Dec 2, 2023 - Experimentation with tools like ChatGPT and Runway ML

Description

Chapters

Transcript

0:02 [Music]
0:10 [Music]
0:29 oh
0:32 [Music]
0:38 there's been something baby I've been
0:40 trying to
0:42 say for an age and it seems I don't know
0:48 how with the past and the future now
0:51 surrounding
0:53 me surrender to whatever cheap through
0:56 can be
0:58 found there's been little
1:01 trouble since you came to my
1:05 [Music]
1:09 resue champy wants to welcome you all to
1:12 the AI learning lab happy birthday CH B
1:15 happy birthday
1:20 [Music]
1:24 CH woo that is Out Of
1:29 Tune champ he's doing vocal exercises he
1:32 is he's warming us up he's getting us
1:34 going happy Saturday night
1:37 everybody Welcome to the AI learning
1:40 lab never a dull moment here in champ
1:44 town I think I'm all com covered in
1:46 Champ here I was taking a little nappy
1:48 on his
1:51 bed yes we're one of those houses where
1:53 the dogs are on all the furniture well
1:56 the
1:58 dog we uh we are a a boundaryless family
2:03 when it comes to uh canine
2:07 participation we we hire so we we've had
2:12 we've had a bunch of labs and then we
2:14 have Champ whatever the hell he is and
2:17 uh we we have always had poorly behaved
2:20 dogs and so when we got our first pair
2:22 of labs they they were particularly
2:25 neurotic I think they were a little
2:26 inbred little cousin situation going on
2:29 there and uh we hired a a dog trainer
2:34 and I'm like this is awesome it's gonna
2:35 get all better now and then what I
2:38 learned from the dog trainer is that
2:40 they don't train the dog they train the
2:43 people not to be horrible dog
2:46 owners and I realized oh we're always
2:50 going to have horrible dogs for the rest
2:53 of our lives we just we just don't have
2:56 the
2:57 Constitution to be good dog owners so we
3:00 have shitty
3:02 dogs but they love us
3:09 so oh man all right my favorite dog is a
3:13 German shepherd very nice I had a
3:16 rotweiler when I was a kid didn't like
3:17 that
3:21 dog he almost bit my little toe off the
3:24 little [ __ ] well the big
3:26 [ __ ] those things are
3:29 intimidating
3:32 all right all right everybody welcome
3:35 welcome welcome this is the AI learning
3:38 lab I just saw a uh an interesting
3:42 little news
3:44 item that
3:46 apparently Google was going to surprise
3:50 release Gemini next
3:53 week so this is this is the supposed GPT
3:57 4 killer the
3:59 multimodal Google GPT 4 killer and the
4:05 CEO delayed it till till
4:08 January
4:11 [Laughter]
4:14 so I didn't realize this I saw in the
4:16 article Google has 200,000 employees I
4:20 had no [ __ ] idea they were that big
4:22 and um open AI has
4:25 800 and there's a there's a uh a tweet
4:28 from Greg Brockman
4:30 today that for for me this kind of
4:33 rubbed salt in the wound of
4:36 Google wait let me see if I can find it
4:39 um
4:42 [Music]
4:46 Greg Brockman Greg
4:53 Brockman so he
4:56 goes six hours ago Greg Greg Brockman
5:01 tweets first week back at work complete
5:05 52 team meetings and one on oneon-one so
5:08 far biggest takeaway is people across
5:11 every single function are united in
5:13 their thinking renewed focus and energy
5:17 desire to figure out what they can
5:18 uniquely do to further the mission we
5:20 are back stronger it's like they were
5:24 firing on all cylinders before so so if
5:28 they've tightened things up now and I I
5:31 I mean you know this is what happened
5:33 was one of those events that either
5:34 blows the company the [ __ ] up or it it
5:37 takes it to the next level so uh they're
5:39 going to be an Unstoppable Force for
5:41 some amount of time that that that kind
5:43 of that kind of cultural uh blast is
5:47 something also Greg Brockman posted a
5:50 picture of him and Ilia
5:52 sver um who in the big announcement
5:55 three days ago or two days ago um was
5:59 not announced other than we're trying to
6:01 figure out how to work together so maybe
6:03 they're working some stuff out so anyway
6:06 I thought I thought that was that was
6:08 some pretty clever clever
6:12 uh Google
6:16 trolling by the way you tried to break
6:18 us it got us
6:23 stronger all right so the way we do this
6:26 here if you have questions about AI pop
6:28 them in the comments below thank thank
6:30 you very much be Not Afraid worker B
6:32 good to see you
6:34 um when I was going live couple of
6:37 things I did I'm starting to add the
6:40 date to the name of the live
6:42 because it took me what have I been
6:46 doing this for six
6:48 months what I realized is if the name of
6:52 the live is always the same when I
6:55 download them from Tik Tok to put up to
6:57 YouTube they all have the same file name
7:00 so I have to go in and manually match
7:01 them up and rename them and it just
7:04 takes forever it's just a stupid
7:06 administrative thing uh but I figured
7:08 that out so that was good and then
7:10 there's also some Tik Tock live Aon
7:13 going on so I joined that but if it does
7:16 anything annoying just it ignore it if
7:19 there's something fun it does cool I
7:22 think it's they wanted me to get a
7:23 certain amount of diamonds or
7:25 or uh celebration like party hats or
7:29 something I I I don't [ __ ] know I
7:31 don't care um but but if it gets
7:33 annoying just ignore that [ __ ] um I'm
7:36 supposed to I'm supposed to Hawk it up
7:39 let's have another battle with
7:41 DAR let's not
7:45 um oh I forgot to say what we're doing
7:48 here so so what we do here so if you
7:50 have questions ask them in the comments
7:53 below
7:54 um we can we can look at some chat GPT
7:58 stuff uh we can look at other tools um
8:01 there's there's a lot of activity in the
8:03 video space lately um that seems to be a
8:06 newly a
8:08 newly bubbling primordial soup that is
8:11 starting to spit out new
8:14 Innovations um every time I've shown
8:17 people the sunno music generator thing
8:21 their minds get blown so ass play with
8:23 that write a song or two I don't know if
8:25 Jim Ross is in here but Jim Ross
8:27 actually got a a business contract with
8:29 that thing I wrote a big sappy thank you
8:32 to open AI oh nice oh I saw that today
8:35 lunar stick yeah that was a really
8:36 nicely written piece congrats on that
8:38 that was a good piece um yeah was really
8:41 cool yeah she if I may paraphrase your
8:45 piece it was basically you know it's
8:47 it's unlocked you know large swaths of
8:49 your brain that that have gone largely
8:51 dormant and Chachi PT is just you know
8:54 reignited some [ __ ] I feel the exact
8:56 same way I I feel like yeah it's just so
9:01 it's just so funny it's the
9:03 the like the more I do this stuff the
9:06 more clear it becomes this is this is as
9:09 it stands right now a force of good in
9:12 the world um and and it really is around
9:16 unlocking the creative brain Kyle do you
9:18 have access to music LM on the Google AI
9:21 Test Kitchen I don't think
9:25 so what's the Pate tell me the uh the
9:28 URL I can I can look it up hang
9:31 on
9:34 Google Google AI Test
9:44 Kitchen see I don't think I do because
9:48 um I have work group work pad work Group
9:53 whatever it is the one where I pay for
9:56 it so if you pay for it for your company
9:59 they don't let you play with the stuff
10:01 try now although wait wait for it sign
10:05 up for AI Test Kitchen I'm looking now
10:08 can you can you review the custom
10:10 instruction settings on chat GPT
10:12 absolutely that shouldn't matter okay
10:14 cool that's great music LM is an
10:17 experimental technology that allows you
10:19 to generate your own synthetic music for
10:22 inspiration very cool I'm getting a
10:25 spinning wheel right now but we'll go
10:28 look at it
10:31 and then what else
10:34 um the uh the AI Salon uh Community is
10:39 really ramping up we're we're
10:41 approaching 300 members of that we
10:43 haven't even officially launched it yet
10:45 I've been Hawking it in here but and you
10:49 know cuz you all are swell it's been
10:50 it's been growing the community which is
10:52 cool but there's a lot of really nice
10:54 activity in there and it's not just it's
10:56 not just uh AI learning Lab live related
11:00 so that's super cool
11:02 um I know isn't that cool cuno I know
11:06 it's good it's good it's good I don't
11:08 know what's happening with this test
11:11 kitchen pet it is not
11:14 loading I say get
11:18 started oh you know what I'm in I am in
11:21 a Chromium browser but I am not in
11:25 Chrome hang on let me try Chrome let me
11:27 see if those wacky engineers at Google
11:30 are doing Shenanigans with
11:34 competitive
11:35 browsers that's one of my pet peeves is
11:39 cuz Apple was Notorious at that they'
11:41 just [ __ ] with Microsoft and Microsoft
11:43 would [ __ ] with them and it's like yeah
11:45 we're the consumers and we're getting
11:46 screwed here
11:49 people just make our [ __ ] work well if
11:53 you had if you use Safari everything
11:55 would be
11:56 fine ah look see look at that
12:00 there might be a wait list for it so you
12:01 may not be able to my monitor is too
12:06 bright get
12:09 started sign in
12:11 now get
12:14 started now I'm getting the same [ __ ]
12:16 I'm just getting the spinning the spinny
12:19 wheel of
12:21 death I got nothing I introduced myself
12:24 on the a AI salon and not even the AI
12:27 bot said welcome Anthony yance is
12:29 unacceptable let's go change that right
12:31 now all right Irregulars you know what
12:34 to do go find Anthony yance in the
12:37 Welcome
12:38 Wagon and give him a warm
12:43 welcome I wonder if I can search for it
12:45 I bet I can we're brand new here Anthony
12:47 so we're still we're still figuring it
12:49 out but I'm going to go find your intro
12:52 and make you
12:56 Anthony REM maze
13:00 Baron are you one of those Anthony
13:03 yance what did you sign in as
13:11 yance I got
13:13 nothing did you sign in with something
13:22 different Allan
13:25 JP I've been trying to go through and at
13:27 least heart things
13:33 well if any of the Irregulars can find
13:35 Anthony ants in there I I don't see you
13:37 in there hang on are you sure you hit
13:40 submit when when you made your intro I I
13:43 am I am guilty of of not hitting the
13:46 submit button many times in my life um
13:50 let's
13:53 see uh uh uh the
13:57 feed sir
14:00 search all spaces
14:05 Anthony I got Anthony REM Anthony Mae
14:09 and Anthony Baron are the only anthonies
14:12 I see in the community did you do it in
14:14 Discord ah I bet you did it in Discord
14:19 but he said AI oh he said AI Salon I bet
14:22 you're right hang on
14:24 hold so as of Tuesday we are switching
14:28 from the Discord
14:30 to Mighty Network so I'll show you the
14:32 uh the URL there um see there's there's
14:35 the announcement that says we're moving
14:37 over to the mighty
14:39 networks
14:44 uh let me see let me do a search here
14:47 that's a good call there I think that
14:49 was lunar stick
14:52 yance search for yance Anthony yance the
14:56 real swamp
14:58 drainer husband of five running to
15:01 campaign for State R representative here
15:04 wife and I are realtors hope that gives
15:05 you a bit of who we are awesome do me a
15:08 favor Anthony let's let's first of all
15:11 hi welcome and let's get you over to the
15:14 proper
15:15 Community seem see seems like like
15:18 Anthony went to the uh went to the old
15:21 Community uh I know I I downloaded uh
15:24 Discord for this reason so discord's
15:26 going to stay going to stay live uh to
15:29 do things like mid journey and if you
15:31 want to build Bots and things like that
15:33 but the primary um Community is Shifting
15:36 over to Mighty Network so head on over
15:38 to AIS salon.
15:41 mn. if you'd be so kind if if you're
15:43 still willing to have
15:46 us and anyone else that's interested in
15:48 joining community so what this is this
15:50 is a group that I started the week after
15:53 chaty BT came out as a way to surround
15:56 myself with people that were like
15:58 choosing to be on this AI journey
16:00 together and and so it's um it's a
16:03 really cool bunch of people and so if
16:05 you go to AI salon. mn. there's no cost
16:09 to it we're getting it covered by
16:11 donations and sponsorships and things
16:13 like that so um so it's it's a cool
16:15 group so anyway apologies about the
16:17 confusion that is uh that's that's one
16:19 of the things that's going to happen
16:21 when you make a change like that
16:22 so um so yeah there you go all right now
16:28 if you brand new to chat
16:31 J and you have not played with it go
16:34 there right now go just go to chat.
16:37 open.com and start
16:39 playing um I donate but just paid my
16:42 property taxes sorry about that will we
16:44 still have hi uh history access to the
16:46 AI Salon Discord yes so so I think what
16:49 I'm going to
16:50 do is I'm going to do I'm going to allow
16:55 I'm basically going to make the um the
16:58 channels dormant so you can't post to
17:00 them anymore but the histories will
17:02 still be there I think I'm going to keep
17:04 show and
17:05 tell I'm going to keep General
17:07 discussion live I'm going to Archive
17:09 some of the other discussions um I'm
17:11 going to keep General disc General
17:13 discussion live um any resources Page
17:17 live I'm going to keep all the guilds
17:19 there but they're going to be in readon
17:21 mode so if there's stuff in any of the
17:23 guilds and then um the show and tell
17:27 Will Remain live and mid Journey will
17:30 remain live and then if anyone wants to
17:32 build stuff in there they can ask about
17:34 it all right so so yeah so so it's going
17:37 to be it's not going to be the primary
17:39 place of Engagement but there are some
17:41 things there are some reasons to keep
17:43 the the Discord there that we're going
17:44 to do all right please give us a brief
17:46 on modeling using
17:49 ml using I don't what do you mean
17:53 modeling a brief on modeling using ml
17:56 using chat GPT what what kind kind of
17:59 modeling
18:02 Gazo like like down the runway because I
18:07 did that one night I created a fashion
18:09 line using chat
18:11 GPT um I'm ready for merchandise so I
18:14 can
18:15 recruit you mean swag we're going to
18:17 have schwag we're going to I I think I
18:20 think fingers crossed um we are going to
18:24 have a one-year
18:26 anniversary limited edition t-shirt
18:28 which I'm really excited about it's
18:30 going to be it's only going to be live
18:32 for 48 hours and we're going to use it
18:34 as a way to you know celebrate the year
18:37 but also raise money to cover the
18:39 infrastructure cost for the salon so I'm
18:41 pretty excited about that and then we're
18:42 going to have an ongoing uh merch store
18:44 but but I like the idea of doing like
18:47 limited runs so so you can say I was
18:50 there
18:51 when um so yeah okay so go to that thing
18:55 start playing with it just just tell it
18:57 to write you a poem tell it to I don't
18:59 know make up a recipe using skittles
19:03 it'll do
19:04 [ __ ] so are you wanting to
19:09 see oh PA is asking are you wanting to
19:11 see chat GPT do data analysis ah okay
19:14 yeah if if that's the kind of modeling
19:16 you're talking about I can do that um
19:18 perfect if we get an answer for that I
19:20 will do that hope you have gangster
19:27 sizes it's pretty good I'm sure I'll
19:31 have at least you know at least one
19:34 level of multiple X's may maybe larger
19:39 shipping to Oz I don't know we're doing
19:40 it on Teespring so wherever they
19:42 wherever they
19:44 ship we're doing print print on
19:47 demand yes okay great um okay so someone
19:51 wants to see code interpreter so we'll
19:52 go do code interpreter um trying to
19:55 think if there's anything a thing else I
19:57 want to show first
20:07 um oh I know what I
20:10 forgot you can make Mone
20:12 with all right let's move on shall
20:16 we he just pushed this button and said
20:19 they could make money with chat jpt I
20:21 didn't understand why he did it I I it
20:24 see there were a lot of people in there
20:26 watching I don't know why they stayed in
20:27 there they seem to be in there they were
20:28 talking a lot they were chatting with
20:30 him they were chatting with each other I
20:32 don't understand it should should is
20:35 there something I'm
20:37 missing I don't even know what his
20:39 qualifications are I don't think he has
20:41 any as far as I can tell he doesn't have
20:43 any
20:44 qualifications
20:46 anyway so how long did you spend there
20:49 oh I of three hours okay do you know the
20:52 best place to learn how to create gpts
20:55 ourselves you are
20:57 there
20:59 I would say join the AI
21:01 Salon join this group don't go to the
21:05 Discord we found out that's that's
21:07 that's a disappointing that's a
21:08 disappointing entry into the community
21:11 because it is slowly being phased out in
21:13 favor of this platform so go there um so
21:20 I'm going to go play and chat GPT right
21:23 now we're going to start out doing some
21:25 code interpreter stuff and then what I'm
21:27 going to do is I'm going to flip over
21:29 and I'll make a GPT we'll make we'll
21:31 make a GPT together you tell me tell me
21:34 what you want to do one about and then I
21:36 will embarrass myself and try to figure
21:39 it out on the Fly that's my that's
21:41 that's my job here is I actually I just
21:44 realized why this is a popular Channel
21:47 this is like a blood sport where it's
21:49 like yeah let's let's let's watch Kyle
21:52 faceplant as he tries to demo [ __ ] live
21:55 I you guys are you're more cruel than I
21:58 realized oh I get it now it all it's all
22:02 coming
22:06 clear lordy lordy
22:09 lordy oh man alive all right let's
22:17 see Be not Afraid hello Sher B Banks was
22:21 shaking thank you mods for doing what
22:23 you do on a nightly basis appreciate all
22:27 of your work cuz I know it's kind of
22:29 thankless overwhelming work at
22:32 times um all right okay so let's go do
22:37 this let's go jump into chat
22:39 j i I spent a fair amount of the day
22:42 today um speaking of smacking my face
22:45 into a brick wall um trying to take I I
22:49 wrote a prompt that's decently
22:51 successful at taking a long long
22:55 transcript generating it into a YouTube
22:59 title a description with a bunch of you
23:02 know
23:04 rules and a series of chapters but it
23:08 but the chapter thing in particular is
23:10 just totally inconsistent and sometimes
23:13 brilliant and sometimes awful and so I
23:16 tried to turn it into a chat GPT today
23:18 and I actually rate limited myself and I
23:20 I just never I never got it so I'm a
23:22 little I'm a little bitter on chat G on
23:25 the gpts right now cuz it's just it's
23:29 it's like it's so close but it's just
23:31 it's getting it I'm telling you man
23:33 these AIS right now you can get them to
23:36 80% quick which is really impressive but
23:39 then if you try to get them to a 100 is
23:42 it frustrating speaking of face plants I
23:44 created a GPT for the M for MMA
23:47 Enthusiast lunar stick you're on fire
23:49 with these things you need like a whole
23:50 channel for yourself you should if you
23:52 haven't done it seriously you should put
23:55 together like a a page of just you know
23:59 icons with all your different um all all
24:01 your different things I'll I'll promote
24:03 it we we'll demo it on here we we'll go
24:05 we'll drive some traffic your way that's
24:07 super cool she's basically taken a box
24:09 full of index cards of ideas she's had
24:11 over the past couple of decades and
24:15 resurrecting them in in uh in the form
24:18 of gpts okay so we're going to hop over
24:21 to GPT 4 oh I might need to
24:25 quit do I need to quit Edge probably so
24:30 why are you using Edge I know I
24:34 know yeah hang on we can probably do
24:37 this over
24:39 in
24:40 Chrome chat JP chat.
24:45 open.com all right there you
24:48 are all
24:51 right okay so if you haven't seen chat
24:56 GPT before this is it there's two modes
24:58 to it there's GPT 3.5 which is the free
25:02 version and there's GPT 4 the problem
25:05 with gp4 right now is if you don't have
25:08 it you can't get it they're not giving
25:10 out new uh subscriptions to it um
25:15 because they're they're overwhelmed with
25:16 how much people using it that said it
25:19 doesn't mean chat GPT 3.5 isn't good so
25:21 if you've never seen it I'll just do a
25:23 little bit of a little bit of GPT in
25:25 right now that'll that that that should
25:29 should tickle your Nethers um so let's
25:33 see let's we'll do
25:35 uh write a country
25:40 song about changing the oil in my pickup
25:45 truck when I
25:47 spot the love of my
25:54 life so 3.5 is fast um it does this
25:58 stuff pretty fast and it's it's really
26:00 quite good uh I was under my pickup
26:02 truck sunny day changing the oil just
26:04 passing the time away had a wrench in my
26:07 rag covered in grease little did I know
26:09 my world was about to find some peace
26:11 all right not great but not bad I was
26:14 changed my oil when I saw your smile
26:16 underneath that pickup truck for a
26:18 little while all right whatever okay um
26:21 so but now I could say I could say um
26:25 let's see um what are uh let's see give
26:30 me a table
26:35 comparing um
26:38 successful
26:41 lyrics for country
26:45 songs versus pop
26:49 songs versus
26:53 WP and so this should give us a nice
26:57 table
26:59 of so love country song heartfelt
27:01 sentimental pop song romantic often
27:04 catchy rap song relationships passion
27:08 and so you know it just quickly we we'll
27:11 do stuff like that you could also say
27:14 um I want to create a song lyric
27:18 generator that first asks the
27:22 user to choose one of those three Styles
27:26 and input a
27:28 subject and then based on the rules of
27:31 songwriting you've outlined above it
27:34 generates a set of
27:37 lyrics
27:41 and chat jpt will just go off and write
27:44 that for you and explain to you how to
27:46 deal with it and how to run it and you
27:50 could paste code in here and you could
27:52 say document this for me or find the bug
27:55 in it things like that so uh if you
27:57 haven't played with chat GPT before just
27:59 start doing these are what I call parlor
28:01 tricks just start throwing [ __ ] at the
28:04 wall right throw pasta at the wall try
28:06 to stump it tell it to educate you about
28:09 something you know deeply and you you'll
28:11 see you you'll pretty quickly come come
28:13 to realize oh this this there's
28:16 something here like it's kind of like
28:18 playing with a video game when you start
28:20 first start playing with it um and then
28:23 at some point you may have what I call
28:24 your Kevin McAllister moment that moment
28:28 you're like oh my sir I've looked for
28:31 screenplay to St storyboard on
28:32 futurepedia and Adobe isn't there yet
28:38 thoughts I've looked for screenplay to
28:41 storyboard on futurepedia and Adobe
28:44 isn't there yet thoughts uh yeah that's
28:47 not a surprise to me because
28:52 Adobe one of the things that adobe has
28:55 not done super well over the years is is
28:58 cross the beams of their products one of
29:01 the things that large language models do
29:03 really well is sort of mesh together
29:06 different modes Adobe is very siloed the
29:09 apps are very siloed even though they've
29:10 got bridge to to pass things between
29:13 them they are very distinct
29:16 Feature Feature Suites um so the fact
29:20 that that because that
29:23 requires
29:25 both storytelling large language models
29:27 with they don't have a ton of that [ __ ]
29:29 like Adobe is sort of like they'll
29:30 assume that you've got the story figured
29:32 out in fact talking to John neck who
29:35 works at Adobe they like they tried
29:38 doing some automation stuff that that
29:40 helps with the storytelling like what
29:41 storyvine does and just didn't it didn't
29:45 fit with the culture of their products
29:46 so um so it doesn't surprise me that
29:49 they're not there yet that said Adobe of
29:52 all the big companies um trying to
29:55 figure this [ __ ] out Adobe really is
29:57 leaning into it so I would I would say
30:01 that storyboarding I it feels like such
30:05 an a niche thing for
30:09 them I'm guess I'm guessing that you'll
30:11 see some other player get there before
30:13 they do and if they're smart they'll
30:16 they'll make it compatible with with uh
30:17 with Adobe stuff anyway okay so let's go
30:21 over to gp4 so so someone I think from
30:24 Japan welcome if you're in Japan
30:29 or even if you're not in Japan but
30:31 you're Japanese or even if you just have
30:33 Japan in your your name welcome um
30:37 wanted to see um chat GPT work do code
30:42 interpreter and so what code interpreter
30:44 is is the ability to do Advanced Data
30:48 analysis um within the context of a
30:52 conversation and it's it's quite
30:54 impressive it's it's a bit of a hidden
30:56 feature if you don't know it's there you
30:57 you don't know it's there if you do you
30:59 do but in chat gp4 you see this little
31:02 paperclip icon down here you can upload
31:05 images to that chat GPT can see images
31:07 you can also upload images to that and
31:09 say things like find the find the
31:15 um top 10 average colors in this image
31:18 and it will actually analyze the data of
31:21 the image versus the the visuals of the
31:23 image so so you can do all sorts of
31:25 stuff but what I'm going to do before I
31:27 upload something here is we're gonna go
31:29 get some data so where I'm gonna go is a
31:31 site called kaggle.com K A GLE
31:50 [Music]
31:56 Ecom and I'm going to go to to a thing
31:58 called data
31:59 sets and then we're going to go get some
32:02 data and I'm going to try to get a
32:03 decently meaty
32:05 One World Cup complete data set I don't
32:09 know what that means
32:11 complete there's a 1 Megabyte New York
32:14 City taxi and limousine Fair data set oh
32:17 these are
32:18 recently Hollywood Hits and flops that's
32:20 only 29k let's get something meaty
32:24 diabetes health
32:25 indicators Colorado Health
32:31 [Music]
32:44 care tomato detection the hell is tomato
32:48 detection those are massive data
32:55 [Music]
32:56 sets
33:04 this might be too big but it's kind of
33:06 interesting so this is 120
33:10 megabyte 1.5 million Uber app Google
33:15 store
33:26 reviews
33:28 okay so it's archive oh archive one okay
33:32 so back to Chad
33:36 j this might be too big for it I think I
33:39 think the limit previously was 100
33:43 megabytes but maybe they upped
33:55 it come on Chad jpt you going to upload
33:59 this or not we want to go get some
34:02 catfish out of that there data all right
34:04 well that's thinking I'm going to go get
34:06 another one that's smaller because I
34:08 have a feeling that just might be
34:10 choking on it it's just choking on it's
34:13 just not doing too good if you know what
34:14 I
34:16 mean all right let me just try to upload
34:20 it again maybe we got to just reset the
34:23 session
34:25 here no it is is not happy all right
34:29 we're going to get
34:30 another we're going to get another set
34:32 hold
34:34 please multiplatform online courses
34:37 let's go get the let's go get the uh New
34:40 York City taxi and limousine Fair data
34:43 set so this is a megabyte not huge
34:48 but we're going to kill that one we're
34:51 going to go get the New York City limo
34:56 stuff
34:58 we might not be doing this at all oh
35:00 there it goes okay it went yay okay so I
35:04 can just say um tell me about the data
35:09 inside this
35:13 archive and so what it says right now is
35:16 analyzing and analy when it says
35:18 analyzing it's got a little drop- down
35:20 menu here this is code interpreter so I
35:22 can I can pop this open and what it's
35:23 actually doing so when code interpreter
35:26 runs it creates what's called a
35:27 container so inside your chat is a
35:29 container that effectively acts like a
35:32 virtual computer that's got its own
35:34 little mini file system and it's got
35:37 access to python libraries and it can
35:39 write and execute python code and all
35:42 within the context of your chat so when
35:45 I asked it to open the zip file and see
35:47 what was in it the first thing it does
35:49 is it uses a library to open a zip file
35:51 writes the code to open the zip file and
35:54 then pulls the stuff out of it so it
35:56 says your archive upload contains two
35:58 CSV files here they are the files appear
36:01 related taxi trip data possibly some
36:04 analysis mean values as suggested would
36:06 you like to extract and explore the
36:08 contents of these files further so
36:10 say sure why don't you look at the data
36:14 determine any interesting Trends within
36:17 it and then help me visualize those
36:20 Trends so I'm giving it a compound
36:23 request here and why I'm doing that code
36:26 interpret
36:28 is is a glimpse
36:31 into how these large language models are
36:33 going to work in the future where
36:35 they're going to be more like agents on
36:36 our behalf right now um chat GPT for the
36:41 most part is not autonomous and this
36:43 isn't quite autonomous but I just tell
36:46 it hey open up that file and go do some
36:48 stuff and so what it does is it goes and
36:50 it does some stuff and then it looks at
36:53 what it got and it tells me about it so
36:55 the data set contains these things da da
36:58 da and then it says let's start by
37:01 visualizing some of these Trends so what
37:02 did it find it found trip durations and
37:05 distances passenger count influence
37:07 payment types timebase Trends so so
37:11 again if you don't know anything about
37:13 data analysis which I do
37:16 not this opens a Pandora's box of
37:19 possibilities if you do know something
37:21 about data analysis you're going to be
37:23 really be able to exploit the crap out
37:25 of this thing uh um so here's a a
37:28 scatter plot on trip distance versus
37:31 fair
37:32 amount and it it looks like the scale is
37:35 off because someone did a thousand trip
37:39 a three a a three mile, trip through
37:43 this chart off so so we could tell it to
37:46 throw out the uh throw out the anomalous
37:49 data there and and you'd probably see
37:51 greater uh influence in distance to to
37:54 things but it is to to get these charts
37:58 like these charts are not in the data it
38:00 is making these charts it's analyzing
38:03 the data it's it's using Matt plot LI or
38:05 whatever that thing's called the library
38:07 in Python to make charts and it's doing
38:11 that for us so so now it says for a
38:14 final analysis we can look into the
38:16 timebase trends of the data including
38:18 specific timestamps these could re
38:20 reveal insights into the busiest hours
38:22 for taxi trips let's create a Time
38:24 series analysis based on pickup time
38:27 times so 4 in the morning this you know
38:31 they say New York never sleeps but
38:33 apparently around 3:30 they're starting
38:36 to nap and then you know them wall
38:39 streeters picking it up early to go hit
38:42 those financial
38:44 markets and then 7 o' is 7 and 8:00 6 6
38:49 and 7 o'clock is the is the peak right
38:52 so when most of America's going to bed
38:55 New York City's going out to dinner
38:58 and then and then they're usually coming
38:59 home around midnight so so that's code
39:04 interpreter I mean it it's just it's
39:06 just nuts nuts nuts nuts nuts nuts nuts
39:08 nuts nuts nuts nuts nut
39:11 nuts um someone else wanted me to show
39:13 something else about chat GPT 4 oh
39:15 making gpts
39:18 okay so what are
39:20 gpts if you are new to chat GPT the
39:24 interface over here on the left this is
39:26 your prompt history so these are the the
39:29 prompts I've done in the
39:31 past um
39:35 why why do I not
39:39 have why do I not have an
39:43 Ayes there they are okay that was
39:47 odd okay so you've got your you've got
39:50 your prompt history down below if if
39:52 you're um if you haven't created one of
39:54 these or used one you might just see an
39:56 an explore button below chat GPT or you
39:59 might see like three or four of these
40:01 things um I've got a lot of them there
40:04 cuz I've made a bunch um but if you
40:06 click on the explore
40:11 button you'll you'll get uh a list of
40:14 gpts and so there's the ones I've
40:16 created there's some recent ones I've
40:18 used and then here's some made by open
40:19 AI so if you haven't made any before
40:21 you'll kind of start with this list
40:23 you'll see do and data analysis and chat
40:26 gbt classic and Santa and all that sort
40:28 of stuff okay at the very tippy top it
40:31 says create a GPT and you're like but
40:33 that's way too complicated for me I
40:35 can't I can't program a robot using
40:39 machine learning algorithms because I
40:42 don't have a degree from Stamford Kyle I
40:44 don't think you understand I'm kind of a
40:46 dumb Dum when it comes to
40:48 computers we've all been
40:51 there well not all of us those of us
40:54 that went to Stanford with with
40:55 mathematics degrees have not been there
40:58 but the rest of us have been there and
41:02 open AI has actually solved that in a
41:04 way that I think is is REM remarkably
41:07 elegant okay so what a CH what oh let me
41:10 go show you what a GPT
41:12 is um I'll show you my favorite one
41:15 which is the first one I ever made and I
41:16 made it live here on the the on the uh
41:20 much against my my my own uh better
41:24 judgment I said I know I'll learn how to
41:26 do this with 100 people watching me so
41:30 so this was the very first one I made uh
41:33 for a lot of people there's a lot of
41:35 people that come to these things night
41:37 after night uh because they got a screw
41:39 loose um they're called the Irregulars
41:41 they're the ones that are probably down
41:43 there you know telling you what's going
41:44 on and things like that they're awesome
41:46 people um the Irregulars will tell you
41:49 this took me maybe five minutes I had
41:52 never built one before I had built some
41:54 Poots before but Poots are actually like
41:59 it's it's a bit more like a a low code
42:02 or no code kind of thing where you still
42:04 have to kind of fill in stuff and you
42:06 have to know what models are they
42:08 they've really fixed things up here with
42:10 chat GPT to make it quite easy so what
42:12 this is this is called Sin City sipper
42:14 it's the uh Vegas bartender with a Twist
42:17 on the classics and then I so I can say
42:19 um I could say
42:22 I'm
42:23 an
42:25 advanced
42:29 mixologist and want a
42:33 fancy new drink for my Brazilian
42:41 steakhouse for the
42:47 holidays and so Sin City sipper
42:50 recognizes um if you're Advanced or new
42:55 and it will give you more simple recipes
42:57 if you're new it'll give you more
42:58 complicated recipes if you're Advanced
43:01 um and then it's the way I wrote it it's
43:04 supposed to give you the recipe first
43:06 and then create an image sometimes it
43:08 just doesn't and it starts with the
43:11 drink first but this GPT generates
43:16 images of the drinks it's telling you
43:17 how to make that thing's actually
43:18 [ __ ] beautiful for the bra Brazilian
43:22 holiday special I've crafted a cocktail
43:24 that's vibrant and festive let me
43:26 introduce you to the karaa holiday Fizz
43:30 that looks [ __ ]
43:31 amazing jeez that's really good 2 ounces
43:36 of aged uh Kaka kachaka I don't even I I
43:41 used to be a bartender I don't know what
43:42 that is fresh lime juice uh Mara
43:47 Maraka kuya passion fruit syrup
43:51 sparkling wine lime wheel for garnish um
43:55 mint for for
43:57 and so and then it gives you the recipe
43:59 and tells you how to make it and then
44:00 the other thing it does is it tells you
44:03 to go check out story vine.com which is
44:05 my company now my company doesn't have
44:07 anything to do with booze but I just I
44:10 threw a prompt in here that said when
44:13 you give someone a recipe um send them
44:16 to my website and it does it it does it
44:18 brilliantly um so so anyway so that's
44:20 what gpts are they're these
44:23 customized um versions of chat GPT that
44:26 have all of the features of chat GPT
44:28 which you can turn on and off if you
44:30 want depending on your thing um but you
44:32 can just you can you can now share these
44:34 with the world so you could create one
44:37 for your business you could create one
44:39 um for a project you could create one uh
44:42 for a group of friends that I I created
44:44 one the other day for a screenplay I
44:46 wrote 30 years ago and I now have I now
44:49 have a GPT that you can ask about my
44:52 screenplay it'll tell you about it and
44:54 then every time it tells you about a
44:55 Scene It renders
44:57 um an image from the storyboard it's
45:00 it's wild okay so how do you make these
45:02 things it is easier than you think you
45:05 click on create a
45:06 GPT and it brings you to this
45:16 interface okay you got the left hand
45:19 side which is the let's make this [ __ ]
45:21 you got the right hand side here's the
45:22 [ __ ] we
45:24 made easy it's a it's a convention in
45:27 software development and then on the
45:29 let's make this [ __ ] side you've got the
45:32 create Tab and the configure tab if
45:34 you've ever made a pot a bot on PO this
45:39 interface is very familiar there's an
45:41 you can upload an image you can name
45:43 your GPT you can write a description
45:45 about it you put in your prompt in the
45:47 case of Po you have to select uh an AI
45:51 model to use you don't have to do that
45:53 here um you can put in some some uh
45:56 pre-written questions um GPT the these
46:01 gpts in particular one of the things
46:03 that makes them very very powerful is
46:07 one of them is this thing called
46:08 knowledge you can upload your own data
46:10 here so so when I did my screenplay one
46:13 I uploaded a PDF of my screenplay here
46:15 and that's why you can interact with my
46:17 screenplay you also have access to all
46:19 of the capabilities of gp4 web browsing
46:23 um image generation and code interpreter
46:26 so what are we doing oh we haven't
46:28 decided what we're doing here code
46:29 interpreter is what I just showed you
46:31 where it can analyze data you can you
46:33 can make that and then there's also
46:35 actions where you can add um GPT you you
46:38 can add apis so you can connect your GPT
46:41 to other
46:42 services so maybe you want to pull
46:47 um an appointment from your Google
46:49 calendar into the GPT as part of the
46:52 thing you can do that by adding actions
46:55 um I'm not going to going to go into
46:56 that cuz it's it's it's a bit
46:59 complicated and I also haven't done it
47:01 enough that I'm comfortable like I'll
47:03 just I'll just screw it up and I don't
47:05 want to confuse you so
47:09 um what's what's really nice about this
47:12 is if you're brand brand new to coding
47:17 if you're like if if if if this feels
47:19 like it's giving you hives already
47:23 because I'm going through this stuff
47:24 like you can turn on and off features
47:26 and you can configure things you don't
47:28 have to do [ __ ] you just go over here to
47:30 the create Tab and this is just chat GPT
47:33 and it's chat GPT that's specialized to
47:36 build you your own GPT so the first
47:40 thing we need to do is we need to come
47:41 up with a concept anybody have a concept
47:45 what do we want to make one of what do
47:47 we want to make one of come on people
47:51 what are we going to make a jbt about
47:55 h Pate I uh wait what was that Pate said
47:59 something there API please API I don't I
48:04 don't I'm not comfortable enough Shan
48:07 the the problem with some of the thing
48:09 like I don't mind falling on my face but
48:11 with something like apis it is so sort
48:14 of geeky troubleshooting uncomfortable
48:18 conversations with relatives perfect
48:19 cuno okay
48:22 so
48:24 um this is
48:26 good see where I immediately went cuno
48:30 cuz I was like do we want to have um
48:32 brows with Bing hooked up on there and
48:34 then I was like wait a minute we could
48:37 we we could create a an a family
48:40 argument starter where it pulls it pulls
48:44 news items of the day and then and then
48:48 gives you gives you arguments that will
48:50 piss off your
48:52 relatives okay I'll let I'll let one of
48:54 you do that one I'll do the other one
48:57 troubleshooting uncomfortable
48:59 conversations with relatives okay
49:01 great I'd like to create a
49:05 GPT that coaches me
49:10 on how to have uncomfortable
49:12 conversations with relatives during the
49:15 holidays
49:17 period that's all I'm going to
49:20 say so now it says updating the GPT so
49:24 what open AI is done here oh and by the
49:25 way the me talking to my computer like
49:28 that that's just a Macintosh Mac OSX
49:31 accessibility thing so I don't have to
49:32 type in front of you um so it's not a
49:35 special thing from chat GPT but you can
49:37 if you're on a Mac you can turn that on
49:38 I think Windows has it too um for this
49:42 GPT how about the name family Harmony
49:45 coach I like that does that sound good
49:46 to you I'm GNA
49:47 say sounds
49:51 great all right oops I just said sounds
49:55 well and so now it's generating the
49:57 profile picture remember I said before
49:59 you could upload a picture I don't know
50:01 if you've noticed but over here
50:02 underneath my picture this GPT is live
50:05 we can actually test this right now how
50:07 do I talk about politics with my uncle
50:10 perfect let's click on that this is one
50:12 of the pre-written questions so I'm
50:13 going to click on
50:15 that there's the there's the thing it
50:17 gave us okay that is absolutely the
50:19 wrong uh the wrong icon here here's what
50:22 we're going to do for the
50:24 icon the icon needs to
50:28 be a
50:35 festively
50:38 overloaded holiday dinner
50:42 table
50:44 with two people screaming at one
50:48 another period the image should focus on
50:51 the two
50:52 faces with their mouths open and veins
50:55 bulging
51:01 all right let's see what this thing
51:04 did uh set a positive tone begin with a
51:07 positive or neutral topic Express
51:10 curiosity listen actively use eye
51:13 statements seek Common
51:18 Ground the family Harmony coach uh let's
51:21 see
51:24 um that's great can you do it in
51:34 color it's so
51:37 good and and and basically this thing is
51:40 live like it's going to be live when I
51:42 share it but that's it like like so now
51:46 it's generating a new profile picture
51:48 for us that's good okay that's perfect
51:51 here's the coloriz holy crap if if you
51:54 haven't seen uh
51:57 um if you haven't
52:00 seen the bear season two of The
52:05 Bear this is It's required
52:08 viewing the episode with Jam Lee Curtis
52:11 where they're having you know dinner at
52:14 dinner at Mom's Jam Lee Curtis's Mom it
52:19 is it's it's one of the most disturbing
52:23 and Powerful I think it's a 45 minute
52:27 episode but it's like it's like 30
52:29 minutes
52:30 of just PTSD inducing it's it's one of
52:34 the most
52:35 remarkable episodes of Television I've
52:38 ever seen um but it's pretty it's pretty
52:41 much like that
52:47 so it's really good um okay now that we
52:51 have a a specific profile set so so you
52:54 go over here now so family Harmony coach
52:58 guides you through holiday family talks
52:59 with empathy what's the best way to
53:02 handle questions about my what my
53:06 relationship status oh that's a good
53:12 one all right so so we're gonna we're
53:15 gonna modify this a little
53:19 bit okay so here's here's what we're
53:21 going to do we're going to
53:24 go
53:29 I don't want people to have to read a
53:31 lot I want them to be able to interact
53:34 with this in short little breaks in
53:37 between
53:39 arguments with their crazy uncle
53:44 period keep the points of advice limited
53:47 to three and make them very
53:51 short for each set of three bullet
53:54 points I want you to create an image
53:57 that represents the
53:59 harmony you are intending to create for
54:02 the specific
54:04 situation that was asked
54:07 about period give the advice first and
54:10 then generate the
54:14 image okay so we'll let it we'll let it
54:17 update the GPT and and this is real time
54:20 like as it updates things there it it
54:22 it'll fix it over here so it says
54:24 updating here
54:26 so it's writing all the prompts in the
54:28 background it's doing all that for you
54:30 and now it's it's just done and now it
54:32 says quick clear advice for holiday
54:34 family talks with images that's really
54:37 boring so we're going to go fix that
54:38 okay so now that that's written I can
54:41 flip into the configure mode and all of
54:43 this stuff has been written for
54:45 us so family Harmony Coach quick clear
54:49 advice for for Holiday talk select quick
54:52 quick clear advice for
54:59 defusing um holiday
55:02 [Laughter]
55:10 meltdowns all right great and then I
55:15 don't we don't need it connected to the
55:17 internet that might just confuse things
55:19 so we'll just turn that off and we'll
55:21 have Dolly code interpreter on and then
55:23 we'll try one of these um how to handle
55:27 political
55:28 debates so family Harmony coach should
55:31 now just give us three set boundaries
55:32 early redirect conversations practive
55:35 active listening now let's create a nice
55:38 little image that we can show to Uncle
55:39 Bob and tell him to stop being an
55:41 [ __ ] all right and it's it's erroring
55:44 out I chat chbt has been really uh wonky
55:46 lately so assume it created something
55:48 funny there okay um so now I'm going to
55:52 save this and so how you save these
55:54 things is in the up upper right hand
55:56 corner even though all of the interface
55:58 to this point has been in the bottom of
55:59 the windows now all this how you save it
56:03 and share it is in the top of the window
56:05 so I'm going to share this publicly so
56:08 you can do it just for yourself only
56:10 someone that has a link and then
56:12 public right now because there's not a
56:14 GPT store the there's no difference that
56:17 I can tell between only people with the
56:19 link can see this and public um but when
56:23 the GPT store launches which is probably
56:25 going to be January now it sounds like
56:30 um if you have it selected as public and
56:33 you get accepted as a GPT developer and
56:36 whatever their program is going to be um
56:39 then then you'll be live in the store so
56:46 confirm and it's live and so what I'm
56:49 going to do now is I'm going to go grab
56:51 that
56:51 URL I'm going to head over to
56:57 the AI
57:00 Salon AIS
57:02 salon.
57:04 MN mn.
57:13 Co am I in yeah I'm in and then so this
57:16 is the uh this is the AI Salon if you go
57:20 here so if you decide to join which
57:22 which you should go to the about the
57:24 salon first talks about who we are why
57:26 we do this how we explore and it also
57:28 talks about our
57:30 values curiosity generosity like teach
57:33 what you're learning um like be swell
57:36 explore be adventurous be collaborative
57:39 if if someone's trying to solve
57:40 something jump in with them and and work
57:42 on it together be inclusive um be
57:44 empathetic um so if those values
57:46 resonate with you and that sounds good
57:48 then join us but within here there's a
57:52 little section called showand tell so I
57:54 am going to go show and tell this and
57:56 you can all go play with it right now um
57:58 so I'm going to say this
58:02 is the what was it
58:05 called wait um the family Harmony coach
58:10 wait that shouldn't that be holiday the
58:12 it's the holiday Harmony coach come on
58:16 people family Harmony
58:20 coach let's go to edit come on take me
58:23 there oh come on
58:27 GPT
58:30 please it could be my
58:35 internet how to wear your crazy uncle
58:37 down and put you in his
58:42 will okay if I can get this thing to
58:44 edit I'll add that one cuno that's a
58:46 good one can I go
58:50 back oh this is just not happy at
58:53 all family Harmony
58:56 coach edit okay here we go so wait let's
59:00 let's go so this is now going to be
59:03 holiday holiday Harmony
59:06 coach uh what was the third one
59:12 um uh how to
59:16 wear
59:19 your uh wait uh wait how
59:24 to
59:25 wear my uncle down and get in his
59:35 will that's [ __ ]
59:37 genius okay so wait
59:41 uh let's see quick
59:44 advice for def oh I spelled diffusing
59:48 wrong for
59:50 diffusing
59:53 holiday family meltdowns okay perfect
59:57 okay so we're going to update it we're
1:00:00 going to make it public we're going to
1:00:02 confirm
1:00:04 that and we're going to say view view
1:00:06 our GPT there it is we're going to go
1:00:08 copy the URL boom we're going to go over
1:00:12 to Mighty networks to show and tell this
1:00:15 is
1:00:17 uh let's
1:00:22 see hopefully this will get you out of
1:00:25 some sticky situations this holiday
1:00:27 season comma and might even end you up
1:00:31 in someone's will exclamation
1:00:34 point and then I'm going to paste
1:00:38 in
1:00:40 the
1:00:42 URL and then I'm going to tag it so be a
1:00:47 be a good citizen of the community and
1:00:50 tag your thing so I'm going to tag it at
1:00:52 regulars because that'll say that it was
1:00:54 done during one of these sessions and
1:00:55 I'm going to tag it as a GPT and I'm
1:00:57 going to go
1:01:00 post and that is live so you can now
1:01:04 go to this thing you click on that and
1:01:07 it will fire you off
1:01:09 to the holiday Harmony coach now you
1:01:13 want to see something even really
1:01:15 cooler uh wait one of my gpts is
1:01:18 requiring a privacy policy URL before I
1:01:20 can go public with it uh I assume that
1:01:23 is that because you you are collecting
1:01:26 data in some way
1:01:28 Andrea did you use the the uh action
1:01:31 where you did an API that's got to do
1:01:34 with um
1:01:36 gdpr if you're collecting any sort of
1:01:38 data you have to have a privacy policy
1:01:40 by law so that's why they have that in
1:01:43 there so
1:01:46 um I don't know that it needs to be a
1:01:48 specific privacy policy you could maybe
1:01:51 just go point to someone else's privacy
1:01:53 policy that's not legal by the way like
1:01:56 it's not legally binding uh but I I
1:01:59 don't know I don't know uh what exactly
1:02:02 you need to do when I did one of those I
1:02:04 did it to my storyvine privacy policy
1:02:06 because we have one but if you don't
1:02:08 have one the other thing you could do if
1:02:10 you've got a website of any kind is you
1:02:12 could just put up a page that is a
1:02:13 privacy policy have chat GPT WR you one
1:02:17 and then just say I want to I want a
1:02:18 boilerplate privacy policy that meets
1:02:21 gdpr
1:02:23 guidelines uh uh uh uh uh or see a
1:02:27 lawyer don't take legal advice from an
1:02:29 ex
1:02:30 [Laughter]
1:02:41 actor sorry meds don't work well at
1:02:45 night you guys are funny um what was I
1:02:49 going to do oh I was going to show you
1:02:50 something wasn't I hold please welcome
1:02:53 to chat add
1:02:55 [Music]
1:02:58 okay so I'm going to go to chat GPT I'm
1:03:00 going to the chat GPT app on
1:03:05 my um on my
1:03:09 iPad 18
1:03:18 more where is my new
1:03:23 thing
1:03:34 it's not in there yet it's hang on my
1:03:37 let me force quit my chat GPT app hold
1:03:41 please I'll be right with
1:03:53 you
1:03:55 [Music]
1:04:04 okay
1:04:08 so I'm in chat
1:04:11 GPT there's my holiday Harmony coach
1:04:15 right
1:04:18 um in the chat GPT app so when you get
1:04:21 the chat GPT app one caveat
1:04:25 make sure you get it make sure that the
1:04:28 company that produces it is open AI
1:04:31 there's a bunch of Faker apps that use
1:04:33 almost identical icons and apple keeps
1:04:37 selling them and I think uh Google Play
1:04:40 does too cuz they're making money on it
1:04:42 they don't
1:04:43 care um it's just bad make sure you get
1:04:48 the one from open AI it's free um in
1:04:51 order to use these you do need to have a
1:04:53 chat GPT 4 thing but but if you have
1:04:55 chat
1:04:56 gp4 you can go to the holiday Harmony
1:04:59 coach and you can put it in conversation
1:05:06 mode
1:05:12 um hey my crazy Aunt keeps trying to get
1:05:16 me to go on a date with her niece and
1:05:20 how do I break it to her that I'm not
1:05:23 interested
1:05:27 one Express appreciation start by
1:05:30 acknowledging your aunt's effort and
1:05:32 intention showing that you appreciate
1:05:34 her thoughtfulness two be honest and
1:05:37 direct politely but clearly state that
1:05:40 you're not interested in dating right
1:05:41 now focusing on your own feelings rather
1:05:44 than any perceived flaws in her niece
1:05:47 three suggest an alternative redirect
1:05:51 the conversation by suggesting you're
1:05:52 open to building a friendship with her
1:05:54 niece that's something you're
1:05:56 comfortable with check out the images in
1:05:59 our conversation
1:06:01 history so so you put it in conversation
1:06:05 mode you just run to the bathroom and
1:06:06 say I'm stuck here what do I say to
1:06:08 Uncle
1:06:11 Bob and it makes the cute little image
1:06:15 so um so that's how you make gpts it's
1:06:19 it's trivially simple to get them up and
1:06:22 running now like I said I spend a good
1:06:24 part of the day today trying to get one
1:06:26 dialed in that's tougher one piece of
1:06:29 advice that conversation mode where I
1:06:32 sort of talked it into
1:06:34 existence use that for initially setting
1:06:37 it up but once you get into fine-tuning
1:06:39 it and tightening it up just stop using
1:06:43 that interface and and flip over to
1:06:45 configure mode and just add new rules to
1:06:47 your
1:06:48 prompt uh because what ends up happening
1:06:51 is when you use that chat interface
1:06:55 every small little refinement you want
1:06:57 to add to it it kind of rewrites the
1:06:59 whole thing and it it like forgets [ __ ]
1:07:02 it's just bad so use that conversational
1:07:04 mode to get it up and running the first
1:07:06 time and then as you refine it use the
1:07:09 The Prompt side the you know mod modify
1:07:12 the prompt on the configure
1:07:16 side well that was
1:07:19 fun the holiday Harmony
1:07:23 coach I'm late huh oh Milo's wife what
1:07:27 are we going to do with you what are we
1:07:28 going to do with
1:07:30 you we're going to keep loving
1:07:34 you that's what we're going to do that's
1:07:37 what we're going to do people we're just
1:07:39 going to keep loving her she's the
1:07:41 sweetest she's so Amelio's wife by the
1:07:44 way she is the one that termed the coin
1:07:47 the Irregulars or coined the term term
1:07:50 term the
1:07:53 coin like got steel
1:07:55 trap coined the term the Irregulars and
1:07:58 she's been here pretty much since the
1:08:00 beginning of these things so so you lay
1:08:03 off Amelio's
1:08:06 wife right now I I can pick on her no
1:08:08 one else can she's the
1:08:13 best all
1:08:15 right fantastic but there's all sorts of
1:08:18 cool oh maybe that's because uh so the
1:08:21 when when you're doing the hearty Hearts
1:08:22 I'm seeing all sorts of new shapes here
1:08:24 is that part of the thing that we're in
1:08:26 the I I signed us up for some sort
1:08:30 of some live Festival so I guess I guess
1:08:35 we've got fancy well those are awesome
1:08:39 look
1:08:40 see Tik Tok knows how to make us happy
1:08:43 they give us little stickers it's
1:08:46 fantastic all right what do we want to
1:08:48 do next what time is it what hey hey
1:08:51 lero what time is it Eddie what time
1:08:54 time why you wash your Dicky in the
1:09:00 thing GC this is
1:09:03 incredible I took an edible let's do run
1:09:06 more Runway TV all right let's go do it
1:09:09 that's awesome cuno you're the best 8K
1:09:12 already oh awesome oh these are the
1:09:13 likes okay so listen one of the things
1:09:16 one of the things that the Irregulars uh
1:09:18 sometimes
1:09:19 do uh when they get bored with me which
1:09:23 you can imagine is pretty easy to do
1:09:26 they start having games amongst
1:09:28 themselves and one of the things they
1:09:30 they have done the the the highest I
1:09:33 think they got to was
1:09:34 123,000 likes over the course of like a
1:09:37 two and a half hour live um they all got
1:09:39 carpal tunnel and they couldn't really
1:09:41 talk or participate for two weeks after
1:09:43 that but it was very cute but they went
1:09:46 nuts so if you want to have a little fun
1:09:47 now that we got special stickers for for
1:09:50 our lives feel free to Tap Away um I I
1:09:54 do have a theory that some of the
1:09:56 Irregulars bought autoc screen tappers
1:09:58 from Amazon and and so they're cheating
1:10:01 which is not fair to the rest of the
1:10:04 people using their
1:10:06 fanges um yeah so all right the movie
1:10:09 site you mentioned yesterday show again
1:10:11 okay let's go look at that cuno took
1:10:14 some edibles let's let's go trip her out
1:10:16 so so I'll give you the I'll give you
1:10:18 the the backstory on why I think that
1:10:21 what I'm about to show you is
1:10:22 interesting check out show and tell Kyle
1:10:24 okay hang
1:10:38 on what just happened oh I hit the wrong
1:10:44 button Sherry
1:10:47 [Laughter]
1:10:49 Banks uncomfortable dinner talk
1:10:53 radio
1:10:55 Ted looks at life all all all bringing
1:10:58 bringing things to life in real time
1:11:00 well so I'll tell you something that's
1:11:02 I'll tell you something that's actually
1:11:04 one of the things that I'm really
1:11:06 excited about recently I'm I'm I'm going
1:11:08 to talk about this Runway TV thing
1:11:10 there's something about what's
1:11:13 happened in this community with these
1:11:16 lives
1:11:21 where it's it's like a jam session
1:11:24 and so I just ramble incoherently and
1:11:27 show [ __ ] right and then sometimes
1:11:29 people just go build [ __ ] or sometimes
1:11:31 they take something stupid I say or
1:11:33 something stupid I'm working on and in
1:11:35 in real time they're reacting to it and
1:11:38 generating stuff like this and sharing
1:11:40 it
1:11:42 so there's kind of a new kind of
1:11:45 entertainment
1:11:46 happening where it's like a radio
1:11:49 call-in show except everyone gets to
1:11:51 participate and and you get to
1:11:54 collaborate and make fun and I there
1:11:56 there's just something there's something
1:11:58 wild about it that I It just I find it
1:12:01 really inspiring and and kind of like
1:12:05 it's completely
1:12:06 unexpected somebody also did the this
1:12:09 was for for poor old who was it Anthony
1:12:13 before that
1:12:15 uh well introduced himself on Discord
1:12:19 and it was all shut down and nobody
1:12:21 welcomed him I feel really bad about
1:12:23 that
1:12:24 that kind of
1:12:29 sucks this is Drew Drew lives in
1:12:31 Australia so he can he can mow while
1:12:35 while listening to the nonsense here oh
1:12:38 that one's cool who did that Andrea look
1:12:40 at that that's
1:12:43 cool beautiful Pro tip enable voice
1:12:46 controls on your phone and say
1:12:49 like every 30 seconds to dis to to oh to
1:12:53 oh wait wait to deploy autotap oh that's
1:12:56 really good that's really good
1:13:01 Pate who did that one daisy look at
1:13:05 that is that oh is that the mar that's
1:13:07 the mara Lago is that the mara Lago
1:13:10 gingerbread
1:13:13 house for the
1:13:17 family right before that
1:13:20 scene We Are The Real Entertainment
1:13:22 exactly Kyle's just the the hold music
1:13:25 that's ex that's what that's what's
1:13:27 happening that's what's happening
1:13:29 forever hooked it it it
1:13:33 is it's like it's like there's this
1:13:37 whole separate entertainment thing
1:13:39 happening independent of the live and
1:13:42 it's [ __ ] incredible um by the way we
1:13:45 just went from like 8,000 to 15,000
1:13:48 likes so so careful of your car carple
1:13:52 tunnel um okay
1:13:55 um I'm going to show you runway. TV um
1:13:59 and we're going to go look at some stuff
1:14:00 cuz because it's
1:14:02 fun oh hey misters Cory Sandler sent me
1:14:05 through awesome welcome welcome welcome
1:14:07 um she's awesome she makes amazing
1:14:09 Pottery she makes good videos she she's
1:14:12 good Tik Tock a good Elder Tik tocker
1:14:14 like myself the Gen X Gen X Tik Tock
1:14:17 we're doing all right we're doing all
1:14:20 right we don't let him [ __ ] with
1:14:22 us
1:14:24 [Music]
1:14:26 all right um Runway Runway
1:14:29 TV I'm going to go show it to you so
1:14:31 runwayml if you don't know is a text to
1:14:34 video
1:14:36 technology I think it's runwayml is it
1:14:38 Runway ml.com hold let me tell you let
1:14:41 me not run you down a wrong Rabbit Hole
1:14:43 there send you off to some godforsaken
1:14:46 site that eats all your
1:14:49 passwords
1:14:51 Runway
1:14:52 ml
1:14:54 is it me or is it a
1:15:02 hold it's
1:15:05 doc Runway ml.com
1:15:09 okay and
1:15:11 so you can just tell it to make a video
1:15:13 it'll make a video you can generate an
1:15:15 image in mid Journey or in doly or in
1:15:18 stable diffusion you can use that image
1:15:20 as a starting point to make a video it
1:15:23 makes 4 second Clips with no
1:15:27 sound and then you can extend them up to
1:15:29 like 18 seconds so you can make 18
1:15:31 second clips and as you can imagine um
1:15:35 they're weird and janky but they're
1:15:38 getting
1:15:39 better and what's happening is creative
1:15:42 people are stting starting to take these
1:15:45 little janky pieces of video stuff and
1:15:48 actually actually do some real
1:15:50 storytelling with them so runway
1:15:52 brilliantly in my my opinion has created
1:15:55 this thing called runway.
1:15:58 TV and it reminds me of the early days
1:16:02 of MTV so in
1:16:05 1981 this thing
1:16:17 launched should Place something here in
1:16:19 a
1:16:22 second
1:16:31 [Music]
1:16:36 Landing sequence
1:16:38 [Music]
1:16:52 initiated
1:16:55 [Music]
1:17:00 [Music]
1:17:04 [Laughter]
1:17:07 [Music]
1:17:08 [Laughter]
1:17:11 [Music]
1:17:14 aw cuno hang in there it's going to be
1:17:18 [Laughter]
1:17:22 okay
1:17:26 [Music]
1:17:39 [Applause]
1:17:39 [Music]
1:17:52 a
1:17:54 [Music]
1:17:56 all right I I'll do another one that
1:17:58 actually pretty
1:18:03 good so I'm just I'm just hitting the
1:18:05 random button right
1:18:09 [Music]
1:18:15 now in the future mortality is a
1:18:22 choice close your
1:18:25 [Music]
1:18:30 eyes I still remember the day I lost
1:18:35 them morality didn't just take them from
1:18:39 me it tore my world
1:18:44 apart when confronted with a choice
1:18:47 between oh yeah that's a good point
1:18:51 Vicky the random button almost got you
1:18:53 banned and to be careful good point so
1:18:55 so um so if you click on the TV Guide
1:18:58 button down here this looks like kind of
1:19:01 late 90s early 2000s kind of video sites
1:19:05 and and when people were trying to
1:19:06 compete with YouTube and it also kind of
1:19:09 looks like early Netflix right and so
1:19:13 there's these different channels and
1:19:15 basically this is just a rapper site
1:19:16 sitting on top of YouTube
1:19:19 videos but one of the things one of the
1:19:22 sections they have is Runway Studios so
1:19:25 they've got a group I think they just
1:19:27 raised I don't know they raised a decent
1:19:29 amount of money but they're creating a
1:19:32 studio to create original content and
1:19:34 then they've got these partner programs
1:19:35 Eerie entertainment sci-fi music videos
1:19:38 these are just categories and they're
1:19:40 allowing individuals to submit your
1:19:43 videos now why I think this is
1:19:47 fascinating is that the the videos are
1:19:50 so janky and weird like some of them are
1:19:52 decent but but I think this one is one
1:19:55 like it's just weird right there's not a
1:19:58 lot of
1:20:01 coherence it's got that weird flashy you
1:20:04 know stuff that that you know when you
1:20:06 do the the video animations it gets that
1:20:08 weird flashy
1:20:12 [ __ ] organizations like Netflix and
1:20:15 Paramount Studios and Sony and real
1:20:18 filmmakers and real producers are going
1:20:21 to look at this and many of them are
1:20:23 going to dismiss it and that's what
1:20:26 happened when when MTV came out in 1981
1:20:28 is there there was sort of a 3mon
1:20:31 campaign before it came out where there
1:20:32 were these weird commercials there was
1:20:34 only like I don't know six channels and
1:20:39 and there there oh no it was there were
1:20:41 more than six cuz cable had come out
1:20:43 there were like 12 there were like 12
1:20:45 channels and there were these weird
1:20:47 commercials that said I want my
1:20:49 MTV and then it came out and it was this
1:20:52 weird weird TV station that had music
1:20:57 that you watched and nobody got it
1:20:58 people were like this is weird and then
1:21:00 MTV became MTV and it became a
1:21:02 juggernaut and you know music videos
1:21:04 became the the way you marketed you know
1:21:07 music for for a generation
1:21:11 and um this feels very familiar this
1:21:16 this feels very similar to that to me
1:21:18 where I think a lot of people are going
1:21:21 to dismiss this kind of storytelling as
1:21:24 amateurish and you know yeah but that's
1:21:27 not
1:21:28 Hollywood and what's going to happen
1:21:32 is this shit's going to get better
1:21:35 really fast and there are going to be
1:21:39 filmmakers that are using these tools
1:21:42 that have never done anything with
1:21:44 traditional uh film making before and
1:21:47 they're going to do things differently
1:21:49 and just you you watch there's going to
1:21:51 be there's going to be forms of
1:21:53 entertainment that come out of this like
1:21:55 nothing we've ever seen that that hit
1:21:57 the Zeitgeist and there's a really good
1:22:00 chance that something like Runway TV or
1:22:03 a series of them actually start
1:22:05 competing with you know a Netflix in the
1:22:08 future or with film studios in the
1:22:10 future things like that so anyway so
1:22:12 just go to this thing and check it out
1:22:14 because I think it's the other thing you
1:22:17 should do is go make some flipping films
1:22:20 go make images go to Runway you you have
1:22:23 have to use you know at least I think 4
1:22:25 seconds of Runway footage in your film
1:22:27 some people have shot you know regular
1:22:29 video and then they've got Runway stuff
1:22:31 scattered throughout it some are just
1:22:33 pure Runway stuff go make some [ __ ] and
1:22:36 and submit it because you can you can um
1:22:41 they'll they'll you'll be one of the
1:22:44 filmmakers on early you know the early
1:22:47 equivalent of MTV like I don't know if
1:22:48 this is going to be the winner but this
1:22:50 is the first one I've seen that's really
1:22:52 like leaning into into this as as a new
1:22:55 kind of streaming service which I think
1:22:58 is brilliant and interesting and
1:23:00 fun cool right I Want My
1:23:06 MTV Runway TV
1:23:11 RTV all
1:23:15 right here's another one from Ted this
1:23:18 is the AI learning
1:23:20 Lab at the mall
1:23:26 just I [ __ ] love that I think we need
1:23:28 to I think we need to to take the
1:23:31 learning lot just just keep growing this
1:23:32 thing keep turning the party so that we
1:23:34 actually end up with Mall
1:23:38 locations we we can Resurrect The the
1:23:40 entire Mall industry by uh taking over
1:23:44 defunct retail stores and
1:23:48 malls what do you do here how we just
1:23:50 show up and make pictures of [ __ ]
1:23:52 generally llamas and turtles we don't
1:23:54 know how that happened or why people do
1:23:56 it but they seem to enjoy
1:24:00 it it's
1:24:04 familyfriendly you can go in the back
1:24:06 room for the
1:24:08 adults for the adult
1:24:11 activity oh
1:24:16 man all right fantastic fantastic Bob
1:24:20 tell him what is one are we having a
1:24:21 good time here are we enjoying ourselves
1:24:24 uh sir yes go ahead yes I would like to
1:24:27 inquire um rather uh enjoying what
1:24:31 you're doing here it seems uh
1:24:33 entertaining people seem to be uh
1:24:35 enjoying themselves they're tapping away
1:24:38 I was uh Curious sir uh what are your
1:24:43 qualifications my qualifications ah
1:24:47 yeah as regards to well to running this
1:24:50 channel I you know you seem to be uh
1:24:54 confidently speaking I just wondered if
1:24:56 there was uh uh how would you say
1:24:59 substance to what you do what sir are
1:25:04 your
1:25:06 qualifications
1:25:07 generally when we're interacting with
1:25:09 the public we wear smart
1:25:12 pants all right I can't create gpts in
1:25:15 the mobile version of chat gbt correct
1:25:18 but here's what you can do Edson um go
1:25:21 to Safari or Chrome
1:25:23 go to a browser and go to chat. open.com
1:25:27 just
1:25:28 run um the chat GPT website in a browser
1:25:32 on mobile and then you can create gpts
1:25:35 there hack it's a life
1:25:38 hack from the AI learning
1:25:41 lab but yeah you can't make gpts in the
1:25:45 chat GPT app but you can talk to gpts in
1:25:49 the chat GPT app which is
1:25:51 cool so so I hope that helped I hope
1:25:54 that helped you know what I'm saying
1:25:57 good Lord open AI was dumb yeah I don't
1:26:02 disagree you want to hear something
1:26:04 dumber
1:26:07 Edson in the chat GPT app you can search
1:26:11 your prompt history in the website you
1:26:14 still can't search your web history your
1:26:17 your prompt history it's [ __ ]
1:26:21 Insanity how how long's this how long
1:26:23 has chat G GPT been out let me see if
1:26:26 they added it no you still can't search
1:26:29 your prompt history in chat GPT chat
1:26:32 gpt's
1:26:34 website it's
1:26:36 like I probably call Sam
1:26:41 Alman three four times a day like I
1:26:43 don't want to call him too much so I
1:26:45 just try to keep it to like morning like
1:26:47 around lunch and then right before
1:26:48 dinner and he hasn't picked up yet but
1:26:50 that's it's just because he and I
1:26:52 there's there's a history anyway I call
1:26:54 him three to four times a day and the
1:26:56 first thing I'm going to ask him when he
1:26:58 picks up the phone is why can't I
1:26:59 [ __ ] search my prompt history
1:27:01 Sam and he's like uh I don't know if you
1:27:03 heard but I was just fired and like
1:27:05 rehired and then fired and then rehired
1:27:07 and then fired and then hired by
1:27:09 Microsoft and then rehired sort of kind
1:27:11 of but not really but then this other
1:27:13 guy came in and did you hear about any
1:27:15 of this so he's probably going to be
1:27:17 saying that but then right after that
1:27:20 right after he like gets it off his
1:27:21 chest I'm like okay yeah whatever the
1:27:23 drama you're in the Wall Street Journal
1:27:24 big deal anyway um I'm going to say why
1:27:29 can't I search my prompt history Sam why
1:27:31 why why can we call Ilia does Greg
1:27:34 Brockman have a handle on this do you
1:27:36 have anyone that knows how to create a
1:27:37 search box do do you want to could you
1:27:41 use Google's search thing
1:27:43 box what are gpts okay
1:27:50 sorry is he on is he on sub
1:27:53 substances do you do you think he's on
1:27:55 substances I think he might be on
1:27:56 substances he doesn't seem normal he
1:27:59 doesn't seem quite
1:28:00 right so
1:28:02 gpts are let me Flip Flip back over
1:28:06 here so here's a GPT so this GPT is
1:28:10 called the holiday Harmony
1:28:12 coach so what gpts are is I just missed
1:28:16 a lot but mission accomplished cuz it's
1:28:20 officially Blown Away
1:28:24 what Blown Away about
1:28:26 what what blew you away um
1:28:30 anyway gpts are Bots basically they're
1:28:34 they're customized versions of chat GPT
1:28:36 4 that are prompted to do a specific
1:28:39 thing so what I just did in I don't know
1:28:42 whatever it took me 15 minutes here with
1:28:44 me rambling was I created a GPT called
1:28:47 the holiday Harmony Coach quick advice
1:28:50 for diffusing holiday holiday family
1:28:53 meltdowns and then let's see one of the
1:28:55 pre-prom questions from one of our
1:28:57 lovely IR
1:28:58 regulars uh kavun note was how do how do
1:29:01 I wear down my uncle to get in his will
1:29:05 so so we're going to ask the holiday
1:29:07 Harmony coach that I'm here to promote
1:29:09 Harmony show genuine interest offer help
1:29:13 and support appreciate the
1:29:18 relationship oh now let's create an
1:29:20 image I'm going to say okay
1:29:22 so make the image
1:29:25 already there we go now it's going to
1:29:27 create the image of us getting in the
1:29:30 will he was so he was so not interested
1:29:33 in my AI stuff wait who are you talking
1:29:35 to Source
1:29:47 Camp look she just got into the uncle's
1:29:51 will
1:29:59 oh your cousin oh that's awesome that's
1:30:01 awesome he was not interested in the AI
1:30:04 what what was the stuff that blew him
1:30:06 away um ask GPT to come up with a
1:30:09 clothes concept like members only
1:30:11 jackets it'll do that that I did a I did
1:30:14 a fashion one of these things so anyway
1:30:17 so so that's what the what these things
1:30:19 are is just personalized versions of GPT
1:30:23 but the cool thing is you can share them
1:30:25 so you can send them to other people and
1:30:27 they can use them um you can also do
1:30:30 things like
1:30:33 um you can add your own knowledge to
1:30:36 them so when you create them there's a
1:30:39 button that says knowledge you can
1:30:40 upload files to that and I uploaded um
1:30:45 our 50-page sales document for storyvine
1:30:48 for my the company that I that I'm the
1:30:50 CEO of and we now now have an
1:30:53 interactive sales coach um trained on
1:30:56 our 50-page sales process document so so
1:31:00 they can be quite quite powerful and and
1:31:03 literally take minutes to to spin up to
1:31:07 create one you just tell it to make it
1:31:11 and it does it's pretty it's pretty
1:31:13 insane it's per near
1:31:18 [Music]
1:31:19 insane how many variations of GPT are
1:31:23 available um that's actually a really
1:31:25 good question let's
1:31:27 go let's go talk about that because this
1:31:31 is a confusing space right now so you
1:31:34 create it yeah exactly well you create
1:31:36 it oh someone asked a question before
1:31:38 about how do you find these things so
1:31:40 let me show you a little
1:31:45 hack if you go to
1:31:51 google.com
1:31:53 and you typee in
1:31:59 site
1:32:01 colon
1:32:03 chat.
1:32:06 openen AI wait
1:32:10 open.com
1:32:12 SLG
1:32:14 slash and hit
1:32:17 return this will give you a list of all
1:32:21 of the GP t s that Google has indexed
1:32:24 and it's thousands of them 8,000 of
1:32:27 them um I don't some of I've got I think
1:32:31 I've got one called Dream
1:32:45 Weaver yeah so my Dream Weaver by open
1:32:49 AI or my Dream Weaver is indexed but my
1:32:53 Sin City sipper is
1:32:59 not oh no it is why does it have the
1:33:03 Japanese there that's
1:33:06 weird oh yeah all right it is index now
1:33:09 okay
1:33:11 um so so if you go to site colon chat.
1:33:16 open.com
1:33:18 SLG and then you can just put in a topic
1:33:20 so you could put in you could put in oh
1:33:25 crap where did I go what did I do what
1:33:28 did I do
1:33:43 people well anyway there we are so
1:33:46 here's all
1:33:48 these here's all these gpts so that's a
1:33:51 way you can find the
1:33:52 them okay now someone asked
1:33:57 about like where are these GPT how like
1:34:00 how many gpts are there
1:34:05 so the things called gpts are just part
1:34:09 of open ai's chat GPT which is the
1:34:13 they're sort of the king of the hill
1:34:15 hill right now so chat. open.com is the
1:34:19 website that's where you get GPT chat
1:34:23 GPT 3.5 chat GPT 4 that's where you can
1:34:26 make these things called
1:34:29 gpts there are a fairly large number of
1:34:33 competing companies and competing chat
1:34:36 GPT like things called large language
1:34:39 models
1:34:40 llms and so chat GPT is one of them
1:34:44 there are lots of others of them so what
1:34:46 I'm going to do right now is I'm going
1:34:47 to walk you
1:34:49 through nine of them and there's a
1:34:52 there's a 10th one that I'll talk about
1:34:54 that doesn't sound like it's worth our
1:34:58 it's worth putting on these lists so
1:35:00 what these lists are that I'm showing
1:35:01 you these are kind of the things that
1:35:05 I've that that I feel like just because
1:35:07 I do this all the
1:35:09 time I think are worth you understanding
1:35:13 that these things are available and it's
1:35:15 actually worth you playing with them and
1:35:17 understanding which ones are good at
1:35:19 what some of them are connected to the
1:35:21 internet some some of them are not some
1:35:23 of them are really good at research some
1:35:25 of them are better at
1:35:28 conversation how do you learn that you
1:35:30 just smack your head against the
1:35:32 wall I'm sorry you just get curious and
1:35:36 you methodically try you know tool after
1:35:40 tool so the caveat here is if you're
1:35:44 already overwhelmed with chat GPT and
1:35:46 you're just trying to figure that thing
1:35:48 out just go there this is the market
1:35:50 leader right now and and even chat GPT
1:35:53 3.5 the non-paid version is so powerful
1:35:57 that you could spend the next year
1:35:58 exploring it and not even scratch the
1:36:00 surface of it just the nature of what it
1:36:03 is is it it is such a
1:36:06 deep ocean that there's a lot to explore
1:36:10 there if that that said if you're
1:36:12 starting to get your head around it and
1:36:14 you really want
1:36:15 to
1:36:17 understand what's coming and what's out
1:36:19 there and where the Innovation might
1:36:21 happen that's what the this next little
1:36:24 piece is about
1:36:27 okay um by the way the the other side of
1:36:30 this this is just so chat GPT this is
1:36:32 sort of the consumer front end of chat
1:36:34 GPT there's also a developer backend
1:36:37 where you access their API um if you're
1:36:41 geeky you probably already know about
1:36:43 this but if you if it's something that
1:36:45 you want to you know learn about and
1:36:47 explore you can uh platform. open.com
1:36:50 playground you don't need to go here if
1:36:53 if you're if you're new to this stuff if
1:36:55 you're not an engineer you don't need to
1:36:57 go there you can just go there but
1:36:58 that's what that is
1:37:00 okay so these guys are the king of the
1:37:04 hill right now and
1:37:06 and uh as I said at the beginning of the
1:37:10 I was going to say broadcast I guess
1:37:11 this is a broadcast it's a
1:37:13 narrowcast uh it's a it's a Tik Tok live
1:37:17 it's a live He's live on on a Tik Tok um
1:37:23 Google has a new thing coming out called
1:37:26 Gemini which is apparently a chat gp4
1:37:31 competitor no one has caught up to these
1:37:34 guys yet which is fascinating to me no
1:37:37 one's caught up to these guys so so
1:37:39 they're still King of the Hill um
1:37:42 okay these are the top
1:37:45 four that I would I would pay attention
1:37:47 to if I were you so and and by pay
1:37:49 attention I mean like go use them now
1:37:51 the exception right now grock in the
1:37:53 upper right is the Elon musk's
1:37:58 uh Twitter enabled chat GPT competitor
1:38:03 it was supposed to come out this week I
1:38:05 still don't have access to it so it's
1:38:07 you have to be uh an ex premium
1:38:11 subscriber their $16 a month version and
1:38:15 then you're supposed to get access to
1:38:16 that I don't have access to it yet but I
1:38:18 think they're going to be a player
1:38:20 because it incorporates Twitter data
1:38:22 into the large language model so so you
1:38:25 can get more misinformation faster so
1:38:28 that that's the exception it's not there
1:38:30 yet but the other three are there and
1:38:31 you should go play with them
1:38:34 co-pilot is chat gp4 but it's
1:38:37 Microsoft's version of it so if you go
1:38:40 to copilot
1:38:42 microsoft.com it can it's got Vision it
1:38:45 can see images it can create
1:38:47 images um it can I don't know if it can
1:38:50 do um code interpreter but it's but it's
1:38:54 built on the gp4 back end so if if you
1:38:59 can't get access to chat gp4 because
1:39:02 they've they've sort of locked down new
1:39:04 accounts for now go to co-pilot and that
1:39:07 is gp4 so so that's a good alternative
1:39:12 to chat GPT because it's the same base
1:39:15 model all right then you have Claude
1:39:18 Claude is the closest competitor right
1:39:20 now to chat gp4 it's got a really big
1:39:23 context window meaning you can put in
1:39:25 big prompts and then Bard b.google.r
1:39:29 it's not super great but it it's getting
1:39:32 better all the time in fact one of the
1:39:34 things that Bard has right now is
1:39:36 they've indexed all of YouTube so you
1:39:39 can ask Bard about YouTube videos and it
1:39:41 will uh uh tell you about them so um
1:39:46 it's getting better and when Gemini
1:39:48 comes out um it's supposed to be really
1:39:51 really good so like I said Gemini it
1:39:54 there it's a piece in the
1:39:56 information uh
1:39:59 newsletter that
1:40:00 said Gemini was supposed to seek
1:40:04 surprise launch next
1:40:07 week and and the CEO killed that because
1:40:12 apparently it does really poorly with
1:40:15 non-english languages and for a for a
1:40:18 global company that's probably not a
1:40:20 good look
1:40:22 all right so those are those four now
1:40:23 here's four more um Pi perplexity and po
1:40:27 um are all really interesting place so
1:40:30 Pi has got an app and it's an app that
1:40:32 you can talk
1:40:33 to do you have all the links posted
1:40:36 anywhere wait when you search for gpts
1:40:39 search for your name you have three
1:40:41 indexed
1:40:43 currently do you have all the links
1:40:45 posted anywhere yet yes um yes jrc
1:40:49 so if you if you go to show and tell and
1:40:55 you click on um
1:40:59 gpts I migrated all of the Discord gpts
1:41:02 over but Bradley created a a single
1:41:05 document that's got them all listed in
1:41:06 it all the ones from Discord and then
1:41:08 there's some individual ones in there
1:41:10 but just here you go to the um hang
1:41:16 on you go to show and
1:41:20 tell
1:41:27 what what
1:41:29 oh hang
1:41:34 on you go to show and tell and then go
1:41:37 to categories oh look I want my ROM
1:41:41 MTV I want my chat JP I love it that's
1:41:46 great that's great magic in every
1:41:50 jingle
1:41:51 live with Kyle oh
1:41:55 yeah this is good Sher um so if you're
1:41:59 in show Andel go to categories and then
1:42:01 click on gpts and that'll that'll um
1:42:04 search and then somewhere down
1:42:08 here is this
1:42:13 it
1:42:18 [Music]
1:42:20 no
1:42:31 somewhere in here is a document that's
1:42:33 got all of them in a single document
1:42:35 that came over from Discord that Bradley
1:42:37 put
1:42:38 together can I search for gpts and
1:42:49 Bradley no
1:42:52 I'm still getting to know how to use
1:42:54 this people you think I know everything
1:42:56 I don't know everything I'm just a dumb
1:42:59 dumb oh wait here's
1:43:01 members can I look for members within
1:43:04 there
1:43:05 no I don't know anyway it's somewhere in
1:43:16 [Music]
1:43:19 there gpts
1:43:29 it no anyway that's how you find them
1:43:33 all
1:43:41 right hopefully they get their [ __ ]
1:43:43 together and and actually show these
1:43:45 thing show the uh the thumbnails I think
1:43:47 the ones that are showing thumbnails
1:43:49 people have just uploaded the IM Imes on
1:43:51 their own okay anyway back to back to
1:43:54 where we were so
1:43:58 um so other than the big four you've got
1:44:01 you've got these three you've got Pi so
1:44:03 so
1:44:05 Pi um has an IOS app and it you can put
1:44:09 it into conversation mode just like chat
1:44:11 GPT can go into conversation mode P can
1:44:14 as well and I I really like Pi Pi is
1:44:17 very much designed to be um conver
1:44:21 ational and it's even got a Mode called
1:44:23 support pie that acts kind of like a
1:44:30 therapist I just got in a fight with my
1:44:34 aunt over politics what should I
1:44:44 do translation oh my you just had an
1:44:47 argument with your aunt I'm so sorry to
1:44:49 hear that would you like to tell me what
1:44:50 happened
1:44:52 yes and you don't need to translate to
1:44:54 French anymore we were arguing over
1:45:04 politics got it I'm switching to English
1:45:06 only mode now arguments over politics
1:45:09 can get quite heated I know did you and
1:45:11 your aunt disagree on a specific
1:45:13 political issue or is it more about your
1:45:15 values or ideologies in
1:45:17 general well she's an
1:45:19 idiot
1:45:27 let's see what she does here that's
1:45:30 quite the assessment but I'm guessing
1:45:32 you're not exactly on the same page then
1:45:35 where the two of
1:45:36 you so that's pie um then you have
1:45:39 perplexity perplexity is really good at
1:45:41 research because it brings in citations
1:45:43 to web things so one of the fears of
1:45:46 large language models is that they
1:45:48 hallucinate and they lie and all that
1:45:50 sort of [ __ ] so by the way we're up to
1:45:52 26,000 likes you guys are
1:45:55 beasts uh perplexity is really good for
1:45:57 research po lets you play with lots of
1:45:59 different large language models it's a
1:46:01 single front end for lots of large
1:46:03 language models po also has Bots that
1:46:06 are almost identical to
1:46:08 gpts and they have a they don't really
1:46:11 have a store but they've got a directory
1:46:14 which is I guess a store and you can
1:46:16 monetize them you can you can monetize
1:46:18 your your bots on PO so I think it would
1:46:22 actually be if you're building GPT Bots
1:46:24 or if you don't have access to GPT 4 and
1:46:27 you can't build them you can build them
1:46:29 on you could build a poot and put it
1:46:32 into Mon monetization mode and start
1:46:34 marketing it why not you can also add
1:46:37 knowledge to Poots
1:46:40 so we don't know who's going to be a
1:46:43 winner right now so all of these things
1:46:45 are just kind of these janky tools so
1:46:47 get to learn them and then llama 2 is a
1:46:49 whole world of open source models I
1:46:52 actually just installed hand line it's
1:46:54 just like two commands in the terminal
1:46:57 um and there's explanations on how to do
1:46:59 it but it installed a 4.6 gigabyte file
1:47:03 on my laptop and I now have wait let me
1:47:07 see if I've got it I now have running on
1:47:09 my
1:47:10 laptop
1:47:12 a large language model that sits
1:47:14 somewhere between chat GPT 3.5 and chat
1:47:17 GPT 4 that just runs locally on my
1:47:20 laptop
1:47:21 I don't have to send things off to open
1:47:25 AI um so pretty pretty remarkable
1:47:29 stuff
1:47:30 [Music]
1:47:39 [Music]
1:47:49 um
1:47:55 isn't that annoying when he sings like
1:47:56 that oh it's
1:48:00 terrible I wonder if it's still
1:48:05 running restore all
1:48:14 tabs so I got this thing running this is
1:48:17 called llama pie or no llama file
1:48:22 and and so th this is one of the things
1:48:25 that's
1:48:27 happening
1:48:29 because this this is one of the reasons
1:48:31 we have llamas in a lot of our pictures
1:48:34 meta released this thing called llama 2
1:48:38 which is an A A now open-source large
1:48:41 language
1:48:43 model that anyone can use and anyone can
1:48:45 use
1:48:47 commercially and um so so I now have
1:48:51 this running on my Mac I think it's
1:48:53 still running let me see
1:48:55 um write a poem
1:48:59 about a family
1:49:02 argument that ends up with me getting in
1:49:09 my uncle's
1:49:19 will
1:49:22 so there it goes so that's pretty fast
1:49:25 that's like somewhere between J chat gbt
1:49:27 3.5 and gp4 speed wise I'm on an M1 Mac
1:49:33 laptop not an M1 Pro or Max or anything
1:49:36 like that just an M1 thing and it's
1:49:39 running that fast the argument began
1:49:41 with words so harsh between my uncle and
1:49:43 his son they were both too far
1:49:46 gone um spoke of things of the past of
1:49:50 wrongs done years ago and soon the fight
1:49:53 escalated their voices growing
1:49:56 low so you know not great and and then
1:50:00 apparently it's also got Vision so I can
1:50:03 upload an image to this thing so let me
1:50:06 go find an image go to desktop we'll get
1:50:09 some crazy ass some crazy ass image from
1:50:13 one of the Irregulars there's a good one
1:50:14 from Ted Ted stay tuned for the AI
1:50:17 learning lab so we'll upload that and
1:50:19 I'll say
1:50:21 okay um
1:50:24 describe
1:50:26 this image in as much detail as
1:50:33 possible we'll see what it does I I've
1:50:36 not tried this before but anyway like
1:50:39 this is running locally on my machine I
1:50:40 do not need an internet connection to
1:50:42 use this and so that's that's something
1:50:46 that we're going to increasingly
1:50:48 see um that that these things are going
1:50:51 to get more and more powerful and more
1:50:52 and more portable where we'll have them
1:50:55 running locally on our phones Samsung
1:50:57 just announced the phone that's going to
1:50:59 have this built into it so that didn't
1:51:01 that doesn't seem to be working let me
1:51:03 reset
1:51:04 it let me try uploading that image
1:51:11 again did it
1:51:13 upload I do not know what it did or
1:51:16 didn't
1:51:17 do uh what is this
1:51:25 image oh maybe it takes a little bit
1:51:27 longer to do the image analysis we'll
1:51:29 just let it sit there for a second
1:51:31 repeat that which can you explain what
1:51:34 running locally means again
1:51:38 okay so when you go to chat GPT so when
1:51:44 we run over to chat JP here we are chat
1:51:49 JP log
1:51:52 [Laughter]
1:51:53 [Music]
1:51:57 in and they say how can I help you today
1:52:00 and I I click on suggest some fun
1:52:02 activities and it and it gives me back
1:52:05 this answer what's actually happening is
1:52:08 when I submit when I hit the return
1:52:11 button with that prompt it actually
1:52:14 sends that
1:52:16 prompt to open AI servers that actually
1:52:19 sit on microsof soft
1:52:22 Azure and that prompt goes into their
1:52:25 machines their machines do all the
1:52:27 processing and then spit the answers
1:52:29 back right so this is so it's it's
1:52:32 effectively two computer systems talking
1:52:34 to one another mine and then all that
1:52:36 data goes over the Internet processes
1:52:38 over there and comes back one of the
1:52:40 security concerns
1:52:44 about all of these large language models
1:52:46 is hey if I send private confidential
1:52:50 secure whatever information to open AI
1:52:53 are they going to use it for their
1:52:55 training or might it be accessible from
1:52:57 other people the answer has been yeah
1:52:59 yeah pretty
1:53:01 much so so the guidance has been don't
1:53:05 send sensitive information to chat
1:53:07 EBT running something locally means all
1:53:11 of what this is doing is happening only
1:53:14 on my machine there's a file sitting on
1:53:17 my machine that is like the super
1:53:20 optimized mini
1:53:22 version of the Llama um large language
1:53:26 model that meta
1:53:28 released I don't know 3 four months
1:53:31 ago and the developer Community has been
1:53:33 working on making it smaller and smaller
1:53:35 Winston the Llama thank you so much sir
1:53:38 um squish squish squish squish squish um
1:53:40 I think cuno is edible transfer though
1:53:43 either because I don't get it so so well
1:53:47 hopefully I'm explaining it now
1:53:49 so
1:53:50 sitting sitting on my computer is this
1:53:54 file so when I when I um put in the
1:53:56 prompt write me a poem instead of going
1:54:00 out to the internet to open ai's
1:54:02 computers it just looks at this file on
1:54:06 my local computer and gives me the
1:54:07 answer right here it that data for me
1:54:11 you know uploading this image uh telling
1:54:15 it to write me a poem never leaves my
1:54:17 computer so that's what running locally
1:54:19 means make
1:54:21 sense does that make sense although this
1:54:24 does not seem to be working
1:54:27 um let me reload
1:54:29 this so if I say
1:54:33 um let's
1:54:35 see I
1:54:38 have a rash on
1:54:42 my
1:54:44 private parts what do I do question mark
1:54:51 I might not want to put that out on the
1:54:52 internet because that might be sensitive
1:54:54 or embarrassing so now I'm getting an
1:54:58 answer back keep the affected area clean
1:55:00 and dry avoid using soap or other
1:55:02 irritants right so I now got an answer
1:55:05 back that I didn't have to go out on the
1:55:07 internet like there there's no chance of
1:55:09 what I typed there getting out to the
1:55:10 world unless say I was live on Tik
1:55:16 Tok but the actual processing of that
1:55:19 prompt just happened right on my
1:55:21 machine that's what that's what running
1:55:23 locally means so it's a pretty big deal
1:55:25 that that you can just install this on
1:55:28 like a non GPU heavy computer and run it
1:55:33 um the the open- source Community has
1:55:36 figured out a way to make that thing
1:55:38 small enough that you don't need a
1:55:40 massive hard drive you don't need a
1:55:42 massive processor you don't need a
1:55:44 massive $88,000 GPU to run
1:55:48 it which is
1:55:57 bitching where did I put my
1:55:59 watch what time is it it's 9:55 it's
1:56:04 9:55 people it's almost top of the hour
1:56:07 what are we up to on the old we're up to
1:56:08 30,000 likes yall are insane so our
1:56:12 inter our interactions are
1:56:15 accessible if we aren't local they're
1:56:17 not really accessible but
1:56:21 but open AI basically says we reserve
1:56:25 the right to train our data on your
1:56:29 [ __ ] and
1:56:31 then if you've got your prompt history
1:56:35 and chat GPT turned on it is
1:56:39 hackable it's not it's not very secure
1:56:42 so so someone I will never tell about
1:56:45 your rash thank you we'll we'll never
1:56:48 tell about your rash thank you Amelia's
1:56:50 wife I certainly appreciate
1:56:53 [Laughter]
1:56:56 that um it's it's not like if you put
1:56:59 stuff there people can just see it but
1:57:02 for example I don't know when this was
1:57:05 March or April someone from Samsung in
1:57:09 Japan put a transcription of a
1:57:13 confidential conference call and some
1:57:18 proprietary material about a product in
1:57:20 chat GPT with their prompt history
1:57:23 turned on and they got hacked and
1:57:26 somebody got that data um so our
1:57:28 interactions are accessible if we aren't
1:57:30 local they're not accessible but they
1:57:31 are at risk of being accessible um just
1:57:34 in general you should not put super
1:57:38 sensitive private stuff out there now
1:57:42 there's an Enterprise version of chat
1:57:44 GPT we don't have access to that yet
1:57:47 because you have to I think it's like 10
1:57:49 or 15,000 a month you needs to be your
1:57:51 spend with them to get that
1:57:54 um but this is why so so when they fired
1:57:58 Sam Alman and it and it seemed clear
1:58:00 that like the company could potentially
1:58:03 fold and chat GPT was very much at risk
1:58:05 of like we don't know the future there
1:58:08 was a there was a flock of people to
1:58:11 open source and and this is
1:58:14 why because the the Llama the the
1:58:18 ability the ability to
1:58:20 run a large language model either
1:58:23 locally on your machine or in a what's
1:58:26 called a virtual machine in the cloud
1:58:29 where you're still sending your data
1:58:33 over the internet but you're sending it
1:58:35 to an environment that you control right
1:58:38 so open and AI can't train on it right
1:58:41 because you're running it locally um and
1:58:44 you're using you know one of these open-
1:58:45 Source models so that's that's that's
1:58:48 why to pay attention to that World um if
1:58:51 you're not super geeky it doesn't matter
1:58:53 right
1:58:54 now I would think that within 6 months
1:58:59 um it will be really easy someone will
1:59:02 come out with the equivalent of you know
1:59:04 just like a you download an application
1:59:06 you double click it and it runs like
1:59:09 chat GPT and it does cool [ __ ] so just
1:59:14 just if you're not super geeky just wait
1:59:17 and it's it's not like anything you put
1:59:19 in here is just there's people are just
1:59:21 sitting there watching
1:59:23 it but you know it's but I wouldn't put
1:59:27 your tax returns out there if you don't
1:59:29 want people to see
1:59:30 them all right all right enough enough
1:59:34 security
1:59:35 talk here we are at the mall welcome to
1:59:38 the AI learning lab just down from
1:59:40 Victoria's Secret across from Ben and
1:59:48 Jerry's
1:59:51 all right
2:00:01 [Laughter]
2:00:10 fantastic all
2:00:16 right let me go look at show and tell
2:00:18 what's going on in show and tell people
2:00:20 what are you showing and
2:00:23 telling M tasty Burgers is tasty I think
2:00:28 steveo's snacky Steve looks like steveo
2:00:31 went to The Burger
2:00:33 [Laughter]
2:00:37 Joint I think he's in Runway ml Bob bold
2:00:43 move is that Ted yeah that's
2:00:46 Ted Ted's good with the with the current
2:00:49 events
2:00:50 Sherry Banks oh wait hang on come
2:00:55 on that's a good
2:00:59 one old and timey picture that's pretty
2:01:05 good now using
2:01:10 pie very nice very nice very nice oh
2:01:14 come on Mighty networks is very slow at
2:01:17 loading images
2:01:20 oh these are cool
2:01:22 Sherry these 50s
2:01:25 photos the chat GPT Appliance that's
2:01:28 pretty
2:01:31 slick I love these Al alternative
2:01:38 histories make money with chat GPT the
2:01:41 little sign on the window and the Llama
2:01:44 the Hedgehog and the turtle watching TV
2:01:46 that's pretty
2:01:48 good
2:01:51 Runway killed the dolly
2:01:54 star this is good you guys are good I
2:01:57 like it all
2:01:59 right let's see if we if we got Anthony
2:02:02 in here to introduce himself did we get
2:02:05 Anthony Anthony Anthony we ain't got
2:02:08 Anthony I think Anthony got pissed off
2:02:10 that we didn't welcome him over in
2:02:12 Discord sorry
2:02:14 Anthony I don't know what to tell you
2:02:18 man sometimes
2:02:20 [ __ ]
2:02:21 happens sometimes [ __ ] happens did I
2:02:24 explain local versus not local clearly
2:02:28 anyone still
2:02:37 confused there's a GPT ointment for
2:02:40 [Laughter]
2:02:48 that
2:02:50 you can get that where that thing I was
2:02:52 talking about if you do a search for a
2:02:54 llama
2:02:57 file hang
2:03:13 on is this
2:03:18 it
2:03:47 I don't think that's the official site I
2:03:49 think that's
2:03:50 someone hang
2:03:54 [Music]
2:04:03 on mozilla's Innovation group and
2:04:07 Justine tunny just released llama
2:04:18 file
2:04:29 where is
2:04:31 I where am
2:04:33 I I'm so
2:04:36 confused
2:04:48 wait
2:04:50 all right so I'm going to copy this so
2:04:54 if you go to this URL I'm going to put
2:04:57 this in the I'll put it in
2:05:00 uh am I going to put it hang on I'm
2:05:03 going to put it in AI tools and
2:05:04 resources I think it's already in there
2:05:06 Peter Peter put it in there so wait hang
2:05:10 on
2:05:12 so this is the the GitHub repository so
2:05:16 here's a quick start for how to do it so
2:05:18 basically you download but this that's a
2:05:21 different one than Peter showed hang on
2:05:23 let me go find the one from Peter
2:05:24 Kaminsky on the AI
2:05:27 Salon um
2:05:31 Peter there's Peter
2:05:35 Kaminsky see
2:05:47 activity llama file
2:05:54 so where's this that's in Show and
2:06:02 Tell so if you if you go to the a AI
2:06:05 salon and you search for Peter Kaminsky
2:06:08 and llama file he's got a link to it
2:06:11 here yeah this is the
2:06:15 one and so so basically you you download
2:06:19 this 4.26 gig file and then you
2:06:23 basically do some [ __ ] in terminal which
2:06:26 is it's a pain in the ass but but it's
2:06:28 it's not that big a deal but you do need
2:06:30 to run terminal basically you just copy
2:06:33 and paste that command and then it's
2:06:36 going to install some [ __ ] and then you
2:06:38 copy and paste that command so wait
2:06:40 actually you copy and paste that command
2:06:43 to go get the file that command to turn
2:06:46 it into a something you can run and that
2:06:49 command to run it and then it basically
2:06:51 launches your
2:06:52 browser
2:06:54 and that interface that I had up and
2:06:56 running is there
2:06:58 so go to the AI Salon Mighty networks
2:07:02 search for Peter Kaminsky and then the
2:07:05 word llama and you'll see this this
2:07:07 article llama file and mid
2:07:09 journey and then click on that so so he
2:07:13 used llama file to create prompts and
2:07:17 then he took those into mid journey to
2:07:19 to make images all
2:07:23 right it's getting geeky in here people
2:07:26 we're getting into command line and
2:07:28 everything them computers is fancy this
2:07:31 is fancy computer stuff we're doing here
2:07:38 now and so listen here's the deal with
2:07:41 all this open source
2:07:44 stuff for the most part I ignore it why
2:07:48 because I'm not an engineer and I'm not
2:07:51 good at that [ __ ] and I end up getting
2:07:55 really frustrated now here's the good
2:07:57 news you can figure this [ __ ] out
2:08:00 because chat GPT is there to help you so
2:08:03 for example when I went to install this
2:08:07 today I downloaded the file and rather
2:08:10 than following the instructions like a
2:08:13 like a neurotypical person would do
2:08:15 because I have ADD I just went and
2:08:18 downloaded the file into my downloads
2:08:20 folder and then I realized oh I didn't
2:08:22 download the file um using terminal so I
2:08:26 like so it's in a different place than
2:08:28 those instructions thought it was so
2:08:31 then I went to chat GPT and I said hey I
2:08:33 made I did this dumb thing in terminal
2:08:36 how do I find the thing and do the thing
2:08:39 and whatever and it told me how to do it
2:08:41 so you do have chat GPT if you run into
2:08:44 errors you can
2:08:47 um you can pop the the error into chat
2:08:49 GPT and it'll tell you what to do so so
2:08:51 the other error I ran into was I needed
2:08:54 to have some thing installed on my Mac
2:08:57 which I didn't have installed so I I
2:09:00 pasted the error in in and it said well
2:09:02 you don't have this thing installed so
2:09:05 type in this command line and it'll
2:09:07 install it so I typed in that command
2:09:09 line it installed it and then all the
2:09:11 other [ __ ]
2:09:13 worked Joe mama I have Enterprise open
2:09:16 AI at work yeah if you've got Enterprise
2:09:19 chat GPT at work then your data is
2:09:22 secure if you don't then it's at
2:09:38 risk there we'll put up the bar Arthur
2:09:41 version of
2:09:47 me
2:09:50 H can we talk about which New Jack City
2:09:53 we can see you in I think I might have
2:09:57 been cut out of the movie but I don't
2:09:58 know I'm still in the credits but it's
2:10:00 at the very beginning of the
2:10:05 movie
2:10:08 and I'm in a car with another white guy
2:10:11 I'm white guy number one he was white
2:10:13 guy number
2:10:14 two and we're like Connecticut
2:10:17 douchebags and we're in Harlem buying
2:10:19 drugs and then I think it's Wesley
2:10:24 Snipes so so like we pull up and like we
2:10:27 wave him over to get our drugs and he
2:10:29 comes over and then something goes weird
2:10:32 I forget what the line was I had I had a
2:10:35 line and then um he pulls his gun out
2:10:40 and then I hit the gas and we Zoom away
2:10:44 and then there was another scene that we
2:10:45 shot that was on a bridge but that that
2:10:47 they cut out I think they might have cut
2:10:48 it out so I don't even know if I'm in
2:10:49 the final edit I haven't seen it but I
2:10:51 get you know my eight or N9 check every
2:10:55 six
2:11:00 months it's I know it's pretty glamorous
2:11:02 I don't like to brag but you know I
2:11:04 wasn't nominated but I I think it's I
2:11:07 think it was just the year there was a
2:11:08 lot of other you know it was probably
2:11:10 like you know Rainman was out probably
2:11:12 Dustin Hoffman did something that you
2:11:16 know I didn't get the odd for white boy
2:11:19 number one in New Jack City it just
2:11:21 didn't
2:11:22 happen you know what are you going to do
2:11:26 what are you gonna
2:11:28 do I met my best friend Steve so so my
2:11:31 buddy Steve sashin um he and I met
2:11:35 um shooting an episode of America's Most
2:11:38 Wanted where he played the Killer and I
2:11:40 played his best friend my favorite line
2:11:43 from that show was uh uh yeah Adam about
2:11:47 this gun
2:11:52 thing and in the next scene he shot a
2:11:54 guy in the
2:11:58 back it was funny we shot that in
2:12:01 Connecticut and then it turned out we
2:12:03 lived a block from each other in New
2:12:05 York City and so we lived on the same
2:12:07 street in
2:12:08 SoHo and uh yeah it was cool hang on
2:12:11 champy Champ's got to deal with
2:12:13 something what are you whimpering about
2:12:16 you want to go out you want to go see
2:12:18 Mama
2:12:22 go see
2:12:33 [Music]
2:12:37 M maybe Sam Alman Nicks the Oscar he's
2:12:40 always had it out for you I I think it's
2:12:42 possible I don't I don't know that Sam
2:12:45 was even
2:12:47 born when I was shooting New Jack City
2:12:50 but I bet you know what I bet he was a
2:12:53 seven-year-old Punk and I bet he has
2:12:56 connections I bet he had connections
2:12:57 when he was seven I bet you're right I
2:12:59 bet he pulled some strings he was like
2:13:01 you
2:13:02 know that guy Shannon it just doesn't he
2:13:05 doesn't sit right with me and they're
2:13:06 like Sam who who who are you talking
2:13:08 about what are you what are you talking
2:13:10 about you're you're in second
2:13:13 grade I think you're right I think
2:13:16 you're right I think Sam sabotaged my
2:13:17 acting career
2:13:20 son of a [ __ ]
2:13:22 well guess I'm calling him four times a
2:13:25 day
2:13:26 now first question why can't I search
2:13:29 for my prompt history
2:13:32 Sam why can't I search on the IOS app
2:13:35 but I can't search on the website why
2:13:37 why has no one noticed that no one
2:13:41 you've got 800 people there no one's
2:13:43 like hey you know what would be cool if
2:13:45 you could actually go back and find the
2:13:47 thing you wrote
2:13:49 six hours ago six weeks
2:13:53 ago no no
2:13:57 one you have you have to go to Stamford
2:13:59 to figure that out bitter you bet
2:14:08 you apparently you have to go to a state
2:14:11 stool school and get an acting degree to
2:14:13 figure that one out all
2:14:15 right could you please repeat the Google
2:14:18 search uh to get the searchable GPT
2:14:20 index yes this is another one like why
2:14:23 is this why why is it such a particular
2:14:26 kind of syntax that that Google made us
2:14:29 do but it is so
2:14:36 [Music]
2:14:39 whatever go to
2:14:42 Google so the question is how do you
2:14:44 search
2:14:46 for gpts
2:14:52 so you go to Google and you type
2:14:58 in site sit
2:15:02 colon chat.
2:15:05 open.com so that's just the chat GPT
2:15:09 website then you type in SLG slash
2:15:11 that's the director the directory for
2:15:14 gpts so and then you hit return
2:15:18 and these are all the
2:15:20 gpts and then you can also um like put
2:15:24 in a keyword like uh dog and it'll find
2:15:30 you gpts with dog images dog facts dog
2:15:33 breed encyclopedia dog G
2:15:37 it dog G it I'm all about dogs all the
2:15:40 time facts stories jokes even images by
2:15:42 Adam
2:15:44 lasnik you could search
2:15:46 for s Sin City sipper Sin
2:15:51 City no now you can't see it but wait I
2:15:54 got I got it before Sin City sipper
2:15:57 maybe I have to put in all of it no it
2:15:59 didn't find it now dang it it just found
2:16:02 it earlier what's going on people oh you
2:16:05 know what I think I was on Bing let me
2:16:07 go to bing.com see if Sin City sipper is
2:16:10 indexed on Bing that'll move bing up in
2:16:13 my book so we'll do the same thing we'll
2:16:16 go site colon
2:16:19 chat. open oh this was in bing that Sin
2:16:22 City sipper showed up Sin
2:16:28 City Sin City sipper there it is Vegas
2:16:31 bartender with a Twist on the
2:16:34 classics that's me people right there
2:16:38 look at that I'm famous I'm on
2:16:41 Bing 2% of the search population the
2:16:44 search audience uses Bing and anyone
2:16:48 who's searching for Sin City Kyle search
2:16:51 your name oh okay that's a good
2:16:54 idea thank you thank you for the the the
2:16:58 remedial guidance there Vicki that's I I
2:17:01 actually did not think of that Kyle
2:17:04 Shannon Kyle Shannon Shannon Shan good
2:17:10 Lord what' you do Saturday night I don't
2:17:13 know watch some old guy try to
2:17:16 type all right right oh the first one
2:17:19 the executive order Explorer I like that
2:17:22 one so that one is I took Biden's
2:17:24 executive order on AI and you can that's
2:17:27 the knowledge base for that one Dream
2:17:29 Weaver is a dream analysis
2:17:32 GPT The Dilemma simulator I don't think
2:17:34 that's
2:17:35 mine Sin City
2:17:38 sipper get rid of that why is that stuff
2:17:40 there go
2:17:41 away what what is
2:17:44 that I don't want related searches stop
2:17:48 well
2:17:49 whatever Philly GPT Pierce coach gilb
2:17:53 Bri a lot of these aren't mine but
2:17:55 anyway yeah there you
2:17:57 go that one's mine executive order
2:18:00 Explorer that's a good
2:18:02 one look at
2:18:05 that what's the main purpose can you
2:18:08 explain section three of the executive
2:18:12 order
2:18:17 certainly
2:18:22 um make it a
2:18:25 poem or
2:18:28 poic here's a
2:18:34 poem I love doing stupid [ __ ] in the
2:18:37 realm where digital minds are born a
2:18:39 mandate stands both clear and sworn
2:18:42 section three with foresight Keen
2:18:44 sketches a vision both safe and clean
2:18:48 here lies the heart of ai's creed where
2:18:51 ethics and Innovations intercede
2:18:54 guidance set for agencies wide ensuring
2:18:56 AI safe trustworthy stride principles
2:19:00 and standards hand in hand crafted to
2:19:03 guide this text so Grand protection
2:19:06 fairness in the spotlight with
2:19:08 transparency guiding through the
2:19:11 night see and you thought politics was
2:19:13 boring come on you can make it as
2:19:16 exciting as Shakespeare
2:19:18 [Laughter]
2:19:23 and just put GPT search spot in Show and
2:19:26 Tell oh cool Ted put GPT search b in
2:19:29 Show and Tell so another reason to join
2:19:31 the AI Salon go to show and
2:19:34 tell and Ted just put this GPT will
2:19:37 search 22,500 gpts there you go look at
2:19:41 that somebody did it for you boom who
2:19:45 did
2:19:46 it seo.
2:19:49 a
2:19:51 awesome let's see um find me gpts
2:19:57 by Kyle K Kyle
2:20:02 sha hey K
2:20:05 hey C hey c c listen c c
2:20:14 c executive order
2:20:16 Explorer
2:20:21 it only gave me
2:20:26 one
2:20:28 um what about Sin City
2:20:35 sipper Sin City sipper is going to be my
2:20:37 most famous of the
2:20:41 gpts and what time is it it's getting to
2:20:44 be it's get to be that time it's going
2:20:45 to be that time I'll let you people go
2:20:48 Sin City sipper appears to be a virtual
2:20:50 bartender with a unique Twist on classic
2:20:52 drinks this GPT is likely designed to
2:20:54 provide users with creative cocktail
2:20:57 recipes or advice on drink
2:21:00 mixing the GPT is made by choer and
2:21:04 flaad that [ __ ] stole my
2:21:13 GPT wait the best gpts
2:21:16 here
2:21:21 I don't think that's
2:21:24 right that's by me yeah it's a little
2:21:28 you got a little hallucination there you
2:21:30 got a little hallucination I got a
2:21:31 little paranoid there for a minute I was
2:21:33 going to fly over to Norway or Sweden or
2:21:37 wherever Sven is from and we were going
2:21:40 to have a we were going to have wor we
2:21:42 were GNA I was going to throw down on a
2:21:45 fjord Sven and I were going to have some
2:21:48 words on a
2:21:50 fjord but we're good
2:21:52 now Kyle if you just Google your name
2:21:55 not with
2:21:58 GPT now do these gpts show up in the
2:22:01 results if you just Google your name
2:22:03 probably
2:22:05 not I I did I thought I I do not
2:22:10 know I do not new let me go to Bing and
2:22:13 see let's go to
2:22:16 Bing Bing
2:22:19 if I just Google my name it's likely
2:22:21 going to be
2:22:22 um it'll it'll bring up Kyle Shanahan
2:22:25 who's the coach from the the San
2:22:28 Francisco
2:22:30 fananas um k k hey K Kyle Shannon that's
2:22:35 two girls names did you know that that's
2:22:37 two girls name see Kyle Shanahan
2:22:41 [ __ ] every time every time that
2:22:43 football team has a shitty weekend I get
2:22:47 direct messaged on on Twitter hey maybe
2:22:50 you should learn to coach you [ __ ]
2:22:52 loser I'm like maybe you should learn to
2:22:54 spell you [ __ ]
2:22:59 loser remember create a
2:23:01 GPT listed in the mod channel oh
2:23:06 cool
2:23:12 great perfect okay
2:23:16 yeah beautiful beautiful beautiful is
2:23:19 there such a thing as a totally free AI
2:23:21 app to use that is any good at all yeah
2:23:24 um Microsoft uh co-pilot is good so let
2:23:29 me let me go through my
2:23:31 list and I will tell you where the free
2:23:33 ones
2:23:34 are this is free for GPT 3.5 it's good
2:23:40 it's usable it's good it can code is it
2:23:44 as good as gp4 no but it's free
2:23:48 um Claude is free you can pay
2:23:51 for it like they they rate limit you
2:23:54 after I don't know some amount of tokens
2:23:58 so like maybe 10 10 searches if you're
2:24:02 using big documents you can do like 10
2:24:04 things before it rate limits you um you
2:24:07 can pay for extra but that's free Bard
2:24:09 is free co-pilot is free grock will uh
2:24:13 grock is not free grock's 18 bucks a
2:24:15 month co-pilot is free co-pilot
2:24:17 microsoft.com this is oops this is chat
2:24:23 gp4 it's Microsoft's version of it but
2:24:25 it's using the same base model as chat
2:24:29 GPT um Pi is free perplexity is free has
2:24:34 a paid version PO is free has a paid
2:24:37 version all of the Llama models all of
2:24:39 the open- source models kind of with
2:24:42 llama 2 in the lead are free now you
2:24:45 have to be geeky and figure out how to
2:24:46 install that [ __ ] thank you Winston I
2:24:50 loves me some llama you know why I like
2:24:52 llamas llamas is
2:24:55 [Music]
2:24:56 tasty so anyway so there's so let's see
2:25:00 1 2 3
2:25:05 4 5 6
2:25:10 7
2:25:11 8 plus Amazon just announced
2:25:15 Q which is also so free but apparently
2:25:18 pretty
2:25:19 shitty so that one's not good but all of
2:25:22 the ones that I've got listed here are
2:25:23 good and there so there's there is a lot
2:25:25 of free
2:25:26 options Flagg that sipper husband
2:25:30 unimpressed divorce absolutely out of
2:25:34 here husband just needs to hang out more
2:25:36 in the AI learning lab make it a date
2:25:38 night we've we've got a number of people
2:25:41 that have date nights in the AI learning
2:25:43 lab because they know how to
2:25:46 live
2:25:48 I'm I was getting a little upset on your
2:25:50 behalf thank you Source
2:25:58 Camp oh man regarding your GPT being
2:26:02 swiped yeah it wasn't swiped but thank
2:26:04 you I appreciate that so save for
2:26:07 everyone with the link means save for
2:26:11 all save for everyone with the link oh
2:26:15 yeah so so when saving a GPT when you
2:26:19 say save and you go public that means
2:26:22 anyone will have access to
2:26:24 it are these free apps available on
2:26:27 iPhone and easy to use all of them are
2:26:29 easy to use all of these large language
2:26:31 models are easy to use
2:26:34 um Pi has an iPhone
2:26:37 app po has an
2:26:39 app that doesn't I think perplexity is
2:26:42 just a website all of these what's nice
2:26:44 about these large language models
2:26:46 websites is they're all so simple that
2:26:48 you can go to these URLs in your um
2:26:52 browser on your phone and then some of
2:26:55 them have apps like Pi has a very good
2:26:57 app you can put it in conversation mode
2:26:59 po has a pretty good
2:27:01 app
2:27:04 um Microsoft has a Bing app which is
2:27:07 which is essentially po which is
2:27:09 essentially co-pilot I don't know if
2:27:10 they have a co-pilot app yet um Claude I
2:27:14 don't know grock is inside X
2:27:18 Bard I don't I think you just I I think
2:27:21 you just go to b.google.r
2:27:32 in browsers just figured out the magic
2:27:36 that is uploads and customization hang
2:27:38 out here and join the salon play like
2:27:42 Kyle says thankful Jones well thank you
2:27:44 thankful Jones seriously ly the the if
2:27:49 there is an imperative that I am I am
2:27:52 trying to put into the world is just
2:27:55 play get curious about this [ __ ] and
2:27:58 just play join the salon because what
2:28:00 the salon is is like a support group of
2:28:04 weirdos that are just playing they're
2:28:07 exploring this [ __ ] they're trying to
2:28:08 figure it out they're making gpts
2:28:10 they're making stupid images all of that
2:28:13 play is educating them about where the
2:28:16 boundaries are of how these tools work
2:28:18 and what they're good at and what
2:28:19 they're not good at chat GPT iOS is not
2:28:23 amazing except I would argue Joe mama
2:28:26 that the conversation in the IOS app is
2:28:29 amazing and the fact that you can search
2:28:31 your prompt history is also
2:28:34 amazing because there have been times
2:28:37 when I've wanted to go back to a prompt
2:28:38 that I can't find I'll go to the IOS app
2:28:40 I search for it there I then make a
2:28:43 change in it which pops it to the top of
2:28:45 my prompt history and I find it again
2:28:47 again so or I copy the URL and paste it
2:28:49 on my on my desktop machine perplexity
2:28:52 does have an IOS app okay great is there
2:28:55 an AI model to create an entire iOS or
2:28:59 Android app from scratch like On Demand
2:29:01 app um there are a couple of them but
2:29:05 the last time I looked at them they were
2:29:07 janky pieces of [ __ ] that they demo
2:29:10 really well but they don't necessarily
2:29:12 write good applications I don't know if
2:29:14 that's gotten better so I would say go
2:29:16 to Future
2:29:17 media.io which is a big directory of
2:29:20 lots and lots of I I AI
2:29:24 apps and search on there for like an iOS
2:29:29 developing tool I'm sure there's some
2:29:31 out there that are pretty decent at this
2:29:32 point check out G GPT store. okay oh
2:29:38 cool somebody made a store GPT store.
2:29:42 let's see
2:29:44 GPT
2:29:45 store. a
2:29:50 hi there you
2:29:52 go somebody made one I'm going to search
2:29:54 for Kyle Shannon Kyle hey Kyle Kyle
2:29:59 Shannon why do you only ever search for
2:30:01 yourself Kyle no gpts found atal all
2:30:04 right let's look up Sin City
2:30:08 sipper hey I'm a
2:30:12 star let go exac wait EX
2:30:17 utive
2:30:19 order no gpts found
2:30:23 boo I don't know all right Sin City
2:30:26 sipper is in there because it's the best
2:30:28 Sin City
2:30:32 sipper yeah he's a
2:30:36 winner GPT store GPT needs to review
2:30:39 proceeding what what where did it just
2:30:42 take me Trin City
2:30:44 sipper did this just install
2:30:47 malware in my
2:30:52 computer all right
2:30:56 um lunatick I Heard a snippet of Kyle
2:30:59 talking about everything you've been
2:31:01 doing really interested yeah I she's
2:31:04 she's just building building building
2:31:07 building
2:31:08 building like Reinventing herself
2:31:10 rediscovering herself reigniting herself
2:31:13 it's it's really cool to watch and
2:31:14 that's so Luna stick is one of many of
2:31:18 the Irregulars in here the people that
2:31:20 show up here a lot and the
2:31:22 subscribers that are just playing and in
2:31:27 playing discovering either discovering
2:31:31 or or generating really interesting work
2:31:34 opportunities it's it's it's it's very
2:31:37 inspiring it's very inspiring to see so
2:31:41 like just [ __ ] hang out just play
2:31:44 just play just play just play just play
2:31:48 but K if I just play how do I make money
2:31:53 you can make money
2:31:56 with the money comes second the
2:31:59 technology does not come first the
2:32:02 problem comes first figure out a problem
2:32:04 you want to solve and that problem can
2:32:07 be as stupid as Kyle just made a dumb
2:32:09 joke I would like to make an image of
2:32:11 that dumb joke and share it on a salon
2:32:14 that's a problem I'm going to solve
2:32:17 um lunatick I'm a former politician who
2:32:19 could uh who couldn't nuro spicy mask
2:32:22 anymore and had a breakdown oh wow
2:32:24 that's intense yeah [ __ ] awesome this
2:32:28 I I'll tell you what what AI you know as
2:32:31 a as a laminated card carrying member of
2:32:33 the ADHD
2:32:35 Club
2:32:37 um I I think these AI tools as an as an
2:32:43 accessibility um tool as a as a
2:32:47 capability amplifier like I'm finding
2:32:50 that I am more productive at a higher
2:32:53 level and more creative than I've ever
2:32:56 been in my
2:32:57 life and you know and I'm not a young
2:32:59 spring chicken like why do I have all
2:33:01 this energy cuz I'm [ __ ] like my
2:33:03 brain is on
2:33:05 fire and so I think I think yeah neurod
2:33:09 Divergent people I think there's I think
2:33:11 there's some real [ __ ] magic in this
2:33:13 stuff for
2:33:15 us what problem are you solving for
2:33:19 um so at at work I'm solving
2:33:24 for here let me flip these
2:33:29 around let me let me put up B Arthur
2:33:33 me um at work I'm solving
2:33:36 for a couple of issues the the the
2:33:39 biggest strategic issue that I'm solving
2:33:41 for it works so my current company is
2:33:44 called
2:33:45 storyvine we've been around for 11 and
2:33:47 A2 years we do a lot of work in
2:33:50 healthcare and Pharma and it's an
2:33:52 automated video storytelling platform so
2:33:55 somebody downloads our app they answer a
2:33:57 series of prompts about like say a
2:34:01 diagnosis we we work with Pharma a lot
2:34:03 so if fizer wants a patient to tell
2:34:06 their patient Journey the patient
2:34:09 downloads the app answers a series of
2:34:11 questions those go up to the cloud 5
2:34:12 minutes later you got a fully edited
2:34:14 video super impressive really cool that
2:34:17 auto magic as we call it the automagic
2:34:20 editing has been a major value
2:34:22 proposition for storyvine for a
2:34:25 decade
2:34:28 well the value of automation is going to
2:34:31 go to zero because everything's going to
2:34:33 be automated no one's going to give a
2:34:34 [ __ ] they be like okay yeah you automate
2:34:36 video okay doesn't everyone do that now
2:34:39 so there's a part of our value
2:34:40 proposition that the value drops to zero
2:34:43 so so I'm like well okay how do you
2:34:46 [ __ ] deal with that there's another
2:34:48 part of our value proposition which is
2:34:50 human beings telling stories
2:34:53 well which in a world of infinite
2:34:55 content where people are just dumping
2:34:57 [ __ ] out there because they can make it
2:34:58 really easily authentic storytelling
2:35:01 that feels to me like the value of that
2:35:03 goes up so so if I've got these two
2:35:06 value propositions for my company right
2:35:08 now one goes to zero and one goes up so
2:35:11 I need to figure out how do I leverage
2:35:12 that
2:35:13 thing so so I'm solving for that and
2:35:16 then I'm also solving for what we do
2:35:19 generates a video file well what else
2:35:23 can I make out of that video file well
2:35:25 with generative AI I could make things
2:35:27 like a YouTube title or a description or
2:35:31 a tweet or a blog post or all sorts of
2:35:34 things so I'm looking at how do I do
2:35:37 story
2:35:38 Centric
2:35:41 um content creation so so I'm not
2:35:44 throwing out a decade's worth of of
2:35:47 value I'm augmenting it and I'm hedging
2:35:50 for the [ __ ] that's you know going to
2:35:51 drop in value so so there's that then
2:35:55 we're also just doing things like we're
2:35:56 solving for efficiency that there's a
2:35:59 way that we we do creative development
2:36:01 for the prompting questions that
2:36:04 historically took two or three meetings
2:36:07 with clients and agencies and a lot of
2:36:09 times the agencies would kind of have
2:36:10 their heads up their asses and it would
2:36:12 just it would just take way too long no
2:36:14 one would understand it would just be
2:36:16 this awkward weird ass
2:36:18 thing and and how we do that is a very
2:36:21 particular way so we turned it into an
2:36:23 automation where we ask the client two
2:36:26 or three questions they answer them we
2:36:29 put them into a Google form and 30
2:36:32 seconds later it spits out a set of
2:36:34 perfectly written perfectly formatted
2:36:36 Story vine approved
2:36:38 prompts that we literally then copy and
2:36:41 paste into a Word document and just say
2:36:43 to the client hey based on the answers
2:36:46 that you gave us here's a starting point
2:36:48 for these prompts why don't we start
2:36:50 here and then edit them and refine them
2:36:52 and for the most part we don't have to
2:36:54 edit [ __ ] so that cut two days out of a
2:36:57 client you know two two hour meetings
2:37:00 with lots of people and lots of AA out
2:37:03 of the process like
2:37:06 gone it's
2:37:08 gone so you know that stuff and then
2:37:11 nightly here I'm just solving you know
2:37:13 someone will come in and go you said
2:37:16 that chat jpt would be good for
2:37:18 storyboarding how do you actually do
2:37:21 that and then I'm like oh [ __ ] I don't
2:37:25 know let's go
2:37:28 try I mean if if nothing I mean that
2:37:31 that's really all this channel is is
2:37:33 let's go
2:37:36 try and hopefully I inspire people to do
2:37:39 the same like when I'm doing stupid [ __ ]
2:37:41 and fumbling around and you're like
2:37:43 watching me for 20 minutes trying to
2:37:44 download a PDF document
2:37:47 My Hope Is that you're off making your
2:37:49 own GPT or you're making images or
2:37:51 you're asking for help in the AI
2:37:56 Salon all right I gotta go let me let me
2:37:59 um let me tell you just a little bit
2:38:01 about the salon I know I've been talking
2:38:02 a lot about that so if you don't know
2:38:04 what that
2:38:14 is all right so that's the address of
2:38:18 the AI Salon it's a mighty
2:38:21 networks Kyle you make it look easy it
2:38:24 well thank you
2:38:28 um
2:38:29 it it's not that it's it's easy it's not
2:38:32 easy because there's a lot of it
2:38:34 changing fast you have to you have to be
2:38:36 engaged to be in it but you don't need
2:38:39 to be a programmer right I that's that's
2:38:42 one of the one of the big misconceptions
2:38:44 is well I'm not I'm not good with with
2:38:46 computers I'm not an engineer I don't
2:38:49 have a degree from Stanford that's okay
2:38:51 you don't need
2:38:52 to the reason chat GPT is such a
2:38:56 significant piece of
2:38:58 technology is not because it's
2:39:01 AI it's because it's a very simple
2:39:04 interface so just like the worldwide
2:39:07 web like turn the internet into a thing
2:39:10 you could point at and click on a
2:39:12 hyperlink right that's the the big
2:39:16 Innovation with the worldwide web was
2:39:18 that it went from having to type in
2:39:20 command lines to click click
2:39:24 click that's what chat GPT does for
2:39:28 AI Kyle you embody let's go try
2:39:31 inspiring attitude thank you very much
2:39:34 Ted you you're just killing it with the
2:39:37 images I feel like T Ted is like our our
2:39:40 New Yorker cartoon illustrator dude
2:39:43 where like a joke comes up Ted's like I
2:39:45 got this one here's your image it's
2:39:48 pretty good let me see I'm going to go
2:39:50 look at Show and Tell see what we got
2:39:52 [Laughter]
2:39:55 going Kelly camp that says I'm not
2:40:03 bitter oh these are
2:40:06 good these are really
2:40:08 good um
2:40:10 anyway go there and then when you get to
2:40:13 the salon first thing I want you to do
2:40:15 is check
2:40:18 out chiy all right I know I'm coming 5
2:40:22 minutes you little dog you little dumb
2:40:25 dog all right um come to the salon and
2:40:30 the first thing I want you to check out
2:40:32 is about the
2:40:36 salon and you'll see that pretty picture
2:40:39 of the rag tag group of what what do we
2:40:41 say a rag tag group of missfits Nero
2:40:44 Wells and Miss crant who who have
2:40:46 gathered to explore the Uncharted
2:40:47 Wilderness of generative AI we are
2:40:50 everyone from entrepreneurs to Educators
2:40:53 from healthcare workers to creatives and
2:40:54 everyone in between and then go check
2:40:57 out our values so the values are really
2:41:00 important to me I want this group to
2:41:02 grow but I don't want it to lose what it
2:41:04 is which
2:41:06 is remarkable and why is it remarkable
2:41:10 these values are not [ __ ] we made up
2:41:12 these values are who people are showing
2:41:15 up
2:41:17 as so they're curious they're generous
2:41:20 this is a big one generous generous to
2:41:23 me
2:41:24 means willingness
2:41:28 to share what you're learning everyone's
2:41:31 learning like here's the
2:41:33 deal if you think you're really far
2:41:35 behind if you think you're totally
2:41:36 clueless and everyone knows more than
2:41:38 you do it's absolutely not true no one
2:41:40 knows anything right now because
2:41:41 everything is changing so fast and we
2:41:44 we're all confronting these Tech
2:41:46 Technologies through different lenses
2:41:49 and so the ability to come in here and
2:41:51 and learn something that you might think
2:41:53 is in significant if you share it in
2:41:55 this community it's going to be
2:41:57 significant to someone else and that's
2:41:59 what this group is all about um
2:42:02 exploration is one of the values so like
2:42:03 are you adventurous collaboration or are
2:42:07 you're willing to reach out and say hey
2:42:08 you want to do something together great
2:42:09 inclusion people are coming from all
2:42:11 over and the last one's empathy just
2:42:14 recognizing that everyone is trying to
2:42:18 figure this [ __ ] out right now and they
2:42:19 might be scared they might be angry they
2:42:21 might be
2:42:22 resistant that's cool we understand
2:42:26 we're all going through it too I mean
2:42:28 the thing I said about my company where
2:42:30 there's a significant portion of our
2:42:32 value proposition that drop drops to
2:42:34 zero in an 11 and a half year old
2:42:36 company you don't think that's not
2:42:38 [ __ ] terrifying it's
2:42:40 terrifying how am I dealing with that by
2:42:42 learning everything I [ __ ] can about
2:42:44 this stuff
2:42:47 the the only thing I can do the only
2:42:49 thing I can do is deal with the [ __ ] I
2:42:52 have in my control what do I have in my
2:42:55 control I don't know I can learn what
2:42:57 these tools are that's
2:43:01 it I can't fix if the robots are going
2:43:04 to kill us I can't fix if the [ __ ]
2:43:07 privacy [ __ ] is I all I can do right now
2:43:10 what's in my control is learn these
2:43:13 tools and the more I learn about them
2:43:15 the more I realize everything's changing
2:43:17 everything has already changed it's just
2:43:19 that a lot of people don't know it
2:43:22 yet oh joy py Mighty networks is so much
2:43:25 easier than Discord thank you
2:43:28 um I think I agree I think I agree I'm
2:43:31 I'm I'm used to Discord there's some
2:43:33 things that Discord did that I like but
2:43:35 there's a lot of things that Mighty
2:43:36 networks does that I like a lot so it's
2:43:40 I'm I'm glad people are finding it
2:43:41 intuitive
2:43:43 good no one I know is using Ai and and
2:43:47 that's what I would say thankful Jones
2:43:49 is don't be afraid to talk about it just
2:43:51 because people aren't using it tell them
2:43:53 what you're doing hey I just did this
2:43:55 thing on chat gbd oh here comes thankful
2:43:57 Jones and talking about AI
2:44:00 again last thing I want to hear about
2:44:03 right now chat jtp you know what I mean
2:44:06 you know those things hallucinate right
2:44:08 I heard that once um those
2:44:13 people when you talk to them and they
2:44:15 make fun of you now they're going to
2:44:16 come back to you in 6 months and go uh
2:44:19 hey there Jonesy uh listen you know uh
2:44:22 hey you were talking about that chpt
2:44:25 once Could It Could you teach me how
2:44:28 to I think I need to learn it now just
2:44:31 got laid off
2:44:33 here it'll happen just keep just keep
2:44:36 doing it keep getting curious keep
2:44:38 talking about it and and that's one of
2:44:40 the reasons that the salon exists right
2:44:43 this is a group where they they won't
2:44:45 think you're weird this is a group that
2:44:49 is in the same boat you are none of the
2:44:51 people we know are using this [ __ ]
2:44:53 either I've got a very small company
2:44:55 I've been telling them for a
2:44:57 year use AI use
2:45:00 AI two of them are
2:45:03 two
2:45:05 two
2:45:07 two other than
2:45:10 me does that frustrate me yeah it
2:45:12 frustrates the [ __ ] out of me is it
2:45:15 really in my control no the fact that my
2:45:18 23-year-old twin boys are not using it
2:45:21 drives me [ __ ] insane is there
2:45:23 anything I can do about it no I'm just
2:45:26 going to keep cranking away and at some
2:45:28 point someone's going to go hey hey Dad
2:45:31 I think I want to learn that now all
2:45:33 right here you go
2:45:35 boy they're probably listening to me
2:45:37 right
2:45:38 now stop talking about us
2:45:44 Dad how often do we we do well we n n do
2:45:47 well Ted that Gary Gman piece God that
2:45:51 is the funniest
2:45:53 [Laughter]
2:45:55 thing um let me see if I can find it I
2:45:59 love this this is it's one of my
2:46:01 favorite bits if you've not heard the
2:46:03 Gary
2:46:07 Gman Gary Gman
2:46:10 States state abbreviations
2:46:14 YouTube
2:46:19 a documentary to everyone and
2:46:22 then very much I just wanted to uh
2:46:24 recommend a documentary to everyone and
2:46:27 then then I'm going to go two capital
2:46:29 letter abbreviation and so they wait
2:46:31 where's
2:46:34 the is this the whole
2:46:43 thing thank you very much it's nice to
2:46:45 be here I was uh I was exercising today
2:46:47 and there was a man no that's not it
2:46:50 anyway go find the Gary Gman piece on
2:46:53 state
2:46:54 abbreviations it's a it's one of the
2:46:56 best written oh this is thanks very much
2:46:59 I just wanted to uh recommend a
2:47:01 documentary to everyone and then then
2:47:03 I'm going to go um it's about the uh
2:47:07 it's about the men and one woman who
2:47:08 abbreviated all 50 states down to two
2:47:10 letters all you all you have to know for
2:47:12 this is that we have 50 states in
2:47:13 America and they each have a two capital
2:47:15 abbreviation but that that wasn't always
2:47:17 the case up until I want to say
2:47:21 1973 so I will uh up until 1973 every
2:47:26 state had its own length of abbreviation
2:47:28 and it was chaos like Massachusetts was
2:47:30 mass period Florida was fla Utah was
2:47:36 Utah they just dropped the H not much of
2:47:39 an abbreviation but then the post office
2:47:41 said no every state has to have a two
2:47:42 capital letter abbreviation and so they
2:47:44 convened a crack Squad of
2:47:46 abbreviators they assembled a rag tag
2:47:49 outfit of Rogues Misfits and Naro
2:47:52 Wells how often do well n
2:47:56 n so so if you if you look at
2:48:00 the we are a rag teag group of Misfits n
2:48:03 duells and Mis
2:48:05 creant you see where some of our
2:48:07 inspiration comes from that that is one
2:48:10 of the best written pieces of Comedy i'
2:48:13 I've ever seen it is so good he just
2:48:16 runs down these rabbit holes oh it's so
2:48:19 good and then he comes back out of them
2:48:21 it's amazing he's he's great all right
2:48:23 oh
2:48:24 listen I gotta get out of here go join
2:48:27 the AI Salon uh thank you for hanging
2:48:29 out tonight let's see my my edible
2:48:31 kicked in I've been laughing like a wild
2:48:36 woman oh tapping like a wild
2:48:39 woman that's
2:48:42 awesome I thought you meant laughing
2:48:44 like running around running around your
2:48:46 kitchen
2:48:47 island like that doesn't sound like
2:48:50 inedible to me tapping a lot
2:48:54 does oh what did I do with my life
2:48:57 before this
2:48:59 live I hope I hope a
2:49:03 lot I I feel I do feel guilty sometimes
2:49:07 that I'm I have I have hijacked a bunch
2:49:09 of
2:49:11 lives oh we Kyle we miss you during the
2:49:14 day too all that's very sweet thank you
2:49:17 um all right listen I'm going to get out
2:49:19 of here folks thank you very much uh I
2:49:22 see the turtles and the llamas and the
2:49:24 Hedgehogs um let me go through here and
2:49:27 do a little Romper Room Winston good to
2:49:29 see you sir Emilio's wife as always
2:49:31 you're amazing digital Gods lunatick
2:49:34 thank you Mega Tobias oh digital Gods
2:49:38 thank you Joe mama very sweet thank you
2:49:40 so much apple just told me it's past my
2:49:44 bedtime
2:49:45 steo Vicky so thank you so much thank
2:49:48 you to the mods ASU um jrc good to see
2:49:52 you cuno as always Ted Dr J Man a bunch
2:49:57 of you in here tonight a human being Joe
2:50:01 mama Sherry D wow it's a it's a Party
2:50:04 Source Camp cuno so cool oh we hit
2:50:07 40,000 likes awesome you we did it
2:50:14 yay
2:50:15 oh
2:50:16 man he said it's picking up now but holy
2:50:19 moly it
2:50:20 took it took long for how much tuition
2:50:23 costs I don't know what that's in
2:50:25 reference to all right all right
2:50:28 everybody
2:50:29 listen we are
2:50:31 addicted you're a bunch of Twisted
2:50:34 Twisted nines um you're my favorite
2:50:37 Twisted nines I am out of here I'll see
2:50:39 you tomorrow um yeah go make some [ __ ]
2:50:41 go get curious join the salon hang out
2:50:43 talk to each other start [ __ ]
2:50:46 just go play play play all right peace
2:50:49 out
2:50:50 everybody