aiLL 8/22/23 - The Magic AI Genie Will Make Our Lives Easier

Description

Chapters

Transcript

0:14 and
0:25 hit that's not what I want to
0:29 do
0:35 oh welcome everybody what is going on oh
0:39 that's pretty
0:41 funny I put on a little I put a little
0:45 nappy time filter that's pretty funny I
0:49 didn't mean to put on a nappy time
0:53 filter I got to turn it
0:57 off ah that was pretty good all right
1:01 we're going live what day is today
1:04 Tuesday right that's correct yes Kevin
1:08 Vasquez greetings earthlings greetings
1:11 greetings
1:15 indeed uh this shirt isn't going to work
1:17 is
1:18 it well well well we're going to have to
1:21 go change the
1:22 shirt we're going to have to go change
1:25 the shirt it happens all right uh it was
1:29 a big news day today
1:31 in AI it happens all right uh it was a
1:35 big news day today in AI
1:41 um but I'm going to have to go change my
1:43 shirt so I will I will go do that y'all
1:46 y'all just hang out here hey everybody
1:49 hey Sher bear what's
1:51 happening what's happening I have a
1:53 lovely waterl blackberry and
1:58 lemonade s
2:02 sparkling
2:03 water
2:05 huh cuz I know that's a big topic of
2:08 conversation on the AI learning lab
2:10 what's Kyle
2:13 drinking all right what's Kyle wearing a
2:15 shitty shirt for this I will be back wow
2:18 there's a lot of people here hopefully
2:19 there's some people when I get back uh
2:21 this is the AI learning lab my name's
2:23 Kyle Shannon I'm going to go put on a
2:24 black shirt so you can see the text
2:26 instead of this you know this lovely
2:29 lovely piece of apparel but not good for
2:32 tick
2:33 tocking you're thinking Kyle you might
2:35 have thought of that before you went
2:36 live I might have but you know this is
2:39 not how we do things here I got champ by
2:42 my side I'm going to be back you just
2:44 all yeah hang on hang on those of you
2:47 that are the Irregulars so that's what
2:49 we call the people that are regulars
2:51 here because it's funny um so if you
2:55 regulars would just let people know I'll
2:58 be back we're not just looking at a live
3:00 of a bunch of shelves this it's going to
3:02 get good I got news I got news I got
3:06 like I was late cuz I was like oh let me
3:08 see if anything happened on Tik Tok holy
3:10 [ __ ] like a bunch of [ __ ]
3:13 happened this is okay so listen I've
3:16 been in the entertainment business this
3:18 is what they call a cliffhanger you're
3:20 here you're like what's this dude
3:23 talking about what's the AI learning lab
3:25 why is he got [ __ ] up here why is he
3:28 leaving cliffhanger
3:59 for
4:15 all right did we lose anyone yeah we
4:17 lost like 30 people that's fine that's
4:19 fine we're not going to sweat it what
4:21 shirt did I go with Brooklyn Only the
4:24 Strong
4:25 Survive how about you go [ __ ] yourself
4:28 get back here hey get back here little
4:30 punk I don't know if that's Brooklyn or
4:32 the Bronx doesn't really matter it's
4:34 sort of a bad New York accent all right
4:38 shirt's a little small we're going to be
4:39 okay though it's going to be good it's
4:41 going to be good thank you for the heart
4:42 hearty Hearts it looks like the
4:44 Irregulars are at it with the uh the
4:46 like train again uh the the Irregulars
4:49 last night in the live it was like a two
4:51 and a half hour live they amassed
4:56 5 4.2 th000 likes or 52.4 4,000 likes in
5:01 like 2 and 1/2 hours that's not
5:04 reasonable we we had like on average
5:07 maybe 75 people something like that 100
5:09 people
5:11 maybe 50 52,000 likes I'm not that
5:15 likable just to be clear um
5:19 welcome we are the shirt we are the
5:22 shirt hey all right someone did a movie
5:25 script you lost 30 people but you gained
5:28 me awesome Steve Well Steve you're you
5:30 are you're a required element of the
5:33 evening so okay let's get started
5:36 welcome everyone my name is Kyle Shan oh
5:38 now we've lost a lot w we went down we
5:40 were up to like 130 when I went and
5:41 changed my t-shirt the Brooklyn Brooklyn
5:44 apparently booted out like 80 people
5:48 what are you going to do what are you
5:49 going to do he hey hey you know hey it's
5:52 no sweat off my back you know what I
5:53 mean hey he what are we doing you
5:58 been
6:00 will this guy ever start he might he
6:05 might okay so my name is Kyle Shannon
6:09 this is the AI learning lab what do we
6:11 talk about here we talk about
6:14 shirts we talk about the
6:16 Irregulars uh we talk about like trains
6:19 I don't know if that's a thing in Tik
6:20 Tok but it is now that's look I got a
6:23 hearty heart I got Hearts going um you
6:26 haven't heard of Like Trains on Tik Tok
6:28 no that's cuz you haven't into the AI
6:30 learning lab so stick around the here's
6:34 what happens when when you combine old
6:36 people and and new technology you get
6:38 [ __ ] like
6:40 this
6:42 um my name is Kyle Shannon I'm a a
6:45 longtime entrepreneur I've started I
6:47 don't know 12 15 companies something
6:49 like that my current gig is I'm the
6:51 founder and CEO of a company called
6:55 storyvine we're a uh an automated video
6:59 storytelling platform we do not use well
7:02 our core technology is not AI uh we do
7:05 our automation the oldfashioned way with
7:09 code and uh I'm completely rethinking my
7:13 platform and the value proposition and
7:15 the product offering and the road map
7:18 thanks to AI so that's been a really
7:19 interesting journey I can talk about
7:21 that um I was also an entrepreneur in
7:23 the early days of the worldwide web so
7:25 back when hey how you doing partner he
7:27 sherbear Sher bear give me a hat there I
7:30 don't I'm talking sort of like George W
7:32 bush
7:33 yeah
7:37 um sometimes I go into like Ross perau
7:40 for those of you that are Gen X gen xers
7:43 remember him um in the early days of the
7:46 worldwide web I started a company called
7:48 uh agency.com which was one of the first
7:50 digital agencies why do I talk about
7:51 that well it was kind of cool cuz cuz uh
7:55 it was one of the first digital agencies
7:57 we built a lot of the early websites but
8:00 the early days of the worldwide web is
8:02 very very reminiscent of the early days
8:04 of this chat GPT [ __ ] what's going on
8:06 with this stuff in my opinion is more
8:10 profound than the impact that the
8:13 worldwide web had which was pretty
8:15 [ __ ] profound it changed everything
8:18 how we did everything so this stuff's
8:20 bigger than that so couple of things we
8:23 do here I I keep these URLs up uh during
8:27 my little chats here uh for a couple
8:29 reason if you're new to generative AI if
8:31 youve not played with chat GPT that's
8:34 what these things are all about so let
8:35 me let me talk about what I got going on
8:37 here I got I got a pile of
8:47 [Music]
8:48 pages
8:51 okay why is that looking
8:54 so hold got a lighting situation here
8:59 not a good one
9:02 there we go
9:04 okay so if you're if you're brand new to
9:07 this AI
9:10 stuff go there chat. open.com that's the
9:14 official chat GPT website if you've not
9:16 played with it go play with it go play
9:18 there until you have what we love to
9:21 call here in the AI learning lab your
9:23 Kevin mallister moment this
9:26 moment remember Home
9:28 Alone go there until you have that
9:31 moment what's that moment I don't know
9:32 it's different for everyone uh you can
9:34 go there tell it to write you a country
9:36 song Tell it to write you a poem tell it
9:37 to write you a python application um ask
9:41 it to create a marketing strategy for
9:43 you just figure out something that you
9:45 know and you know how hard it is to
9:47 create and then go there and see if you
9:50 can get it to do what you think is
9:53 hard and then at some point you will
9:55 have that moment oh my God oh my God
9:59 what happened
10:01 um Claude and Bing the two below it what
10:04 are those um they're competitors to that
10:06 one well Bing is a sort of Co-op
10:10 Cooper co-option right this is from
10:13 Microsoft this is GPT micros Bing chat
10:18 is M good
10:23 Lord all right I don't know that we
10:25 should be doing this tonight the just
10:29 what just
10:32 the muscles in my jaw just malfunction
10:36 okay B
10:39 chat uses
10:42 gp4 that is behind chat GPT gp4 so this
10:46 is also gp4 but it's Microsoft's version
10:49 of it it's connected to the internet it
10:51 can make images thank you manry uh it
10:55 can read images uh it's generally
10:58 crappier than this one because it's got
11:01 a lot of safety guard rails on it cuz
11:03 Microsoft is tired of being embarrassed
11:05 on the front page of the New York Times
11:07 so it's a little shittier than than gp4
11:10 up here but it's still good and it's
11:12 free Claude is a competitor to chat GPT
11:16 it's a bunch of xop AI people that
11:19 invented gpt3 they went off and said we
11:22 don't like how you're doing safety we're
11:23 going to go do our own and they created
11:26 claw and that's got a really big context
11:28 window which basically means you can put
11:29 in giant prompts okay so so if you're
11:32 new to this [ __ ] go
11:34 there um prompt. chat will teach you how
11:37 to use these things okay that's a little
11:39 just instruction manual on how to talk
11:41 to these things there there's lots of
11:43 ways you can prompt these things but
11:44 that's just a good one to start okay so
11:46 that's
11:47 that if you have used some of these
11:49 things before here's some others to play
11:51 with Bard is Google's entry in this
11:53 space uh perplexity doai is a search
11:56 engine that does citations it it sort of
11:59 mixes search and AI in an interesting
12:03 way and then labs. perplexity uh lets
12:06 you play with meta llama 2 models for
12:08 free why because they're awesome I guess
12:11 I don't
12:13 know everyone's got their
12:16 reasons
12:17 pi. the Irregulars here on the learning
12:20 lab have nicknamed chatty GPT and I
12:23 think that's
12:24 appropriate chatty GPT uh if you if you
12:29 put p in conversation mode it's really
12:31 interesting to talk to I was I was with
12:33 a uh a buddy of mine today um guy named
12:36 townzen Ward law who's a a sales
12:39 consultant and he he actually helped
12:41 design the sales um our sales strategy
12:45 for Story vine and I haven't talked to
12:47 him in years and he was in Denver and he
12:49 said let's get together so so we hung
12:51 out and he was talking about his wife is
12:55 has been working with AI on loneliness
12:59 having having people that are lonely
13:02 talk to GPT models and I said have you
13:06 seen piie or have you tried Pi or have
13:08 you talked about pi and he said no and
13:11 so I opened up the PI app and I asked it
13:15 you know who is Townsen War word law and
13:18 what is his theory on on you know
13:22 sales and it nailed it his jaw he he had
13:26 a Kevin mallister moment he had this
13:28 moment cuz he knows a decent amount
13:30 about AI but it it like nailed who he
13:34 was what his philosophy was and it was
13:36 like the latest up to the minute who he
13:38 was uh so anyway so Pi's really
13:40 interesting po uh you can make your own
13:43 Bots there that's really interesting
13:45 this the like train is [ __ ] hilarious
13:48 it's totally distracting me but I love
13:50 it okay so so that's all that [ __ ] okay
13:55 now pictures pretty pictures you're like
13:58 but Kyle I don't like words I want
14:01 pictures unless you're demo Demo's a
14:04 troll that's I think demo secretly loves
14:07 the AI learning lab he comes in every
14:09 night and trolls me and says I don't
14:10 know what I'm talking about and I'm an
14:12 [ __ ] and that um well he hasn't
14:15 really said I'm an [ __ ] I've
14:16 interpreted it that way um but he also
14:19 says that images are a waste of
14:21 everyone's time so if you're not demo
14:24 and you think images are not a waste of
14:26 time then uh these things might be
14:28 interesting in to
14:30 you I don't know why I'm calling out and
14:33 making my trolls more famous but he's
14:36 he's he's he's at least consistent in
14:38 his trolly MC trollington uh mid journey
14:41 is the best Firefly is the most
14:43 corporately safe for lack of a better
14:46 term in fact it it's still early so I
14:48 don't I don't think you can actually use
14:50 that images from
14:52 Firefly in corporate applications yet
14:55 but you will be able to um Leonardo and
14:58 clip drop are both stable diffusion
15:00 they're just fancy front ends in front
15:01 of stable diffusion um and then open
15:04 art. a/ promptbook will teach you how to
15:07 use these things okay and then future
15:10 pedia that's like 4,000 it's probably
15:13 more now they add hundreds each day
15:16 4,000 AI tools that if you ask me Kyle
15:20 what's the best way to do accounting for
15:24 circuses in Lithuania I go there there's
15:28 probably an app for that I just don't
15:29 know them
15:30 all okay that's that final
15:34 thing um and welcome I I see some new
15:37 people have joined probably because
15:39 there's just an incessant amount of
15:40 likes the the the Irregulars are at it I
15:43 think it's hilarious and I love them
15:46 um you might be thinking AI learning lab
15:49 I think I've heard of that well that's
15:51 because I'm a celebrity and this is a
15:52 famous channel it's so famous in
15:56 fact that we have a [ __ ]
16:00 scammer and this just pisses me off
16:02 because Tik Tok won't deal with it so if
16:05 you know anything about Tik Tock scammer
16:09 account reporting or uh if you work at
16:12 Tik Tok or if Tik Tok you're listening I
16:14 know you're listening because you shut
16:16 down my live occasionally because I say
16:17 some trigger
16:19 word so so I've learned a few of the
16:23 ones you don't like um but if if you are
16:28 following an account that is AI learning
16:30 lab that has two L's in it for learning
16:33 AI L learning lab it has a stutter the
16:36 stuttering account is not me but you're
16:39 like but Kyle it's got all your videos
16:41 and it looks like you and it's taken
16:43 everything about your account
16:45 exactly and then what they do is they
16:47 reach out to people that follow them and
16:49 they try to sell them some multi-level
16:50 marketing scam [ __ ] so [ __ ] them
16:53 [ __ ] the horse they rode in on this is
16:55 the only recourse I have is
16:57 to
16:59 take advantage of my like train and say
17:02 screw those guys but but if you follow
17:04 them you just check you have to sort of
17:06 dig in and like go past the initial page
17:09 cuz they've copied like like the initial
17:11 page that they've done looks exactly
17:13 like mine so you have to go into their
17:15 uh profile and see the two L's so
17:18 unfollow them report them and this is
17:21 the real
17:22 deal okay good Lord that's a lot it's a
17:25 lot it's a lot to open up the the show I
17:27 had to change my shirt we did that we
17:30 went through all of this the like train
17:32 is going crazy here's the other thing
17:34 that happens here hang on hold hold
17:47 please beautiful all right so that's
17:50 just going to sit there so that's the
17:51 chat GPT website go there if you haven't
17:53 gone there that's why we have that there
17:54 it's just like a little Public Service
17:57 Announcement um
18:00 um W thank you very much steo I like
18:03 those little dvo glasses remember dvo
18:06 from the 8S you old people you jet
18:09 xers you old cranky
18:13 bastards and basts
18:19 bastards um so if you have if you have
18:22 questions about AI comments about AI pop
18:25 them in the comments below it's a Tik
18:27 Tock hat thank you very much for that
18:30 um I will uh I'll do my best to answer
18:33 them now here a couple of things I try
18:35 to answer all the questions some people
18:37 really appreciate that new people
18:39 [ __ ] hate it they're like which
18:41 questions is he answering I'm answering
18:43 the ones way up above wherever the [ __ ]
18:45 you are that's just the way it is I've
18:49 tried to like do a reset with timers it
18:52 just I have ADD I can't I can only do
18:54 what I can do um so I will eventually
18:57 get to your question question
19:00 um I make funny voices um I I sometimes
19:05 cuss if you don't like cussing get the
19:07 [ __ ]
19:09 out um what else uh I don't know that's
19:13 it I'm I'm going to show some stuff
19:15 there there was some some decent news so
19:17 if you hang out I'll go through some of
19:18 the news but what I'm going to do right
19:20 now is I'm going to I'm going to scroll
19:22 up into the comments I'm going to see
19:24 what people have been saying as I have
19:26 been
19:27 rambling rambling rambling
19:30 rambling all right welcome welcome
19:35 everybody all right Tobias is here Kevin
19:38 Vasquez is
19:40 here uh uh uh uh uh sherbear is here we
19:44 got some regulars okay farmer this is
19:46 cool Lor C LOL Lura C is Ling
19:51 tonight and major gift fundraising
19:55 what's everyone up to with AI good
19:57 question yeah pop pop that in the
19:59 comments below we'll talk about whatever
20:02 you're up to as well sup dude steo is
20:04 here Tobias is here hey everybody stand
20:07 by star star star star star star star
20:09 star star star
20:12 star all right let's
20:14 see wait the current sag strike is about
20:17 salary and AI royalties yes it is it's
20:21 about yes well not just AI
20:27 royalties
20:29 it it's more like
20:34 um so here's what
20:38 the oh big corporations are so
20:41 predictable aren't they so sag was about
20:45 to go on strike and they're they're in
20:46 their negotiations and then the
20:48 producers Guild or whatever the [ __ ] is
20:50 the movie Producers
20:53 thing said uh we'll get back to you on
20:56 whatever your requests were
21:00 and then they ghosted them for like two
21:02 weeks and then they said okay we have a
21:06 response to all of your
21:07 requests and for like almost all of
21:10 their requests the response was
21:12 denied [ __ ] you [ __ ] you [ __ ] you [ __ ]
21:16 you they went down the list [ __ ] you
21:18 [ __ ] you [ __ ] you but but then toward
21:21 the bottom there was one but we have a
21:24 groundbreaking
21:26 proposal and our groundbreaking proposal
21:29 is this from The Producers to The Screen
21:32 Actors
21:33 Guild for background actors so extras
21:36 right the ones that don't have all the
21:37 lines in the
21:39 movies how about this how
21:43 about we as
21:45 producers pay those extras
21:49 $200 and on that day that we pay them
21:51 $200 we will scan their likeness and we
21:55 will digitize it and then we will put it
21:58 in our our magic AI computer
22:01 box and we
22:03 will use their likeness imp perpetuity
22:07 Forever Without ever compensating them
22:09 ever
22:11 again
22:13 um and that's what that's when sag went
22:16 fantastic [ __ ] you and went on strike uh
22:20 so so it it's a big deal now here here's
22:24 the [ __ ] up
22:27 part
22:29 if I take emotion out of it and the fact
22:31 that I'm an you know a trained actor and
22:33 I've written screen plays and you know I
22:36 have empathy for for that
22:38 crowd for my peeps
22:45 um there's nothing stopping the
22:47 producers from paying nothing like what
22:50 what the producers are saying is we'll
22:51 at least pay them a day rate of 200
22:54 bucks um they could just say hey for for
22:57 extras in any feature film we're not
23:00 going to ever use human beings ever
23:02 again we're just going to synthesize
23:03 them all we'll we'll we'll hire a bunch
23:05 of people to go generate them in mid
23:07 journey and and the extras will always
23:09 be computer generated they won't be
23:11 based on any likenesses and that's the
23:12 way it's going to be um so
23:16 so so the leverage that sag has is is
23:20 not great because because the tools are
23:23 really powerful so so it's it's kind of
23:24 a fascinating thing um you know I I
23:28 think the producers will ultimately do
23:31 something um sag and The Writers Guild
23:34 have leverage a little bit right now but
23:38 a year or two from now won't like this
23:40 probably will be the last strike where
23:42 the Screen Actors Guild and The Writers
23:44 Guild or any any kind of unions like
23:45 that have any kind of Leverage
23:47 whatsoever so it's a fascinating
23:50 fascinating time uh and and also I think
23:53 that both sag and the wga are are
23:56 actually being very shortsighted about
23:57 how they're approaching ing AI stuff all
24:01 right um did a movie script we are the
24:05 [ __ ] sherbear the the Irregulars are
24:08 [ __ ] awesome you guys are nuts we're
24:10 already up to 177,000 likes I'm not I'm
24:12 like like what am I 15 minutes in you're
24:15 amazing uh did I miss the delay
24:18 answering time hey toos
24:22 um no I I don't think so I don't know
24:26 what the delay answering time is but
24:27 you're not late I'm just getting
24:31 started okay all important people are
24:34 here uh ah K oh look at fantastic I got
24:39 the little I got the magic spell I feel
24:41 blessed I feel Zapped that was awesome I
24:45 don't I don't know what it is about me I
24:47 am I am uh what do they call it I'm
24:50 peppy tonight like last night I was a
24:52 little
24:55 like tonight I'm
24:57 like
24:59 I don't know what you prefer or so we'll
25:03 see all
25:05 right I would have a question all right
25:07 simaga have it can we sell the artwork
25:11 created through
25:27 leonardo.da
25:28 I'm going to go find something here
25:31 we'll
25:37 do fact so I'm on
25:57 leonardo.da
26:01 that we've never ever ever read you know
26:05 this document that
26:08 um Facebook and Google just slightly
26:12 updated every 3 weeks for the first 20
26:15 years of our existence and we just kept
26:17 going oh yeah sure sure sure this is the
26:20 document where we gave away all of our
26:23 um data for the past couple of decades
26:26 so here's a fun thing you can do I'm I'm
26:28 going to grab the URL from the terms of
26:32 service I'm going to go to chat GPT we
26:35 could also do this on I'll show you two
26:38 ways to do this we'll go to chat GPT
26:40 we're we're going to do a plugin on chat
26:42 GPT and then we're also going to do
26:45 Claude because Claude is one of one of
26:48 the uh models I show
26:53 here so for chat GPT I'm going to go to
26:56 plugins and and I don't have any enabled
26:59 so I'm going to go enable one if they
27:05 load I had a plug-in enabled so
27:09 something's
27:13 weird plugin store did they get rid of
27:16 all of my plugins wait
27:22 what hang
27:24 on hang
27:27 on
27:28 no plugins
27:31 enabled oh okay this is okay here's a
27:34 good lesson I'm logged
27:39 in let me reload this they obviously
27:42 reset their plugins
27:44 database oh no okay I just reloaded okay
27:47 they're back Jesus okay
27:51 so so uh that that that well-earned
27:55 Panic is we're so early in this AI stuff
27:59 that every piece of AI software right
28:01 now is a janky piece of duct taped spit
28:05 toothpicks and um dog food is what
28:09 they're made of so
28:12 sometimes sometimes stuff just uh breaks
28:16 because it's bad software and then
28:18 sometimes it's just like reload your
28:19 browser uh hang on one second 7 3 6 8 my
28:24 turn
28:26 wait uh uh
28:32 uh okay
28:34 y all right that's cool
28:38 okay so I've got a plugin enabled called
28:43 uh web pilot now what web pilot can do
28:45 is it can go read a website so I'm going
28:48 to say
28:50 please
28:52 read and then I'm going to pop in that
28:55 URL and I'm going to hit shift return
28:57 for a line break and I'm going to say
29:05 um summarize the five most
29:09 important parts of this
29:12 agreement with regards
29:15 to my ability to use any images I
29:21 create for commercial
29:26 purposes
29:28 and so it's going to go read that legal
29:37 document now um who was it I think it
29:41 was it retr
29:42 Punk retro I think it was retr Punk sent
29:45 me a a DM so one of the Irregulars
29:48 ownership of content you own all of the
29:51 input for the output generated based on
29:54 your input Leonardo assigns to you all
29:57 its rights title and interest in and to
30:01 that
30:02 output it doesn't quite say you own it
30:05 right they say you own the
30:07 input and I guess because they're
30:10 generating there was an error in
30:11 response oh it just crapped out
30:14 okay for the output generated based on
30:17 your input okay oh okay so that's
30:21 interesting so what they're saying is
30:24 you own your
30:26 prompt and if you generated something
30:29 from your
30:31 prompt Leonardo assigns that to you but
30:34 I think what that also implies is if
30:39 you use someone else's prompt they don't
30:42 so just make sure you modify your prompt
30:45 right change a word in it so that you
30:46 can you can claim it's yours now is
30:48 anyone going to check that the the
30:50 people that check whether or not you
30:52 modified or prompt or not are the same
30:54 people that check the signatures on our
30:56 credit card receipts that that we signed
30:58 at the convenience store I'm pretty sure
31:00 no one is checking those signatures
31:03 because everyone has the same signature
31:06 and it does not resemble their name in
31:08 any way um so anyway so that was that
31:10 was kind of an interesting thing so
31:12 that's that now we're going to go over
31:14 to Claude and do the same
31:17 thing now at Claude so I'm going to pop
31:20 in the The Prompt so summarize this
31:23 thing and then I'm going to do a line
31:25 break I'm going to go back to the terms
31:26 of service and I am going to just
31:29 highlight them
31:31 all so I just highlighted all of the
31:34 terms of service I'm going to come back
31:35 to Claude this is this is one of the
31:37 reasons to use Claude Claude is free and
31:41 Claude has 100,000 token context window
31:44 so I can paste in like a 75,000 Word
31:47 document it could be a legal document it
31:49 could be a screenplay it could be a
31:51 novel and I type in paste and then I
31:55 submit it and so now it's going to go
31:57 read all of the text of that
32:01 document retr Punk said he uploaded a
32:05 PDF to
32:06 Claude and he wanted to translate a
32:09 permission slip from English to Spanish
32:12 and it transl it it basically like made
32:15 up an entirely new document my suspicion
32:18 of what happened with retr Punk was and
32:21 if it wasn't you retrop punk if it was
32:22 someone else sorry about that it it
32:24 might have been Tobias I don't know it
32:25 was it was one of the Irregulars
32:28 um what might have happened is it might
32:30 not have been able to read the PDF so it
32:33 might not have actually extracted the
32:34 text out of the
32:35 PDF um depending on how the PDF was
32:39 constructed right so if it was just a
32:42 scan instead of actual text it might not
32:44 have seen it and then and then it
32:46 probably just made some [ __ ] up so
32:48 that's that's my suspicion okay based on
32:51 my review of the terms of service images
32:54 created by paid users retain full
32:56 ownership copyright and intellectual
32:58 property rights you have the commercial
33:00 rights to images you create as a paid
33:03 user okay that's interesting for free
33:07 tier
33:08 users Leonardo owns all the rights to
33:11 the images created but grants
33:13 free users a royaltyfree license for the
33:18 commercial use of images they create
33:20 okay
33:21 so I'm glad we did this so it summarized
33:26 what the document said in two slightly
33:28 different ways so if you're trying to
33:31 figure out if you own the rights from
33:32 Leonardo I would actually go read the
33:34 terms of service yourself but this is a
33:36 good way to to get the short answer the
33:38 short answer is yes if you pay for it
33:40 you own the image and you can do
33:42 whatever the [ __ ] you want with it if
33:44 you're on the free tier you've
33:47 got free royaltyfree license to use the
33:51 image now let's see any images made by
33:55 public by any user can be used
33:59 commercially any images made public by
34:03 any user tier can be used by Leonardo so
34:09 so you can use it but Leonardo can also
34:11 use it you represent having rights to
34:14 any inputs so if you if your input is
34:17 copywritten material that's not good
34:19 there are restrictions on offensive
34:21 illegal and harmful content so yeah so
34:24 basically you own it or can use it but
34:26 they can so use it there you
34:30 go but there's a that was a that was a
34:33 fun little fun little rabbit
34:39 hole I haven't gotten to the news yet I
34:42 haven't even gotten to the news I I'll
34:46 tell you what you people that are like
34:47 oh I'll watch this old man for like five
34:49 minutes you're going to miss some [ __ ]
34:51 the Irregulars [ __ ] hang out in here
34:53 see all those hearts going that's not
34:55 cuz they like me that's because they're
34:57 bored as [ __ ] and they have created a
35:00 whole series of games to entertain
35:02 themselves while I ramble that's what
35:04 that is it's not
35:06 likes but you have to hang out here cuz
35:08 [ __ ] happens [ __ ] goes down steo see I
35:11 got the dog ass I got the dog thank you
35:13 very much I loves me some dog well I
35:15 let's not say it like
35:17 that dogs are lovely where's my doggy
35:22 champy Champ's not around he's Champ's
35:25 done with me tonight he's like no no
35:27 no we're good I'm
35:31 good okay so it's a mad mad world by the
35:36 way 9 9.5 out of 10 on the name ideally
35:41 what you would do is you would put in
35:43 the five mads like the original movie
35:44 had that would get you a 10 out of 10
35:47 but it's a call back to a fantastic well
35:50 a horrible film that I thought was
35:52 fantastic when I was a
35:53 kid what about storyboarding um there
35:57 are lots of people doing um oh that is
36:01 so cool thank you Sher bear the lovely
36:03 heart the lovely heart um at the at the
36:08 AI Salon which by the way I don't think
36:10 I talked about that if you want to join
36:12 a community of people just geeking out
36:13 on AI the AI
36:15 Salon uh that's the URL for it the
36:18 salon. A and uh that's a group I created
36:22 back in uh December and that's the link
36:26 that's the UR to the link tree that's
36:29 got the Meetup where you can you know
36:32 subscribe to the Meetup we meet every
36:33 other week so that'll give you
36:34 information about that it's also an
36:36 invite to the Discord server uh where
36:39 you can geek out and hang out with us um
36:42 one of the early guys that spoke at the
36:45 salon his
36:49 name what's his first
36:52 name free tag I don't know it doesn't
36:54 matter well it does because I can't
36:56 remember his name
36:57 um but he he um he was doing
37:01 storyboarding with mid Journey um four
37:04 four movies right so so producers and
37:06 directors were calling him up saying hey
37:08 we want we want concept boards for this
37:10 film and he was using he was knocking
37:13 out full-on story boards um in like 3
37:17 days that that would have taken him
37:20 weeks he was doing it in 3 days and they
37:23 were they were so cool so yeah you can
37:27 do
37:28 storyboarding and you know you can do
37:30 it with with a lot of different tools I
37:32 mean mid journey is the one that I think
37:34 is is the uh is is is the one of the day
37:39 it's not going to be for long all of
37:41 these tools as exciting and sexy as they
37:44 are today uh crypto Joe welcome as as
37:47 exciting as they are today like every 3
37:49 months this list gets kind of rejiggered
37:52 I've I've had to redo this list now a
37:54 bunch it's it's more complicated than I
37:56 want it to be but there's there's a lot
37:57 going on Nancy's here what's
38:00 happening cam kaken hello
38:04 hello um I would love to ask you some
38:07 questions then please do simaga here's
38:10 the deal simaga when you asked I would
38:12 love to ask you some questions that was
38:14 probably 20 minutes
38:17 ago so now I'm telling you I there's a
38:20 large possibility you're not here
38:21 anymore there's a large possibility you
38:23 asked your questions and then I didn't
38:25 answer them and then you asked them 30
38:27 eight more times and then people in here
38:29 told you dude chill out he's going to
38:31 answer them he's going to get to them
38:32 this is the deal this is the thing where
38:34 the regulars understand this Dynamic of
38:36 this channel the new people don't
38:39 appreciate it so much but I will answer
38:41 all your questions dragonfly C Alchemy
38:44 is here hey dragonfly what's
38:46 happening
38:48 uh I'm using it every day awesome that's
38:51 simaga what's your
38:53 question uh you don't need to you don't
38:55 need my permission to ask questions just
38:57 just ask questions although here's what
38:58 I would say don't ask questions like um
39:02 could you explain it some
39:04 more I don't I never know what it is
39:07 because I'm not reading the comments in
39:08 real time so if I'm talking about uh I
39:11 don't know puppies say could you explain
39:14 puppies some more and then that way when
39:16 I get to your question I'll go back and
39:18 talk about the things some more you need
39:20 coffee I don't think I need [ __ ] Co
39:22 well I guess with ADD maybe it'll calm
39:23 me down tonight maybe I do need coffee
39:26 maybe I I think you're right okay farmer
39:29 I think you're right I think I'm I think
39:31 I'm [ __ ] jacked up on not
39:35 coffee oh add you're such a special
39:40 condition yeah because stimulant Focus
39:43 me and calm me down and and now that I
39:45 think about
39:48 it I don't think I had any coffee today
39:52 I'm [ __ ]
39:55 just I it it it actually does feel like
39:58 I just did a couple of
40:01 bumps are you taking guests um I I can
40:05 I'm happy to bring a guest up I I tend
40:07 to just answer the questions uh down
40:10 here but uh just let me know what you
40:13 want to do you can ask the question and
40:15 text I'll get to it open AI my best
40:16 friend Claude and Bing are also I know I
40:19 know that that that thing I did today
40:21 with with my friend towns in wlaw where
40:24 I I asked piie to
40:28 describe who he was and it did it so
40:30 perfectly he was just like he was like I
40:33 got to install this that was pretty
40:36 amazing you need some apple cider
40:37 vinegar verified human I well oh by the
40:40 way if if you didn't know this about
40:43 me is if it wasn't obvious
40:47 enough I'm a verified
40:50 human oh you don't believe me hang on
40:54 you know the trolls with their you don't
40:55 believe you what you're you're not a
40:59 verified human how would you even know
41:01 that how would I know that how would I
41:04 know
41:06 that the the the doubt the
41:09 doubt how I know that is I have this
41:14 verified I'm an anonymous verified human
41:17 what's it
41:18 say um proof of
41:22 personhood
41:24 person verified at orb what is this
41:28 you're
41:29 saying that's not important right
41:32 now my biometric data has been given to
41:35 the board hey
41:38 Clint um that that is actually a
41:42 um my nft from that's my world ID from
41:47 Sam alman's New World coin and World ID
41:51 company to verify humanhood
41:54 personhood um which is going to be
41:57 really important whether or not Sam's
41:59 company does it I think things like that
42:02 thing I just had behind me are going to
42:04 actually be really
42:05 important all right um why do we need to
42:08 have to plead with folks to use these
42:12 tools oh just the like the permissions
42:15 cuz everything's in persistent beta mode
42:17 and you have to get on [ __ ] lists I
42:19 don't know it well there's there's two
42:20 reasons one is it's kind of a hangover
42:22 from the nft world where the crypto Bros
42:25 are like you know if we if we make if we
42:28 make it seem like it's it's harder to
42:31 get in and you got to jump through
42:33 [ __ ] hoops and it and actually a a
42:36 lot of the nft Bros are like Okay click
42:39 the heart button you know if you get to
42:41 50,000 Hearts you know you'll get into
42:44 this next layer of the Discord and and
42:46 then so people just do that sort of Enga
42:49 engagement mining is what they call it
42:51 to get the communities to do [ __ ] our
42:54 like train just organically emerged out
42:56 of the Irregulars cuz they're so [ __ ]
42:58 weird but but um but the the Crypt Bros
43:03 like gamified all that [ __ ] and so I
43:05 think a lot of the you got to sign up
43:07 for the weit list is kind of generating
43:10 um demand but some of it is also a lot
43:14 of these startups are probably like you
43:16 know two college kids in a case of Red
43:18 Bull right and
43:20 so they're only paying for so much
43:23 bandwidth and compute power so they want
43:26 to they want to test the thing before
43:27 they go live or maybe before they get
43:29 funding so that's probably what's going
43:30 on if that's what you're asking about
43:32 felonius
43:34 uh he goes in order of comments thank
43:37 you are you happy now are you happy now
43:40 is I am trying to get into the box oh
43:42 Thelonius is trying to get into that box
43:45 oh that's that's a rare thing felonius I
43:48 don't really bring bring much people up
43:49 on stage does he take guests I don't
43:51 have that much time well what do you
43:53 want to talk about simaga Saga simaga
43:58 Saga simaga Saga
44:02 simaga Nancy I've never seen except
44:04 guests I brought um hair growth geek up
44:08 once I brought I brought I've brought a
44:11 couple of people up
44:13 um I I won't bring you up like just tell
44:15 me what you want to talk about like what
44:17 I don't want to do is bring you up and
44:19 just have you hawk some piece of
44:21 technology or God forbid some
44:24 crypto um
44:27 so simaga no guess ask your question in
44:29 comments it might be 10 minutes Sher
44:32 Bear's Sher Bear's got it worked out
44:34 she's like no ask your questions and use
44:38 your words use your words
44:44 simaga ah Saga Nancy thanks now I can
44:47 move on he does not provide much info
44:49 absolutely correct I am completely
44:52 unqualified uh so you should absolutely
44:54 not be here if you're looking for for
44:56 information I tried Pi it's very cool
44:58 yeah Pi so the the thing about Pi is um
45:02 it's super chatty it's chatty GPT so if
45:05 you want to get [ __ ] done Pi's not your
45:08 one but if you want to kind of have a te
45:11 and talk to your computer in a
45:14 significantly meaningful way then Pi's
45:17 really interesting really
45:20 interesting you
45:22 Einstein uh I don't know what that means
45:24 Steve Pennington am I Ein Stein am I a
45:28 genius we'll have to ask chat GPT about
45:31 that it's possible I still haven't
45:32 gotten to the news you're all like where
45:34 am I going to get to the news I don't
45:36 know maybe if you hit enough likes I'll
45:39 get there what we're only at 33,000 good
45:42 Lord you
45:44 people 33,000 likes we're less than an
45:47 hour
45:48 in guys are wild I like it it's it's
45:52 solid I'm I'm a little I'm a little
45:54 flustered oh my I do D de CLA is that a
45:57 little woman he isn't me
46:00 oh y'all are
46:04 funny all right I love that you're
46:07 entertaining yourselves uh some people
46:09 have told me that they they use me to go
46:11 to sleep so so this isn't just about the
46:14 information this is sometimes about the
46:16 entertainment value and sometimes about
46:19 the narcotic effect that my voice has to
46:22 to knock people out on a regular basis
46:25 all right I a bad experience with Claude
46:27 it was youth lonus
46:29 um uh yeah I gave it a PDF to translate
46:33 instead of doing that CLA got creative
46:35 and just made stuff up I have a feeling
46:38 it it couldn't read the PDF I would
46:40 check that PDF see if the PDF actually
46:42 has text in it or if someone took a
46:44 picture of a permission slip and the PDF
46:47 is just an image because if it's that
46:50 then you would probably need to use
46:51 something like
46:53 Bard or Bing that can read images so
46:57 this is one of the other things these
46:58 things are all kind of screwed up
47:04 um chat GPT can't really read documents
47:09 except with its plugins it sort of
47:12 can and with code interpreter it sort of
47:15 it can but in in kind of really specific
47:17 ways Claude can read PDF text documents
47:21 csvs things like that Bing can read
47:25 images Bard can read images um what are
47:29 the other ones
47:31 um perplexity I don't think can read you
47:34 can't upload anything Pi I don't think
47:36 you can upload anything po you can
47:38 upload things but I think it's just
47:39 based on whatever models you have
47:41 selected uh I think po can read like
47:43 csvs and text files and [ __ ] like that
47:46 so
47:47 so they're they all sort of act like
47:51 they're going to be really cool but
47:53 they're all sort of janky this goes back
47:55 to that comment about all of these tools
47:56 are kind of you know spit in duct tape
47:58 right now um you kind of need and it's
48:01 the reason I have so many of these
48:02 things up on screen all the time is you
48:05 kind of need
48:06 to if you really want to make use of
48:09 these things you kind of have to know
48:10 what they all do which is it's not going
48:11 to be like that for long it's give it a
48:13 year and this is going to be a
48:15 remarkable time right now it's a little
48:17 janky what is scammer oh oh a scammer
48:21 okay so toas
48:24 um the this is scammer um scammer stand
48:30 so a scam in English is you know someone
48:33 trying to um I'm going to use more
48:37 vernacular rip someone off some someone
48:39 trying to basically steal from people so
48:42 the scam is they
48:45 copied my channel and my content so
48:49 they're
48:50 getting you know folks that are
48:52 following me to follow them and then
48:54 once people follow them they reach out
48:56 to them they basically DM
48:58 them acting like me and say hey would
49:02 you join this you you know you should
49:04 really join this awesome marketing
49:06 opportunity and it's a m multi-level
49:09 marketing scheme or a pyramid
49:12 scheme and
49:14 so they're trying to make money off of
49:17 our audience off of the Irregulars it
49:20 was the Irregulars that that found this
49:22 out for me like I didn't figure it out
49:24 they figured it out so thank you all um
49:27 and and so that's the scam is that is
49:31 you know ripping people off and scammer
49:34 is the [ __ ] that does
49:39 it all right the Irregulars need
49:41 t-shirts and hat merchandise I know I I
49:44 I really do have to get I'm going to
49:45 figure out how to do I looked
49:49 up two days
49:51 ago um it looks like printful and one
49:55 other one
49:57 are
49:58 um have a Tik Tock integration so I can
50:02 so I can I can use Tik Tock shop to to
50:05 do print on demand shirts so I'm going
50:08 to do a you can make money with chat GPT
50:10 um I'll probably do a couple of others I
50:12 think I'm going to absolutely do the
50:14 Irregulars um I'm an irregular that I
50:18 think that's it I'm an irregular um and
50:21 then I don't know maybe we'll have like
50:23 some little asterisk bu it and it you
50:25 know the the little thing will say you
50:27 know in a regular is a member of the AI
50:29 learning lab or something like that
50:31 we'll figure it out but yeah that's a
50:32 good idea Laura I'll get some I'll get
50:34 some merch going so and then I actually
50:37 oh you know what be cool is I'll do an
50:40 Irregulars
50:42 um I'll do an a regular thing but maybe
50:44 I'll put like a box like a white outline
50:47 with a black middle it'll the t-shirts
50:49 will all be Black by the way so we can
50:51 wear them on the channel um and I'll
50:54 make like a box which is where the the
50:56 text goes I don't know well all right
50:58 I'll think about that we'll do
51:01 something okay I know now oh okay cool
51:03 uh WN
51:05 radio yes scammer messaged me today I
51:08 blocked and reported thank you marid D
51:10 and I apologize for that I it I guess
51:12 it's not for me to apologize but it just
51:14 it it it gets my goat it gets my goat
51:18 you know what I mean it gets my goat
51:21 that
51:23 that you know these people that spend
51:26 more
51:27 time imitating me in the amount of time
51:31 that they can figure out how to imitate
51:33 me and maintain that
51:35 ruse ruse is a
51:38 word they could have started their own
51:40 [ __ ]
51:44 channel right
51:52 like did I miss the shirt changing yes
51:54 you did Tommy you missed the shirt
51:57 changing it was a fantastic it was so
52:00 fantastic so we got in here and and they
52:02 were like like we went from like 20
52:05 people to 130 people I'm like awesome
52:07 people I'm going to go change my shirt
52:09 and I lost 80 people it was fantastic
52:13 we're slowly working it back up I think
52:14 because of the like train that the
52:16 Irregulars are are are smash that like
52:19 button the Irregulars are entertaining
52:22 themselves at this point this channel is
52:24 no longer about me which is that's
52:26 probably how it should be this is
52:29 awesome um of course now I play song
52:31 quiz with Alexa oh that's funny that's
52:34 awesome um Pi is really good at at games
52:37 too so if you get Pi into conversation
52:40 mode and tell it you want to learn
52:42 something and tell it to make it make a
52:44 game up and and teach you the thing in
52:47 whatever game it comes up with comes up
52:49 with trivia games it comes up with all
52:51 sorts of [ __ ] Tik Tok just jumped me to
52:53 the bottom of the comments so I'm trying
52:55 to figure out where I
53:00 was wait a cheap way to know if you're
53:04 in real if you're in a real live is to
53:06 see your comment in a mirror screen or
53:10 mirror I don't know what that
53:17 means uh I emailed a complaint today
53:20 Google the email address thank you Nancy
53:22 yeah anyone who can report those scammy
53:26 MC scaming
53:27 Ton's um maybe Tik Tok will listen it it
53:31 it's really frustrating I can't it just
53:34 it's staggering to me like like with all
53:37 the [ __ ]
53:38 algorithmic you know let's prevent hate
53:41 speech stuff that is you know very very
53:45 flawed it seems like it's a relatively
53:48 straightforward thing to see if somebody
53:50 copied someone else's
53:52 account it's the same
53:54 videos on their
54:01 account 848 in Denver am I late Amelio's
54:04 wife what's happening uh I don't think
54:07 maybe I don't know it's 9:20
54:11 now
54:13 um and I'm still not back to where I was
54:16 well I like it looks like I lost my how
54:19 can I get a longer movie script
54:22 um use um so there's there there's a
54:27 couple of things about writing long form
54:30 documents chat
54:33 GPT um only has a 4,000 token context
54:37 window right now it it should at some
54:41 point have a 32,000 context uh token
54:45 context window but it's not there yet
54:47 because open AI is busy they're cutting
54:50 deals and buying companies whatever
54:52 they're doing they're spending open AI
54:55 spending $700,000 a day is what open
54:58 open AI is
54:59 doing
55:03 um so so there's a couple of things you
55:05 need to do if you're in chat PT you need
55:07 to chunk it
55:10 up trying to think I'm trying to think
55:13 of some other ways you could do
55:14 it in chat GPT you could use code
55:18 interpreter code interpreter can write
55:21 much longer documents I don't know if it
55:23 would be any good though you could yeah
55:25 I wouldn't use code interpreter no so so
55:28 what I would do what I would do if you
55:30 want to write a
55:36 screenplay you're going to have to chunk
55:38 it up the first thing that you do is you
55:39 write an
55:40 outline and it'll be able to do the
55:42 outline and then maybe you say based on
55:44 this outline you
55:48 know let's create um a set of character
55:52 description you're going to need an
55:53 outline a set of character description
55:55 desps a set of place
55:58 descriptions so you're going to have to
55:59 kind of have those elements and then and
56:01 then you're going to write the scenes
56:03 from
56:04 there and you might even need example
56:07 scenes right if you if you write
56:08 dialogue in a particular way or or if if
56:12 you've if you haven't quite described
56:14 your characters well enough you might
56:15 need example dialogue so
56:18 outline sorry having written some
56:20 screenplays I'm I'm trying to think how
56:22 I would do it this this is how I would
56:24 do it this is not how you do it or how
56:26 people do it this is how I would do
56:28 it I I would I would get it to write a
56:32 concise outline that I agree with a
56:35 character sheet a place sheet I would
56:38 probably have some sample dialogue like
56:40 you you want to se it with good
56:44 dialogue those three things right there
56:47 are probably going to take up most of
56:48 your context window so 4,000 tokens is
56:52 about 3,000 words so so those elements
56:56 right
56:58 there one way you could do it is you
57:00 could put in those elements and then
57:02 tell chat GPT to summarize all of those
57:06 into the least amount of words possible
57:08 so you kind of make it less
57:11 tokens and then I would just do sort of
57:14 scene at a time or even subscene like
57:16 here's here's the first half of scene
57:18 one here's the second half of scene one
57:21 and then just copy and paste those
57:23 out or
57:27 Claude has a 100,000 token context
57:31 window so it's got a 75,000
57:34 word um memory basically window that you
57:37 can play Within So within Claude you
57:39 could take those three those long
57:41 elements that I described outline
57:43 characters places sample dialogue and
57:47 then you could just say write me act one
57:50 of this screenplay and it'll probably
57:51 write the whole
57:54 thing I think that's probably a bad
57:57 approach
58:00 because you know you remember the the
58:02 Looney Tunes commercial I knew I should
58:04 have taken a left at
58:05 alui
58:07 um let's say that that Claude or chat
58:11 GPT writes scene one and right at the
58:14 end of scene one it kind of goes in a
58:16 direction you didn't want it to go and
58:18 then it'll write you it it'll sort of
58:20 take the action from that kind of went
58:22 off a little bit and it'll write you all
58:24 the other scenes and like you know it it
58:26 it'll I don't know it'll just be bad
58:30 so there's a there's a project out there
58:33 it's if you're
58:35 geeky check out a project called novel
58:40 GPT novel GPT will write you a full
58:43 novel and what it does is you give
58:47 it what you want the basic story to be
58:50 about what you want the writing style to
58:52 be about and the number of chapters and
58:54 then the the way it does what it does is
58:56 it writes an
58:57 outline and then it writes chapter one
59:01 and then it takes the outline and
59:02 chapter one and it writes chapter 2 and
59:05 then it takes the outline in chapter one
59:07 and chapter two and writes chapter 3 so
59:10 so it's building upon what it wrote so I
59:12 would take a similar approach with
59:14 writing a screenplay I'd get all of your
59:16 sort of planning documents done you can
59:19 have chat gbt or Claude help you write
59:21 those and then I would just chunk it out
59:24 um don't expect to write the whole
59:26 screenplay because it'll it'll just [ __ ]
59:27 it up it'll just be forever
59:30 disappointing
59:32 um what's interesting about that novel
59:34 GPT project is that it's it's in a
59:38 collab a Google collab notebook you can
59:40 see all the code that they wrote all the
59:42 prompts that they wrote to to write all
59:44 of that if you if you if you're code
59:47 friendly and can go look at the scripts
59:49 within the within the notebook and you
59:52 could modify you could modify novel GP P
59:55 to become screenplay
59:57 GPT might be
1:00:01 interesting the other thing you should
1:00:04 do is not only have it right the
1:00:07 screenplay but have it WR the coverage
1:00:09 right so so when screenplays get read
1:00:12 you know at agencies they they do these
1:00:14 coverage documents like a two-page
1:00:16 explanation of it uh and log lines and
1:00:19 all that sort of stuff so anyway all
1:00:24 right
1:00:26 still I need an Oki impersonation I
1:00:28 don't know who Oki is but I'm sure I've
1:00:31 done it at some point I just don't know
1:00:32 who it is John Wayne will do oh Oki when
1:00:34 you put the Hat on me yeah howd it
1:00:37 partner uh doing Microsoft security
1:00:41 analyst certificate on corsera yeah
1:00:44 that's interesting I think I think
1:00:46 Microsoft uh is going to be a big player
1:00:50 with with all of their uh integration of
1:00:54 chat GPT with within everything they're
1:00:56 doing so I think that's how a partner
1:00:59 thank you Dragon
1:01:02 Fly chip chip chip chip chip
1:01:07 chip anyone who hasn't been here before
1:01:09 is just got to be like this dude is
1:01:11 [ __ ]
1:01:12 insane yeah I haven't had my coffee what
1:01:15 do you
1:01:18 want an Adder without coffee is like a
1:01:22 normal person on cocaine apparently
1:01:26 It Is What It Feels Like Tonight um I
1:01:28 just tried to report the fake AI
1:01:29 learning lab for pretending to be
1:01:31 someone else only choices were celebrity
1:01:34 or me I that exactly so Professor Sonia
1:01:37 you're going to have to call we're going
1:01:39 to have to just accept the fact that I'm
1:01:41 a celebrity okay I look kind of like
1:01:44 celebrities people have mentioned that
1:01:46 doesn't that count and I have fans that
1:01:51 like it enough to how much do they like
1:01:53 it 46.6 th000
1:01:57 likes so that's celebrity likes I just
1:02:00 say I'm a
1:02:01 celebrity I I'll back you up you tell
1:02:04 you tell Tik Tok I'm a celebrity I'll
1:02:06 back you up best AI to manipulate a
1:02:11 photograph I would say generative fill
1:02:15 within Photoshop so get the Photoshop
1:02:17 beta and then you can just you know you
1:02:19 can use all of photoshop's selection
1:02:21 Tools in a fairly sophisticated way and
1:02:25 and then just manipulate the [ __ ] out of
1:02:26 it or expand
1:02:29 it and then you know just you can
1:02:32 protect the stuff you want to protect
1:02:33 and obliterate the stuff you want uh
1:02:36 were me and celebrity I typed in Kyle
1:02:38 Shannon's name and a football player's
1:02:40 came name came up actually no a uh
1:02:43 that's Kyle Shanahan and it's he's a
1:02:45 coach of the San Francisco 49ers and
1:02:49 what's funny about Kyle Shanahan and
1:02:51 Kyle Shannon cuz my Tik Tok is uh Kyle
1:02:55 Shannon so whenever the 49ers [ __ ] up
1:02:58 and have a bad game I get hate mail from
1:03:01 football people going yeah maybe you
1:03:03 should learn to coach
1:03:04 Shannon and I'm like maybe you should
1:03:07 figure out the last name of the person
1:03:08 that actually coaches
1:03:14 loser so yeah that's pretty funny and
1:03:16 football season's about to start so I'm
1:03:19 about to get some hate mail unless the
1:03:22 49ers do really good this year um so I
1:03:26 had to report them for trying to sell
1:03:27 multi-level marketing
1:03:31 yep it's so frustrating I'm sorry you're
1:03:34 all having to deal with that [ __ ] but I
1:03:36 don't I don't know any other way to deal
1:03:38 with it I guess Tik Tok only lets you
1:03:39 report people pretending to be
1:03:40 celebrities
1:03:42 exactly um I emailed a complaint today
1:03:45 Google the email address okay that's a
1:03:47 good idea oh maybe I should put that on
1:03:49 there hang on wait Nancy can you do you
1:03:52 have the email address can you put it in
1:03:53 the comments
1:03:55 because maybe I can hang on let me
1:03:58 see
1:04:04 um is that working no Tik
1:04:11 tock tick
1:04:13 tock
1:04:16 uh customer support
1:04:23 email
1:04:29 support
1:04:32 no well anyway pop it down there if you
1:04:34 found it I don't I don't want to take
1:04:36 the time here to just Google
1:04:39 something this this Channel's
1:04:41 insufferable enough as it is um I just
1:04:44 feel like it's getting more and more
1:04:46 complicated all the
1:04:48 time like I haven't even gotten other
1:04:50 news yet um but that's okay we'll get
1:04:52 there how to convert p torch
1:04:54 functionality into tensor JS um I'm not
1:04:59 geeky enough not the right channel for
1:05:01 this anything like that I would just go
1:05:03 to Reddit and and um find communities
1:05:06 that are working on tensor JS I'm so so
1:05:09 think about the AI learning lab as post
1:05:12 chat GPT like using AI rather than
1:05:16 developing AI um one of the big pieces
1:05:18 of news today however is that um open AI
1:05:24 announced fine-tuning for for GPT 3.5
1:05:29 turbo so the API so you can now train um
1:05:34 GPT on your own data um so that's a
1:05:38 really big deal um so that's in that
1:05:40 neighborhood but but that specific
1:05:42 problem pytorch functionality into
1:05:44 tensor JS um I would do a couple of
1:05:47 things Bard is connected to the internet
1:05:50 Mech fly so I'd go to b.google.r
1:05:55 you can go find the documentation for
1:05:59 for both of those things and just pop it
1:06:02 in there and then ask the question um
1:06:05 but Bard being connected to the internet
1:06:07 you might be able to just ask that
1:06:08 question at b.google.r
1:06:24 application layer stuff I'm not down at
1:06:26 the at the more um machine
1:06:29 learning uh foundational stuff that's
1:06:32 that's not my expertise that is a large
1:06:34 area of ignorance for me like like for
1:06:38 years I was like oh I'm going to go play
1:06:40 with tensorflow every time Google made
1:06:42 an announcement about tensor flow I'm
1:06:43 like I'm going to go do Ai and machine
1:06:46 learning and then I'm like I'm not a
1:06:47 mathematician this is
1:06:50 hard so so so I'm a different different
1:06:54 kind of
1:06:56 expert I'm a I'm a special expert well
1:06:59 in he special and he special coming on
1:07:02 here talking about being expert in
1:07:05 artificial intelligence and then someone
1:07:06 ask a very basic question about tensor
1:07:09 JS and he's like I don't know well well
1:07:13 well who's he inspired by who could it
1:07:16 be oh I know could it be
1:07:20 Satan yes it is according to Tik Tok
1:07:24 Kyle Shannon is an NFL football player
1:07:27 Kyle
1:07:28 Shanahan where can I learn about prompt
1:07:31 training cat GPT right
1:07:35 there prompt.
1:07:40 chat start there it's not it's not the
1:07:44 definitive document but it's got It's
1:07:46 got some good primer documents in it
1:07:48 that we talk to you about how you talk
1:07:49 to these things here's the biggest
1:07:52 philosophical thing Jason
1:07:56 that th this took me a while to get
1:08:00 Google for years has trained us how to
1:08:02 talk to search engines right and what
1:08:05 they've trained us is don't put too much
1:08:07 [ __ ] in the prompt keep it really
1:08:10 concise just ask for exactly what you
1:08:12 need and like we've gotten clever at how
1:08:14 you write a Google Search right you if
1:08:17 you're good at Google searching you you
1:08:19 kind of get the it's somewhere between
1:08:22 natural language and Boolean logic
1:08:26 right these things are the complete
1:08:29 opposite they need a lot of
1:08:31 context so don't think of these things
1:08:34 like I'm going to put in the perfect
1:08:35 prompt and get out the perfect answer
1:08:37 think of it much more like I'm having a
1:08:40 conversation with the world's most
1:08:43 frustrating intern who happens to be a
1:08:47 savant and knows
1:08:50 everything so you know when you have an
1:08:52 intern and you're like oh could you go
1:08:54 make me a a permission slip and then
1:08:58 they come back and it's like you're like
1:08:59 oh oh God I got to explain what a
1:09:01 permission slip is oh oh Christ okay
1:09:04 yeah uh so Jimmy uh what a permission
1:09:09 slip is is could you go make one now
1:09:12 right and you spend more time explaining
1:09:14 to the
1:09:15 intern than it would have taken you to
1:09:17 do it but you know you're like but you
1:09:19 got to give the intern a chance and
1:09:20 someone's got to learn somewhere and
1:09:22 it's like everything's going to be fine
1:09:24 it's not frustrating at
1:09:26 all that experience that's what these
1:09:29 things are like except with these things
1:09:32 you can say hey Jimmy go make me a
1:09:34 permission slip uh by the way here's an
1:09:37 example of a permission slip um but I
1:09:41 want you to make the all the permission
1:09:42 slips in I Amic pentameter and then I
1:09:44 want you to translate the them into
1:09:47 Romanian
1:09:49 Japanese and and I want one permission
1:09:52 slip written in Python code in
1:09:54 executable python code and it will just
1:09:56 go do all that [ __ ] um so so you have to
1:10:00 give you have to give these things a lot
1:10:02 of framing right and and
1:10:06 also rather than sort of create a single
1:10:10 Mega
1:10:11 prompt you can do what's called prompt
1:10:14 stacking where I was I was sort of
1:10:15 talking about this before where if I'm
1:10:17 going to write a
1:10:19 screenplay the first step might be hey
1:10:22 I'm thinking about writing a screen
1:10:24 screenplay um can you share with me the
1:10:28 top five story Frameworks that
1:10:31 screenwriters like to use oh sure here's
1:10:33 Sid field and here's save the cat and
1:10:35 here's all these other things hero's
1:10:38 journey Joseph Campbell right and then
1:10:40 you can say oh I like the Joseph
1:10:42 Campbell one let's let's can you help me
1:10:44 write a an outline you know using the
1:10:46 Joseph Campbell thing sure here's how we
1:10:48 would write an outline blah blah blah
1:10:50 and just kind of Stack prompts one after
1:10:52 the other that context that the the
1:10:56 conversation um will make these tools
1:10:59 get better and better the further you go
1:11:03 the one caveat is with chat gbt because
1:11:05 it's got a small context
1:11:08 window it will sometimes forget the
1:11:10 early part of the conversation so you're
1:11:12 going along and things are going great
1:11:14 and then all of a sudden it's just like
1:11:15 chat GPT just like flakes out and seems
1:11:18 to have forgotten whatever the hell you
1:11:19 were talking about that's because the
1:11:21 top of the conversation went out of the
1:11:23 context window so you may need to go
1:11:25 back to the early part of the
1:11:26 conversation grab some of that early
1:11:27 stuff and put it down later um and then
1:11:30 just keep it going or use Claude which
1:11:34 has like a 75,000 word context window
1:11:36 you can have much longer
1:11:40 conversation we'll take you as you are
1:11:42 Professor Shannon thank you Emilio's
1:11:44 wife that is very lovely of You by the
1:11:47 way uh 10 out of 10 Emilio's wife well
1:11:49 no I'm going to give you 9.5 out of 10
1:11:51 cuz it's hard to say but but I like it I
1:11:55 like you and I like the flowers cuz I
1:11:57 can always see your cuz your flowers I
1:11:59 can always see it's you it's very
1:12:04 nice I think Amelio's wife is actually
1:12:07 the one that came up with the uh the
1:12:10 called us Irregulars right you're the
1:12:11 one that actually named named it called
1:12:14 it us Irregulars must also have ADD
1:12:17 constant screen tapping yeah exactly
1:12:20 exactly rather than having to pay
1:12:22 attention here you're like I got a game
1:12:24 there's a game there's a game I like the
1:12:25 game games are good games are good oh
1:12:27 Kyle said something funny more
1:12:29 game holy crap we're at 56,000 likes I
1:12:32 think we're about an hour
1:12:34 in um you're cutting off oh no
1:12:38 Irregulars we're
1:12:39 riding we're riding rigging till
1:12:44 midnight uh of course we all read the
1:12:46 terms of service yeah exactly please do
1:12:48 John Wayne I'm an Oki oh I see okay
1:12:51 please do John Wayne I gotta get oh you
1:12:53 know what um we'll go do PI for this
1:12:59 one pi.
1:13:12 a let's
1:13:20 see give me the top 10 John Wayne oh
1:13:24 Siri I found this on the web no if you
1:13:27 approach it with the right mindset you
1:13:29 can still focus on being authentic
1:13:31 genuine stop shut
1:13:34 up they're so exhausting okay we're
1:13:37 going to go to Bard hang
1:13:40 on all these agents they're really
1:13:48 exhausting
1:13:52 hello uh uh top
1:13:55 10 John
1:14:00 Wayne Uh movie
1:14:13 quotes courage is being scared to death
1:14:16 but saddling up anyway I didn't know
1:14:18 that was a quote that's a really good
1:14:19 quote a man deserves a second chance but
1:14:21 keep an eye on him if everything isn't
1:14:23 black black and white I say why the hell
1:14:26 not that's
1:14:29 close I won't be wronged I won't be
1:14:32 insulted I won't be laid a hand on I
1:14:34 don't take kindly to them that do
1:14:38 that a man's got to have a code a Creed
1:14:41 to live by that's pretty good a big
1:14:44 mouth oh here we go all right next time
1:14:46 I get the Hat a big mouth don't make a
1:14:48 big man that's good you can't stop
1:14:52 what's coming but you can slow it down
1:14:54 that's good that's for AI you can't stop
1:14:56 what's coming but you can slow it
1:14:59 down you can't stop what's uh you can't
1:15:03 stop what's yeah see now the Hat's not
1:15:05 there they got it okay when you when you
1:15:08 when you report the scammer to Tik talk
1:15:10 also tell them to keep the hats on
1:15:11 longer I can't I can't do a John Wayne
1:15:14 quote with a hat on for half a
1:15:17 second like oh this is good life is hard
1:15:21 it's harder if you're stupid a man's got
1:15:23 to know his
1:15:24 limitations all right I got to I got to
1:15:26 put these somewhere oh all right let's
1:15:29 see you can't stop what's coming but you
1:15:32 can slow it
1:15:33 down all right I'll I'll I'll get it
1:15:36 I'll I'll work on it I got to have my
1:15:38 script I'll get my script
1:15:40 going all
1:15:42 right let's
1:15:44 see we could do some news you want to do
1:15:46 some
1:15:48 news anyone else here in Minnesota New
1:15:51 York City Professor Sonia San Francisco
1:15:55 for
1:15:56 salonius a cheap way to know if you're
1:15:59 in a real life is to see your comment in
1:16:02 a mirror screen or
1:16:04 mirror we saw that
1:16:06 earlier I would just take the
1:16:09 permission and a picture to Google lens
1:16:12 do the translation
1:16:16 later is Claude trained on current data
1:16:19 what constraints does it have Claude
1:16:21 Claude is more recent than chat GPT
1:16:23 claws up to like somewhere in
1:16:26 2023 um or late 22 I think it's in I
1:16:30 think it's like early 23 but all all
1:16:32 these large language models have a cut
1:16:33 off at some point cuz they're they're a
1:16:36 beast to train for the foundation model
1:16:38 or the Frontier Model they're calling
1:16:40 them now um so it's more recent but Pi
1:16:45 Bard and Bing are all connected to the
1:16:49 internet so those three
1:16:52 um make up for the the that Gap in data
1:16:57 by doing searches internet
1:17:01 searches Claude AI is UK us only that's
1:17:04 a drag Laura I I heard that before I
1:17:07 yeah I don't even think it's you can't
1:17:09 even get it in in
1:17:12 Canada I would just take the permission
1:17:14 okay that we said help me write a catchy
1:17:17 title for all my
1:17:19 employee
1:17:21 article on generative AI oh yeah I catch
1:17:25 you yeah that that would be a good use
1:17:28 case have you done that what was the
1:17:30 result
1:17:32 Ciro to goas
1:17:34 2201
1:17:36 2301 thanks for the Claude AI unravel
1:17:40 Kyle you're
1:17:41 welcome it's likes my dog is
1:17:45 lovely what breed is Champ champ is oh
1:17:50 let me see if I have a picture of champ
1:17:51 that I can bring up up here I
1:17:55 should I don't know if I can find one
1:17:57 hang
1:17:59 on
1:18:08 plus there he
1:18:09 is there's
1:18:14 champy look at that big dumb dumb see
1:18:17 he's got the
1:18:18 see he was in the big house he shanked
1:18:21 someone in the big house
1:18:24 that's that's that's a tear tattoo he
1:18:27 Shanks someone he's a badass um he's
1:18:31 part
1:18:33 Beagle part French
1:18:35 bull The fral Frugal I don't
1:18:39 know he's he's beagle and not French
1:18:45 bull maybe French bull yeah something
1:18:48 like that he's got some beagle he's got
1:18:50 some
1:18:51 bull and he's like
1:18:54 he's 60 lb he should be about 40 lb but
1:18:59 we don't have any boundaries with our
1:19:00 pets this this this is something this is
1:19:03 something I learned many many years ago
1:19:05 we we had a very badly behaved dog and
1:19:07 so we hired a dog trainer right cuz you
1:19:10 do cuz that's cuz you're going to be
1:19:12 good people you're going to hire a dog
1:19:13 trainer here's what you learn in dog
1:19:15 training school they don't train the dog
1:19:19 they train the people if you have a
1:19:21 shitty dog it's cuz you're a shitty
1:19:23 dog owner so once I realize that I'm
1:19:27 like oh I'm just always going to have
1:19:29 horribly behaved dogs all of my life
1:19:31 because I have no boundaries they're
1:19:33 getting on the bed they're getting on
1:19:34 the couch they're eating people food and
1:19:36 my wife's worse like she cooks for him
1:19:40 so yeah I should just this should be the
1:19:43 background the whole time for the for
1:19:45 the learning lab I'll make myself really
1:19:47 small just sit sit back
1:19:51 here oh good lord good Lord good Lord
1:19:55 good Lord I sound I'm a getting a little
1:19:58 horse I sound like
1:20:01 Clinton he's a lovely
1:20:04 dog um what are the chances chat GPT
1:20:07 Returns the same exact response actually
1:20:09 very very low inspiring dad so the way
1:20:12 these things work
1:20:14 so if you go if you go to the GPT
1:20:18 playground there's a thing called
1:20:21 temperature that it used to slide
1:20:23 between 0 and one but now it slides
1:20:25 between 0 and two I don't know why they
1:20:27 made that change because things that use
1:20:29 it haven't made that change it's it's a
1:20:31 bit of a nightmare but anyway doesn't
1:20:33 matter this thing called temperature if
1:20:35 you slide it to zero it gives you sort
1:20:38 of a predictable response chat gpts
1:20:42 defaulted to like 6 out of a scale of
1:20:45 one so if you go to all the way to one
1:20:49 um it's like fully creative like it'll
1:20:52 it'll give give you answers that might
1:20:54 not even resemble what you asked for so
1:20:56 six gives you mostly what you asked for
1:20:59 but adds some creative in there if you
1:21:00 dial that to zero you can get
1:21:02 predictable responses so if you're
1:21:04 developing applications that slider is
1:21:06 really important because sometimes you
1:21:08 want it to give you you know sort of
1:21:10 just the facts and you want it to be
1:21:11 predictable and that's that's what
1:21:12 that's that's about if you go to
1:21:22 Bing
1:21:23 go to Bing and then go to their
1:21:26 chat
1:21:30 button wait what's going
1:21:34 on no start
1:21:38 over
1:21:40 Okay what is okay there we go okay so on
1:21:46 Bing if you go to chat this is chat GPT
1:21:50 it's welcome to the new Bing right your
1:21:52 AI high powered co-pilot for the web
1:21:55 Bing has more precise more balanced and
1:21:58 more creative so they're sort of
1:22:00 revealing that idea of temperature so
1:22:02 more creative is probably like a 0.9 out
1:22:05 of one more balanced is probably 6 or
1:22:08 0.5 more precise is probably 0.2 um so
1:22:13 so if you want something more
1:22:14 predictable you can go to Bing put it in
1:22:16 precise mode or you can go to the open
1:22:19 AI playground which is platform .op
1:22:24 a.com
1:22:28 playground and this is where you can
1:22:34 um change the temperature right
1:22:38 here so you go into chat mode you can do
1:22:42 GPT 4 you can do GPT 3.5
1:22:45 turbo and then you can change your
1:22:48 temperature from zero to two why they
1:22:51 changed the scale of that
1:22:53 I don't know cuz again all these things
1:22:56 are duct tape pieces are poopy right now
1:22:59 they'll get better they'll get better
1:23:02 it's going to get better people it's
1:23:03 going to get
1:23:04 better you're like but Kyle isn't this
1:23:07 incredibly more complicated than it
1:23:09 needs to be shouldn't it just be simple
1:23:12 Kyle Kyle Kyle listen listen Kyle Kyle I
1:23:15 have a question why is this so
1:23:18 hard I want it to be magic I want it to
1:23:22 be be a magic trick and it's
1:23:29 hard you
1:23:32 remember this might be an add thing now
1:23:34 that I come to think about
1:23:37 it trying to think what age I was I was
1:23:40 probably like eight or
1:23:42 10 and I really got into magic
1:23:45 conceptually like I saw it on TV or I
1:23:49 saw a magician on
1:23:51 TV and I'm like I'm going to get a magic
1:23:54 trick
1:23:56 kit and so I asked for a magic trick kit
1:23:59 and I got the magic trick kit at
1:24:01 Christmas and I'm like I'm going to be a
1:24:05 magician and then what you learn when
1:24:07 you get a magic trick
1:24:09 kit is that ain't no magic it's all you
1:24:14 have to [ __ ] practice it's like the
1:24:16 next 10 years of your life are are
1:24:19 manipulating cards and I still
1:24:21 manipulate cards but I'm shitty at it do
1:24:23 I have any cards here
1:24:27 um magic takes a lot of work and these
1:24:30 tools right now when you know how to use
1:24:34 them they're [ __ ] magic but like
1:24:37 getting your head around how to use them
1:24:38 it's a lot of work right now it's not
1:24:40 going to be like two years from now
1:24:42 three years from now this shit's going
1:24:43 to be
1:24:44 [ __ ] like I like I don't know like I
1:24:48 I can't imagine that we're going to
1:24:50 normalize this you know I feel like kids
1:24:52 kids that are being born today bornn
1:24:54 into this world where these tools exist
1:24:57 it's just going to be normal for them
1:24:58 that you just you just say I want to
1:25:02 play a video game that has John Wayne in
1:25:05 it that's you know got racing game
1:25:08 mechanics
1:25:09 and I don't know gambling and it'll just
1:25:12 spin up a world for you and you'll start
1:25:14 playing the game that's going to be
1:25:16 normal for people you know that are born
1:25:18 today but anyway it's going to be quite
1:25:21 remarkable
1:25:23 I don't even know if that answered your
1:25:26 question me without coffee is really not
1:25:29 good I should probably go make a pot but
1:25:30 what time is it it's too late what time
1:25:33 is it 9:52 severe weather what the
1:25:36 [ __ ] going on we got another storm
1:25:38 coming I
1:25:39 guess uh it's 11:03 I'm predicting I'll
1:25:43 see this at
1:25:44 11:47 52 you were pretty close whoa oh
1:25:48 that's
1:25:50 awesome share bear with the Black
1:25:53 Panther claw that was super cool thank
1:25:57 you um you were actually very close
1:25:59 Apple user 336 uh 3388
1:26:05 6368 um I'm going to give you a five out
1:26:07 of 10 just because you're an Apple user
1:26:10 so you get you get bonus points for
1:26:12 being an Apple user but you get you get
1:26:15 really dinged for the just having the
1:26:17 the number there 338 that's actually the
1:26:20 first
1:26:21 three digits of my social which I
1:26:24 probably shouldn't say on a live but
1:26:26 whatever I got my Iris scammed what's it
1:26:29 matter you encouraged me yet me ignore
1:26:32 me no no oh do John Wayne
1:26:36 now um I will I have my quotes wait hang
1:26:40 on pause for a second give me a
1:26:43 hat where's
1:26:46 my which one is
1:26:48 it all right give me a give me a hat
1:26:54 that's a that's a music note give me
1:26:56 give me the cowboy hat somebody give me
1:26:58 the cowboy hat H yeah a big mouth don't
1:27:01 make a big man it it needs to stay there
1:27:04 longer it I I just look like an idiot
1:27:06 when it goes away when I lose the
1:27:08 mustache a big mouth don't make a big
1:27:11 man A man's got to have a code a Creed
1:27:14 to live
1:27:16 by well well well partner I got to go
1:27:20 back and watch some John Wayne movies my
1:27:22 my John Wayne I'm just doing like
1:27:24 generic Southern dude I'm not doing John
1:27:26 Wayne Right Now Steve I'll I'll get i'll
1:27:29 get John Wayne dialed in it's the the
1:27:31 cowboy hat deserves better than what I'm
1:27:33 giving it right now I'll give you that
1:27:35 it's it's a solid criticism it's a solid
1:27:37 criticism by the way constructive
1:27:39 criticism is not trolling creative
1:27:42 trolling is is uh enjoyed on this
1:27:45 channel um constructive criticism
1:27:47 enjoyed on this channel uh unimaginative
1:27:50 trolling like what demo does
1:27:53 it's just a waste of time waste
1:27:55 everyone's time chamomile tea instead of
1:27:57 coffee not for
1:27:58 me can't stand chomal te I think I'm
1:28:01 Jacked Up on not coffee I I think that's
1:28:04 what's going on P is my best friend one
1:28:07 love very cool um he's gone all right I
1:28:10 don't know who's gone but verified human
1:28:12 in the house yes that's
1:28:15 me although it says verif what does it
1:28:18 say uh ver verif uh proof of personhood
1:28:23 proof of personhood is what that is what
1:28:25 my little card says World coin yeah it's
1:28:27 it's a it's a worldcoin world ID um I
1:28:30 knew you were one of us
1:28:33 cryptos listen I I I was very deep in
1:28:37 the crypto world I'm a member of
1:28:38 neotokyo which was a which which is a
1:28:41 really cool crypto Gaming
1:28:44 Community um I've got an S1 citizen I've
1:28:47 got an S2
1:28:48 citizen I got deep in the blockchain
1:28:51 world um
1:28:53 I started learning about AI because of
1:28:55 crypto I got in a project called pixel
1:28:57 mind which is an nft project sitting on
1:28:59 top of stable defusion it's like you
1:29:01 create these art series and then you can
1:29:03 mint them it's it's a really cool
1:29:04 project um and I met the I met the uh
1:29:07 the founder of
1:29:09 that and I just got really disillusioned
1:29:13 by all the crypto Bros and all the
1:29:14 [ __ ] like the the neotokyo community
1:29:17 is actually quite strong um but so much
1:29:19 of that stuff is just it it it deserves
1:29:22 the [ __ ] blood bath that it took
1:29:25 because we just we got to get the
1:29:26 garbage out of that like the technolog
1:29:28 is really interesting I think you know
1:29:31 nfts as access are really interesting
1:29:34 the the world ID thing um you know using
1:29:37 blockchain for verifying authenticity of
1:29:40 people of content I think is going to be
1:29:43 actually
1:29:44 critical after the 20124 election I
1:29:47 think it's going to be a [ __ ] [ __ ]
1:29:48 show and people are going to be like we
1:29:50 need to do something about this fake
1:29:53 content I I wish someone would have
1:29:55 anticipated
1:29:57 this someone a celebrity isn't there
1:30:00 some celebrity talking about this well
1:30:02 Kyle Shannon over at the AI learning
1:30:04 lab's been talking about it for a year
1:30:05 he's not a celebrity he kind of looks
1:30:08 like fat Alec
1:30:09 Baldwin um Say It Isn't So Kyle it might
1:30:13 be so what's up Kyle what's up Clint I
1:30:16 know you said that an hour ago 99 new
1:30:19 messages should I jump to the bottom I
1:30:21 think I might have to jump to the bottom
1:30:23 verified Human by Machine definition of
1:30:25 ironic that actually is pretty ironic
1:30:27 isn't it digital Gods
1:30:31 yeah yeah and you know you would think
1:30:33 there would be other other ways to do it
1:30:37 but but you know so back to the Crypt
1:30:38 Bros
1:30:40 thing one of the things that the Crypt
1:30:42 Bros [ __ ] hate is kyc right and so
1:30:47 for those of you that don't know kyc it
1:30:50 it stands for know your customer so it's
1:30:51 a it's a law that basically says Banks
1:30:55 you know o over a certain amount of
1:30:58 money in
1:31:00 play or in certain
1:31:04 countries banks have to have your
1:31:07 identity like your address and like
1:31:10 your your driver's license or your
1:31:14 passport right in order to be able to
1:31:16 use certain Services you have to kyc you
1:31:19 have to you have to give your identity
1:31:21 so that's why something like World ID is
1:31:24 interesting
1:31:26 where yeah they're they're taking your
1:31:28 biometric data but they're they're
1:31:30 creating a uniqueness database and it's
1:31:33 it's Anonymous right they're not tying
1:31:35 your biometric data to you they're just
1:31:39 saying that the person that
1:31:41 possesses that ID has that Iris and they
1:31:46 don't give a [ __ ] who you are so that's
1:31:48 actually pretty brilliant I don't know
1:31:50 of another way to do it and and have it
1:31:52 be
1:31:53 anonymous and then to to to the
1:31:58 point we have to trust in the case of
1:32:01 world ID we have to trust that world ID
1:32:04 is actually disposing of the biometric
1:32:06 data and not tying it
1:32:08 to who I
1:32:10 am but you know who
1:32:13 knows at this point I don't give too
1:32:16 much of a [ __ ] about it it's it it's
1:32:18 it's it's all irrelevant at this point
1:32:20 it's all just right at this point it's
1:32:21 all just kind a monopoly money and just
1:32:23 just play stuff cuz there's not really
1:32:24 an infrastructure there what do you
1:32:26 think of all the tools that are coming
1:32:27 out at the same time oh my God crypto
1:32:29 Joe it's it's it's it's mind meltingly
1:32:34 it's it's impossible to keep up with
1:32:35 let's let's go do the news because
1:32:38 there's some new [ __ ] that came
1:32:40 out and we don't normally do the news
1:32:43 but I want I I was looking at stuff so
1:32:47 this is one
1:32:49 thing so there's a there's a uh there's
1:32:53 a technology out there called Wonder is
1:32:56 it Wonder Labs wonder wonder something
1:32:59 where you take a video and you upload it
1:33:02 to this thing and like you take a video
1:33:04 of a human walking and then like 30
1:33:07 minutes later it will give you back a
1:33:09 video of like a robot walking right it
1:33:11 does really remarkable rotoscoping this
1:33:16 technology is doing it in real
1:33:19 time so this is a an app
1:33:24 called simulan s i m u l
1:33:29 n it's a new IOS app it's in invite only
1:33:32 mode right so you have to know someone
1:33:35 there but it's it's doing human
1:33:38 replacement in real time they're walking
1:33:41 around with this device around a human
1:33:44 and it's replacing the human with that
1:33:45 robot in Real
1:33:47 Time in real time
1:33:50 people do you not get it do you not get
1:33:53 it do you not get it look at what's
1:33:56 coming it's I it's just
1:33:59 I and you know to your point Krypto Joe
1:34:03 like like how do you keep up with this
1:34:06 [ __ ] you can't like I didn't know about
1:34:09 this the the 10 minutes before I started
1:34:12 the live and then all right I'm like
1:34:14 well the the Irregulars need to know
1:34:16 about that so that's
1:34:17 one
1:34:20 um
1:34:24 Salesforce just invested in hugging
1:34:28 face at evaluation wait is this just
1:34:32 yeah August 22nd Salesforce leads the
1:34:35 financing of AI startup at more than $4
1:34:38 billion valuation so hugging face is
1:34:42 it's kind of like GitHub but where you
1:34:44 put AI models like so like all these
1:34:46 different models and the open source
1:34:48 models and all this all this AI stuff if
1:34:50 you want to go geek out and play with
1:34:52 with this AI stuff you basically go to
1:34:54 hugging face somewhere between GitHub
1:34:56 and hugging face is all the code you're
1:34:58 ever going to need to do remarkable [ __ ]
1:35:01 well Salesforce just in is leading a
1:35:04 financing round valuing hugging face at
1:35:06 $4 billion 44 billion doll it's it's
1:35:10 basically a code
1:35:12 bucket4 billion doll so there's
1:35:15 that um I thought that was pretty
1:35:19 significant um this is the new GPT 3.5
1:35:25 turbo fine-tuning and API update so this
1:35:28 is from open
1:35:30 AI um talking about um the you now have
1:35:34 the ability
1:35:36 to um
1:35:40 train GPT 3.5 on your own data now it's
1:35:45 pretty expensive it's like four times
1:35:47 the price
1:35:49 of of GP T4 calls which are are pretty
1:35:54 expensive but this doesn't require you
1:35:58 to to um use Lang chain and things like
1:36:01 that right so so I've talked before
1:36:05 about Lang chain like if you want to
1:36:06 train stuff on your own data You' got to
1:36:08 kind of jump through these hoops and and
1:36:09 be geeky this is still happening at the
1:36:12 API level this is not in chat GPT this
1:36:14 is on the back end with the with the
1:36:16 development playground so but but you
1:36:20 could probably I'm guessing
1:36:23 I'm just learning about it so I haven't
1:36:25 I haven't dug into it but I'm guessing
1:36:27 that you'll be able to do a zappier
1:36:30 automation where you're training GPT on
1:36:33 your data just through calls right and I
1:36:38 so big deal big big deal big big deal
1:36:42 big deal big deal big
1:36:45 deal this thing I've heard about before
1:36:48 but it it's getting written about this
1:36:50 is a thing called re rewind for
1:36:53 iOS
1:36:55 that um tracks everything you do on your
1:36:59 iPhone and remembers it and so you can
1:37:01 basically chat
1:37:04 GPT anything you've done so you know
1:37:07 when you're like I know I wrote an email
1:37:09 to someone at some point now you can
1:37:11 just sort of say what was the email I
1:37:12 wrote to some who did I send the email
1:37:14 to that talked about SP spacex's rocket
1:37:17 oh here are the four times that you
1:37:18 mentioned that you know twice in a
1:37:21 PowerPoint deck and twice in emails so
1:37:25 that's amazing that's called
1:37:27 rewind what's this oh this is just one
1:37:32 this is a legal um there there's
1:37:35 probably a ton of these but this just
1:37:36 one I saw on Twitter it's called
1:37:39 Spellbook and it will automatically
1:37:42 Redline legal contracts so you know how
1:37:46 lawyers Redline you know legal contracts
1:37:49 in very specific ways well this now just
1:37:51 does it for you so that's called
1:37:54 Spellbook and
1:38:00 then more more large language model data
1:38:04 retrieval for the Enterprise this is a
1:38:06 project called Mano so this is a a y
1:38:10 combinator company that is they're
1:38:14 creating a whole company around making
1:38:17 it easy for companies to train on their
1:38:21 own data right so this is going to be
1:38:23 the next absolutely the next big thing
1:38:26 is chat
1:38:27 GPT like ugb you GPT right or your
1:38:31 company GPT is going to be the next big
1:38:34 thing I think that was it yeah that was
1:38:36 it so holy crap man those are those are
1:38:41 all big or interesting kind of things
1:38:44 that like I didn't know about before I
1:38:45 got on here so that one this one
1:38:49 simulan holy crap
1:38:53 what's the website let's see if we can
1:38:55 find the
1:38:56 website simula simon.com for easy access
1:39:00 let's go get Early
1:39:03 Access don't you all get in there before
1:39:05 me hi what's your
1:39:11 name did you all hear that that they
1:39:14 figured out a way to um on MacBook Pros
1:39:19 they can take the audio of someone
1:39:22 typing and um reconstruct what they
1:39:26 typed just based on the audio of the
1:39:30 keyboard [ __ ]
1:39:33 crazy uh any social okay so I'm G to put
1:39:38 on my Tik
1:39:50 tock
1:39:58 um what else do I want to put in
1:40:01 there um I'll just do that Kyle
1:40:07 Shannon yeah that's
1:40:11 good to avoid spam and make sure you're
1:40:14 a real
1:40:15 person type the last letter of the word
1:40:18 simulan
1:40:20 okay thanks for registering all right go
1:40:23 register for simulan and you too can be
1:40:26 making robot pictures of people in real
1:40:28 time zoom calls are going to get really
1:40:31 interesting aren't they everyone's going
1:40:32 to be a flipping robot but not like a
1:40:36 shitty fake Google you know sticker
1:40:41 robot like a real
1:40:46 robot all right you know what I'm going
1:40:48 to do the
1:40:49 regulars get to get in thex box before
1:40:52 randoms yeah exactly kruno and you know
1:40:56 we should
1:40:59 that's yeah I think
1:41:02 I I think I need to experiment with
1:41:05 bringing people up on on stage these
1:41:06 things are long enough that that um it
1:41:09 shouldn't just be me talking so let me
1:41:11 think about
1:41:19 that okay I am wait he's gone wait he is
1:41:25 gone he left after his
1:41:26 comment
1:41:28 um so here's what I just did I just
1:41:31 jumped to the bottom because I was like
1:41:34 I was I was way far behind I was uh oh
1:41:37 uh oh dang it dang it dang it jig jig
1:41:43 jabbit I can't remember my quotes
1:41:47 yes let's see what were they do I have
1:41:53 them let's see a big mouth don't make a
1:41:56 big
1:41:57 man howy partner howy partner a big
1:42:01 mouth don't make a big man that's more
1:42:02 John Wayne a big mouth don't make a big
1:42:04 man I was too high I was too high in my
1:42:07 register all right big mouth don't make
1:42:10 a big
1:42:12 man oh wait this is okay this is my
1:42:15 favorite one life is hard it's harder if
1:42:18 you're stupid that's very AI learning
1:42:20 lab life is hard it's harder if you're
1:42:23 stupid all right all right we got it
1:42:25 life is hard it's harder if you're
1:42:27 stupid don't let me forget that um all
1:42:29 right so so I've kind of jumped to the
1:42:31 bottom of comments so I probably just
1:42:33 skipped a bunch of questions so if you
1:42:35 had questions that I didn't answer Reas
1:42:37 them down here if you still want to know
1:42:40 if I didn't cover them and then I'm just
1:42:43 going to jump to the bottom because I
1:42:44 was too far behind thank you cuno for
1:42:46 the
1:42:48 trumpet um I used to play the trumpet I
1:42:51 was a
1:42:52 trumpeteer I like the trumpet the
1:42:54 trumpet was an obnoxious add friendly uh
1:42:58 uh life is hard it's harder if you're
1:43:02 stupid all right we're getting there
1:43:05 we're getting there this good all right
1:43:07 it's good solid this is not quite
1:43:10 NPC um level stuff but it's
1:43:14 [Laughter]
1:43:17 close all
1:43:18 right um hug my face
1:43:22 movie making gigs are in trouble I
1:43:25 absolutely agree well I don't know that
1:43:27 they're in trouble movie making gigs in
1:43:30 the way they're done today are in
1:43:33 trouble but I think the world that's
1:43:36 opening up to us
1:43:39 is the democratization of high
1:43:45 quality creative
1:43:48 execution for lack of a better term so
1:43:50 like film making is still going to be
1:43:52 hard like telling a story that can keep
1:43:55 someone there for two hours is that's
1:43:58 hard and that's it's it's going to be
1:44:01 hard for a while to do to do really well
1:44:03 at world class but but like the
1:44:06 execution of like that real time being
1:44:09 able to swap someone out for a
1:44:12 robot we're going to be able to
1:44:15 do people who don't know anything about
1:44:17 film making are going to be able to do
1:44:19 really high quality Film Production
1:44:24 with you know little to no money little
1:44:27 to no tools um and it just it you know
1:44:32 it just just like the video camera kind
1:44:35 of democratized that ability and and we
1:44:37 got you know filmmakers like Spielberg
1:44:40 because of that
1:44:42 like we're gonna have a whole new rash
1:44:45 of of creative professionals that just
1:44:47 do things in a completely different way
1:44:49 it's going to be wild hugging face is
1:44:52 wildly useful yep hugging face is really
1:44:55 good well and now it's worth $4 billion
1:44:57 dollars and it's not that old I don't
1:44:59 think what is it like three years old
1:45:00 four years old something like
1:45:02 that uh hug my face I know hugging faces
1:45:06 are really they're G at some point
1:45:08 they're gonna have to deal with the name
1:45:09 wow oh yeah thank you Winston that was
1:45:13 awesome thank you very
1:45:16 much that was really
1:45:19 good life is hard but it's harder if
1:45:23 you're stupid all right I'm gonna go
1:45:25 find that clip oh you know what maybe we
1:45:27 can go find that clip here look here's
1:45:29 add at its finest let's have an add
1:45:31 moment shall
1:45:35 we um John
1:45:40 Wayne life
1:45:43 is
1:45:45 hard it's harder if you're
1:45:49 stupid
1:45:54 he's
1:45:54 here well what do you know Striker hello
1:45:58 Thomas thought you were in the first I
1:46:01 was sit
1:46:03 down just sit down Thomas anywh this is
1:46:06 two minutes I just want the
1:46:12 quote this life's hard man but it's
1:46:15 harder if you're stupid well that's not
1:46:17 John
1:46:19 Wayne
1:46:22 done where's my
1:46:23 [Music]
1:46:38 brother yeah all
1:46:42 right welcome to bangit news today life
1:46:47 is
1:46:48 hard yeah I just oh I just Clos close
1:46:51 that window can I get it back I think I
1:46:54 can six
1:47:04 tabs all right I can't find it whatever
1:47:08 someone will find
1:47:09 it annoyingly fine-tuning 3.5 is way
1:47:13 more expensive than fine-tuning three
1:47:15 yeah I know it's no it's it's four times
1:47:18 more expensive than uh G pt4 calls which
1:47:21 are expensive so yeah it's it's annoying
1:47:26 um so yeah and I haven't played with it
1:47:29 yet so I don't I don't really know how
1:47:32 useful it is but it's still you know all
1:47:35 of this stuff right now all of the
1:47:36 things that I'm talking about all the
1:47:38 tools that I'm showing just all take all
1:47:40 of them with a grain of salt even the
1:47:41 pricing with a grain of salt because as
1:47:44 they start to figure this [ __ ] out these
1:47:46 things are going to get it's just going
1:47:47 to get more and more sophisticated and
1:47:49 transparent and cheaper and all that
1:47:51 sort of stuff so
1:47:52 just learn learn what you can by the way
1:47:55 checking in on the the gamification of
1:47:57 the AI learning lab we're up to 77,000
1:48:00 likes you all are insane thank you
1:48:02 that's awesome I don't know what it does
1:48:04 I hopefully it's entertaining you it's
1:48:07 really funny holy [ __ ]
1:48:09 rewind um I hadn't seen doing that now
1:48:13 yeah
1:48:14 exactly um I use many AIS greater than
1:48:17 10 or about 10 I would say the worst one
1:48:20 I've used used is hugging face really
1:48:22 well but hugging face isn't a single AI
1:48:24 hugging face is thousands of AIS and
1:48:28 whatever one you used you might have
1:48:30 used hosted on hugging face um but it's
1:48:33 just a place for you to to try the
1:48:34 models download the models things like
1:48:36 that so that it's one of the challenges
1:48:39 with hugging face is it doesn't know
1:48:41 what it wants to be like it can run code
1:48:43 it's a code repository it's a model
1:48:46 Library it's it's it's a dessert topping
1:48:49 and a floor wax
1:48:51 sorry I left for a sec did you do John
1:48:53 John Wayne I did uh uh life is hard it's
1:48:58 harder if you're stupid uh but um I but
1:49:02 I tried to find the actual quote from
1:49:04 John Wayne and it was just it was too
1:49:05 much of a pain in the ass it's [ __ ] like
1:49:08 reading
1:49:09 keystrokes from audio or thoughts or
1:49:11 brain waves that scares me I know Sam I
1:49:14 I agree it
1:49:17 like the thing that that is not
1:49:21 a apparently obvious when you play with
1:49:25 chat
1:49:26 GPT is
1:49:28 that the technology that makes chat GPT
1:49:31 work it doesn't it doesn't care what
1:49:34 data it does what it does on meaning
1:49:39 chat GPT aimed it at words brain
1:49:42 scientists are aiming it at pictures of
1:49:45 brain scans um you know audio people or
1:49:49 hackers are aiming at at audio of
1:49:51 keyboards it's the same basic
1:49:54 technology um where they're just
1:49:55 training it on these massive data sets
1:49:57 of whatever data and then if you have
1:50:00 the right data set in there now it can
1:50:03 read your keyboard Strokes from audio so
1:50:05 this stuff's going to be everywhere and
1:50:07 this is this is the reason the other
1:50:08 night I was saying I don't know someone
1:50:10 asked something about you know the open
1:50:13 source models and what if the hackers
1:50:14 you know get to them I'm like well like
1:50:17 they're out like it was probably not
1:50:22 Super Wise for meta to release their
1:50:25 their their weights their core large
1:50:27 language models but they
1:50:29 did and they're good and
1:50:32 so hacker gonna hack developers going to
1:50:35 develop and you know hopefully the the
1:50:37 white hat hackers out haacker the black
1:50:40 hat hackers um but yeah we're in and and
1:50:44 it's it's not even the hackers we need
1:50:46 to worry about right it's it's the
1:50:48 government it's corporations it's all
1:50:50 that that sort of
1:50:52 [ __ ] all right 10:17 I'm going to go for
1:50:55 a bit longer I'm getting a little sleepy
1:50:57 now now the lack of coffee um adrenaline
1:51:02 rush is wearing off I'm
1:51:04 like perplexity chat GPT po Claude And
1:51:08 soon I hope xai yep so in your
1:51:12 list I'm trying to think if I would add
1:51:17 anything I would add Bard in there Bard
1:51:20 or Bing just because they're connected
1:51:22 to the
1:51:24 internet if you can get access to them
1:51:28 and then Pi Pi is worth putting on your
1:51:31 list this is Chris Martin um Pi is worth
1:51:34 putting on your list just for to
1:51:36 experiment with it being really
1:51:39 conversational chatty GPT is is what the
1:51:42 Irregulars have nicknamed Pi so I would
1:51:45 add pi and Bard or Bing to your list
1:51:48 anyone working on Aid driven object
1:51:51 recognition um I'm not but but you know
1:51:55 meta released they op sourced that there
1:51:59 that crazy ass object recognition thing
1:52:02 that like in real time was was
1:52:05 recognizing like every object in a video
1:52:07 like in Crazy real time um so I'm sure
1:52:12 there I'm sure a lot of the a lot of the
1:52:15 projects on GitHub and uh and hugging
1:52:18 face are are taking advantage of that
1:52:20 stuff
1:52:21 the Mind rating stuff is definitely
1:52:22 scary simulan is IOS only that's okay
1:52:25 because iOS is the
1:52:30 best make all you Android people
1:52:37 crazy wasn't it wasn't it a swell time
1:52:40 when when the the biggest political
1:52:44 divide in our country was Windows versus
1:52:48 MAC that was quaint
1:52:52 it would be nice if we went back to
1:52:55 that because it's ugly right now I don't
1:52:58 like our country right now my daughter
1:53:00 was a trumpet performance major oh
1:53:02 that's
1:53:03 awesome yeah being a trumpet performance
1:53:05 major she was she must have been
1:53:08 good I recall your trumpeting
1:53:13 past who are my favorites my favorites
1:53:16 were Maynard Ferguson obviously you have
1:53:18 to you have to
1:53:20 um Chuck
1:53:22 mangon uh those are the two
1:53:26 biggies Tommy what's his name from The
1:53:28 Tonight
1:53:30 Show uh uh uh uh we were all yelling at
1:53:33 you
1:53:34 L wait what were you yelling at me for
1:53:38 welcome to Rock
1:53:44 Bottom this is not
1:53:47 bottom do you did you ever do drum core
1:53:49 I did marching band I never did drumcore
1:53:52 though drumcore looks cool I I just did
1:53:53 marching band and then I got into the
1:53:56 theater I got on the stage I entered
1:54:00 upon the stage and I never
1:54:02 left and here we are I'm a
1:54:06 celebrity in my own
1:54:08 mind and in the mind of Tik Tock because
1:54:10 that's all I can be it's either myself
1:54:12 or a
1:54:13 celebrity no one else could possibly
1:54:16 have anything they'd want to report to
1:54:17 Tik
1:54:19 tock
1:54:25 good
1:54:26 Lord Silver Fox gave me a peach thanks
1:54:29 for the peach nice
1:54:32 Peach toas the little what's that little
1:54:35 Dino dude he's he's got a name doesn't
1:54:38 he thank you for
1:54:40 the by of a name oh unbrace Sun reached
1:54:44 level 15 nice thank
1:54:48 you um
1:54:54 content creation will change content
1:54:56 creation is going to be is it's not
1:54:58 going to change it's
1:54:59 gonna it's gonna it's gonna explode it's
1:55:03 GNA explode and be completely
1:55:05 reconfigured everything is gonna
1:55:11 explode I don't know if you've been if
1:55:14 you've been with me for a while like I
1:55:15 feel like I'm going through these phases
1:55:17 it's not quite phases of grief it's like
1:55:19 it's like phases of understanding the
1:55:21 implications of
1:55:23 AI like like when I first started this
1:55:27 channel it was definitely like we got to
1:55:30 keep up with the tools we got to keep up
1:55:31 with the
1:55:32 tools and then it and then it was like
1:55:35 we got to keep up with the implications
1:55:37 of the tools
1:55:39 and and now it's like I
1:55:46 just I want to understand what tools
1:55:50 [Music]
1:55:53 to I want to I want to know what tools I
1:55:56 need to understand to be able to glimpse
1:55:59 into what's
1:56:01 coming and it's really hard and every
1:56:04 time I get a little glimps of it I'm
1:56:07 just like holy [ __ ] like what is the
1:56:09 world going to be like and it's not that
1:56:12 far
1:56:14 away like that simulan thing where
1:56:17 you're just real time replacing people
1:56:21 in video with like
1:56:26 worldclass
1:56:31 CGI I don't know I just I
1:56:34 can't open AI just acquired a like a
1:56:37 World building game
1:56:40 company so so that says to me that chat
1:56:43 GPT you're going to be able to just say
1:56:46 build me a world I want to enter
1:56:48 it write me a game build me a world fly
1:56:52 me over the Eiffel
1:56:54 Tower design a new jet that
1:56:57 is you
1:57:00 know you know
1:57:03 um you know scientifically sound but
1:57:07 looks really cool
1:57:09 Okay and like all the engineering will
1:57:11 be in it I just I don't know
1:57:16 man I don't know it's wild it's just
1:57:19 it's just
1:57:21 wow like like I want I want to be able
1:57:24 to to provide some level of I don't know
1:57:29 it's not even Guidance just like Clarity
1:57:32 on what's what's important to pay
1:57:34 attention to what's coming all that sort
1:57:36 of stuff right like I like I feel like
1:57:38 that something I have in my life or in
1:57:40 my brain is is the ability to do that
1:57:42 but this stuff is so complicated and
1:57:45 it's moving so fast and what's coming is
1:57:47 so
1:57:48 profound I like I just feel like I I I
1:57:51 got nothing feel like I got
1:57:54 nothing hang on I gotta look at
1:57:56 something
1:58:07 here all right fellow trumpeteers
1:58:11 subgroup of Irregulars exactly I finally
1:58:14 sold I bought a really expensive uh
1:58:18 trumpet I bought like a $1200 trumpet
1:58:21 when I when I got some money when I sold
1:58:23 a previous company and
1:58:26 uh what I learned
1:58:30 was
1:58:31 like some [ __ ] you should just let go
1:58:34 like I just never played it I bought
1:58:35 $1,200 trumpet like with the fantasy of
1:58:38 how much fun I had with it in high
1:58:40 school but like I wasn't in high school
1:58:41 anymore I was like 35 and it just sat
1:58:43 there for like two years and I was like
1:58:45 oh [ __ ] you just kind of learned this
1:58:47 [ __ ] in life that like you Tred to
1:58:49 recapture certain parts of your life
1:58:51 when you have money and then it's just
1:58:53 like it's not the same but anyway so I
1:58:56 am I am I am in a trumpet free zone
1:58:58 right now watch out you've got some in
1:59:00 your
1:59:01 class going for 80 we're at 83,000 likes
1:59:06 good
1:59:08 Lord um overactive class overacting even
1:59:13 that that's what this channel is all
1:59:14 about this is the overacting channel for
1:59:16 sure 80,000 coming up fast guys tap tap
1:59:20 I love that this this channel is no
1:59:21 longer about AI this is just a
1:59:24 game it's the
1:59:27 best uh seeing a bit of backlash to chat
1:59:31 interfaces how do you see interfaces to
1:59:33 llms evolving that's a really good
1:59:36 question true human true human AI
1:59:39 James um 8.5 out of 10 on the name
1:59:43 you're close it's good I like it it
1:59:46 could you could tighten it up make it a
1:59:47 bit more memorable but it's good I like
1:59:48 it um
1:59:54 I'm not sure what what backlashes you're
1:59:57 seeing the chat like it you you
1:59:59 know the interface right now is
2:00:01 effectively Google right so so I could
2:00:04 see I could see a backlash
2:00:18 to my best
2:00:22 guess is
2:00:24 that our
2:00:28 devices it's it's just so
2:00:34 weird we'll still have the ability to
2:00:37 use
2:00:38 keyboards but voice is going to become a
2:00:41 lot more natural I
2:00:44 think one of the things that makes Siri
2:00:47 and Alexa and those kind of tools weird
2:00:49 right now is you can't have true
2:00:51 conversations with them some people try
2:00:53 to but they're they're they're manually
2:00:57 programmed right they're they're they
2:01:00 they can't just talk Pi can just talk
2:01:02 you know putting any of these chat
2:01:04 things into voice mode you can just talk
2:01:09 so so I think
2:01:13 that voice is going to become a more
2:01:16 important component I think Apple's
2:01:18 spatial computer
2:01:21 Paradigm or metaphor or or redefinition
2:01:25 of what computers
2:01:28 are kind of goes
2:01:40 there I think I think I think the input
2:01:43 for these devices will be whatever we
2:01:45 need it to be it could be a keyboard it
2:01:47 could be voice it could be
2:01:50 images it could you know just like you
2:01:52 know hold hold a book up to the camera
2:01:55 right
2:01:58 um so I think the input
2:02:01 becomes
2:02:03 everything it's just so weird and then
2:02:06 the output becomes
2:02:08 everything meaning we're going to have a
2:02:11 single
2:02:13 application that is
2:02:16 just our entry into this whatever this
2:02:20 magical [ __ ] machine
2:02:22 becomes
2:02:25 um
2:02:26 and I'll watch movies in it I'll listen
2:02:29 to music in it I'll write music in it
2:02:31 I'll make movies in it I'll write work
2:02:34 stuff in it I'll send emails like all of
2:02:36 the stuff that we do will do but it's
2:02:38 just it it is going to be through a
2:02:40 single interface right now it looks like
2:02:42 a chat
2:02:43 window but I think that kind of expands
2:02:46 and it you know it takes
2:02:48 in voice visuals it it's multimodal in
2:02:52 multimodal out and generative in the
2:02:58 middle so I don't know like I sure
2:03:01 there's a backlash on chat gbt right now
2:03:04 and people are like I don't like chat
2:03:05 gbt it hallucinates it spells something
2:03:08 wrong okay whatever
2:03:14 um this ain't going away this is this
2:03:17 is the the the prediction of the death
2:03:20 of of uh AI is is I find it kind of
2:03:25 humorous um it ain't this ain't getting
2:03:27 uninvented this is this is way too
2:03:30 powerful stuff to to go away
2:03:34 um and so I just think it's I think it's
2:03:36 Omni in and Omni out Omni Omni like
2:03:40 anything input anything output
2:03:44 generative in the
2:03:46 middle what's it generate whatever the
2:03:48 [ __ ] we want it to
2:03:53 the guy Townsen that I mentioned earlier
2:03:55 towns and wlaw we we're playing around
2:03:57 with
2:03:58 pi he was talking about the concept of a
2:04:01 magic Genie and and we're talking about
2:04:04 ai ai is a magic Genie it's like you ask
2:04:06 it you ask at anything and it grants you
2:04:09 a wish right and right now the wishes
2:04:11 are kind of limited right they're
2:04:14 they're a little little
2:04:16 limited but it can spew out a novel in a
2:04:19 minute
2:04:21 well in two three years it'll spew out a
2:04:27 novel
2:04:30 instantly and just start reading it to
2:04:32 you in the voice of your choice whatever
2:04:35 celebrity you like even if it's
2:04:39 me and then if you're like well I'd like
2:04:42 there to be an audio soundtrack to this
2:04:44 novel it'll just start playing and it'll
2:04:46 play so that it matches the novel
2:04:50 and is something that is emotionally
2:04:54 Pleasant to you because it will
2:04:56 know if that song is working or not and
2:04:59 it will dynamically rewrite the song as
2:05:02 you listen to it to make you
2:05:06 happier like that kind of
2:05:09 [ __ ] so I don't know man like I'm I'm
2:05:13 just right now that that box for me is
2:05:16 just the metaphor of it's you're going
2:05:19 to be able to dump anything in it and
2:05:21 your desire and what's going to come out
2:05:24 is what you
2:05:28 want you guys are making me think it's
2:05:30 making my brain hurt it's making my
2:05:32 brain hurt there's not a lot going on up
2:05:35 here this prefrontal cortex is like a
2:05:38 piece of cheese and there's something
2:05:40 going on somewhere back here I don't
2:05:42 know what's happening don't make me
2:05:44 think what's your juicer today uh it's
2:05:48 uh
2:05:51 let me see
2:05:52 uh when I'm feeling a little parched I
2:05:55 go with water loop it's the BlackBerry
2:05:58 lemonade sparkling water
2:06:07 Costco can I have a meeting with you um
2:06:11 the [ __ ] the [ __ ] listen
2:06:14 let me tell you something Mr Shannon I
2:06:16 think we're going to have to have a
2:06:17 meeting my my name's the [ __ ]
2:06:20 you on the other hand consider yourself
2:06:22 a celebrity we got to work some things
2:06:24 out I don't think you're a celebrity I
2:06:26 don't think you're even close so we're
2:06:28 gonna have a meeting I'm gonna do the
2:06:31 talking you're G to do the listening all
2:06:33 right you [ __ ] oh I'm sorry I'm
2:06:35 the [ __ ] you're the [ __ ]
2:06:37 well my name's the [ __ ] name's
2:06:39 actually Bob name's actually Bob but
2:06:41 yeah but we should definitely talk
2:06:44 because I'm the [ __ ] you know
2:06:45 what I mean you know where this is going
2:06:47 neither do I uh yeah just send me a DM I
2:06:50 got my link tree with my contact info in
2:06:54 the proof H happy to chat um exactly
2:06:58 what is your T-shirt today Brooklyn Only
2:07:00 the Strong
2:07:02 Survive uh hey let's organize a group to
2:07:05 send the hat at once oh maybe time for
2:07:09 his quote Oh that's pretty funny you're
2:07:12 see the IR okay here's here's a problem
2:07:15 when the Irregulars start organizing
2:07:18 some shit's going down I don't think
2:07:20 it's going to be good but uh you guys
2:07:22 are
2:07:23 funny um how many names you know of huh
2:07:27 I don't know what you're talking
2:07:29 about uh uh uh uh uh uh uh the
2:07:33 [ __ ] toilet paper John Wayne
2:07:36 toilet paper Okay um uh don't know what
2:07:39 you mean where's the like button I have
2:07:41 no idea how to Tik Tok true human AI I
2:07:44 don't either this is like Elder talk
2:07:46 welcome to Elder talk old people try to
2:07:49 use new
2:07:50 technology uh you just double tap the
2:07:52 screen uh true human and uh the hearts
2:07:55 just go and we are now up to Heart check
2:07:59 like check 85,86 you guys are just weird
2:08:03 all right um 82 new messages I just
2:08:08 jumped to the bottom of the messages
2:08:09 good Lord 880,000 I applied for an AI
2:08:13 job unraced son congratulations that's
2:08:17 awesome according to Snopes that wasn't
2:08:20 J John Wayne actor that said the life is
2:08:22 hard quote Oh
2:08:25 interesting yeah because that that there
2:08:27 was another movie that had it it wasn't
2:08:29 John Wayne so that was probably the one
2:08:32 that said it but I another thing came up
2:08:35 when I was looking at
2:08:38 um at
2:08:40 YouTube life is tough but tougher when
2:08:44 you're
2:08:47 stupid
2:08:50 gun where's my brother right there he
2:08:53 [Music]
2:09:16 hit I told you you'd need this bad
2:09:21 someday thanks
2:09:23 Marine no that wasn't it it looks like
2:09:27 it must be in the Sands of
2:09:29 eima life is tough but tougher when
2:09:32 you're stupid John
2:09:38 Wayne I don't know I don't know we'll
2:09:42 find
2:09:43 it unbrace sun pi. a is my BFF 80,000
2:09:49 strong um I'm going to jump down to the
2:09:52 bottom again because I'm way far behind
2:09:54 so I'm just going to do it I'm just
2:09:56 doing
2:09:57 it um I've really for Kyle's llm heating
2:10:01 up with my question wait I've
2:10:06 really I don't quite know what that
2:10:08 sentence means that's the thing I feel
2:10:10 like nobody gets yet it's just a toy no
2:10:13 no it's really not
2:10:15 Sam this is what's Sam corn said
2:10:21 there he says the thing that I that that
2:10:24 I feel like people don't get is they
2:10:26 look at they look at chat
2:10:28 gbt and you look at something like pie
2:10:32 and it looks like a
2:10:34 toy and if you go to chat GPT and you're
2:10:37 like write me a song you do the parlor
2:10:39 tricks right write me some code write me
2:10:41 a song and you do the parlor tricks it
2:10:43 seems like a
2:10:44 toy these are not toys they're not toys
2:10:47 they're like there's
2:10:49 so they're so much more profound and
2:10:52 it's like I can't quite articulate it
2:10:54 yet I can't quite articulate why but the
2:10:57 thing I keep going back to is what were
2:10:59 they trained on they were trained on the
2:11:01 public
2:11:03 internet what's the public internet well
2:11:05 it's the place where for the last four
2:11:09 decades we've
2:11:12 all expressed ourselves we've
2:11:15 self-expressed in idiotic ways in
2:11:18 brilliant ways in moving ways in Stupid
2:11:21 Ways in
2:11:22 right so there's 40 and probably more
2:11:26 there's probably more like you know
2:11:28 hundreds of years of
2:11:31 human
2:11:33 output that's been put on this internet
2:11:36 thing and has been ingested into these
2:11:39 machines in a
2:11:44 way that is not like
2:11:47 Google
2:11:50 these things have not indexed all of the
2:11:52 [ __ ] they've learned right Google said
2:11:56 we're going to go look at all the
2:11:57 websites and we're going to index them
2:11:59 so we can rank them so when you ask for
2:12:01 a search we're not going to give you
2:12:03 what's in the website we're going to
2:12:04 give you the link to the website with a
2:12:06 waiting right that's the index this this
2:12:10 is not
2:12:11 indexed
2:12:15 words it's it's like I I feel like what
2:12:19 what scientists have actually done with
2:12:22 the combination of neural networks
2:12:25 Transformer
2:12:26 architecture and you know the the let's
2:12:30 give open AI credit for the you know
2:12:32 they were the ones that really went
2:12:33 aggressively after let's put as much
2:12:37 data into this thing as we can let's go
2:12:40 get as much of the internet as we can
2:12:42 and jam it in this
2:12:46 thing they've created
2:12:49 kind of a brain right kind of a brain
2:12:51 that knows a lot of what Humanity has
2:12:55 contributed so when you talk to this
2:12:57 thing it's it's like all it's like the
2:13:01 history of humanity responding back to
2:13:03 you right there's a there's a collective
2:13:05 knowledge in here the connective
2:13:08 knowledge of us is in
2:13:11 here and it looks like a toy because
2:13:14 it's so simple it's not a toy it's just
2:13:19 [ __ ] wild man I don't know why my
2:13:21 mind is so blown
2:13:24 tonight it could be the lack of
2:13:27 coffee but I like I I
2:13:33 just I feel like I'm in a transition
2:13:35 period I'm in a Lial State
2:13:39 between kind of having my head around
2:13:41 this [ __ ] and not having a [ __ ]
2:13:47 clue
2:13:49 I just I'm trying to imagine even a year
2:13:51 from now what these things are going to
2:13:53 look
2:13:56 like I don't know I don't know man my
2:13:59 mind's
2:14:02 blown you got a lot way more than
2:14:04 nothing Kyle you are providing comedy
2:14:06 and info against a humorless well thank
2:14:09 you for that a humorless backdrop I so
2:14:13 thank you for that I do I feel like
2:14:15 we're living in a bit of a humorless
2:14:16 world right now like everything is so
2:14:18 [ __ ] serious the [ __ ] climate is
2:14:21 literally exploding around us you know
2:14:24 the political discourse in the States
2:14:26 has infected you know the rest of the
2:14:28 world or I don't know maybe it was a
2:14:30 global global phenomenon we're still in
2:14:33 [ __ ]
2:14:35 PTSD mourning like [ __ ] dealing with
2:14:39 coid and Co now it's resurgent like it's
2:14:41 just like the world right now is like
2:14:43 this humorless serious [ __ ] place and
2:14:47 then on top of it
2:14:49 this [ __ ] is going to obliterate what we
2:14:53 currently like know as work and how we
2:14:57 work and what we work and what is work
2:14:59 and what
2:15:01 is oh my God so yeah I think I think
2:15:05 part of what we have to do is laugh
2:15:06 about it because I don't
2:15:09 know but thank you for the kind
2:15:12 words oh it's exciting it is it really
2:15:16 is exciting where where will our World
2:15:18 being in 10 years with AI cam cam
2:15:22 katkin okay I'm giving you a 9.5 out of
2:15:25 10 on the name I assume that's your
2:15:26 actual name that's an awesome name
2:15:29 name's cam katkin it sounds like you
2:15:32 should be in some sort of you know
2:15:34 Escape movie like an escape room movie
2:15:37 he you know cam cap capkin is the one
2:15:40 that like leads the the crew to uh to
2:15:44 figure out the code to get out of the
2:15:46 room to save the world
2:15:51 um where will our world be in 10 years I
2:15:54 I I I have no idea I I I'll give you an
2:15:57 answer that I'm going to steal
2:16:00 from um a friend of mine named Matt baa
2:16:03 who's a who's a technologist and he's a
2:16:06 CTO at a a Pharma
2:16:09 company and I said I said on a panel
2:16:12 that I didn't know like like I felt like
2:16:15 in the early days of the worldwide web I
2:16:18 could look at what the technology was
2:16:21 and I could Envision what the world
2:16:24 looked like 30 years out and it like and
2:16:26 it essentially looks like that right
2:16:27 like it like I could I could see that
2:16:30 like I could see you know if you
2:16:33 collapse geography and you collapse time
2:16:36 with instant
2:16:37 Communications you can kind of play
2:16:41 out you know even back in the mid 90s
2:16:44 what what was going to come you maybe
2:16:46 couldn't see all the stuff but you could
2:16:47 generally see
2:16:48 we're going to have instant
2:16:49 Communications you know we're going to
2:16:52 have society's going to become a global
2:16:54 Society right that for the most part has
2:16:56 happened um and and you know just how we
2:16:59 do everything is everything's instant
2:17:01 right so everything's there so you kind
2:17:02 of see that this stuff I can't see so so
2:17:06 after I mentioned that on a panel that I
2:17:08 was on Matt called me afterwards he goes
2:17:10 I think I've got it and his his timeline
2:17:13 was like 50 years out but I think we'll
2:17:15 we'll start to see this in 10
2:17:17 years is he said
2:17:19 that imagine a world where all of the
2:17:23 complexity is
2:17:25 gone like the complexity of
2:17:28 systems so if you think about how
2:17:30 complicated our world is and if you
2:17:32 think about what many many companies
2:17:35 are there's a lot of companies that
2:17:38 exist that their sole purpose in life is
2:17:40 to deal with the complexity of the world
2:17:43 right so you've got H&R Block tax
2:17:46 company what what is their purpose in
2:17:48 life well someone there reads the
2:17:50 [ __ ] tax code every year and then us
2:17:54 dummies walk in there and go I I don't
2:17:56 know what to do could you make my
2:17:59 numbers could I get money back
2:18:02 like what so what have they done they've
2:18:04 taken the complexity of tax preparation
2:18:07 and turned it into a you know you can be
2:18:10 a dumb Dum
2:18:12 right um Insurance Brokers um uh
2:18:16 financial advisors uh just think think
2:18:19 of any complicated system um you call up
2:18:23 any [ __ ] company in the world today
2:18:25 and you sit on hold for 45 minutes and
2:18:28 then you get put into this Byzantine
2:18:31 [ __ ] phone tree this logic tree of
2:18:34 well you're at level one support and
2:18:35 Level One support can only read this
2:18:37 script and then if you get past them
2:18:39 then you get to level two support and
2:18:41 like the whole [ __ ] thing is designed
2:18:44 to a get rid of the people that could
2:18:46 actually answer questions for cheaper
2:18:48 people right but now it's this Byzantine
2:18:52 complicated [ __ ] awful thing there's
2:18:56 companies that build that Byzantine
2:18:58 complicated thing there's people that
2:19:00 operate their businesses within that
2:19:02 complicated thing what if that went
2:19:05 away what if when you called your
2:19:11 company they picked up the
2:19:14 phone and talked like you talked because
2:19:18 they know where you're calling
2:19:20 from so they say okay he's in Denver
2:19:22 probably talks kind of like a Denver
2:19:25 person oh he's in Tuscaloosa Alabama
2:19:28 probably talks like a Tuscaloosa person
2:19:31 right so someone that that talks like
2:19:33 your
2:19:34 region answers the
2:19:37 phone has a conversation with you you
2:19:39 can ask it anything including like hey
2:19:42 how's your day going oh pretty good
2:19:43 how's yours
2:19:45 going right go talk to pie
2:19:48 you'll you'll start to experience this
2:19:51 and then that machine will be
2:19:55 authorized to do all the [ __ ] it will
2:19:57 know all the [ __ ] and it will be
2:19:59 authorized to do all the [ __ ] except for
2:20:01 the
2:20:02 anomalies and it will be smart enough to
2:20:04 know oh Kyle's actually got a problem
2:20:07 that's outside of what I can
2:20:10 do why don't I connect you with a human
2:20:12 being that can actually help you and it
2:20:14 automatically will route me with a
2:20:17 summary of my conversation to the human
2:20:20 being that can actually make some
2:20:21 judgment
2:20:23 calls so in 10 years imagine where you
2:20:28 know someone before mentioned that that
2:20:30 there's there's kind of a a push back on
2:20:32 the chat
2:20:34 interface but imagine in 10 years where
2:20:37 every single task that you do that's the
2:20:39 hard [ __ ] to
2:20:41 do is just like a
2:20:44 button refinance my house get a car
2:20:49 loan
2:20:52 um uh pay my
2:20:55 taxes and then maybe with pay my taxes
2:20:58 there's like pay my taxes and then
2:20:59 there's like what's your risk tolerance
2:21:01 and you slide it left and right like I
2:21:04 want you can get more money back but you
2:21:06 know higher chance of an audit or less
2:21:08 money back and you're safe so most of us
2:21:11 are sort of keep it somewhere in the
2:21:12 middle just I I just want to get a
2:21:14 little bit money back it' sure be nice
2:21:17 if I could get that
2:21:20 PS5 so
2:21:23 so so I think in 10 years we will likely
2:21:28 see a layer an interface
2:21:32 layer that hides all of the complexity
2:21:36 all the complex systems sitting beneath
2:21:39 it and then what that does to the
2:21:43 economy I don't have a [ __ ] clue
2:21:45 because think about all of the companies
2:21:48 that their their their sole purpose in
2:21:51 life is to deal with those complex
2:21:54 systems and have human beings deal with
2:21:56 them and greet you and read the [ __ ]
2:21:59 manuals and read the tax code and do all
2:22:02 that what happens to all those companies
2:22:05 I they go away they like they have they
2:22:08 they'll have to become the application
2:22:10 layer like H&R Block will
2:22:12 become the agent where instead of
2:22:15 walking into the office you just say I
2:22:17 want to do my taxes and up might pop
2:22:20 here's you know four options do you want
2:22:21 H&R Block do you want the chat GPT you
2:22:25 want chat GPT just to do it
2:22:28 dynamically do you want to use TD Merit
2:22:31 trade because you trust the brand but
2:22:33 it's all doing the same [ __ ] but then
2:22:35 you just push the button slide your risk
2:22:37 tolerance and your taxes are
2:22:40 done so I and
2:22:46 then what happen happens to us it so
2:22:49 think about how many hours a week you
2:22:55 spend dealing with complex
2:22:58 systems the 45 minutes on hold the
2:23:03 the you know all of
2:23:08 it what if all those hours were given
2:23:11 back to
2:23:13 you what if instead of 38 hours a week
2:23:16 of just dealing with nonsensical over
2:23:19 complicated [ __ ] that our society
2:23:22 has either purposefully or just through
2:23:25 inertia made
2:23:27 everything way too [ __ ] complicated
2:23:29 what if you had 38 hours where you
2:23:31 weren't abjectly frustrated with the
2:23:35 state of the
2:23:36 world I think people would be less
2:23:40 [ __ ] it might give us a chance to
2:23:42 reconnect with one another as human
2:23:46 beings we might have the energy and the
2:23:49 passion and the clarity to be able to go
2:23:51 you know what I've always wanted to
2:23:53 do needle point I think I'm going to
2:23:56 teach myself needle point like how many
2:23:59 like passions aren't
2:24:03 happening because the world's just so
2:24:06 [ __ ]
2:24:07 exhausting what if it
2:24:10 weren't now are we going to be there in
2:24:12 10 years no but I think in 10 years
2:24:14 we'll we'll at least have a layer where
2:24:17 you can start to see glimpses of it and
2:24:19 then 30 years out it probably starts to
2:24:22 be pretty seamless where you just ask
2:24:25 for what you want and it essentially
2:24:30 materializes who do we get to be in a
2:24:33 world where we have a magic
2:24:36 Genie we get to be whoever the [ __ ] we
2:24:38 want to be like that that to me is the
2:24:40 really exciting part like that's why I'm
2:24:41 excited about this because that's the
2:24:44 future that you know the doomers aren't
2:24:48 seeing the doomers are like the robots
2:24:49 are going to [ __ ] kill us okay they
2:24:51 might but what
2:24:53 if the world actually just got
2:24:56 better it's a
2:25:00 possibility reality is always likely
2:25:03 somewhere in between the polls
2:25:05 but um people freaked out when Web 2.0
2:25:08 started too yep exactly how long do you
2:25:10 think an AI project manager will last as
2:25:13 far as
2:25:15 time I think AI operations and project
2:25:19 management around AI projects we got we
2:25:22 got a lot of years of so so getting to
2:25:25 the 10 years where that Simplicity layer
2:25:27 is there all that shit's got to be built
2:25:30 even though it's going to be easy to
2:25:31 build it because of these tools project
2:25:33 management is going to be really
2:25:34 important and AI operations is going to
2:25:37 be really important so I think I think
2:25:39 AI project management or AI operations
2:25:43 um having a clue being able to keep up
2:25:46 with the change in technology and like
2:25:49 like a person who can sit between the
2:25:53 business needs and the and the
2:25:55 technology and and and be a translator
2:25:58 between those two I think they're going
2:25:59 to
2:26:00 be you'll be able to demand your price
2:26:03 because every company's going to be
2:26:05 trying to figure out how do we keep up
2:26:07 with it and how do we reinvent our
2:26:10 business using these tools so someone
2:26:13 that can sit in the middle and
2:26:14 understand both sides of that
2:26:16 conversation are going to be invaluable
2:26:19 so I would I would expand it Beyond
2:26:21 project
2:26:23 manager to that more strategic kind of
2:26:27 thing but a long time I think a long
2:26:35 time I find so little humor and good
2:26:38 info on TV I completely quit watching
2:26:40 the idiot box and now you got the idiot
2:26:43 channel on Tik Tok we're all here we're
2:26:45 all just being we're just we're just
2:26:47 being cards we're just being cards it's
2:26:50 hilarious it's
2:26:51 hilarious um What About Bob what's going
2:26:56 on I'm busy Bob likes Mania song
2:27:00 creation apps on AI are mostly toys they
2:27:03 are for now but have you seen um have
2:27:07 you seen uh what's it called
2:27:11 uho chaos generation so sunno is
2:27:15 a text a music generator but it's a
2:27:19 lyrics to singing generator so it
2:27:21 generates the song you you write the
2:27:23 lyrics or no sorry you have catb write
2:27:26 the lyrics and then you put it in this
2:27:28 thing and describe the kind of song you
2:27:29 want and it just makes
2:27:31 [Music]
2:27:35 ITB and
2:27:44 SEL [ __ ] leg I'm like a [ __ ] chicken
2:27:51 fresh
2:27:53 remix so still kind of kind of crappy
2:27:57 and toy like but you know
2:28:00 again don't forget that a year and a
2:28:02 half ago text to image generators were
2:28:05 comically
2:28:07 bad and now they're terrifyingly good
2:28:10 that was a year and a half so we now
2:28:13 have a thing that can write a 20 second
2:28:15 or 30 second song
2:28:18 with a beat with background music with
2:28:23 [Music]
2:28:26 vocals um at this level of
2:28:29 quality is it great no but again throw
2:28:33 this a year in the future throw this
2:28:35 five years in the future what does this
2:28:37 thing become what are the inputs the
2:28:40 inputs could just
2:28:42 be I've got sensors near me that that
2:28:47 understand my emotional state and when
2:28:49 you say well what sensors would you have
2:28:50 well if you have an Apple Watch that's
2:28:52 full of sensors imagine the amount of
2:28:54 sensors on this thing in five
2:28:58 years imagine you know those Vision Pro
2:29:01 goggles you know get to a form factor
2:29:03 not dissimilar to what I'm wearing right
2:29:06 here so all of these sensors are going
2:29:09 to know everything about us they're
2:29:11 going to remember what we've said
2:29:12 they're going to remember what we're
2:29:13 talking about they'll likely know some
2:29:15 version of what we're thinking what our
2:29:17 emotional state is and I can just say I
2:29:22 want to play my Spotify make me feel
2:29:24 better list and it will just
2:29:27 dynamically create music that it knows I
2:29:30 need to
2:29:34 hear like that's where this stuff goes I
2:29:39 think again Omni in Omni out you take
2:29:42 any input put it through this generating
2:29:45 machine and any out put and all of it
2:29:48 happening in real
2:29:50 time it's just [ __ ] wild isn't
2:30:00 it Mandela affected Mandela affected
2:30:04 Mandela affected Mandela
2:30:08 affected the quote is when they're in
2:30:11 practice bayonet practice okay I'll go
2:30:13 look at it so it is the the sanss of
2:30:15 eima
2:30:17 the the John Wayne thing I'll find it
2:30:21 but what's wait what's the one that I
2:30:23 liked it wasn't that
2:30:27 [Music]
2:30:32 one is
2:30:35 it oh I lost it oh
2:30:39 well yeah lost it what are you gonna
2:30:43 do almost to 100K oh my God 100,00 th
2:30:47 likes all right I gotta go this this is
2:30:50 ridiculous it's 11 o'clock I do have to
2:30:52 go let me answer two more questions two
2:30:54 more just two more all right so what
2:30:56 would you say is a good name for this
2:30:58 sort of AI job AI project consultant
2:31:11 um
2:31:15 H I'm I'm
2:31:17 I I'm reticent to to put AI in it
2:31:20 because AI is going to be a ubiquitous
2:31:22 thing it's like saying you know like
2:31:23 having a chief digital officer like
2:31:26 everything's digital right so it becomes
2:31:29 kind of this weird redundant thing like
2:31:31 like it's necessary now because it's
2:31:33 new so I guess we keep it in there for
2:31:36 now
2:31:42 um AI strategist AI business strategist
2:31:46 um
2:32:04 you you know if you want to keep it in
2:32:05 the in the project management realm if
2:32:07 you want to keep it in in the more like
2:32:09 getting [ __ ]
2:32:15 done
2:32:17 oh you know you know could be
2:32:18 interesting is something around so
2:32:21 actually it would be interesting play
2:32:23 around with chat GPT like put in put in
2:32:27 like who you are what you
2:32:29 do and and start having a dialogue about
2:32:32 it there could be something interesting
2:32:34 around playing around with the idea of
2:32:36 like AI change manager you know change
2:32:40 management that that all this AI stuff
2:32:44 is gonna is going to precipitate
2:32:47 massive amounts of change and so if if
2:32:50 you want to if you put yourself at the
2:32:52 center of that that I'm an AI change
2:32:53 manager what's that mean well shit's
2:32:55 changing so I need to understand the
2:32:57 technology side of this I need to
2:33:00 understand the business side of this so
2:33:03 these Technologies make these things
2:33:05 possible which means this part of the
2:33:06 business needs to change so I'm going to
2:33:08 be the a AI change manager that I can
2:33:12 manage both of those so maybe maybe that
2:33:16 something like that I would go play with
2:33:18 chat GPT and and and dig in there with
2:33:22 it
2:33:23 because uh wait what is making music
2:33:27 what AI it's called suo s n
2:33:31 o and I think it'sso doai hang on you
2:33:35 have to make the music in Discord um
2:33:39 sun. I mean I like it funny you know
2:33:42 wait wait what was making that
2:33:45 music yeah
2:33:48 sun.
2:33:49 um that's the thing these tools are
2:33:51 coming so fast and so Furious that like
2:33:55 like you you if you want to keep up with
2:33:58 it which at this point I've given up
2:33:59 trying to keep up with it like I'm I'm
2:34:01 like checking in
2:34:03 occasionally to see what's new because
2:34:06 it's you can't you literally can't like
2:34:09 the only way I think you could
2:34:10 realistically keep up with this stuff is
2:34:12 pick a vertical say I'm just going to
2:34:15 look at a
2:34:17 legal applications that that deal with
2:34:20 law [ __ ] and even then I don't think you
2:34:23 can keep up with it because you have
2:34:24 General tools like chat GPT that can do
2:34:26 a bunch of law [ __ ] too and then you got
2:34:29 these legal specific things like I I
2:34:31 just think I don't know it's
2:34:34 crazy I have very little clue what I'm
2:34:36 doing with this open- Source digital
2:34:38 audio workstation also other
2:34:42 AI yeah I
2:34:45 don't I I told myself for years I'd get
2:34:47 good at music production but it was
2:34:49 always too complicated for me actually
2:34:51 music production is one of those things
2:34:53 where it is incredibly complicated where
2:34:57 an AI simplification layer on top of it
2:35:00 is going to it's going to there's going
2:35:02 to be a lot of people that write really
2:35:05 amazing music without having music
2:35:09 theory knowledge without having
2:35:11 production knowledge and they're going
2:35:12 to be able to make worldclass [ __ ] and
2:35:15 some of it's going to be derivative but
2:35:17 you're going to you're going to find
2:35:19 people that Embrace these
2:35:21 tools that maybe aren't
2:35:24 musicians but they Embrace these tools
2:35:26 in such a way that they become
2:35:28 masters of generating highquality hit
2:35:34 music just like the DJs and and the
2:35:37 hip-hop era like the musicians like well
2:35:40 that's not really music they're sampling
2:35:42 or that's not really music they're just
2:35:43 mixing you know this beat with that beat
2:35:46 with
2:35:47 DJs yeah but certain DJs take it to
2:35:51 another level certain rappers took it to
2:35:53 another level right where it became a
2:35:55 new genre I think that's going to happen
2:35:58 where it'll probably be young people
2:36:01 that don't know any better and they'll
2:36:04 come in and they you know [ __ ] some [ __ ]
2:36:05 up they're going to make really cool
2:36:07 music and they're going to do it a lot
2:36:08 and consistently and I don't know like
2:36:11 how it gets into the world is going to
2:36:12 be a different thing and and I don't
2:36:15 know it's going to be it's it's going to
2:36:16 be that's going to be fun to
2:36:18 watch what in tarnation did we just
2:36:20 accomplish 100,000 likes it's
2:36:23 unbelievable sharebear unbelievable you
2:36:26 guys are just wild all right I'm gonna
2:36:29 go um I'm down here at the bottom thank
2:36:32 you everybody for hanging out let's I I
2:36:35 want to do some shout outs to to some
2:36:37 Irregulars so chaos Generation Um cam
2:36:40 I'm going to throw you in there because
2:36:41 I've seen you a couple of times uh who
2:36:44 else do we have down here tonius Silver
2:36:47 Fox
2:36:49 sherbear um Steve
2:36:52 Pennington um y'all are awesome thank
2:36:55 you thank you thank you Emilio's wife uh
2:36:58 who else do we have in here Laura C has
2:37:02 been in here a lot thank you um I'm I'm
2:37:05 missing a ton of people this is like
2:37:07 this is like why you don't thank people
2:37:09 on award shows because you miss someone
2:37:11 like hey you didn't mention me I'm in
2:37:13 here all the time little son of a [ __ ]
2:37:15 she I'm not even going to in there
2:37:16 anymore well I guess I will If he if he
2:37:18 shows up I'll go back
2:37:20 in um but but this has been a lot of fun
2:37:22 tonight thank you very much uh I I guess
2:37:25 I'll bring back my coffee so I'm not so
2:37:27 [ __ ] cracked out like I am tonight
2:37:29 thanks bear um thanks everybody uh steo
2:37:33 steo I forgot to mention uh Silver Fox
2:37:36 thanks so much peace peace peace um I
2:37:39 got a board meeting tomorrow for Story
2:37:41 vine
2:37:42 so prepared for that I'm I'm good with
2:37:45 that and then yeah I'll be back tomorrow
2:37:47 night and uh peace out everybody sleep
2:37:50 well it was fun thanks Laura next time
2:37:52 Bob well Bob well Bob tell him what he's
2:37:55 won you wanton a new car you can make
2:37:58 money with
2:37:59 chbt all right bye
2:38:01 everyone