aiLL 9/15/23 - Embracing AI: Becoming "AI Native"

Description

Chapters

Transcript

0:27 welcome everybody hey Dr J shaking
0:31 happy Friday
0:32 night it's a party party
0:39 party what is
0:41 happening so there is that good evening
0:45 hello how are you happy Friday
0:49 night got on a new t-shirt a little
0:53 little binding around the
0:59 neck oh
1:01 happy Friday cuno what's happening
1:03 what's shaking what's shaking what's
1:05 going down what's happening what's
1:07 happening with you yeah yeah see yeah
1:10 all right yeah all right we're going to
1:13 I think I put the makeup on a little too
1:14 thick tonight what do you say all
1:17 right all right welcome to the AI
1:20 learning lab this is Ali the AI learning
1:23 lab we'll let some people get in
1:25 here say hello say hello good day good
1:29 evening good evening welcome to the AI
1:32 learning lab my name is Vincent
1:36 Price that's a very old reference older
1:39 than all of us
1:43 combined all right
1:46 welcome I don't know if you knew this
1:49 but this is the AI learning
1:51 lab and this is an AI learning lab get
1:55 it it's comedy it's a visual comedic
1:59 element welome all right so what are we
2:02 going to do here uh we will uh we'll let
2:05 some people come in then we'll talk
2:07 about Ai and then we'll uh talk about
2:11 going AI native I put a new video up
2:13 tonight if you haven't seen my latest
2:15 video I think it's a fun one I had a
2:17 good uh I had a good talk in my uh AI
2:20 office hours on LinkedIn
2:22 today um Jim Ross who's been in here
2:28 before and a guy named G they took a
2:31 class
2:32 called um being AI native in business or
2:36 something like that and so we talked
2:38 about that and it it hit me
2:41 that I never really I I really
2:44 considered the idea of a what like what
2:46 is AI native you know what does that
2:49 mean other than you know the kids that
2:52 are born today are going to be AI native
2:54 right the kids the kids that are sort of
2:56 two two and under today are not going to
2:59 know a world without a
3:01 I but what's it mean for us old farts
3:04 that are just starting out on this
3:05 journey and what hit me
3:08 was it was it was actually a combination
3:11 of what we did last night so so Amelio's
3:14 wife hey Ted looks at life what's
3:16 happening Amelio's wife came in last
3:18 night and she said I want you I want to
3:20 see if you can use chat BT to help
3:22 figure out my stupid open enrollment
3:24 health insurance [ __ ] and I'm like yeah
3:26 we can do that and I don't know we spent
3:28 like 45 minutes just [ __ ] around with
3:30 chat gptt and we ended up getting it to
3:33 sort of work in Claude and then Claude
3:35 ran out of credits and it you know we
3:36 didn't quite get there but it was like a
3:38 good 45 minutes of of the live was spent
3:43 doing something that if you had just
3:45 spent the same 45 minutes looking at the
3:48 piece of paper you'd have probably
3:50 gotten to the same result right and then
3:53 when I heard these guys talking about
3:55 this AI native class and and and the the
3:58 basic idea was you basically try to do
4:01 everything that you do using
4:04 Ai and I thought oh that's kind of
4:06 interesting and then so what hit me what
4:08 the video is about if you go watch it
4:12 is okay if it's going to take you 45
4:14 minutes to read the document
4:16 traditionally and 45 minutes to dick
4:18 around with AI and maybe or maybe not
4:21 get the answer which one do you
4:24 choose I think you choose [ __ ] around
4:27 with AI for 45 minutes because at a
4:30 minimum you'll learn something and it'll
4:32 probably be able to do what you want it
4:33 to do and and that that I'm really
4:37 excited about uh clear warned AI I'm 53
4:41 because of AI learning lab and my
4:43 inclination in multimedia I'm all in
4:46 that's awesome AI native go AI
4:49 native I'll be 54 next month I'll be 59
4:54 in
4:55 May so got a few years on you but
4:59 awesome awesome awesome welcome um so
5:03 welcome everybody this is the AI
5:04 learning lab that's Ali the AI learning
5:06 lab and what we do here is as you can
5:10 imagine we wax poetic about this this uh
5:14 generative AI stuff and
5:17 and what it's what it's going to change
5:20 and what we can choose to have it change
5:23 for us personally and that's that's the
5:26 thing that
5:28 I'm I mean I suppose that's what this
5:31 Channel's always been about right choose
5:33 to step into it choose to if if you're
5:35 here and you haven't played with ch
5:37 Chachi BT one of the things I do is I I
5:40 leave the address of chat GPT up in my
5:44 live for most of the time although now
5:46 that we have cool images we also we also
5:49 put those up there um but the reason I
5:54 put that there is if you've not played
5:55 with chat GPT go there right now chat.
5:57 open.com and just start playing with it
6:00 ask it to write you a poem ask it to
6:01 write you a song ask it to write you a
6:03 marketing plan doesn't [ __ ] matter
6:05 anything a gardening
6:09 strategy
6:11 um and just see what it does and then go
6:14 to prompt. chat this is a little like a
6:16 Word document that's got a bunch of
6:18 prompts in it you can copy and paste but
6:19 it also has some primers that teach you
6:22 how these large language models work and
6:24 how to talk to them and and and and and
6:27 what you can do what they make possible
6:29 and just start on the journey and if
6:31 you're lucky within the first I don't
6:34 know 10 or 15 minutes or an
6:36 hour you'll have what we lovingly call
6:39 in here your Kevin mallister moment
6:41 right your moment from Home Alone where
6:43 you're
6:44 [Laughter]
6:46 like and that's what that's what you
6:49 want to get to because in having that
6:52 moment you'll realize holy crap chat GPT
6:57 is it's more than just hype
7:00 it's more than just fancy Google it
7:02 looks like a fancy Google search bar it
7:04 looks like a stupid Google search bar
7:07 doesn't look fancy at
7:09 all and yet when you use it you realize
7:12 there's some profound depth here I got a
7:15 good parlor trick for you you know
7:16 there's the meme going around on Tik Tok
7:18 all these wives asking their men how
7:20 often do you think about the Roman
7:21 Empire which by the way I don't think
7:24 about the Roman Empire but whatever a
7:26 lot of men do go to chat JP and ask it
7:29 about the Roman Empire ask it why men
7:32 might be so obsessed with the Roman
7:33 Empire maybe it knows um but just go
7:37 start playing with
7:38 it and and start to reveal to yourself
7:42 why people are so excited about this cuz
7:45 it's it's more than just hype this is
7:47 there's some profound [ __ ] going on
7:48 these things are called large language
7:50 models because they've been trained on
7:52 like lots and lots of language on the
7:54 public internet for the past many years
7:58 30 4050 50 60
8:01 years and so they're reflections of us
8:04 we are large language models right words
8:07 go in they process ideas come
8:11 out that's what these
8:14 are reflecting our Humanity amplifying
8:17 our
8:18 Humanity it's not just a search engine
8:20 it's not just copying and pasting and
8:21 plagiarizing which is one of the
8:23 misconceptions if you understand how
8:25 they actually work it's [ __ ]
8:26 mind-blowing It's Magic Arthur C Clark
8:30 any sufficiently advanced technology is
8:33 indistinguishable from Magic that's what
8:36 these are right champy champy just came
8:39 in he got excited about Arthur C Clark
8:42 all right and so men think about the
8:45 Roman Empire limits equivalent is that
8:47 we think about murder yeah I saw that
8:48 cuno uh I saw someone else posted on on
8:51 Tik Tok it was she said she said you
8:54 know either that were going to be
8:55 murdered or other people's murdered or
8:56 us doing the murdering it's always about
8:58 murder I thought that was awesome um by
9:01 the way if you if you see people in here
9:04 and you're like Kyle you you haven't
9:05 really said anything yet and there's all
9:07 these likes what the hell are these
9:09 people doing meet the Irregulars so the
9:12 Irregulars are people that just show up
9:13 here
9:15 regularly and uh they play games with
9:17 themselves they entertain themselves and
9:20 they some of them actually mod the
9:21 channel and they will support you if you
9:25 have questions and I can't answer it I
9:26 haven't gotten to it they'll do that
9:28 they're super awesome people
9:30 um and they're the ones that are getting
9:31 carpal tunnel by doing
9:33 this on the Tik Tok screen okay so if
9:38 you have questions about generative AI
9:40 pop them in the comments below I'll do
9:41 my best to answer them happy Friday
9:43 everybody welcome everyone and let's get
9:46 get DED let get started and and listen
9:49 sometimes sometimes in the AI learning
9:51 lab I act a little dum dum sometimes I
9:54 act a little dumb Dum sometimes I act
9:56 super smart sometimes you'll be very
9:58 intimidated I'll be all like
9:59 intellectual like like you'll be like oh
10:02 my God he's got qualifications and then
10:04 you realize H maybe not I wonder if he's
10:07 qualified is he a dum dum it's possible
10:09 he's a dumb Dum you just don't know
10:13 that's part of the joy of the lab you
10:16 never know what you're G to get we like
10:17 to call it chat
10:20 add why is that well you know hang out
10:24 we'll see he it's also there's a lot of
10:28 time spent talking about my
10:31 hair no no one else actually points it
10:34 out I'm apparently obsessed with
10:37 it but sometimes it just goes crazy and
10:40 then sometimes it looks like I'm a
10:41 carpet salesman it's just I don't know
10:44 it's just it's something that entertains
10:48 me all right let's go all right so what
10:50 I'm going to do I'm going to go back and
10:53 I'm going to start answering some
10:54 questions I'm going to see some comments
10:56 and I'm going to comment on the comments
10:59 this is is you know what this channel is
11:01 this is like one of those reaction
11:02 channels except reaction instead of me
11:05 reacting to people who've Got Talent I'm
11:08 reacting to people who are commenting
11:11 and then and then sometimes I'll call
11:12 them a dum dum and then sometimes I'll
11:15 be like oh my God that's a deep
11:16 philosophical question let me talk for
11:19 15 minutes that's how it goes you never
11:22 know what you're going to get here life
11:24 is like a box of chocolates you never
11:26 know what you're going to
11:28 get
11:30 go Forest
11:31 [Laughter]
11:34 go all right I think I have a coffee
11:37 being made hold I'll be right
11:41 back I smell
11:44 coffee
11:58 I
12:28 than
12:48 all right Champion let's do this let's
12:51 do this let's do this all right I have
12:55 my coffee did we lose everyone yep we
12:57 lost everyone 30 people left this is the
12:59 good stuff this is a small intimate
13:02 crowd this like this like seeing the
13:04 stones in a little corner bar well if
13:09 the stones were like a fat old
13:11 58-year-old talking about computer [ __ ]
13:14 it was just like
13:17 that very small audience very small
13:20 pinch your
13:23 head
13:25 M all right that is tasty freshly
13:28 brewed what is the best AI to translate
13:31 audio files to text what is the best AI
13:35 to translate audio files to text then I
13:38 want AI to
13:40 summarize the conversation and find key
13:42 points
13:43 okay um there's probably other tools for
13:46 this go to futur pedia.com
13:59 it a couple of ways you can download the
14:00 official open aai app like on the iOS or
14:04 Android
14:06 stores the official chat GPT app you
14:09 sort of turn on the microphone I don't
14:11 know if you can upload a file in the app
14:13 but you can talk into it and it does
14:15 really good transcription but if you
14:17 want to upload a file oh by the way
14:19 there's a new tool out let me show you
14:22 this um stable diffusion the company
14:26 behind um or stability AI the company
14:29 buying stable diffusion launched this
14:31 thing called stable
14:34 audio and you just describe music and
14:37 then it plays it so I just put in Moody
14:39 finger style guitar and it's not
14:44 [Music]
14:52 bad oh here comes champ he's a
14:57 singer you like that
15:00 [Music]
15:07 champy all right buddy it's done it's
15:10 good that's good it's a good boy you're
15:11 a good singing dog you're a good singing
15:18 doggy champy approves champy
15:21 approves okay so anyway that is go to
15:24 stable audio.com and you can play with
15:26 that it looks like their servers are
15:28 very overwhelming right now um but
15:30 what's interesting about this is they're
15:33 doing for music what stable diffusion
15:35 does for images so it's trained in in an
15:37 identical way they took 800,000 songs
15:40 they tagged them and then they turned
15:42 them into essentially mathematical
15:44 equations and sort
15:46 of um how do how do I put it like math
15:49 mathematical versions of song signatures
15:53 it's not really the songs it's song
15:55 signatures and then they um when you put
15:58 in your prompt it sort of amalgamates a
16:01 bunch of those signatures and then
16:02 effectively denises them goes from a
16:05 diffused model to a more focused model
16:07 over a bunch of steps and it generates
16:10 music so um so it's it's it's a really
16:14 interesting approach okay uh
16:17 text to
16:19 transcription so what we're going to do
16:21 is we're going to go
16:24 to you're going to go to
16:27 playground or plat form so write this
16:30 down because there's a pain in the ass
16:33 there's a pain in the ass URL thank
16:35 thank you
16:37 um um open AI all right so you ready you
16:41 got your pens get your pens get your
16:44 pens come on listen most of you in here
16:47 are are of an older Affliction and I
16:50 know that you need papers and pens and
16:53 reading glasses so go get go go get your
16:57 autra mon
16:59 um so you're going to go to platform.
17:03 open.com
17:05 playground and then basically this is
17:07 this is where developers develop tools
17:10 sitting on top of GPT 3.5 GPT 4 it's
17:14 going to default you into chat mode so
17:17 in the right hand side here here let me
17:18 make this bigger so you can see it oops
17:21 that was too big in the right hand side
17:23 here you've got mode and it's chat you
17:26 don't want to be in chat mode you want
17:27 to be in comp complete mode it says
17:30 complete Legacy which basically means
17:32 this is going to go away at some point
17:34 but then you have a little window here
17:36 and in this window you could type uh
17:38 anything you could say um
17:44 um why are men so obsessed with the
17:47 Roman Empire question mark it seems that
17:50 almost every man thinks about it on a
17:53 regular
17:55 basis and women are mystified why this
17:58 is happening happening period can you
18:00 shed any
18:06 light men are fascinated by the Roman
18:08 Empire because of its impressive
18:10 achievements in politics Warfare
18:12 engineering art literature and
18:14 architecture so so this works just like
18:16 chat gbt but you can do other things
18:18 like you can Muck with the temperature
18:20 blah blah blah blah blah but cool thing
18:22 here is this little teeny microphone in
18:25 the corner right here this is access to
18:28 chat gpt's whisper technology so you
18:31 click on that and then you can either
18:33 start recording or you can upload audio
18:36 from your
18:38 computer and
18:41 um let's see I wonder if I'm trying to
18:45 think if I've got an audio
18:53 file
18:57 MP4
19:00 not name MP4
19:03 MP4 EG movie this
19:10 [Music]
19:21 Mac me see let me see if I have an MP3
19:26 file MP3 audio
19:37 a steady
19:39 climb I don't think I have anything
19:42 here I didn't really prepare for
19:44 this oh yeah baby yeah
19:48 baby yeah
19:51 [Laughter]
19:53 baby you've got mail all right where's
19:57 the beef hey anyway um so basically so
20:01 just upload an audio file here it'll
20:04 turn it into a
20:06 transcription and then depending on the
20:08 length of the transcription you may need
20:10 to break it up into into little segments
20:12 um there's probably other tools that do
20:14 this but but once it's once you've got a
20:16 transcription in fact I probably do have
20:18 a transcription in here
20:21 somewhere let me go find a
20:25 transcription
20:27 transcription
20:35 all right here's an AI learning lab
20:37 transcription wait why did that
20:41 open oh there we go
20:43 okay so I'm just going to grab a a good
20:46 chunk of this actually you know what I'm
20:49 going to do
20:51 okay so this is a really long
20:54 transcription so this is one of my you
20:56 know 2 and 1 half hour rambling sessions
20:58 right so this is just words from me
21:00 talking now where I got this
21:03 transcription was I
21:05 uploaded my video from Tik Tok to
21:08 YouTube and then YouTube automatically
21:10 generated this so another way to get
21:12 this transcription is via YouTube but
21:15 the whisper API from chat GPT is
21:17 actually better so for for what it's
21:19 worth um okay so now we're going to go
21:22 to chat
21:25 GPT and I'm going to try to paste that
21:28 transcription into chat GPT and I'm
21:31 going to say
21:34 um let's see um
21:38 summarize
21:40 this and it's going to do a lot of that
21:44 it's going to read it and it's going to
21:45 go the message you submitted was too
21:47 long chat GPT has a um a limit of about
21:52 3,000 words and a 2 and 1/2 hour
21:55 transcript is probably more like 15 to
21:57 25,000 words so it doesn't work there
22:01 but we have options we can go to
22:04 claw.
22:06 a we can do the exact same thing I can
22:09 paste in that whole thing so I'm I'm
22:12 just in the chat window I paste in and
22:14 notice what it did was if the if what
22:17 you paste in is larger than a certain
22:18 amount certain amount it basically
22:21 attaches it as a text file and so now I
22:23 can say um
22:26 please summarize
22:31 this and tell me what is going on with
22:38 it and so it'll think for a little bit
22:41 here so so Claude this tool is from a
22:45 company called anthropic this is a bunch
22:47 of xop AI people so these are the people
22:50 that develop
22:51 gpt3 the the the model that sits
22:54 underneath chat GPT
22:56 3.5 these are the people that develop
22:58 Ved it they quit open Ai and started
23:01 anthropic um they're they're more
23:04 concerned about safety than than
23:06 creating um than creating models but
23:09 what they realized was they had to
23:10 create models in order to do the safety
23:12 stuff they wanted to do so the
23:14 transcript is a long video of Kyle
23:17 Shannon's AI learning lab on Tik Tok the
23:19 main points are Kyle talks
23:21 enthusiastically about AI especially
23:23 chat GPT and how it's changing
23:26 everything he believes AI have a
23:28 profound impact this is kind of
23:30 embarrassing but good he encourages
23:32 people to try out jet gbt um he gives
23:35 advice on how how to get started he
23:39 discusses potential applications in AI
23:42 then he discuss then there's a
23:44 discussion around jobs Kyle talks about
23:47 the pace of change ethics risks
23:49 regulation all of it let's see um
23:54 does um Kyle ever swear
24:01 let's see what it
24:06 says so anyway I mean this this tool is
24:09 quite remarkable at at being able to
24:11 this was the tool last night when we
24:13 were working with um with um Amelio's
24:16 wife with the PDF document that she she
24:19 shared with me about health insurance
24:22 plans and it actually it parsed that PDF
24:26 better than chat chpt did
24:29 yes based on a scan of the transcript
24:32 Kyle uses some profanity at time stamp
24:36 11629 he says Jesus
24:45 Martha oh man my fears they've got their
24:49 heads up their asses all
24:53 right when discussing insurance
24:56 companies he says but you
24:58 suck now to be fair when we're referring
25:02 to insurance companies but you suck is a
25:06 technical term that's not offensive
25:09 language that's just
25:11 fact so anyway that's how you do that
25:14 all right I hope it helps did it help
25:16 did it help I hope it helps I'm here to
25:18 help I'm just here to help help people
25:21 you know if I didn't help then that
25:23 would that would be so bad it would be
25:25 so embarrassing if I didn't help but if
25:27 I didn't help we'll blame Ally Ally did
25:29 it hair growth gate what's happening
25:32 what's shaking lady hope you're well all
25:38 right I'm 49 messages behind thank you
25:41 if anyone's modding just feel free to
25:43 pop up a comment now that I'm out of the
25:45 uh transcription hole thank you
25:49 steo um I had it make a plan for fall
25:52 plantings that will bloom after June oh
25:54 that's super cool actually that will
25:55 bloom after June so fall plantings that
25:58 will bloom after June so I assume that's
26:00 bulbs I don't know much about gardening
26:03 but I assume you're not putting in
26:06 delicate little flowers as we go into
26:08 the cold season June Mary I want to
26:10 create an evaluation methodology for
26:12 International Education
26:16 records That's Heavy um it's totally you
26:19 can do it playground AI is giving us
26:22 Graphics like never
26:26 before how hm
26:30 how
26:31 how
26:33 how um hair growth geek how are you
26:36 getting images from it are you having it
26:38 generate um prompts or using Dolly
26:46 here I didn't know you could
26:52 do I didn't know you could do images
26:54 here do share do share how you're using
26:57 playground to get
26:59 images that's pretty
27:02 cool uh and then and then we will share
27:04 oh by the way that's that's one of the
27:07 one of the I guess one of the purposes
27:10 of this channel
27:13 is I'm doing everything I can to share
27:15 what I'm learning in real time right
27:17 that like it to a great degree the AI
27:20 learning lab is is something for me to
27:22 be able to process all the crap I'm
27:24 learning and get it out of my head one
27:26 of the things I encourage is wait we can
27:28 customize and optimize using Google
27:30 image analyzer oh cool oh very
27:33 nice that's super cool um as you start
27:38 learning as you start playing with this
27:40 stuff even if you're like brand new to
27:41 it talk about it talk about it at work
27:45 talk about it with your family talk
27:47 about it with your friends even if they
27:49 think you're [ __ ]
27:51 weird it's important I just keep playing
27:54 with it and learning and then I get
27:55 better all text okay cool got it
27:58 yeah it's amazing there was Ethan mullik
28:00 did a a an interesting LinkedIn post I
28:03 don't know four days ago about a study
28:07 where
28:09 um one of the one of the research groups
28:13 compared human prompting and um chat gbt
28:18 prompting and and the quality of the
28:20 results and chat
28:22 gbt writing prompts for itself beat the
28:25 humans excuse me beat the humans so
28:27 that's
28:28 that's pretty cool and that's what
28:29 you're up to there I like it I dig it I
28:32 support it anyway as you learn this [ __ ]
28:34 share it and and there's a couple of
28:36 reasons for it one is just that thing
28:39 of you learn more when you teach than
28:44 then when you have someone tell you what
28:45 to do if you have to figure out how to
28:48 articulate an idea of something you
28:50 learned it helps you learn um so do that
28:53 here let me pull up hair growth
28:56 gey
28:59 if she's around I don't know if you're
29:01 if you're camera friendly I don't know
29:02 if you're camera ready right
29:04 now if you're not that's cool
29:08 um but but you know one of the things is
29:11 you'll learn better the other thing is
29:14 with with um with teaching is you'll
29:17 teach people like people in your life
29:19 even the ones that think you're really
29:21 weird hang on settings no that's not
29:24 what I wanted to do hold please
29:28 dual camera there we go bang all right
29:31 had to get my doggy
29:33 back I like doggies doggies is
29:37 good uh my family is planning an AI
29:40 intervention they think I don't know
29:42 about
29:44 it yeah be willing to be intervened by
29:48 your family be willing to be that weird
29:50 because because here's what's going to
29:52 happen some people you're going to teach
29:54 it to they're going to be really
29:55 appreciative they're going to want to
29:57 learn they're going to jum jump on the
29:58 bandwagon they're going be like woohoo
30:00 other people are going to be like
30:01 whatever I don't care about AI blah blah
30:03 and they're going to go away and then
30:04 other people are going to be resistant
30:05 to it and think you're weird and you
30:07 know try to diminish you and and put out
30:10 your spirit and a don't let them [ __ ]
30:13 do that sorry in PJs no no that's good
30:17 um but what's going to happen is
30:21 this the people that are resistant to it
30:23 right now are doing that either out of
30:25 fear or
30:26 ignorance and at some point when these
30:29 tools start
30:32 invading their
30:34 work and it becomes clear to them that
30:36 something's shifting and they don't know
30:39 what to do they're going to come to you
30:42 because they know that you're you're
30:44 someone in their life that knows about
30:46 this [ __ ] oh that's something she was
30:48 talking about at dinner last you know
30:50 six months ago maybe I should call her
30:53 can you help me learn this right so so I
30:56 think it's I just think it's an
30:57 important service I think I think the
31:00 more people that are prepared to at
31:01 least understand you know the nature of
31:06 what these tools are how profound they
31:09 are what they can do well what they
31:11 can't do well um and and really start to
31:15 start to to live in this idea of being
31:19 AI native that whenever you got a
31:22 problem to solve the first thing you
31:23 think of is how can I use AI to do it
31:26 not because it can necessar do it better
31:28 though it likely
31:30 can but because in you trying to figure
31:33 out how to use these tools to solve that
31:36 problem you'll do what we did last night
31:39 we tried chat GPT two different ways and
31:41 I couldn't get it to work and like I'm I
31:44 I think I'm pretty good at this [ __ ]
31:45 right and and we couldn't get it to work
31:47 and then we went over to Claud and got
31:49 it to work and and just just in doing
31:51 that the amount of time I spent on that
31:53 I learned a ton um I don't know if
31:56 anyone in here learned a ton cuz I was
31:58 kind of all over the place uh bing went
32:00 Bonkers today with pictures and videos
32:02 it took an hour to allow
32:05 searches I'm done yeah Nancy so here's
32:09 the deal with Bing
32:13 Bing
32:14 um Bing got kicked off my adults table
32:17 right my adult table now has chat gbt
32:20 Claude and Bard and if if you've been on
32:23 this channel for a while you know I made
32:25 fun of Bard hard for a bunch of months
32:28 and Bing is down on the kids
32:32 table with pie and Po and
32:35 perplexity perplexity actually probably
32:38 deserves to be up at the up up at the
32:39 adult table but Bing is definitely down
32:42 here so because yeah it's the the safety
32:46 guard rails on being some whatever I
32:48 don't know what's going on with
32:49 Microsoft but you know they're being
32:51 Microsoft the funny voices not only help
32:54 but do something I didn't know what that
32:56 said oh but heal my my soul that's
32:59 that's good Winston I'm glad I'm Soul
33:02 healing I will never contradict you
33:04 again that's cute uh you can customize
33:07 to
33:08 optimize wait we we can customize to
33:11 optimize wait what is going on with my
33:15 comments are just jumping all over the
33:16 place I learned a lot from the chat GPT
33:19 failure last night oh good Andre thank
33:20 you I'm glad thank you for saying that
33:22 like sometimes when I do stuff like that
33:24 I don't know if it was a complete waste
33:25 of time cuz I know it was it was long it
33:27 was
33:28 like someone was flipping through Tik
33:31 Tok it it wasn't as good as like honey
33:33 is this thing on you know that'll get
33:36 him to stop all right
33:41 um so thank you for that let's
33:49 see I don't know where I am am I toward
33:52 the bottom or toward the top let me see
33:55 oh very cool I love that one I I don't
33:57 know what that's called but that thing's
33:58 so cool thank you very much for that and
34:02 uh so there's that um just made it to
34:06 level 11 and it is a competition so so
34:09 in case you're in case you're keeping
34:11 track I don't know what level you're at
34:12 but if you're not at level
34:14 11 you know time to step up your game
34:18 apparently um I learned a lot from the
34:20 chat gbt failure the other night thank
34:22 you for that hair growth geek thank you
34:23 for the CR by any other name would smell
34:26 is sweet oh very lovely Winston I love
34:30 it it's the
34:31 cosmos the cosmos Beyond billions upon
34:35 billions of stars in the
34:38 sky oh very nice little puppy thank you
34:42 very nice oh hair growth geek level 11
34:46 see you got there all right not
34:49 necessarily bulbs many weren't oh very
34:51 cool I guess it could be seeds too right
34:54 um I had issues with words in Discord
34:58 I don't know what that means but if you
35:00 tell me a little bit more about that
35:01 I'll do what I can to um to do something
35:04 with
35:06 it
35:10 um AI
35:13 intervention listen listen we we need to
35:16 talk oh all my friends are here yeah uh
35:20 Kyle sit down thank you digital Gods I
35:23 appreciate that K life's tough it's
35:26 tougher when you're stupid
35:32 did you see the video of the dude
35:34 there's like some dude in New York and
35:36 he's doing the NPC thing he's like and a
35:39 hat and a hat thanks for the Hat thanks
35:40 for that he's doing all that stupid [ __ ]
35:42 on the street and some guy standing
35:44 behind him like in the camera view and
35:47 he turns around just loses his [ __ ] on
35:49 this guy he goes I'm trying to feed my
35:51 family it's pretty
35:55 funny oh good lord
35:58 shocking how few people are curious yes
36:01 one shy girl I totally agree I talk
36:04 about it daily and one out of 10
36:06 actually ask questions exactly so and
36:09 and and this is the thing I know right
36:11 now like sometimes it it feels like
36:13 you're you know you're shouting at the
36:15 ocean
36:16 right sometimes this channel feels like
36:18 that although it doesn't because I mean
36:20 the comments like like you guys are are
36:22 clearly engaged other than the Trolls
36:24 but they're fun to pick on um
36:30 that just means we're early it just
36:32 means we're early right this chat GPT is
36:35 9 months old you know November 30th 2022
36:38 is when that thing um came
36:41 out and so it's only been nine months
36:44 and and you know if you've if you've
36:47 been in the tech scene for a while you'd
36:48 be like but Kyle machine learning's been
36:50 around for many years I I won't even
36:52 argue decades yes it has that that date
36:57 the date that thing came out is the
36:59 equivalent of when the worldwide web
37:01 came out for the internet right the
37:03 internet was all command line Tim
37:05 burners Lee comes out with this thing
37:06 called the worldwide web and once we
37:08 have browsers now you can just point and
37:10 click on websites right you can click
37:13 here and jump to another website or jump
37:15 to another country that was the big
37:17 innovation it was actually an interface
37:19 thing that's exactly what that is so
37:22 there's an inflection point in history
37:24 and we're only nine months into it these
37:27 tools as profoundly powerful as they are
37:30 are we're just scratching the surface
37:33 with what they're able to do they're the
37:35 worst they're ever going to be but
37:37 they're brand brand new so what I what I
37:39 would say to you is just keep talking
37:41 you know keep keep putting it out there
37:44 and people will will opt in to your
37:47 Social Circle one shy girl um who are
37:52 into it and that's cool and then 6
37:55 months from now someone will come I'm
37:57 knocking on your door could you teach me
37:59 I know I made fun of you before I know
38:01 we had that fake intervention I was just
38:03 being a jerk I was scared could you help
38:06 me and then and then if they were too
38:09 much of a jerk you can choose whether
38:11 you help them or
38:12 not like well you were douchy MCD
38:15 dington so I'm going to go on ahead and
38:16 choose to let you suffer or you could be
38:19 empathetic and say okay it was fear I'll
38:22 help
38:23 him it's funny when my my um in my
38:27 company
38:28 storyvine our our VP of account she's
38:31 she's brilliant she's she's awesome
38:34 she's sometimes in here um whenever she
38:37 she'll she'll sometimes come to me and
38:39 says hey this this is going on with a
38:41 client and and you know they want a
38:43 favor right the client wants a favor and
38:45 my first question to her is are they
38:48 [ __ ] cuz it it plays a part if
38:52 you're an [ __ ] uh you don't get the
38:55 you don't get the uh the warm and
38:57 fuzzies as much as if you're not an
38:59 [ __ ] so it's just you know life is
39:03 actually not that hard if you just
39:05 follow this one rule don't be a douche
39:08 it's it's actually simple it's actually
39:11 simple everyone's going through [ __ ]
39:13 just don't be a douche just try to have
39:15 some empathy just try to do it no no all
39:19 right or just be a
39:21 douche all
39:22 right uh Winston Winston Friday night
39:26 gift night
39:28 hello Emilio's wife good to see
39:31 you did did you see amelos did you see
39:34 the video that I
39:36 posted I acknowledged what we did for
39:39 you last night and a conversation I had
39:41 today for the idea in that video so you
39:44 got a shout out in in a in a video on AI
39:46 learning lab channel it's very cool is
39:49 there a recommended wait is there a
39:51 recommended text on
39:54 gpt3 I don't know what that means ask
39:56 that again I don't I don't quite know
39:58 what you mean there um but ask it again
40:01 I did not W yeah go check it out you got
40:03 a little shout
40:04 out um I don't know if I actually
40:07 shouted out your name I think I did I
40:09 don't know I might not have but I said
40:11 we were we were talking about your uh
40:13 the health insurance
40:16 plans I think I said Amelio's wife
40:19 though uh what's up talk hey Kyle hello
40:24 hello to-do list Shout at the ocean
40:27 exactly
40:29 AIML came out in
40:32 95
40:33 yep yeah and and you know like like this
40:36 stuff's been around for a while and
40:38 similar to the the you know the internet
40:40 had been around for decades before the
40:42 worldwide web came out this is these
40:45 tools generative AI all this stuff
40:49 is
40:53 um democratizing that power to the rest
40:57 rest of the world to the rest
40:59 of humanity you don't need to be a
41:02 mathematician douchy
41:05 McDon book is there a book Zippy to pin
41:09 head by the way Zippy to pin head 10 out
41:14 of 10 on the name I would appreciate you
41:17 camel casing the words so it's easier to
41:19 parse but it's solid if you'd have done
41:23 Zippy the pin head it would have been
41:25 8.5 n but Zippy duin head 10 out of 10
41:29 solid like
41:31 it telegram upon sending 21 I do not
41:35 know what that means but that's probably
41:37 just cuz I'm old doesn't allow
41:39 screenshot
41:40 oh telegram upon sending if you're
41:43 trying to screenshot on a Samsung 212 or
41:46 3 it doesn't allow it because Ticky talk
41:50 decided to say screw you to
41:53 screenshotting I don't know why I don't
41:55 care I don't I listen I have stopped
41:57 trying to figure out Tik Tok it's true
41:59 people are starting to ask questions I
42:01 have helped with a I had oh I I have you
42:05 you helped with a business plan today
42:06 Obi-Wan oh okay by the way
42:09 Obi-Wan so it's not Obi-Wan it's OB Wan
42:14 Juan 10 out of 10 on the
42:17 name oh man you you almost break a 10
42:21 that's a great name I love it it's
42:25 awesome I need to figure out a good
42:27 assignment in my stats class that still
42:30 aligns with my course
42:33 outcomes I need to figure out a good
42:35 assignment in my stats class so brother
42:39 52
42:41 um I
42:44 mean so so so going back to this idea
42:47 that that I hit on today of of or that I
42:51 kind of internalized today someone else
42:53 hit on it cuz they're teaching a [ __ ]
42:54 class on it but this idea of being um AI
42:59 native what I would say um is wait where
43:05 did uh
43:07 uh oh digital Gods finally got the shove
43:10 the robot stuff right in ideogram cool I
43:13 think you posted that in Discord
43:15 too
43:18 um Tik Tok just jumped me somewhere oh
43:20 here we go brother 52 okay I need to
43:23 find out a good assignment in my stats
43:25 class so so what would say is this I
43:27 would go to go to chat GPT right now or
43:30 whatever Claude And whatever the the
43:32 large language model of your
43:34 choice tell it about your stats
43:37 class tell it about your course
43:41 outcomes and then have it give you 10
43:44 ideas for a good
43:46 assignment
43:48 that you know with with some criteria
43:51 like you you define what good is like
43:54 like come up with 10 assignments that
43:55 accomplish you you know the following
43:57 goals or that you know provide the
43:58 following challenges or make it
44:00 interesting whatever it is have it give
44:02 you 10 and then when it gives you 10
44:05 ideas pick the top two and say I like
44:08 three and seven you know give me 20 more
44:11 that are more like those and and just
44:14 see if you can get chat GPT to give you
44:16 all that and this is this is the thing
44:20 this is my new thing the the video that
44:22 I just posted today this afternoon is
44:25 the challenge
44:27 is rather than only going to chat GPT
44:30 when it's obvious to go to chat GPT
44:33 whenever you have a problem go to chat
44:36 GPT or whatever mid Journey or you know
44:38 whatever the tool is that go to some AI
44:41 tool um spend the time no matter how
44:44 ridiculous it seems trying to solve it
44:47 with
44:47 AI because you'll probably be able to do
44:50 it for one thing and then if if if not
44:53 then you just learn something so anyway
44:55 I I that that I'm really excited that's
44:57 my that's my new ex I'm I'm I'm I'm all
44:59 excited I'm all excited I'm all excited
45:02 can you tell that I'm excited I'm all
45:05 excited I really want to have it
45:07 summarize a video transcript but worried
45:09 that's too complicated brother 52 I just
45:11 did it I I just popped in a 2hour 2 and
45:13 a half hour transcript into Claude do it
45:16 in Claud it's not too
45:18 complicated like literally copy and
45:21 paste the transcript copy the transcript
45:23 out of whatever thing you've got it in
45:25 paste it into Claude claw.
45:28 A and tell it to summarize it and then
45:30 start asking it questions about it it's
45:33 it's remarkable it's
45:36 remarkable um oh do I recommend any book
45:40 on gpt3 language no so Zippy to pin pin
45:43 head here's the deal okay we're gonna
45:48 we're gonna we're going to go we're
45:49 going to go a little smarmy for a
45:51 moment so all of these tools right now
45:55 all of the AI tools are what I lovingly
45:57 refer to as janky pieces of [ __ ]
46:01 including Chachi BT not because the
46:03 companies are bad not because the
46:05 technology is actually a piece of [ __ ]
46:07 it's just that it's janky why is it
46:08 janky because it's all
46:11 new generative AI the the ability to
46:14 interact with chat GPT there are likely
46:16 lots of books out there about
46:19 prompting most of them were probably
46:21 written with chat
46:23 GPT or GPT 3.5 or GPT before um no one
46:28 is an expert right now these tools are
46:31 changing incredibly fast in fact um I
46:36 heard last night on the live here
46:37 someone someone said to me
46:42 um the Google's new
46:46 Gemini model has been released to select
46:50 customers that's their new replacement
46:52 for for Palm 2 the stuff sitting
46:54 underneath bard um which it's actually
46:57 pretty good right now but apparently
46:59 Gemini is quite remarkable I haven't I
47:01 haven't played with it I haven't touched
47:03 it but whatever book you would have read
47:07 on chat GPT when Gemini comes out if
47:10 it's multimodal and if it's better and
47:12 if it does certain things better and
47:13 maybe it uses a different kind of you
47:16 know natural language maybe it likes uh
47:20 structure or it doesn't like structure I
47:22 like we don't know but when it comes out
47:24 whatever you learn for chat GP
47:26 is is likely going to be not as relevant
47:29 all of a sudden so so my advice right
47:34 now is if you if you can find you know a
47:38 channel like the AI learning lab or
47:40 Rachel Woods or McKenzie Bose or you
47:43 know you find some Tik tocker or some
47:44 YouTube channel where the voice of the
47:46 person doesn't bug you I don't know why
47:49 I would bug you why don't you listen to
47:51 my channel it's not fair you know you
47:55 got someone like that you can't listen
47:56 to them or maybe you love that find
47:59 someone you like and and you know you
48:00 can start doing it that way um the the
48:03 thing about watching you know Tik Tok or
48:05 YouTube channels where they're
48:06 constantly putting out new content is is
48:09 unlike a class or a book where you sort
48:11 of have to you know put a stamp in
48:15 time um they're going to be updating
48:18 things the the the thing that I think is
48:20 is the the better approach though is get
48:23 curious yourself start start
48:26 understanding how prompting Works start
48:29 understanding the difference between you
48:31 know a Google prompt which you know a
48:34 Google prompt Google has sort of trained
48:36 us that you put in sort of the minimal
48:38 amount of words because you if you over
48:40 complicate a Google prompt it tends to
48:42 just give you really shitty results
48:44 these things are kind of the opposite if
48:46 you give it too sparse a prompt it
48:48 doesn't really have enough context and
48:50 the way they work it just gives you
48:52 generic answers and it's just it's just
48:54 bad but if you give it lots and lots of
48:56 context if you start to have a
48:58 conversation with this thing it starts
48:59 to get better and better and better you
49:02 can learn that from a book but the book
49:04 is going to be so out of date that I
49:06 just I I would you know um uh shy girl
49:10 or I forget what your name actually is
49:13 the what your handle is but shy girl was
49:15 talking about you know she's you know
49:17 one out of 10 people she talks to even
49:20 asks a question and probably you know
49:23 one out of five of those is actually
49:24 truly engaged
49:26 um but the one out of five that's truly
49:29 engaged now she's got a community like
49:31 start a community of people that are
49:33 curious about this stuff and exploring
49:34 it join the AI salon so this is a group
49:37 that I started in December of last year
49:41 um so if you go to the salon. there's a
49:44 there's there's a bunch of links there
49:46 it's a link tree the the first one talks
49:48 about the salon and what our values are
49:50 the second one is the Meetup we have
49:52 bi-weekly meetups and then we've got a
49:55 Discord Channel Channel where you can go
49:56 in and um ask questions you can play in
49:59 mid Journey things like that I would
50:01 just start you know just sort of dive in
50:05 um get curious learn on your own because
50:09 the stuff's moving too fast to feel what
50:12 I think the risk is of taking a class or
50:14 reading a book is that you'll think
50:16 you've learned enough you you'll you'll
50:19 think that the actual answers are in the
50:22 book or in the class they're not the
50:24 only thing that's in the book or in the
50:26 class is sort of a step up to get you
50:30 curious enough to start learning more
50:33 because it's going to be the next two
50:34 years is just going to be you're going
50:35 to have to learn this is this is not a
50:38 low energy next two years by the way you
50:41 can sit back and not participate and let
50:44 this stuff just blow through the blow
50:45 through the economy but if you want to
50:47 participate in this stuff you're going
50:49 to have to be fairly
50:51 active all
50:54 right cut and paste a syllabus yes do
50:57 that as I'm experiment oh yeah uh yeah
51:00 cut and paste a syllabus for um for that
51:03 stats class absolutely pop that in there
51:06 pop in yeah have it yeah exactly good as
51:10 I'm experimenting I'm finding that chat
51:11 GPT is giving me better info than barter
51:14 Bing yes so gp4 in particular is so so
51:21 so to put it in context
51:23 um Bard and GPT
51:27 3.5 are based on um um 175 billion
51:33 parameter models it doesn't really
51:35 matter what that means but GPT
51:38 4 is actually based on eight different
51:42 200 billion parameter models so it's
51:46 however much that is 1.6 trillion
51:49 parameters um so gp4 is markedly better
51:53 now the gp4 at open.com is very very
51:59 different or the the chat
52:00 gp4 at
52:02 open.com is very very different than the
52:06 chat GPT at Bing chat so both of them
52:10 are based on the gp4 foundation model
52:13 but however they coded this app and
52:15 whatever safety guard rails they put
52:17 into it have have essentially made it
52:20 functionally unusable I think Bing chat
52:22 at this point is useless so I'm just the
52:25 more I'm using it and the more it keeps
52:27 telling me well I'm not even going to
52:29 answer that you're just too nasty you're
52:32 just too nasty I'm good enough I'm smart
52:35 enough and dog on it people like
52:38 me it it I mean it it it honest to God
52:41 Bing is like the steuart SMY of of the
52:44 chat GPT
52:50 models oh well you don't need to BU so
52:53 offensive well
52:57 so yeah yeah and and you know what what
53:01 you're hitting on and also
53:06 is different projects like different
53:09 things you try to get it to do like like
53:11 like you'll get chat GPT you'll think
53:13 you'll have it totally dialed in and
53:15 then you'll have some new kind of
53:16 project and you're like oh great I'm
53:17 going to use chat GPT and then it'll
53:19 just fail
53:21 miserably why I don't
53:24 know so go try Claude or go try Bard or
53:28 go try whatever so anyway so yes I do
53:31 not um couldn't custom instructions be
53:35 used for the stats class question it
53:38 could Firefly although custom
53:40 instructions the idea with custom
53:42 instructions is to to sort of set your
53:45 chat GPT up to to be an ongoing thing if
53:48 if you wanted to basically if you wanted
53:51 to use chat
53:52 GPT as kind of a stats teaching assist
53:56 istant actually that's kind of cool
53:58 let's go do that let's go play with this
54:00 okay that's a really good
54:03 idea it's a good idea Firefly that
54:06 Firefly very clever it's very clever
54:09 very clever all right uh chat chat chat
54:14 GTP anyone else have a pet peeve when
54:17 people say chat
54:18 GTP come on it's it's right there I know
54:21 the letters rhyme but it you're reading
54:25 it
54:26 the the P's before the T all
54:29 right he seems a little bitter I don't
54:32 know why why is he so testy about people
54:34 making mistakes sometimes people say
54:36 chat gbt they didn't mean it yeah okay
54:40 listen listen listen people okay so one
54:45 of the things that you may or may not
54:48 know for both GPT 3.5 and GPT 4 for all
54:52 of them you have a thing called custom
54:54 instructions and with custom
54:58 instructions you can tell chat GPT to
55:02 act like something so so this is my joke
55:04 custom instructions but we're going to
55:06 replace this but but I'll show you how
55:08 they work so my joke custom instructions
55:11 are you're an overbearing pompous ass
55:13 who uses large words that even you don't
55:15 understand as a way to intimidate those
55:17 that you talk to and then how would you
55:19 like chat gbt to respond whenever anyone
55:22 asks you any anything I want you to use
55:24 an obscure literary ref reference to
55:25 belittle their idea and then redirect
55:28 the conversation to talk about yourself
55:30 it sounds like a nice guy right and so
55:33 now we can go in here and we can say um
55:37 i'
55:39 I'd like hang it like advice for my
55:45 stats
55:50 class all statistics a
55:54 subject all often misconstrued by the
55:56 uninitiated it reminds me of Mark
55:58 Twain's humorous observation facts are
56:01 stubborn things but statistics are
56:03 pliable now allow me to indulge you with
56:06 a profound Insight in my extensive
56:09 intellectual repertoire statistics are
56:12 but a mere footnote nevertheless I shall
56:14 Grace you with my wisdom right so so the
56:17 custom instructions basically turn
56:19 Chachi BT into whatever the hell you
56:21 want it to be so I'm going to what I'm
56:24 going to do is I'm going to turn
56:27 off the pompus ass for a
56:30 moment and I'm just going to use GPT
56:33 3.5 and I'm going to say
56:43 um I want you to describe a very
56:49 helpful statistics
56:53 professor in a way that gives gives
56:56 their
56:57 background their teaching
57:02 philosophy and their areas of
57:14 expertise all right I think that's too
57:16 many words so I'll
57:20 say that's great but the professor
57:24 doesn't need a name and can
57:29 you consolidate this description into
57:32 two or three paragraphs so that it's
57:35 more
57:40 concise all right so statistics
57:44 Professor holds a PhD in statistics a
57:46 master's in Applied Mathematics their
57:48 research particularly in basian
57:50 statistics and hierarchical modeling is
57:52 widely recognized in the field with a
57:54 deep understanding of mathematical
57:56 foundations blah blah blah blah blah
57:57 blah blah blah okay perfect so now we're
57:59 going to
58:08 say now I'd like this
58:17 professor to be very helpful whenever
58:20 anyone asks a question so can you
58:24 describe
58:25 how this professor should
58:29 respond in the different ways people
58:33 might ask them questions about
58:38 statistics and this will all make sense
58:40 in a second I I assume you you know
58:42 where I'm going with
58:43 this so what did it do here um certainly
58:46 here's when students seek Clarity on
58:49 statistical Concepts okay responds with
58:52 patience and Clarity when they apply
58:55 okay so I don't I this might be too if
58:57 if they're too long we we'll have to uh
59:00 concise it down
59:03 again
59:13 [Music]
59:15 um okay so now we're going to go into
59:18 custom
59:20 instructions I'm going to turn it on and
59:22 I'm going to replace you're an
59:24 overbearing pomp
59:25 s with the statistics thing and then I'm
59:29 going to save
59:30 it and then I'm going to go grab
59:36 the how to respond
59:46 stuff come on come on copy paste don't
59:50 be a dumb
59:51 dumb there we
59:53 go da
59:58 da paste okay I have too many I thought
1:00:02 that might be the
1:00:03 case so we're going to go save we're
1:00:05 going to come back out here
1:00:08 um um again please
1:00:13 consolidate the
1:00:20 instructions what's going
1:00:23 on into two to three
1:00:29 paragraphs so let's see known for
1:00:32 unwavering
1:00:33 commitment
1:00:39 uh
1:00:50 okay
1:00:51 um rewrite the instructions to refer to
1:00:55 the professor as
1:00:59 you you are a
1:01:05 renowned okay when faced with conceptual
1:01:08 questions you excel at simplifying
1:01:09 complex ideas and relatable concepts for
1:01:12 practice okay so
1:01:14 basically why why I went through that
1:01:18 long and really boring exercise I'm
1:01:21 sorry I wasn't talking enough about what
1:01:23 the hell I was doing there
1:01:25 so if if it wasn't obvious I had chat
1:01:28 GPT describe a statistics
1:01:31 professor and then I had chat BT write
1:01:34 us how that Professor would respond
1:01:37 anytime someone asked questions so now
1:01:39 we've got a chat GPT that's been trained
1:01:43 up to be that
1:01:47 so now I'm going to go to gp4 this is
1:01:50 going to be a bit slower but but let's
1:01:52 see um
1:01:59 my name's Kyle and while I'm curious
1:02:02 about statistics I have a degree in
1:02:04 acting so you can make some assumptions
1:02:07 about my mathematical skills and
1:02:10 competency
1:02:12 period I'd like to understand where I
1:02:15 should even
1:02:16 begin understanding
1:02:20 statistics
1:02:23 for a second year graduate
1:02:31 program now let's see how it does hey
1:02:33 Kyle great to hear you curious here's
1:02:35 where to start basic terminology con uh
1:02:37 conceptual understanding simple
1:02:39 visualization practical
1:02:41 steps online courses inclusivity and
1:02:44 open
1:02:46 dialogue by starting with these
1:02:47 foundational aspects okay so so let's
1:02:51 say
1:02:53 um
1:02:55 great can
1:02:58 you teach me some of the
1:03:02 terminology by creating a back and forth
1:03:08 game where you ask me questions and I
1:03:12 answer period don't move on to the next
1:03:15 question until I've successfully
1:03:18 answered the
1:03:22 previous absolutely Kyle let's make this
1:03:25 fun and interactive question one what do
1:03:28 you think the term mean refers
1:03:31 to well like in the movie Mean Girls
1:03:35 Mean is how they act to other people as
1:03:38 in they're Mean
1:03:42 [Laughter]
1:03:45 Girls haha I appreciate the movie
1:03:49 [Laughter]
1:03:52 reference um I was just joking I think
1:03:57 mean means the average is that
1:04:01 right you got it Kyle the mean and
1:04:05 statistics is indeed the average of a
1:04:07 set of numbers you find this by summing
1:04:09 up the numbers in the data set and
1:04:10 dividing by the number of items so
1:04:12 anyway so so I now have a chat GPT that
1:04:16 that I could probably you know and and
1:04:19 you know if you know anything about
1:04:20 statistics you could obviously you know
1:04:22 make make this much more specific much
1:04:24 more interesting so anyway yeah that's a
1:04:26 really good idea Firefly thank you for
1:04:28 that that was
1:04:29 cool okay sis I'm a little winded I'm a
1:04:34 little tired I'm a little tired what'
1:04:36 you do Friday night I don't know I
1:04:38 listen to some old guy talk about janky
1:04:40 pieces of [ __ ] and he turned chat GPT
1:04:43 into a statistics Professor what have
1:04:45 you done with your life I don't know
1:04:47 it's not good it's not pretty it's not
1:04:50 pretty at all oh yeah oh yeah y'all see
1:04:53 this from last night night this is a
1:04:56 good
1:04:58 [Music]
1:05:01 one the nasty lady the nasty lady nasty
1:05:05 she nasty Sam Corn's in the house what's
1:05:08 happening sir good to see you good to
1:05:11 see you good to see you all right let's
1:05:15 see Tik Tock I lost my place all right
1:05:18 thanks for the stats demo need to take
1:05:20 the class need to take the class
1:05:23 undergrad degree yeah I know if I wanted
1:05:26 to learn anything I mean my problem
1:05:28 about statistics is I just I got no
1:05:30 interest in
1:05:33 it although with all this AI [ __ ] I
1:05:35 probably should because it's pretty good
1:05:37 at that stuff I should probably get
1:05:40 actually actually that might not be a
1:05:42 bad idea I you know this idea that you
1:05:44 can you can basically take any subject
1:05:47 and get up to speed on it instantly and
1:05:50 then get you know decently competent
1:05:52 with it in a very short amount of time
1:05:55 there's something I don't know something
1:05:57 pretty amazing about
1:06:00 that all right I find that Bard and Bing
1:06:03 give me summaries of the sites they
1:06:05 search yep oh oh yeah that's interesting
1:06:08 I dm'd you I'm an IR regular and still
1:06:10 not a mod what's up talk let me go do
1:06:14 that
1:06:15 now where do I do it
1:06:18 settings
1:06:20 mods Plus what's cuz I know I followed
1:06:23 you what
1:06:25 why am why is it not finding
1:06:28 you
1:06:30 huh well well well a that special
1:06:34 well it appears we've got a little
1:06:37 problem with Satan doesn't it a that
1:06:41 [Music]
1:06:42 terrific what's up talk Ed okay confirm
1:06:47 you are now a mod thank
1:06:51 you all
1:06:52 right I don't know why I couldn't search
1:06:55 for the word what in what's up talk but
1:06:58 you know what are you going to do agree
1:07:01 not to be on the bleeding edge after 40
1:07:04 years and retired
1:07:07 yep good thoughts cart before the horse
1:07:10 seems to be a thing nowadays in it
1:07:13 absolutely Winston I just shared the
1:07:16 live with someone by accident when I
1:07:18 picked up my IPA I didn't notice
1:07:23 who
1:07:25 that's awesome they'll be like why did
1:07:27 you send me this old guy like rambling
1:07:30 on about Winston are you
1:07:35 okay Winston needs an
1:07:39 intervention I think maybe that should
1:07:41 be all of our goals is to is to have you
1:07:43 know our friends and loved ones give us
1:07:46 interventions Kyle we need to talk about
1:07:47 the AI okay listen it's getting serious
1:07:51 listen
1:07:53 um I'll speak on behalf of all of
1:07:57 us know that we all love you we listen
1:08:00 Kyle you're just you're
1:08:02 amazing we've known you for years as you
1:08:04 know
1:08:05 I oh I'm getting choked up oh
1:08:09 my anyway um about this AI stuff it's
1:08:14 listen we know you're passionate
1:08:18 um the Tik Tok Channel
1:08:23 um it it's cute that you you you think
1:08:27 you're
1:08:28 youthful yeah exactly we should all have
1:08:30 one of those interventions in our life I
1:08:32 I like
1:08:33 it all right and major gift giving major
1:08:37 gift
1:08:39 fundraising that's a that's a that's a
1:08:41 lot of words in that name and it's it's
1:08:43 still I appreciate it it says what you
1:08:45 do and I don't think you should change
1:08:47 it but for this tongue it's not so good
1:08:50 uh just interested in new new tech
1:08:52 always
1:08:53 curious well this is Tech to be it's
1:08:56 good to be curious about this
1:08:58 Tech because it is not going away and it
1:09:01 is going to affect oh you know
1:09:06 everything hair growth geek thank you
1:09:09 for the lovely
1:09:10 heart I like to wear the heart on my
1:09:13 sleeve
1:09:16 huh that's facts all right I think
1:09:20 prompting will change as the models
1:09:21 improve as well I'm a librarian so yeah
1:09:26 yes I think that learning prompting
1:09:29 right now is kind of the equivalent
1:09:34 of learning to do math lonand knowing
1:09:37 that calculators exist now the
1:09:41 calculators don't quite exist yet
1:09:44 right two years from now I don't think
1:09:47 we're going to have to do much prompting
1:09:49 we're just going to sort of say what we
1:09:51 want and it's going to go figure all
1:09:52 that [ __ ] out it's going to prompt it
1:09:54 itself they're going to be autonomous
1:09:56 agents so I don't think long term that
1:09:59 you need to understand prompting that
1:10:01 said if you're in the development side
1:10:03 of AI I think it's going to be an
1:10:04 ongoing
1:10:06 requirement but even if you're not as
1:10:09 these tools get more
1:10:11 sophisticated if you've taken the time
1:10:13 to learn prompting right now you're
1:10:16 going to be much better at using those
1:10:18 tools when they get more sophisticated
1:10:20 because you're going to understand how
1:10:21 they work and how they do what they do
1:10:23 and how it's different from from how it
1:10:24 used to be all that sort of stuff little
1:10:27 green dude I like that started following
1:10:30 me thank you
1:10:32 um it's funny my my friend David LaVine
1:10:36 from back in my New York days he had a
1:10:37 band called little green and he and he
1:10:40 always said dude so little green dude
1:10:43 you've got Yoda there but it it reminds
1:10:45 me of David LaVine and his band little
1:10:47 green little green little green little
1:10:50 green little green so yeah I agree with
1:10:52 you I'm a librarian I think I think
1:10:53 prompt is going to evolve pretty
1:10:55 dramatically over the next two years
1:10:57 watching Kyle here will improve
1:10:59 prompting 100% yep I hear him in my head
1:11:02 as I type my prompts I'm so sorry Ted
1:11:05 but but good I think I think that's a
1:11:07 good thing volunteered on Alice back in
1:11:10 the day oh wow very very cool that's
1:11:13 super cool yeah it's it's it's it's
1:11:15 really cool I mean you know I mean
1:11:19 listen there there are advantages to
1:11:21 being young like you know would it would
1:11:23 it be good to be young again it would be
1:11:25 good to be young again you know with
1:11:27 with this much life
1:11:28 experience um but I think I think anyone
1:11:32 who went through anything in the early
1:11:34 days of the web and and you know having
1:11:38 experience starting in the early days of
1:11:40 the web and kind of writing that thing
1:11:42 up and and not thinking there'd be
1:11:44 another one of those in my lifetime and
1:11:46 knowing that we're at the beginning of
1:11:48 this one like I can't tell you like the
1:11:51 level of EX well no I can you obviously
1:11:53 can tell tell I'm a little excited about
1:11:55 this [ __ ] but it's
1:11:56 like you know people sometimes say to me
1:11:59 like how do you have so much
1:12:02 energy it's because like I Don't Want to
1:12:05 Miss a [ __ ] moment of this cuz this
1:12:07 here's the deal I mean this may be
1:12:11 self-evident you don't get to be at the
1:12:13 beginning of something this
1:12:17 transformative ever
1:12:21 again like like I was in in my in my um
1:12:25 AI office hours on LinkedIn
1:12:27 today Jim Ross was in there who's
1:12:30 sometimes in the
1:12:31 lives and um I said and and he he um he
1:12:36 owns and runs um uh storage
1:12:40 facilities and I said you know what are
1:12:42 you doing with AI and he said oh I
1:12:44 started back in November well [ __ ] he
1:12:46 started at the beginning right and even
1:12:49 if you're just getting in now you're n
1:12:51 like we're nine months into this thing
1:12:53 this is the beginning like you you you
1:12:57 you don't get to come back here you
1:12:59 don't get to say I knew what it was like
1:13:02 before these tools got
1:13:04 good I knew what it was like to
1:13:07 understand how they
1:13:10 worked
1:13:13 so yeah it's an amazing time it's a dang
1:13:16 amazing time it's a dang amazing time
1:13:19 dang it you know what I'm saying you
1:13:21 know what I'm saying do you know what
1:13:23 I'm saying
1:13:25 listen listen we got to talk about this
1:13:27 AI stuff Kyle it's just not right it's
1:13:31 just not right you know we go down to
1:13:34 the VFW you ain't been hunting forever I
1:13:37 do you even fish anymore do you own a
1:13:39 fishing pole do you own a fishing pole
1:13:42 or is it all like I'm AI
1:13:44 now you can't catch a fish with AI can
1:13:47 you oh I suppose you're going to tell us
1:13:49 you could do that
1:13:51 too good Lord
1:13:54 you want a
1:13:56 paps you want a blue ribbon no yeah I
1:13:59 figured not because you're probably like
1:14:02 oh I got to have my wits about me to
1:14:03 learn about the
1:14:07 AI well anyway we're thinking about you
1:14:09 come on down to the bar if you want to
1:14:11 you know I'll buy a drink or something
1:14:12 maybe if you know the paycheck clears
1:14:14 you know it's rough it's rough but
1:14:17 anyway I don't know I I ain't learning
1:14:20 this AI stuff I mean them [ __ ] robots
1:14:23 going to they're going to take us out
1:14:25 you know they're going to take us out
1:14:27 and uh I don't know what you're doing
1:14:29 there but uh but you always been a
1:14:32 little you always been a little
1:14:33 different you know you know I mean I
1:14:35 like you but you know the the guys are
1:14:37 talking you're just not the
1:14:40 same so anyway that's how I grew
1:14:46 up Gemini will believe five times GPT
1:14:49 for it launched 20 times in a
1:14:52 year yeah it but is that coming from is
1:14:55 that coming from Google or is that
1:14:56 coming from people that are using it cuz
1:14:58 I don't know what five times GPT for at
1:15:00 launch is cuz gp4 is 1.8 trillion
1:15:03 parameters it's eight different models
1:15:06 they're calling it a mix of experts
1:15:08 model like is it multimodal like I don't
1:15:11 know enough about Gemini I want to know
1:15:13 about Gemini tell me about Gemini
1:15:14 anybody know about G Gemini Pate are you
1:15:16 in here pate's been in here pate's doing
1:15:20 some some
1:15:22 uh some stuff at Google he's got one of
1:15:25 them their uh what do they call them
1:15:30 phds I called him a mathematician he's
1:15:32 like I'm not a mathematician I'm like uh
1:15:35 you have a PhD and you work at Google
1:15:37 I'm I'm thinking you might be a little
1:15:39 better at math than
1:15:42 me but anyway I'm sure Pate can't talk
1:15:44 about it eventually they may have to
1:15:46 standardize prompts or part of promps to
1:15:48 get results I.E GPT bullion you know
1:15:51 what's interesting I'm a librarian is
1:15:53 that
1:15:54 um I've heard a number of people now um
1:15:57 Rachel Woods talks about this in in uh
1:16:00 in her Channel um where adding structure
1:16:04 into prompts like like um if you if you
1:16:08 say hey based on this article and then
1:16:10 you paste in an article if you put like
1:16:13 a text like an HTML bracket that says
1:16:16 text at the at the top and then a closed
1:16:19 bracket at the bottom like an HTML Clos
1:16:21 tag um apparently
1:16:25 it um these tools act better because it
1:16:28 really clearly demarcates your the
1:16:31 different sections of your text so
1:16:33 apparently adding structure to um chat
1:16:37 GPT prompts is effective so that might
1:16:39 be something worth experimenting with
1:16:42 guilty King
1:16:44 48 no way I was just about to use it I
1:16:46 don't know what it is but that's super
1:16:48 cool glad to hear it's not just me with
1:16:50 Bing it says here's a website I know
1:16:53 Bing is just
1:16:57 Bing is a janky piece of
1:17:02 [ __ ] they got to get their stuff
1:17:04 together gp4 also has specialized
1:17:06 plugins to customize your project search
1:17:08 yep chat GPT does
1:17:10 not yep well chat GPT 4 does but
1:17:15 yeah um fun yes guilty King if you want
1:17:19 to upload your own image as a base to
1:17:22 start with which which is a good one to
1:17:24 use
1:17:27 oh you can do it in mid
1:17:30 Journey um it's a little
1:17:33 weird you can do it in Leonardo you can
1:17:37 use a base image in Leonardo so Leonardo
1:17:39 is just a website sitting on top of
1:17:40 stable diffusion and you can upload a
1:17:43 source image and basically do imageo
1:17:45 image prompting
1:17:48 um you can use photoshop's generative
1:17:52 fill so depending on what you want to do
1:17:53 with your base image like if you've got
1:17:56 a person in your image wearing a red
1:17:58 coat you want to change it to a yellow
1:18:00 coat you can just go into Photoshop
1:18:02 highlight the red coat and type in
1:18:04 yellow coat and it will replace it
1:18:08 um which one's the best it depends on
1:18:11 what you want to do I I would say the
1:18:14 one where you can get the
1:18:16 most are they predictable results the
1:18:18 one where you've got the most control is
1:18:20 photoshop's generative fill stable
1:18:23 diffusion and mid journey in terms of
1:18:25 their imageo image stuff are are both um
1:18:30 fairly unpredictable so now if what you
1:18:34 want is to like start with an image of
1:18:36 yourself and and kind of have it take
1:18:39 creative Liberties and make like a
1:18:41 really stylized version of yourself then
1:18:43 I would say stable diffusion and mid
1:18:45 Journey if you want to take like you
1:18:47 know the you can make money with time if
1:18:51 you want to take a photograph of this
1:18:52 button and put it in the middle of a
1:18:54 field somewhere then I would use
1:18:55 generative Phill for that right so so it
1:18:58 depends on what you're trying to
1:19:00 accomplish Chad add that's that's it
1:19:03 don't she just love chat ad well hello
1:19:06 there can't stay along because I've got
1:19:07 rashash Shana tomorrow Services early
1:19:09 tomorrow Happy rajash Shana um Sam but
1:19:13 thank you for hanging out for a bit um
1:19:16 can't resist stopping in I'm so sorry
1:19:18 hey there Firefly it's cool you're
1:19:20 connecting with each other I love that
1:19:23 hey regulars I hope you're all having
1:19:24 lovely evenings to you too sir um great
1:19:28 to see you Thelonius uh can it take on
1:19:31 the personality of the skipper and SS
1:19:34 minow and treat me like first mate
1:19:36 Gilligan yes and I think that has to
1:19:40 happen so we're going to go do exactly
1:19:42 the same thing we did with the stat
1:19:44 statistician thing so wait I got to go
1:19:47 turn Okay so for those of you that are
1:19:49 new here what I'm doing is I'm going to
1:19:51 go play in chat gpt's custom instruction
1:19:54 so before I made it a a
1:20:01 statistician
1:20:03 and I'm going to kill all that and then
1:20:07 okay
1:20:11 so we're going to go expert stats
1:20:14 Professor so basically I'm going to go
1:20:16 into why didn't this load Come on load
1:20:20 load
1:20:22 load
1:20:24 okay so here's a cool thing you can do
1:20:27 so if if you've been here for a bit so
1:20:29 what I did before was I had chat GPT
1:20:32 describe a professor a a
1:20:36 a professor of statistics and then I had
1:20:39 it describe how it was going to behave
1:20:42 and answer questions so now what I'm
1:20:43 going to do is I'm going to
1:20:51 say okay now I want
1:20:54 you to
1:20:58 rewrite the
1:21:02 description of the professor but instead
1:21:06 of being a
1:21:07 professor you're going to be the skipper
1:21:10 from Gilligan's Island period And I want
1:21:13 you to rewrite the description of how
1:21:17 the skipper will
1:21:20 respond and I want the skipper to resp
1:21:23 respond only as
1:21:25 if you're
1:21:27 Gilligan and you're a bumbling idiot
1:21:30 that frustrates the heck out of
1:21:34 you all right so hopefully like I what I
1:21:38 found in the past is that Chachi PT is
1:21:40 actually really good at
1:21:41 this all right so certainly here's the
1:21:44 Rewritten description of the skipper as
1:21:46 the character the skipper known for his
1:21:49 bumbling yet endearing personality
1:21:52 that's not really his thing that's
1:21:54 Gilligan's thing oh often finds himself
1:21:57 exasperated when dealing with Gilligan's
1:22:00 mishaps when faced with questions or
1:22:02 challenges his responses are typically
1:22:04 marked with a mix Mi mix of frustration
1:22:06 and resigned Amusement for example when
1:22:09 Gilligan asks about a
1:22:11 construction constructing a rescue
1:22:13 signal the skipper might reply Gilligan
1:22:16 it's as simple as tying two sticks
1:22:18 together but please don't mess this
1:22:21 up when Gilligan Ires about okay
1:22:26 um okay that's not quite right
1:22:31 so okay for the Skippers
1:22:34 description I want you to start with you
1:22:37 are the skipper from Gilligan's
1:22:39 Island
1:22:42 and then
1:22:45 describe the Skipper's personality
1:22:49 period for the description of how the
1:22:52 skipper will
1:22:53 respond I want you to use the term you
1:22:56 will
1:23:00 respond as
1:23:04 Gilligan asks
1:23:11 questions period um provide
1:23:20 details of the kinds of things Gilligan
1:23:23 might ask and the kinds of
1:23:25 responses the skipper might
1:23:28 provide let's see if those are a little
1:23:31 better and again you
1:23:34 know the the the the big lesson I've
1:23:37 learned with chat GPT is it's less about
1:23:39 prompting and it's more like you're
1:23:41 working with an intern it's like it's
1:23:44 more like you're talking with like a
1:23:45 really thickheaded intern that's like
1:23:47 well you write fast and you know
1:23:49 everything but you're kind of dumb uh
1:23:53 okay you are the skipper from Gilligan's
1:23:54 Island a character known for strong
1:23:56 leadership rugged demeanor and a
1:23:57 Perpetual frustration by Gilligan's
1:23:59 Antics perfect your personality is that
1:24:01 of a seasoned sailor often exasperated
1:24:03 by Gilligan's well intention but you
1:24:06 care deeply for the well-being of The
1:24:08 Castaways on the island and do your best
1:24:10 to keep them safe and organized despite
1:24:12 the chaos caused by your first mate
1:24:14 Gilligan that's perfect as Gilligan asks
1:24:18 questions and seeks guidance your
1:24:20 responses are colored by the history of
1:24:22 Gilligan's Miss
1:24:23 TS perfect for example if Gilligan
1:24:26 inquires about building a boat to escape
1:24:28 okay when Gilligan ask seeks your advice
1:24:30 your responses okay this is perfect so
1:24:33 I'm going to grab the
1:24:36 description of the skipper and that's
1:24:38 going to go in our first thing
1:24:40 here I'm going to save it and then I'm
1:24:43 going to describe I'm going to grab as
1:24:45 Gilligan asked
1:24:49 questions we're going to say this is how
1:24:51 you're going to respond
1:24:55 and we're going to save that and so now
1:24:57 we have chat gbt is the skipper and
1:25:00 we're Gil
1:25:02 again hey Skipper how's it going today
1:25:05 anything I can
1:25:08 do oh Gilligan you always mean well but
1:25:11 you're asking if there's anything you
1:25:12 can do just reminds me of all those
1:25:14 times okay it's a little bit on the nose
1:25:16 um but you know what wait am I in GPT
1:25:18 3.5 I am okay so we're going to go to
1:25:20 GPT 4
1:25:24 gp4 so so this is one of the subtlety
1:25:27 subtleties that you'll find between 3.5
1:25:29 and four is that GPT
1:25:33 3.5 writes structurally well but
1:25:37 shitty and gp4 just tends to be a better
1:25:40 writer okay so
1:25:45 um hey
1:25:49 Skipper how's it
1:25:51 going I'm
1:25:54 ready to solve any problem you want me
1:25:57 to
1:25:58 solve that was horrible question ah
1:26:02 Gilligan good to see your enthusiasm but
1:26:03 let's not forget the time you tried to
1:26:05 solve our food storage Problem by
1:26:06 burying the canned goods you ended up
1:26:09 creating a treasure map for the coconuts
1:26:11 but all right we have something that
1:26:13 needs fixing the radio is acting up
1:26:14 again can you take a look without
1:26:16 electrocuting yourself this
1:26:19 time that's
1:26:21 funny
1:26:25 but I just put the radio in a bucket of
1:26:29 water Gilligan bucket of water why on
1:26:31 Earth would you do that all right so
1:26:34 it's not quite as funny as I wanted it
1:26:36 to be
1:26:37 but I suppose the writing on Gilligan's
1:26:40 Island wasn't all that great either but
1:26:42 anyway that's a good example of just you
1:26:45 know [ __ ] around with these
1:26:48 tools all
1:26:51 right
1:26:56 hey chat GPT right yeah well right have
1:26:59 chat GPT write a Gilligan's Island
1:27:01 treatment where they get off the island
1:27:03 in the same day series
1:27:05 over um actually that's not a bad
1:27:12 idea
1:27:15 uh I'm gonna say new
1:27:21 chat
1:27:29 I don't think I can do it as the
1:27:37 skipper hey
1:27:40 Skipper I want you to come up
1:27:43 with a plan for getting us off the
1:27:46 island today do you think that's
1:27:51 possible Gil boy remember the time you
1:27:53 tried to take charge of our Escape Plan
1:27:55 yeah no I can't do it I can't I can't do
1:27:58 it with with this as the as the custom
1:28:00 instruction I I would I would in that
1:28:02 case I would basically say you know
1:28:05 you're you're a writer that that you
1:28:07 know is writing the sequel to Gilligan's
1:28:11 Island and it's it's a realtime you know
1:28:15 three-hour exercise that gets them off
1:28:17 the island they patch the boat and leave
1:28:19 something like
1:28:21 that uh
1:28:23 moderator Pinn comment oh man I was
1:28:24 wondering why the writing was a bit
1:28:26 mediocre because it was still 3.5 and
1:28:28 not four
1:28:33 yep thanks for the stats demo I need to
1:28:35 take the class that was interesting I
1:28:38 use chat GPT to write a song but how do
1:28:40 I get the keyboard chords for the songs
1:28:43 using chat GPT so um you just you just
1:28:47 ask for it so let me do gp4 because it's
1:28:51 better
1:28:54 write me a country song
1:28:58 about rich men north of
1:29:07 Richmond verse
1:29:10 one we went down on the coast where the
1:29:13 water's fine but up north of Richmond
1:29:15 there drinking expensive
1:29:18 wine I'm singing it like he does rich
1:29:21 man North rich man got your gold and
1:29:23 your bling but you don't understand a
1:29:27 ain't money money ain't
1:29:30 everything you can't buy a mansion you
1:29:33 can't U buy a friend Rich Man and north
1:29:36 of Richmond one day you'll see in the
1:29:37 end um all right so there it wrote the
1:29:41 you know the verse in the choruses so
1:29:42 now I can say
1:29:46 um I'd like you to add choruses to this
1:29:51 and
1:29:53 have it be more sophisticated than a
1:29:55 three chord country song starting in the
1:29:59 key of
1:30:00 G and using major 7ths and minor chords
1:30:05 as appropriate to keep it
1:30:09 interesting and so now it
1:30:13 should rewrite yeah so so it's rewriting
1:30:17 it in tab right so this is G major 7th C
1:30:20 major 7th D7 and G or
1:30:24 7th and so it's actually sort of burying
1:30:27 the chords within the signals right so
1:30:29 this is this is basic You Know Tab
1:30:31 structure chord you know chord
1:30:34 structure
1:30:37 um and it you know it it understands
1:30:39 basic music theory my experience with
1:30:42 this is that um the more you know about
1:30:45 music theory and can look at what it
1:30:48 creates and go no that's bad or that's
1:30:51 cliche or that's not following music
1:30:53 theory like like actually an interesting
1:30:55 way to do this would be um let's see
1:30:57 let's see
1:31:00 um pretend like you're a music theory
1:31:03 teacher and explain to
1:31:07 me the difference between a basic
1:31:10 country song and a sophisticated country
1:31:15 song ah music theory oh wait is this do
1:31:19 I still have
1:31:20 on oh yeah hang on I I still have on the
1:31:25 skipper
1:31:27 here let me stop
1:31:35 generating stop
1:31:37 generating custom instructions turn that
1:31:40 [ __ ]
1:31:42 off okay so let's say
1:31:47 um I want you to act as a music theory
1:31:51 professor from the Berkeley School of
1:31:54 Music
1:31:55 period I'd like you to quickly educate
1:31:57 me in basic music theory for country
1:32:02 songs
1:32:05 period And I want you to
1:32:08 describe two
1:32:11 different structures of a basic country
1:32:14 song compared to two different
1:32:17 structures of sophisticated country
1:32:20 songs and
1:32:23 explain how the chords make a
1:32:26 difference I don't know something like
1:32:28 that let's see what it
1:32:31 does welcome to country music 101 major
1:32:34 and minor
1:32:37 chords common chord
1:32:42 progressions the 251 is a more Jazzy
1:32:45 progression but can still be
1:32:48 found verse course verse course
1:32:51 Bridge
1:32:52 verse vers Bridge so so it's it's given
1:32:55 us you know not only not only chord
1:32:58 structure but song
1:33:01 structure
1:33:02 um Extended chords like seventh and
1:33:05 Ninth or even suspended chords add
1:33:14 color this is great can you write me the
1:33:18 lyrics for a country song called rich
1:33:22 men from
1:33:38 Richmond on the banks of the James of
1:33:40 the river bends grew up with the stories
1:33:42 the myth the Legends Boys in blue jeans
1:33:44 dreaming of gold hoping they'd be tale
1:33:46 hoping they'd be the tales that are told
1:33:48 rich men from richm with dreams so tall
1:33:51 they got to their
1:33:53 Mansions but they've seen their Falls
1:33:55 all right so all right still still
1:33:58 writing GPT 4 is slower but better okay
1:34:01 so there's that so
1:34:04 now interpret these
1:34:07 lyrics
1:34:11 and add in chords using the more
1:34:15 sophisticated country
1:34:19 approach that you describe above in a
1:34:22 way that brings these lyrics to
1:34:31 life certainly here's my interpretation
1:34:34 starting in the key of G so this is g
1:34:37 cad9 g overb d sus 4 and G that's pretty
1:34:41 cool on the banks of the James so so it
1:34:45 didn't quite format it the same but you
1:34:47 know now you've got something that is at
1:34:50 least
1:34:55 got interesting chords but it's got the
1:34:57 chorus and the
1:34:59 verse using the exact same
1:35:05 chords so what I would do in this case
1:35:08 the bridge actually does something
1:35:10 interesting with the
1:35:17 chords uh uh
1:35:20 uh d d
1:35:26 d outro
1:35:30 nice so I'm going to
1:35:33 say can you
1:35:36 analyze what you wrote period I noticed
1:35:39 that the chord changes in the
1:35:42 verse and the chorus are
1:35:47 identical and in most songs I know they
1:35:51 are different
1:35:52 period did you do that on
1:35:55 purpose or not question mark if not how
1:35:59 would you improve the song after you
1:36:01 analyze
1:36:06 it
1:36:08 analysis you've rightly noticed that the
1:36:10 verses and choruses have the same chord
1:36:12 progressions that similarity can create
1:36:14 a continuous flowing feeling throughout
1:36:16 the song but it might also result in a
1:36:19 lack of contrast between the sections in
1:36:22 many songs it's important to distinguish
1:36:25 the
1:36:26 chorus extended
1:36:28 chords all right improvements to add
1:36:32 contrast we can change the chord
1:36:34 progressions in one of the sections
1:36:36 since the chorus is the song's
1:36:37 centerpiece I'd recommend altering the
1:36:39 verse
1:36:41 chords so the verse is now G E minor c
1:36:45 and d and the chorus is all the fancy
1:36:48 [ __ ] um this may or may not be any good
1:36:51 but anyway that's that's how I would
1:36:52 approach that so yes it can do all that
1:36:54 can it do it good I don't
1:36:57 know and again this is a place where
1:36:59 custom instructions could be really
1:37:01 interesting right where where you could
1:37:02 do custom instructions where you know um
1:37:07 tell Chad GPT to act like you know
1:37:10 Master country songwriter and you could
1:37:12 say you know like like you know Brandy
1:37:14 Carlile or John Hyatt or you know take
1:37:17 people that you know and then in how
1:37:19 they respond tell them that you want
1:37:21 them to songs that are interesting and
1:37:23 and compelling and whatever it might be
1:37:25 and I I don't know not derivative and
1:37:28 and so so yeah you could fine-tune this
1:37:31 to to actually probably be a really good
1:37:33 songwriting partner or or or don't even
1:37:36 do uh Brandy carile you know describe
1:37:39 yourself right here's the kinds of songs
1:37:41 I like here's the kinds of chord
1:37:43 progressions I like here's the kinds of
1:37:45 you know metaphors I like using or or
1:37:48 you know I like more straightforward
1:37:50 lyrics I Don't Like ryy rhymey [ __ ]
1:37:52 right I like more sophisticated Rhymes
1:37:55 um all that sort of stuff and basically
1:37:57 if you get that dialed in right you'll
1:37:59 have a partner that you can basically
1:38:01 pop in an idea and it'll get you to an
1:38:05 80% version of your
1:38:07 song in 3 seconds and you do 30 of
1:38:12 those that would be awesome you know
1:38:15 wouldn't that be awesome that would be
1:38:17 awesome yeah smik Tyson here yeah that
1:38:20 would be awesome
1:38:22 that'd be
1:38:26 awesome all right let's see what time we
1:38:29 got 10:30 good Lord people what are you
1:38:31 doing with my life what are you doing
1:38:33 with my life it's Friday night Friday
1:38:35 night what are you doing with your life
1:38:36 why you on a Tik Tok why you on a Tik
1:38:38 Tok watching an old guy why are you
1:38:42 watching an old guy on a Tik Tok it
1:38:45 don't make no sense we're going to have
1:38:47 to have an
1:38:48 intervention you know I don't know if
1:38:51 you knew this is but uh you talk a lot
1:38:54 about artificial
1:38:57 intelligence you know artificial
1:38:59 intelligence was there before last year
1:39:02 you know that
1:39:04 right oh I
1:39:06 know all right I'm going back up into
1:39:09 the questions I'm jumping up in there if
1:39:10 any of the mods uh see any interesting
1:39:12 questions go ahead and pin one and I'll
1:39:15 do that but also go look for some oh man
1:39:17 I was Wonder okay that was what we saw
1:39:19 before while I've been watching I had
1:39:22 chat GPT write me an entire booklet for
1:39:24 my industry as a download perfect Jim oh
1:39:28 so Jim Jim was on the uh Jim was the one
1:39:30 I was talking about that was on the uh
1:39:32 the LinkedIn office hours today who's
1:39:35 been using chat gbt for like since
1:39:37 November and yeah awesome so he's he he
1:39:41 created a a booklet for his industry
1:39:44 that's fraking awesome yeah and and you
1:39:47 know Jim because you know your industry
1:39:50 so well you're going to be able able to
1:39:51 look at what chat GPT wrote and go oh
1:39:54 Section 3 is [ __ ] and it you know
1:39:56 that's that's old thinking and we do
1:39:58 things differently now and then you can
1:39:59 just tell it that rewrite section three
1:40:01 so that it includes the new ways of
1:40:03 doing things um the other thing to do
1:40:06 Jim if you haven't done it I think I
1:40:07 showed this today no I didn't let me
1:40:11 show you something cool Jim if if if you
1:40:14 you probably know this you've been using
1:40:16 this long enough but just in case not
1:40:17 other people may not all right so we're
1:40:19 going to go did I I turned off custom
1:40:22 instructions right oh no wait did I yeah
1:40:25 I did okay welcome to chat add all right
1:40:30 so I'm going to say
1:40:33 um let's
1:40:37 see you are an expert at the logistics
1:40:41 of
1:40:43 starting
1:40:45 managing and
1:40:47 expanding a storage unit business period
1:40:52 I want you to write a
1:40:55 booklet for beginners who are getting
1:40:59 ready to invest in their first storage
1:41:01 business
1:41:04 period and then and then here's the
1:41:06 thing that if you if if you don't know
1:41:08 this trick this is this is a good one um
1:41:10 sometimes it will just do this um
1:41:12 automatically but if it doesn't this is
1:41:14 a good one to add so so here's here's
1:41:16 what I'm adding to
1:41:17 it use
1:41:20 markdown to
1:41:22 format the booklet to make it easier to
1:41:25 read period be sure to include titles
1:41:29 subheads and bulleted lists as
1:41:34 appropriate and then what it will do is
1:41:38 like give you all the sort of big
1:41:40 Bolding welcome to the exciting world of
1:41:42 storage unit business table of contents
1:41:44 right so here's the Bold here's the
1:41:46 numbered list right introduction
1:41:49 subtitle you know preliminary
1:41:54 considerations so anyway it's probably
1:41:57 something you know but if you don't know
1:41:58 that the the the term format using
1:42:01 markdown markdown is basically like a
1:42:04 kind of simplified version of HTML for
1:42:06 lack of a better metaphor um and you
1:42:09 basically just tell it to do that and it
1:42:11 kind of knows it it knows how to bold
1:42:13 stuff and make bulleted lists and things
1:42:15 like that so
1:42:18 anyway welcome Jim mm
1:42:24 [Music]
1:42:29 three classes left uh what app generated
1:42:33 that cool Saloon pick the cool Saloon
1:42:38 pick this was done in um
1:42:44 ideogram and then all of the outside of
1:42:48 this was done in um adob Adobe Photoshop
1:42:52 generative fill like she was done in
1:42:54 Adobe Photoshop gener fill I put like
1:42:58 cranky
1:42:59 bartender and and that was what it
1:43:01 created which is she's freaking awesome
1:43:04 look at look how angry she
1:43:07 is uh wait why is it what's going on
1:43:11 here there we
1:43:14 go she's so
1:43:17 cranky and then I actually went in and
1:43:19 photoshopped this this is for the the AI
1:43:22 Salon oh no I didn't I didn't Photoshop
1:43:24 this one this ISS the AI salani um I did
1:43:26 a Photoshop version of this where I took
1:43:28 all that
1:43:29 out but yeah a combination of ideogram a
1:43:34 um that does the
1:43:35 letters and then Photoshop generative
1:43:39 fill for that I know that's a pretty
1:43:41 slick little image isn't it I dig that I
1:43:43 dig that one pretty good although this
1:43:46 dog in a space helmet ain't bad it ain't
1:43:48 bad you know dogs are all right you know
1:43:51 they're tasty they're tasty
1:43:55 champ mhm I like dogs dogs are tasty M
1:44:01 mhm they called it a sling blade
1:44:07 mm mod power
1:44:11 activate digital Gods oh o I got a got a
1:44:16 little dancing star in my head what's
1:44:18 your favorite GPT is it chat um um it's
1:44:21 a good question so here let me show
1:44:28 you I have a top
1:44:31 seven and it's not that I necessarily
1:44:33 have a favorite one um gp4 is by
1:44:39 far in my
1:44:41 opinion um the
1:44:44 best the best let's see it's by far the
1:44:49 one that performs consistently better
1:44:52 than the others but it's got a lot of
1:44:56 limitations all of these things
1:44:58 hallucinate and they they all have
1:44:59 different advantages so I'll kind of
1:45:01 walk through I'll walk through my top
1:45:03 seven these three tend to be the ones
1:45:06 that I kind of bounce between the most
1:45:08 and and Bard I I don't use a ton unless
1:45:10 I need something connected to the
1:45:12 internet if like a plugin like um web
1:45:16 pilot fails up here so most of the time
1:45:21 if I've got a long document I need to
1:45:23 analyze I'm using Claude because this
1:45:25 has 100,000 token context window which
1:45:27 basically means 75,000 words so The
1:45:31 Great Gatsby the novel The Great Gatsby
1:45:33 is like 55,000 words so you could put
1:45:36 the entire Great Gatsby into the prompt
1:45:39 of Claude and have it write you a new
1:45:42 chapter
1:45:46 U um chat
1:45:49 GPT has GP 4 which is good it's got
1:45:53 plugins which are okay and then it's got
1:45:55 code interpreter or now called Advanced
1:45:57 Data analysis which is [ __ ]
1:46:00 remarkable so this one is still the King
1:46:03 of the Hill Bard for most of this year
1:46:07 was unusable it was so bad it was
1:46:11 unusable it was it was so bad and now
1:46:14 it's not now it's actually pretty good
1:46:17 um and Gemini is is probably going to be
1:46:21 out within a month or two excuse me and
1:46:25 um Gemini is purportedly going to be
1:46:28 five times better than gp4 I'll believe
1:46:30 it when I see it
1:46:34 but Google kind of got their ass handed
1:46:37 to them from a PR standpoint with with
1:46:40 chat GPT and
1:46:41 Microsoft for a lot of this year so they
1:46:45 have a chip on their shoulder which they
1:46:47 should they invented the Transformer
1:46:48 that makes all this [ __ ] possible okay
1:46:52 then these are the other ones that I
1:46:53 think have got some some Merit to play
1:46:55 with Bing we've talked about a lot
1:46:57 tonight it seems to be getting shittier
1:46:59 the the safety guard rails on this make
1:47:01 it you know decently unusable unless
1:47:04 you're doing sort of pablum kind of you
1:47:06 know steuart SMY level
1:47:10 interactions I'm good enough I'm smart
1:47:12 enough and dog on it people like me
1:47:15 yeah
1:47:17 um Bing is connected to the internet it
1:47:20 is built on GPT 4 it can generate images
1:47:23 it can analyze images so there's there's
1:47:25 things going for it but the safety guard
1:47:28 rails right now make it make it I don't
1:47:30 I I I I stopped using it um and I was a
1:47:33 fan of it I I sort of pushed it for a
1:47:35 while and I just stopped because they
1:47:37 got to get their [ __ ] together um
1:47:39 perplexity doai on its own is a really
1:47:41 interesting tool because when you do
1:47:44 searches uh when when you do um
1:47:47 interactions with it it provides sources
1:47:49 which are actual
1:47:51 links to actual sites so it's good for
1:47:55 research and then labs. perplexity lets
1:47:58 you play with meta llama models then
1:48:00 you've got Pi AI which we lovingly in
1:48:03 the lab here call that chatty
1:48:06 GPT um the thing to do with pi is put it
1:48:10 in
1:48:12 um get the app get the IOS app or the
1:48:15 Android app and put it in voice mode so
1:48:18 you talk to it and then basically when
1:48:20 you stop talking it just you know takes
1:48:22 what you said and it runs it through its
1:48:25 model and then it responds with its
1:48:27 voice um it's designed for talking and
1:48:30 it's it's really interesting and then po
1:48:33 lets you play with multiple of these
1:48:35 models in the same interface and it's
1:48:36 also got Bots and a bot maker so those
1:48:39 are the those are the seven that I that
1:48:41 I think that that I use on a decently
1:48:43 regular basis I think they're all worth
1:48:45 knowing something about um someone
1:48:48 something else was was uh was oh there's
1:48:51 there's something uh pinned right now
1:48:53 wonder how accurate chat GPT is with
1:48:55 respect to syntax any measurements from
1:48:59 Pilot or
1:49:02 beta
1:49:03 um well chat GPT can code so so one of
1:49:08 the things that happened when they
1:49:09 trained so this was this was an
1:49:11 interesting
1:49:17 thing when they first trained GPT
1:49:21 3.5 um they they trained it on sort of
1:49:24 like all the data on the Internet or as
1:49:26 much as they could get in
1:49:28 there and it and it was a certain level
1:49:31 of good and then at some point they
1:49:33 created the code fine-tuning where they
1:49:35 they trained GPT 3.5 to do coding right
1:49:39 they trained it on a bunch of
1:49:40 programming
1:49:41 languages Apparently after they trained
1:49:44 it on the coding the language side got
1:49:47 dramatically better um so so there's
1:49:50 there's some interesting thing where the
1:49:53 logic of
1:49:54 coding um because it understands that it
1:49:58 got better at the logic of English or or
1:50:00 language um so it's quite good at coding
1:50:05 so it's quite good at syntax so um I
1:50:08 would just go experiment with it go have
1:50:09 it write you some code um what I've
1:50:12 heard from engineer friends of mine and
1:50:15 people that are working on it is that
1:50:18 the code quality for gp4 has gone down
1:50:21 pretty dramatically in the past 3 months
1:50:23 4 months um but anytime I've interacted
1:50:27 with it recently it's still pretty good
1:50:29 but I'm not doing a lot of coding like
1:50:30 if I doing coding dayto day it might be
1:50:33 ugly Bard b.google.r
1:50:35 good at coding so I would play with that
1:50:39 and especially if Gemini is indeed going
1:50:42 to be five times better than
1:50:46 gp4 it's probably going to be five times
1:50:48 better at coding first right right
1:50:50 Google's an engineering company not a
1:50:52 Humanities company right Google ain't
1:50:55 apple apple hires creatives right who
1:50:59 are also technologist creative
1:51:01 technologist or or creatives and
1:51:03 technologist Google hires Engineers
1:51:07 brainiacs right so so my
1:51:11 guess is that Gemini is going to be
1:51:14 better at logic coding syntax things
1:51:17 like that now that
1:51:19 said Google owns
1:51:22 YouTube so Gemini might be way better
1:51:26 than gp4 because it was trained on all
1:51:29 of YouTube that's a possibility here so
1:51:32 I don't know what it was trained on but
1:51:35 we'll we'll have to wait and see anyway
1:51:37 all right Bing is my boy um bing bing
1:51:40 also had a survey today interesting oh
1:51:42 my God generative Phil new
1:51:45 mascot generative Phill exactly oh Sam
1:51:48 go make a generative fill all right
1:51:50 anyone who wants to make a generative
1:51:52 fill make it in ideogram so by the way
1:51:55 ideogram is the thing that can do words
1:51:57 so if you see all these things with
1:51:58 words on them that's all from
1:52:01 id. um so um yeah go make a generative
1:52:07 fill and throw it in the uh the share
1:52:10 the the showand tell channel in Discord
1:52:13 in the salon Discord that's really funny
1:52:16 SS minow and away we go all
1:52:19 aboard
1:52:21 all right come sit right back and you
1:52:23 hear a tale A Tale of a faithful trip
1:52:25 that started on this traic wait on
1:52:28 this aboard this tiny ship the mate was
1:52:33 saing man a skipper tried and
1:52:40 true I used to know it cold oh my God
1:52:44 Gilligan's
1:52:46 Island I was I was as much time time as
1:52:50 I put into the AI learning lab I put
1:52:54 that much time into Gilligan's Island as
1:52:56 a kid I [ __ ] watched that show
1:52:59 whenever I could is AI able to predict
1:53:02 likely likely numbers using a set of old
1:53:05 data using probabilities or Series so
1:53:08 Prince
1:53:10 K yeah well I don't know enough about
1:53:13 statistics and able to predict likely
1:53:17 numbers using a set of old data but
1:53:20 Advanced Data analysis can do full-on
1:53:24 data analysis so I assume it can do
1:53:26 that um how you play
1:53:30 with um Advanced Data analysis is you
1:53:34 have to be a chat GPT plus
1:53:37 member you have to turn it on so if you
1:53:39 don't see it you have to go to your
1:53:41 settings and then go to Beta features
1:53:44 and beta features you will then turn on
1:53:47 plugins and data
1:53:49 analysis
1:53:51 and then now you go to data analysis and
1:53:54 what it allows you to do see that plus
1:53:56 button you can now upload a
1:54:01 file your system is run out of
1:54:03 application memory oh Microsoft Edge
1:54:06 you're a real pain in the ass how about
1:54:08 if we
1:54:09 just pretend like that didn't happen
1:54:15 Okay the technologists on this thing are
1:54:17 going no you got to restart nah you got
1:54:20 to live on the edge people you got to
1:54:22 live on the edge um let's go to Archive
1:54:28 archive name contains archive
1:54:32 desktop archive 4
1:54:35 okay here's an archive for.zip
1:54:40 so we now have a zip file here and what
1:54:44 you can do here is you can
1:54:47 say please open this and tell me what
1:54:51 files are inside and if they are data
1:54:55 files examine the data and tell me what
1:54:57 might be interesting about
1:55:02 it so this is
1:55:06 um what this is doing oh see it just it
1:55:10 just crapped out again this this might
1:55:11 not work let's see yeah it's good we're
1:55:15 good here let me quit something else
1:55:17 we'll quit that
1:55:21 [Music]
1:55:29 it didn't quit or did it I don't
1:55:32 know yeah we'll quit keynote too keynote
1:55:35 a big
1:55:38 one oh wait I don't want to Quick
1:55:40 keynote oh well we just
1:55:44 did okay so so what what advanced data
1:55:50 analysis is doing is it's creating
1:55:52 container I used to say it's creating a
1:55:54 virtual machine and and and the
1:55:56 technology Boys in here [ __ ] slapped me
1:55:58 instead it's not really a virtual
1:55:59 machine it's technically a container all
1:56:02 right so it's a container anyway it's
1:56:04 creating some hermetically sealed little
1:56:06 Computing
1:56:07 area um with with within um within
1:56:11 memory basically and it's G it's been
1:56:15 given access to a number of python
1:56:17 libraries so what ends up happening is
1:56:21 code interpreter or Advanced Data
1:56:23 analysis writes and executes its own
1:56:26 python code so to open the zip file it
1:56:29 it imported the library to do that it
1:56:31 wrote the code and it opened it up then
1:56:34 it says now I'll examine the CSV file
1:56:36 that's in there so it brings in the
1:56:38 Panda's library and writes the code to
1:56:41 look at the first few lines of the thing
1:56:43 and then it says the CSV file contains
1:56:46 data on unemployment in various US
1:56:48 states for specific months and years
1:56:51 here's a breakdown of the columns so it
1:56:52 gives you a breakdown of the column so
1:56:54 it knows what's in there let's see given
1:56:57 this structure we can analyze various
1:56:59 aspects such as states with highest is
1:57:01 there a specific
1:57:03 analysis I'd like you to tell
1:57:07 me what you think are the most
1:57:10 interesting Trends in the data based on
1:57:14 your analysis of
1:57:18 it
1:57:20 all right I'll perform a comprehensive
1:57:23 analysis now so notice what it's doing
1:57:26 here I just said I want you to tell me
1:57:28 what's interesting about this all right
1:57:31 I'll perform a comprehensive analysis of
1:57:33 the data set to uncover interesting
1:57:35 Trends here are the steps I'll follow
1:57:37 analyze the overall trend of
1:57:38 unemployment across States over the
1:57:40 years find the states with the highest
1:57:42 and lowest
1:57:44 unemployment volatility identify states
1:57:46 with the most volatile unemployment
1:57:49 rates long long-term performance check
1:57:51 the things now look it's it's already
1:57:53 it's creating charts and graphs right so
1:57:56 it brings in the matplot library right
1:57:59 the P plot as PLT and then it writes
1:58:03 this [ __ ] to to write that
1:58:06 chart
1:58:08 so
1:58:10 amazing overall the trend late 70s to
1:58:13 early 80s this was a significant rise in
1:58:15 unemployment early 2000 there's a
1:58:17 noticeable spike in the early 2000 which
1:58:19 can be attributed to the dot bubble
1:58:21 burst
1:58:23 2008 corresponded to the global
1:58:25 financial crisis post 2010 a decline in
1:58:28 unemployment rates a slight uptick in
1:58:30 recent years here's the states with the
1:58:33 highest rates this is doing full-on data
1:58:36 analysis fullon data analysis so yeah
1:58:39 you can also have it right you can have
1:58:42 it Recon you can have it clean up data
1:58:44 and rewrite you CSV or Excel files um
1:58:48 with cleaned up data
1:58:50 um you can make it have it make pretty
1:58:52 charts and graphs it's still going right
1:58:55 so it's prompting
1:58:57 itself to go look at all this stuff and
1:59:00 find out what's interesting about the
1:59:02 data summary the overall unemployment
1:59:03 rate across the US experienced
1:59:05 noticeable spikes certain States and
1:59:08 areas New York City Michigan West
1:59:10 Virginia have been more susceptible to
1:59:12 ex economic
1:59:13 fluctuations some states Nebraska South
1:59:15 Dakota North Dakota have consistently
1:59:18 maintained low unemployment rates
1:59:20 others have higher averages would you
1:59:22 like to an explore I'm going to say
1:59:26 um what is the most interesting way you
1:59:29 could visualize some of the trends
1:59:32 you've identified that will be beautiful
1:59:34 to look at and make the information
1:59:41 clear visualizations so a heat map of
1:59:44 the average unemployment rates by state
1:59:46 a line plot a bar plot with top and
1:59:50 bottom States area chart for volatility
1:59:54 I'll start by creating a heat map of the
1:59:57 average unemployment ratees so I know
1:59:58 nothing about statistics and data
2:00:00 analysis I I mean I've heard of some of
2:00:02 these charts but I don't know that like
2:00:05 I don't know what off the top of my head
2:00:07 so it does and so now what's it going to
2:00:10 do here let's see what it does I
2:00:12 apologize for the oversight it seems
2:00:14 there was an error in plotting the
2:00:15 boundaries of the
2:00:16 USA instead of using the previous
2:00:19 approach I'll use a different method to
2:00:21 create the heat map for the average
2:00:23 unemployment rate by state let's try
2:00:25 that so it may not have access to State
2:00:30 boundaries yeah so like whatever that is
2:00:33 is
2:00:34 weird so that's a blank
2:00:41 chart
2:00:42 [Music]
2:00:43 um yeah that's fine so
2:00:46 anyway that's Advanced Data analysis
2:00:48 crazy isn't it it's crazy I tell you
2:00:51 it's crazy I tell you it's just crazy
2:00:54 it's crazy man what are you doing what
2:00:58 are you doing running your own
2:00:59 spreadsheets anymore put it in the
2:01:03 Chet I don't know if you knew this boy
2:01:06 you can make money
2:01:08 with you say what I'm
2:01:10 saying oh boy all right what time is it
2:01:14 it's 11: isn't it
2:01:17 10:56 all the Castaways had last name
2:01:19 but Gilligan but was Gilligan a first
2:01:23 name or a
2:01:28 last I don't know I do not know the
2:01:31 answer to that I got to go to sleep good
2:01:33 night all good night Sam have a good
2:01:35 have a good
2:01:36 weekend Professor Saves the Day see this
2:01:39 is just what the wga writers were afraid
2:01:42 of isn't it
2:01:46 well
2:01:48 yeah
2:01:52 I don't know
2:01:54 listen I have such mixed feelings about
2:01:57 the writer strike I
2:02:00 support everything they're
2:02:03 doing except for the contention that AI
2:02:07 is evil and shouldn't be
2:02:09 used they they just it it just feels
2:02:12 like both them and the SAG positioning
2:02:15 on the role of AI is the misguided piece
2:02:19 like I want them to have you know rights
2:02:22 for their likeness rights if they use AI
2:02:25 to write scripts like all the stuff that
2:02:27 they want is good but they've got this
2:02:29 attitude of like AI is evil and we don't
2:02:31 want you to ever use it and I just think
2:02:33 it's really
2:02:34 shortsighted um but yeah they they they
2:02:37 should they should be both scared and
2:02:39 excited like the smart WR the smart
2:02:42 writers should be spending the
2:02:45 strike getting masterful at using
2:02:50 generative AI to help with the writing
2:02:52 process they should because the writers
2:02:55 that can turn out work faster better all
2:02:58 that sort of stuff are going to be in
2:03:00 demand in the
2:03:02 future um and I think it's still going
2:03:05 to require writers right it you know
2:03:08 the writing isn't about the act of
2:03:11 blocking and tackling language
2:03:14 formatting writing is about curating
2:03:17 ideas and refining ideas and expressing
2:03:20 ideas and while chat GPT can blurt out a
2:03:23 bunch of
2:03:25 [ __ ] it's ultimately the writer that
2:03:27 goes that's the good idea that's the bad
2:03:31 idea and the ones that are the most
2:03:33 facile with that and can do
2:03:35 that quickly and effectively are going
2:03:39 to be in demand and the ones who can't
2:03:42 won't be so anyway I I think you know
2:03:46 just proclaiming AI is evil is is is
2:03:50 idiotic I think it's
2:03:52 idiotic like not for nothing but chat
2:03:56 gppt is the most powerful writing tool
2:03:58 in the history of humanity and the
2:04:01 writer's Guild is saying we should never
2:04:03 use
2:04:05 that
2:04:10 really we will only use typewriters for
2:04:13 here on out these new fangled computers
2:04:16 with their memory and cassette tapes
2:04:20 that store data it's not a piece of
2:04:23 paper is
2:04:25 it it's exactly the same as
2:04:31 that why would you not use a computer if
2:04:34 you're on
2:04:35 typewriters let's have a strike and
2:04:37 let's tell the studios to that that they
2:04:41 can never hire writers that use
2:04:43 computers they we only want them to hire
2:04:46 writers that use typewriters because
2:04:48 that's real writing makes
2:04:51 sound every time you strike a letter the
2:04:54 letter imprinting upon the page is part
2:04:56 of the writing
2:04:58 process that's the equivalent of what
2:05:00 they're what they're arguing right now I
2:05:03 think it's just [ __ ] idiotic I use
2:05:06 chat GPT to write a song but how do I
2:05:08 get the keyboard chords we did that wow
2:05:11 I'm I'm behind on my little questions
2:05:13 here can anyone name a famous song that
2:05:15 starts with the bridge didn't doesn't
2:05:18 Taylor Swift start a lot of her songs
2:05:20 with the bridge she's really into
2:05:22 Bridges I don't but I don't know I don't
2:05:24 know enough of her work to know that I'm
2:05:27 I'm in total
2:05:29 mansplaining territory here interesting
2:05:32 I'll try to keep interesting I will keep
2:05:34 trying the results needed too much work
2:05:37 yep that's a mega glass of wine
2:05:41 this that is
2:05:47 coffee or you meant this one that's just
2:05:50 water that's that's uh Rocky Mountain
2:05:53 tap
2:05:56 water this is the intervention you're
2:05:59 right
2:06:00 Nancy this is the
2:06:02 intervention wow results using format
2:06:05 markdown and bulleted list prompt neat
2:06:07 yeah Isn't that cool Firefly yeah print
2:06:09 AI is the
2:06:11 past that's good um love the salon piic
2:06:15 seems like some AI have problems
2:06:17 generating text so Hamilton so this is
2:06:20 an important thing um to my knowledge
2:06:24 well there's one other tool that sort of
2:06:25 does it but to my knowledge the and it's
2:06:28 brand new it's like a three- week old
2:06:30 tool so if we go to
2:06:33 id. get rid of our dead
2:06:37 memory bug oh let's see the memory might
2:06:41 might have bit Us in the
2:06:47 ass stop
2:06:50 stop all right are we in we're in okay
2:06:56 so ideogram does text it does it
2:07:00 well um these guys are new they they're
2:07:05 they're brand new I'm going to have to
2:07:07 quit
2:07:09 Edge oh that's a cool one let's let's
2:07:12 steal that one so so one of the things
2:07:14 that that I like to do is I'll come in
2:07:18 here oh this is going to be
2:07:22 unusable and I'll copy a prompt let me
2:07:26 just move this out of the way and then
2:07:28 keep going as if it's not there oh I
2:07:30 can't
2:07:32 uh
2:07:34 copy and we'll go out here yeah this is
2:07:38 this is about to explode it's about to
2:07:45 explode 3D render typography
2:07:51 poster vibrant and uplifting fuzzy wool
2:07:54 style word
2:07:57 text AI
2:08:01 learning lab 3D render and I've already
2:08:05 click those little buttons so I think I
2:08:06 can get rid of that
2:08:08 crap and we're going to make it vertical
2:08:11 and we're going to try to get rid of
2:08:12 that and then we're going to say
2:08:13 generate and let's see if we get
2:08:16 something they look like AI learning
2:08:17 loveb now here's the deal
2:08:20 the this
2:08:23 tool gets the letters right maybe 15% of
2:08:28 the
2:08:29 time so you have to keep messing with it
2:08:33 um but if you can get it close then
2:08:39 um
2:08:41 like AI learning lab is close learning
2:08:46 is
2:08:47 close
2:08:51 um that one's kind of cool so if you get
2:08:54 something that's close but not quite
2:08:56 there um what you can do is remix
2:09:00 it and just sometimes it'll just don't
2:09:04 do anything just keep just make another
2:09:06 four based on that one and then
2:09:09 sometimes it'll get the the words right
2:09:11 yeah like this this one wasn't quite
2:09:12 close enough but anyway this is this is
2:09:16 uh ideogram a
2:09:19 [ __ ] this I'm going to
2:09:21 Mars I love that awesome all right I'm
2:09:25 quitting Microsoft
2:09:27 Edge uh I don't need to force quit just
2:09:30 quit yeah there we go did it go it's
2:09:33 going it's going oh I might need to
2:09:34 force quit
2:09:37 ah oh you
2:09:39 Macintosh all right
2:09:41 listen my Max had
2:09:43 it we got the turtles we got the janky
2:09:47 stuff yeah we'll go out we'll go out
2:09:49 with a nice AI learning lab Ali the AI
2:09:51 learning lab it's how we started the
2:09:53 night it's how we're going to go
2:09:57 [Music]
2:10:03 out all right everybody listen for those
2:10:07 of you that modded tonight thank you for
2:10:09 your modding I really do appreciate it
2:10:11 Kyle is there a way to delete pictures
2:10:12 on ideogram I do not know I don't know
2:10:17 um just go into your count
2:10:20 and I don't
2:10:22 know see you tomorrow all right Silver
2:10:25 Fox all right so let's let's uh let's
2:10:27 let's say goodbye who who have we seen
2:10:29 tonight uh Nancy and Silver Fox and
2:10:33 Zippy to pin head
2:10:35 thonus um Pate was in here I knew it hey
2:10:39 Pate how's it going girl with
2:10:43 needles solid shared the live thank you
2:10:46 for sharing the live I do appreciate
2:10:48 that Firefly Prince K good to see you
2:10:52 who else we got in here tonight Emilio's
2:10:54 wife welcome or good night um big fish
2:10:58 good night good night good night steo
2:11:01 and cam
2:11:02 katkin winning AI with cam katkin cam
2:11:06 katkin signing off for the night welcome
2:11:08 I'm cam katkin this is the news traffic
2:11:11 on the eights with Cam
2:11:13 catkin I hope cam appreciates me every
2:11:17 time I see Cam's name I'm like I love it
2:11:20 I love it I gotta I just got to go
2:11:23 there uh digital Gods thank you so much
2:11:27 all right all right everybody peace out
2:11:29 have a great F Friday night it's been
2:11:32 fun it's been a good one I had a good
2:11:34 time hope you learned something I'd
2:11:35 appre I I uh oh Cam katkin I appreciate
2:11:38 you love great thank you um I think
2:11:41 you're all well this is fun all right
2:11:44 I'll see you tomorrow