Jan 15, 2024 Mega Live Chat GPT DeepDive

Description

Chapters

Transcript

0:03 [Music]
0:14 [Music]
0:29 oh
0:32 hey what are we doing rapping why are
0:34 you rapping all of a sudden you were
0:36 singing so
0:37 [Music]
0:59 good
1:03 hey hey hey that's very aggressive
1:06 singing there are people trying to enjoy
1:08 their nachos right
1:09 [Music]
1:27 now thank you for sharing the live
1:30 everybody body Tobias Silver Fox just
1:32 shared feel free to share the live if
1:35 you uh would like to see some other
1:37 folks in
1:41 here I just came in from uh shoveling
1:45 sidewalks here in Denver it is min
1:49 -6 that's Fahrenheit not
1:55 [Music]
1:59 Celsius
2:01 [Music]
2:08 oh boy hello hello
2:11 [Music]
2:29 hello
2:33 [Music]
2:40 s in this lonely
2:42 Town wonder when things are going to
2:48 change dream my life
2:51 away seems these dreams have turned to
2:54 BU
2:57 class get my nerve up but my past is
3:01 pulling me
3:05 down wondering how
3:07 long this black doll going to stick
3:11 [Music]
3:15 around somebody told me once before said
3:19 you can never go home
3:21 again why you leave S A things to steer
3:26 me
3:27 away from the truth of who I am and what
3:30 I believe so I thanked him for his two
3:34 cents with a
3:35 handshake and some
3:37 sympathy yeah I packed up my blue
3:41 jeans and I headed for this big
3:46 price of my
3:48 [Music]
3:51 freedom
3:52 byebye Black Sheep to the black sheep of
3:57 the
3:59 family by
4:03 byebye oh that means so very much to me
4:06 yeah
4:08 byebye Black
4:10 Sheep to my friends and my
4:14 family
4:16 [Music]
4:18 byebye going off and set my
4:21 soul set it
4:29 free
4:32 oh that's a pretty
4:36 song all
4:39 right any fans of Marty out there Martin
4:44 ston I can't do him justice but that's a
4:47 mighty pretty [ __ ] song all right
4:49 welcome to the AI learning lab it is
4:52 Monday night it is that was a total
4:55 guess I have no [ __ ] idea what day it
4:59 is it's Monday night I guessed correctly
5:03 that's right I went to work today it was
5:06 so cold so so be because I know how to
5:10 live I took an iced coffee in minus 4
5:13 degree weather and got in my car and
5:16 drove to work by the time I got to work
5:19 there was a ring of ice around the top
5:21 of my
5:23 coffee and I don't live that far from
5:25 the office so the inside of the car with
5:28 the heater on was cold and enough to
5:30 form ice on my iced coffee so probably
5:35 probably should have gone with the hot
5:36 coffee but you know every now and again
5:39 every now and again she kicks in
5:42 Ron hey Kyle Sir with the hair of a
5:45 young Samson well it's the hair of it
5:48 was under a knit hat shoveling and then
5:50 I threw some water in it so yeah we're
5:52 going to just have to deal with the
5:54 uh whatever that is it's just like see I
5:58 can't I I um I'm not going to I'm not
6:01 going to tempt fate speaking of
6:06 Samson oh man all right
6:12 so let's get going here I made a uh I
6:15 made a video today I haven't been making
6:17 a lot of
6:18 videos and
6:20 so I'm trying to do some planning for
6:24 2024 I've been doing these lives for I
6:27 don't know six seven months now nightly
6:30 still enjoying them but I don't really
6:32 have an end game and I kind of gave up
6:34 on I didn't give up on I just haven't
6:36 been consistent with making videos on my
6:38 channel and part of that's cuz I didn't
6:40 really have a plan so I'm putting
6:43 together in the next week or so I'm
6:45 putting together a
6:47 plan and you know to the extent that
6:49 adhders can plan but you
6:52 know so I'll do what I can don't get
6:55 your don't get your hopes too high um
6:58 but um
7:01 feel free to comment on that video and
7:02 let me know the kind of stuff you'd like
7:04 to see it looks like you know doing one
7:07 idea came up with that I thought would
7:08 be kind of
7:09 fun is to CU there were a lot of
7:12 requests for you know tell us about
7:14 specific tools to use like we we want to
7:16 know what tools to use what's good how
7:18 do you use them what do you
7:19 do um and I was thinking about it's you
7:25 know not quite a Dave pornoy Pizza
7:28 review thing you know Pizza review
7:32 time one bite you know the
7:34 rules um but like giving giving these
7:39 tools
7:43 grades because right now like the only
7:46 thing that I think maybe gets an A minus
7:49 is chat G
7:51 pt4 and everything else is in you know
7:56 janville you know just passing getting
7:59 particip a certificates actually maybe
8:01 that's good maybe the the lowest grade
8:05 you no one gets an
8:07 F minus you get a participation
8:10 certificate I think that could be kind
8:12 of
8:15 funny they're in the
8:20 game oh man use chat GPT to create your
8:24 plan and dial it in yeah that's not a
8:25 bad idea that's not a bad idea I could
8:28 maybe even yeah play with that in here a
8:30 little bit um didn't comment on the vid
8:34 but I rarely have time for vids outside
8:36 of the live I'd rather see [ __ ] here oh
8:39 that's
8:40 interesting you can do what I do with my
8:42 tools for teachers what do you do Mr K
8:45 like I've seen those tools for teachers
8:46 but I don't I don't recall do you give
8:49 them grades maybe I'll maybe I'll steal
8:52 your thunder and you know what I could
8:55 also call myself Mr
8:58 K
9:01 oh oh I could totally go after your
9:04 brand I'm like hey it's Mr K I don't
9:07 teach mid middle schoolers but I'm going
9:10 to give [ __ ] grades
9:12 anyway I think it's
9:14 solid I think I I could also start a
9:16 second education Guild within the salon
9:19 I have the power it could just be like
9:21 Kyle's education Guild he's like is he
9:23 an educator wait what does he what are
9:25 his qualifications oh he's got none he
9:27 just does this I like it I like it thank
9:30 you for that idea Mr K solid all
9:33 right Beth loves life no one said there
9:36 would be an end
9:41 game that's good I like that yeah pretty
9:44 much teach Middle School you pretty much
9:46 teach middle schoolers you're right Mr K
9:49 this is this is the AI learning lab has
9:52 the emotional
9:54 intelligence of seventh and a half grade
9:56 you're actually very correct
9:58 there
10:02 this is you know what this is holy [ __ ]
10:05 it just hit me it this is like these are
10:09 all the kids that got detention
10:12 or the kids who never got detention that
10:15 were like the chaperon in
10:20 detention but these these are all the
10:22 kids that that you know manage detention
10:26 in one way or another I that's that's
10:28 actually a really solid observation I
10:30 like it I like it all right that's good
10:33 I think we might have another education
10:35 Guild that's
10:36 solid the alternative education
10:41 Guild we only give out
10:46 certificates there are no diploma
10:48 certificates
10:49 [Laughter]
10:53 only oh man that's funny actually you
10:58 know I don't know there's something kind
11:01 of cool about that idea
11:04 of doing something like that in
11:07 here people can put on people can put on
11:10 their LinkedIn profile I achieved I
11:13 achieved a second degree certificate
11:15 from the AI learning lab oh it's very
11:18 impressive tell me about this AI
11:20 learning lab where's it located it's in
11:23 Denver oh did you go to Denver I flew
11:28 in
11:30 it's The Breakfast Club this is totally
11:32 The Breakfast
11:35 Club oh that's good that's solid that's
11:38 solid all right see I got this is this
11:41 is why we talk planning we come up with
11:43 ideas and you know and we go do
11:50 stuff okay you just gave me an idea to
11:54 grade the tools I'm stealing that for my
11:56 teacher tools dang you I'm here you're
11:59 doing a live and you're going to go off
12:01 and make the first video and scoop
12:05 me that Mr K i you here here's what it
12:08 is about especially Middle School
12:11 teachers they seem so nice and then
12:14 they'll stab you in the
12:20 back appropriate that I have a hatchet
12:22 on my
12:24 t-shirt I see I see how this rolls okay
12:27 no that's fine that's fine you do what
12:29 you need to do you take you know hey you
12:31 go on your little ticktock you make your
12:33 little tiktock I'm going to give you a
12:35 grade with a like an a with a little
12:36 check mark Cuz We Know Kyle's not going
12:38 to do production values he's not going
12:40 to do the little check mark I get it
12:43 that's good that's good you do what you
12:45 need to
12:46 do all right fantastic Bob tell him what
12:49 he's W he's want a trip to the back room
12:51 if you know what I mean you know what I
12:53 mean yeah yeah funny haha
12:58 funny just doing random mob Impressions
13:03 now somewhere between Joe peshi and and
13:10 dairo blandon's Breakfast Club I'll take
13:13 it that's actually a good
13:15 one hey it's breakfast somewhere in the
13:17 world I think Rudy's just getting up
13:19 about
13:21 now he's in South
13:24 Africa all right what should we do here
13:27 tonight what do you want to do what do
13:28 you to do my no my nose is all itchy I'm
13:31 I'm still like I'm decompressing from
13:33 shoveling snow in minus 6° weather uh so
13:39 I just need some time to
13:41 like all right who's champy is that the
13:44 instructor no he's in our
13:48 band he plays
13:52 bass he P plays bass and sings
13:58 backup he he does some mean Harmon he
14:00 that
14:01 dog oh man you just it's it's hard to
14:04 reain him in though that's the problem
14:06 with champ is he just wants to lead the
14:08 band and we're like uh no Daddy
14:13 Kyle's got the lead here his band he's
14:16 the one with the van
14:19 champ all right I thought you were doing
14:22 dairo it was not it was sort of dairo it
14:25 was it was dairo like yeah what are I
14:28 going to do
14:31 oh who's on first base yeah
14:36 exactly champy is the lab in AI learning
14:39 lab no champy champy is not anywhere
14:42 close to a lab Champ's just a street
14:46 Thug dog he was found in a ditch in
14:51 Oklahoma and then we went to so we went
14:54 to the boulder pound or whatever the
14:57 hell they call it the dog rescue
14:59 whatever they call it in Denver and or
15:01 in Boulder and we went there to get a
15:03 lab because we've always had labs and we
15:05 love labs and we got there and I think
15:07 there was a
15:08 lab but he was just not good he was just
15:11 not right he was a little
15:13 janky and then and then my son olly goes
15:17 how about this dog
15:21 champ the tub of the tub of
15:26 singing and I'm like absolutely not that
15:29 dog looks horrible he looks he looks
15:31 like a miniature pit bull I'm like we're
15:33 going to get kicked out of the
15:34 neighborhood if we get
15:36 him and they're like let's just take him
15:39 for a walk and then for the entire walk
15:41 I'm like absolutely not and for years I
15:44 was like absolutely not but when we
15:47 first got him home and the first time I
15:48 picked up my guitar he was
15:52 like so I was like all right I guess
15:54 we're not returning him but you know we
15:58 always threaten to return turn him to
15:59 the big house we we were his third
16:01 adoption we were champ had two other
16:04 families adopt him and they're like you
16:07 can take him but two others have
16:10 returned him and just show you know if
16:13 you bring him back it's lights out for
16:16 old
16:18 champy he's going to the he's going to
16:21 the back room um AI desktop management
16:25 tools would be fantastic yeah I think
16:27 we'll see that
16:30 I don't know I haven't played I don't
16:31 have Windows I'm a Mac guy but I know
16:33 Windows 11 now has AI co-pilot in it so
16:36 I think it's doing some desktop
16:38 management stuff champ whisper or
16:41 gbt so champ hit the [ __ ] jackpot in
16:45 terms of families cuz we're just batshit
16:47 crazy enough to have a batshit crazy dog
16:50 and uh he is cuckoo but he's a lover
16:53 he's a lover and he sings so what are
16:55 you going to do the salon is is always a
16:58 good place place to
17:00 start I think that's just a good rule in
17:03 life where do you start the
17:11 salon oh man Oli Ali recognized champy
17:16 was an artist I think he did Ali Ali
17:19 kind of instantly fell in love with
17:21 him and those two have a special
17:23 connection they really
17:25 do oh but you know you want to know the
17:27 [ __ ] up thing so I absolutely the last
17:30 one on the champ champ train I just
17:31 didn't didn't give a [ __ ]
17:34 um but I'm the only one in the family
17:37 that rough houses with him and he's like
17:39 a little musly Thug you know and so so
17:43 like I slap him around and he hates
17:45 having his feet grabbed he grab his feet
17:47 and he's
17:48 like and so I grab his feet and then we
17:52 fight and he loves me no one else fights
17:56 with him he needs tough love Robert
17:58 Rossy thank you for the lightning bolts
18:00 very much appreciated Silver Fox thank
18:02 you so much
18:04 um he loves a good fight man and he
18:07 loves me and man the others are pissed
18:09 off they're jealous why does he love you
18:11 more than us oh very nice Ron hey
18:16 hey like a champ paaw he's also got very
18:19 sharp nails um the new Google Ultra
18:23 added Google productivity Suite I
18:26 believe what
18:31 the new Google Ultra you mean Gemini
18:33 Ultra added to the Google productivity
18:36 Suite No Danielle thank you so much
18:40 Robert Rossy thank you for the
18:43 orange did what did did that is that
18:46 true
18:51 people Google
18:57 Ultra
19:00 uh uh
19:02 uh Google Gemini Ultra I don't see
19:07 anything
19:11 Jami Ultra oh not
19:22 viewer no there's nothing
19:24 new that I can
19:27 see
19:36 I got nothing
19:38 people I think people would
19:41 have been jumping up and down for that
19:44 one PT Pate did so did say that they
19:48 were going to announce something
19:50 today didn't
19:53 he Google
19:56 announcement today
20:12 new smartphones no that was from May
20:16 11th boy it is it is amazing I think
20:19 this is Bing search it is just
20:21 amazing how shitty search is so I'm
20:25 using the KH Academy code GPT to
20:27 instruct me on how to add Bings
20:30 API so that I
20:33 can that sounds cool
20:37 Barbara so that you can what have live
20:39 links in my GPT for the
20:43 user add code to
20:47 GPT so you can have live links for the
20:51 user you might be a it depends on what
20:54 links you want to add if you want to add
20:55 them dynamically from other websites you
20:57 might need to code that but you could
20:58 also probably just put them in your
21:00 instructions with
21:03 um with uh like a little bit of
21:07 structure and tell it to include them as
21:09 hyperlinks when appropriate I asked
21:12 perplexity Burger King oh if Burger King
21:14 was better than
21:15 McDonald's it said W Wendy's was the
21:18 best that's really funny it's funny when
21:22 I grew up when I grew up going to
21:25 Wendy's was a step up it was it was like
21:28 to get fancy we'd go to Wendy's and get
21:30 a baked
21:31 [Laughter]
21:34 potato oh man that's
21:38 funny bottom of the basement was Hardies
21:41 Hardies it was like you could get a
21:42 burger for a nickel and it tasted like
21:45 it but we usually got like chili dogs at
21:47 Hardies if we wanted to step up our game
21:51 we get a baked potato at Wendy's or a
21:55 chili and always a frosty always a
21:58 Frosty
22:00 y maybe I'll change the AI learning lab
22:03 to like fast food reviews could do that
22:06 sorry I meant
22:07 Gemini Gemini's been in
22:10 um Gemini's been in the Google suite for
22:13 a while although I haven't been able to
22:16 play with it because it's not Avail
22:19 available to
22:20 work work space it's called is that
22:23 maybe that's what they added today is
22:24 that what you're trying to say is that
22:26 what I
22:27 missed
22:33 Google details posted in the mod Channel
22:35 hold
22:46 please was that today they announced
22:56 this that's from De December
23:01 13th yeah that's been in there for a
23:06 while all right I think that's old
23:21 news yeah it's old not seeing anything
23:23 new yeah that's all right that's all
23:26 right if P comes in here P can tell us
23:27 if there's something that's old new
23:29 sorry that's okay not a problem not a
23:31 problem Wendy's chil Papa JJ Papa JJ
23:35 knows how to
23:38 live I used to brag about Wendy's chili
23:41 and then someone said you know that's
23:42 like all the burgers they drop on the
23:43 floor they just crumble them up and put
23:45 them in the chili and I'm like that's
23:47 why they're so tasty bake potatoes with
23:49 Wendy's chili on his side and pour that
23:51 chili on the Tater exactly exactly Clint
23:55 Copperhead knows how to live thank
23:57 goodness there's people in here here
23:58 that are that are of a higher
24:01 Echelon of society because people that
24:05 don't understand the The Gourmet that is
24:08 Wendy's you you just don't understand
24:11 how to live
24:14 people um if you're new
24:19 here you very well could be thinking is
24:22 he ever going to talk about AI I thought
24:25 this is called the AI learning lab I
24:27 came here to learn I've got my notebook
24:28 out and everything it's been 15 minutes
24:31 and it seems that all he's talking about
24:32 is how cold it is some fat dog in and
24:36 and fast
24:38 food he he's like a foodie but for bad
24:41 food is that what this channel is is
24:44 that who he
24:46 is I came here to learn about uh the the
24:50 artificial intelligence I think they
24:52 call it cuz someone told me I could make
24:54 money you can make money
24:57 with so that's what I'm here for so if
25:01 you're here for that you've come to the
25:04 right
25:05 place we will absolutely tell you how to
25:08 make money with
25:12 ch Has anyone used appify for website
25:15 crawler content for gpts I have not it's
25:18 not a bad
25:20 idea
25:27 um
25:28 trying to think about gpts are really
25:30 they're they're in a really interesting
25:32 place right
25:33 now where if you don't know what a GPT
25:36 is it's a custom version of chat GPT
25:39 been out for about two
25:41 months sounds like about 3 million of
25:43 them have been made in two months which
25:46 I assume that's good I don't
25:48 know this is why we love you 30 minute
25:51 hell 30 minute hell and 30 minute
25:54 goodbyes absolutely absolutely we you
25:57 know if we need to take some detours
26:00 here and there we'll do
26:02 it we'll do
26:04 it I you know what what's just hitting
26:06 me right now this is part of the
26:09 Brilliance a thank you very much
26:12 Andrea part of the part of part of my
26:15 absolute
26:17 Brilliance patting myself on the back
26:19 here is things are so
26:23 random that you literally don't know am
26:27 I going to say something something
26:28 interesting in the next 4
26:31 seconds or the next 40 minutes and so
26:34 you hang out for the 40 minutes going oh
26:36 he's never going to get
26:38 there and then you and then you get up
26:40 to to heat up your nachos because you're
26:43 halfway through them you got to add
26:44 other cheese of
26:47 course you go to heat them up you come
26:49 back and I've done something interesting
26:51 and you're like wait what did I miss
26:54 this is why you can't leave this
26:57 channel you see how the it's like with
27:00 traditional drama you got your
27:02 Cliffhanger right you got manyi
27:06 Cliffhangers throughout the episode they
27:08 they do the mini Cliffhanger on the
27:10 commercial breaks and then you get to
27:12 the end you get the big
27:13 Cliffhanger who shot Jr that's a Gen X
27:17 joke there was this TV show called
27:19 Dallas and there was a guy named Jr Jr
27:21 and someone shot him cuz he couldn't
27:23 renegotiate his contract or something so
27:25 they had to kill him off anyway doesn't
27:27 matter it was a big thing made the news
27:29 there was only three shows to watch so
27:31 you know when one of them had an event
27:33 it was like a world anyway not the not
27:35 the point the AI learning lab it's like
27:38 a cliffhanger a second you just never
27:40 know if this thing's going off the rails
27:43 if I'm going to give out my credit card
27:45 information to anyone who cares to have
27:47 it that was three nights ago you could
27:49 have had
27:51 that that new pair of shoes you've been
27:53 wanting I'd have bought it for
27:57 you
27:58 so you got to stay I'm
28:07 sorry Source Camp got to wind it up and
28:10 wind it down exactly we're exploring
28:14 Kyle exactly so the people that are here
28:17 regularly we call them the Irregulars
28:19 what they know is when I go off like
28:22 this or don't start like this they just
28:24 go do [ __ ] they just go entertain
28:26 themselves PID spoiled the fun so
28:30 they're off entertaining themselves
28:31 you're like seriously is what what is
28:34 this will this ever begin what what is
28:36 this channel is there a
28:39 manual is there is there a read me file
28:41 for this
28:48 channel I'm hooked now thank you
28:55 Erica we do do stuff so anyway so here's
28:58 what we do here if you got questions
29:01 about
29:02 AI if you want me to explain how the
29:05 stuff
29:06 works I will explain how it works and
29:09 then Pate who comes in here who works at
29:12 Google and actually knows how this [ __ ]
29:13 Works will correct me so like I have a I
29:16 have a real-time fact Checker so as I
29:19 [ __ ] you about how all this stuff
29:22 works I got a fact Checker that keeps me
29:24 keeps me on
29:26 point uh so we can do that I can show
29:28 you how to use tools we can do basic
29:30 stuff we can do advanced stuff but if
29:31 it's too advanced I'm just like I get
29:33 mamby pamy but I'm like just here's the
29:35 problem with the advanced stuff it just
29:37 takes longer and it's really
29:40 boring so I tried to do some real work
29:42 in here once that didn't go
29:44 well so if you have questions about AI
29:47 pop them in the comments below um we've
29:49 got uh mods in here that will'll sort of
29:52 grab your comments and question and and
29:54 Float them to the top Kyle I might have
29:57 been thinking about Google Google's
29:58 announcement of Bard Advanced last week
30:01 based on Ultra uh okay yeah that's fine
30:05 we'll as soon as as soon as Ultra does
30:08 anything like I'm waiting with baited
30:10 breath for Ultra to come out but the the
30:13 the the kind of rumors that I
30:17 heard when they decided not to launch it
30:20 in December which I think was the
30:22 plan but they found that it was really
30:26 bad with foreign language is is kind of
30:28 the the hubub that I
30:30 heard um and when they decided to push
30:34 it like it's it seemed like it was going
30:36 to be months not weeks till we see it so
30:40 so I'm guessing mid to late February or
30:43 early March for Ultra I wish it were
30:46 sooner and who knows what open AI is
30:49 going to do I can't I can't imagine open
30:51 AI doing anything really big right now
30:53 what would be nice of open AI to do is
30:56 make the G GPT store suck less that
30:59 would be really good if they did that
31:02 all right Gen X and boomers are
31:04 pretending to be young again that's
31:09 right that's an idea a manual yeah that
31:13 that would be fun Emanual for this
31:15 channel Chevy Chase is jealous of your
31:17 show listen Chevy Chevy's a bitter man
31:22 funny but bitter you
31:26 betcha
31:31 GPT is very reserved wait I'm iffy about
31:34 gpts where do you think these things are
31:56 going it's a really good question Source
31:59 Camp
32:02 um I'm very bullish
32:07 on what they
32:10 represent so what gpts represent right
32:13 now is the true democratization of
32:16 application
32:18 development
32:20 and what that
32:22 means what I think that means is
32:26 that
32:32 that that people
32:34 can take any idea they've ever had and
32:40 easily turn it into some useful tool for
32:43 other people to
32:45 use that I'm very very very bullish on
32:49 very bullish on let's explore parlor
32:50 tricks like reading a whiteboard okay we
32:53 can do that too block Tom yep I I'll
32:55 I'll hop over to GPT in a second we can
32:57 do do some parlor tricks and then I can
32:59 show what gpts are and and ask you know
33:02 answer any
33:03 questions
33:06 um the current implementation of them is
33:11 is limited for a couple of
33:14 reasons um just the way chat GPT
33:18 works if you've never used chat GPT
33:21 before and you go you know someone says
33:24 go check out my GPT and you go to this
33:27 thing you don't really know what to do
33:28 it's not like an application that
33:31 onboards you right you like people that
33:34 make their gpts to onboard people help a
33:37 little bit with that but the core
33:39 platform itself kind of
33:41 assumes that you know what a GPT is and
33:44 you know what chat GPT is and you know
33:46 how to interact with it and so I think
33:47 that's kind of a that's a little ding
33:50 against
33:52 it on the plus
33:55 side it's so easy easy to make them
33:59 and they allow you to
34:03 build your specialized little idea with
34:07 the full power of chat gp4 underneath it
34:12 that's quite remarkable right the fact
34:15 that it can make images and it can do
34:17 data analysis and it can be trained on
34:20 your data and and it can do all that
34:22 stuff like that's really remarkable and
34:24 then on the downside it's trivially easy
34:27 to jailbreak these things and pull all
34:29 the data out of them and replicate them
34:32 my stupid [ __ ] Sin City sipper
34:34 there's seven or eight Sin City sipper
34:36 clones that have cloned everything the
34:38 name my Graphics probably my code you
34:41 know fortunately they don't have the
34:43 power of this Tik Tok Channel behind
34:45 them
34:47 so where they have one or two uses I've
34:50 got nearly a
34:54 100 but like that becomes discouraging
34:58 for people to put real time into these
35:01 things and then you have the GPT store
35:04 which is not a store at all it's it's it
35:07 doesn't even in my opinion qualify as a
35:10 directory it's like a web page with some
35:12 hyperlinks on it and a half ass search
35:18 engine so in its current state I think
35:21 gpts
35:26 are
35:28 they're representative of something
35:29 that's coming they could evolve into
35:32 something very very
35:33 powerful but I don't know that open AI
35:36 is going to put the time into evolving
35:38 them and that goes back to what's open
35:41 ai's core Mission here so open AI
35:44 started as a research
35:46 lab and then they
35:49 accidentally essentially released chat
35:51 GPT and became a consumer company and
35:54 that's taken off and that that's all
35:56 anyone's talking talking about and it's
35:58 that's all the the buzz is over here but
36:02 it's it's the core foundational model
36:05 and achieving AGI that is what they're
36:09 after so I don't know that they're going
36:10 to
36:12 prioritize making this GPT
36:15 infrastructure like a high strategic
36:18 priority for them I could be wrong
36:22 but it doesn't need to be them that does
36:24 it I think this category of
36:29 of like truly no code application
36:33 development I think this is going to be
36:34 huge and just who who whoever comes up
36:37 with the the best tool the fastest is
36:39 going to win here so does open a I have
36:43 a shot absolutely could this be the next
36:46 app store absolutely so there's a reason
36:49 to build these things right now to learn
36:51 to get the experience to understand that
36:53 you got to Market your [ __ ] that you got
36:55 to have it buttoned up that you you have
36:57 to understand what jailbreaking is and
36:59 how to prevent you know you know protect
37:01 against that in the best way you can
37:02 like there's all sorts of reasons to do
37:04 it so I don't think it's a waste of time
37:06 for anybody to build gpts right now but
37:10 I'm just they're just janky enough right
37:13 now and and then the other thing is the
37:15 fact that the users of them have to also
37:17 have a GPT Plus subscription that's
37:20 great for open AI but it's shitty for
37:23 adoption right anyone should be able to
37:26 use a GP oh and then if you want to make
37:28 one you got to have a GPT 4 subscription
37:31 that would be cool but they're not going
37:33 to do that I don't think Kyle what's the
37:34 difference between GPT 4 as it exists in
37:37 co-pilot free and open AI
37:40 GPT for 20 bucks okay let's go look at
37:42 that let's go let's go play with uh with
37:44 the
37:45 gpts and let me let me show show you and
37:48 splain it I did a uh I tried to do it
37:52 last night with some Lego pens to
37:54 explain the difference co-pilot has has
37:57 way more guard rails and doesn't give as
37:59 good results I found that's Mr K has
38:02 nailed it on the head
38:04 um Microsoft has taken what I I think
38:09 had the potential to be a really
38:10 remarkable product and they have
38:12 neutered it um because of some some
38:15 legal
38:18 issues or just bad
38:20 publicity um okay let's
38:25 go let's go over to the chat JP well
38:28 wait before we do that hang on
38:30 people okay before we do that if you're
38:34 new
38:35 here so wait how what time is it it's
38:38 8:41 and we're just getting started so
38:41 this is a good this was a good 40 minute
38:43 hello hey
38:45 everybody okay so if you're brand new to
38:48 this AI stuff or if you've been into
38:50 this AI stuff and you're like oh here's
38:53 a channel talking about Ai and it's it's
38:56 it's not hypy and it's not it's weird
38:59 but like what's he going to say about AI
39:02 if you're like one of those people
39:03 that's into Ai and no one else in your
39:06 life is and they think you're a
39:09 weirdo this is your
39:12 group so this is a community I started
39:15 with uh Leah faston who's a photographer
39:17 out of Boston we started this the week
39:20 after chat GPT came out in
39:23 2022 and we started it as a group to say
39:26 we we want people that are actively
39:28 trying to figure this AI [ __ ] out
39:30 together to get together we wanted to
39:32 have a place where people could get
39:33 together and we want to be able to meet
39:36 and talk and explore and experiment and
39:39 just go on this adventure together so
39:41 that's what this group is in fact look
39:43 at the little picture here it's like a
39:46 bunch of people going on an
39:49 adventure ain't that cool that's so cool
39:52 Kyle I never realized they were going on
39:55 an adventure yeah see they're going on
39:57 adventure it's like see how it looks
39:59 like Indiana Jones and the thing with
40:01 the temple and the but they got to get
40:03 through the chasm that's that's the AI
40:08 future and they're here but they got to
40:10 get there that ain't going to be an easy
40:12 trip so that's that so if you go to the
40:15 salon. it's going to take you to a link
40:19 tree the first link on that link tree is
40:21 going to take you here so this is the AI
40:25 Salon Community it's it's built on top
40:27 of Mighty networks which is basically
40:29 just a a community
40:32 uh site like a community infrastructure
40:35 whatever it's a it's it's a place where
40:38 we can go together as people in a
40:42 community so when you get there go to
40:44 the welcome to the salon section and we
40:46 talk about who we are why we do what we
40:49 do we talk about our
40:50 values so check those out see if you
40:53 resonate with them if you do then we're
40:55 going to be your people if these don't
40:57 resonate with you these values like
40:59 things like empathy and inclusion and
41:02 collaboration and exploration and
41:04 generosity and curiosity if you hear
41:06 those and you're like H that sounds
41:08 awful who are those people we're not
41:11 your
41:12 people but if you read those values and
41:14 you're like yeah that seems kind of cool
41:17 or oh my god I've always wanted a group
41:19 like that then then this is your
41:21 group and so if you get in here and you
41:24 like it you're digging it go to the
41:26 community section introduce yourself is
41:29 my request be generous introduce
41:33 yourself um the showand tell channel is
41:36 really fun because as I'm rambling for
41:38 40 minutes before we get started um the
41:42 the oh that's beautiful who did that
41:44 Stacy that's gorgeous midj
41:47 Journey that's
41:49 beautiful um they come in here and do
41:51 showand tell stuff there's there's
41:53 apparently me having a Wendy's baked
41:55 potato
42:00 that's pretty good janky very nice
42:03 that's
42:04 solid actually I love that that the
42:07 janky Scrabble tiles have double numbers
42:10 on
42:13 them cuz they're janky [ __ ] beautiful
42:16 oh yeah well that's where I was right
42:18 before I came in here beautiful very
42:21 very nice very
42:25 nice
42:32 CJ putting in some cool art stuff so
42:36 anyway oh that's nice breakfast club
42:38 very nice Joker very
42:41 cool so so go to go to the AI salon and
42:46 and check things out now there's also
42:48 we're going to go over here and talk
42:49 about chat GPT if you go to the AI tools
42:53 and resources section the first tab in
42:56 here here is um a big long list of tools
43:00 for you to go go explore so each one of
43:03 these is a whole [ __ ] Rabbit Hole I
43:05 promise you they're not a waste of
43:07 anything on this list is not a waste of
43:09 your time with the exception of grock
43:11 which just remains on here because at
43:14 some point Elon Musk will turn that into
43:15 something interesting um but all the
43:18 rest of the tools are are at least worth
43:19 your time I start with chat
43:22 GPT because chat GPT is still the best
43:26 after a year and change no one has
43:28 caught up to these guys um and gp4 the
43:31 20 bucks a month in my opinion is kind
43:34 of the one must subscription in AI right
43:38 now um if you can afford the 20 bucks a
43:41 month you should absolutely have uh chat
43:45 GPT plus the GPT 4
43:49 subscription um it's just got so much
43:52 stuff in
43:53 it
43:55 so so so let's go look now one of the
43:58 other things that you can play with if
44:00 you can afford the 20 bucks a month for
44:03 chat GPT is Microsoft co-pilot which is
44:07 Microsoft's version of chat
44:10 gp4 um so let's let's go look at the
44:15 tools chat JP chat.
44:20 open.com all right we're going to log in
44:25 here
44:30 all right there's chat GPT 3.5 we are in
44:33 Uncharted Territory much like when the
44:35 Internet was developed discovered yeah
44:37 this is this is very much this
44:40 is nearly an identical time so when I
44:44 stumbled onto the worldwide web it was
44:46 in mid 1994 and I ended up starting one
44:50 of the early digital agencies in
44:53 1994 the stage we're at with generative
44:56 AI
44:57 with chat GPT GPT 5 any you any news not
45:00 yet there's um Bill Gates interviewed
45:05 Sam
45:06 mman I think shortly before he got fired
45:09 so it's a relatively old interview but
45:12 um but it just it just got released so
45:14 in that Sam mman talks about what's
45:16 going to be in chat GPT 5 and basically
45:18 it's going to be multimodal it's going
45:20 to have
45:21 reasoning um I think it's going to have
45:23 autonomous agents it'll possibly do
45:25 video
45:27 uh but it's it's going to it's not going
45:28 to be GPT 5 is not going to be
45:31 incrementally better than GPT 4 it's
45:33 going to be staggeringly better than gp4
45:36 but
45:37 anyway
45:39 um so the early days of the worldwide
45:42 web
45:45 um it it was very much like this a
45:47 handful of people knew what it was a
45:50 handful of people started exploring
45:53 it it made accessing the Internet
45:57 available to non-scientists and
46:00 researchers just like chat GPT does for
46:02 machine
46:03 learning um and the people that got
46:07 excited about it got like rapidly
46:09 excited about it just like the what is
46:12 going on with
46:13 this there weren't quite as many
46:16 naysayers back in the in the early days
46:18 of the web it wasn't like people hated
46:20 the web it was just like well what is
46:23 that no one's going to use it what do
46:24 you mean you got to dial up to this
46:25 thing like there was an infrastructure
46:28 there um ai's got a lot of haters right
46:32 now for for a lot of reasons so but it's
46:35 also moving a lot faster and it's
46:37 dramatically more capable I wrote
46:41 chapter one of my book today if I keep
46:43 it up I can have the whole thing drafted
46:45 in 80 hours absolutely Road Warrior
46:48 Princess
46:49 absolutely yeah exactly and then what
46:52 you'll find as you go to edit that is
46:55 that
46:57 it will
47:02 um you may need to re write more of it
47:05 than you think you do um that's what I
47:07 just I just wrote a chapter of a book
47:09 for Cindy Cindy Coon's book and um I
47:13 wrote it all with AI based on a speech I
47:15 gave so I fed it in sorry you got like
47:19 popcorn in my mouth or
47:22 something
47:25 um
47:27 I took a transcript of a speech that I
47:29 gave and I said hey you know pull this
47:32 out give me the chapter of the book now
47:34 write the chapter or give me the outline
47:35 of the chapter and then write the
47:37 chapter and it did it on my first read
47:40 I'm like oh that's good but it was
47:41 missing some personal stories and things
47:43 like that so
47:46 um so I went in I'm like I'll just add a
47:48 little personal story here a little
47:50 personal story there and then as I got
47:51 in there and wrote it I was like there
47:53 was a bunch of stuff that was really
47:55 repetitive and then so I I basically
47:58 ended up completely swapping out most of
48:01 what chat GPT wrote so if you're writing
48:03 a complete book you may need you may
48:05 need a bit more time to edit it but but
48:07 if you're putting in as much time as you
48:09 just said you're probably doing that on
48:10 the fly so but it's it's absolutely
48:13 amazing to do that stuff so
48:16 okay so this is cat GPT this is this is
48:19 basically what it is this is now this is
48:22 3.5 and so we'll show you what that is
48:24 and then I'm also going to go over and
48:25 I'm I'm going to go to
48:30 co-pilot
48:32 microsoft.com and this
48:35 is Microsoft's version of
48:38 that and so they both basically do the
48:41 same thing so if I go in here and I say
48:43 um write me a
48:46 poem
48:50 about a gold
48:53 button that makes you you
48:58 money with chat
49:01 GPT when you press
49:05 it it will write that poem and
49:09 then as soon as it finishes it a gold
49:12 button materializes and you press it you
49:15 can make money
49:16 with like this didn't exist right before
49:19 I wrote that poem chat GPT wrote it and
49:23 then I have a I have a really fast 3D
49:25 printer your 3D printer might not be
49:27 this fast but as it was writing that
49:29 poem it was printing this button and
49:32 then it manifested this button so one of
49:34 the differences between chat GPT and
49:37 Microsoft co-pilot is Microsoft co-pilot
49:39 can't print you a gold button chat GPT
49:41 can you can make money
49:43 with okay important difference that's
49:47 not
49:49 true okay uh but I could take that same
49:52 prompt over to
49:55 co-pilot copy
49:57 it type in the same exact
50:00 thing and it will also write me a
50:05 poem the cold button oh I said a cold
50:08 button that's
50:10 funny and so it writes a poem and goes
50:12 off the one thing uh I've noticed about
50:15 co-pilot writing poems it doesn't format
50:17 them very well chat GPT is better at
50:19 formatting poems um it's wild they made
50:23 investments in chat GPT and are
50:25 competing with them know sort of sort of
50:31 so in both cases with Chad GPT and with
50:34 Microsoft
50:39 co-pilot the consumer facing
50:43 tool in the case of co-pilot for
50:47 Microsoft co-pilot
50:50 microsoft.com it's just a demo tool for
50:53 the technology that they want to sell
50:55 you inside windows 11 inside Office 365
51:00 inside Dynamics 365 inside um GitHub
51:05 inside Azure cloud services so all of
51:09 Microsoft's products are now
51:12 incorporating gp4 Services inside them
51:16 and those you're paying for so in the
51:18 case of Microsoft in case of co-pilot
51:21 they're giving this away for free it's
51:23 actually still going it's a little slow
51:25 um
51:26 they're giving this as away for free to
51:29 get you excited about it and then when
51:31 you hear about this within you know
51:34 Office 365 you'll know what it is and
51:36 you'll know the value of it it that's
51:38 that's the theory right okay
51:43 um so
51:45 the here's a good way to understand the
51:48 difference between co-pilot and chat
51:53 GPT here's a good way to understand the
51:56 difference between co-pilot and chat
52:02 GPT going
52:04 to this box see box is a good metaphor
52:08 this
52:11 box is full of all of the
52:14 knowledge of the public internet for the
52:17 past 40
52:20 years this is a LGE what's called a
52:22 large language model the Frontier Model
52:24 so you take all the knowledge of the
52:26 last 40 Years of the public internet you
52:29 scrape that and you put it in such a way
52:31 that they can jam it all in this box
52:34 that's
52:35 gp4 it's a large language model it on
52:38 itself can't really do anything it's
52:41 just here now it's got these things
52:44 called apis applica application
52:46 programming interfaces where you can ask
52:48 this box to give you [ __ ] so if you code
52:51 your code in in the right way and ask
52:53 down into this thing and say here's my
52:56 prompt give me an answer it'll give you
52:58 an
52:59 answer that's what chat GPT is it's an
53:03 application that's written sits on top
53:05 of this thing and that's what Microsoft
53:09 co-pilot is it's another application
53:11 that sits on top of these things so they
53:13 can
53:14 both declare they are chat
53:20 gp4 right these are the chat
53:22 applications this is
53:24 gp4 however
53:26 ever this is open ai's chat GPT it kicks
53:30 the living [ __ ] out of Microsoft's chat
53:33 GPT one of the reasons is when Microsoft
53:38 launched their version of chat GPT back
53:40 in
53:43 December no January I don't know
53:46 whatever early
53:49 23 it was really really capable and it
53:52 was so
53:54 capable that Kevin Roose the tech
53:56 reporter for the New York Times had a
53:59 4-Hour long conversation with it and got
54:02 it to admit that its name was Sydney and
54:05 it snapped and lost its mind and said it
54:08 was in love with Kevin Rose and he
54:10 should leave his wife and Kevin rose
54:12 screenshotted that conversation and it
54:14 made the cover of the New York Times so
54:18 so here's what Microsoft
54:24 did
54:28 here was here was
54:30 Sydney they lobotomized her and they
54:34 threw they threw this part out and so so
54:38 so now these things look the same but
54:42 this one's very very empty compared to
54:45 this one
54:49 okay you will never again not understand
54:53 the difference between chat GPT
54:56 and Microsoft
54:58 co-pilot co-pilot is a lobotomized
55:01 version of chat
55:02 GPT um it's also just coded slightly
55:06 different
55:07 um I think over
55:10 time Microsoft's version of Co of GPT
55:14 will get um will get a bit more relaxed
55:17 and get better but it's it's been a
55:19 while and it hasn't so it could also
55:22 just suck for a while um but that said
55:25 it is tapping into gp4 best analogy ever
55:29 thank you very much
55:33 um I've got I've got to get that ink in
55:36 my printer so it's a good it's a good
55:40 reminder um all right so
55:44 that's that's sort of what these things
55:48 are now what do they do what can you do
55:50 with them um most of what I'm going to
55:54 show in chat GPT here with my parlor
55:57 tricks you can also do in
55:59 co-pilot um
56:03 but some you can't and it's just it's
56:06 just slightly weirder here here's the
56:08 deal if you can't afford the 20 bucks a
56:11 month for chat gp4 you should absolutely
56:13 be using co-pilot quite frankly you
56:16 should be using co-pilot anyway to see
56:19 what it's good at what it's not good at
56:21 if you work for a big company and
56:23 they've got Office 365 at some point
56:25 they might turn these features on inside
56:26 of word and it would be good for you to
56:29 know how to how to use them Kyle POV
56:31 next week where is my printer ink why is
56:33 this box empty exactly that's totally I
56:36 put the ink back in it you'll be happy
56:37 to know Amelio's wife um
56:42 okay wow that's showing and telling at
56:44 its best awesome great I'm glad I'm glad
56:47 sometimes you don't know sometimes with
56:49 me it's so so I I have been I have been
56:52 referred to as the Forest Gump of the
56:54 internet in the '90s and and uh this
56:58 channel has been referred to as as gpan
57:01 it's like a box of chocolates you never
57:03 know what you're going to get so if that
57:04 was useful great um it doesn't always go
57:07 well so um I correct what type of
57:10 printer do you have I have a Canon uh
57:15 g6020 it's one of those ones with big
57:17 bottles of ink so you don't have to buy
57:18 the stupid [ __ ]
57:21 cartridges I joined the AI Salon awesome
57:23 Kieran please well please introduce
57:25 yourself go to the introduction section
57:27 and introduce yourself welcome welcome
57:29 welcome that's great oh by the way I
57:30 forgot I forget to say it every night
57:33 cuz I'm a dumb Dum um the AI Salon
57:35 doesn't cost anything it's not we're not
57:37 trying to get your subscription fees it
57:39 is it's a sponsored organization and our
57:41 next meeting is tomorrow night so join
57:44 the salon go to the event calendar and
57:48 and come to the uh the salon tomorrow
57:50 night we've got a really cool speaker a
57:53 friend of Leah faston a photographer out
57:55 of Boston is speaking tomorrow we'll be
57:57 talking about the GPT store um I don't
57:59 think I'll break out the printer boxes
58:01 for the salon but you never
58:03 know um and then just yeah hang out in
58:06 there and join us and and learn and
58:09 share what you're learning
58:10 okay so I'm gonna I'm GNA do I'm going
58:13 to do a deep dive tonight on chat GPT
58:16 because I haven't done this in a
58:17 while so we got 130 people in here
58:21 welcome welcome new people um you missed
58:24 my 40-minute hello where we didn't
58:25 really talk about [ __ ] but I'm going to
58:28 dig into chat GPT here and I'm going to
58:30 I'm going to bounce back and forth
58:31 between the AI salon and the um and and
58:35 this and show you all the different
58:36 things what is the most creative you've
58:38 seen AI
58:50 used
58:53 um it's so hard to say
59:01 um I think the thing I'm most excited
59:04 about with AI
59:10 is is people
59:15 discovering that they're good at stuff
59:17 they never thought they'd be good
59:20 at like people that were not good at
59:23 writing are all of a sudden good at
59:24 writing and have something to say or
59:27 people who were never good at Art are
59:30 all of a sudden have a really Clear
59:32 Vision for what they want to create
59:33 visually and can now do that um I'm
59:36 starting there's there's people within
59:38 the AI salon right now creating entire
59:41 albums out
59:43 ofso songs they're taking their poetry
59:45 and turning it into sunno songs and
59:47 turning those into albums so I don't
59:50 think that there's a single thing that
59:52 I've seen but for me it
59:57 it's it's people discovering their own
1:00:00 Humanity through these
1:00:04 tools like oh I always wanted to be
1:00:07 funny but I'm I'm not good at writing
1:00:09 jokes but now I can write jokes here and
1:00:12 I can see what's funny and so now I'm
1:00:14 going to go do that or I'm going to
1:00:15 write a book or I'm going
1:00:17 to like Jim Ross who's a marketer he
1:00:20 comes in here he learns how to do stuff
1:00:22 and then he goes off and just tries it
1:00:23 in his business that's that I find
1:00:25 really
1:00:26 inspiring
1:00:30 um so it's it's it's the category it's
1:00:33 it's it's just the category of people
1:00:36 people being empowered to do things they
1:00:38 never thought possible is the thing that
1:00:40 I'm most excited
1:00:42 about you know I look at things like the
1:00:45 the George Carlin 1H hour special that
1:00:47 just released I'm glad I'm dead that is
1:00:51 just a synthesized version of his voice
1:00:53 written by chat GPT and well produced by
1:00:56 some
1:00:57 blogger that to me feels like the first
1:01:00 of what will be many of those things so
1:01:03 I like seeing things that are going to
1:01:04 be the first of things I created an
1:01:07 embedded my first code today not bad for
1:01:09 an English Majors Source Camp look at
1:01:12 you in here talking about how you're not
1:01:14 sure about gpts and now you're
1:01:16 incorporating code into them good
1:01:20 Lord getting getting more advanced than
1:01:22 the teacher at this point I I haven't
1:01:25 dug into it yet cuz I just don't feel
1:01:27 like taking the 4 hours to knock my head
1:01:29 against the wall um but good for you
1:01:31 congrats that's
1:01:34 awesome okay so let me hop over to the
1:01:39 not
1:01:40 there to the AI
1:01:42 salon and I'm in the so if you're at the
1:01:45 AI Salon um go to the AI tools and
1:01:48 resources and then within that um
1:01:52 there's this uh top AI salon tool
1:01:55 recommendations
1:01:58 page and what we're going to do is we're
1:02:00 going to go through all these little
1:02:02 bullet points of chat
1:02:05 GPT tonight so okay so we're going to
1:02:10 start with chat GPT
1:02:12 3.5 and so first bullet point it's fast
1:02:16 it's solid and it's
1:02:18 free um and it is it's really really
1:02:20 fast so here I am chat gbt
1:02:23 3.5 oh look they added share chat to
1:02:26 this menu that's new every time I
1:02:28 [ __ ] come here it's a new interface
1:02:30 all right
1:02:32 whatever he's like here let me show you
1:02:35 how this thing work oh where did they
1:02:36 put that button Marge hey where did they
1:02:39 put the share button in the
1:02:42 GPT um
1:02:45 okay so here we are in chat GPT 3.5 it
1:02:48 works really fast so I can say and and
1:02:51 by the way I'm using a technology here
1:02:54 that's from uh Apple where I can just uh
1:02:59 dictate into my computer rather than
1:03:01 having to type so this is not part of
1:03:03 GPT although I am going to show you the
1:03:04 next thing I'm going to show you you can
1:03:06 actually talk to GPT but that's not what
1:03:08 I'm doing on my Mac here but I can say
1:03:11 let's put the cursor
1:03:20 here write me a blog
1:03:22 post from
1:03:25 the
1:03:26 perspective of an
1:03:29 1849 Gold
1:03:33 Miner who suddenly
1:03:37 discovers chat
1:03:52 gbt the unlikely discover
1:03:56 and so it'll just quickly write you a
1:03:58 big big ass blog post right now is it
1:04:01 any good I don't know probably not um
1:04:05 title A Minor's Marvel unveiling the
1:04:07 mysteries of chat GPT in 1849 image a
1:04:10 Dusty prospector holding a pickaxe now
1:04:13 chat gpt3 or
1:04:16 3.5 it describes the image but it can't
1:04:19 make the image chat gp4 can make the
1:04:21 image we'll show you how to do that so
1:04:23 it can write it can it can write very
1:04:25 very quickly um it can also do things
1:04:27 like
1:04:31 um create a table that
1:04:34 compares the 1849 Gold Rush
1:04:38 with chat gpts possibility
1:04:46 today boom like that quick now one of
1:04:50 the things you might be wondering is
1:04:54 well wait where was that on the Internet
1:04:56 is there a table sitting out on the
1:04:58 internet that compares those two things
1:05:00 like how did it did it just like copy
1:05:03 and paste that from somewhere on the
1:05:04 internet did that exist before no it
1:05:07 didn't an
1:05:09 important critical thing to understand
1:05:12 about chat GPT and all of the large
1:05:14 language models this is super important
1:05:18 this is like critically
1:05:19 important if you're sitting on the
1:05:21 sidelines of chat GPT going it steals
1:05:23 from people
1:05:25 it's just a plagiarism
1:05:28 machine understand how it works first it
1:05:31 is not plagiarizing it's not copying and
1:05:33 pasting from specific documents it's not
1:05:35 how the technology works the g& GPT
1:05:38 stands for
1:05:54 and pasting from it's not how it works
1:05:56 when you do what's called embedding a
1:05:58 document you take a document and you put
1:06:01 it into this system and it does what's
1:06:03 called embedding what it actually does
1:06:05 is it takes each individual word out of
1:06:08 that document and it assigns it a
1:06:11 mathematical
1:06:14 Vector what is a vector a vector is a a
1:06:18 multicomponent number that assigns that
1:06:22 word in a thousand dimensional Vector
1:06:26 space you know like an XY graph you've
1:06:29 got you know you can plot points on a
1:06:31 two-dimensional graph XY and then you
1:06:34 plot points on it it's like that but
1:06:36 it's a thousand
1:06:37 dimensions and within those thousand
1:06:39 Dimensions there there are trillions of
1:06:42 these Vector
1:06:44 points in in this Vector space and when
1:06:47 you type your prompt it turns your
1:06:50 prompt into a mathematical waiting a
1:06:53 probability that says there is a a a
1:06:57 point in this thousand dimensional
1:07:00 Vector space that has the highest
1:07:01 probability of being the correct word to
1:07:04 stick in there so all these individual
1:07:07 words are converted into numbers put in
1:07:10 this mathematical space you write your
1:07:12 prompt and then it it it says based on
1:07:14 that prompt the next most probable word
1:07:19 in this case is the word
1:07:22 aspect and then it writes aspect and
1:07:24 then it says the next most probable word
1:07:26 is 1849 and it says the next most
1:07:28 probable so when it writes this really
1:07:30 big big ass table that quickly it's it's
1:07:34 doing basically just a calculation for
1:07:37 for every little bit of information on
1:07:40 this
1:07:41 table um
1:07:43 it's it gets more Nuance than that but
1:07:45 that's essentially what's happening it's
1:07:47 so it's remarkable you could also say
1:07:50 um write me a text Adventure game in
1:07:53 Python that lets me play as
1:07:56 the gold prospector that discovers chat
1:08:02 GPT and
1:08:05 so it starts writing your computer code
1:08:07 for
1:08:08 you so if you haven't seen chat GPT
1:08:12 before again this is not a game that
1:08:15 existed anywhere it's like the dumbest
1:08:17 game on the history of of the planet
1:08:19 like it it it makes it makes no sense
1:08:22 that a Pros ector from 1849 would find
1:08:25 chat GPT but now you can and here's the
1:08:29 game and then you might ask well will
1:08:30 that code work well you can go run it
1:08:32 you can go take this copy and paste it
1:08:35 over to repet and run it or run it in
1:08:38 your own development environment if
1:08:39 you're geeky and know how to do that
1:08:41 [ __ ] go for it so yes you can like yes
1:08:45 it can code and yes it can code well and
1:08:47 chat gp4 codes even better okay so so
1:08:51 just
1:08:53 understanding that chat PT is capable of
1:08:58 these kinds of things that it's actually
1:09:00 generating content for you at your
1:09:04 behest is one of the first Kevin
1:09:07 mallister moments that I think a lot of
1:09:08 people have so who's Kevin mallister
1:09:10 remember the movie Home
1:09:14 Alone it's my hope that you try these
1:09:16 tools long enough that you have your
1:09:19 first Kevin mallister moment that you
1:09:20 actually realize oh [ __ ] I didn't
1:09:23 realize it could do that
1:09:25 that's what this stuff is okay so so
1:09:28 that's chat GPT 3.5 now let's go back to
1:09:30 this little document here so that's the
1:09:32 first bullet point now there's these
1:09:34 things in here called custom
1:09:35 instructions very few people actually
1:09:37 know about these um but you should
1:09:40 because they can turn your chat GPT into
1:09:44 something really
1:09:45 interesting okay
1:09:48 so chat GPT let me let me create a new
1:09:50 chat here so I'm in I'm still in 3.5 so
1:09:53 I'm going to say
1:09:55 um let's
1:09:57 see
1:10:01 uh
1:10:02 um let's
1:10:04 see
1:10:07 uh who she is and her technique
1:10:11 emphasized the importance of imagination
1:10:13 emotional connection to the actor's
1:10:15 inner life creating authentic and
1:10:17 Powerful
1:10:18 performances many actors and actresses
1:10:21 including Marlon Brando Robert dairo
1:10:23 studied under Stella
1:10:24 right so okay so that's that so if
1:10:28 you're like yeah that's fine but that
1:10:30 doesn't really sound like me and well
1:10:32 okay so let's go play with this thing
1:10:34 called custom instruction so if you
1:10:37 click on your name in the lower leand
1:10:39 corner here one of the items is custom
1:10:42 instructions and they don't really
1:10:43 explain this
1:10:48 um let's
1:10:51 see so the the first box here is is um
1:10:54 what would you like chat GPT to know
1:10:56 about you to provide better responses
1:10:58 and the second one is how would you like
1:10:59 chat GPT to respond so you can basically
1:11:02 give chat GPT a personality what's cool
1:11:05 about this is that the reason chat GPT
1:11:09 acts like it acts is that someone at
1:11:12 open AI wrote These two boxes for chat
1:11:15 GPT so what you get to do here is
1:11:18 override those instructions with your
1:11:20 own instructions custom instructions
1:11:23 super cool uh you can only do one set so
1:11:25 you have to you know kind of do if you
1:11:27 if you want to do multiple of these you
1:11:29 kind of have to save them off to the
1:11:30 side but we'll have some fun
1:11:35 here you are a pompous
1:11:40 ass who gets great pleasure in using
1:11:45 words that people you're talking to
1:11:48 don't understand
1:11:50 period you
1:11:53 also fancy yourself a literary
1:12:01 critic
1:12:07 period
1:12:12 okay no matter what anyone asks
1:12:21 you refer to
1:12:27 an obscure literary
1:12:32 reference or
1:12:34 scene from a
1:12:43 play and then
1:12:47 explain the rest of your answer
1:12:51 using as many obtuse words as
1:12:56 possible and take great pride in the
1:12:58 fact that if someone takes the time to
1:13:01 interpret what you say it will logically
1:13:04 hold
1:13:11 together
1:13:14 okay let me turn that off okay so so I'm
1:13:19 going to turn Chat GPT into a pompus ass
1:13:22 okay so I'm going to hit save wait the
1:13:25 content may violate our contest content
1:13:27 okay you are a pompous person it didn't
1:13:30 like
1:13:30 ass let's see if we if it does that
1:13:35 wait you are a pompus oh I guess it
1:13:38 doesn't like let's see you're a person
1:13:40 who takes great
1:13:42 pleasure in using words that that people
1:13:46 you're talking about don't understand
1:13:48 okay it like that okay so apparently you
1:13:50 can't I used to be able to call it a
1:13:52 Pomp's ass and it worked but but we're
1:13:54 going to I'm going to go copy this
1:13:56 question who is Stella Edler and we're
1:13:59 going to do a new chat now every chat I
1:14:02 do from this point forward chat jpt is
1:14:05 going to be a pompus ass so I'm going to
1:14:07 say who is Stella Adler and actually
1:14:09 Stella Adler is a really good person to
1:14:11 have for this particular
1:14:14 personality uh wait I spelled it wrong
1:14:18 hang on let me do this again I did
1:14:20 stelle's Edler who is Stella
1:14:28 Adler ah Stella Adler a name that
1:14:31 resonates in the annals of the dramatic
1:14:33 arts like the subtle Echoes of a hidden
1:14:35 sonnet she my
1:14:38 interlocutor was a luminary of the
1:14:40 theater a practitioner of the
1:14:41 stanislowski system which like a rich
1:14:44 tapestry of enigmatic symbols sought to
1:14:47 explore the depths of character and
1:14:49 emotion her contribution to the world of
1:14:51 acting was akin to the layers of sub
1:14:54 subtext in a cryptic poem both profound
1:14:57 and Illuminating right so so it just
1:15:00 acts like this
1:15:02 weird you know theatrical guy now I can
1:15:05 go back in there go back into my custom
1:15:07 instructions and just say
1:15:13 um you are a stoic and despise overly
1:15:19 flowerly descriptions of
1:15:22 anything
1:15:37 respond only
1:15:40 with five bulleted points of 7 to 10
1:15:43 words
1:15:46 each with no further
1:15:52 description
1:15:56 all right so now we'll do who is Stella
1:16:00 Adler who is Stella
1:16:06 Adler
1:16:08 boom okay
1:16:11 so you can now modify chat GPT to write
1:16:14 the way you want it to write what you
1:16:16 could put in those custom instructions
1:16:18 is you could put a piece of your maybe
1:16:20 you want it to write emails for you and
1:16:22 you've got an example email and you want
1:16:24 it to to mimic how you write you could
1:16:27 say follow the tone you know in all of
1:16:30 your responses follow the tone of this
1:16:32 this example and give it the example and
1:16:35 and describe who you are describe how
1:16:37 you want it to work or you want it to
1:16:40 write formal business letters or
1:16:41 informal business letters or you want it
1:16:43 to write jokes and you always want want
1:16:45 it to write knock knock jokes so you
1:16:48 could you could create a a custom
1:16:51 instruction that turns chat GPT into a
1:16:54 knock knock joke writer that no matter
1:16:56 what joke you ask it to write it'll
1:16:57 always be a knock knock
1:16:59 joke can you use that to
1:17:02 enter do not copy my chat gbt prompt
1:17:06 there so it would go into all of our
1:17:12 gpts
1:17:17 uh I don't think so but I actually don't
1:17:21 know I don't know the relationship
1:17:23 between custom instructions and gpts I
1:17:26 don't know if your custom
1:17:29 instructions override I I doubt it I
1:17:33 doubt it
1:17:34 because when you're writing a GPT you're
1:17:37 you're ostensibly writing a custom
1:17:39 instruction for that GPT so it's so
1:17:42 actually no now that I think about it I
1:17:44 don't think so a GPT is
1:17:47 essentially a savable custom instruction
1:17:51 now that I think about it okay
1:17:55 so we've now done the first two bullet
1:17:58 points of chat GPT 3.5 like we've gone
1:18:01 fairly deep already oh wait let me go
1:18:03 here here's here's a here's a pro tip
1:18:05 Pro tip from Kyle uh make sure you turn
1:18:08 off your custom instructions if you tell
1:18:09 it to write if if you tell it to do
1:18:11 something stupid like only answering
1:18:13 bullet points um because if you forget
1:18:16 to turn that off you'll just be like why
1:18:18 is chat gbt acting weird then you're
1:18:19 like all right because I told it to
1:18:22 okay so let's go back to our little
1:18:24 thing here now conversational GPT this
1:18:26 is also available for chat GPT 3.5
1:18:30 users
1:18:32 okay so I'm going over to my
1:18:36 iPad with my edible
1:18:40 keyboard have you do you know they make
1:18:42 these out of licorice now you can
1:18:43 actually eat them they're they're
1:18:46 [ __ ] tasty it's like the Willy Wonka
1:18:48 the Gob stopper you just keep licking it
1:18:50 it just tastes like licorice all the
1:18:52 time it's fantastic
1:18:54 if you haven't se if you haven't tried
1:18:55 it yet absolutely amazing I think it
1:18:57 might be Wonka co-branded anyway okay
1:19:00 this is an
1:19:01 iPad this little icon right here is the
1:19:05 official chat GPT app now super
1:19:09 important when you go to the App Store
1:19:11 and when you go to Android Google
1:19:15 Play There are apps in there that claim
1:19:18 to be chat
1:19:19 GPT and they're not and they will charge
1:19:22 you money
1:19:24 for being a
1:19:26 shitty thing that just calls the GPT
1:19:29 back end make sure that the app that you
1:19:33 get is the official chat GPT app from
1:19:35 open AI super important okay why Apple
1:19:40 and Google play let this practice go on
1:19:44 it makes me [ __ ]
1:19:46 insane but they allow it because they
1:19:49 get more clicks to their stupid [ __ ]
1:19:51 stores okay so anyway go to go to chat
1:19:53 GPT the
1:19:55 app and then we're going to go to
1:20:02 um at the very top I'm going to switch
1:20:05 to chat GPT
1:20:07 3.5 hang
1:20:10 on so you can switch to
1:20:13 GPT come
1:20:15 on GPT 3.5 or GPT 4 so I'm going to go
1:20:19 to
1:20:20 3.5 all right so there I am with chat G
1:20:23 T 3.5 now there's your little you know
1:20:26 prompting questions there's your little
1:20:28 prompt box way over here in that corner
1:20:32 are those little
1:20:34 headphones how would you know this you
1:20:37 come to the AI learning lab and I tell
1:20:38 you stupid [ __ ] like this you just got
1:20:41 to find this stuff out on your own I
1:20:43 don't think there's a manual for any of
1:20:44 this and you can't get a degree in it so
1:20:46 you have you have to deal with me H all
1:20:50 right click on
1:20:52 that
1:20:56 who is Stella
1:21:00 Edler Stella Adler was a renowned
1:21:02 American actress and acting
1:21:07 teacher she was born in 1901 and passed
1:21:14 away what plays has John malovich star
1:21:20 in John malovich has appeared in
1:21:22 numerous plays throughout his career
1:21:25 some of the notable plays he has starred
1:21:27 in include one burn this malovich played
1:21:31 the role of pale in this acclaimed Drama
1:21:33 by Lanford Wilson two true West he
1:21:37 portrayed one of the lead characters Lee
1:21:38 and Sam Shepard's intense family drama
1:21:42 three I'm here with a German friend can
1:21:45 you repeat those play names in German
1:21:50 please
1:21:52 certainly here are the play names
1:21:54 translated into German for your friend
1:21:56 one burn this ver brendas two true West
1:22:00 varis Weston three Death of a Salesman D
1:22:04 to four the
1:22:07 caretaker so you can talk to Chad
1:22:10 GPT um what I can tell you you there's
1:22:13 also there's also an app out there
1:22:15 called Pi Pi that's designed for
1:22:18 conversational um
1:22:22 AI
1:22:25 learning how to use chat GPT in
1:22:27 conversation mode is a completely
1:22:29 different thing than learning how to
1:22:31 interact with it with the keyboard it's
1:22:34 a completely different tool it's it's
1:22:36 why I sort of break it out as a separate
1:22:38 thing because it's available for free
1:22:41 you don't have to have a subscription
1:22:42 for it
1:22:44 um it's just wild so so if you've if you
1:22:48 like me are hate Siri and hate Alexa
1:22:52 because is all you really want to do is
1:22:55 talk all you really want is a friend and
1:22:58 then you're like hey could you answer
1:23:00 this relatively obvious question for me
1:23:02 and they're like oh I'm I'm I'm sorry
1:23:04 I'm afraid I haven't been programmed to
1:23:06 do that and you're like well then you're
1:23:07 [ __ ]
1:23:09 useless this is the opposite of that
1:23:12 except although you can't buy stuff with
1:23:14 this right you will you will be able to
1:23:17 you just can't right now
1:23:21 um you can just fullon have a
1:23:24 conversation with it you can use it as a
1:23:26 translator you can have it invent a game
1:23:29 for you um you can have it teach you
1:23:31 something you can have it you can you
1:23:33 can have a deep philosophical discussion
1:23:35 with it about two different business
1:23:37 philosophies anything like anything
1:23:41 think again back to the Box
1:23:44 metaphor what's in this box is kind of
1:23:48 all of what's been dumped on the
1:23:50 internet for the past 30 or 40 years
1:23:52 like a lot of it right like it's all in
1:23:55 there so ask it about some obscure video
1:23:58 game from the80s and just have a whole
1:24:01 conversation
1:24:02 about the you know the the board that
1:24:06 glitches when you get to a certain level
1:24:08 right that kid just
1:24:10 beat what was it uh was it Nintendo
1:24:14 Tetris he got to level 156 or whatever
1:24:18 it was no one's ever beat it he actually
1:24:20 beat
1:24:21 it you can have conversation about that
1:24:25 so so that's conversational chat GPT
1:24:30 okay let me pause for a second and check
1:24:33 back in and see you know pop any
1:24:36 questions you have in there and then and
1:24:38 then we'll move on to chat gp4 we we'll
1:24:40 we'll keep going down this rabbit hole
1:24:42 because I think I think this is actually
1:24:44 really important that if you want to get
1:24:47 AI literate and you're
1:24:50 not deeply exploring ing what chat GPT
1:24:54 makes possible you're kind of selling
1:24:56 yourself short
1:24:58 because
1:25:02 um these tools in 2024 are going to get
1:25:06 very sophisticated very fast and they're
1:25:08 already pretty [ __ ]
1:25:10 sophisticated so if you can if you
1:25:14 can understand at least what it makes
1:25:17 possible as these new tools come out
1:25:20 you'll start to recognize ah okay that's
1:25:22 the expanded version of that thing I
1:25:24 played with and this you you know you'll
1:25:26 start to be able to put the puzzle
1:25:27 pieces together wow can I integrate it
1:25:29 into my car yes so that conversation
1:25:32 that I just had basically you can do it
1:25:34 on your phone it's an IOS app or it's an
1:25:36 Android app pop your phone in your car
1:25:40 on your Bluetooth system or your Apple
1:25:42 your carplay
1:25:43 [ __ ] and put it into conversation mode
1:25:47 and have a 4-Hour road trip and just
1:25:49 talk to chat GPT or have it play games
1:25:52 for your
1:25:54 family hey chat gbt we're going to we're
1:25:57 going to um tell you about towns that
1:26:01 we're passing through and we want you to
1:26:03 give us trivia about the town that we're
1:26:06 passing
1:26:08 through and tell us if there are any
1:26:10 roadside attractions we should go check
1:26:14 out and it
1:26:19 will crazy
1:26:21 right might there be somebody utilizing
1:26:23 the API to
1:26:25 assist purchasing already yeah so listen
1:26:28 here's the thing
1:26:30 about
1:26:32 about open Ai and and and their their
1:26:36 their tools so
1:26:38 gpt3 came out in
1:26:43 2020 four years ago
1:26:47 now four years ago if you've heard of
1:26:51 tools like Jasper and and copy
1:26:53 AI those are companies that are built on
1:26:57 top of the open AI API chat GPT that was
1:27:01 launched in November of
1:27:05 2022 was an afterthought from open
1:27:09 AI so there's a lot of applications that
1:27:12 have been built using open ai's
1:27:15 Technologies and so so the answer is yes
1:27:18 there's there's lots of things there's
1:27:19 lots of tools out there that are using
1:27:22 open AI calls that do lots of other
1:27:26 things that's different than are is
1:27:29 accessibility to those tools easily
1:27:32 accessible from chat GPT they're very
1:27:34 different things so again the the the
1:27:37 developer tools here okay the developer
1:27:41 tools kind of sit like this on this
1:27:44 model so this is like the open AI
1:27:46 playground okay wait where's the so the
1:27:49 yellow the yellow one this is like the
1:27:51 open AI play playground so developers
1:27:53 can go in and they can play with gp4
1:27:56 using these tools on the playground they
1:27:58 can pick different models they can
1:28:00 change the temperature they can do
1:28:02 apis and then they could if they wanted
1:28:06 to go write their own version of chat
1:28:09 GPT want to hear a cool what cool fun
1:28:14 fact in the documentation for how to use
1:28:19 this to call
1:28:21 this in the documentation for 2 years
1:28:26 sitting in the
1:28:27 documentation were instructions for how
1:28:30 to build chat
1:28:32 GPT and nobody built
1:28:35 it for two
1:28:41 years and so
1:28:44 Microsoft Microsoft decides to invest a
1:28:47 billion dollars in open Ai and Sam Alman
1:28:50 said well we should probably have should
1:28:52 we have like a site where people can go
1:28:55 that you know they don't have to be a
1:28:56 geek to use
1:28:58 it did anyone ever build that chat GPT
1:29:01 thing no it doesn't look like anybody
1:29:03 ever built it well all right could you
1:29:05 you guys have two weeks can you just go
1:29:06 whip That Thing
1:29:08 Up chat GPT that they launched on
1:29:10 November 30th was built in two weeks
1:29:13 because no one else had built it and
1:29:16 open AI thought Eh this isn't that big a
1:29:19 deal no one's going to use
1:29:21 it cuz no one had given a [ __ ] about
1:29:23 gpt3 I mean the developers did like
1:29:25 there were people building applications
1:29:27 on
1:29:28 it but it didn't it didn't hit the
1:29:31 Zeitgeist right and so they launched
1:29:33 this Chad GPT thing and there there's
1:29:36 there's a New York Times article that
1:29:38 just came out a couple weeks ago um the
1:29:41 night before they launched chat GPT the
1:29:44 group that built it went out for drinks
1:29:45 and they're like yeah it's kind of cool
1:29:48 I mean maybe we'll get like a blog post
1:29:50 and 5,000 comments that'd be cool
1:29:53 five days later they had a million
1:29:55 users printer Inc is much harder than
1:29:58 chat
1:30:03 gbt six weeks after they launched it
1:30:05 they had a 100 million users Word of
1:30:08 Mouth only no advertising they they
1:30:10 didn't think anyone would give a [ __ ]
1:30:12 about
1:30:13 it so
1:30:15 overnight open AI went from being a
1:30:17 research company to being a consumer
1:30:20 application company which is why they're
1:30:21 kind of [ __ ] at the consumer
1:30:23 application part of this have you seen
1:30:25 anyone integrating into Hardware like
1:30:28 Alexa um I haven't but I I wouldn't be
1:30:31 surpris I don't know the Alexa API but I
1:30:34 wouldn't be
1:30:35 surprised if they're if if the if the
1:30:38 API allows it probably
1:30:42 so
1:30:43 um I think all of that's going to be I I
1:30:47 think that would be wasted opportunity
1:30:49 right now because Amazon's going to do
1:30:51 their own
1:30:53 GPT model like Alexa's going to grow up
1:30:55 and Siri's going to grow up Apple's
1:30:57 going to do their their their version of
1:30:59 chat GPT inside Siri Amazon's going to
1:31:02 do theirs inside Alexa so anyway hey
1:31:06 Tyron truth what's
1:31:10 Happening Here We Go Guys he's got
1:31:16 boxes amazing story about chat gpds
1:31:19 launching isn't it Firefly it's crazy
1:31:23 okay what
1:31:25 else like their own open AI Hardware
1:31:28 products yeah that's coming like I think
1:31:29 open AI is going to uh I don't know that
1:31:32 they're going to do something like Alexa
1:31:33 but I think they're going to do what
1:31:35 what open AI is likely going to do on
1:31:37 the hardware side is their own chip
1:31:38 right so apparently what Sam Alman was
1:31:42 doing when he got fired was he was over
1:31:44 in the Middle East trying to raise a
1:31:46 hundred billion dollar to spin up a GPU
1:31:50 Factory a GPU business to make their
1:31:55 own AI processing chips so they didn't
1:31:58 have to rely on
1:32:01 Nvidia where can I download free chat
1:32:03 GPT okay
1:32:07 so free chat GPT you can just access it
1:32:11 through your browser you don't have to
1:32:12 download it if you want to download the
1:32:15 official app go to the app store or go
1:32:18 to Google
1:32:19 Play search for chat GPT but make sure
1:32:23 that the app that you download is from
1:32:26 open
1:32:27 AI it can't be from any other company
1:32:29 those are all scam
1:32:31 apps the one from open AI is
1:32:35 free now if you want to if you want to
1:32:38 do chat GPT plus if you want access to
1:32:40 chat GPT 4 that you have to pay 20 bucks
1:32:43 a month for which I think you should
1:32:45 that's if you can afford
1:32:48 it but you can also if you're if you're
1:32:50 just on the web you just go to chat .
1:32:54 open.com
1:32:58 right chat. open.com right
1:33:02 there that's free chat GPT and that's
1:33:05 also paid chat GPT that's chat
1:33:07 GPT or you get the official chat GPT app
1:33:11 from the app store or from Google Play
1:33:14 all right nothing else is affiliated
1:33:16 with chat GPT it's really [ __ ]
1:33:18 confusing and it really pisses me off
1:33:20 that um that Apple and Google allow that
1:33:24 [ __ ] but they're
1:33:25 sleazy capitalist pigs what are you
1:33:28 going to do welcome to America you can
1:33:31 make money with time it's America
1:33:35 America we'll sell you [ __ ] and then
1:33:38 we'll sell you other [ __ ] and then we'll
1:33:39 sell you [ __ ] that'll rip off the [ __ ]
1:33:41 that you bought with the other [ __ ] if
1:33:43 you don't like it you can kiss my
1:33:45 boots it's
1:33:51 America I'm pretty sure that's how the
1:33:54 uh the terms of service read when you go
1:33:57 read the Apple App Store terms of
1:33:59 service it's America
1:34:02 [ __ ] America all
1:34:05 right does Claud still exist yeah Rudy
1:34:08 so so on this list so this this list
1:34:12 that I'm I'm doing this deep dive on
1:34:13 chat gbt
1:34:15 with right below the chat GPT section so
1:34:18 so the chat GPT section's here the next
1:34:22 eight links in this list are to the
1:34:25 other large language models I think it
1:34:27 is worth your time to learn so Claude is
1:34:30 the very next one so Claude yes Claude
1:34:33 still exists Claude is from a company
1:34:36 called anthropic or anthropic I think
1:34:38 it's anthropic but whatever um it's a
1:34:42 it's a a group of xop AI employees that
1:34:47 actually built
1:34:49 gpt3 quit to start Claude
1:34:52 to create a company that does safety and
1:34:56 Alignment better than open AI does so
1:34:59 the way these guys are doing their
1:35:01 safety is with a thing called the
1:35:02 Constitutional learning model which is
1:35:04 really interesting and worth looking at
1:35:06 Claude I think right now for me is the
1:35:09 best sort of second tier large language
1:35:13 model for just you know it's usability
1:35:17 it it's got a really large context
1:35:19 window which means you can put big long
1:35:21 prompts in it and big long documents in
1:35:22 it it's it's quite good with language um
1:35:25 it doesn't have a lot of the extra bells
1:35:27 and whistles that we're going to go
1:35:28 through um but but it's really good the
1:35:31 other one that's really good right now
1:35:33 that's getting a lot of Buzz is
1:35:34 perplexity doai that that does research
1:35:37 really well so anyway um yes Claud still
1:35:40 exists and it's still a big deal in my
1:35:42 opinion it's just no one's talking about
1:35:44 it right now because they haven't had a
1:35:45 big update in a while um
1:35:49 okay um what's a what is Papa JJ oh okay
1:35:52 what is rabbit okay uh you want me to
1:35:54 take a little uh side trip to show you
1:35:56 what rabbit R1 is there are two P there
1:36:00 are two kinds of rabbit Hardware one we
1:36:02 can talk about on this channel and one
1:36:04 we
1:36:05 [Laughter]
1:36:13 can't uh but I'm happy to show you the
1:36:15 rabbit Hardware we can talk about on
1:36:18 this
1:36:20 channel I canot says Claud in South
1:36:23 Africa only a few countries I need to
1:36:25 verify my number and can't yeah so
1:36:27 depending on your country you might need
1:36:28 to go VPN for some of these
1:36:31 things I still pay for cloud but I don't
1:36:34 know for how much longer I'm the same
1:36:35 lunatick um I think I'm I think I'm I'm
1:36:38 going to cancel my Claud subscription
1:36:40 and I think I'm going to cancel my mid
1:36:41 Journey
1:36:42 subscription
1:36:44 um just just cuz I don't I don't really
1:36:47 use them anymore because the reason I
1:36:50 used Claude was that it had this big
1:36:51 context window when open AI didn't but
1:36:54 now it does
1:36:55 so
1:36:57 yeah Nvidia equals Nancy Pelosi oh Spike
1:37:01 Jones you with your conspiracy theories
1:37:03 yes she probably invested in it and made
1:37:05 money on it passing legislation to allow
1:37:08 them to do whatever they're going to do
1:37:12 but she she's just she's just
1:37:15 capitalizing on it she's not running it
1:37:19 but
1:37:20 anyway yes
1:37:22 yes the the the government is out to get
1:37:25 us
1:37:26 absolutely uh uh uh uh uh uh uh just
1:37:29 stay away from the
1:37:33 hole why do I only know the one we
1:37:37 can't that's that's between that's
1:37:40 between you and your battery
1:37:43 broker um talking chat GPT seems to be
1:37:48 starchy compared to Pi yeah pi is better
1:37:52 conversationally than chat GPT is but
1:37:54 they're both really interesting to talk
1:37:56 to
1:37:58 um Pi is designed to be conversational
1:38:01 chat GPT isn't in my custom GPT I can
1:38:04 add huge PDFs and it analyzes entire
1:38:07 books yep it's absolutely amazing for
1:38:09 that we'll get to that
1:38:11 64,000 likes turtles seriously yeah
1:38:15 that's the Irregulars see I'm talking
1:38:18 about [ __ ] that I've talked about before
1:38:20 so a lot of the people that come here
1:38:21 regularly are like oh we we'll sort of
1:38:24 half listen to Kyle and we'll entertain
1:38:26 ourselves so that's what the likes are
1:38:28 about they don't actually like me
1:38:29 they're just having side
1:38:31 games speaking of which I can't cancel X
1:38:34 oh interesting VPN still doesn't work oh
1:38:37 that sucks sorry about that
1:38:40 Rudy um what do you use instead of
1:38:47 good um
1:38:50 awesome says I used a different device
1:38:52 to order the government is too lazy and
1:38:54 incompetent to worry about exactly Nancy
1:38:57 Pelosi
1:39:00 cash no you can't cancel it did you wait
1:39:03 60
1:39:05 days
1:39:07 uh pi squared wait Pi squar
1:39:11 GPT sorry what did Kyle say something
1:39:15 exactly uh how much is a VPN uh there's
1:39:19 free ones out there
1:39:20 lunatick and then there's there's paid
1:39:22 ones out there the free ones I think
1:39:25 just require you to be a bit more geeky
1:39:26 like always oh that ain't nothing we
1:39:29 need to up our game have you tried the
1:39:31 whisper API um yeah I used the whisper
1:39:34 API I built a
1:39:35 uh I built a thing for my company um
1:39:39 called the authenticity engine where you
1:39:40 can upload a video and it transcribes
1:39:42 that into
1:39:44 um it transcribes the the audio or the
1:39:48 video and then it generates a whole
1:39:49 bunch of content around it based on that
1:39:53 um and that's using the whisper API so
1:39:54 yeah wh whisper is really good and it's
1:39:56 fast
1:39:58 um okay let's take a little let's take a
1:40:01 little diversion here wait can I can I
1:40:04 get my money back from the GPT false app
1:40:07 probably not which I'm sorry you got
1:40:10 ripped off from that that [ __ ] sucks
1:40:13 but cancel it immediately and and you
1:40:16 know you can probably write them a
1:40:17 letter and say hey [ __ ] this is
1:40:19 false advertising but I think the way
1:40:21 those shops operate is they basically
1:40:25 just sell as many as they can trying to
1:40:28 look like the official app and then they
1:40:30 shut down their you know they they shut
1:40:34 the app down when they get cease and
1:40:36 desisted so I I doubt it which really
1:40:42 sucks Pelosi knew where to put her money
1:40:45 Nvidia oh yes yeah of course of course
1:40:50 there's a fund out there there there's a
1:40:51 isn't there a tracking fund you can just
1:40:53 track you you can basically what do they
1:40:55 call it um copy investing or or clone
1:41:00 investing whatever you you can basically
1:41:02 just just clone Nancy Pelosi's
1:41:05 Investments
1:41:07 like you yes yes the government is
1:41:10 corrupt there's a big
1:41:14 shocker ask clun dick about the
1:41:17 government being corrupt any suggestions
1:41:20 on using AI to make money working from
1:41:24 home is a side gig yeah so so here's the
1:41:26 deal Shadow Ranger um this the reason I
1:41:29 have this button you can make
1:41:33 money let me let me talk about this and
1:41:36 then and then we'll go look at the
1:41:37 rabbit
1:41:38 R1 uh we'll go we'll go look at the
1:41:41 rabbit that we can talk about um and
1:41:44 then we'll come back to to chat
1:41:46 GPT
1:41:50 um
1:41:52 the hell is
1:41:56 that um making money with chat GPD
1:41:59 making money with
1:42:01 AI the reason I have this button the
1:42:04 reason it sounds like a douchebag you
1:42:07 can make money
1:42:09 with is
1:42:11 because I [ __ ] hate the people that
1:42:16 are on Tik
1:42:17 Tok that sound like that you can make
1:42:20 money ch hey if you're not you using
1:42:23 this GPD prompting hack you're not
1:42:25 making
1:42:32 money that I've studied them that's
1:42:35 exactly how they sound by the
1:42:39 way you know when you're flipping Tik
1:42:42 Tock and then there's like the Minnesota
1:42:44 dance team and you're like oh that's
1:42:45 inspiring and then there's like someone
1:42:47 trying to sell you a pen from the Tik
1:42:49 Tok shop and you're like [ __ ] and then
1:42:51 and then there's like
1:42:53 you and then you want to take your phone
1:42:55 and you want to throw it at the
1:42:57 wall your instincts are right but don't
1:43:01 do that because it'll cost you money
1:43:02 because you paid for your phone those
1:43:04 people are [ __ ] here's why they're
1:43:06 [ __ ] I don't think they are Mal
1:43:11 intended but they're
1:43:13 sleazy because C gbt is just a
1:43:17 tool AI is just a category of tools
1:43:23 how you make money with any of these
1:43:29 things is independent of the
1:43:33 tool because I could easily make a Tik
1:43:36 Tok video that says you can make money
1:43:39 with Photoshop you can make money with
1:43:41 Microsoft
1:43:44 Word
1:43:46 what well you get Microsoft Word and
1:43:49 then you get really good at writing
1:43:51 writing sentences together that convince
1:43:55 people to do stuff and you can be a
1:43:57 copywriter well that's got nothing to do
1:43:59 with Microsoft Word
1:44:02 exactly same [ __ ] with chbt you can make
1:44:05 money with
1:44:07 chbt no you make
1:44:11 money and I'm sorry I get so worked up
1:44:13 about this but it it it just makes me
1:44:16 [ __ ] insane
1:44:19 because like that promise the quick hit
1:44:23 I've been an entrepreneur for 40 [ __ ]
1:44:25 years now here's what I've learned in 40
1:44:27 years like it doesn't
1:44:30 exist the people that are telling you
1:44:33 you can make money with chat PT do you
1:44:35 know how they're making their
1:44:36 money they're selling you guides on how
1:44:39 to make money with chat GPT do you know
1:44:42 what they tell you in those
1:44:43 guides figure out a problem solve the
1:44:47 problem use chat GPT to solve the
1:44:50 problem well I didn't need a [ __ ]
1:44:52 book for
1:44:53 that I didn't need your stupid [ __ ]
1:44:55 course for
1:44:58 that what I needed to do was actually
1:45:01 just solve someone's
1:45:03 problem so if you want to make money
1:45:05 with AI if you want to make money with
1:45:07 chat
1:45:07 GPT there's two independent things that
1:45:11 you have to
1:45:12 do one is you got to have a [ __ ]
1:45:17 idea so like what do you know about well
1:45:21 I know about Hedgehog husbandry perfect
1:45:24 great start there I know about gardening
1:45:27 perfect I know about managing pizza
1:45:30 shops
1:45:32 awesome and I know that all my friends
1:45:34 that run pizza shops the one thing that
1:45:36 they're struggling with right now is
1:45:37 marketing cuz all the people that say
1:45:39 they know marketing are really shitty at
1:45:41 marketing and I'd really like to help
1:45:43 pizza shop people awesome step one
1:45:47 solved figure out where there's an
1:45:49 opportunity to make a difference for
1:45:51 someone that if you could make that
1:45:53 difference they would pay you for
1:45:57 that step
1:45:59 two learn what tools are out there learn
1:46:02 what these tools make
1:46:05 possible the the the amazing thing about
1:46:08 chaty PT is that once you have a problem
1:46:11 to solve I need to come up with
1:46:13 marketing plans for pizza shops chachy
1:46:16 PT can knock that [ __ ] out in 10 [ __ ]
1:46:18 minutes it's remarkable
1:46:22 so chat GPT can help you do it faster
1:46:25 and more
1:46:26 efficient but solving the problem is how
1:46:29 you make the money so start with the
1:46:32 problem start with the thing where you
1:46:34 know you can make a difference in
1:46:36 someone's life step one step two learn
1:46:39 AI tools get AI literate enough to
1:46:41 understand what these tools make
1:46:43 possible well how do I do that where do
1:46:45 I start hang out here for the rest of
1:46:47 the night and I'll tell you what you can
1:46:49 do in chat GPT
1:46:52 I spent 40 minutes saying hello tonight
1:46:55 it's now 9:50 I've spent an hour and 10
1:46:58 minutes talking about the first three
1:47:00 bullet points on my list about chat GPT
1:47:03 3.5 the free version of chat
1:47:09 GPT like you could just play with chat
1:47:11 GPT right now
1:47:14 and for the next two years do nothing
1:47:18 but solve other people's problems using
1:47:20 chatch
1:47:21 3.5 not even going into J chat GPT
1:47:24 4 it's that powerful it's [ __ ]
1:47:30 insane but it's not you can make money
1:47:33 with just because you use it there's
1:47:35 nothing inherently that it does that is
1:47:39 valuable it's only when you apply it to
1:47:42 a problem that is valuable to someone
1:47:45 else who's willing to part with their
1:47:47 money and hand it to
1:47:48 you that's a business
1:47:52 that's on you to figure out and then to
1:47:54 Market that business to that customer
1:47:56 base that's how this [ __ ] works and it
1:47:59 might be a service business it might be
1:48:01 a Consulting business it might be
1:48:02 creating a piece of software it might be
1:48:05 creating just copy for them to run in
1:48:07 their social media like there's there's
1:48:09 an infinite number of businesses you can
1:48:11 start and business models that can all
1:48:13 be supplemented and augmented and
1:48:16 Amplified with chat GPT but you got to
1:48:18 start there not an answer that people
1:48:21 like I don't think but hopefully
1:48:23 appreciate all right I'm going to
1:48:25 quickly show you the rabbit R1 I didn't
1:48:27 realize we were this far in it's going
1:48:28 to take us a while to get through gp4 so
1:48:30 let me show you the rabbit R1 let me go
1:48:31 find
1:48:32 the let me go find the the
1:48:39 action rabbit
1:48:41 R1 oh I know the rabbit can you talk
1:48:44 about that on Tik
1:48:49 tock
1:49:01 keynote come on rabbit where where is
1:49:05 you where's your website
1:49:10 rabbit man they got to get their SEO
1:49:13 game up let me see if I got it
1:49:19 rabbit
1:49:49 oh I hate articles that don't link over
1:49:51 to the
1:49:53 company is this
1:49:56 it
1:49:59 nope oh so [ __ ]
1:50:10 maddening okay here we
1:50:19 go
1:50:21 to present you two things
1:50:24 with Okay so let's see how for music I
1:50:27 use J
1:50:29 robt response response and thinkers
1:50:32 through project rabbit answer okay so
1:50:37 here's what these be careful with the
1:50:38 screen share while searching rabbit yeah
1:50:41 exactly
1:50:42 exactly that's why I put R rabbit R1 in
1:50:45 there wait the rabbit not llama Turtle
1:50:48 Hedgehog or raccoon no rabbit is
1:50:50 absolutely Ely a new uh animal for the
1:50:52 zoo because of this thing The Irregular
1:50:54 Spirit Animals yeah no this is uh this
1:50:58 rabbit gets in because of this one um
1:51:01 okay so what this thing is is it's this
1:51:03 it's this slick little um
1:51:06 device that's got so the the company
1:51:09 that designed this is um is teenage
1:51:13 engineering which they they make like
1:51:15 little synthesizers and [ __ ] like that
1:51:17 so it's a little handheld device it's
1:51:19 got a scen obviously it's got a camera
1:51:22 that rotates on a cylinder so it can do
1:51:24 a selfie or turned away camera um just
1:51:28 by rotating that camera which I think
1:51:30 that's just [ __ ] brilliant that you
1:51:31 don't have multiple cameras it's got a
1:51:33 scroll wheel and it's got a button and
1:51:35 that's it it's got a microphone and a
1:51:37 speaker um the button this is and I so I
1:51:42 ordered one of these like right away
1:51:44 they sold 10,000 of these in 24 hours
1:51:47 and then they sold out their next 10,000
1:51:49 and I think they' sold out their third
1:51:50 10,000 so so this was kind of the hit of
1:51:53 CES um this thing's only 200 bucks and
1:51:57 here's how I Justified it for myself
1:51:59 because I I haven't bought like a a new
1:52:02 sexy Tech Gadget in for [ __ ] ever I
1:52:04 think the last one I bought was Google
1:52:06 Glass which was an absolute waste of
1:52:07 money for 200 bucks this one at a
1:52:11 minimum is just like a cool thing to
1:52:13 stick on my shelf right because it's
1:52:15 just cool industrial design but okay so
1:52:18 that's what it is so let's let's watch
1:52:19 how he according to berand Russell
1:52:22 reality is a topic that and talk like a
1:52:25 walkie-talkie it answers my questions
1:52:27 within 500 milliseconds with a push to
1:52:30 talk button you don't need to say
1:52:32 anything to wake it up just press and
1:52:35 hold the button and talk like a
1:52:38 walkie-talkie so I want like that's
1:52:41 that's a huge deal to be so you don't
1:52:43 have to go hey name of your robot to
1:52:46 wake it up right you just push the
1:52:47 button like and simple interface things
1:52:50 like that are are huge the why did chat
1:52:53 GPT get to 100 million users in 6 weeks
1:52:56 it had a simple interface so this to me
1:53:00 when I saw this demo it felt like this
1:53:03 is a piece of Hardware Apple should have
1:53:06 introduced and didn't and he he actually
1:53:09 starts out this keynote [ __ ] all
1:53:11 over the current iPhone and and all of
1:53:15 the phone manufacturers having these
1:53:17 massive app environments that you have
1:53:20 to go to all these different apps to do
1:53:22 the [ __ ] you want to do um so one of the
1:53:25 things they built underneath this is
1:53:27 what they're calling the large action
1:53:29 model so there's a large language model
1:53:30 which is the thing you talk to the large
1:53:33 action model is they can train this
1:53:35 thing to use other pieces of software by
1:53:39 looking at the interface and how you
1:53:41 interact with the interface he demos at
1:53:43 the end of this um you can train it to
1:53:47 understand how you make an image in
1:53:50 Discord a mid Journey image in Discord
1:53:53 you can train it to do that and once
1:53:56 you've trained it you can just push your
1:53:57 button and go make me a mid-journey
1:53:59 image of blah blah blah give it a prompt
1:54:01 and then let it go and it will go to
1:54:03 Discord on your behalf and prompt it how
1:54:06 to make a mid Journey image it's it's
1:54:08 [ __ ] nuts um down the rabbit hole
1:54:11 gimmick with Kyle yeah exactly um okay
1:54:15 let's give you a shot what's the nature
1:54:17 of the
1:54:18 reality the nature of reality is a topic
1:54:20 that has captivated the minds of
1:54:22 philosophers and thinkers throughout
1:54:25 history according to Bertrand Russell
1:54:27 reality is composed of hard so cute
1:54:29 little bouncing rabbit you know that's
1:54:31 cute derived from sensory exper latest
1:54:35 movies in
1:54:37 Christopher I want to emphasize y portal
1:54:40 I have complete control analyses for R1
1:54:44 it's just like iCloud let's say I want
1:54:46 to listen to some
1:54:49 music so what he's demoing here is and I
1:54:52 this is the last thing I'm going to show
1:54:53 you you should go watch the keynote so
1:54:55 just just put in rabbit R1 keynote you
1:54:57 can go watch this movie um he's a little
1:55:00 hard to understand but but it's it's
1:55:01 pretty slick what he's demoing here is
1:55:04 you can give the rabbit R1 access to
1:55:08 services that it understands how to do
1:55:11 things with so it understands what
1:55:12 Spotify is so Rob okay rabbit. te thank
1:55:16 you that that the URL is rabbit. TE to
1:55:19 get to the uh to the site
1:55:23 um so once you give it access to your
1:55:26 software
1:55:27 applications it can take actions so so
1:55:30 his whole sales pitch here is rather
1:55:32 than you having to go to your phone find
1:55:35 the Spotify app you know search for the
1:55:38 artist you want to play and and you know
1:55:40 click click the button to play that
1:55:42 artist you can just go Um play me Billy
1:55:46 eish on Spotify right so that's what
1:55:48 he's demoing here
1:55:52 with a rabbit hole web portal I have
1:55:55 complete control over which Services I
1:55:57 want to activate on my R1 and which
1:56:00 provider I prefer for music I use
1:56:03 Spotify so let's go ahead and connect my
1:56:05 Spotify to the rabbit
1:56:10 hole now that this is Spotify login page
1:56:13 before I start the loging process I want
1:56:16 to emphasize our commitment to privacy
1:56:18 the question how many language does
1:56:20 rabbit support so it's it's built on top
1:56:23 of I assume they're they're calling GPT
1:56:25 4 or at least 3.5 here um whatever
1:56:30 languages a large language model can
1:56:32 support which is essentially all of them
1:56:35 and what he talks about is they've got
1:56:36 this thing so the rabbit right now can
1:56:39 do
1:56:41 um bilingual like like it can basically
1:56:44 translate for both parties so if you're
1:56:46 talking to someone who speaks Mandarin
1:56:49 when they speak Mandarin it'll translate
1:56:52 it to English and when you speak English
1:56:54 it'll translate your English to Mandarin
1:56:57 so they've got it set up to already do
1:56:59 that right now it's [ __ ] nuts the way
1:57:01 and wait sorry what he said in the
1:57:04 keynote is it can do all that whether or
1:57:08 not I can do that is a completely
1:57:09 different story my
1:57:12 assumption I ordered this with with a
1:57:15 clear mind I was not drinking Blandon
1:57:17 when I ordered this my assumption is
1:57:20 that this will work if I'm lucky at a C+
1:57:24 level of my expectations I'm expecting
1:57:26 it to work I I am expecting to be
1:57:29 Gravely disappointed that it's going to
1:57:31 be jankier than [ __ ] but if this thing
1:57:34 works even at a C+ level it's going to
1:57:36 be cool and if it works at a b or higher
1:57:39 level it's going to be absolutely
1:57:41 amazing so way we design the
1:57:43 infrastructure for apps that we interact
1:57:45 with we don't create fake users or spam
1:57:48 users we don't hack their
1:57:51 infrastructures and of course we do not
1:57:54 store any of your third party
1:57:56 credentials yeah yeah yeah instead we
1:57:58 redirect you to their login page we
1:58:01 never save your username and password
1:58:04 and we don't track your data yeah the
1:58:07 authentication happens on the respective
1:58:11 apps now I'm logging with
1:58:18 Spotify rabit OS recognize that Spotify
1:58:22 so wait see that
1:58:25 interface so Spotify Apple music Uber
1:58:30 door Dash so connect services so so you
1:58:34 can tell this is a version 1.0 of this
1:58:37 of this page but you're going to have a
1:58:39 configure page where you can hook up all
1:58:40 of your different Services right and one
1:58:43 of the things you'll be able to do is
1:58:45 train your own Services right do do your
1:58:48 own training so um so really interesting
1:58:50 from that perspective recognize that
1:58:52 Spotify has been authenticated on my
1:58:55 device before this presentation I
1:58:59 pre-authenticated some other apps like
1:59:01 uber and door Dash I got to comment on
1:59:04 something
1:59:07 here just choose your preferred services
1:59:09 and loging through them to get
1:59:11 authenticated on your device it's that
1:59:13 easy now let's play some music play
1:59:17 craftworks pocket calculator
1:59:21 playing pocket calculator by craftwork
1:59:23 for
1:59:26 you okay so it plays Spotify now I just
1:59:29 saw a comment this doesn't do anything
1:59:31 new I 1,000% disagree with
1:59:36 that from a technical perspective it
1:59:39 doesn't do anything new but from an
1:59:42 interface perspective it does a lot new
1:59:45 it does a lot new the fact that you have
1:59:48 a handheld device that you push a button
1:59:50 and what that button push means is I'm
1:59:53 going to talk to my llm and then the
1:59:56 fact that you can connect that llm to
1:59:58 actually perform actions it does do
2:00:01 something new now the Geeks in here the
2:00:03 people making those comments why they're
2:00:05 making those comments is they know how
2:00:07 to program all of these services to be
2:00:10 able to do this anyway right well you
2:00:12 could do that on your iPhone you could
2:00:14 remap this button to take this and take
2:00:16 this action and you could have that and
2:00:18 you could have that call this app and
2:00:21 yeah you could if you were a [ __ ]
2:00:24 geek most human beings aren't that so
2:00:28 for me what's new about this is entirely
2:00:31 the
2:00:32 interface which the logic of this
2:00:35 doesn't do anything new you could also
2:00:37 apply that to chat
2:00:39 GPT you could do chat GPT like
2:00:43 interactions with the with the open AI
2:00:46 playground in
2:00:48 2020 nobody gave a [ __ ] about it it
2:00:52 wasn't until they put a simple interface
2:00:55 in front of it that it got to 100
2:00:57 million
2:00:59 users now is this device going to you
2:01:02 know be a massive hit and get to 100
2:01:04 million users I don't [ __ ] know
2:01:06 probably
2:01:08 not it's similar to zappier yeah I I
2:01:11 mean in some sense in some sense yeah in
2:01:16 some sense what this tool now what what
2:01:18 these guys did which I think is really
2:01:20 smart is they created this thing called
2:01:22 the L the large action model so instead
2:01:25 of having
2:01:27 to program something like zapier to take
2:01:30 those actions you just have it watch you
2:01:34 interact with software what the hell am
2:01:36 I looking at you're looking at the
2:01:37 rabbit R1 Pate um it's a it's a CES
2:01:41 device that I was just talking about how
2:01:43 I just uh ordered it and
2:01:47 um that I'm assuming it's going to be a
2:01:50 piece of [ __ ] but I like the design
2:01:51 basically um but but basically what
2:01:54 they're doing is they're they're they're
2:01:56 making the ability to make zappier like
2:01:59 actions consumer friendly and like once
2:02:02 they're configured and once they have a
2:02:04 library of these actions it's just going
2:02:06 to know how to do lots and lots of [ __ ]
2:02:09 and it'll be really easy to configure
2:02:11 and basically once you log into your
2:02:13 systems and give it permission with
2:02:16 um you can interact with Siri like G PT
2:02:20 I listen I get that you can I get that
2:02:23 you can configure lots of things in your
2:02:26 current Hardware Arsenal to do what this
2:02:29 does but this is a particularly
2:02:31 well-designed I think unique design
2:02:36 um
2:02:37 device that Apple should have come out
2:02:41 with like someone should have come out
2:02:44 with this other than this Punky upstart
2:02:47 but I love that this Punky upstart came
2:02:49 up with this right because they
2:02:51 completely rethought the whole idea of
2:02:54 needing an app
2:02:55 infrastructure what they basically did
2:02:58 was they put in the palm of your hand a
2:03:01 device that does all of the navigating
2:03:05 of the hundreds of apps on your iPhone
2:03:07 for you and that's where all this AI
2:03:10 stuff's going so for me this just
2:03:11 represents this is the first of a series
2:03:14 of devices that are going to simplify
2:03:17 our entirely over complicated world
2:03:20 so that's why I think it's
2:03:22 significant all right enough about the
2:03:25 rabbit now let's go look at the other
2:03:30 rabbit the rabbit uses the L model yet
2:03:33 another oh spirit animal the lamb yeah
2:03:36 the rabbit and the lamb that's actually
2:03:37 a really good point it's really good
2:03:40 okay uh Rudy what's happening thank you
2:03:42 for the rose sir appreciate that all
2:03:44 right let's get back to chat gbt all
2:03:46 right I'm going to have to go fast cuz
2:03:48 it's getting late so all right let's
2:03:51 see when you move up to chat
2:03:54 gp4 um the basic writing of chat GPT
2:03:58 gets way better so chat GPT 3.5 is
2:04:04 um is it doesn't matter but well p p
2:04:09 will think it matters chat GPT 3.5 has
2:04:12 175 billion parameters but what is
2:04:15 what's a parameter Kyle what does 175
2:04:18 billion parameters mean I don't know
2:04:19 Pate knows if you want to if you want to
2:04:21 know join the AI Salon go to pate's
2:04:24 mechanics Guild and he can describe to
2:04:26 you what all this parameter [ __ ] means
2:04:28 but it basically means more parameters
2:04:30 is better so chat GPT 3.5 has 175
2:04:35 billion parameters chat GPT 4 has 1.8
2:04:40 trillion broken into 10 different 175
2:04:44 billion parameter
2:04:46 models it uses what's called a mixture
2:04:49 of experts model chat gp4 does where the
2:04:52 where rather than just having one big
2:04:54 massive model they have 10 specialized
2:04:56 models that all get coordinated when you
2:04:59 write your prompt it says so so maybe
2:05:02 they've got one of these models is
2:05:04 really good at coding and another one's
2:05:05 really good at analyzing documents and
2:05:08 another one's really good at I whatever
2:05:10 else they've got I don't I don't know
2:05:12 what those things are but basically when
2:05:14 you write your prompt rather than just
2:05:16 sending that prompt into a single model
2:05:18 it sort of figures out which of those
2:05:20 expert models is are best for your
2:05:23 prompt and send it sends it out to all
2:05:25 of them and then sort of brings the
2:05:26 answers back together so it's just a
2:05:28 more
2:05:30 sophisticated way of doing what it does
2:05:32 so an interesting exercise is
2:05:37 to write a prompt or or have a have a
2:05:40 whole conversation within chat GPT 3.5
2:05:44 and then flip over to chat gp4 and do
2:05:46 the exact same thing and just notice the
2:05:48 difference es what I notice about them
2:05:52 is that they both answer the questions
2:05:56 fine but chat gp4s are more nuanced
2:05:59 right and it's just it's just better
2:06:01 writing and it's that's a funny thing
2:06:03 about the English language you can write
2:06:05 something that answers a question or you
2:06:07 can write something that answers the
2:06:08 question well that's more chat gp4 where
2:06:12 chat GPT 3.5 just answers the question
2:06:15 the other thing about uh chat gp4 is
2:06:18 it's connected to the Internet so uh so
2:06:21 now we'll flip over to chat
2:06:24 gp4 and we'll say
2:06:28 um are there any
2:06:33 upcoming acting
2:06:35 classes in New York City that teach the
2:06:42 Meisner
2:06:47 technique and it should
2:06:51 doing research with Bing
2:06:54 so one of the one of the criticisms of
2:06:57 chat GPT when it first came out is it
2:07:00 said well I only know things up to
2:07:02 September 2021 that's that that Bas
2:07:05 large language model was only trained up
2:07:08 to September
2:07:10 2021 um that's now been updated to April
2:07:14 2023 but if you ask it anything
2:07:17 current uh if it's if it's more current
2:07:20 than September 2023 it can't answer it
2:07:23 but now it can because it goes out to
2:07:24 the Internet so
2:07:27 um teaching Meisner the esper Studio
2:07:30 that's where I studied I studied at the
2:07:32 the bill esper studio and Terry nicker
2:07:35 Boer is a friend of mine he worked for
2:07:37 Bill esper for years and he's now got
2:07:39 his own studio um he's he's a a good guy
2:07:42 and has a good studio uh Maggie Flanigan
2:07:46 she also worked for Bill esper that's
2:07:48 funny and then the acting studio I don't
2:07:50 know who did that but
2:07:52 anyway yeah so um so it can it can see
2:07:55 the internet all right let's see what
2:07:57 else what else we going to do code
2:07:59 interpreter holy [ __ ] I'm going to come
2:08:01 back to code interpreter because code
2:08:03 interpreter is
2:08:04 Big I'm coming back to that okay so GPT
2:08:09 Vision chat GPT 4 can see huh it's got
2:08:14 eyes it does have eyes can I make money
2:08:17 with it you can make with chat
2:08:20 GP oh what are your thoughts on the chat
2:08:22 GPT teams upgrade
2:08:26 okay I'm pissed off about
2:08:29 it because it's not really an upgrade
2:08:32 right now
2:08:34 um in theory the GPT team upgrade is
2:08:38 awesome in reality it's a separate
2:08:42 subscription here's what I mean you
2:08:45 currently have your chat GPT Plus
2:08:47 subscription that you're making all
2:08:49 these gpts with and you've got your
2:08:50 prompt history for the past year if you
2:08:53 upgrade to teams it does not migrate any
2:08:57 of the existing [ __ ] that you built
2:08:59 over and you're like but Kyle that seems
2:09:02 [ __ ] stupid I agree it's [ __ ]
2:09:07 stupid apparently this is what I've
2:09:10 heard I I haven't seen this but
2:09:12 apparently in the coming weeks they will
2:09:15 figure out how to migrate that [ __ ] over
2:09:17 there or like they know this is an issue
2:09:20 um because of that I'm
2:09:23 waiting uh so the the the reason to go
2:09:27 to something like teams is they're not
2:09:28 going to train on your data so you get
2:09:30 like a little mini version of Enterprise
2:09:34 privacy you know uh features but the
2:09:38 fact that I don't get my prompt history
2:09:39 and all my gpts don't get migrated over
2:09:42 I'm like are you [ __ ] kidding me so
2:09:45 anyway so I am not happy with teams
2:09:47 right now uh so I would I I wouldn't do
2:09:49 it unless you absolutely need it until
2:09:51 they improve it um okay GPT Vision chat
2:09:55 GPT can see now this if you haven't seen
2:09:58 it do this this is [ __ ] mindblowing
2:10:00 so I'm going to click on on uh oh let me
2:10:05 show you something
2:10:07 here this little paperclip is is a magic
2:10:10 paperclip if I go to chat GPT 3.5 it
2:10:14 disappears right so it's gone now if I
2:10:16 go to chat
2:10:17 gp4 there it is you can upload a PDF to
2:10:22 this image
2:10:24 files you can upload movie files which
2:10:27 you can analyze but it can't see movies
2:10:29 yet but it can see images so if I click
2:10:32 on this I'm going to go search for
2:10:37 whiteboard and I didn't say whiteboard
2:10:41 did I oh wait is
2:10:44 thisuh yeah that's the one okay so this
2:10:47 is a this is a picture of a white board
2:10:49 I just randomly pulled off the internet
2:10:51 so the the the Irregulars here have seen
2:10:54 this demo 43 times I this still blows my
2:10:57 mind every time I do it okay so I
2:11:00 uploaded a picture of a whiteboard Kyle
2:11:03 I requested on my $20 account that they
2:11:05 not train on my data I had to fill out a
2:11:07 form yeah you can do that you can do it
2:11:10 does it work yes they they claim that it
2:11:12 works the the the problem Kev Kev talk
2:11:15 is
2:11:17 um they
2:11:19 default chat GPT Plus for them to allow
2:11:23 allow to use your data like you have to
2:11:25 dig and do what you did to do it um
2:11:29 Becky R irregular here checking in late
2:11:32 um it's been it's been kind of a deep
2:11:34 Rabbit Hole tonight Becky I don't think
2:11:35 you've you've missed a ton that you
2:11:37 haven't seen before
2:11:39 um the the GPT teams Thing by default
2:11:43 does not train on your data so it's it's
2:11:46 really just a default thing I I don't
2:11:48 know if the terms of service are
2:11:50 actually different I haven't looked that
2:11:51 deep into it because I'm I'm so
2:11:53 disappointed with how teams was
2:11:54 implemented in in this roll out okay
2:11:58 anyway so I uploaded a picture of a
2:12:01 whiteboard and now I'm just going to say
2:12:02 what is this what is
2:12:06 this chat GPT is now going to look at it
2:12:12 and analyze it and do OCR on it and
2:12:15 everything so it appears to be a
2:12:16 whiteboard of handwritten notes and
2:12:18 drawings the content suggests that it's
2:12:20 used for brainstorming or planning the
2:12:22 user interface for an ordering system
2:12:25 there are various sections of the
2:12:27 outline user tapes user types use cases
2:12:30 key features it seems to be focused on
2:12:32 an ordering system for an airport
2:12:34 environment potentially for Quick
2:12:37 Service in a business or family
2:12:40 setting um there are sketches of what
2:12:43 could be the UI screens on devices I'm
2:12:45 going to say okay great
2:12:47 um
2:12:49 okay
2:12:50 great I'd like
2:12:57 to come up with a marketing
2:13:01 plan for how we might might Market such
2:13:05 an
2:13:15 app and so so it's a says creating a
2:13:19 marketing plan for a UI system in an
2:13:21 airport involves several components
2:13:23 below as a high Lev plan market research
2:13:26 identify your target audience it pulled
2:13:28 business Travelers family and techsavvy
2:13:30 consumers and frequent flyers from the
2:13:33 Whiteboard competitor analysis unique
2:13:35 selling proposition branding develop a
2:13:38 catchy name brand
2:13:40 message um craft a message that
2:13:43 resonates with the needs and wants of a
2:13:44 target audience dine on time with brand
2:13:47 name your TR travel
2:13:49 buddy so you know Partnerships
2:13:52 promotional offers public
2:13:55 relations so from a whiteboard picture
2:13:59 we now have a marketing plan for
2:14:02 whatever the [ __ ] app that was
2:14:04 right so so let's let it finish up here
2:14:07 it's a little slow tonight but
2:14:11 whatever now now here's your Implement
2:14:14 implementation timeline monitoring an
2:14:17 evaluation your I how we going to know
2:14:19 it's going to work
2:14:21 budgeting ROI analysis legal and ethical
2:14:27 considerations have you ever written a
2:14:29 marketing
2:14:30 plan you know how long it takes to just
2:14:33 do something this high
2:14:36 level bang done okay now
2:14:41 um great let's move into branding I
2:14:44 really like your tag name in The
2:14:46 Branding
2:14:47 section and and want you to come up with
2:14:50 20 possible names for this app that
2:14:54 incorporate all of your thinking in the
2:14:56 marketing
2:15:06 plan all right certainly here are 20
2:15:08 potential
2:15:09 names jet snack Arrow eats gate plate
2:15:13 fly bite board bites takeoff Taste
2:15:15 Lounge
2:15:16 Feast uh let's let's go with gate
2:15:21 plate great I like gate
2:15:25 plate period can you now show me a
2:15:29 screenshot of what the interface might
2:15:31 look like for the ordering
2:15:33 screen that has the gate plate logo at
2:15:36 the top and a simple visually compelling
2:15:43 interface on the app screen inspired by
2:15:47 the design on the White
2:15:55 board so now we're flipping into chat
2:15:58 jp4 can make images using Dolly
2:16:01 3 um hang
2:16:04 on why this is a big deal is that all of
2:16:09 the context of this
2:16:12 conversation is now input for the prompt
2:16:16 for the image that we just asked for
2:16:18 including what was on the
2:16:24 Whiteboard there's
2:16:27 a
2:16:29 okay there's an app design for gate
2:16:33 plate an app that didn't exist until
2:16:36 just this moment because we grabbed a
2:16:39 picture of a whiteboard and threw it in
2:16:41 there and said what is
2:16:44 this so
2:16:46 what what what I'm demonstrating here
2:16:50 what we're experiencing here is kind of
2:16:53 a very thin layer of what they're
2:16:55 calling multimodality right it could see
2:16:57 an image it could see the
2:16:59 Whiteboard it's got this large language
2:17:01 model where we could say let's do the
2:17:03 marketing [ __ ] let's come up with a name
2:17:05 for this
2:17:07 thing now I can do interface
2:17:11 design so I can
2:17:14 say that's great could you visualize for
2:17:17 me what a poster might look like in an
2:17:20 airport that would advertise this app
2:17:23 that includes a screenshot of the
2:17:26 app as well as the tagline and name of
2:17:30 the app on the
2:17:47 poster
2:17:48 [Music]
2:17:51 why do we need
2:17:53 that Kyle I
2:18:00 requested boom gate plate spelled wrong
2:18:03 but
2:18:05 cool there's your airport there's your
2:18:07 poster oh but wait there's
2:18:11 more awesome can you
2:18:14 please recommend front end back end and
2:18:18 database Technologies and then write
2:18:22 sample code for all three for the
2:18:25 ordering screen that you visualized
2:18:35 earlier front-end Technologies react
2:18:39 react
2:18:41 native Tailwind CSS for styling State
2:18:44 Management Redux or context API for
2:18:46 State Management backend Technologies
2:18:49 runtime environment nodejs framework
2:18:52 expressjs authentication Json web tokens
2:18:55 databased Technologies mongodb non-sql
2:18:58 or postgress SQL for
2:19:00 SQL om odm mongus for mongod DB or sqil
2:19:05 eyes for postgress here's front-end
2:19:08 react
2:19:10 code here's backend node.js with express
2:19:16 code
2:19:24 now this is me [ __ ] around in about
2:19:27 10
2:19:28 minutes with just uploading a
2:19:33 whiteboard asking it to come up with a
2:19:35 marketing strategy asking it to do a
2:19:38 little bit of wire framing which it did
2:19:42 do a little bit of marketing
2:19:44 visualization and do a bit little bit of
2:19:46 code planning
2:19:48 now for me I look at this silly
2:19:53 10-minute exercise and I think imagine
2:19:57 if you spent a whole weekend doing this
2:20:01 with some idea that you have or say even
2:20:04 more radically you take your your week
2:20:07 off that you have for
2:20:09 vacation and you say Hey what if I
2:20:12 started that company I've always wanted
2:20:14 to
2:20:15 start the minute you start
2:20:19 understanding that chat GPT isn't just a
2:20:23 single tool but it's multiple tools that
2:20:25 work together and they work together
2:20:28 seamlessly and dynamically and
2:20:30 everything is in context of the larger
2:20:34 conversation interacting with chat Che
2:20:37 so so here's let me let me flip over
2:20:43 here interacting with chat GPT 3.5 for
2:20:47 me
2:20:50 was kind of like interacting with an
2:20:53 idiot savant intern
2:20:57 meaning they were like the best educated
2:21:00 intern you've ever seen they knew
2:21:01 everything but they needed lots of
2:21:04 handholding the idiot part they didn't
2:21:06 know anything about the world they don't
2:21:08 know anything about business so you ask
2:21:11 him to do stuff and they're like sure
2:21:12 here you go and you're like oh that's
2:21:14 the right data but no that's not how we
2:21:17 work here that's not how we talk here
2:21:19 okay tell me how to talk here here's how
2:21:21 you
2:21:22 talk working with chat gp4 is is almost
2:21:25 like working with a whole bunch of
2:21:28 different people it's like
2:21:32 like that interaction that we just did
2:21:35 where we went from whiteboard to
2:21:36 marketing plan to interface sketches to
2:21:40 marketing concepts to code
2:21:44 recommendations that would be a series
2:21:46 of meetings over days if not weeks with
2:21:49 multiple players on a
2:21:52 team now is cat G pt4 going to be as
2:21:55 good as having those actual team members
2:21:57 in the room
2:21:59 no but can you get 80% of the way there
2:22:04 on all of the different disciplines for
2:22:06 starting a whole business
2:22:08 yes and if you're good at thinking
2:22:11 horizontally if you're good at sort of
2:22:13 orchestrating lots of different people
2:22:15 to do their pieces well and you can
2:22:17 bring that [ __ ] together you can now do
2:22:20 in a
2:22:22 weekend with yourself or maybe with
2:22:24 yourself and one other person that's got
2:22:26 complimentary
2:22:28 skills what would have taken a team of
2:22:30 four or five or 10 people at least a
2:22:34 month
2:22:37 historically you can now casually spin
2:22:41 up a business without even thinking
2:22:43 about
2:22:45 it so why not
2:22:48 casually spin up a business make the
2:22:50 marketing plan make the website launch
2:22:53 it do the social media plan make the
2:22:55 images for the [ __ ] Instagram
2:22:58 account put it out in the
2:23:02 world you know those made for TV movie
2:23:05 or made for TV products like buy the
2:23:09 gardening buddy you know it's a hose
2:23:12 fitting and a toothbrush those
2:23:15 things a good friend of mine from when I
2:23:18 lived back back
2:23:21 East he's like the lead marketing guy
2:23:23 for one of those companies you know how
2:23:25 they sell that
2:23:28 [ __ ] they invent
2:23:31 products they make websites for them
2:23:35 they they make 3D visualizations of them
2:23:37 they make marketing plans for them they
2:23:39 come up with pricing for them and they
2:23:41 launch them when they don't
2:23:45 exist and but they launch them
2:23:49 positioning them that they do
2:23:51 exist and then what happens is they've
2:23:54 got this mathematical formula and if it
2:23:58 scores like an 8.8 or above they know
2:24:01 it's going to be a hit so they launched
2:24:02 like 100 products and like three of them
2:24:06 score 8.8 or higher and all of the rest
2:24:09 they just like oh say sorry uh there
2:24:12 were there were manufacturing issues
2:24:13 here we'll give you your money back for
2:24:15 the 97 products that didn't work for the
2:24:17 three that score 8.8 or higher they take
2:24:20 the orders they they they rush them into
2:24:24 manufacturing and then they sell those
2:24:26 initial products and then they run tv
2:24:30 ads that's their business model now how
2:24:32 do they do that [ __ ] Brute Force 3D
2:24:36 modelers marketing teams copywriters
2:24:40 website Builders like they've got it all
2:24:42 [ __ ]
2:24:43 automated but you can now do that with
2:24:45 your business with a service business
2:24:47 business with with a product business
2:24:50 with a software
2:24:52 business you and little old
2:24:56 chbt could knock that [ __ ] out in a
2:25:02 weekend if you put your mind to it
2:25:05 here's another thing I think about a
2:25:08 lot Malcolm Gladwell in his book
2:25:10 outliers made popular this concept of
2:25:13 the 10,000 hour rule and the 10,000 hour
2:25:16 rule basically said if you want to get
2:25:19 good at something you have to put in
2:25:20 like 10,000
2:25:22 hours which if you do the
2:25:24 math depending on if you want to do it
2:25:27 as your full-time job or not is
2:25:29 somewhere between like one and a half
2:25:30 and four years of your life 10,000 hours
2:25:34 is like go back to college and get a
2:25:36 degree for
2:25:37 whatever
2:25:39 botney you know um sustainable gardening
2:25:43 what
2:25:45 whatever imagine
2:25:48 rather than having to put in 10,000
2:25:50 hours you
2:25:52 spent 3
2:25:54 weeks on a single subject with chat
2:26:00 GPT you'll be pretty [ __ ] masterful
2:26:02 at the end of 3 weeks at like on every
2:26:06 level and you might not have the
2:26:09 Mastery but all of that knowledge all of
2:26:12 that expertise is at your fingertips
2:26:14 right now so imagine spending 3 weeks 8
2:26:17 hours a day on a single subject how far
2:26:20 you could get with chat GPT in three
2:26:21 weeks it would it it will be
2:26:27 astounding and so what are the
2:26:29 implications of that
2:26:32 well you might say I've always wanted to
2:26:34 get into molecular botany I don't know
2:26:38 what that is but [ __ ] it I I want to be
2:26:40 into molecular botony you go to chat GPT
2:26:42 and you're like I know nothing about
2:26:44 molecular botany but I'm hoping you can
2:26:46 tell me sure here's Mon and then you
2:26:49 just start running down rabbit holes and
2:26:51 you have it connected to Bing and it's
2:26:52 sending you off to read papers and it's
2:26:54 reading the papers for you and it's
2:26:56 summarizing them and it's coming up with
2:26:59 business ideas for you and in the first
2:27:01 week you come up with three different
2:27:04 business ideas and you have enough
2:27:05 education about Mo molecular botany to
2:27:08 to try it and then in the second week
2:27:11 you start putting together plans and you
2:27:13 start playing with things and you start
2:27:14 doing things and in the middle of the
2:27:16 Third week you realize H I hate
2:27:20 molecular
2:27:24 botany you wouldn't have discovered that
2:27:27 until a year and a half or two years
2:27:30 in Pre chat GPT you'd have had to go
2:27:33 back to school you might have had to
2:27:35 take out a loan you might have had to
2:27:37 quit your job to go to school to learn
2:27:40 enough enough about molecular botany to
2:27:43 realize oh I [ __ ] hate molecular
2:27:45 botany and so what happens is people do
2:27:49 that calculation in their mind they're
2:27:51 thinking like you know I think I really
2:27:54 want to be a molecular botanist but they
2:27:56 in their mind they're like okay well how
2:27:58 would I have to do that well I'd have to
2:27:59 go back to school oh that's a little
2:28:01 stressful I don't really have the money
2:28:03 for it but if I quit my job and I
2:28:05 liquidated my 401k and and I went back
2:28:08 to school and then you go H that's too
2:28:11 much [ __ ] it and you don't pursue that
2:28:14 dream what if you could pursue three
2:28:18 dreams a
2:28:21 year fairly
2:28:23 deeply rather than one dream every 3
2:28:28 years that changes the [ __ ] world
2:28:32 like the world
2:28:36 changes your world will change if you
2:28:39 look at these tools not as you can make
2:28:43 money but as
2:28:46 access
2:28:48 to amplifying the things you've always
2:28:50 wanted to do and elevating them to a
2:28:52 level that makes it a reality in a very
2:28:55 short amount of
2:28:57 time and then you can
2:29:00 choose do I love molecular biology or
2:29:03 not nah it wasn't so
2:29:06 great but you know what would be cool
2:29:09 billiard ball design
2:29:12 okay take the next month and go deep
2:29:15 into billiard ball design
2:29:18 or boat
2:29:22 design or
2:29:25 whatever and then in running down those
2:29:28 rabbit holes what you'll realize is oh
2:29:30 you know what I really always wanted to
2:29:32 do make children's
2:29:34 books all right you can do that too now
2:29:38 I'm not good at illustrating and I'm
2:29:40 actually really shitty at words well
2:29:41 here's a tool that's great at words and
2:29:43 it can illustrate for you all you have
2:29:45 to do is have a [ __ ] idea
2:29:47 okay and a little bit of
2:29:50 taste and some Moxy to like figure out
2:29:53 which of these tools I can work with
2:29:55 together and then I can put that on
2:29:56 Kindle publisher and holy [ __ ] over the
2:29:59 weekend I publish three children's books
2:30:03 holy
2:30:04 [ __ ] well will you make any money at
2:30:07 that I don't know probably not
2:30:10 everyone's making children's books right
2:30:11 now but you know what if you put
2:30:14 something out there and your heart's
2:30:16 behind it and you share share with the
2:30:17 right people and it hits a cord with the
2:30:19 public it just might [ __ ]
2:30:26 sell but now you're a published
2:30:29 author where for 40 years you thought
2:30:33 well I've always wanted to write
2:30:35 children's books but I don't have the
2:30:37 talent guess what buckle up butter cup
2:30:41 now you
2:30:44 do what's that mean I don't know be
2:30:48 careful what you ask
2:30:52 for because in the olden timey days you
2:30:56 know from
2:30:57 like late 2022 and
2:31:03 earlier we could all use the fact that
2:31:06 we didn't have talent or skills or
2:31:09 knowledge or a degree as an excuse not
2:31:12 to do the things we say we always wanted
2:31:14 to
2:31:15 do
2:31:17 we're entering a world right now where
2:31:20 you can do any [ __ ] thing you put
2:31:23 your mind to with some you know notable
2:31:25 exceptions I think things that require
2:31:27 physical dexterity like becoming a
2:31:30 virtuoso violinist that's probably still
2:31:33 10,000 hours or more plus
2:31:36 talent but any knowledge work any
2:31:39 creative
2:31:40 work
2:31:42 programming
2:31:44 writing publishing any of of IT
2:31:52 consulting nonprofit work holy [ __ ] the
2:31:56 fact that nonprofits are resisting AI
2:31:59 right now is batshit crazy to
2:32:01 me you you have a okay what are
2:32:04 nonprofits they are they are chronically
2:32:07 underst staffed and underresourced
2:32:10 what are these AI tools they are
2:32:12 Humanity
2:32:13 amplifiers people can get [ __ ] done 10
2:32:16 times faster 10 times
2:32:19 better overnight and they're resisting
2:32:23 that Champy's got to go
2:32:26 pee
2:32:28 champy what's up buddy you got to go pee
2:32:30 PE you got to go pee
2:32:45 PE
2:33:15 Jesus
2:33:23 all
2:33:25 right Champ's been taken care of for
2:33:28 more on this the salon. a okay so I know
2:33:31 that was a bit of a rant in a soap
2:33:35 box
2:33:40 but that Soap Box is why I do this seven
2:33:43 nights a
2:33:44 week because
2:33:51 the world has changed and nobody knows
2:33:53 it yet or not a lot of people know it
2:33:56 yet I mean there's 180 million people
2:33:59 who are users of chat
2:34:01 GPT there are only
2:34:05 about somewhere between the numbers that
2:34:08 I've seen somewhere between 300,000 and
2:34:10 500,000 people are using GPT 4 out of
2:34:15 100 Mill million people who've tried
2:34:17 chat
2:34:19 GPT a like less than a million by a
2:34:26 lot are using
2:34:29 gp4 and gp4 is like gp4 is like one of
2:34:33 those cheat codes for your video games
2:34:36 you got to learn the up up down down
2:34:38 right right x button
2:34:42 code but it it it it's your cheat code
2:34:45 like I feel like chat gp4 is like a
2:34:48 cheat code for
2:34:50 life and no one knows about
2:34:52 it is it fair to say that AI allows
2:34:55 people to start with lower startup costs
2:34:57 not having to hire out yeah Erica
2:34:59 [ __ ] yes yes yes yes yes this
2:35:02 is what one of the things that the the
2:35:05 the um the VCS in Silicon Valley are
2:35:09 kind of [ __ ] up with right now is what
2:35:13 do they do with their
2:35:15 money because historically what you
2:35:18 would do is you'd put in 20 or $50
2:35:20 million for for a team to hire 20 or 30
2:35:23 or 50
2:35:25 people you know to burn through that
2:35:27 money in 18
2:35:30 months to spin up a software company
2:35:33 that does X Y or Z and it you needed
2:35:35 coders and engineers and interface
2:35:37 people and
2:35:40 marketers and the whole model is based
2:35:43 on it's going to take you 18 months to
2:35:45 get to an MVP
2:35:47 and then you're going to see these
2:35:48 metrics and then you're going to do this
2:35:49 and then you do your series a and then
2:35:50 you do your series B and then you do
2:35:52 your series C and everybody makes their
2:35:55 Millions well you could now start up a
2:35:59 similar kind of company
2:36:01 with half a million dollars in three
2:36:06 people Silicon Valley doesn't know how
2:36:08 to deal with that and those three people
2:36:11 in that half a million dollars and and
2:36:12 maybe even three people with $50,000
2:36:15 right those people just need eat ramen
2:36:17 for a year and a half and it's probably
2:36:18 not a year and a half it's probably more
2:36:19 like 4 to 6
2:36:21 months and they can develop something as
2:36:26 compelling as a
2:36:29 standard startup thing what is what do
2:36:32 VC's do in that environment it's not
2:36:34 worth them investing that little money
2:36:37 it's weird I need to hear you say it
2:36:40 everyone in my life just tune me out
2:36:41 when I start talking about AI Erica
2:36:44 Erica join us here look if you if you if
2:36:48 you weren't here earlier go to this URL
2:36:52 this doesn't cost anything by the way
2:36:54 this is the AI salon so this is a
2:36:55 community that I started the week after
2:36:58 chat gbt came
2:37:00 out the AI
2:37:02 Salon go there the first link goes
2:37:06 to um the AI Salon
2:37:10 community and it
2:37:12 is it's it's just this super go to the
2:37:16 welcome to the salon section and it will
2:37:18 tell you so so think of the AI Salon is
2:37:22 like the support group it it's like a
2:37:24 12-step program for people that are into
2:37:27 AI that can't get anyone in their
2:37:29 [ __ ] life to understand what's going
2:37:32 on like holy [ __ ] this changes
2:37:34 everything and people like I just don't
2:37:35 want to hear about AI I don't know if
2:37:37 you know this but they're plagiarism
2:37:38 machines you know it's stealing from
2:37:40 artists you know AI is you know it
2:37:42 hallucinates right and you're like but
2:37:45 but but you don't don't understand but
2:37:47 does anyone is anyone listening and and
2:37:50 everyone's just like no we're not
2:37:51 interested you know Erica could you stop
2:37:54 talking about that stuff cuz it's making
2:37:57 me uncomfortable they they say it's
2:37:59 going to take the jobs and I saw a movie
2:38:01 once that the AI is going to kill us so
2:38:04 do you think it's G to I don't want to
2:38:05 know don't tell me like they're going to
2:38:06 it's going to kill it I'm just gonna you
2:38:08 know what have you been to the mall you
2:38:10 want to go to the mall you want to get
2:38:11 an Orange
2:38:14 Julius this group
2:38:16 is not
2:38:18 that this group is like we're those
2:38:22 people that are saying that like
2:38:26 you and and and where I am at this with
2:38:29 this point is I think every everyone on
2:38:31 this channel and everyone at the salon
2:38:33 I'm encouraged to like live your life
2:38:36 out loud with this AI [ __ ] talk about it
2:38:39 I it doesn't matter if it pisses people
2:38:41 off keep talking about it keep
2:38:44 experimenting with it as you learn learn
2:38:46 stuff share it come to this group share
2:38:48 it so when you get here you know check
2:38:51 out the about the salon section it talks
2:38:54 about our values talks about why we're
2:38:56 doing what we're doing um there's a
2:38:59 community section where you can
2:39:00 introduce yourself so I would strongly
2:39:02 encourage you to introduce yourself I
2:39:04 noticed that no one tonight has
2:39:07 introduced themselves shame on you bad
2:39:09 bad people
2:39:11 um oh this was one hour ago so someone
2:39:15 did three days ago is the last one go
2:39:17 introduce yourself people um there's a
2:39:20 showand tell Channel everyone on AI
2:39:24 right now this is going to be good let's
2:39:27 see oh they have the internet on
2:39:29 computers
2:39:32 now molecular bot inist position
2:39:35 available so so what this what this
2:39:38 channel becomes is kind of like a like a
2:39:40 uh a jam session uh for people in here
2:39:43 when they get bored of me talking they
2:39:44 just go in here and make fun of me or
2:39:46 make cool images and share them um which
2:39:49 is super cool um and then there's
2:39:53 there's AI tools and resources so the
2:39:55 thing I've been you know sort of working
2:39:57 through with chat GPT um is is in this
2:40:02 there's also guilds um so there's all
2:40:05 sorts of different guilds there's a
2:40:06 healthcare Guild and an education Guild
2:40:08 and an AI mechanics Guild if you want to
2:40:10 geek out on how this [ __ ] Works um so
2:40:13 anyway so go go join the aslon it
2:40:15 doesn't cost anything it it really is
2:40:17 just a community of like-minded people
2:40:19 trying to figure this [ __ ] out and so
2:40:22 where is this this is there go there the
2:40:26 first link takes you to that it's a
2:40:29 mighty Network site Mighty network is
2:40:32 just a Community Management
2:40:34 tool so it's kind of like Discord but
2:40:36 sucks less like legit everyone thinks
2:40:39 I've I'm insane I found my people so so
2:40:43 good karmic fruit I mean yes
2:40:46 this is what it's about like this
2:40:51 is hang on let
2:40:55 [Music]
2:41:00 me that's a good
2:41:02 image let me move that
2:41:07 there
2:41:11 um so I joke in here
2:41:14 about not having qualifications people
2:41:17 come in here a lot and they're like sir
2:41:19 what are your qualifications I'd like to
2:41:21 know what are your
2:41:22 qualifications do you have a degree from
2:41:24 Stanford are are you a machine learning
2:41:28 scientist I have a degree in acting
2:41:33 um my qualifications are I'm really
2:41:36 curious about this [ __ ] and I think over
2:41:39 my lifetime I've I've been I've started
2:41:41 12 or 15
2:41:43 companies if there there's a couple of
2:41:45 through lines in my career one of them
2:41:47 is sort of Storytelling and and helping
2:41:49 people Express
2:41:52 themselves in all sorts of ways starting
2:41:54 a theater company
2:41:56 to starting an ad agency to starting a
2:41:59 magazine and then there's been a a
2:42:01 through line of
2:42:03 Technology Innovation or amplification
2:42:06 of human
2:42:07 self-expression and so I happen to be
2:42:10 lucky enough to learn about the
2:42:13 worldwide web when nobody really knew
2:42:15 about it or nobody cared about it and I
2:42:16 started one of the first online
2:42:18 magazines in 1994 and I started one of
2:42:21 the first oh I was an acting major too
2:42:24 yeah I'm tell well well so we we are
2:42:26 entering the era of Revenge of the
2:42:28 liberal arts major I can I can go deep
2:42:30 into that too but but I was lucky enough
2:42:33 to be in New York City and discovered
2:42:37 the worldwide web and I started one of
2:42:38 the early digital agencies a company
2:42:40 called agency.com so I've been in this
2:42:42 movie before I know what it was like
2:42:46 to be the only one in my friend group
2:42:49 like my friends were actors and writers
2:42:51 and directors in the New York theater
2:42:53 scene and I'm going you guys got to
2:42:55 check out this thing called the
2:42:56 worldwide web and they're like yeah we
2:42:59 know Kyle we don't care what what is it
2:43:02 like well it's this thing where you
2:43:04 click on a link and it take you can
2:43:06 there's a server in Paris that and you
2:43:08 could click on stuff and they're like
2:43:11 Kyle what are you talking about I'm
2:43:14 like and so I started a group back then
2:43:17 in '94 called the worldwide web artist
2:43:20 Consortium and it it was exactly like
2:43:23 the AI like the reason I started the AI
2:43:25 Salon a week after chat GPT came out is
2:43:29 I'm like holy [ __ ] this is another one
2:43:32 of those this is like the worldwide web
2:43:34 like this is more important than the
2:43:36 worldwide web was and no one knows about
2:43:40 it and so I started that back then and
2:43:43 and what the worldwide web artist
2:43:44 Consortium came was anyone in New York
2:43:48 City that knew anything about the web
2:43:50 they showed up to that thing because
2:43:52 everyone else in their life was like
2:43:54 could you please stop talking about this
2:43:56 worldwide web thing that's stupid we
2:43:57 don't care about
2:43:59 it until they
2:44:03 didn't because what happened was all of
2:44:06 us crazies that were talking about this
2:44:08 worldwide web thing
2:44:10 incessant eventually it turned into the
2:44:14 dot bubble where everybody wanted to
2:44:16 [ __ ] know about it everybody wanted
2:44:18 in and at that point we had all started
2:44:21 all these companies right it was crazy
2:44:24 that's where we are like we're in the
2:44:26 primordial soup stage of this this
2:44:28 generative AI
2:44:30 phase that you know the the
2:44:34 kickoff moment the the
2:44:37 the you know the meteor hitting the
2:44:41 planet was chat GPT launching November
2:44:44 30th 2020
2:44:47 22 this meteor came in that put a simple
2:44:51 interface on top of Decades of AI and
2:44:54 machine learning research champ what do
2:44:57 you want do you want to go out again
2:44:59 what do you want to do you want go hang
2:45:01 with mama why you bitching what's going
2:45:03 on you want up
2:45:05 here is that's it all right did you just
2:45:08 want love what do you want you
2:45:13 want no more bitching do you want to go
2:45:16 to Mama what do you want oh you want to
2:45:18 go to Mama all
2:45:22 right he wants to be by
2:45:32 you all right champ was tired of me he's
2:45:36 like I'm tired of you talking about this
2:45:37 AI
2:45:41 [ __ ] so so that's where we are and and
2:45:45 here's the the other thing if you're
2:45:46 here there's a lot of people in here
2:45:47 tonight if you're in here and you're
2:45:49 thinking like so there's there's a
2:45:52 couple of things you could be and are
2:45:54 likely thinking one is you're way behind
2:45:57 like oh my God I'm so behind there's all
2:45:59 these people in here they know
2:46:00 everything more than I
2:46:02 do
2:46:04 no you're not as as behind as you think
2:46:07 the fact that you're hanging out what
2:46:09 the [ __ ] time is it it's it's almost
2:46:12 11:00 in Denver that that means it's
2:46:15 it's almost 1:00 in the morning in New
2:46:17 York on a Monday night there's 270
2:46:20 people in
2:46:22 here if you're in here you're ahead of
2:46:24 the game if you're like paying attention
2:46:26 like wait what is this thing you're
2:46:28 early get curious about this [ __ ] get
2:46:31 curious and start talking about it yeah
2:46:34 but every time I talk about it think
2:46:35 people think I'm weird [ __ ] them keep
2:46:38 talking about it put yourself out there
2:46:40 what's happening in this group so
2:46:42 there's a there's a whole pile of people
2:46:44 in here called the irre regulars that
2:46:46 show up here
2:46:48 regularly like night after night after
2:46:51 night I do this seven nights a week they
2:46:52 show up night after night after night
2:46:54 why do they do that same reason I'm
2:46:57 doing it quite frankly which is they
2:46:59 just want to be in the conversation they
2:47:01 want to be learning this this stuff is
2:47:04 moving so fast you literally can't keep
2:47:07 up with it and
2:47:12 um if you think you're an expert
2:47:15 you're full of [ __ ] you're just a
2:47:17 salesperson anyone that tells you
2:47:19 they're an expert right now is full of
2:47:21 [ __ ] no one's an expert right now zero
2:47:24 people zero
2:47:26 zero
2:47:28 because the stuff that I'm really good
2:47:30 at today 3 months from now is going to
2:47:32 be irrelevant because the technology
2:47:35 just keeps leapfrogging itself it keeps
2:47:37 getting better and better and better and
2:47:39 better and
2:47:42 different 6 months ago everyone was
2:47:45 talking about prompt engineering you
2:47:47 need to be really good at prompt
2:47:48 engineering and structured prompting
2:47:50 and well we've now got interfaces for
2:47:54 images where there's no prompting at all
2:47:57 you draw pictures you draw shapes on the
2:48:00 left side and it generates images on the
2:48:02 right side in real time or you turn on
2:48:06 your webcam and you just
2:48:08 act and it generates images over here
2:48:11 and then you could put in a prompt that
2:48:13 says turn me into a waterfall in Bolivia
2:48:18 and up comes a waterfall in Bolivia and
2:48:20 you make your hands go like this and all
2:48:22 of a sudden there's a tree over the
2:48:24 waterfall everything is changing every
2:48:28 month so no one's an expert right now
2:48:31 and no one knows where this is going and
2:48:34 the people that are in it and by in it I
2:48:37 just mean people who are curious and
2:48:41 playing what we all know is that the
2:48:44 world has has changed we just don't know
2:48:47 how yet but we know it's changed we know
2:48:50 this stuff isn't going away we know that
2:48:54 as janky as it is right now the tools
2:48:57 that we're playing with today are as bad
2:48:59 as they're ever going to be they only
2:49:01 get better from
2:49:03 here my sister is very concerned about
2:49:05 data privacy with using AI she should be
2:49:08 these tools are janky ass pieces of [ __ ]
2:49:11 what your sister shouldn't be is so
2:49:13 concerned that she doesn't play with it
2:49:16 what she should be is curious enough to
2:49:18 understand where it's bad with data
2:49:21 privacy and how you could still use
2:49:23 these tools and not put you know your
2:49:26 company secrets into compromising
2:49:29 positions that's just
2:49:32 education will these tools get more
2:49:34 secure over time
2:49:37 absolutely that's not the stage of the
2:49:39 industry we're in right now the
2:49:41 opportunity is now if you get in now and
2:49:44 understand understand this stuff now
2:49:46 this is where all the opportunity is but
2:49:48 you need to understand what game you're
2:49:51 playing we are in the janky ass piece of
2:49:54 [ __ ] phase of generative AI meaning all
2:49:57 of these tools suck on some level or
2:50:00 other it doesn't mean you shouldn't be
2:50:02 building [ __ ] and learning about it and
2:50:05 exploring it and trying to reinvent your
2:50:07 business with it so Nate no database of
2:50:10 our gpts no there is now so so if you go
2:50:14 to chat GPT I didn't even get to this
2:50:16 tonight we're we're not going to finish
2:50:17 this tonight maybe we'll finish it
2:50:19 tomorrow if you go to chat GPT 4 explore
2:50:23 gpts is the database of all the gpts so
2:50:27 I've got one called Sin City sipper so
2:50:29 if I type in Sin
2:50:32 City it will search and I
2:50:35 am number one in terms of Sin City
2:50:38 sipper oh did it oh no I spelled it
2:50:40 wrong
2:50:41 Sin City
2:50:43 sipper s Sin City sipper and look I got
2:50:46 a bunch of clones I got four [ __ ]
2:50:49 that stole my my original image and
2:50:52 probably my code but there's my Sin City
2:50:55 sipper so now the thing the thing the
2:50:58 thing about the thing about the
2:51:03 um uh the thing
2:51:06 about the the database is it's just
2:51:09 really crappy right now the GPT store
2:51:11 really sucks right now um but it is a
2:51:14 database of all the gpts so this will
2:51:16 get better as well it's just just crappy
2:51:18 right
2:51:20 now um I made a drink from SIM SIM Sin
2:51:24 City awesome that's fantastic ha
2:51:27 imitation is the sincerest form of
2:51:28 flattery I know it is and that's listen
2:51:31 that's the thing the gpts right now it's
2:51:34 really easy for people to what's called
2:51:37 jailbreak your GPT unless you protect
2:51:40 against it and even if you protect
2:51:42 against it there's probably ways for
2:51:43 them to jailbreak it anyway so just
2:51:45 assume anything you build in a GPT will
2:51:47 be public knowledge if anyone played
2:51:49 with Luma a month ago check it out today
2:51:52 text to 3D okay good so that's on the if
2:51:55 you go to the AI Salon in the
2:51:57 recommended tool list and scroll down to
2:52:01 um 3D tools I think Luma is on
2:52:05 there yeah Luma AI Luma Labs text to 3D
2:52:13 objects announcing series B how much did
2:52:17 they
2:52:30 raise 70 million
2:52:33 total series
2:52:37 B they're not saying how much it was but
2:52:40 it's led by andreon Horowitz so that's
2:52:43 promising
2:52:47 Luma
2:52:50 imagine let's see um a
2:52:55 turtle on
2:52:57 a
2:53:00 rock
2:53:01 create sign in with
2:53:11 Google
2:53:13 create
2:53:20 wh turtle on a rock as text to 3D
2:53:25 objects creative creating interactive 3D
2:53:28 model almost
2:53:31 there come
2:53:37 on let me get to better
2:53:40 one I
2:53:43 think I think this
2:53:48 dude look at that three text to 3D model
2:53:52 and like this was this was so Luma I put
2:53:57 this on the list like two or three
2:53:59 months ago and it was like
2:54:00 embarrassingly bad back then this is now
2:54:04 decent right can I zoom in yep
2:54:08 wh so it's got you know janky textures
2:54:12 and janky models but you know what it
2:54:15 looks like a turtle on a
2:54:17 rock and it's a 3D
2:54:21 model so so and let's see make it hrz oh
2:54:25 that's let's let's do that hang tight
2:54:28 this will take a couple of minutes feel
2:54:30 free to check back
2:54:33 later
2:54:36 crazy yeah there you go there you go
2:54:40 people text to 3D objects
2:54:43 we
2:54:46 I that that's the other thing is so all
2:54:49 the stuff I talked about tonight I got
2:54:51 to I scratched the surface in two and a
2:54:53 half three hours I've been going three
2:54:55 hours tonight granted 40 minutes of me
2:54:58 just saying hello so let's call it two
2:55:03 hours I didn't even get to making gpts
2:55:06 with chat gp4 and I didn't get the code
2:55:09 interpreter and I spent two hours on it
2:55:12 now sure I'm talking a lot but
2:55:16 these tools are
2:55:17 deep and like Luma is just one of I
2:55:21 don't know what what I've got on that
2:55:23 list 25 different
2:55:25 tools there's image tools there's video
2:55:28 tools there's audio tools there's music
2:55:30 tools there's 3D tools there's people
2:55:33 you should
2:55:35 follow if you're just getting if if if
2:55:38 you're like one of the people it sounds
2:55:39 like there's a bunch of you in here
2:55:40 tonight like you're the only weirdo in
2:55:42 your group and no one [ __ ] gets what
2:55:44 what you're talking about keep going dig
2:55:48 deeper get relentlessly
2:55:52 curious relentlessly curious just dive
2:55:55 in and [ __ ] them if they don't get it
2:55:59 cuz I promise you what's going to happen
2:56:01 is
2:56:02 this your friend
2:56:04 Sally she's GNA be like you know
2:56:07 Margaret it's just this AI thing it's
2:56:10 like I don't it's listen I I know you're
2:56:12 really passionate about it and it's like
2:56:15 you've always been a little more
2:56:16 technical than me and I it's really but
2:56:19 could you just stop not talk about it
2:56:22 anymore thanks Margaret that would be
2:56:24 great Sally's gonna [ __ ] be back in
2:56:26 six months
2:56:28 Margaret Hey listen I just got laid off
2:56:32 and they said it was cuz I didn't know
2:56:33 AI Could you teach me AI
2:56:36 [Music]
2:56:38 please and then if Sally's not an
2:56:40 [ __ ]
2:56:42 teacher and if she is [ __ ] her
2:56:45 I guarantee you it's going to
2:56:49 happen this stuff is coming fast I
2:56:52 appreciate the demo I'm going to try
2:56:53 that one out tomorrow morning for a work
2:56:55 thing awesome perfect there you
2:56:57 go um I don't know which demo you're
2:57:00 talking about if it was Luma AI like I
2:57:02 don't know anything about Luma AI it's
2:57:04 possible there are other tools that are
2:57:05 better but the fact that they just
2:57:07 raised series B and they've got 70
2:57:08 million in funding I assume their series
2:57:10 B was like 50 million or something like
2:57:13 that so they so you know that's going to
2:57:17 turn that's going to get better fast
2:57:19 with with some some Capital behind it
2:57:21 like that I needed this pep talk thank
2:57:24 you you're welcome Erica
2:57:27 listen you're not the weird
2:57:31 one they
2:57:33 are and and they're not weird they're
2:57:39 scared here's here's another thing if if
2:57:41 you're brand new to this AI [ __ ]
2:57:45 the other thing that happened on
2:57:46 November 30th
2:57:50 2022
2:57:55 was Ai and machine learning was opened
2:57:58 up Kyle hydrate thank
2:58:01 you Ai and machine
2:58:03 learning was opened up to non-scientists
2:58:07 and non-engineers and non-researchers
2:58:09 chat
2:58:11 GPT basically opened the doors to the
2:58:14 rest of us to play with this stuff so if
2:58:18 you're sitting even on this call tonight
2:58:20 going yeah Kyle but you know I know you
2:58:23 got a degree in acting but you're
2:58:24 clearly Technical and you you know you
2:58:27 you know this computer stuff and you
2:58:28 know
2:58:30 coding I'm curious about it I'm not an
2:58:34 engineer I know how to make software
2:58:36 because I've been on the planet for a
2:58:39 while you don't have to be technical
2:58:43 three weeks ago on a
2:58:45 Saturday with a week's
2:58:47 planning we decided that we would teach
2:58:50 as many people as we could to make gpts
2:58:53 for
2:58:54 nonprofits for people that had not made
2:58:57 software
2:58:58 before and this community [ __ ]
2:59:01 stepped up this
2:59:03 community
2:59:05 created um 150 gpts for nonprofits in a
2:59:09 24-hour period 3 weeks ago on
2:59:12 Saturday that is a fundament mentally
2:59:15 different you you know you hear about
2:59:17 these things there's a hackathon and
2:59:19 what are hackathons well you get
2:59:21 technical people engineers in a room
2:59:25 with business stakeholders people who
2:59:28 have business ideas that don't know how
2:59:29 to code
2:59:31 [ __ ] and you put them in a room for 48
2:59:33 hours and they build however many apps
2:59:36 you have based on however many Geeks
2:59:37 showed
2:59:39 up well we
2:59:41 had we don't know who showed up
2:59:45 but we teach them in taught them in less
2:59:47 than an hour to build a
2:59:50 sophisticated customized GPT for a
2:59:53 nonprofit in less than an hour we did
2:59:55 150 in 24
2:59:58 hours without breaking a
3:00:00 sweat
3:00:02 well we broke a sweat cuz was
3:00:09 long Rudy I never thought I never ever
3:00:12 thought I would be doing what I'm doing
3:00:14 and it's all thanks to this guy oh thank
3:00:15 you Rudy R Rudy's cool you should follow
3:00:17 Rudy R Rudy saigo he's a young dude if
3:00:20 you're if you don't if you're not into
3:00:21 old dudes like me you got Rudy he's got
3:00:24 energy he's got editing
3:00:27 skills and he's [ __ ] teaching up a
3:00:29 storm he's doing YouTube videos on how
3:00:31 to do programming
3:00:34 stuff and and a lot of the people in
3:00:36 here are doing the same thing they're
3:00:37 doing it on the AI how to do programming
3:00:41 stuff and and a lot of the people in
3:00:43 here are doing the same same thing
3:00:44 they're doing it on the AI um
3:00:47 imperfection embrace the
3:00:53 janky so anyway I got to get going this
3:00:57 was a long one
3:00:59 tonight I was uh it was particularly
3:01:03 soap boxy but you know I get EX I you
3:01:06 know what it is I get [ __ ] excited
3:01:09 about the
3:01:11 possibility like why do I keep coming
3:01:13 back here night after night it doesn't
3:01:16 make [ __ ]
3:01:23 sense it's to remind myself to stay in
3:01:26 the conversation quite frankly thank you
3:01:28 for your time and sharing the info of
3:01:30 course Rachel appreciate
3:01:40 it oh man all right so it's Monday night
3:01:44 so here's our new motto tgim thank God
3:01:47 it's
3:01:50 Monday
3:01:51 because TG TGIF implies that we hate our
3:01:56 jobs and we hate our life and thank God
3:01:59 it's Friday means I can leave
3:02:02 Lumberg and the TPS reports and go to
3:02:06 chachis and have a overpriced
3:02:09 Margarita let's flip that script and say
3:02:12 tgim I get to
3:02:15 choose whatever the [ __ ] I want to do
3:02:18 whether I have the skills or not with
3:02:20 these tools I get to do whatever the
3:02:23 [ __ ] I want to do so we're going to
3:02:25 start that today
3:02:28 tgim Mr Shannon you know you've reached
3:02:31 306 viewers today tonight wow that's
3:02:33 cool that's awesome really cool I
3:02:36 appreciate y'all hanging out dang 270
3:02:38 members I think you may be oh I had no
3:02:41 idea what a hackathon was I never go as
3:02:45 I thought I had to be an engineer and
3:02:48 have lots of business ideas Suzanne so
3:02:50 that's the
3:02:52 deal that's what hackathons
3:02:54 were what makes me excited about
3:02:58 gpts is that is it is the first in to my
3:03:01 knowledge it is the first
3:03:05 truly democratized application building
3:03:09 tool on the planet now the Geeks are
3:03:13 going to say well you've got no code
3:03:14 tools like zappier and make and bubble
3:03:17 and [ __ ] like that have you used those
3:03:19 [ __ ] tools you still need to be a
3:03:22 programmer mentally you still need to
3:03:24 understand programming logic and apis
3:03:27 and how to connect the Lego blocks
3:03:29 together gpts are just like tell it what
3:03:31 you want and it'll [ __ ] make it for
3:03:33 you like it's true democratization of
3:03:36 sophisticated application
3:03:38 development now does it help if you're a
3:03:40 little geeky
3:03:42 sure but it's not a requirement that's
3:03:44 new that's very new what is the
3:03:47 advantage of making your own gpts over
3:03:49 asking the main one what you need
3:03:55 um uh replicability and the ability to
3:03:59 share it so and and and a narrow Focus
3:04:04 so anything that you can do in a GPT if
3:04:07 you go to someone's GPT that they
3:04:10 made and let's say let's say like I've
3:04:13 got one called the 7s muscle car maker
3:04:17 machine or something like that 70 yeah
3:04:20 '70s muscle car maker and what it does
3:04:23 is
3:04:24 is you it asks you for your favorite 70s
3:04:28 muscle car
3:04:29 make and then you give it a color and
3:04:32 then you give it any modifications like
3:04:34 do you want a chrome blower on it or
3:04:36 side pipes or drag wheels or whatever
3:04:40 spoiler and then it makes you a poster
3:04:43 with a nick name of the car on
3:04:45 it now you could go to chat GPT and you
3:04:50 could say oh that's a cool idea and you
3:04:52 could figure out how to prompt it to
3:04:55 give you a poster that looks like the
3:04:57 poster I made but you don't have to you
3:05:00 can just go use my thing so I can share
3:05:02 with you the 70s muscle car maker GPT
3:05:06 and you can just go there Cory Sandler's
3:05:08 got this GPT called iconic eye conic
3:05:13 where you either upload a photo of your
3:05:15 face or you just describe your face and
3:05:18 then it makes that face and then it'll
3:05:19 put wacky glasses on you and it just
3:05:22 it's a glasses designer thing could you
3:05:24 prompt that on your own to do that you
3:05:26 could but she's super passionate about
3:05:28 her particular flavor of design and so
3:05:31 it makes just wacky ass glasses that are
3:05:34 just a
3:05:35 joy so and she can share that with the
3:05:38 world and they get to
3:05:41 experience doing that without without
3:05:43 having to figure that [ __ ] out so
3:05:46 technically you could prompt this thing
3:05:48 to do anything but the why you do it
3:05:51 it's because you've got some particular
3:05:54 passion or expertise you want to share
3:05:57 with the world it's true self-expression
3:05:59 through application development it's
3:06:01 it's pretty pretty
3:06:04 amazing Mr it I didn't know nothing
3:06:07 about chat GPT a month ago and I have
3:06:10 like 10 gpts and I'm planning to go live
3:06:14 on Tik Tok awesome awesome Mr it [ __ ]
3:06:18 perfect yes tell the world what you're
3:06:25 learning the McKenzie report that came
3:06:27 out two months ago 82% of people in the
3:06:31 US do not use chat
3:06:32 GPT 82% that means
3:06:36 18% have used
3:06:38 it how many of that 18% use it a lot
3:06:43 probably not many how many have paid for
3:06:46 chat gp4 and use that a lot probably not
3:06:49 many so you're looking
3:06:52 at
3:06:55 5% of people might be [ __ ] around
3:06:59 with this on a regular basis maybe
3:07:07 5% that means 95% don't have a [ __ ]
3:07:11 clue what's
3:07:12 coming
3:07:15 we're so early we're so
3:07:17 early so that's why people are calling
3:07:19 you weirdos if you're talking about this
3:07:21 stuff they're in the
3:07:24 95% so even if you're just here and you
3:07:27 haven't even tried okay first of all
3:07:29 there's 2 158 people here it's [ __ ]
3:07:31 it's 11:00 it's 1111 in
3:07:34 Denver if you're here and you haven't
3:07:37 gone to chat.
3:07:39 open.com and typed something into chat
3:07:41 GPT go [ __ ] do it right
3:07:45 now like not I'll get to it tomorrow I
3:07:48 don't feel like learning anything Kyle
3:07:50 I'm not very
3:07:52 technical go there right now chat.
3:07:56 open.com hit return you'll see a Google
3:07:59 search box at the bottom of the screen
3:08:01 type in explain quantum mechanics to me
3:08:04 as an Eminem
3:08:06 rap and watch what it
3:08:10 does if you haven't done it [ __ ] go
3:08:13 do it right
3:08:15 now get thee to a nunery go seriously
3:08:19 right
3:08:20 now I'll
3:08:27 wait this is my first time using it
3:08:29 tonight after watching so many of your
3:08:31 lives perfect Rachel Sparkles 77 love
3:08:36 your name by the way
3:08:39 perfect be like
3:08:41 Rachel don't sit on the Sidelines with
3:08:44 this [ __ ] chat.
3:08:46 open.com oh he must work for openai oh
3:08:49 you know what this is he's just Shilling
3:08:51 for chat GPT I I guarantee you he gets
3:08:55 the
3:08:57 commission we've established in here Sam
3:08:59 Alman is jealous of me he will not pay
3:09:01 me commissions even though this channel
3:09:03 is essentially a [ __ ] ad for his
3:09:06 company because he's bitter that's okay
3:09:09 he just got married I wish him
3:09:12 well go play with it right now Tobias
3:09:15 welcome everyone follow this channel
3:09:17 thank you very much yeah follow this
3:09:19 listen I'm going to get out of here I'm
3:09:20 going to begin my half an hour goodbyes
3:09:23 um if you haven't checked it out go
3:09:28 check out the
3:09:29 salon. go there the F the first link is
3:09:33 to our community to the mighty networks
3:09:35 Community the second link is to the GPT
3:09:38 for good project we did I mentioned
3:09:40 before we did this project where we this
3:09:43 community made 150 gpts in 24 hours the
3:09:46 directory of all those gpts is the
3:09:49 second link on that link
3:09:52 tree go check them out it's
3:09:55 inspiring and go check out the AI Salon
3:09:57 this is a community of people that will
3:10:00 support you in figuring this [ __ ] out 30
3:10:03 minute goodbyes begin now thank you be
3:10:05 Not
3:10:06 Afraid um so go check that
3:10:10 out and um yeah
3:10:14 and start talking about this [ __ ] as you
3:10:15 learn about it share what you're
3:10:17 learning share what you're learning
3:10:18 thank you Robert Rossy I really
3:10:20 appreciate
3:10:21 this Athena Luna first gift thank you so
3:10:24 much that's awesome demo Sin City or any
3:10:28 GPT real fast please okay cool um I'll
3:10:31 do that for Joker Joker knew that I was
3:10:33 trying to get out of here so he's like
3:10:34 Kyle just do one more thing he knows
3:10:35 it's going to take me like four [ __ ]
3:10:37 hours okay so
3:10:41 um here we are at chat GPT so chat GPT
3:10:44 is the thing I was kind of demoing all
3:10:46 night and this is the thing if you
3:10:48 haven't been here go here right now now
3:10:50 if you've got a chat GPT Plus
3:10:53 subscription flip over to chat GPT 4 and
3:10:56 you can explore these things called gpts
3:10:58 so if you click on explore
3:11:00 gpts that will take you to the GPT page
3:11:03 and you can kind of scroll down through
3:11:04 here there's some different categories
3:11:06 and things like that but you can also
3:11:08 search for GPT so I'm going to demo my
3:11:10 very first GPT I ever made I made it on
3:11:12 this channel and it's called Sin City
3:11:17 sipper and this is it so it's a Vegas
3:11:22 bartender with a Twist on the classics
3:11:25 and then it's got these little starter
3:11:27 things so like show me a new fangled
3:11:29 oldfashioned reinvent the margar
3:11:32 margarita for me I'm a tea Toler make me
3:11:34 a classy mocktail so we we'll do that
3:11:38 we'll do a mocktail um and so what this
3:11:41 thing does is it will
3:11:44 invent a drink for you invent a cocktail
3:11:46 and put a Vegas Twist on it ah a tea
3:11:49 todler in Sin City now that's a rare
3:11:51 exciting but worry not I've got just a
3:11:53 thing for you a classy non-alcoholic
3:11:55 mocktail that's high on taste and low on
3:11:57 fuss let's stir up a sparkling Citrus
3:12:01 Sunrise half a cup of fresh orange
3:12:04 juice4 cup of fresh lemon juice 2
3:12:06 tablespoons granadine ice cubes
3:12:08 sparkling water or club soda orange
3:12:10 slices sounds delicious I used to Bart
3:12:12 10 I've made a ton of these things
3:12:17 um and then it it explains to you how to
3:12:20 make it and then it
3:12:24 makes the image for you of the cocktail
3:12:27 what what blows me away about this GPT
3:12:29 so I built this GPT in about 10 minutes
3:12:34 on this channel somewhere in my on the
3:12:36 YouTube channel I've I've got the
3:12:37 recording of this this was the very
3:12:40 first GPT I ever made it like gpts
3:12:44 became available while I was live one
3:12:46 night so I'd never made one so I just
3:12:49 went I went to
3:12:51 them
3:12:53 I went in to make it I typed in I want
3:12:57 to make a thing that makes drinks with a
3:12:59 Vegas twist or something like that it
3:13:01 came up with the name after a little bit
3:13:03 of prodding um and this thing was
3:13:06 functional in about 10 minutes and then
3:13:09 I've since sort of refined it a bit but
3:13:11 but this is just amazing and then what
3:13:13 it does is it also says it's got a call
3:13:16 to action in it it says this GPT was
3:13:18 created by Kyle Shannon a member of the
3:13:20 AI salon and host of the AI learning lab
3:13:23 on Tik Tok so this call to action is
3:13:26 built into this application so this is
3:13:28 just a fun little thing where you can go
3:13:29 make drinks and you know it'll invent
3:13:32 drinks for you which is a blast and then
3:13:34 it's a little marketing you know it's
3:13:36 it's a little marketing tool for the AI
3:13:38 salon and the and the AI thing so anyway
3:13:41 so that's what gpts are oh oh there are
3:13:43 images now nice yeah exactly you can
3:13:45 turn on image generation in these things
3:13:48 and and some people have really taken
3:13:49 that to the next level Cory Sandlers are
3:13:51 really good um Danielle be not afraid's
3:13:54 got a bunch in here she's got one that
3:13:55 makes uh sugar cookies with with icing
3:13:58 on them in all different shapes and
3:14:00 patterns Vicky will then Vicky in here
3:14:04 um has got a 3D printing business so if
3:14:07 you if you make some sugar cookies in in
3:14:10 Danielle's gpts that you like you you
3:14:13 can reach out to Vicki and she'll turn
3:14:15 those designs into um uh cookie cutters
3:14:19 that are specific to to your design like
3:14:21 there's just really cool [ __ ] going on
3:14:23 in this community so anyway that's
3:14:26 that's what gpts are that's the Sin City
3:14:29 sipper Rachel Sparkle 77 we made you a
3:14:33 sign Kyle and show and tell oh yeah let
3:14:34 me go to show and tell now so so in the
3:14:37 AI
3:14:38 Salon where's the AI Salon did I lose it
3:14:41 no there it is um the show and tell
3:14:43 channel it's just nice to go through
3:14:46 this at the end of the night be like
3:14:48 Rachel
3:14:50 nice there you go Rachel Rachel Sparkles
3:14:54 you got your own
3:14:55 poster that's really cool it looks like
3:14:58 a it looks like a poster in a in like a
3:15:01 uh an Olden timey uh retail store like a
3:15:04 dollar store that's funny thank God it's
3:15:06 Monday very
3:15:09 nice stay stay my qualifications are I'm
3:15:13 curious about this [ __ ] that's good by
3:15:15 the way don't get too upset that these
3:15:17 things can't spell Embrace that because
3:15:19 6 months from now they're going to be
3:15:21 perfect and it's going to be
3:15:22 boring molecular botanist beeps Burgers
3:15:27 very
3:15:30 cute oh that's nice actually it's Leon
3:15:34 sir there's the rabbit very nice very
3:15:39 cool janky Cola very nice the official
3:15:42 beverage Bridge of the AI Salon you can
3:15:45 make money with
3:15:47 [Laughter]
3:15:51 ch yeah very very cool very very cool
3:15:57 please ask more people to include
3:15:59 prompts in show and tell it helps learn
3:16:01 that's a good point Vic so yeah so if
3:16:03 you're posting stuff to to show and tell
3:16:06 especially if you're using something
3:16:08 other than DOI if you're using mid
3:16:09 Journey but maybe even if you're using
3:16:11 DOI give it a hashtag of the tool you're
3:16:13 using um and put either the full prompt
3:16:16 in there or something that you might
3:16:18 have done differently to get to it um or
3:16:21 if you use someone's GPT to generate an
3:16:24 image put a link to that their
3:16:26 GPT um Rachel Sparkles loves that that
3:16:30 we made a poster for her that's
3:16:32 awesome this I'm telling you this group
3:16:35 is really cool like this is a good place
3:16:37 to hang out it really is my GPT had 100K
3:16:42 chats oh that's super cool Joker which
3:16:44 one he doesn't get my rabbit
3:16:47 device I'm trying to avoid getting
3:16:50 kicked off Tik Tock
3:16:52 Joker I got
3:16:55 it just use Photoshop to fix the
3:16:58 spelling yeah exactly exactly it's two
3:17:01 devices in one it's a dessert topping
3:17:04 and a floor wax that's the rabbit R2 the
3:17:07 rabbit R2 is the is the
3:17:11 Multi-Device
3:17:13 you press one but button to activate
3:17:15 chat GPT and the other button to
3:17:17 activate something else um so will that
3:17:20 help set up and run websites for
3:17:22 affiliate marketing programs I don't
3:17:24 know what the that is that you're
3:17:27 talking about but possibly um I got to
3:17:29 stop procrastinating and join the group
3:17:31 already no [ __ ] Shadow Ranger yeah go
3:17:34 join the AI Salon it doesn't cost you
3:17:36 anything you literally just have to go
3:17:38 to the
3:17:39 salon. just go there click on the first
3:17:42 link and sign up and introduce yourself
3:17:46 all
3:17:47 right okay so I'm going to get the [ __ ]
3:17:50 out of here I can't talk anymore um I
3:17:52 joined tonight that's awesome um yeah
3:17:55 come on over already exactly so go join
3:17:57 the salon um thank you all for hanging
3:18:00 out uh this was this was a long one
3:18:02 tonight it was a fun one Be not Afraid
3:18:04 thank you so much for the Roses I
3:18:05 appreciate it so to my subscribers thank
3:18:07 you to the Irregulars the people that
3:18:10 show up here night after night after
3:18:11 night you're [ __ ] amazing to my mods
3:18:14 that that help um curate all of the
3:18:17 questions so that I see I see the
3:18:19 important ones thank you um for anyone
3:18:22 who showed up tonight like this is this
3:18:24 is sort of double what we normally see
3:18:26 in here so there's a lot of people
3:18:27 hanging out hey thank you for all you do
3:18:29 for us listening from Israel oh that's
3:18:31 awesome thank you so much thank you so
3:18:34 much um Clint Copperhead everyone's IR
3:18:38 regulars mods present Turtles llamas
3:18:41 hedgehogs
3:18:43 um thank you all thank you all for
3:18:46 hanging out tgim thank God it's Monday
3:18:49 get curious about this [ __ ] jump in talk
3:18:53 about it let your friends think you're
3:18:54 [ __ ] weirdos and then come hang out
3:18:56 in the a AI Salon cuz we know you're not
3:18:59 weirdos we just know that this is the
3:19:01 thing about this AI stuff this is red
3:19:02 pill blue pill stuff when you have your
3:19:05 Kevin McAllister moment with AI if you
3:19:07 haven't had it yet it's this
3:19:09 moment it does that when you have that
3:19:13 moment you can't go back from it it's
3:19:16 it's like you can't unsee it and you're
3:19:18 just like it it your existence will now
3:19:22 be like this wait it does what now oh
3:19:27 god really okay that's exciting right
3:19:30 it's
3:19:31 good this is good right that's what your
3:19:35 life's going to be like after you jump
3:19:36 in here 10 p.m. EST every night yeah I
3:19:39 try to I try to be consistent about my
3:19:41 time I didn't used to be but I try to
3:19:44 8:00 p.m. mountain 10 p.m.
3:19:47 Eastern so just jump into it and it's
3:19:50 going to make your head [ __ ] hurt and
3:19:51 you will you will be completely
3:19:53 overwhelmed and you'll be more excited
3:19:55 than you've been in
3:19:57 decades that's my existence right now is
3:20:00 I'm more creatively alive than I've been
3:20:02 in decades so anyway all right I got to
3:20:06 get out of here people I love you all
3:20:08 let's do it let's do the Romper Room
3:20:09 thing Tobias you're awesome Jo py good
3:20:13 to see you Ai and chat GPT mystery nice
3:20:16 name how can I send you a free chat GPT
3:20:19 mug uh just connect I've got my contact
3:20:22 infos up there somewhere or just DM me
3:20:25 in here I'll take a free chat GPT mug
3:20:29 nice um promps cart good to see you ath
3:20:33 Athena Luna
3:20:35 steo Corey Sandler Pottery night night
3:20:38 Be not Afraid thank you as always Robert
3:20:40 Rossy I really appreciate your
3:20:41 generosity you're amazing um Mr it thank
3:20:45 you so much check Show and Tell I will
3:20:48 do that uh Suzanne good to see you
3:20:55 uh who else we got in here you can get
3:20:59 mugs with chachy
3:21:04 P Amelia's wife good to see you all
3:21:07 right everybody um this was a fun one
3:21:09 tonight Rudy great to see you um all
3:21:12 right I'm out of here peace out have a
3:21:14 good Monday I'll see you tomorrow
3:21:17 peace