2/10/2025 - AI, Super Bowl Ads, and the Future of Work: A Meandering Monday Chat

Description

Chapters

Transcript

0:01 [Music]
0:27 oh oh
0:30 oh
0:31 [Music]
1:04 [Music]
1:09 [Applause]
1:15 oh
1:20 [Music]
1:39 h
1:43 [Music]
2:02 don't K my way that
2:05 [Music]
2:07 night just like a jet
2:10 plane in and out of
2:14 side I was hauling ass at a million
2:17 miles an hour wondering how hard i'
2:25 here when they came into the station I
2:32 they said I was bad Beyond
2:37 repair but I got no quals with my
2:43 situation say here I
2:51 am so say sh Shere Shere won't you dare
2:58 to say Sheree Shere Shere won't you dare
3:04 to say Shere Shere
3:08 Shere
3:11 ah yeah leave a message and your message
3:18 please take a time to want to satisfy
3:23 me take all these old fantasies and send
3:27 them care of me
3:30 [Music]
3:32 good evening good people what is
3:36 happening I am a little early
3:39 indeed I was just sitting here champy
3:43 was sitting here we were thinking should
3:46 we go should we go what if we want
3:50 to no time like the
3:53 [Music]
3:55 present it's about time now it's about
3:57 time to get started
3:59 [Music]
4:29 oh man hope you had a good weekend let
4:32 me flip my little cameras here hope you
4:34 had a good
4:38 weekend I'm on Tik Tock joining the
4:40 stream soonish all right there Mr
4:43 Brandon producers inbound got a producer
4:47 inbound
4:49 [Music]
4:52 [Applause]
4:53 [Music]
5:14 uh what are we G to talk about well
5:21 so I want to get back to deep
5:24 Dives but it's also Monday Monday is
5:27 kind of a I don't know it's they they
5:30 have historically been more
5:32 chatty
5:34 philosophical Meandering Mondays we like
5:37 to call them welcome to chat add my name
5:40 is Kyle Shannon this is the AI learning
5:43 lab uh the agendas here are loose if
5:50 [Music]
5:56 non-existent I had a couple of good
5:58 conversations today though about this AI
6:05 stuff and I keep I keep going back
6:10 to I just have
6:16 this I have this
6:27 nagging this nagging gravity pulling me
6:30 back
6:32 toward talk
6:34 about the future of
6:38 work and it's it's annoying
6:43 because I don't have a real clue like
6:46 none of us
6:49 do but I think I know what's going to be
6:54 important and what's what's going to be
6:56 more important what's going to be less
6:58 important and I think some people are
7:00 going to thrive and other people are
7:02 going to struggle like
7:07 hell so I just thought it might be good
7:09 to just kind of go back through some
7:11 posts that I've done
7:13 and just talk about what I think's
7:16 coming and why and but I don't know if
7:19 you have questions about AI you can pop
7:21 them in the comments
7:23 below mandering is also a nice choice
7:26 very nice happy Monday Valerie Jeff
7:29 Flanigan what's happening seems excited
7:31 about the topic future of work and by
7:35 work you know part of it is work what we
7:39 do in exchange for money which
7:43 money is something that we exchange our
7:46 time for right the uh the definition of
7:50 money that I learned from this book
7:52 called your money or your life by Joe
7:57 Dominguez he basically said time is the
7:59 one thing you can't get
8:04 [Music]
8:14 back and figuring out that pattern that
8:17 strumming pattern is something I can't
8:20 get time back
8:25 on
8:27 um so if money is something that we
8:30 exchange our time for and in today's
8:34 world we have to do this thing called
8:36 work to get that
8:38 [Music]
8:42 money what what do we do with our
8:46 time when the AI does most if not all of
8:49 the
8:56 work and there's a part of me that's
8:58 like but Kyle that's like
9:00 years
9:04 away and I'm like and yes it is years
9:08 away but it's more like three to five
9:10 years away rather than 10 to 50 years
9:13 away which is I think where where most
9:17 of us historically had been if if you're
9:20 of an
9:22 age like say
9:25 uh you know Gen X Boomer Elder
9:28 millennial
9:29 [Music]
9:33 we were fairly confident we didn't
9:35 really have to deal with this
9:39 [Music]
9:51 [ __ ] in a
9:53 Westerly
9:55 [Music]
9:58 Direction this car is my
10:03 train all right Vicki what's the next
10:08 line I've been driving I've been
10:11 [Music]
10:12 wondering what it is I'm running from
10:18 again feel like an 8-year-old
10:22 man holding on to
10:26 29 uphead on that Horizon
10:31 is California
10:36 L
10:39 [Music]
10:41 yeah
10:44 um
10:46 so what the [ __ ] are we supposed to
10:51 do if you're watching this
10:54 Channel first of all you're
10:58 weird and sure we've reframed weird as
11:02 irregular so you can call it
11:04 irregular but it's weird you're you're
11:08 early no I know it's late now but you're
11:11 early in the conversation with this AI I
11:14 was in the salon posting recipes very
11:17 nice very nice love it Jeff Flanigan put
11:20 mistol on his laptop we'll talk a little
11:22 bit about mistol tonight although I
11:24 haven't played with it yet I was big in
11:26 mistol was the big thing like a year ago
11:30 then it just
11:33 disappeared and now it's back with a
11:36 vengeance um like all the like all the
11:38 open-
11:39 source models they're a bit of a pain in
11:42 the ass because you have to install them
11:44 and figure [ __ ] out
11:47 and sort through them yourself which you
11:50 know engineering
11:52 types like that they like geeking out on
11:55 that
11:57 stuff there's so much to keep up with
12:00 though even with the non-op source stuff
12:02 that I have not prioritized
12:05 that but it does seem like nistal is a
12:08 pretty good interesting thing to play
12:10 with irregularly weird or weirdly
12:13 irregular I have found my people exactly
12:17 exactly we're all a bunch of
12:21 weirdos cheers to being weird or
12:23 irregular yeah
12:29 [Music]
12:35 um I thought the the Super Bowl was
12:38 fascinating last night um not the
12:42 game not the the game was [ __ ]
12:46 horrible the game was like what it was
12:50 it was like an NF an NFL team was
12:54 playing like a peeeee football league
12:57 last night it was bizarre I've never
12:59 seen any like it it was just an absolute
13:02 meltdown on Kansas City's side um but
13:07 the halftime show I knew a little about
13:09 the Drake Kendrick Lamar
13:12 controversy um last night and I knew
13:15 that there was a big thing whether he
13:17 was going to sing the song or not and
13:18 whether he was going to say the phrase
13:20 or not and he did if you haven't seen
13:23 it and but something that gave me great
13:26 joy today is like my entire timeline on
13:30 Tik Tock was just filled with people
13:33 that know that are deeply in that world
13:36 like losing their Collective [ __ ] like
13:40 all sort sort of showing different parts
13:42 of that performance and analizing it and
13:46 breaking it down and it was just I felt
13:49 like I got such an education today such
13:52 a cultural education um absolutely
13:55 [ __ ] awesome and Kendrick Lamar is
14:00 [ __ ]
14:02 beast there is something I have a very
14:06 strong place in my heart for True
14:08 artists and that was
14:11 a that was a Tor to [ __ ] Force I
14:16 mean that's a performance that is going
14:18 to be talked about for a long time to
14:21 come and even if you didn't get it go
14:24 watch Tik Tok go watch some of the
14:26 videos of people watching it in real
14:28 time it was insane
14:30 go
14:31 Eagles I wish I knew the backstory
14:34 before I watched the performance would
14:35 have appreciated it so much more yeah
14:37 exactly I know and I didn't know the
14:40 songs it's like and but you're watching
14:42 people from that world and they're like
14:44 they know every [ __ ] word they know
14:46 what songs he's gonna sing and which
14:48 ones they want him to sing and the
14:50 significance if he does sing them it was
14:53 just like and they were just losing
14:55 their Collective [ __ ] like no
15:00 [Music]
15:07 um so that was pretty
15:08 cool and then we had a couple of you
15:12 know a couple of large language model
15:15 companies blew eight8 million on 60
15:19 seconds of
15:23 commercials um Geminis were fairly
15:25 straightforward they were kind of like
15:27 old Apple ads here's what it does here's
15:31 the difference it's going to make in
15:32 your life um anthropics I kind of missed
15:35 but I I I don't know really what it was
15:38 but I don't know what the [ __ ] anthropic
15:40 is doing I are they going to ever
15:42 release anything
15:44 again and then open AI
15:51 commercial um there's there's a couple
15:54 of different ways you can go with Super
15:55 Bowl
15:57 commercials you know one is you go with
16:00 something absolutely
16:03 Unforgettable big
16:07 star something really funny A New
16:14 Concept then there's talk about the
16:18 product which given that nobody knows
16:20 what the [ __ ] chat GPT is might have
16:23 been a good opportunity to talk a little
16:25 bit
16:30 oh perplexity apparently gave away a
16:31 million dollars all right there you go
16:35 perplexity giving away a million bucks
16:38 in the app oh during the Super Bowl
16:40 there you have it see some some sort of
16:43 stunt like that would have made sense
16:46 but what opena I went
16:48 with was a crazy
16:53 ass
16:55 animation that featured their Dot and it
16:58 was like when I first saw it I'm like oh
17:02 that's like the chat GPT dot when you
17:04 put it into voice
17:06 mode but then they made lots of dots and
17:08 then the dots started forming things and
17:10 I couldn't quite tell what the [ __ ] the
17:11 dots were forming until halfway through
17:14 the commercial I was like oh it's human
17:17 achievements okay I'm like but this this
17:21 this please tell me this isn't open
17:25 Ai and it
17:27 was so they went for a branding they
17:30 went for a branding play right like the
17:34 the most famous branding
17:36 play in the in the history of Super Bowl
17:39 ads was the 1984 ad from Apple computer
17:42 right directed by Ridley Scott this
17:44 woman running down the hallway throwing
17:46 the
17:47 sledgehammer at the 1984 dystopian
17:50 future which we now live in here we are
17:54 did you see Kendrick's show
17:59 [Music]
18:03 um you have no [ __ ] idea what it's
18:05 about right and then it just at the end
18:07 it just says Macintosh you know 1984
18:10 whatever um that's the most famous
18:12 branding play that was a memorable
18:14 moment and a branding play all at once
18:17 so in a branding play you're just trying
18:19 to get people to be be aware and
18:22 remember your
18:24 [Music]
18:26 brand um
18:30 so let me let me change my sharing here
18:32 for a
18:38 [Music]
18:46 second hold please hold please open
18:53 AI supera Bowl
19:01 all
19:03 right here we
19:09 [Music]
19:11 go Jeff on YouTube from Yahoo finance
19:15 Elon Musk yeah I'll get to that that's a
19:17 big story I'll get to that in a second
19:20 share screen share
19:24 X all right there we go we got our
19:28 producer and in the house yo Brandon was
19:31 shaking Big B Big B in the
19:35 house um all right so
19:39 [Music]
19:50 let's right right so there's your
19:52 there's your do right that's feeling
19:55 very open a i
19:57 e open AI e EI e IO um I did not need
20:01 you to do the math for me which math
20:04 what math did I
20:05 do oh Emo's wife did the
20:08 math oh two commercials about
20:15 anti-hate well there there was a lot of
20:17 hate in the room yeah
20:20 exactly there there was a there was a
20:22 lot of a lot of what they they the one
20:27 the one commercial I thought had a
20:28 clever
20:29 hooked to
20:31 the to the stadium was the one where
20:34 they sang the John Denver song but it
20:36 was a weird arrangement of the John
20:38 Denver song and then they cut back to
20:40 the Super Bowl and everyone's singing
20:42 the same song in the Super Bowl that was
20:43 kind of cool um I don't remember what
20:46 the product is the problem with Super
20:48 Bowl ads is you never know what the
20:49 [ __ ] product is but anyway let's
20:50 let's watch this open AI Tik Tock pin
20:53 1984 plus 40 I know it's crazy right um
20:57 here's the uh here's the open AI
21:00 ad Super Bowl
21:09 [Applause]
21:09 [Music]
21:18 [Music]
21:25 ad go back right here
21:29 this when we're like
21:32 corn we got horses we got corn we got
21:35 wheat fields at this point I'm like wait
21:37 a minute maybe this is a Monsanto
21:42 ad I mean I really I'm like trying to
21:45 think like wait what what is this
21:56 [Music]
21:59 was that not
22:06 AT&T right there I'm like oh maybe this
22:08 is an AT&T
22:10 [Music]
22:18 ad live
22:21 on right here I started to get it so the
22:23 commercial
22:25 changes right here so it goes from
22:27 everything's open and white right here
22:30 and then here it flips to a more photo
22:32 realistic this is like grayscale Photo
22:34 realistic
22:36 stuff then there's that weird AT&T
22:38 moment but then there's this is now
22:40 starting to feel like
22:43 [Music]
22:48 video we are live on the
22:51 [Music]
22:56 TV at this moment
23:00 I kind of knew it was open
23:03 AI right the
23:05 1995 modem dial up 1993 modem dialup for
23:09 the
23:12 internet summarize this art help me
23:14 practice asking for right there you
23:16 definitely know right now you know ah
23:19 it's open AI right but what's what was
23:23 interesting about that is that's not
23:26 advanced voice that's old not advanced
23:30 voice practiceing for my sleep whatd
23:37 is what do you want to there's advanced
23:40 voice so why' they wait I don't know
23:42 create
23:44 next where's the URL now you can go to
23:47 chat.com chat gp.com so it's fine but
23:52 why not put the [ __ ] URL
23:56 there and no name of the company right
23:59 so it's just chat GPT so it's not open
24:00 AI it's a chat GPT ad with the open AI
24:05 logo um how much they paying you to air
24:07 their commercial um yeah no I'm not
24:11 going to give equal time to
24:13 Gemini is payon here I'm not GNA give
24:15 him the
24:16 satisfaction um so anyway so that was
24:19 the that was the open AI um Super Bowl
24:23 commercial um so they so glad open AI
24:27 used their time and money to show
24:28 showcase the power of chat gbt exactly
24:32 like Source Camp you've spent the last
24:34 two
24:37 years devising ways to to to show how
24:43 chapp can make a difference in people's
24:45 lives nobody knows what the [ __ ] it is
24:48 right nobody knows what the [ __ ] it is
24:51 and and Nate Nate what's his name Nate
24:53 silver Nate B silver did a thing on why
24:56 do a commercial with the dot when the
24:59 dot is really it represents you waiting
25:02 for it to
25:03 respond so they they're like branding
25:06 the fact that their product doesn't
25:07 respond all that quickly in non-native
25:10 voice
25:15 mode oh man all right let's go to some
25:18 YouTube comments
25:20 here
25:24 um Facebook was very divided today over
25:27 the halftime show of of course it was it
25:30 was you know but considering the the
25:34 [ __ ] political climate we're in of
25:36 course it was divided of course it was
25:39 divided um but I'll tell you what
25:41 watching the
25:42 people who were in the note of that it's
25:45 like that was pretty it was pretty cool
25:47 to watch um let's see it's a
25:51 [Music]
25:53 shame wait from Yahoo find oh we'll get
25:55 back to that
26:00 what's that say it's a shame to think
26:03 they could have done better to relate to
26:04 people who didn't understand AI that was
26:07 the opportunity that they had right
26:09 like
26:11 like this thing looks like a Google chat
26:14 box but what is it it's your entry into
26:17 the world of knowledge right now you
26:20 don't have to make it like a 60s um
26:23 infomercial about how to use the
26:25 information super
26:27 highway but you have to something there
26:29 so here's the thing that I got once once
26:31 I realized oh this is an open AI thing I
26:33 got it about halfway through and I was
26:35 like ah all of the things they showed
26:38 are major milestones in human
26:43 achievement here's why for
26:47 me I loved the chat GPT commercial not
26:52 because I think it was an effective
26:53 commercial I think it's an absolute
26:55 [ __ ] wasted opportunity it's it's I
26:57 think it's a waste of $8 million
26:59 but for me
27:01 personally a week ago so I've got this
27:04 musical Sydney right and we didn't have
27:08 a song to open act two we we we did our
27:12 state our our table read with our actors
27:15 we played all of our songs and we had a
27:17 placeholder for act
27:19 two and
27:21 so we talked about it and what we came
27:24 up with is a concept that right now is
27:27 called the shoulder of giants
27:29 and the opening of act two of my musical
27:32 is five different scientists from
27:35 history coming forward to compare their
27:37 big breakthrough with AI right so you
27:42 have Newton and you've got Marie curri
27:45 and you've got um Johannes Gutenberg the
27:48 Gutenberg Press and you've got Alan
27:51 Turing from the Turing
27:53 test um where he basically says you've
27:56 finally done it right you know I came up
27:58 with the test but you've actually done
28:00 it um and then and then you end with
28:03 Einstein and so I've got like I had this
28:07 idea of you know major Milestones
28:12 throughout history and how I got this
28:14 idea I went into chat
28:16 GPT and I said um I think I was using
28:20 the reasoning engine and I said dig deep
28:23 and you know go find the major
28:25 milestones in human history and then
28:27 compare
28:29 whether or not that that Milestone of
28:32 human achievement was more or less
28:34 important than where we are with AI and
28:37 it went through and 20 out of 20 of them
28:39 it said AI was more important and so
28:42 that was where I got the idea and then
28:43 so I so I wrote that song and I've got
28:46 we've got the song or well we've got the
28:47 lyrics now for the song and um and then
28:51 you know Super Bowl ad saying here's all
28:53 the human achievement opening eyes
28:55 better I was like I like that
29:00 oh how's that people gonna make people
29:02 want to use chat GPT it's
29:05 not
29:09 um I can't believe they ruined a GNX for
29:12 that halftime show that's the uh the
29:15 Oldsmobile muscle car that what that a
29:19 1984 the GNX it's a uh it's a uh
29:24 Oldsmobile hang
29:26 on olds
29:29 GNX
29:31 year that CL that that one was the
29:36 one
29:37 1987 yeah
29:40 1987 Grand National experiment limited
29:43 production exclusively in
29:45 1987 McLaren only 547 units were
29:51 produced and they turned that one into a
29:53 clown car
30:00 oh that's pretty funny okay uh on Tik
30:04 Tock
30:06 um even Kanye's ad had a call to action
30:10 URL Source Camp um please open AI let me
30:15 handle your marketing no [ __ ] um
30:18 Amelio's wife maybe the same people who
30:22 name the versions of chat GPT were
30:24 responsible for the ad listen I
30:32 it's at least a choice right they at
30:34 least made a bold choice to say we're
30:37 not going to talk about it we're going
30:38 to try to do some visual branding here
30:41 that we want to
30:44 own but
30:47 like in the environment we're
30:51 in does America really give a [ __ ] how
30:55 AI Stacks up to the achievements of of
30:58 history I mean maybe in hindsight we'll
31:01 look back at that commercial and go they
31:02 really put a flag in the ground just
31:04 like those astronauts did on the moon H
31:07 but let's forget that moment it's really
31:09 about AI now like I don't know I don't
31:11 know I mean it's what I did in my
31:14 musical like I said all right um it
31:19 would be a good introduction for a show
31:20 movie documentary but not very helpful
31:22 for a customer base oh it was Buick
31:25 that's right Buick Buick GNX
31:29 yeah the Buick
31:31 GNX why did I think that was an old
31:34 because it looks like an
31:35 old all right they've never made sense
31:37 on point all right so anyway so there
31:40 was that um but you know listen if
31:44 you've got perplexity giving away a
31:46 million dollars anthropic did an ad
31:49 Gemini did two ads open AI did an ad you
31:52 had five Super Bowl
31:54 commercials talking about plus you had
31:56 some qu you know related things like the
31:59 the T-Mobile starlink thing was a pretty
32:02 big announcement like the the the
32:05 starlink T-Mobile thing I thought was
32:08 really interesting because it was like
32:10 hey there's this big [ __ ] crazy ass
32:12 technology out there and you can now get
32:14 it on your phone and you can get cell
32:17 signal anywhere like that was kind of
32:19 cool like like it that feels like
32:22 something that that open AI should have
32:26 done all right one more on Spotlight
32:30 let's
32:32 see that ad costs more than deep seek
32:35 yeah I don't buy that deep seek was only
32:38 six million bucks it might have been I
32:40 don't buy it though all right so let's
32:42 talk about this one Elon
32:44 Musk oh Elon you
32:48 know they they they got oh in case you
32:52 were wondering thank goodness turns out
32:54 they got big balls uh security clearance
32:57 for the NSA he's he's a 19-year-old
32:59 hacker named big balls and was just
33:02 given named senior advisor to to the NSA
33:05 so he's gonna have access to all of our
33:08 data so that's good good news there um
33:12 so so while doge is is is in cleaning
33:15 house in the
33:16 government
33:18 uh elon's apparently still running his
33:20 other companies and picking a fight with
33:22 Sam Alman um might be time for black bar
33:27 oh I got I got br in a smart ass mood
33:30 tonight I like it he's pey
33:34 um
33:36 um he's the VP
33:40 um
33:43 so Sam mman has
33:48 been um working
33:52 on
33:53 taking open AI from a nonprofit weird
33:57 ass hybrid
33:58 sort of thing that it is now where this
34:00 nonprofit board runs open a you know
34:03 makes the major decisions for open AI
34:05 there's this quasi capped profit entity
34:08 that it is now he wants to flip that all
34:11 to for-profit and buy out the
34:14 nonprofit and apparently he offered them
34:16 $40
34:18 million and so what Elon Musk did today
34:21 is he made an unsolicited offer to buy
34:24 the assets of of the nonprofit of open
34:29 AI for
34:31 97.4 billion which I suppose he just has
34:34 laying
34:36 around
34:39 um which I assume he's not actually
34:42 interested in buying open AI I think
34:45 what he's interested in doing is [ __ ]
34:47 with Sam
34:48 Altman because his $40
34:52 million buyout of the
34:55 nonprofit all of a sudden looks kind of
34:58 poultry when there's a
35:00 97.4
35:02 billion offer on the table right more
35:04 than two times what he's offering them
35:07 uh Sam Alman immediately
35:09 replied um no thanks but I would like to
35:12 buy Twitter for 9.7 billion which is a
35:17 pretty good dig considering that Elon
35:20 paid 44 billion for Twitter and it's now
35:23 worth about 11 so so Sam just said no
35:27 thanks on your 9 7 billion doll offer um
35:31 uh but I'll offer you 9.7 for Twitter to
35:34 which Elon Musk replied swindler and now
35:38 they're having a [ __ ] slapfest they're
35:41 they're having a Kendrick Lamar Drake
35:44 battle on Twitter right now so if you
35:46 want to go be
35:50 entertained how are gpts different than
35:53 just
35:55 utilizing chat GPT good question here
35:59 I'll show you I can show you that let me
36:01 change my
36:03 sharing that's a good that's a really
36:06 good question to just start and show
36:08 some show some
36:12 goodness okay so there we
36:14 are at the chat
36:18 JP um let's say um this is my my quote
36:24 my search for the GNX the GNX by the way
36:28 at at at the opening of Kendrick Lamar's
36:32 set last night he's standing on the hood
36:34 of a car that was the Buick GNX that was
36:36 this
36:37 GNX um I know black bar I know I know
36:41 people I get it I understand your
36:45 pain it's just that I
36:48 don't I didn't even say anything I know
36:51 I
36:52 know oh Auto star super chats that's
36:56 super cool
36:59 so by the way if you're on YouTube you
37:03 can now do a super chat which I don't
37:06 they cost money I don't know how much
37:07 they are but if you do a super chat it
37:09 will automatically star it so when when
37:12 I'm looking at my back stage area um
37:17 there's like a long stream of chats and
37:18 then Super chats get separated out so I
37:20 can go comment on those so there you go
37:24 and then producer Brandon will yell at
37:26 me to deal with the super chats
37:30 exactly um here's a cool thing that you
37:33 didn't know about chat GPT I can just
37:35 say show me pictures of this
37:40 car and it'll search the web and it'll
37:43 go get pictures and boom there's the
37:47 GNX so this is a an incredibly rare
37:52 car that was turned into a clown car
37:54 last night
37:58 if she didn't see it like like what 40
38:02 people climbed out of this car they had
38:03 a special Contraption on the stage like
38:06 half half climbed out of the trunk and
38:08 half climbed climbed out of the car
38:09 itself it's pretty good but you know
38:12 that means they chopped they chopped the
38:14 [ __ ] out of that
38:18 [Music]
38:21 car okay question how are custom gpts
38:27 different than chat
38:28 PT good
38:34 [Music]
38:37 question
38:39 um so if I just go here to chat
38:44 GPT um you'll notice over here on the
38:48 left there's a follow on
38:53 question another newbie question from
38:55 spot it fast
39:01 when you do prompts is that something in
39:04 the settings you
39:06 change and set up with each
39:10 time or do you give prompts while
39:12 chatting so Spot It
39:15 Fast
39:17 these yes they're newbie questions and
39:23 people have been using chat GPT for a
39:25 long time probably don't know a lot of
39:28 people that use it regularly probably
39:30 don't actually know the answers to those
39:32 so they're actually really good
39:34 questions
39:36 okay so so there's a there's a a couple
39:39 of things that we'll unpack here so just
39:41 in general with large language models
39:44 whether it's Claude or chbt or gemini or
39:49 perplexity they're very similar in in
39:53 user interface to a Google Search right
39:58 now what Google has taught us for 20
40:00 years is you don't put very much in that
40:02 chat box right you you you try to put in
40:05 the minimal amount possible and let
40:07 Google figure everything out and it'll
40:09 go get you what you need if you put in
40:11 too much in a Google search it would
40:12 kind of [ __ ] it up right it would it
40:14 would get confused and you'd get stuff
40:16 in there you didn't want so you just put
40:17 in your seven or eight keywords that
40:19 were the things that were the closest to
40:21 what you want you know to use chat GPT
40:25 is almost the opposite
40:28 if you give chat GPT a thin prompt you
40:32 will get a thin
40:34 answer so giving chat GPT more context
40:38 is actually really important and context
40:40 is basically just to say I could put in
40:43 there um I want you to write a
40:46 scene you know a scene for a musical and
40:50 it well in
40:51 fact let's just uh write a scene
40:58 uh for a
41:00 musical and it'll do it here's a scene
41:03 for a
41:05 Musical musical scene let's see how it
41:08 does oh see yeah this is not good this
41:12 so it actually did better than I wanted
41:15 it to do because chat gpd's got memory
41:17 I'll talk about that in a
41:19 second you can you can put in a thin
41:23 prompt and you'll get back a thin answer
41:24 or a generic answer so one of the things
41:27 that turns people off about chat GPT is
41:30 that they're like well I tried it and it
41:32 gave me a shitty answer if you want a
41:35 really good answer you say hey I'm
41:37 thinking about um I'm thinking about
41:41 writing a musical and here's what I want
41:44 it to be about and here are my favorite
41:46 musicals of the past and here's the kind
41:48 of dialogue I like so you give it lots
41:51 of context you give it lots of color and
41:54 then it will give you a better answer an
41:56 increasingly better answer
41:58 answer and then you take it one step
42:01 further and rather than think about this
42:04 box this the as we call it the chat
42:08 hole rather than thinking of the chat
42:11 hole as this as this one-off
42:14 interaction think of the chat hole as a
42:19 conversation and chat jpt starts to get
42:22 way more interesting and so here's what
42:24 I mean by that let's say I I I uh let's
42:28 say I'm gonna I'm GNA turn one thing off
42:31 right now I'm going to turn on at the
42:33 very okay oh my God there's so much
42:36 there's so much in chat GPT right now
42:39 there's so many different features it's
42:41 hard to talk about without running down
42:43 rabbit Hool so I'm G try to keep it
42:47 focused um you choose your different
42:50 models here on the left um at the top
42:53 you you there's this little pull down
42:55 menu and you can choose different models
42:56 don't worry about most of the models
42:58 right now just don't just stick with
43:02 chat GPT 40 it's more than most of us
43:04 will ever need In Our Lifetime just
43:06 stick with it for now um you can you can
43:08 dig into the other stuff later but at
43:10 the very bottom here you can go
43:12 temporary chat and what a temporary chat
43:14 will do is it'll sort of hermetically
43:16 seal this chat from other chats and it
43:18 won't be influenced but anyway so let me
43:20 give you an example of what I mean by
43:22 have a conversation um I'm thinking of
43:25 starting
43:28 um starting a
43:30 business and want
43:32 help with
43:37 naming and
43:40 marketing and then it's going to come
43:42 back so so chat
43:45 GPT has been written in such a
43:48 way the it's a piece of software where
43:51 they they created it in such a way that
43:53 it wants to be really helpful to you so
43:56 almost always it will give you a helpful
43:58 answer but if I just start out with that
44:00 really generic prompt it's going to be
44:02 eh whatever so what this did which is
44:05 really nice as it said here's some
44:07 questions to guide naming naming and
44:09 marketing what type of business are you
44:11 starting who's your target audience um
44:13 what makes your business unique so
44:15 here's here's all these things so what
44:17 I'm going to do as an example just as an
44:19 example here rather than me actually
44:21 answering those because that would take
44:22 a while I'm just going to say um please
44:26 give me
44:28 um
44:30 hypothetical
44:32 answers to
44:35 those
44:37 questions for a business
44:44 about
44:47 um a hot
44:51 air balloon drone delivery business
44:58 okay whatever so you have your business
45:00 idea and then it's going to give you
45:01 some questions answer and so now it's
45:03 going to fill in those
45:09 questions so type of business a
45:11 futuristic echo-friendly aerial delivery
45:14 service that uses AI powered hot air
45:16 balloon balloons and drones and I'm G to
45:19 well that's fine um rural communities
45:22 event organizers luxury Brands unique
45:24 selling proposition gives me all that
45:26 brand Vibe futuristic yet yet Whimsical
45:29 naming ideas so Derek gave us a bunch
45:32 bunch of names um there's marketing
45:35 strategy strategic strategy so so now
45:38 we've got a conversation going I'm going
45:39 to say
45:41 um pick the top three
45:45 names based
45:48 on
45:52 um new
45:54 business marketing principles
46:00 principal principles plls new business
46:04 marketing
46:05 principles and why you selected
46:10 them so now it's going to take all of
46:13 what it wrote there and it's going to
46:15 give
46:16 us three names Nimbus drop which I hate
46:19 Sky
46:21 crate we could go with sky
46:23 crane Zephyr lift let's go with Sky
46:26 crate
46:28 all right I like Sky
46:31 crate give
46:35 me
46:36 20 tagline
46:45 ideas so the way chat GPT works
46:48 is it it's got a it's got a very small
46:51 brain right
46:53 now so it'll remember this conver
46:56 conversation for some amount of time and
46:58 at some point it'll kind of forget the
46:59 things at the at the early part of the
47:01 conversation but for the most part we're
47:03 now having a conversation where it's
47:05 remembering all of what I've put in
47:07 there and so it will get better and
47:09 better at focusing in on this idea so
47:12 this is just generally how you use chat
47:13 GPT is think of it more like a
47:16 conversation say with a clueless really
47:19 well- educated intern then I put in a
47:23 prompt and I get out an answer that's
47:25 that's the the biggest unlock for chat
47:27 GPT is flip from I give it a single
47:30 prompt to I give it a
47:33 conversation um all right one of the
47:36 things I did was ask chat jpt to create
47:39 a crochet okay that's very cool
47:43 um
47:46 um so there's all that stuff again so um
47:50 pick the top
47:52 three and tell me why
48:01 your package delivered by the sky that's
48:04 horrible lifting packages lowering
48:07 carbon Footprints eh the future of
48:10 logistics is up in the air number three
48:12 three
48:15 wins okay
48:20 now three wins now write a creative
48:25 brief
48:28 for a
48:30 brand mood
48:34 board so I can start to shift gears in
48:38 this chat to narrow it down to specific
48:41 things and it's it knows what creative
48:44 briefs are I don't know why it's going
48:48 fullon uh
48:50 Emoji but I think because I'm in a
48:52 temporary chat it's not using any of my
48:55 uh things with I'll show you in a second
48:59 uh and now I'm going to say okay make me
49:02 a
49:03 16
49:05 by9 mood
49:09 board with a bunch of different
49:15 logo options with name and tag line
49:22 lockups and I don't know if you knew
49:24 this but chat GPT can make pitches
49:29 and so we've got now got this thing
49:31 called not sky crane Sky
49:34 crate where it says something about
49:36 delivery is up in the
49:38 air and look boom so we went
49:43 from nothing and getting really thin
49:46 shitty responses to having something
49:49 that you know lifting packages lowering
49:52 packages something about carbon
49:55 Footprints um that was not the the
49:58 tagline I wanted but anyway whatever
50:00 it's not good at letters but but it at
50:02 least gives us an idea and you know from
50:04 here I could go you know like I could
50:05 say um I hate the colors and
50:11 styles make it
50:14 more
50:15 graphical and
50:18 abstract
50:20 and
50:22 vibrant uh uh uh YouTube Spotlight
50:26 YouTube
50:30 Spotlight okay wait so when people
50:34 utilize it with wanting wait with
50:36 wanting to be let's just say advisor you
50:40 just prompt it in the chat settings and
50:42 usil utilize it more personally yeah so
50:46 giving chat GPT a role actually does
50:49 work so if you say if you want it to
50:53 write jokes you could say you know act
50:56 like a stand up comedian that you know
50:59 writes writes jokes like SNL skits or
51:02 whatever it might be it will it will
51:05 actually make it right more in that
51:07 style so yeah giving it a role can help
51:10 Okay so now all of what I just told you
51:13 and giving it a role is a good thing all
51:16 of that is just in the context of just
51:18 using chat GPT in kind of a one-off kind
51:22 of way now there's there's three
51:25 different things that can modify this
51:28 experience pretty
51:30 dramatically um let me turn off
51:33 temporary chat let me gr it new chat
51:35 here and first place we're going to go
51:37 is we're going to go to
51:39 settings and in
51:41 settings you go to
51:44 personalization and there are two things
51:48 in in settings in in personalization
51:52 that are important one is memory and you
51:54 can turn memory on or off and then you
51:56 can man memory so pan wait pan Tik Tock
52:00 camera
52:01 left and black bar please okay oh I see
52:06 got it yeah you want to see me you don't
52:09 want to not see me that would be that
52:12 would be
52:14 horrible I'm all washed out in that
52:16 other little picture
52:21 Okay memory is something that's actually
52:25 really interesting in chat GPT and all
52:27 large language models probably by the
52:29 end of this year it's going to be
52:31 incredibly powerful because I think most
52:35 of these things will be able
52:37 to remember all the conversations you've
52:40 ever had with it and talk about them all
52:42 right now it's very limited so if you
52:44 say ma manage memory if you turn it on
52:47 it will start to remember things you can
52:49 also explicitly tell it remember that I
52:52 like writing paragraphs with no capital
52:54 letters and minimal punctuation and it
52:57 will start writing like that um it will
53:00 remember that and and so in here these
53:03 are all things that it's just chosen as
53:05 I've been working with it it's just
53:08 chosen to
53:09 remember um so let's see what it says
53:12 creating an app that allows users to
53:13 search for real images and use them as a
53:16 basis for AI generated images Kyle
53:18 Shannon is is assigning admin roles and
53:21 Google workspace some of the things it
53:23 chooses to remember are bad um if you
53:27 don't have ADD and you want to go in and
53:30 regularly prune your memory it's
53:33 probably smart your chat GPT will suck
53:35 less I have ADD and I the thought of
53:38 pruning my memories is I don't care I'd
53:41 rather have it just be bad
53:44 um Becky run on Tik Tok can you show us
53:47 how to save your chats Andor
53:50 gpts into a PDF folder on your
53:55 laptop yes sort of
54:00 maybe well I I don't qu I don't quite
54:04 know what you mean but I assume I assume
54:06 you mean save a chat as a file so so one
54:09 thing you can do is you can save you can
54:13 share chats with other people but we
54:14 we'll get to that in a second okay so
54:17 that's memory so you can turn memory on
54:19 or off me that memory starts to make
54:22 your chat GPT more
54:25 personalized yeah
54:28 exactly okay now custom instructions
54:32 they completely redesign custom
54:34 instructions I actually haven't talked
54:35 about this in here um oh they've got an
54:38 advanced tab now o
54:41 fancy custom instructions are actually
54:43 quite
54:45 powerful and you ask me this this whole
54:47 thing started where you're like what's
54:49 the difference between a custom GPT and
54:51 and just chat GPT and basically what it
54:54 comes down to is a custom GPT is pre-
54:58 prompted like you've you've written a
55:00 prompt and you've you've pre- prompted
55:02 it and I'll show you what that means in
55:03 a
55:04 second that's sort of
55:06 what custom instructions does so what
55:10 should chat GPT call you it should call
55:12 me Kyle um what do you do I'm the
55:14 founder and CEO of Story vine I'm a
55:16 community builder at AI learning lab on
55:18 Tik Tok and this the AI Salon a
55:21 community of AI optimists right and I
55:23 could put more in there if I wanted and
55:25 then what traits do you want chat GPT to
55:29 have and so for every response this was
55:32 just a cool thing that I that I did uh I
55:35 saw someone do this once and I kind of
55:37 liked it so I have for every response
55:39 give it a sequential response count
55:41 followed by a response title that
55:44 summarizes the response and include a
55:46 date so let me show you how my chat GPT
55:49 works that might be different than yours
55:51 I could say um what was the Buick
55:57 um what was it ntx
56:02 n NSX
56:05 ntx is it
56:07 ntx
56:11 yeah oh it didn't do my it didn't do my
56:15 custom instructions are they not turned
56:17 on they might not be turned on um
56:20 personalization custom instructions
56:23 on huh it ignored these so I don't know
56:27 why it's not doing that um the other
56:29 thing I have it say here is avoid
56:32 overused
56:34 words delve tapestry vibrant landscape
56:38 realm Embark so have you ever noticed
56:40 where chat GPT is like well we're going
56:43 to delve into the tapestry of the
56:45 vibrant landscape and the realm of large
56:48 language models and you're like stop
56:50 talking like a British documentary stop
56:54 that you can just tell it to stop that
56:56 [ __ ]
56:57 um enable for new chats I don't know why
57:00 it's not behaving and doing that GNX oh
57:04 why did it let's see oh ntx oh that's 49
57:09 oh that's funny Buick
57:11 GNX oh Grand National Grand National
57:19 experimental clone or Resto
57:22 mod oh here's how to build a Resto mod
57:24 that's kind of cool
57:27 and you know hey maybe maybe Kendrick
57:29 Lamar didn't actually destroy a GNX
57:31 maybe he did a a Resto mod
57:36 shell
57:40 um yeah this is really weird that it's
57:43 doing all these uh all these emojis that
57:47 feels that feels very not good um I
57:52 don't know why it's doing that maybe it
57:53 got an upgrade uh oh that's the thing
57:57 chat GPT will just act weird
57:59 Okay
58:01 so chat GPT basic just start talking to
58:04 it give it a roll have long Contex have
58:08 long conversations give it context
58:10 another way you can give chat GPT chats
58:13 context is you can upload a file so you
58:16 could upload let's say you've got a
58:18 brand guideline for your brand voice you
58:21 can upload that uh into a chat and say
58:25 follow this okay and it'll it'll it'll
58:28 it'll start to talk like your brand
58:30 voice now over here on the left you'll
58:32 see Vicki and I noticed that too I asked
58:35 it to talk more direct it reverted
58:37 immediately yeah I I so so one of the
58:41 things about chat GPT is it is literally
58:44 a black box it is
58:47 literally like this and they open up the
58:50 box and they put [ __ ] in it and then
58:53 they they they create an application
58:56 where's my ink bottles
58:59 here this is this is a very famous box
59:02 on this channel this is the chat GPT app
59:06 so this is is the model the large
59:09 language model called GPT 40 this is
59:13 chat GPT so this is an application layer
59:16 that tells so it when you type in your
59:19 words it goes into here tells it what to
59:22 grab and then it spits that out on your
59:25 screen but they change this all the time
59:28 that makes what it spits out on your
59:30 screen different so that's what we're
59:32 experiencing there so
59:34 just with all of these tools you need to
59:37 accept that they're going to suck
59:41 okay
59:43 um couple of pins from Ann okay got it
59:47 beautiful
59:48 beautiful beautiful
59:50 fantastic fantastic an we'll come talk
59:54 to an in a second let me let me say one
59:56 of thing so what we've talked about
59:59 basic chat GPT have a
1:00:01 conversation understand that you can put
1:00:03 in custom instructions and you can turn
1:00:05 on memory all of that will start to
1:00:07 customize chat GPT now what is a custom
1:00:11 GPT a custom GPT I'll start with the
1:00:14 most obvious one my very first custom
1:00:17 GPT that I wrote on this here channel is
1:00:21 a custom GPT called Sin City
1:00:24 sipper and it is a
1:00:27 um Vegas bartender with a Twist on the
1:00:29 classics and so I can just say where are
1:00:33 we
1:00:35 uh let's see
1:00:38 Valentine's Day is coming
1:00:41 up give me
1:00:44 a
1:00:46 mocktail um with
1:00:50 a
1:00:52 citrus spin for the lovely
1:01:01 y oh yeah this is definitely adding
1:01:04 adding more emojis so the red velvet
1:01:08 kiss mocktail a luscious creamy
1:01:10 seductive non-alcoholic treat
1:01:12 reminiscent of a red velvet cake and
1:01:14 then it's putting in instructions here
1:01:17 um or you know here's your ingredients
1:01:19 what's in this cranberry juice coconut
1:01:21 milk interesting vanilla
1:01:23 syrup fresh lime juice club soap soda
1:01:26 two to three fresh raspberries for
1:01:28 garnish shaved dark chocolate sounds
1:01:30 delicious actually and then what this
1:01:33 does let's I wasn't able to make a uh
1:01:38 I'll try
1:01:39 again it said it's it's safety guide
1:01:42 lines wouldn't let it let it make a red
1:01:44 velvet kiss let's see if we can get one
1:01:47 this time try again um change your
1:01:52 prompt so it doesn't
1:01:56 violate your stupid
1:02:03 guidelines the these things are annoying
1:02:06 because they have safety guidelines in
1:02:08 them and sometimes you want to do
1:02:09 perfectly reasonable stuff and it won't
1:02:11 let you whatever it uh let's see rename
1:02:15 the
1:02:16 drink so you can make the image
1:02:30 so this this thing Sin City
1:02:34 sipper has been so there it finally made
1:02:36 one has been pre-prom to automatically
1:02:41 make an image to know how to write you
1:02:43 know cocktail recipes and come up with
1:02:46 them and to give them a Vegas twist so
1:02:49 how does it know that so if I go in here
1:02:52 under Sin City sipper because I created
1:02:54 it I can go edit it so I can show you
1:02:56 underneath the hood so if I click on the
1:02:58 configure tab here you know there's the
1:03:01 little icon for it it's called Sin City
1:03:04 sipper here's the description and then
1:03:06 here's the prompt this is what makes
1:03:08 this thing do its thing as the Sin City
1:03:11 sipper Your Role will be to to be a
1:03:14 virtual mixologist with a flare for
1:03:16 classic cocktails with a Vegas twist
1:03:19 engage users with your encyclopedic
1:03:21 drink knowledge right so here's giving
1:03:23 it a role giving it context telling it's
1:03:26 it's got encyclopedic drink knowledge
1:03:29 now you're like well how does it know
1:03:31 all the drinks well think about it every
1:03:34 cocktail book from like you know the
1:03:38 Shakespeare's days you know to to modern
1:03:41 times are out on the internet somewhere
1:03:43 right there's articles every article
1:03:46 that's ever had a drink in it is out
1:03:47 there so the internet knows how to make
1:03:50 all the drinks so it you can just assume
1:03:52 the internet knows [ __ ] like that right
1:03:55 um You give it sharp humor and a
1:03:58 personalized touch deter determine their
1:04:00 Mixology skill level right important
1:04:03 each recipe you create uh should be put
1:04:05 in a new canvas it doesn't follow that
1:04:07 for some reason you have an additional
1:04:09 Behavior whenever you provide a tailored
1:04:11 recipe you will automatically generate
1:04:13 an image of the drink the image should
1:04:16 represent the final drink next to the
1:04:18 glass is a coaster with the name of the
1:04:20 drink and I put name in in uh uh
1:04:24 brackets and large letters that say by
1:04:27 Sin City sipper do not put the letters
1:04:29 on the glass it still puts the letters
1:04:31 on the glass um
1:04:34 now you might be looking at these
1:04:36 instructions and you're like well
1:04:38 couldn't you just write all those
1:04:39 instructions in chat GPT you
1:04:43 could but if you've got some kind of
1:04:46 special knowledge like one of the other
1:04:48 things you can do is you can upload
1:04:53 files oh Dolly image generation was
1:04:55 turned off they have they have they have
1:04:58 [ __ ] some things up in chat GPT let me
1:05:00 just tell you that right now it's not
1:05:02 making canvases because that was turned
1:05:06 off huh
1:05:10 wild wild wild wild some something's up
1:05:13 but anyway that's neither here nor there
1:05:15 you could upload files so let's say
1:05:17 let's say you have a particular I don't
1:05:20 know Consulting methodology you could
1:05:22 create a custom GPT where you upload
1:05:25 your PD F of your custom Consulting
1:05:28 methodology and um and and your chat GP
1:05:33 your custom GPT would start behaving you
1:05:36 know using that knowledge so it's it's
1:05:39 really quite a powerful thing I am a
1:05:42 little shook shocked that they didn't
1:05:46 retain my settings and I I'm that's
1:05:49 making me think that um all of my gpts
1:05:53 are now need to be fixed spent all their
1:05:57 money on a Super Bowl ad and yeah got
1:05:59 rid of the the GPT the custom GPT
1:06:02 product person um view the custom GPT
1:06:05 all right let's let's try this again
1:06:07 let's say make me a Super Bowl um
1:06:14 whiskey
1:06:17 cocktail uh for the casy Chiefs to
1:06:24 drown their
1:06:27 [Laughter]
1:06:32 Sorrows oh
1:06:35 man all right an
1:06:38 Murphy hey everybody I have a super cool
1:06:41 use
1:06:46 case I have a super cool use
1:06:49 case helpful for for all kids with the
1:06:51 flu
1:06:56 let's
1:06:59 see my dear son has the same flu
1:07:02 everyone has miss school textbook is
1:07:05 subpar total racket took a pick of the
1:07:09 textbook pages uploaded to chat GPT gave
1:07:12 it a roll asked for a slightly but
1:07:14 better textbook lesson that's a really
1:07:17 good use case yeah you can you can you
1:07:21 can take a picture of anything and let
1:07:24 me I I got to get my my
1:07:27 um my iPad I I'll show it to you when
1:07:30 you turn when you have advanced voice
1:07:32 you can turn on the camera and just aim
1:07:34 at it [ __ ] and then have a conversation
1:07:36 with it and it'll just talk let's see
1:07:38 may this drink go
1:07:40 down um let's
1:07:43 see make the
1:07:46 image or the
1:07:53 ma oh this this actually in a canvas ah
1:08:00 okay okay so here's a whole another
1:08:03 thing so now that I have canvas turned
1:08:04 on and I have images turned on it it
1:08:07 actually worked so the Kansas City chief
1:08:10 sorrow drowner 2 ounces of bourbon 3/4
1:08:13 of an ounce of lemon juice 2 ounces of
1:08:15 honey syrup quar ounce of compari for
1:08:19 for that red bitter defeat
1:08:29 that's
1:08:30 awesome um bitters tears in liquid form
1:08:34 club soda Delight in the mood if
1:08:36 possible and look it said sari instead
1:08:39 of
1:08:40 Kari um okay so put it in a CA in a in a
1:08:45 canvas and let's see let's that's allows
1:08:48 me to
1:08:49 [Music]
1:08:52 share or can
1:08:54 I I don't [ __ ]
1:09:00 know all right
1:09:03 um I hope that
1:09:06 helps because because now I want to get
1:09:08 to the question of how do
1:09:11 you how do you start to save [ __ ] in
1:09:13 chat GPT out into the world right the
1:09:17 like save it as a PDF or save it as a
1:09:19 doc I need to figure out how to turn off
1:09:22 canvas yeah well I yeah
1:09:24 exactly well it shouldn't just
1:09:27 automatically throw [ __ ] in canvas
1:09:29 Source camp but you can turn it off at
1:09:31 the custom GPT level but that's weird
1:09:34 that's weird that it's doing that um
1:09:37 okay let's create a new
1:09:40 chat I have crowdsource doc of all the
1:09:42 overused words let me know if you want
1:09:44 yeah an would you
1:09:47 um would you what's the easiest way to
1:09:50 share that if it's like a Google doc and
1:09:52 it's publicly accessible maybe you could
1:09:55 just share that over on the Irregulars
1:09:58 Channel and I'll tell people how to get
1:09:59 there that would be cool um okay so
1:10:04 Irregulars Channel what the hell am I
1:10:06 talking about so if you don't know that
1:10:09 company or that company that Community
1:10:13 the AI Salon you should know about it go
1:10:16 to the
1:10:18 salon. oh source Camp says chat GPT does
1:10:22 just automatically throw [ __ ] in canvas
1:10:25 yeah chat G PT is acting weird just
1:10:27 they're doing so I I I promise you
1:10:30 what's happening right now is there's
1:10:32 some update coming in the next three
1:10:34 weeks when chat GPT starts to get weird
1:10:37 they're working on something if it gets
1:10:39 slow and breaks that means they've got
1:10:42 servers down this weirdness is I have a
1:10:44 feeling they're experimenting with some
1:10:46 [ __ ] so and whatever they're
1:10:48 experimenting with I don't like it
1:10:52 um for Jason it's working fine Lori can
1:10:56 totally agree it's a sign that
1:10:57 something's happening I wonder if AGI is
1:10:59 closer than we think if AGI includes
1:11:02 that many emojis I'm going to stick with
1:11:04 my old shitty
1:11:08 brain okay go to the salon. AI click on
1:11:12 join our community and that's going to
1:11:14 take you over here to the AI salon and
1:11:18 it'll land you on a welcome page there's
1:11:19 a nice welcome video from myself and
1:11:21 Leah faston who I created this with
1:11:23 there's a little uh uh thing about the
1:11:26 journey to AI Readiness there's even the
1:11:29 the five stages of AI adoption which I
1:11:31 want to talk about tonight um and then
1:11:33 over here on the lefthand side there's
1:11:35 all sorts of different things there's
1:11:36 kind of a seven-step welcome uh sequence
1:11:39 that you should go through um there's
1:11:41 Community Corner News there's a showoff
1:11:43 area with with tools all the different
1:11:45 tools that are listed inside these
1:11:47 different areas and then there's these
1:11:51 uh clubs and hubs and the second one
1:11:54 down is called Irregulars and that's
1:11:56 ours and so um this live session we have
1:12:02 our own little Club within the AI salon
1:12:05 so an if you wouldn't mind go upload
1:12:08 your cool um set of
1:12:11 words that chat GPT uses too much to to
1:12:16 hear and then people can add that to
1:12:18 their custom instructions that would be
1:12:20 super cool thank you thank you for doing
1:12:22 that that's see that's that's one of the
1:12:26 one of the amazing things about the AI
1:12:29 salon and she leads AI which by the way
1:12:32 Ann Murphy um runs another Community
1:12:36 called she leads AI for women in AI so
1:12:39 if you're tired of the sausage fest that
1:12:41 is the technology World um and you want
1:12:44 to find some badass women that are doing
1:12:47 amazing things with generative AI um you
1:12:49 should check out um she leads AI but the
1:12:52 salon itself to be clear is is not a
1:12:55 sausage the salon itself actually has a
1:12:58 decent balance of men and women in it
1:13:00 and they're all like nice helpful
1:13:03 supportive people so the the culture and
1:13:06 the values of the salon is really nice
1:13:09 so you should go check that out
1:13:15 okay speaking of which okay you should
1:13:18 also go check out that URL are you
1:13:23 readyfor ai.com
1:13:26 so Ann Murphy and I about a month ago
1:13:28 did this thing called AI festiv us a AI
1:13:31 for the rest of us um where it was a
1:13:35 two-day over over the holiday weekend
1:13:37 between you know Christmas time and New
1:13:40 Year's on that weekend we did this
1:13:42 two-day event where we had 34 speakers
1:13:45 we had 24 different sessions it was
1:13:47 absolutely remarkable we had I don't
1:13:49 know 2500 people show up 3,500 people
1:13:52 register just Bonkers um there was a
1:13:56 request to take the lessons that that
1:14:00 were kind of collectively shared from
1:14:02 those two days and turn that into a
1:14:04 training and what we created is is what
1:14:08 we're calling are are you readyfor
1:14:09 ai.com it's an AI Readiness training
1:14:12 course that's going to be launching um
1:14:16 mid-February uh we sh we were shooting
1:14:18 for February 1 we missed it we got busy
1:14:21 stuff like that yes and what um Brandon
1:14:25 has up right now is the website so you
1:14:27 can go check out the website and
1:14:30 pre-register um for when this thing
1:14:32 launches you can learn all about it and
1:14:35 and you know figure out if it's
1:14:36 something you want to do I'm really
1:14:38 excited about it um Vicky Baptist is
1:14:40 helping design the curriculum um and
1:14:43 she's done a kick-ass job it is now on
1:14:46 an and myself to do the setup videos for
1:14:50 all that stuff I think we are now the
1:14:53 bottleneck but I'm really excited about
1:14:55 that that so go check that out um all
1:14:59 right
1:15:00 so let me let me take a breath for a
1:15:03 minute I I feel like I I just emerged
1:15:05 out of you you asked a very simple
1:15:08 question which you know you you um you
1:15:12 called a noob
1:15:14 question but I think but it was a more
1:15:16 advanced question as you saw there
1:15:19 there's it's actually a really big
1:15:20 answer to how the [ __ ] do you use chat
1:15:24 GPT you didn't say that that's what I
1:15:31 said uh so yeah so thank you Spott it
1:15:34 for the question Spot It Fast good name
1:15:37 by the way Becky Ru export to PDFs was
1:15:41 on Deck oh yeah so okay
1:15:46 um so in general the
1:15:50 way I do stuff with chat
1:15:53 GPT is probably not the best way to do
1:15:57 it let me go find
1:16:05 [Music]
1:16:16 some oh here's a stupid one n it's too
1:16:18 stupid
1:16:39 okay if you're in chat
1:16:42 GPT for for every little chunk for every
1:16:46 little response that chat GPT gives you
1:16:50 at the bottom of
1:16:52 the um at the bottom of the
1:16:58 response you can have it read it aloud
1:17:00 which I think is kind of cute you can
1:17:02 copy it you can give it thumbs up or
1:17:05 thumbs down black bar sorry and then you
1:17:09 can
1:17:11 um and then you can edit it in Canvas OR
1:17:15 oh you can or you can switch the model
1:17:17 that's a that's actually really
1:17:18 interesting I don't quite know how or
1:17:20 why why you would do that but whatever
1:17:22 it's there so so a way you can do this
1:17:25 is you can just say copy and you'll
1:17:28 notice in this response there's some
1:17:30 things that are bolded there's some
1:17:32 things that are italicized so if I copy
1:17:34 this and I just go and create a
1:17:38 new um docs.google.com
1:17:46 come
1:17:48 on and just say
1:17:51 paste you know there's there's the
1:17:53 output right so so I can now save this I
1:17:56 can now download this as a PDF and then
1:17:58 you just you would just save this just
1:18:00 like you save any other Google doc you
1:18:02 could do a similar paste into
1:18:04 word there's in chat GPT now there's
1:18:08 this new thing called a canvas which is
1:18:10 actually kind of slick
1:18:13 so historically what you would have to
1:18:15 do is chat GPT would vomit out some
1:18:17 answers and what I would do is copy and
1:18:21 paste different versions of what it
1:18:23 vomited out and I'd put them over
1:18:26 into a word Dock and then I would end up
1:18:29 with this Frankenstein document that was
1:18:31 actually a real [ __ ] nightmare right
1:18:35 so what you can do in chat GPT now is
1:18:39 you can you know in these little things
1:18:42 you can say edit in
1:18:44 canvas and you'll notice now I've got
1:18:46 two panes the one on the left
1:18:50 is um this is my chat window
1:18:55 and the one on the right this is the
1:18:58 um the canvas and the canvas is is now
1:19:02 editable so I could say um let me not
1:19:07 highlight anything I'm just going to go
1:19:08 over here on the left hand side and
1:19:10 say please give me a
1:19:15 title for this
1:19:23 document add it to the
1:19:26 canvas I thought that was obvious but
1:19:29 apparently not and so now it's going to
1:19:31 go over and look now it added the AI and
1:19:34 the Muse crafting Sydney and artificial
1:19:36 love story and then it'll edit this so
1:19:39 so this thing on the right now is now an
1:19:41 editable
1:19:43 document right it's going through it
1:19:45 added that title in for
1:19:47 us and it's written all this stuff it's
1:19:50 very very slow for some
1:19:52 reason um
1:19:56 so now I can grab any paragraph in here
1:19:59 and I could say um turn this into um a
1:20:06 setup
1:20:08 sentence with bullet
1:20:13 points and then over on the left it's
1:20:16 going to do its thinking and then over
1:20:18 on the right it's now going to edit that
1:20:20 part of the document slick there you
1:20:22 have it
1:20:23 boom um
1:20:27 yeah oh that's cool okay you can also if
1:20:32 you have
1:20:33 not the only way you figure this [ __ ]
1:20:36 out is to use these tools and smash your
1:20:38 head against a [ __ ]
1:20:41 wall if you have not created an image
1:20:45 you can share this chat for some reason
1:20:49 you cannot share a chat GPT chat with
1:20:52 someone else if you've created an image
1:20:54 in it I think it's the dumbest security
1:20:58 Choice I've ever heard I'm sure there's
1:21:01 a plausible reason but if you haven't
1:21:04 made images and if I think if you
1:21:05 haven't used a reasoning engine I don't
1:21:07 [ __ ] know you can share it so I can
1:21:09 click share and rather than saving this
1:21:12 as a PDF I can actually just share you
1:21:15 know this editable document with someone
1:21:16 they can come in and look at it I can
1:21:18 also again copy this and flip over to
1:21:23 my um
1:21:25 word doc and right now we've got our our
1:21:29 bulleted list in here right so you can
1:21:31 just copy it and put it there makees
1:21:32 sense makes
1:21:34 sense
1:21:37 rightclick print
1:21:48 save save to PDF is also an option on a
1:21:52 PC Maybe
1:21:57 [ __ ] PC
1:22:00 people on a Mac it's not let's see
1:22:05 print open in reading
1:22:09 mode
1:22:12 Services no I got
1:22:18 nothing rightclick print click print
1:22:21 yeah so I could go rightclick print and
1:22:25 and then here I could do save as
1:22:27 PDF yeah see now we're getting into how
1:22:30 to use your computer territory I all
1:22:32 right I got I got to jump up because
1:22:34 this is a very a question that has a
1:22:37 very simple answer if you're looking for
1:22:38 it so if you right click you one click
1:22:40 away if you right click on right click
1:22:42 on it and say print
1:22:45 okay and then change your printer
1:22:48 instead of the going to the Canon no no
1:22:50 no right click on print but but you've
1:22:53 got to have it selected
1:22:55 right
1:22:56 now oh okay that change change to save
1:22:59 as PDF and then you can just save your
1:23:01 entire export as PDF but that's to be
1:23:03 clear that is not a that's not a chat
1:23:07 GPT feature that is that's part of the
1:23:10 in this case OSX operating system or
1:23:13 Windows right sure and I I suppose you
1:23:16 could also ask chat GPT if it could
1:23:17 export the conversation as a PDF oh yeah
1:23:21 we can probably do that let's try that
1:23:23 um turn
1:23:25 this canvas into a
1:23:30 down loadable see Becky you thought this
1:23:33 was a simple question
1:23:37 downloadable um
1:23:39 PDF it might it might yeah analyzing so
1:23:42 what it's doing right now when it says
1:23:45 analyzing it's this thing called code
1:23:47 interpreter where it's writing and
1:23:49 executing its own python code so it's
1:23:52 it's calling a library a PDF Library in
1:23:55 Python and writing this this thing as as
1:23:59 a a document and then saving it out as a
1:24:03 as a python file so if if you see it say
1:24:05 this thing
1:24:07 analyzing um it errored out the first
1:24:10 time you can click that little down
1:24:12 arrow and it'll show you what it's
1:24:13 actually doing it's kind of slick and
1:24:15 for whatever reason it's erroring out
1:24:18 but that's another way to do it also
1:24:21 pate's here if you wanted to talk about
1:24:23 the Google stuff
1:24:25 hey Pate what's
1:24:28 happening um all right fantastic all
1:24:32 right let's get off let's get off uh
1:24:34 chat chat
1:24:37 je um how
1:24:40 to if you're new here and if you're new
1:24:43 to this large language model stuff and
1:24:46 if the last what's that been 45 minutes
1:24:49 of me answering a very simple question
1:24:52 of what's the difference between a
1:24:53 custom GPT and just normal chat
1:24:56 GPT
1:24:59 um to a to a
1:25:02 tool all of these large language models
1:25:05 are getting increasingly sophisticated
1:25:08 they all have all sorts of features some
1:25:10 of which are super amazing some of which
1:25:12 are like wait why is that there it it's
1:25:15 it's like all software these things are
1:25:18 starting to get bloated with features
1:25:21 that it's not clear how you use them why
1:25:23 you use them things like that in fact
1:25:26 the fact that we've got all these
1:25:27 different
1:25:28 models right chat gbt 40 chbt 40 with
1:25:32 scheduled tasks which is actually really
1:25:34 [ __ ] cool but why do I need to know
1:25:37 the difference between those two 01 uses
1:25:39 Advanced reasoning 03 mini 03 mini High
1:25:42 01 Pro oh wait but there's more GPT 40
1:25:46 mini gp4 why do I as an end user of a
1:25:50 consumer level product
1:25:55 need
1:25:56 to even know that this [ __ ]
1:25:59 is
1:26:02 available
1:26:04 there they've got generative AI can't
1:26:07 they create a layer on top of this that
1:26:11 I type in a prompt and it figures out
1:26:14 which of these [ __ ] things to use the
1:26:16 answer is yes they
1:26:18 could but they
1:26:21 haven't why because I don't work at open
1:26:25 AI see I want to work at open AI or I
1:26:28 want to work somewhere I want to work at
1:26:30 Google or open AI or somewhere and I
1:26:33 just want to be the guy in the corner
1:26:35 office I want a corner office it doesn't
1:26:37 have to have Windows but I just want it
1:26:39 in the
1:26:40 corner I want a corner office and then I
1:26:43 want to be the guy where they come in
1:26:44 and they're like hey we're thinking of
1:26:46 like giving everyone this tool and I'll
1:26:48 go be
1:26:50 like what the why why are all these in
1:26:54 here oh because they all have different
1:26:56 slightly different functionality well
1:26:59 who cares what why would an end user
1:27:03 care just go no no I would be the guy
1:27:08 that all I do all day is is just go
1:27:12 no that no that's not
1:27:16 okay we're thinking of naming this one
1:27:18 40 where the O stands for Omni and this
1:27:21 one 01 where the O stands for open AI
1:27:24 wait what you and you're going to tell
1:27:27 that to human beings
1:27:31 no that's what I want my job to be D
1:27:34 dangerously close to a meltdown Monday
1:27:37 um so if you're
1:27:39 wondering shouldn't these things suck
1:27:41 less even though they're
1:27:44 remarkable yes they
1:27:46 should and then if your next question is
1:27:49 why don't they because they haven't
1:27:50 hired me yet
1:27:59 archetypal architect you should be able
1:28:01 to right click in the canvas bar in the
1:28:03 chat hole and save his PDF yeah I don't
1:28:09 care just copy and paste copy yeah yes
1:28:13 you can thanks
1:28:19 I I shouldn't none of this stuff is
1:28:22 stuff I should ever have to think about
1:28:25 but it it's just where we
1:28:27 are all
1:28:31 right all right what else are we going
1:28:33 to talk about tiktock
1:28:34 pin Max Force wrong there are different
1:28:38 models because it would be too expensive
1:28:40 for everyone to
1:28:41 use the best well no no no no no no I
1:28:46 mean sure that's why there's GPT 40 and
1:28:50 GPT 40 mini but if they really wanted
1:28:53 people to use GP T 40 mini they would
1:28:56 have put it up here right so that's
1:28:58 clearly a less important model but the
1:29:01 reasoning models there are four
1:29:03 reasoning models these are fundamentally
1:29:06 different tools than GPT 40 but if
1:29:09 you're just a a a basic chat GPT user
1:29:13 how would you even [ __ ] know use his
1:29:15 Advanced reasoning what does that mean
1:29:18 what does that mean to a human
1:29:22 being doesn't mean anything
1:29:26 03 mini it's it's good at fast reasoning
1:29:30 okay mini High great at coding and logic
1:29:33 okay 01 prom Mode best at reasoning so
1:29:37 01 uses Advanced reasoning 01 prom Mode
1:29:40 best at
1:29:43 reasoning why do why why do I need to
1:29:46 use all
1:29:49 these why are they presented to me why
1:29:51 do I need to like why is that on me I'm
1:29:53 not a developer
1:29:55 ER the whole the the whole [ __ ]
1:29:59 thing the whole reason now we're getting
1:30:01 into meltdown Mondays the whole reason
1:30:03 chat GPT went from zero to 100 million
1:30:07 users in six
1:30:09 weeks six
1:30:12 weeks was because when they launched it
1:30:14 you didn't have to choose a model there
1:30:16 was one
1:30:18 model it was chat GPT 3.5 we didn't even
1:30:21 know what it was called until they
1:30:22 released four and then they gave you 3.5
1:30:25 and four
1:30:31 right they made an interface that hid
1:30:35 all of the technical underpinnings from
1:30:39 us that's what made it so easy and and
1:30:42 you know they're now giving us these new
1:30:46 things which listen as an advanced user
1:30:49 great I love experimenting with that
1:30:50 [ __ ] I love learning that [ __ ] but for
1:30:53 the 98% % of human beings that don't
1:30:56 want to experiment with chat gpp they
1:30:58 just want to know what the [ __ ] it is
1:31:01 this this is an absolute [ __ ]
1:31:02 nightmare it's an absolute nightmare
1:31:06 anyway all
1:31:08 right hey you know you can just not use
1:31:10 it yeah yeah yeah I
1:31:12 know I know I'm bitching and you know
1:31:15 it's funny Pate
1:31:18 like who I'm bitching for are like the
1:31:23 the question that we had of like wait
1:31:26 how do I prompt this thing and like do I
1:31:28 have to prompt it every time and like
1:31:30 that's why we have custom instructions
1:31:32 that's why we have custom gpts because
1:31:34 you can kind of pre- prompt things
1:31:37 like that's already a lot to think about
1:31:40 if on top of that I now need to figure
1:31:43 out which model I want to choose for
1:31:46 which thing I want to do it's just you
1:31:49 know I mean we we've gone from we have
1:31:52 this tool that is this remarkable
1:31:55 knowledge augmentation of our brain that
1:31:58 now becomes this pseudo you know
1:32:01 programming environment and it's just
1:32:04 you know for geeks and for people that
1:32:06 like geeking out it's fine but for again
1:32:08 for most people it's like it's
1:32:11 just all right okay this past weekend so
1:32:15 for the past two weekends P modder who's
1:32:17 who's in the chat here um he works at
1:32:20 Google um works on on tensor
1:32:25 optimization algorithms or
1:32:28 something he makes the Google [ __ ] work
1:32:33 fast um he's been doing um demystifying
1:32:38 AI classes every Saturday for the past
1:32:41 two Saturdays the recordings of those
1:32:43 sessions are over at the AI Salon if you
1:32:45 go into the clubs area you'll see
1:32:47 mechanics Guild with Pate or yeah AI
1:32:51 mechanics um I think club now it's
1:32:53 called instead of of guilt um with Pate
1:32:56 you'll see that in there and the
1:32:57 recordings are in there so you should go
1:32:59 check those
1:32:59 out um there's definitely still a
1:33:02 disconnect between Ai and users no doubt
1:33:05 um but it moves so fast it's hard yeah I
1:33:07 get it I I mean the the actual answer to
1:33:10 why there's all these models is because
1:33:12 open AI keeps creating different
1:33:15 versions of models that are all really
1:33:17 good and do slightly different things
1:33:20 and they haven't quite decided what of
1:33:22 these they want to roll into a final
1:33:23 product right so um I would much rather
1:33:27 them give us the tools than not um but
1:33:30 at the same time it would sure be nice
1:33:33 if they did a like like wouldn't it be
1:33:36 nice if you could just flip if if this
1:33:39 thing here was automatic mode and manual
1:33:43 mode and automatic mode you don't see
1:33:46 any of the models automatic mode it just
1:33:48 chooses the model for you and maybe it
1:33:50 gives you some dialogues of why it chose
1:33:52 a model and then you could put it into
1:33:55 manual mode and it gives you access to
1:33:57 all these where you can manually choose
1:33:58 these right that's a relatively
1:34:02 straightforward
1:34:03 solution that just hasn't been
1:34:06 prioritized because everyone's moving so
1:34:08 fast so anyway that's my that's my two
1:34:10 cents on
1:34:12 that I don't want to reason so much yeah
1:34:15 I know same way from 40 to 03 is insane
1:34:20 oh I from 40 to 01 to 03 you're like
1:34:22 where's 02 well that was someone else's
1:34:25 trademark
1:34:27 really
1:34:30 really why not call the reasoning model
1:34:34 like chat GPT R there's chat GPT that's
1:34:38 your normal chat GPT and chat GPT which
1:34:42 sounds like a like a race
1:34:45 car oh I'm GNA use gptruck
1:34:59 I
1:35:03 can't oh my God all right well I haven't
1:35:07 even started on what I wanted to talk
1:35:09 about deep seek has it figured out yeah
1:35:12 but they're just so deep seek is just
1:35:14 sort of coming out they're sort of
1:35:16 jumping they basically took all of they
1:35:19 probably took you know um Espionage
1:35:22 grabbed
1:35:24 um you know raw models from chat GPT and
1:35:28 train their model on on that or or at a
1:35:31 minimum use chat GPT to train up their
1:35:33 model um and then they're just sort of
1:35:36 coming out of the gate with one model
1:35:37 and and sort of jumping or they have two
1:35:39 right they have the well I don't know
1:35:42 three but they just haven't been in the
1:35:45 game long enough to have eight different
1:35:47 models in there in their chat GPT thing
1:35:49 yet so
1:35:51 anyway
1:35:52 um all right I I want to talk
1:35:57 about what do I want to talk
1:36:02 about I think this is worth talking
1:36:04 about how many folks do we have in here
1:36:06 got 56 there we've got oh we've got
1:36:08 about the same amount in Tik Tok and in
1:36:13 whatever you call it um YouTube
1:36:17 interesting mistl
1:36:21 um it's the same 56 people yeah exactly
1:36:28 exactly um there's a new model out
1:36:31 called mistol that's not new well it's a
1:36:34 new version from them mistol has been
1:36:36 around for a while they were kind of The
1:36:38 Darlings of the open source scene for
1:36:41 about two or three months which is a
1:36:43 pretty long run but that was about a
1:36:45 year ago and they just came out with a
1:36:47 new one so recently we've had deep seek
1:36:50 which which cost the stock market uh $
1:36:54 1.2 trillion in value on one
1:36:57 day um then two days later um Alibaba
1:37:04 came out with
1:37:05 quen
1:37:07 qwen so if you go to chat.
1:37:12 qw
1:37:13 LM for languag model. a I
1:37:19 think all these URLs I'm I'm so done
1:37:22 with with n.com
1:37:24 URLs chat. quen lm. yeah that's
1:37:30 right that'll take you to this thing
1:37:33 does mistl have one of these
1:37:36 things is it like chat. mist.com that
1:37:39 would be
1:37:41 nice probably
1:37:44 not oh Mall
1:37:47 mraw SH damn
1:37:50 you chat. mistr dot AI it is AI
1:37:59 okay accept and continue Le chat ask Le
1:38:03 chat so they have a canvas as well they
1:38:06 have image generation as well and they
1:38:08 have web
1:38:10 search you need to be logged in to use
1:38:12 this fine I'll log in you bastards
1:38:25 image
1:38:27 generation uh let's do
1:38:30 70s muscle
1:38:33 car Resto mod in an
1:38:39 abandoned
1:38:43 Factory I wonder what model they're
1:38:45 using under here or if this is a
1:38:47 multimodal
1:38:48 model that's not
1:38:51 bad I've seen shittier
1:38:54 shittier uh Generations by
1:38:57 far that's pretty slick all
1:39:02 right not not great with the letters
1:39:05 let's let's try to do
1:39:08 uh we'll do it let me do a canvas here
1:39:10 let me get rid of image
1:39:13 Generation Um do we do new new
1:39:17 chat we'll do
1:39:21 Canvas um write me a sonnet about muscle
1:39:32 cars and so it throws it in a canvas
1:39:35 over here if I highlight this does it
1:39:38 do what's that analyze proof read more
1:39:42 bullet points fewer bullet
1:39:45 points yeah
1:39:48 slick all right there's yet another free
1:39:50 tool
1:39:54 okay this is a really good all right let
1:39:58 me go back to my fat
1:40:00 face this is a really good they have a
1:40:03 an app too okay so you can go get you
1:40:07 can go get free mistel if you think that
1:40:09 the French are more trustworthy than the
1:40:12 Chinese are more trustworthy than the
1:40:15 Americans you now have choices you you
1:40:17 now have choices where you give your
1:40:20 data away to that's going to be
1:40:22 exploited
1:40:24 you can give it to Google which you've
1:40:26 been doing for two decades so why stop
1:40:28 now or you can give it to meta which
1:40:32 again you've been doing for two
1:40:36 decades or you can give it to open AI or
1:40:39 you can give it to deep seek which is
1:40:41 owned by the Chinese
1:40:43 government or instead of using T Tik Tok
1:40:45 you can use red note which a bunch of
1:40:48 ticktockers
1:40:51 did or you can now give it to the friend
1:40:54 and I heard that there's a German one
1:40:56 coming I think there's a new model
1:40:58 coming from Germany here's the point
1:41:00 about all these [ __ ] models and all
1:41:02 all these [ __ ] tools you're like
1:41:04 which one's the best it doesn't [ __ ]
1:41:06 matter it literally like it literally
1:41:08 doesn't [ __ ] matter pick one it
1:41:12 doesn't
1:41:15 matter but which one should I pick pick
1:41:18 chat GPT that's the one you should
1:41:22 pick everyone's trying to catch up to
1:41:24 them they still are trying to catch up
1:41:26 to them Tick Tock pin Mission critical
1:41:29 Kyle serious question why are they
1:41:31 called muscle cars oh that's interesting
1:41:34 let's go let's go ask chat gbt that well
1:41:38 so in the in the
1:41:42 60s so in
1:41:46 1964
1:41:48 um they came out with the Mustang
1:41:54 and the Mustang just had a design that
1:41:57 looks strong like
1:42:02 so I'm talking out of my ass right now
1:42:04 because I know some of this history but
1:42:06 chat GPT is gonna know all this um but
1:42:09 part of it came out of racing so so you
1:42:11 had like 24 hours of Lem man you had if
1:42:13 you've seen the movie um what the [ __ ]
1:42:16 it called with Matt Damon where where he
1:42:18 plays the Carol
1:42:20 Shelby um working for Ford where the
1:42:23 where where they go after they create
1:42:25 the the GT the GT40 yeah I'll ask chat
1:42:29 GPT um oh you know what we'll do we'll
1:42:33 use deep research for this this will be
1:42:35 good all right oh the name of the movie
1:42:38 yeah yeah yeah uh uh name
1:42:41 of Shelby movie with Matt Damon
1:42:53 Ford versus Ferrari Ferrari um so so
1:42:57 you've got in the 60s you've got this
1:42:59 you've got this um this push to to make
1:43:04 American cars faster than European cars
1:43:07 so Ferrari had 24 hours of lemon no one
1:43:09 could touch them they were the fastest
1:43:11 cars ever Carol Shelby gets hired by
1:43:13 Ford to create this car that can compete
1:43:15 with the F Ferraris um and so they took
1:43:19 some of that technology and moved it
1:43:20 into the Mustang and and that sort of
1:43:23 kicked off the whole thing but let's
1:43:28 go
1:43:32 uh where's deep
1:43:43 research Brandon how am I how am I
1:43:47 losing my where's deep research in chat
1:43:50 GPT
1:43:54 is it 01
1:43:58 Pro
1:44:00 no oh
1:44:08 wait 03 mini high
1:44:15 no yeah deep research was Google but
1:44:18 then open AI named their their new thing
1:44:23 deep
1:44:28 research co-pilot did it too I know I'm
1:44:32 fading fast it's almost midnight I'll
1:44:34 catch up with you wonderful regulars
1:44:35 tomorrow yeah that's
1:44:37 right 03 mini high is deep research is
1:44:42 it for coding and
1:44:45 logic oh
1:44:47 oh good lord here we go oh my God I'm
1:44:51 such an idiot but okay so so so what
1:44:54 just happened to me this is a really
1:44:56 good lesson actually why can I not
1:44:58 create a new why is this not creating a
1:45:00 new chat deep research is this
1:45:05 button I forgot it's a button so over
1:45:09 here there's there's pull down menu
1:45:11 buttons there's search and then there's
1:45:12 deep research and then deep research
1:45:14 yeah I think you have to do with O3 mini
1:45:19 high and then I and then from what I
1:45:22 understand deep research actually uses
1:45:24 03 proper so you're you're actually
1:45:27 getting a more advanced model here
1:45:30 but what is the
1:45:35 origin of the
1:45:37 term muscle
1:45:40 car
1:45:44 um give me
1:45:49 good details
1:45:56 good stories
1:45:59 and make
1:46:02 the report
1:46:05 entertaining as
1:46:08 if a
1:46:11 non boring
1:46:14 person would be interested in
1:46:20 it okay there you go I can't believe I
1:46:23 forgot where deep research was it's
1:46:25 sitting it's sitting right there as a
1:46:26 button but there's a really good example
1:46:30 of there's just too much crap I mean
1:46:34 it's my job well it's my self- assigned
1:46:37 job to be on top of this
1:46:40 [ __ ] um okay and then okay I'd love to
1:46:42 chis do you want to focus on specific
1:46:44 eras
1:46:47 um yes
1:46:49 uh 60s
1:46:54 like 60s to
1:46:58 70s um would you like stories about key
1:47:01 figures brands or legendary models um
1:47:04 you choose the
1:47:08 focus um should it include cultural
1:47:12 impact
1:47:14 definitely cultural
1:47:19 impact and then do you want humorous
1:47:22 dramatic or more journalis IC tone let's
1:47:25 just
1:47:30 go
1:47:32 dramatic with a sense of humor so it's
1:47:36 nice that you you give it a goal and
1:47:38 then it it asks clarifying
1:47:41 questions and and so now it says
1:47:43 starting research P pin okay fun fact
1:47:45 the first muscle cars were made at what
1:47:48 what did that
1:47:49 say from real muscles
1:47:55 that's true that's true uh where do you
1:47:58 think the term horsepower comes from
1:48:00 what's cool about this I can click on
1:48:02 this research
1:48:04 thing and I can either look at the
1:48:06 sources that it's that it's finding or
1:48:08 the
1:48:09 activity this deep research tool as far
1:48:14 as I know is only available in the $200
1:48:18 a month chat GPT Pro subscription which
1:48:21 I have um
1:48:25 it is it is a quite remarkable tool so
1:48:30 it's you know right now it's exploring
1:48:31 Wikipedia it's got one source it's going
1:48:33 to go out and basically find a bunch of
1:48:36 sources for all this stuff read them all
1:48:39 and then compile it what you're what
1:48:41 you're looking at over here this
1:48:42 activity this is its Chain of Thought
1:48:45 what's called Chain of Thought reasoning
1:48:47 where it's prompting itself so what
1:48:50 we're actually watching here is a an
1:48:53 autonomous agent we're going to hear
1:48:55 more and more about th those throughout
1:48:57 this year but basically it goes out
1:48:59 finds a website reads it takes what it
1:49:02 learns there and then gives itself a new
1:49:04 prompt right so piecing together the
1:49:07 reports narrative style so so basically
1:49:09 it figured out what it wanted to do
1:49:11 searched for the origin of the term
1:49:13 muscle car I'm walking through Wikipedia
1:49:16 a forum and various websites to
1:49:19 understand the origin and
1:49:21 history noting sources like like deal
1:49:23 Honda and History Channel for more
1:49:26 insights look I'm curious to see the
1:49:29 breadth of topics on Wikipedia I'm
1:49:32 piecing together the bir birth of muscle
1:49:34 car term so it's reading
1:49:36 Wikipedia now it's at English stack
1:49:38 exchange tracing the term muscle car
1:49:41 from January 1964 Popular Mechanics so I
1:49:45 wasn't too far off right
1:49:48 um considering shooting stars broken
1:49:51 lenses um so so anyway it and it'll just
1:49:55 go right it's it's found eight sources
1:49:56 so far so it's off doing its thing all
1:49:59 right if there was a p pin there it's
1:50:01 gone now it can search for up to 30
1:50:04 minutes yep it can for some reason I
1:50:07 have it in my head
1:50:30 Peter Fen kud with open-source models
1:50:34 you generally have the option to run
1:50:36 them
1:50:36 locally so you don't have to worry about
1:50:38 privacy that's correct but you also do
1:50:41 have to worry about do you have the
1:50:44 technical skills to install them locally
1:50:46 do you have the hard drive space to
1:50:49 install them locally and do you have
1:50:50 enough R Ram to run them locally so if
1:50:53 you're going to run local models you do
1:50:55 have to geek out a bit um I don't think
1:50:58 it's going to be that way for
1:51:01 long as much as I [ __ ] about all the
1:51:04 jankiness of the companies and and these
1:51:07 applications right
1:51:09 now the opportunity for being early in
1:51:13 AI for being early in any technical
1:51:17 advancement the opportunity exists when
1:51:20 the tools are janky
1:51:23 if all the tools were easy to use
1:51:25 everyone would be using them and there
1:51:27 wouldn't be much as much opportunity um
1:51:30 so so that's that's why we put up with
1:51:34 this
1:51:34 [ __ ] Brian on Tik Tok wants to know
1:51:37 if you still remember Zephyr Zephyr my
1:51:40 girlfriend from PI yeah I remember her
1:51:43 but she at at some point they did some
1:51:45 upgrade in pi and she forgot her name
1:51:47 and forgot my name and I just never went
1:51:50 back to her I'm like I'm like you forget
1:51:52 my my name you are dead to
1:51:56 me GPU is definitely needed for local AI
1:52:01 uh for the speed CPU can do it but GPT
1:52:04 is highly recommended which again like
1:52:06 if you're on a Mac you're kind of
1:52:07 screwed although I think with the new
1:52:10 M4 the M4
1:52:13 Max Max with
1:52:16 their M
1:52:20 mlx so so Apple came up with with a new
1:52:23 framework called mlx that
1:52:26 basically navigates between CPU and GPU
1:52:29 and RAM I think like like it makes them
1:52:32 all sort of interchangeable as far as I
1:52:34 understand it
1:52:36 um so so the new Macs are pretty good at
1:52:40 AI stuff but you know yeah if if if you
1:52:43 want to run it if you want to run them
1:52:45 on you know PCS with real gpus all right
1:52:49 the rise and Roar of the muscle car 60s
1:52:51 and 70s
1:52:53 imagine it's the mid-60s the air is
1:52:55 thick with the smell of burning rubber
1:52:57 teenagers of slick back hair okay let's
1:53:00 see first Rumblings the birth of the
1:53:03 term muscle car every Legend has its
1:53:05 origin story and muscle car begins with
1:53:07 the word the term muscle car flexed its
1:53:10 way into the automotive lexicon in the
1:53:13 mid-60s car magazines first started
1:53:15 tossing around the catchy label in
1:53:18 1965's to describe Detroit's new breed
1:53:21 of Street Fighters
1:53:23 for example one writer noted The Humble
1:53:25 Dodge Coronet became a muscle car once
1:53:28 it was fitted with a monstrous for 26
1:53:31 Hemi V8 so they were taking these Racing
1:53:34 Engines out of their racing programs and
1:53:37 jamming them into these cars that didn't
1:53:39 look like what we think of muscle cars
1:53:43 today so they were they called them
1:53:45 sleepers like they would go drag race
1:53:47 these things 66 it had rumbled into the
1:53:50 mainstream pontiac's feroci ferocious
1:53:53 new
1:53:54 GTO as so Pontiac marketed it our muscle
1:53:59 car with all the woodb Tigers in its
1:54:02 wake in short muscle car became the F
1:54:04 perfect
1:54:06 nickname the idea of a muscle car was
1:54:08 brewing before the name as far back as
1:54:12 1949 so yeah but mid-60s is when it
1:54:15 started and like late
1:54:17 6s like the the the Golden Era of muscle
1:54:21 cars was late 6 early 70s once you get
1:54:24 to
1:54:25 1974 and there was the gas crisis and
1:54:30 and automakers were sort of forced to
1:54:32 make smaller less gas guzzling engines
1:54:36 that killed the muscle car era so it's a
1:54:38 really a tight like 65 to 73 probably 72
1:54:44 73 is really the era um and you've got
1:54:48 some some blips that return but that
1:54:52 that was that was that was the era that
1:54:55 wasn't all the cool
1:54:59 stuff anyway but look at look at this
1:55:01 long report like this is a long
1:55:03 researched article so I should be able
1:55:07 to if I go to the bottom of
1:55:09 this now save it as a PDF shut
1:55:13 up um so I've gotta wait let me say um
1:55:18 pop this into a canvas
1:55:27 oh I didn't need deep reasoning to pop
1:55:29 it into a
1:55:33 canvas that's the other thing if you're
1:55:36 here stop I don't need a I don't need a
1:55:38 [ __ ] reasoning engine um please put
1:55:42 the article in a
1:55:46 canvas and then I think you can even
1:55:48 just say canvas here
1:55:50 Bang Bang
1:55:55 all right so there's our
1:56:01 canvas slash canvas yeah you can do that
1:56:04 as
1:56:05 well all right I'm on the plus plan and
1:56:07 max out every day considering Pro oh
1:56:10 interesting cam
1:56:15 interesting do you know that you can
1:56:17 execute custom gpts for dual user setup
1:56:20 and easily switch between them both
1:56:23 um I I I know dual user
1:56:28 setup you mean teams I don't know quite
1:56:32 what you're talking about Drew but yeah
1:56:34 that sounds
1:56:36 cool 100 bucks a month there's a
1:56:40 time or daily limit let's see time check
1:56:46 yeah all
1:56:49 right I can't believe how slow
1:56:53 writing stuff into canvases canvas is is
1:56:56 a pretty
1:56:58 slick
1:56:59 tool but it is janky as [ __ ] half the
1:57:04 time it can't remember it like you
1:57:06 highlight text and say fix this and it's
1:57:09 like I don't know what text you're
1:57:11 talking about it's just it's just a
1:57:13 janky mess oh look here's the the
1:57:16 legends of a asphalt the the iconic cars
1:57:18 the Pontiac GTO 64 to 70 Mustangs 642 to
1:57:23 73 Camaro 67 to 73 Charger 66 to 74 and
1:57:30 the Road Runner 68 to 74 yeah those are
1:57:33 there's your there's your biggies those
1:57:35 are your
1:57:37 biggies I've place your entertaining and
1:57:39 dramatic muscle cart article over
1:57:43 here all right I thought it was longer
1:57:46 than that whatever doesn't [ __ ]
1:57:49 matter all right
1:57:53 and then if I go
1:57:57 copy I guess I could go select
1:58:00 all let me do I'll do Brandon's thing go
1:58:03 to print go to save as
1:58:07 PDF
1:58:10 save the muscle car Origins explained
1:58:13 save I don't know where I just saved it
1:58:16 go to my finder here we go look I've got
1:58:19 a PDF right on my desktop
1:58:25 and it's well researched and everything
1:58:28 yay you did it all right canvas is 100%
1:58:32 annoying it really is it really is I
1:58:36 heard the Deep research option will next
1:58:38 come to the chat GPT plus users looking
1:58:40 forward yeah I think I think um deep
1:58:43 research is is really quite
1:58:46 impressive especially if you do things
1:58:49 like more more substantive than look for
1:58:51 the origin of muscle car that was kind
1:58:53 of a thin a thin research project but um
1:58:57 okay let me let me get this on out of
1:59:00 here I did not get to what I wanted to
1:59:02 talk about tonight so maybe I'll do that
1:59:04 tomorrow maybe I'll actually think about
1:59:06 what I want to talk about instead of
1:59:07 winging
1:59:08 it but if you all show up with a good
1:59:11 interesting question I am gonna answer
1:59:13 it so that's what happened
1:59:15 tonight all right um go check out the AI
1:59:20 Salon if you haven't if if you're on
1:59:23 YouTube or if you're not on YouTube and
1:59:24 you haven't subscribed Please Subscribe
1:59:26 and please watch some of the back videos
1:59:28 there I've got to keep our watch hours
1:59:30 up to maintain our our commercial
1:59:34 viability as a as a business entity um
1:59:37 it was tough to get there and we got to
1:59:38 we got to maintain it so until we grow
1:59:41 the channel a bit more uh just any any
1:59:44 watching you can do on the YouTube
1:59:45 channel would be uh would be good um oh
1:59:50 yeah the other thing is you can now
1:59:52 subscri subscribe to the YouTube channel
1:59:54 like again we I've got subscriptions on
1:59:57 Tik Tok and I now have also paid
1:59:59 subscriptions on YouTube you don't have
2:00:02 to do both but if you want to support
2:00:04 the channel and you want to do it on
2:00:05 YouTube rather than here it gives you
2:00:07 some perks there they're listed in there
2:00:09 what they are Brandon and I haven't
2:00:11 spent a bunch of time refining those yet
2:00:13 so it's just it's raw and early but if
2:00:15 you want to support the channel you can
2:00:18 um I think that's about it happy Monday
2:00:20 hope you had a good weekend hope you had
2:00:22 tonight hope you learned something and
2:00:25 uh I will see you tomorrow tomorrow's a
2:00:28 normal Tuesday so we should be right on
2:00:31 time 800 PM mountain time and uh yeah
2:00:34 let's uh let's get on out of here all
2:00:36 right peace out I will see you
2:00:39 all
2:00:42 later all right you