for
1:39:20 the most part no one cared no one really
1:39:22 talked about it handful of people in the
1:39:24 AI Salon talked about it
1:39:29 um two weeks
1:39:31 ago the the product team well some time
1:39:35 ago they stumbled upon an internal
1:39:37 Google project that had this advanced
1:39:40 voice [ __ ] in it and they said hey why
1:39:43 don't we prompt it to make a little
1:39:45 podcast and we'll just put that in as a
1:39:46 feature in Notebook LM and this is
1:39:49 completely taken off so I look at this
1:39:51 and I say okay this is just one prompt
1:39:54 With You Know Two Hosts and talk like
1:39:56 NPR things imagine if you had a dozen
1:39:59 different outputs where you could have a
1:40:01 long form newscast a short form newscast
1:40:04 a Punchy Sunday morning you know package
1:40:08 kind of style thing um you know a
1:40:12 training session with a with a
1:40:14 inspirational Tony Robbins type
1:40:17 trainer
1:40:20 um th this is going to be really really
1:40:23 powerful just take this concept and
1:40:25 Abstract it up up a level and say what
1:40:27 are all the kind other kinds of outputs
1:40:29 you could have when you have this
1:40:31 natural voice thing and then abstract it
1:40:33 one layer further and say okay not only
1:40:36 can I have this as voice I could have
1:40:38 this as my digital Avatar right where I
1:40:43 could have a version of me talking about
1:40:47 any
1:40:50 document like I do now except the the AI
1:40:54 version of me would actually know what
1:40:57 he's talking about as opposed to the
1:40:59 mansplaining that I do night after
1:41:05 night all right I'm using podcast from
1:41:07 notebook in articles to boost s SEO
1:41:10 aspect interesting and to stay current
1:41:14 wow that's fantastic I like it Google
1:41:16 Labs has illuminate as a podcast
1:41:19 Creation with a bit more than LM nice
1:41:23 that's super cool all right um
1:41:27 so I want to give you one other thing
1:41:30 that I want you to take a look
1:41:31 at before I get out of here if you're
1:41:35 not a part of the AI Salon the
1:41:37 salon. go there right now scroll down
1:41:41 and click on the button that says join
1:41:43 our
1:41:44 community and go check out what we're
1:41:47 about there's a welcome video there from
1:41:49 myself and Leah faston who are the
1:41:51 co-founders of this group this group was
1:41:53 started the week after chat GPT came out
1:41:56 this is a group of really cool very
1:41:59 powerful AI
1:42:01 optimists a lot of them hang out
1:42:05 here and just F there's like a
1:42:08 seven-step onboarding process it's
1:42:10 really easy it's really fun and if you
1:42:13 like what we're up to join us it doesn't
1:42:15 cost
1:42:17 anything
1:42:19 and down toward the bottom you'll see
1:42:22 Community clubs and there's a club
1:42:24 called the AI L Irregulars that's us so
1:42:31 these lives we hang out here when we're
1:42:34 doing these lives and we can make
1:42:35 pictures and stuff like that I haven't
1:42:37 been driving a bunch of activity there
1:42:39 recently so I'm I'm going to pick that
1:42:40 up next week but anyway like this is a
1:42:44 cool group go hang out
1:42:47 there go
1:42:49 join especially if you're like the only
1:42:51 one in your community
1:42:53 who's into this AI [ __ ] and everyone
1:42:55 else is like stop talking about AI you
1:42:58 dumb dumb go here we don't say that
1:43:01 we're like could you talk about AI more
1:43:03 you dumb
1:43:05 dumb all right
1:43:09 um be ready to invest time into becoming
1:43:12 an
1:43:14 irregular I would say the only cost I
1:43:17 would say that's the only cost well
1:43:19 worth
1:43:20 it yeah you got to you got to tolerate
1:43:23 me that's the problem
1:43:28 um so anyway go do that um I'm going to
1:43:31 get out of here I don't I'm trying to
1:43:34 think if there's any homework for the
1:43:38 weekend I don't know I
1:43:43 mean if you haven't played with the
1:43:45 video tools I would say see if you can
1:43:47 go find a video tool where you can make
1:43:49 videos for free Korea Deco here I think
1:43:52 does free videos Leonardo I think lets
1:43:55 you do free
1:43:57 videos um if you haven't made a podcast
1:44:01 from notebook LM if you haven't played
1:44:03 with notebook LM I would strongly
1:44:05 encourage you to do
1:44:08 that
1:44:12 um if you have not started a musical
1:44:18 yet what are you waiting
1:44:21 for uh k I I don't I don't like musicals
1:44:26 well so write one
1:44:30 anyway you can do anything you want
1:44:34 now if you're like if you're like an
1:44:36 auto mechanic why are you not writing a
1:44:39 knitting
1:44:42 guide you cuz you
1:44:46 can you know kitten husbandry could be a
1:44:49 thing you do
1:44:51 now you're like I don't even think k
1:44:53 kitten husbandry is real well you can
1:44:57 find out if it's real or not and if it's
1:44:59 real you could just go start a kitten
1:45:01 husbandry business
1:45:03 tomorrow you'll be down at the bar and
1:45:05 they'll be like hey Bob what's going on
1:45:07 down at the shop you're like I got a
1:45:10 side gig now they're like Bob that's not
1:45:12 like you you're like yeah I'm doing
1:45:13 kitten husbandry like what's that mean
1:45:16 you're like I don't know yet but I
1:45:17 started the
1:45:19 business
1:45:21 so do something ridiculous this weekend
1:45:25 with AI something that doesn't make
1:45:28 sense for you to do it but you
1:45:31 enjoy do that there are two kinds of
1:45:35 people AI curious and then well yeah
1:45:37 them others
1:45:40 exactly like I I would say if if if
1:45:45 you're using AI to only make what you
1:45:49 currently do more efficient
1:45:56 that's kind of [ __ ]
1:46:00 boring and listen I know like efficiency
1:46:03 is cool I love it you can automate [ __ ]
1:46:06 yeah you could you can make like zappier
1:46:09 zaps and make and automations and
1:46:12 automate your email writing but you know
1:46:13 what else you could do write a [ __ ]
1:46:17 musical why not go make a movie this
1:46:20 weekend I'm on the reader app but it is
1:46:23 not give an option to upload a PDF yeah
1:46:27 it should hang
1:46:28 on let me go look there Omar or Ali is
1:46:32 it Ali Omar I think it's Ali 11
1:46:36 reader so you go to library I think go
1:46:40 to
1:46:41 library yeah go to library on the bottom
1:46:45 see library right there and then on the
1:46:47 top click that plus
1:46:50 button Plus
1:46:54 write text import URL upload
1:46:57 file there you go right there those
1:47:02 things
1:47:03 done and then that'll be in your
1:47:09 library embrace the Jank here's my book
1:47:11 embrace the Jank right in there as a PDF
1:47:15 hooked and that's when the real fun
1:47:18 begins mindfully create all right you
1:47:21 want to hear me read it watch this
1:47:27 this all right play time's over now it's
1:47:30 time to get [ __ ] done but here's the
1:47:33 deal mindfully creating isn't about
1:47:35 asking AI to solve your problems it's
1:47:37 about you armed with your new AI
1:47:40 superpowers tackling real world issues
1:47:42 head on that's a bad model that's on me
1:47:46 for recording a shitty
1:47:48 model let's have Lawrence Olivier read
1:47:52 it
1:47:54 head on remember all that crazy stuff
1:47:56 you discovered while playing that's your
1:47:58 secret weapon now these tools are
1:48:00 Humanity amplifiers not human replacers
1:48:03 so when you're faced with a gnarly
1:48:07 problem I've been still trying to get
1:48:09 notebook LM to do an interview with
1:48:15 me I don't know what you mean by that
1:48:17 Wolf Man Clint I think the only way you
1:48:20 could really do that is write a script
1:48:22 for it to
1:48:24 interpret but it's not going to
1:48:27 I that would have to be some fairly
1:48:29 serious prompt
1:48:31 hacking Kyle the Sam podcast wait that
1:48:34 just went away oh the Sam podcast was
1:48:37 the same one we didn't get to Last
1:48:38 Friday date night well we didn't get to
1:48:40 it again tonight I'm trying to get chat
1:48:42 GPT to tell me how to go past high on my
1:48:46 electric
1:48:48 blanket don't do that side up so Mimi
1:48:52 you'll burn the you'll burn the house
1:48:54 down don't [ __ ] with electric
1:49:00 blankets cuz because it'll tell you how
1:49:02 to it'll tell you how to you know take
1:49:04 the stop off your
1:49:06 potentiometer you [ __ ] light that
1:49:09 that thing on fire that will not be fun
1:49:11 Kyle what did you think of Elon Musk we
1:49:14 robot
1:49:18 um the robo taxis are a really
1:49:21 interesting concept that you're going to
1:49:23 be able to just buy a robo taxi for
1:49:25 $30,000 and then it will take you
1:49:28 wherever you want to go and in the 235
1:49:31 hours you're not using it it's out
1:49:33 making money for you um so I think
1:49:36 that's pretty [ __ ] cool I would say
1:49:38 that is about as disruptive to the
1:49:41 entire Transportation sectors you can
1:49:44 get and then the 20 passenger bus thing
1:49:48 that looks like a toaster from a bad
1:49:52 sci-fi movie
1:49:56 um is [ __ ]
1:49:59 hideous but holy [ __ ] it's
1:50:04 cool and then the robots parading around
1:50:07 and sort of talking to people and
1:50:09 pouring them drinks was pretty [ __ ]
1:50:12 impressive I mean you know we're not
1:50:16 that far
1:50:17 into robot building and because we have
1:50:20 large language models now that are super
1:50:22 capable and because we've got advanced
1:50:24 voice I mean they're clearly using some
1:50:26 advanced voice [ __ ] I don't I don't know
1:50:28 if that's open ai's advanced voice or if
1:50:31 it's proprietary to Tesla but pretty
1:50:37 slick I wish Elon weren't such a [ __ ]
1:50:40 douchebag but the man is I mean he is
1:50:45 the new Steve Jobs right he's the only
1:50:47 one right now like really
1:50:49 innovating and uh and just [ __ ] just
1:50:53 pounding
1:50:55 it and he's like [ __ ] it I'm a
1:50:57 billionaire I'll make a toaster shaped
1:51:00 20 person minivan Mini
1:51:05 Bus no one else is doing that [ __ ]
1:51:08 car companies aren't doing that car
1:51:10 companies are like well we've got the
1:51:11 incremental update of the well we're
1:51:14 we're going to make an electric Ford
1:51:16 Mustang but we don't want to cannibalize
1:51:18 the actual Mustang so we're going to
1:51:20 make it look like a shitty SUV so no
1:51:23 nobody buys it
1:51:25 okay oh I love Corporate America
1:51:28 wouldn't want to cannibalize the uh the
1:51:31 cash cow or the Mustang right you know
1:51:33 what I'm
1:51:36 saying everyone sees that new Mustang
1:51:38 there like what is that AIA
1:51:53 oh Corporate
1:51:55 America all
1:51:56 right it was way better than advanced
1:51:59 voice I'm 100% sure it was remote
1:52:03 audio and realizing motion
1:52:06 capture yeah I don't know I don't know
1:52:08 what of that was real what was not you
1:52:10 know maybe the people talking were off
1:52:13 in a control room somewhere and it was
1:52:15 fake maybe it was real I don't know
1:52:18 seemed pretty good
1:52:23 a billionaire with eyes
1:52:27 rolling next thing you know they'll tell
1:52:29 us aliens are real and our vote
1:52:37 matters oh but the voice seems like it
1:52:39 was a human at the office talking back
1:52:42 yeah it very well could have been that
1:52:44 where all complex
1:52:46 AI looked like he had David Guetta off
1:52:50 get off third to last on the bus
1:52:53 that's really
1:52:54 funny um I totally want a robo taxi yeah
1:52:57 me too safety of course th those things
1:53:00 will get safe pretty soon like like Elon
1:53:03 Musk said you know those Robo taxis are
1:53:05 going to be 10 times safer than human
1:53:07 drivers that doesn't surprise me at all
1:53:10 those things are going to get really
1:53:11 really
1:53:13 safe and if anyone thinks they're not
1:53:18 like just they are technology gets
1:53:21 better I mean they're already driving
1:53:23 people around so all right I'm out of
1:53:26 here follow my channel go sign up at the
1:53:30 salon um if you want to support me
1:53:32 directly subscribe to the lives and uh
1:53:36 let's see who's in here tonight Kyle you
1:53:38 need to
1:53:41 watch a AFR Al 2024 this weekend oh
1:53:46 basically the sequel to her great movie
1:53:51 okay oh afraid I get it AFR a i d afraid
1:53:58 arra afraid
1:54:01 afraid
1:54:04 AI all right I'll look it
1:54:07 up um all right who we got in here
1:54:10 tonight we got Joey in the house we got
1:54:12 Drew in AI calling him from Down Under
1:54:15 steo also from Down Under against the
1:54:18 norm I think from Hawaii yes right aloha
1:54:22 indeed to buas in the house abundance
1:54:26 flow manifestation on Netflix I guess
1:54:29 the question for afraid digital Gods
1:54:32 Lord digital Gods good to see you sir
1:54:35 Brandon thank you so much as always
1:54:37 frumple side husto
1:54:40 Mimi Joker in the
1:54:43 house hair loss remedy ULU
1:54:47 RX very very nice very nice ULU RX by
1:54:51 the way is the
1:54:54 if you want one of these things which
1:54:57 you're like but Kyle where do you get
1:54:58 the pink bow I was gifted this pink bow
1:55:03 by one ULU
1:55:05 RX she's the one that told me Kyle
1:55:07 you're wearing it wrong and I'm just
1:55:10 like sorry uh I know what I'm doing so
1:55:15 anyway I'm right she's wrong and you
1:55:17 know what it is she just won't admit
1:55:19 she's wrong and that's fine if she wants
1:55:21 to live her life like that
1:55:23 absolutely fine by me Mike leros in the
1:55:28 house what aliens are
1:55:31 real said ULU
1:55:35 RX abundance flow all right who else we
1:55:38 got did I get everyone frumple frum's in
1:55:41 the house kif good to see you
1:55:45 Shaina DB best
1:55:51 tech Vicki's in the
1:55:54 house I think Serena was in
1:56:01 here Who's down at the bottom yeah
1:56:04 there's Amelio's wife she's in here of
1:56:06 course she's in
1:56:07 here all right Kirk T rock paper
1:56:11 scissors rock paper scissors me say my
1:56:15 name you got a comment or I can't say
1:56:18 your name I don't know you're in here
1:56:19 although I suppose I could look at the
1:56:21 list of people that that would be the
1:56:23 easier way to do it but no you know what
1:56:26 the names I read are the people that are
1:56:27 commenting thank you Lord digital gods
1:56:29 for that gift that's
1:56:32 appreciated um all right I'm getting out
1:56:35 of here have yourself a fantastic
1:56:37 weekend just go do something fun with AI
1:56:39 do something unexpected do do something
1:56:41 where you're like this doesn't make any
1:56:43 sense why I would do this but then have
1:56:45 fun with it all right thank you side
1:56:47 hustle Mimi I appreciate that have a
1:56:50 fantastic weekend everybody I will be
1:56:52 back Monday Monday night 8:00 p.m.
1:56:53 mountain time and we will start this
1:56:56 thing over again all right next Tuesday
1:56:58 is an AI Salon meet and greet so when
1:57:01 you go in there I think the third or
1:57:03 fourth little step in onboarding is RSVP
1:57:06 for upcoming events next Tuesday is a
1:57:10 meet and greet where you can come meet
1:57:12 people that are here all right so go do
1:57:15 that peace have a great weekend