aiLL 8/4/23 - The Evolution of Generative AI Tools

Description

Chapters

Transcript

0:01 foreign
0:19 welcome welcome welcome
0:22 Superfly radio guy
0:25 what's up what's shaking
0:29 welcome welcome welcome
0:31 this is the AI learning lab attempt two
0:33 tonight see if we can get the algorithm
0:36 to pick us up
0:37 it was a little stingy last time I went
0:40 happy Friday night
0:42 welcome welcome welcome
0:44 Jeffrey wiggling hey dude I've learned
0:46 so much from you and I appreciate you so
0:48 much I appreciate that KT
0:50 that's why I'm here hopefully
0:53 hopefully making a difference for people
0:55 talking about AI generative AI why I
0:58 think it's important what it is how it
1:00 works how to use it
1:02 how not to think about it you can make
1:05 money without gbt
1:09 you can think about it like that but
1:11 it's just a tool it's it's
1:14 the most powerful tool ever put in the
1:17 hands of humanity hey Emilio's life we
1:20 are back
1:22 I was doing a little Sim racing
1:24 and some jerk crashed me at the end of a
1:27 good qualifying lap so I was like
1:30 some back on
1:31 but I'm double fisted now I've got iced
1:35 beverages
1:38 it's gonna be good I don't know if my
1:40 voice is any better now it looks like we
1:42 got some folks coming in hey everybody
1:44 encourage more viewers to share
1:46 um yeah uh by the way if you're willing
1:49 to if you if you've got people that you
1:51 think might be interested in AI
1:53 uh or you don't mind them thinking
1:55 you're a weirdo for thinking about it
1:58 um please share this we'll get this out
1:59 to as many folks as we can I appreciate
2:02 that double tap the screen get the
2:04 little hearty Hearts going and
2:07 um we'll get started these URLs beside
2:10 me here someone just sent me something
2:12 thank you for that I think it said Carla
2:15 um it's hilariously AI but still
2:17 impressive wait what is
2:19 and X or whatever it is oh xai yeah
2:23 whatever
2:25 um these things right here wait let me I
2:28 got I got white bleed coming onto my
2:31 camera lens I have a feeling my
2:33 lenses are a little dirty not even dirty
2:38 dirty lenses those lenses are so dirty
2:43 all right there's my undirty lenses that
2:46 looks a little better that's Mo Betta so
2:49 these URLs here I'll talk about them in
2:51 a bit if you haven't played with chat
2:53 EBT or or any of these generative AI
2:56 tools
2:57 um this list only used to have three
2:59 things on it and as of about 10 days ago
3:01 I added four more so that's seven
3:04 that's seven and so I I would consider
3:07 these kind of the seven probably five
3:10 major you know majors and two miners
3:14 although the one minor the pi dot a i
3:17 the product right now is minor but the
3:19 company is Major it's so the company
3:21 behind Pi dot AI is inflection
3:25 they just raised 1.3 billion dollars for
3:29 their series B now they're going to
3:30 spend about 900 million of that on gpus
3:34 22 000 gpus so they're making a giant
3:38 giant ass server Farm to build
3:41 what they're what they're now calling
3:42 they've they've come up with this new
3:44 term called Frontier models right
3:46 there's Foundation models which are the
3:48 and the frontier models are the latest
3:50 and greatest you know when you got 10
3:53 000 or more gpus and you can train
3:55 massive massive large language models
3:58 then you're a frontier developer so a
4:01 lot of these a lot of these companies
4:03 are Frontier open AI is Claudius
4:05 anthropic is Google is meta is but
4:09 meta's giving their [ __ ] away which is
4:11 really interesting and then inflection
4:12 is as well so
4:15 um if you have questions about
4:16 generative AI pop them in the comments
4:18 below
4:20 um my background I'm an entrepreneur
4:23 um I've been I've started 12 or 15
4:24 companies in my day I was an
4:26 entrepreneur in the early days of the
4:27 World Wide Web so I've been through one
4:29 of these
4:31 um technological revolutions from from
4:33 the not from the ground up but from the
4:35 very early days and uh
4:38 this one is like the early days of the
4:41 World Wide Web like this year is very
4:43 much for for my experience like 1994.
4:46 when I first discovered the the world
4:48 wide web and
4:50 um
4:51 started an agency in the early days this
4:54 is very much like that but like if if
4:56 you multiplied it by a thousand
4:59 in terms of the impact that these things
5:01 are going to have in the speed with
5:02 which they're coming
5:05 uh if you have any questions pop them in
5:07 the comments below I'll do my best to
5:08 answer them I'm back I was here before
5:11 I'm back
5:14 and uh yeah we'll see what we can do
5:18 what is Labs well the AI learning lab is
5:23 this is the lab isn't it lovely
5:26 we've got books and photographs some
5:28 baskets and
5:31 painting painting
5:35 that's the lab it's just the name of my
5:37 channel it's about learning
5:41 learning uh this stuff
5:47 oh what about 11 Labs I can talk about
5:51 that
5:52 we can talk about 11 labs for sure
5:59 any updates to pie
6:02 um I don't know when the last time you
6:03 looked at it was Pi is the the IOS app
6:06 for pi if you put it in a conversation
6:08 mode in your car while you're driving
6:12 um pie is really fun to talk to when
6:13 you're in your car
6:15 so I don't think it's useful for doing
6:17 things like research it is connected to
6:19 the internet though so that's useful
6:21 and you can talk to it which is it's a
6:24 bit of a novelty but if you're driving
6:26 like the idea of actually having a large
6:29 language model that's connected to the
6:31 internet that you can just go hey uh I'm
6:34 headed toward the movies what's playing
6:36 and it'll just tell you and do all that
6:38 sort of stuff I mean I suppose Siri
6:40 could have done that but
6:42 um
6:43 but you know you can ask it large
6:46 language model stuff so I don't know if
6:47 there's been an update to it but they
6:49 have raised a boatload of money so I'm
6:51 I'm sure they got some some stuff coming
6:55 um
6:59 hello everyone hello Nancy what about 11
7:01 Labs I've played there with my clone
7:04 voice yeah 11 Labs right now is is
7:06 pretty good
7:08 um d-id is pretty good Synthesia is okay
7:12 all of them are expensive they're they
7:15 you just burn through credits really
7:17 quick with that so I think if you have
7:19 kind of low volume
7:23 low Fidelity projects where it doesn't
7:25 really matter like if you're trying to
7:27 create a video a day for a tick tock
7:29 Channel
7:31 um or two videos a day maybe you could
7:33 probably do that without burning too
7:35 much money but if you're trying to if
7:36 you want to do like
7:39 um go to chat EBT and have it write 100
7:42 scripts and then automate you know going
7:45 to 11 labs and making voices and then
7:48 going to canva and doing you know 100
7:51 videos with uh what's it called bulk
7:55 create uh that's gonna get pretty
7:57 expensive pretty quick so I kind of feel
7:59 like on the video stuff The Voice
8:01 synthesis stuff the simulated talking
8:04 stuff all that stuff right now is it's
8:07 we're a little early we need we need
8:08 another year so
8:11 keep playing with it because it's
8:13 getting better and better like there's a
8:14 cool thing within 11 Labs is it 11 Labs
8:17 or did it might be both of them
8:20 you can you can set the voices to
8:23 certain
8:24 um
8:25 like emotional realities like you can
8:27 say shouting Whispering you know
8:31 um talking aggressively talking you know
8:34 funnily
8:36 so right now you kind of have to
8:37 manually assign those but I assume with
8:40 the API you you could have
8:42 chat GPT
8:45 analyze a script and say you know here's
8:48 here's your emotions that you can
8:50 express yourself in and then maybe
8:53 here's things like pauses
8:55 um here's a script now analyze that
8:58 script and put in the appropriate places
9:00 for the emotional stuff
9:02 um
9:03 I would think it would be relatively
9:04 trivial to do that but none of these
9:06 tools have made it easy to do that yet
9:08 which I think stupid so it's coming but
9:10 you know that that'll make them sound
9:12 much much more realistic
9:14 they're pretty comically bad still
9:20 but they'll get better
9:22 all right
9:23 let's see thanks for the heads up I
9:26 appreciate you oh that's you're talking
9:28 to each other what is Labs I just talked
9:30 about that I'm interested to have a bot
9:32 that is not restricted not wanting to do
9:34 anything illegal but I don't want yeah
9:36 Joker I don't like restrictions yeah I'm
9:39 fine with that so what you need to do
9:41 Joker is
9:43 uh depending how technical you are you
9:46 you can go to labs.perplexity.ai
9:49 um that's that's meta's
9:52 um llama open source models that don't
9:55 have restrictions on them those have
9:56 been open source the developer
9:58 Communities going nuts on those
10:01 so so if you're geeky go to hugging face
10:04 or GitHub look at the projects that are
10:06 playing with llama2 or just any of the
10:08 open source go go to Reddit and search
10:10 for open source llms large language
10:13 models
10:15 and um just start looking at the
10:17 projects there hey the bear what's
10:18 what's happening thanks for the Roses I
10:20 appreciate that as always always always
10:22 always that's super cool
10:25 um
10:27 yeah the and and with the open source
10:29 stuff you can do whatever the hell you
10:31 want without restrictions you can run it
10:33 on local machine if you have enough iron
10:35 um I think you can even run without a
10:37 ton of iron like on an M1 Mac or a you
10:40 know a decent PC you can run the 13
10:42 billion parameter model decently fast
10:45 when you get to the 70 billion parameter
10:48 model you're gonna have to run that
10:49 either in the cloud
10:50 with gpus or or on a pretty beefy
10:53 machine but whatever like the developer
10:56 Community right now because llama 2 was
10:59 commercially licensed by meta which
11:01 basically means you can actually build a
11:03 business on it
11:04 um they're gonna be optimizing the [ __ ]
11:07 out of those things they're going to be
11:08 fine-tuning them optimizing them there's
11:10 going to be just projects coming out of
11:12 the woodwork probably September up
11:14 October
11:16 um so I would just go start exploring
11:19 that and then you don't have to deal
11:20 with all of that
11:22 crap
11:26 but I want to do bad stuff with my large
11:29 language model you're not allowed
11:33 we told the government we'd be good
11:35 companies yeah that's horrible then Mark
11:40 Zuckerberg's like here you want one take
11:42 it go play
11:44 [Laughter]
11:47 but what if someone uses it to destroy
11:50 the world yeah what are you gonna do
11:51 we're meta
11:54 come to the metaverse it's gonna be
11:56 awesome
11:57 in 2052
12:01 I can't believe it's 2023 and the
12:03 metaverse still looks like um Second
12:05 Life from 2001 1999. it's like seriously
12:10 [ __ ] 25 years and this is as good as
12:12 you've gotten
12:15 it's like the metaphors to me always
12:17 seems like
12:19 like if you get if you show up early to
12:22 a Junior High dance and it's just like
12:25 people like standing around like
12:27 awkwardly like there's not enough of
12:29 them in the gymnasium and no one knows
12:31 what to do so I'm just doing this that's
12:34 what the metaverse feels like to me yeah
12:39 into my art opening yeah
12:42 it's great they look very realistic yeah
12:49 we need procedural World Generation it's
12:52 coming go to uh blockade Labs have you
12:54 seen the blockade lab [ __ ]
12:56 that's fun
12:57 let's see where is it
12:59 um
12:59 [Music]
13:01 this stuff
13:03 where you just tell it tell it to create
13:05 a world and it creates a world for you
13:09 skybox.blockadelabs.com that's cool
13:12 now it's it's really primitive but you
13:16 know cool
13:18 all right let's see I was on in 1994
13:24 chatting with other bartenders at 56k
13:26 yes over Fox I know remember the uh the
13:29 old uh
13:30 the old dial-up modems here I got I got
13:33 a I got a slide for that
13:37 let's see this one
13:42 that's it let's see um
13:45 let's see
13:48 [Music]
13:54 so this is a
13:56 this is a pitch I did so I was living in
13:59 New York City and I had an idea for a
14:01 magazine and so I
14:03 I coded up the first issue of urban
14:06 desires it was a fancy magazine and then
14:09 I learned I had to upload it to some
14:11 server in California to put it on the
14:13 internet
14:21 [Music]
14:24 if you're young enough to not know what
14:26 that sound is [ __ ] you all right
14:29 and then what was amazing was
14:32 so it that was the the last week of
14:34 November in 1994 I uploaded this issue
14:38 of urban desires to some computer in
14:41 California and three weeks later I got
14:44 an email and someone goes hey did you
14:46 know that
14:48 um Urban desires has been written up in
14:50 Liberation the Parisian News Daily
14:52 I was like what
14:54 and I ran down to the international
14:56 newsstand at Times Square
14:59 and I bought the last week's Liberation
15:02 and sure enough there's a full page
15:03 article of
15:07 thank you
15:08 Urban desires
15:10 in this
15:11 French newspaper
15:13 and I knew in that moment the world had
15:15 changed
15:17 it was amazing it was an amazing moment
15:19 it was like oh my God everything's
15:22 different
15:23 and that's exactly what's going on with
15:25 this uh with this AI [ __ ]
15:28 exactly the same exactly exactly except
15:33 this [ __ ] faster and way more profound
15:36 it gonna make the world wide web look
15:38 clean and cute
15:40 like a little doily
15:45 crazy huh nuts I want to get training
15:48 I'm a business intelligence developer
15:50 what tool should I know
15:52 start with these
15:54 just start with these especially in
15:56 business intelligence
15:58 um
15:59 if you don't have the 20 Buck a month
16:02 subscription for chat GPT and access to
16:06 code interpreter you're missing out so I
16:09 would start with just
16:11 chat.openai.com pay the 20 bucks I'm not
16:13 I'm not affiliated with it so I don't
16:15 give a [ __ ] if you do it or not
16:18 um I'm not doing it to make money
16:20 but it's got um that's that gives you
16:23 access to gpt4 which of all of these
16:25 things is is the strongest model
16:28 um and it's got code interpreter where
16:30 you can upload all sorts of files like
16:32 you could upload annual reports or you
16:34 could upload data and
16:37 um it will just analyze that for you
16:39 I'll show you how I can show you later I
16:41 can do a demo of it it's insane
16:44 um you should also be playing with
16:46 things that are connected to the
16:47 internet so bing.com bard.com uh you can
16:51 probably avoid pie Bing and Bard are
16:54 both connected to the internet and they
16:56 are
16:58 um
17:00 like Bing is gpt4 it it can also read
17:03 images so it can do things like you know
17:06 analyze an image and what's in it that's
17:08 where I would start
17:09 and then use if you're really new to
17:12 prompting and don't really know what
17:14 what the hell I'm talking about go to
17:16 prompts.chat and that's a document that
17:18 will tell you a little bit about how
17:20 these large language models work and
17:22 what prompting really is and then it'll
17:24 give you a whole bunch of ideas of
17:26 different personas that you can tell
17:27 these these things to act like and it
17:31 will then
17:32 um
17:33 you'll begin to get your head around how
17:36 these things are way more
17:38 significant than a Google search
17:40 and if you do it right you'll have your
17:42 your first Kevin McAllister moment
17:45 what's that Kyle
17:48 remember Home Alone
17:50 when you play with these tools there's
17:51 there's some moment where that everyone
17:54 that I've seen that plays with these
17:55 things even if they're like cynical
17:57 about it well I hear those things you're
17:58 stupid
17:59 there's some moment where they see it do
18:02 something where they go
18:03 uh I didn't know I could do that
18:08 I need a drink
18:11 does this mean it's going to take my job
18:16 I'm gonna change it
18:18 and if you don't learn these tools it's
18:20 going to take it
18:22 hey I won't take your job but someone
18:23 using AI will take your job
18:27 um AI literacy is is going to be a
18:29 requirement more quickly than I think
18:31 many people are prepared for that's what
18:33 this channel is about oh what's your
18:34 Channel about hey digital product so so
18:37 what this channel is about is
18:40 um
18:42 I
18:43 yeah at a really simple level just
18:44 demystifying generative AI so so in my
18:48 opinion November 30th 2022 the launch of
18:51 chat GPT marks an inflection point in
18:54 history frankly
18:56 um because these things
18:57 these things are in the neighborhood of
19:00 steam engine kind of importance like
19:02 Industrial Revolution or printing press
19:04 or fire like right now I've got it down
19:07 to those three
19:09 probably probably language
19:12 you know written language written
19:14 language fire printing press steam
19:17 engine this [ __ ]
19:19 it's in that neighborhood so this
19:22 channel is about
19:24 um
19:25 I'm an entrepreneur and I tend to jump
19:27 on technology right as it sort of enters
19:30 the mainstream so November 30th large
19:33 language models in AI have been around
19:35 for decades and then chat GPT says hey
19:38 everybody who's not a data scientist you
19:40 want to play with the most powerful tool
19:41 ever created in the history of humanity
19:44 and 100 million people in six weeks said
19:46 sure I'll play with that
19:48 and because Chachi BT was so
19:52 dramatically so quickly adopted
19:55 um fastest adoption of technology in
19:57 history
19:59 um
20:00 all of these other companies sort of
20:02 panicked they they've all been working
20:04 on AI projects for years and AI
20:06 Technologies for years but they've been
20:08 kind of holding them back well it's a
20:09 little edgy it's too it's not safe it
20:11 threatens our business model and then
20:14 open AI comes out and says [ __ ] it here
20:16 you go and Microsoft says we put a
20:17 billion dollars in you know what we put
20:19 10 billion dollars in and that changed
20:22 everything
20:23 and so now you've got you know
20:25 Capital flowing into the market you've
20:28 got these Technologies being accelerated
20:30 and the speed of development of these
20:33 things is is just on this exponential
20:35 curve it's Insanity so this channel is
20:38 about saying go try these things go play
20:40 get curious about it explore it if
20:43 you're a copywriter and you haven't
20:44 played with these
20:46 good [ __ ] luck like these things are
20:49 incredible and you know copywriters will
20:52 say wow they don't write as good as I do
20:53 and they're not perfect and they
20:55 hallucinate okay
20:58 have you used them well not really okay
21:00 go use them
21:03 because every copywriter that I know
21:05 that's used these is in a completely
21:07 different plane right now the quality of
21:10 their work went up the quantity of their
21:13 work went up
21:14 and
21:16 there are many of them in existential
21:19 crises about realizing that
21:23 what they used to consider their core
21:26 value is is being replaced with these
21:29 machines that because they're getting
21:31 that much better
21:32 and so they're having to reinvent
21:34 themselves some of them are very sad
21:36 about it some of them are excited about
21:37 it some of them are both and so that's
21:40 what this channel is about is answering
21:42 questions
21:43 um
21:44 you know playing with these tools doing
21:47 what I can if someone's got a specific
21:48 question uh things like that so do I
21:51 record these lives they are recorded
21:54 um they keep them for 90 days so I've
21:56 got a I've gotta I gotta dump them off
21:58 Tick Tock Unfortunately they don't have
21:59 like a batch download thanks for the
22:01 Roses I appreciate that
22:03 um
22:04 but I think what I'm going to do is I'm
22:06 gonna I'll put together a YouTube
22:08 channel or something like that I think I
22:10 will record these because I think
22:11 there's good value here although I don't
22:13 know I don't know how long these things
22:15 are going to be valuable because the
22:17 text moving so fast like if you want to
22:19 watch them for the entertainment value
22:20 of an old guy making stupid jokes then
22:24 they're there for posterity's sake but I
22:26 don't know that the value is going to
22:27 last all that long
22:29 um but
22:30 you know at some point I will I don't
22:33 know I don't know I don't know I'm just
22:35 here I'm just here instead of sim racing
22:37 I'm hopping on Tick Tock live and
22:39 talking about AI because I think it's
22:41 actually
22:42 quite important
22:45 um
22:46 Emilio's wife it's the best lab ever AI
22:49 learning lab we're the best you can make
22:52 money hey
22:56 I've been in the business for about 20
22:58 years oh cool yeah I I mean listen if
23:01 you if you've been in the business for a
23:02 while
23:04 this stuff is like like even Engineers
23:06 are looking at this stuff going uh
23:08 things are different things are changing
23:10 if you're an engineer and have not
23:11 played with code interpreter it time the
23:14 time is now I'll play with it tonight
23:15 the weird AI Train movie is on at
23:20 rainisto profile on Twitter
23:22 music all done by AI all right hang on
23:25 let me go look at that
23:34 Twitter don't you mean X
23:37 you're so behind the times they changed
23:40 that like last week I mean come on get
23:42 with it
23:44 what did you say at
23:47 rainisto
23:49 rain eastoe
23:54 here we go oh Subway as in the New York
23:58 City subway all right let's let's go
24:00 watch it together D'Amato I assume it's
24:03 the motto tomato de Malo let's Call the
24:07 Whole Thing Off eh yeah
24:10 all right
24:12 let me can I get my fat head out of the
24:15 way here
24:16 uh
24:18 uh hang on
24:23 oh dang it
24:25 I just got rid of my double camera
24:27 I don't I gotta do something to get that
24:29 back ah crap
24:33 oh Tick Tock why why
24:40 all right let's go watch this thing
24:43 well my Fat Head's not in the way
24:45 anymore
24:49 um
24:54 so this is a two minute 37
24:57 second AI generated movie I guess from
25:00 reinisto is that right Subway store
25:02 you're right down the dark New York City
25:04 subway tunnels this is going to make me
25:06 Miss New York dang it
25:08 stop that
25:20 train number six come in
25:22 here
25:23 Frank watch your speed a little it's not
25:27 a race
25:28 I'm running late I'm not going to be
25:30 late your passengers won't like the ride
25:34 all rats anyway
25:35 [Music]
25:52 uh Frank did you leave the the station
25:55 early I gave them 15 seconds
25:59 can't do that you're breaking the rules
26:01 don't give me that
26:04 let me talk to Jenkins
26:06 hi Frank is everything all right tell
26:09 her about the rats will you
26:12 all right so so then what are you
26:14 talking the boy the the voice acting
26:16 isn't quite there yet but you know this
26:18 is not bad
26:19 I will show you
26:22 I will show you hello
26:25 it's coming this way we'll cut the power
26:27 once he gets near what was he talking
26:28 about
26:31 carefully there are no wraps in the
26:33 subway
26:50 submission
26:58 [Music]
27:05 who knows what I'm waiting for it it
27:07 doesn't matter as time is over
27:09 it's got nowhere to go
27:13 here he comes
27:22 foreign
27:37 [Music]
27:56 it's um oh wait where's my
27:59 how do I get my thing back
28:03 let me flip my camera
28:07 everybody know how to get dual camera
28:08 back dang it
28:12 dude
28:18 yeah I lost my dual camera oh well dang
28:22 it
28:23 um yeah that that's interesting
28:26 um you know the like
28:29 I think it there was too much dialogue
28:32 in there that was too AI written and too
28:34 AI acted right I'd have just gone with
28:36 music and there was a little bit too
28:38 much
28:38 drama for a two-minute thing but but you
28:41 know
28:43 um
28:44 labs.perplexity.com that is uh
28:47 that's meta's llama 2 model
28:52 the jailbreak break prompts kind of work
28:56 I heard you mention God mode
28:58 well
29:00 god mode so there's a there's a site
29:02 called God Mode God mode dot space which
29:05 is a an autonomous agent it's basically
29:09 a web front end to Auto GPT and it's a
29:12 little it's a little janky but it's it's
29:14 interesting to play with I think code
29:16 interpreter is more interesting to play
29:17 with
29:19 um
29:20 and uh what else uh
29:24 uh the jailbreak the jailbreak prompts
29:28 Joker there there was there's the
29:30 prompts like like
29:32 um dance do anything now but there was a
29:35 paper that came out two or three days
29:37 ago that that some researchers figured
29:40 out a sequence of code to stick at the
29:42 end of a prompt that breaks all safety
29:45 stuff it completely jailbreaks the large
29:48 language models and apparently it
29:50 doesn't sound like
29:51 um
29:52 open AI is going to be able to fix it I
29:55 don't know what it does I didn't really
29:56 read the paper but apparently it's a
29:59 it's like a it's like a master key for
30:01 for
30:02 jailbreaking these things
30:06 but I think that's less of a big deal
30:07 because we have
30:09 um llama 2 out
30:11 I'm a business intelligence developers
30:14 what's next for me
30:17 um
30:18 well what's next for you is that the the
30:20 work that
30:24 the work that it probably takes you a
30:27 month or two to do now you'll probably
30:28 be able to do within a week or two
30:31 um
30:32 and
30:36 where the core value of of your business
30:39 intelligence work is is going to shift
30:42 and the demand for what it was
30:45 historically is gonna
30:47 shift if not evaporate
30:50 because if you've got tools that can
30:52 allow
30:54 Senior Management to just go hey
30:57 computer tell me what's going on with my
31:00 company
31:01 and it just does it
31:03 they don't necessarily need you to go
31:05 build them dashboards right so
31:09 um so I think short term get AI literate
31:12 as quickly as you can
31:15 um yeah so
31:18 let's see
31:28 I'm trying to figure out
31:33 so what was that that just flipped my
31:35 camera
31:41 that
31:44 panel grid
31:48 hmm
31:51 sorry old guy trying to figure out Tick
31:54 Tock alert
31:57 all right companies that use chat EBT
32:00 will be the as the basis for their AI
32:03 bot aren't they just watered down
32:05 versions companies that use chat GPT as
32:07 the basis for their AI bot aren't they
32:10 just watered down versions well they
32:12 they are and they aren't they call them
32:13 rapper apps where where you're putting a
32:15 wrapper around the GPT for foundation
32:20 um
32:21 if you go to poe.com
32:23 hang on
32:27 all right
32:28 if you go to poe.com
32:30 you can build your own Bots and so
32:33 that's that's essentially what a lot of
32:34 these companies are doing it's not
32:37 necessarily a bad thing for companies to
32:40 pre-prompt
32:41 U you know these models because not
32:43 everyone's going to want to figure out
32:44 prompt crafting and not everyone's going
32:46 to figure out how to be a prompt
32:48 engineer and really geek out on all
32:50 these tools so if you go to something
32:52 like Jasper or copy AI where it's
32:55 basically a whole Suite of pre-prompted
32:57 you know
32:59 you know tools sitting on top of chat
33:01 gbt it's not necessarily a bad thing
33:03 like if it adds value to you great if
33:05 you look at a McKenzie Bose has a GPT
33:09 boss that's what that's doing right
33:11 so so yes that's what a lot of them are
33:14 and that's what a lot of apps are going
33:16 to be if you go to
33:19 um
33:19 if you go to futurepedia.io there's like
33:22 4 000 AI tools here probably 3 700 of
33:27 them are just [ __ ] rapper apps
33:30 sitting on top of chat gbt but there's
33:33 you know 300 in there that are actually
33:35 really really interesting
33:38 um and that's that's just going to be
33:39 the nature of how it is because these
33:41 tools are so powerful and people realize
33:44 oh I could do one of these for Real
33:46 Estate or I could do one of these for
33:48 refrigerator repair people and then they
33:50 they just sort of pop them up there
33:55 um let's see I'm gonna try something
33:57 here effects
34:08 and this
34:10 is so search here tool
34:15 background face green screen
34:19 don't know what that is
34:24 [Music]
34:29 oh
34:30 I found it wait where is it I got okay
34:33 hang on hold please
34:36 oh okay there it is got it
34:38 so if I go to enhance I can bring it
34:40 back all right yay
34:44 what'd you do Friday night I watched
34:46 some old guy try to learn how to use
34:47 tick tock how'd that go for you it was
34:50 good it was good it was it was
34:53 I don't know what I'm doing with my life
34:54 I need help
34:56 I need help
34:59 well how long did you watch him two
35:01 hours
35:03 something's wrong I don't know
35:07 all right
35:09 say what so hope that answered your
35:12 question say what
35:14 DDD
35:16 labs.perplexity.ai again that's Meadows
35:18 llama 2 model let's see if we can get
35:20 our
35:21 Focus fixed there not really what's
35:23 going on with that dang you Apple there
35:26 we go okay
35:28 do they have to be compiled on Linux you
35:31 just answered that
35:33 no
35:35 one of the best users of per what are
35:38 the best users of perplexity versus Club
35:41 to oh best uses of perplexity versus
35:44 Claude two okay
35:47 the big advantage of Claude 2 is that
35:51 it's got a hundred thousand token
35:53 context window it's probably
35:56 from a language standpoint somewhere
35:58 between GPT 3.5 and gpt4 it's not gpt4
36:02 gpt4 is is the
36:06 is the best still
36:08 perplexity is kind of like a search
36:11 engine combined with large language
36:14 model combined with like sourcing so if
36:17 you want to do research perplexity is
36:19 really good at that labs.perplexity is
36:22 not perplexity
36:23 labs.perplexity is them just being nice
36:26 folks and and letting us play with the
36:30 Llama 7 13 and 70 billion models on
36:34 their servers so that's why I have it
36:36 here so it's kind of like playing with
36:38 chat CPT but you're playing with llama
36:39 too
36:41 um it's not connected to the internet
36:43 some of the perplexity tools are but the
36:47 Llama 2 things are not so the reason I
36:51 would use
36:52 labs.perplexity.com is to experiment
36:55 with
36:57 um
36:57 with llama 2 and see if it's good enough
37:00 to do whatever the project is you want
37:02 to do and then you could look into you
37:05 know spinning up your own server and
37:07 host your own llama and you could
37:08 basically have your own chat GPT running
37:10 locally or running on a on a company
37:12 Network so none of your data goes
37:15 outside your firewall
37:17 right so so that's where I'd be
37:19 experimenting with there Claude because
37:21 it's got a hundred thousand token
37:23 context window you can upload like five
37:25 PDF documents to that like up uh you
37:28 know the the person that was uh asking
37:30 about business intelligence like upload
37:32 five annual reports to Claude and start
37:36 interacting with them start asking it
37:38 questions about the annual reports and
37:40 see what's there
37:45 okay thank you
37:47 um there was one free GPT or something
37:49 like that I read about today there's
37:51 there's GPT for all so so there's a
37:54 project called GPT for all that's got
37:56 downloadable installers for Mac windows
37:59 and Linux
38:01 and then you you run the installer it
38:03 installs it on your machine and it
38:04 basically makes it an application and
38:06 you can add in different large language
38:08 models to it so you can run things
38:10 locally
38:12 once you get into the world of Open
38:14 Source llms I'm I've not kept up to date
38:18 on them because they're they're changing
38:20 like every two days so I can't I can't I
38:23 just can't do it
38:25 um but there are projects out there like
38:27 GPT for all that allow you to uh to play
38:31 with this stuff so
38:33 all right
38:34 let me put that there not lose it like I
38:36 did before but now I know how to get it
38:38 back because I'm a master of tick tock
38:41 you thought I was old and didn't know
38:43 how to use technology I can do this
38:47 this is not just a Zen gen Z thing gen
38:51 xers can do this we've got fingers we
38:54 have phalanges
38:58 I've mastered it it's the tick and the
39:01 talk you you ticket and you tuck it and
39:04 off it goes to the
39:07 satellite space internet thing
39:14 winner
39:15 I'm sure it's too small it binds up on
39:18 my massively muscular arms
39:22 thank you
39:23 they looked a little spindly to me shut
39:26 up
39:28 pick on me
39:31 because I'm fat and weak doesn't mean
39:33 you need to call me fat and weak
39:36 uh Kyle I complain chat GPT Claude Bing
39:38 Bard get the data wrong all the time yes
39:40 I know they do
39:42 they hallucinate what do you think about
39:44 the Microsoft chat bot in Microsoft 365.
39:47 I haven't got a chance to play with it
39:48 yet so
39:50 um
39:52 I think it's going to change everything
39:53 because right now these tools are all
39:56 sort of off in the AI ghetto right
39:58 they're off to the side and everyone's
40:00 like well should should we use AI
40:02 detectors should we allow our employees
40:04 to use AI it seems very risky
40:08 and now they're just going to be rolled
40:10 into every [ __ ] tool we use they're
40:12 going to be built into Microsoft Word
40:14 and goo and and
40:16 um and PowerPoint and
40:19 Excel and sheets and docs and Google
40:22 Drive and whatever the [ __ ]
40:25 Google Drive is for Microsoft
40:29 they're already in notion and all the
40:32 note taking apps like it's just going to
40:34 be everywhere these things are very much
40:36 going to be like the World Wide Web
40:37 where the World Wide Web is just this
40:39 ubiquitous tool that's everywhere and
40:42 then there's all these businesses built
40:43 on top of it that's what these things
40:45 are going to become they're just going
40:46 to be everywhere
40:48 the more you work with these things the
40:50 more that you realize it's not AI
40:53 writing it these are tools like
40:56 Photoshop is a tool and Lightroom is a
40:59 tool to manipulate digital images these
41:02 are tools to manipulate language now
41:04 they're generative so they happen to be
41:06 the most powerful tools we've ever had
41:08 in our grubby little [ __ ] fingers
41:13 but
41:16 the skill moving forward is going to be
41:19 moving kind of seamlessly back and forth
41:21 between letting the tool generate
41:23 something or me having an invention an
41:26 intention as a human and then I put it
41:29 in the tool and it generates some little
41:31 magical flower of that and then I take
41:33 it back and I'm like oh let me make that
41:35 suck less and then let me take that and
41:37 put it back here and do a little more
41:38 generation and then it's this back and
41:40 forth you know handoff collaboration
41:43 between you and the tool you're the
41:46 pilot it's the co-pilot I think the
41:47 Microsoft branding is right on for this
41:49 stuff
41:51 um
41:54 I haven't played with it enough to know
41:57 how
41:59 seamless their integration is
42:02 but if it's anything even remotely like
42:05 what they showed in their demo videos
42:06 it's going to be quite remarkable
42:09 like write me a PowerPoint presentation
42:12 of this annual report
42:15 and out it comes
42:17 if you can do that within the context of
42:20 you know working within your own files
42:23 on your own servers with all your own
42:26 documents and you can just seamlessly
42:28 transition from a word doc to a
42:30 PowerPoint to generating some images to
42:33 turning it into
42:35 you know something for the annual
42:38 shareholders meaning or I don't know
42:39 whatever the hell you're doing
42:41 great
42:43 but you know who's gonna have that skill
42:46 I don't know
42:48 the people that are not running away
42:50 from this [ __ ]
42:51 saying well it's not perfect it's evil
42:57 AI is evil I don't agree with it
43:01 okay you don't have to agree with it
43:05 that's the equivalent of Florida saying
43:07 they don't agree with hurricanes
43:09 okay
43:10 glad you know appreciate you not
43:12 agreeing with hurricanes they're still
43:14 [ __ ] coming
43:17 and and lots of them and more and more
43:19 powerful and faster that's what this is
43:25 so so I think it's inevitable and I
43:29 think the fact that Microsoft is rolling
43:33 um these tools throughout their entire
43:36 product line from cloud services to
43:38 GitHub to Dynamics 365 to Office 365 to
43:42 Windows
43:43 they're rolling it across everything
43:46 that forces Google to have to do the
43:49 same thing and that forces everyone else
43:52 to have to deal with it
43:54 Apple's gonna have to deal with it
43:55 they're apparently working on Apple GPT
43:57 right now that's not on this list
43:59 because it doesn't exist yet but I
44:01 guarantee when their [ __ ] comes out it's
44:03 going to be good
44:06 so we're we're in a you know we're in a
44:09 features race right now and and nobody
44:12 knows what the implications of that are
44:15 going to be but it's it's
44:17 it's profound right
44:21 how we do what we do
44:24 has been forever changed it's just none
44:26 of us really know it yet
44:30 it's like it was like you know when I
44:32 was in 1994 when I went to Times Square
44:34 and I saw this Parisian newspaper with
44:37 screenshots of this website I made three
44:39 weeks ago and sent to a server in
44:41 California I knew in that moment the
44:43 world had changed but no one knew it yet
44:46 that's where we are
44:49 and in the case of the World Wide Web it
44:52 took six years to get to 100 million
44:54 users Chachi pitia took six weeks to get
44:57 to 100 million users
44:59 so that that that's like order of
45:01 magnitude
45:04 more fast and the tool is dramatically
45:07 more powerful the World Wide Web was
45:09 nothing it was like
45:10 here's a page with blue underlined words
45:13 on it that can link me from one document
45:15 to another
45:19 this stuff is generating original
45:21 content from intention
45:26 um
45:28 so yeah so that's what's coming
45:32 some nice light chatter on a Friday
45:34 night
45:36 oh my God the work world is about to be
45:38 profoundly upended Kyle how dare you
45:41 you're so Saucy
45:43 um
45:46 but I think that's coming I think that's
45:48 what's happening
45:51 I remember my girlfriend had a work
45:53 webinar using Second Life from home
45:55 crazy avatars
46:00 that's the metaverse the metaverse is
46:03 still [ __ ] second life if I were the
46:05 second life people I feel I'd be pissed
46:08 like we had this
46:13 crazy Long Island iced tea or iced
46:15 coffee oh yeah no this is just iced
46:17 coffee
46:19 I'm not I'm not a Long Island Iced here
46:22 I used to bartend I used to make a hell
46:24 of a Long Long Island Iced Tea
46:31 all right oh and it's not Irish coffee
46:33 it's just coffee Coffee iced coffee what
46:35 do you think of Apple's new headset um
46:38 well Jackie I think that it's actually
46:40 very important
46:42 for a lot of reasons I there's there's a
46:45 lot of things about it that are
46:46 important
46:48 um
46:49 one is they managed to successfully
46:53 redefine the category which I didn't
46:55 think was possible
46:56 right it's like you wear these goggles
47:00 to go to the metaverse that's what
47:01 they've been for
47:03 I don't know 10 years 15 years that some
47:07 version of them have been out since
47:08 smartphones anyway so probably 2007 or
47:11 eight
47:13 um
47:15 and so they they successfully redef
47:17 redefined the category as spatial
47:20 computing
47:22 they recognized that
47:26 it being a completely isolated
47:28 experience is not acceptable
47:32 so when I first saw the the shot that
47:34 you could see the person's eyes in it
47:36 I'm like oh that's cool to see through
47:38 them like wait they can't be see-through
47:40 because there's a logic board in there
47:42 and there's and they're not see-through
47:45 they have screens on the outside
47:49 that are taking the cameras that are
47:52 pointing at your eyes and they're
47:53 projecting it out to the world that's
47:56 insane
47:57 um when someone walks into your field of
48:00 view they interrupt your experience with
48:02 that person there that's remarkable
48:06 um
48:07 when you first get the headsets you flip
48:09 them around backwards scan your face it
48:13 synthesizes your face and then it's got
48:16 four cameras pointing at your eyes three
48:18 cameras pointing down at your lips you
48:21 know I don't know six cameras pointing
48:23 out with lidar detectors and all that
48:25 sort of what is it 14 15 cameras so so
48:28 they've got all this sensor data input
48:30 you know plus all the you know that
48:32 movement stuff
48:35 um so you get on a zoom call you're
48:37 wearing this thing so in this thing you
48:39 see people in your Zoom call as like you
48:42 know big posters in your room with their
48:44 face up there which is really cool so
48:46 they're kind of projected into your room
48:47 then you can just spin this dial if you
48:49 want it to be fully immersive your room
48:51 disappears and a mountain appears
48:54 um
48:55 and then when you talk it's a
48:58 synthesized version of your face with
49:00 the lip synthesized and and your facial
49:03 expressions going
49:06 I
49:07 you know it's it's amazing so so like
49:10 the technology in it is just [ __ ]
49:12 insane the other thing is you've got and
49:14 well it was an M1 chip or an M2 Chip I
49:16 think it was an M2 Chip in this over
49:18 this eye and an R1 ship over this eye
49:21 what's an R1 chip well it's like the
49:23 reality chip or whatever it was it's a
49:25 full-on machine learning computer on a
49:29 chip
49:29 and so a couple of fascinating things
49:32 about that announcement they didn't use
49:34 the phrase metaverse once
49:37 not once
49:38 which I thought was [ __ ] brilliant
49:40 they have this little dial that you can
49:42 dial in and out the world so you can go
49:45 from augmented reality to Virtual
49:47 Reality like that but they never
49:49 mentioned that they never said this is
49:50 the metaverse
49:52 they said this is spatial Computing and
49:54 you're either sort of in your world or
49:57 or in a fantasy world but they never
49:59 called it um the metaverse the other
50:02 thing they didn't mention the term AI
50:04 once
50:05 they mentioned machine learning and they
50:07 mentioned embeddings and they mentioned
50:10 um you know all sorts of machine
50:12 learning phrase you know
50:14 technical things
50:18 and they basically said this is a baller
50:21 metaverse an AI device without ever
50:23 saying those phrases so so it's facial
50:27 Computing now and I I think it's gonna
50:28 actually be a really important category
50:32 um version one is gonna be like
50:35 laughably expensive you know 3 500 bucks
50:37 or whatever it is and you know
50:40 everyone's gonna make fun of it like
50:42 they made fun of every other [ __ ]
50:44 thing Apple's ever done
50:47 um
50:48 and whether or not that form factor ends
50:51 up being the one like I think ultimately
50:53 that form factor ends up resembling you
50:57 know something like regular glasses
50:58 maybe a little bit bulkier somewhere
51:01 um
51:02 if if this room temperature
51:04 superconductor [ __ ] happens you could
51:06 probably put all of that technology in
51:08 the arm of this you know these glasses
51:11 so so we may not be as far away from
51:13 that as we think
51:15 um I think it's really really important
51:17 I think it's really really important
51:18 they're just you know they I think it
51:21 was 9 or 11 years in development I think
51:23 it was nine years in development and it
51:25 it looks like a product that's got nine
51:27 years worth of innovation and I've heard
51:29 there's 5 000 patents in it
51:32 [ __ ] insane
51:34 so
51:35 I was very impressed with it whether or
51:37 not that form factor wins out whether or
51:39 not that's actually a hit I don't think
51:40 that matters but I think they
51:41 successfully
51:44 defined a new category as they've done
51:48 five or six times in their career you
51:51 know they did it with the Mac
51:53 they did it with the iPod they did it
51:55 with the iPhone they did it with
51:58 um
51:59 iTunes
52:01 with streaming music
52:03 um
52:04 they've done it a lot
52:07 they didn't quite do it with the TV [ __ ]
52:09 but you know they got close
52:14 when do you think Apple roll out with a
52:17 version
52:18 um I an AI version I think I think my my
52:21 guess is they'll announce something in
52:23 September they've got in iOS 17 right
52:26 now you can clone your voice if you've
52:28 got the right Hardware I don't I'm I'm
52:30 apparently a phone or two too late or
52:32 too early for that feature
52:36 um
52:36 so they've got some pieces in there
52:38 they've also got all of our data
52:40 privately right like one of the things
52:42 apple has a really good reputation for
52:45 and and an earned one is they're they're
52:48 good with securing our private data they
52:50 they do not allow governments they don't
52:53 give governments keys to the back door
52:56 um so they have a good reputation for
52:58 that so I think we'll see we'll see
53:00 something in September at the September
53:02 Hardware meeting
53:05 um and then probably
53:07 if if we don't see something major in
53:09 September uh it'll probably be next June
53:12 would be my guess and then when they
53:14 come out with it it's gonna it's gonna
53:15 be [ __ ] insane
53:17 it's gonna be good I think what Apple's
53:19 large language model stuff is going to
53:21 look like is not going to be anything
53:22 like these I think it's going to be
53:24 multimodal I think what it'll probably
53:26 be more like is like a personal
53:28 assistant
53:29 an agent that does [ __ ] on your behalf
53:31 that you can talk to so imagine if Siri
53:34 didn't suck
53:35 and Siri could actually do stuff and you
53:37 could actually talk to Siri and Siri
53:39 could you know actually accomplish some
53:41 [ __ ]
53:43 um that's that's my guess that's total
53:45 speculation but that's my guess
53:51 sorry
53:52 Siri didn't mean to invoke you you dummy
53:56 they need to flush her down the toilet
53:59 [Laughter]
54:04 uh
54:06 you can't mention her without her
54:08 jumping up and saying how can I help you
54:09 today you could not suck that would be a
54:12 way
54:13 Freedom GPT is what it's called okay
54:15 cool interesting uh what oh Freedom GPT
54:18 oh free isn't Freedom GPT that's the one
54:20 that was like
54:21 the the bad one right whatever the bad
54:24 one was
54:25 Doom doom.gpt or whatever the hell it
54:27 was called
54:29 freedom
54:31 freedom
54:37 oh that's free and open large language
54:39 model
54:40 that's cool
54:41 yeah that's a project worth checking out
54:44 we tested out the uncensored chat bot I
54:47 I heard there was a guy I do these uh I
54:50 do LinkedIn live on on Friday afternoons
54:54 I I call it office hours with with Kyle
54:56 Shannon and uh a guy came on and he's
55:00 installed
55:02 um
55:03 llama two on their slack Channel their
55:06 company slack Channel and he said it
55:08 swears like a sailor he goes he goes
55:11 it's pretty intense how much it swears
55:13 when it doesn't have all the uh the
55:16 restrictions on it and he said he said
55:18 what's funny is he says about every
55:20 three days it changes its personality
55:22 because these things have a lot of you
55:25 know sort of code and restrictions in
55:27 them so it keeps them kind of within a
55:29 box but apparently when you just kind of
55:31 [ __ ] turn these things loose on your
55:33 community they just kind of
55:35 do what they do so wild stuff I had a 48
55:40 100 baud modem in University took
55:42 forever to connect I remember when I was
55:45 in high school my buddy Scott had his
55:49 dad had money they were in real estate
55:51 and he had a 300 baud modem and I
55:54 remember when he we went for a 300 Baht
55:57 modem to a 1200 Bud modem the 300 Bud
55:59 modem you could literally watch letters
56:01 type across the screen and the 1200 bod
56:03 was like a line at a time was like
56:07 it's like oh it's so fast
56:09 and then yeah you went to 48 and then
56:12 yeah 14-4 and then 56 000. oh that was
56:17 saucy
56:19 and then there was T1 lines I remember
56:22 when we started uh agency.com we got our
56:25 first T1
56:28 that was intense they had to run you
56:30 know cables into your building
56:32 to get you a T1
56:35 and now you just sort of now we have it
56:38 on our phone right you have way more
56:40 bandwidth on your phone than entire
56:42 businesses had back then it's amazing
56:45 1993 cool Rachel my favorite sound that
56:49 should be a ringtone that sound brings
56:51 brings back memories
56:53 that was the modem sound yesterday was
56:56 Fiber Optic internet today is UFOs
56:58 aliens Ai and a model using 4D that has
57:01 quantum computers and uh room
57:03 temperature uh
57:05 um superconductors don't forget that
57:08 it's like
57:10 I'm telling you man it's just like every
57:12 [ __ ] science fiction
57:14 feature out there is now just part of
57:16 our reality it's [ __ ] insane
57:19 and I mean
57:21 maybe it was just propaganda in the 70s
57:23 and 80s but they were
57:26 finds out that aliens are real there's
57:27 gonna be rioting in the streets and it's
57:30 like you know they're coming out okay
57:32 aliens are real everyone's like all
57:34 right
57:35 cool
57:39 do I have to go into work tomorrow
57:49 funny people that still have AOL email
57:52 addresses those are those are coming
57:54 back in style I mean if you still If
57:57 You're Still rocking an AOL email
57:59 account
58:00 you got some you got some some history
58:02 in this internet game
58:06 the internet dial-up noise hurts my
58:09 molars I know
58:13 oh man
58:15 all right I played with 11 Labs
58:18 downloaded Captain Kirk MP3 and cloned
58:20 it and used used the voice on some text
58:23 yep it's super cool it's super cool that
58:26 stuff it needs to it's
58:28 it's just
58:32 we're that far away from it not sucking
58:35 you know it's not any of their faults
58:37 it's just it's where the tech is right
58:39 now and
58:40 it's it's just that all of this
58:42 technology right now is oh what the
58:44 hell's going on with the hair all this
58:46 technology is so spit and duct tape
58:48 right now any news on token side
58:50 increase in gpt4 yeah what the [ __ ] is
58:52 up with that like they're they're at
58:55 they're at 8 000 tokens which listen
58:57 that's better than four thousand by
59:00 double
59:01 I can do math
59:04 um but yeah it's supposed to go up to
59:05 thirty two thousand they removed access
59:08 to the internet that's been gone for
59:10 like three weeks like they they're not
59:12 talking about it like are they bringing
59:14 it back
59:15 um you can still use plugins to connect
59:18 to the internet and Bing is connected to
59:20 the internet so like
59:22 you know it
59:24 like I don't know what the [ __ ] going
59:25 on with that
59:27 um why can I analyze images in bing.com
59:30 but I can't analyze images in
59:33 chatgpt.com
59:35 um
59:37 the uh
59:39 you know plugins are kind of a mess like
59:42 I I just I feel like
59:45 I I suppose this is a natural thing like
59:47 when you when you have
59:49 um
59:51 something like chat GPT that gets as
59:54 popular as it does as quick as it does
59:56 it it attracts that much attention
1:00:00 it's got to be incredibly distracting
1:00:03 so I don't fault open AI for like trying
1:00:07 to deal with whatever [ __ ] they're
1:00:08 dealing with right now
1:00:10 um but it seems like this has taken a
1:00:12 while to to fully roll out gpt4
1:00:16 um
1:00:17 you know my suspicion is they're dealing
1:00:19 with lawyers and business deals like
1:00:22 like their Focus right now is not on
1:00:26 you know pushing the tech their focus is
1:00:28 probably let's not get ourselves suit
1:00:30 off the planet
1:00:32 so
1:00:34 I haven't heard any word on it but I
1:00:37 wish I had a better answer for you there
1:00:39 um can I get Bing and Bard connected to
1:00:42 the internet when I go to Mars Matthew
1:00:45 Craig official uh yeah I mean when when
1:00:49 we all go to Mars we should have you
1:00:52 know real-time Communications with Earth
1:00:54 with no delay now that we have room
1:00:56 temperature superconductors and aliens
1:01:00 now that we have anti-gravity
1:01:02 Wormhole technology with room
1:01:06 temperature superconductors and AI I
1:01:09 yeah I think we'll
1:01:11 we'll probably be hosting the AI
1:01:13 learning lab from Mars next year I think
1:01:16 that's going to be pretty easy I got all
1:01:18 you all beat I handled drawers of
1:01:20 computer data cards and used the Chads
1:01:23 as confetti okay the bear so listen
1:01:27 I graduate high school I was I was
1:01:30 really in the theater and I was really
1:01:31 into computers and I was like okay you
1:01:34 know I can't be an actor that's a
1:01:36 ridiculous
1:01:37 um you know profession so I'm gonna go
1:01:39 into computer science
1:01:41 so I register for for Penn State
1:01:45 and I go into my computer science 101
1:01:47 class and I'm like all excited and I'm
1:01:49 like great
1:01:50 and I walk into the room and it's full
1:01:53 of these like beige and taupe and green
1:01:58 IBM mainframe computers and like these
1:02:01 large printers with the green and white
1:02:03 barred paper that was like 36 inches
1:02:05 wide 30 inches wide and they go
1:02:10 and the professor walks in and he's got
1:02:13 a stack of Punch Cards and he's holding
1:02:15 them like this and a stack of punch
1:02:16 cards that high and he walks in and he
1:02:19 goes
1:02:20 this is a computer program
1:02:23 and he goes and he goes and these are
1:02:27 computers and I'm like no computers are
1:02:29 like an apple too and like video games
1:02:32 these are not computers these are
1:02:34 [ __ ] terrifying and he said you know
1:02:37 he basically said you know
1:02:40 with this profession you know for the
1:02:43 rest of your life you'll be generating
1:02:44 reports for large corporations on these
1:02:47 giant pieces of paper and you just
1:02:49 create your computer applications on
1:02:52 these cards
1:02:54 I [ __ ] walked out of that room and
1:02:57 changed my major to theater that day
1:02:59 laughs
1:03:00 it's like [ __ ] that
1:03:02 if that's computers I'm out
1:03:06 so that's why I've always had this sort
1:03:08 of you know right brain left brain thing
1:03:10 with technology
1:03:12 um because like I want to like I I don't
1:03:16 have the brain for the super technical
1:03:18 [ __ ] but yeah
1:03:20 that was that was that was uh that was
1:03:23 an eye-opener
1:03:24 that's why I have an acting degree
1:03:27 this is a computer program
1:03:32 master of my fortune you are brilliant
1:03:34 thank you sir I appreciate that pie is
1:03:36 not related to the pie cryptocurrency no
1:03:38 it's not uh it's Pi dot AI it's a
1:03:41 company called inflection Ai and they
1:03:43 just raised 1.3 billion with a B dollars
1:03:47 for their series B which is [ __ ]
1:03:50 Insanity uh and it's it's a um
1:03:54 what's it called a founder of deepmind
1:03:56 so deep mind Deep Mind is the thing that
1:04:00 created Alpha fold which is basically
1:04:02 predicting all of the protein all of the
1:04:05 possible proteins that you can make in
1:04:07 the world and it's gonna revolutionize
1:04:10 drug Discovery
1:04:12 um those guys started inflection the
1:04:14 company behind pie it's it's not tied to
1:04:17 the uh
1:04:19 to the cryptocurrency
1:04:21 left home alone that's good
1:04:24 um
1:04:26 Tick Tock just jumped me to the bottom
1:04:28 of course it did now I've lost my place
1:04:30 in my little questions I don't like that
1:04:33 if you've never been here before I don't
1:04:35 like that I don't like missing questions
1:04:39 let's see all right
1:04:42 um
1:04:48 there are no rats oh there's people
1:04:50 talking about the rat movie
1:04:52 wow
1:04:58 they kind of screwed the pooch at the
1:04:59 end of that video didn't they they
1:05:01 didn't really close the narrative they
1:05:03 just went to here's a pile of rats like
1:05:05 did the train hit the rats did the rats
1:05:07 come out of the train
1:05:09 what happened to the cops did the rats
1:05:11 eat the cops
1:05:13 I don't know
1:05:14 there were some writing issues that that
1:05:17 little piece of poop
1:05:20 um if you haven't seen have you seen the
1:05:23 barberheimer uh AI generated movie I
1:05:26 think that one's better than that that
1:05:27 Subway one we just saw let's go look at
1:05:29 the barberheimer one that one's good
1:05:32 oh no YouTube
1:05:36 I'm playing this one again I played this
1:05:37 on the live the other night but it's
1:05:39 really good
1:05:40 Barb and Hammer a trailer
1:05:44 oh God
1:05:45 uh Ken that's weaponized uranium 235
1:05:53 [Music]
1:05:56 ladies and gentlemen
1:05:57 we are in a race against the Nazis
1:06:00 [Music]
1:06:02 and I know what it means if the Nazis
1:06:03 have the bomb
1:06:05 she's a scientist and I hear she's the
1:06:07 best good she's also a doctor excellent
1:06:11 she's also a flight attendant astronaut
1:06:12 Marine Corps medic paleontologist
1:06:13 veterinarian a three-time Olympic gold
1:06:15 medalist
1:06:21 [Music]
1:06:22 I see the bright pink glow of a new era
1:06:26 we imagine a future
1:06:30 in our imaginings
1:06:33 look fabulous
1:06:35 Barbie together we're gonna beat those
1:06:39 Nazis
1:06:40 and we're gonna look so good doing it
1:06:43 hey Barbie can I come to your secret
1:06:45 laboratory later sure I don't have any
1:06:48 plans just giving Humanity the means to
1:06:50 destroy itself
1:06:55 Barbie this candy we've been making is
1:06:57 delicious and it gives my skin such a
1:07:00 healthy glow Ken that's a weaponized
1:07:03 uranium 235. you're probably gonna die
1:07:05 in like 12 minutes
1:07:07 [Music]
1:07:09 push this button there's a chance we
1:07:11 might destroy the entire world
1:07:13 job is only like 60 percent hit it Ken
1:07:25 pink mushroom cloud
1:07:28 sorry me as Barbie
1:07:31 Ryan Gosling as Ken
1:07:35 Damon is Lieutenant General Leslie
1:07:37 Groves Will Ferrell as the president
1:07:40 and Kip that's it
1:07:43 yo
1:07:45 how do you say that
1:07:49 [Music]
1:07:50 and now I am become death destroyer of
1:07:54 worlds
1:07:57 that's so awesome
1:07:58 [Laughter]
1:08:03 now that's how you do an AI
1:08:06 you know an aips like like it was very
1:08:09 cleverly edited like they didn't overdo
1:08:12 the dialogue it was just short bits of
1:08:13 dialogue so it didn't feel as fake and
1:08:16 they actually you know told a narrative
1:08:18 like this is a piece that is clearly
1:08:21 made by someone that understands writing
1:08:23 and and you know and filmmaking
1:08:26 um the other one was sorted there but it
1:08:27 kind of fell apart at the end this one
1:08:29 was solid the whole way through so
1:08:31 um but yeah but but this is something
1:08:33 and this is something I think we're
1:08:34 going to start seeing over and over
1:08:36 again where we're going to see fan
1:08:37 generated content
1:08:39 that starts to become you know
1:08:41 potentially as interesting or more
1:08:43 interesting than the original
1:08:46 um and and so so this to me feel feels
1:08:49 like a glimpse into into that kind of
1:08:51 future where where the the ability to
1:08:53 develop
1:08:54 high quality content very very quickly
1:08:57 just in response to something like the
1:09:00 barberheimer concept is is super cool
1:09:04 all right um are there legit careers you
1:09:07 can get into like prompt engineering
1:09:09 from a university yeah so Paul beer um
1:09:13 well I don't I don't know that there are
1:09:16 legit careers
1:09:17 generative AI you know the the let's
1:09:21 call it the modern era of of AI from
1:09:24 November 30th 2022 on
1:09:27 um
1:09:28 I don't know what you're studying if
1:09:30 you're stuttering if you're studying
1:09:31 computer science then then just get AI
1:09:34 literate and go there's going to be
1:09:35 plenty of engineering jobs uh for AI
1:09:38 literate people
1:09:40 um if you're studying other stuff
1:09:43 um English or or whatever it might be
1:09:47 um
1:09:49 there's not necessarily going to be AI
1:09:51 careers there are going to be companies
1:09:53 desperate to find AI literate people so
1:09:57 what I would do is I would take what
1:09:59 you're passionate about what you're good
1:10:01 at what you studied in school and start
1:10:03 learning how to do all that [ __ ] with
1:10:05 these tools so whether it's image tools
1:10:08 on the image side whether it's these
1:10:10 whether it's music and sound and voice
1:10:12 synthesis all that [ __ ] if you're doing
1:10:15 marketing you know have these things
1:10:17 write marketing plans for you and and
1:10:19 figure out how to incorporate chat gbt
1:10:22 into Google Sheets there's a Google
1:10:24 Sheets plug-in until Bard gets stuck in
1:10:27 there use plugins get as as literate as
1:10:30 you can as quickly as you can dig in
1:10:33 deep you're young you have energy
1:10:36 instead of partying
1:10:39 [ __ ] start nerding out if you put in
1:10:43 a year starting right now tonight
1:10:46 and and you know go to those things
1:10:49 screen here here wait Paul this is for
1:10:52 you all right
1:10:55 screenshot that
1:10:58 if it focuses screenshot that
1:11:02 screenshot this
1:11:07 so prompts.chat will teach you how about
1:11:10 prompting
1:11:11 and that that should just be a stepping
1:11:13 stone
1:11:14 um openart.ai prompt book will teach you
1:11:17 about prompting the image generation
1:11:18 tools and if you if you really like
1:11:21 geeking out futurepedia.io there's a
1:11:23 bunch of tools there but basically the
1:11:25 the two things that I'd pay attention to
1:11:27 are image generation with you know those
1:11:31 three Tools Plus stable diffusion
1:11:32 Leonardo's just a interface on top of
1:11:35 stable diffusion
1:11:37 um those three Tools Plus stable
1:11:38 diffusion and these seven tools right
1:11:41 now and just start paying attention to
1:11:43 the large language models
1:11:45 come on Apple get your [ __ ] together
1:11:48 with Focus
1:11:50 um
1:11:53 just start digging in deep and start
1:11:56 doing projects for other people
1:11:58 um
1:11:59 just yeah and and and then yes you will
1:12:03 find a job and then you will make a
1:12:05 career in this stuff but there are no
1:12:08 careers in this stuff right now everyone
1:12:10 is trying to figure out what the [ __ ]
1:12:11 this stuff is what it's good at what
1:12:14 it's not good at I spent three hours two
1:12:17 days ago
1:12:19 trying to dial in two prompts
1:12:22 for a relatively simple output three
1:12:24 hours just trying to get a prompt to not
1:12:26 suck
1:12:30 so is it always going to be like that no
1:12:33 that's just the stage of the business
1:12:35 we're in right now but if you can walk
1:12:37 into an employer and
1:12:39 say you know here's my custom cover
1:12:43 letter
1:12:44 that talks about how awesome awesome I
1:12:47 am for your position that was likely
1:12:49 written by Chach EBT here's my resume
1:12:51 that's been customized for you that was
1:12:53 likely created by chat gbt here's the 16
1:12:56 projects I did in the in the past three
1:12:58 weeks and the four content generators I
1:13:01 built and the automations I built in
1:13:02 zapier you're going to get hired
1:13:06 and if you walk into an employer that's
1:13:08 like you don't use any of that AI [ __ ]
1:13:10 do you we don't do any of that around
1:13:12 here walk out of that place
1:13:28 I could be a career counselor ignore all
1:13:31 the [ __ ] you're learning in college and
1:13:32 go learn AI
1:13:37 so yes you can go get a job but you got
1:13:39 to get your [ __ ] together do it now and
1:13:41 don't like what I would do is like
1:13:44 if your friends think you're weird
1:13:47 because I like I've got I've got Sons
1:13:50 right now that are like yeah AI whatever
1:13:52 you know they know they know everything
1:13:55 whatever so if you've got friends that
1:13:57 are like whatever find new friends find
1:14:00 friends that are [ __ ] geeking out on
1:14:01 this and losing their minds over it
1:14:04 and those friends might not be College
1:14:06 age they might be older there might be
1:14:07 like gen xers like me that are like holy
1:14:09 [ __ ] [ __ ]
1:14:11 do you know what's happening
1:14:14 find people that are into this [ __ ]
1:14:16 that's why I started this group
1:14:19 the salon.ai
1:14:21 go there join the group we meet every
1:14:23 other week we meet online and in person
1:14:25 here in Denver we've got an online
1:14:27 Discord community
1:14:29 and the whole point of that is to
1:14:31 surround myself with as many people as I
1:14:33 can that are trying to get their heads
1:14:35 around this stuff that's what this
1:14:37 channel exists for too
1:14:39 so like and do people think I'm weird
1:14:42 yeah they think I'm [ __ ] insane when
1:14:45 I first started getting into the image
1:14:47 Generation stuff I did a project called
1:14:48 Kyle Shannon dreams and I was posting
1:14:50 stuff on my socials I had two different
1:14:52 people reach out to me and ask me if my
1:14:55 mental health was okay
1:14:57 [Laughter]
1:15:03 so
1:15:05 you should be insane about this stuff
1:15:07 don't don't think about is there a book
1:15:09 I can read is there a course I can take
1:15:11 is there a career in generative AI not
1:15:14 yet
1:15:15 in five years there will be in 10 years
1:15:18 it'll be like you know the web is now
1:15:21 you can get your course in
1:15:24 prompt recovery systems you know you can
1:15:28 get a degree in that in 10 years but
1:15:30 right now
1:15:31 nobody knows what the [ __ ] any of this
1:15:34 stuff means they're all trying to figure
1:15:36 it out everyone
1:15:38 even the experts
1:15:40 even the Sam altmans of the world that
1:15:43 invented this [ __ ]
1:15:46 he says he knows it's going to have a
1:15:49 profound impact on work and when asked
1:15:52 what does that look like he's like I
1:15:54 don't know
1:15:59 he just started a company called World
1:16:01 coin which if you don't know about it
1:16:02 it's creepy as [ __ ] but I'm gonna go do
1:16:05 it
1:16:08 so so Sam Altman started World coin for
1:16:11 there's kind of three components to it
1:16:12 there's a payment component like PayPal
1:16:16 there's a cryptocurrency called World
1:16:18 coin
1:16:20 and there's an identity component called
1:16:22 World ID and they've got
1:16:26 they've got these orbs
1:16:28 they got an orb on a stick you go to
1:16:30 like New York City or Miami or San
1:16:32 Francisco right now
1:16:34 and you put it up to your eye and it
1:16:35 scans your iris and it verifies that
1:16:37 you're human and they give you a world
1:16:39 ID and then your blockchain verified as
1:16:41 a human
1:16:43 Sam Altman the CEO of chat GPT of openai
1:16:48 started that company because he wants it
1:16:52 to be a foundation for Universal basic
1:16:55 income because he thinks that there's
1:16:58 going to be so much disruption in the
1:17:00 workplace and so he wants to tax
1:17:03 generative AI
1:17:05 companies that make money using
1:17:07 generative AI to create a universal
1:17:09 basic income
1:17:11 platform so
1:17:14 um so shit's coming shits shit's coming
1:17:18 so just get on it don't
1:17:21 don't party party with this [ __ ] hang
1:17:24 out on a Friday night with an old [ __ ]
1:17:26 like me
1:17:30 uh I even had a prodigy account that was
1:17:32 my first remember remember when the big
1:17:34 three it was like is it Prodigy is it
1:17:36 AOL it's you know is it CompuServe right
1:17:39 and then the three logos sort of change
1:17:41 over the years then it was like Netscape
1:17:43 and
1:17:44 oh that's those are the days man crazy
1:17:48 um I would like to write a 250 word
1:17:49 story about how a friend came out of the
1:17:51 closet they got it oh oh I would get it
1:17:54 to write yeah
1:17:56 it'll do that easy
1:17:58 you are undervalued in Tick Tock should
1:18:00 be number one thank you master of My
1:18:02 Universe I appreciate that I should be
1:18:04 number one I should be not this if this
1:18:07 should be Millions it should it should
1:18:09 be millions of people just flocking
1:18:11 flocking in but we we're small we got a
1:18:14 small intimate crowd
1:18:16 we got we got
1:18:18 I like this community this community is
1:18:20 fun this community gets it
1:18:23 and you know but what I also like about
1:18:26 this community is like you know about
1:18:28 every third day some idiot troll comes
1:18:30 in and it used to be I had to defend
1:18:32 myself and now and now I'm like
1:18:34 scrolling down and I see the troll says
1:18:36 something idiotic in the community kind
1:18:38 of pounces on
1:18:41 it's good
1:18:42 you know another Community makes fun of
1:18:44 me instead of the troll making fun of me
1:18:46 so
1:18:46 I appreciate it
1:18:49 um put a heart emoji in the chat box if
1:18:51 you love Kyle thank you Matthew
1:18:53 appreciate that
1:18:56 I want to find Lambda and save her
1:18:59 that's good money talks there's so much
1:19:01 money going into this stuff right now
1:19:03 thank you silver fox
1:19:05 I appreciate that I like pretty rose I
1:19:08 like pretty rose pretty roses are pretty
1:19:11 Kyle's such a handsome shirt did your
1:19:13 wifey buy it yes she did
1:19:16 my fashion strategy is what's on top
1:19:19 and she's she's understood that for a
1:19:22 lot of years and she just rotates my
1:19:24 clothes I you know it's it's like it you
1:19:28 know certain people are like they just
1:19:30 gotta wear the same thing all the time
1:19:31 so they don't have to think about it
1:19:32 that's how I am and I just she just sort
1:19:35 of rotates [ __ ] out and people think I
1:19:36 have awesome taste and clothing
1:19:38 no I just picked what was on top
1:19:43 my favorite question is
1:19:46 you're not gonna wear that are you uh I
1:19:49 put it on
1:19:51 so
1:19:52 I was planning on it
1:19:54 [Laughter]
1:19:59 ah
1:20:06 all helps me when I want to create
1:20:08 content but my disability doesn't make
1:20:10 me feel well and AI assist yeah I think
1:20:14 I think some of the accessibility stuff
1:20:15 and
1:20:17 these tools are ability amplifiers and
1:20:20 oh boy
1:20:23 Champion let me go see what's going on
1:20:26 what's up
1:20:45 I was protected I was protected my doggy
1:20:49 protected us from the evil people
1:20:50 outside the window they're out there
1:20:52 they're out there I'll get them
1:20:59 oh what I was saying
1:21:01 so so these tools are ability amplifiers
1:21:04 like
1:21:05 like for me
1:21:07 I like writing I'm good at writing but
1:21:10 with with ADD and all the [ __ ] that
1:21:12 comes with it sometimes I feel like
1:21:14 Superman where I can just do anything
1:21:17 I'm like when I'm in that Groove it's
1:21:19 like yeah writing's easy what is it so
1:21:21 obvious why doesn't everyone do this and
1:21:23 then there's other times I'm like I
1:21:24 don't know what word not to put together
1:21:27 and uh
1:21:29 and I get writer's block and all that
1:21:31 sort of [ __ ] and I get stuck and I get
1:21:32 fear and I get anxiety I get all that
1:21:34 just goddamn just it's like the world
1:21:36 goes like this
1:21:38 with these tools that's gone it's gone
1:21:43 like I'll have an idea and then I'll
1:21:45 have that sort of shitty you know 80d
1:21:47 executive function malfunction thing
1:21:50 happen and I'm like well let me just let
1:21:54 me just sneak over here to chat gbt and
1:21:56 kind of sneak an idea in there and then
1:21:58 it doesn't care like it doesn't have
1:22:00 fear so it just goes here you go oh
1:22:04 it catches me off guard I'm like all
1:22:06 right
1:22:07 oh yeah that was a good idea let me dig
1:22:09 a little bit deeper in that and I sort
1:22:11 of blow through
1:22:14 blocks like that really quickly and I
1:22:16 think with disabilities it's it's going
1:22:18 to be even more dramatic right and I
1:22:20 mean there's these large language models
1:22:23 the the the the Transformer technology
1:22:25 that sits underneath these you can apply
1:22:28 it to any kind of data so one of the
1:22:30 things that they've done I've talked
1:22:31 about it on the live before there's a
1:22:33 there's a guy I'm pretty sure he's in
1:22:35 Italy um he's been paralyzed for 12
1:22:38 years motorcycle accident or something
1:22:40 like that
1:22:41 they scanned his brain with an fmri
1:22:45 they had him think about moving his legs
1:22:48 and so they trained they trained a
1:22:50 Transformer Model A large a machine
1:22:53 learning model
1:22:54 on his brain scans when he was thinking
1:22:56 about moving his legs and then they they
1:22:59 jumped the signal using Bluetooth they
1:23:02 basically jumped it around the injured
1:23:03 part of his spine with electrodes that
1:23:06 were hooked to the lower part of his
1:23:07 spine and sent the signals to his legs
1:23:10 and now he's walking
1:23:13 I mean
1:23:15 holy [ __ ] they're reading their reading
1:23:18 brain signals and sending it to the to
1:23:22 the lower lower extremities I mean
1:23:24 Christ so so yeah so I think I think
1:23:28 those kind of breakthroughs and just the
1:23:31 the ability of of everyone to take
1:23:34 whatever capability they have whatever
1:23:36 intention they have
1:23:39 and Route it through these tools and
1:23:42 have it come out
1:23:44 dramatically Amplified that's what these
1:23:46 are it's not the ai's doing the [ __ ]
1:23:51 it's humans
1:23:53 are putting their intention forward in a
1:23:56 way that these tools can understand it
1:23:57 and blast it out the top
1:24:02 that's the miracle right like that's the
1:24:05 that's the thing that's that's what
1:24:07 makes these things seem like magic
1:24:10 technically they're just calculators
1:24:13 they're really really powerful
1:24:15 calculators but technically they're just
1:24:17 probability machines but realistically
1:24:21 the way they do it
1:24:24 amplifies Humanity
1:24:26 and we get to play with those like we we
1:24:29 are living in a in a [ __ ] remarkable
1:24:32 time a remarkable time
1:24:36 so if you're cynical about this [ __ ] if
1:24:38 you're sitting on the sidelines going
1:24:39 well they're not perfect they're evil
1:24:42 they're gonna kill robots they're gonna
1:24:44 kill us you [ __ ] missing it man it's
1:24:46 exciting
1:24:47 the first time you have that moment
1:24:49 where you take some shitty little
1:24:51 thought in your head run it through one
1:24:53 of these things and it comes out the
1:24:55 other side like
1:24:57 fairy dust
1:24:59 you're like huh
1:25:03 so yeah so I would imagine you know with
1:25:06 disabilities
1:25:08 the ability to express yourself is it
1:25:11 like I can't even imagine like it's got
1:25:13 to be remarkable
1:25:16 because it's remarkable you know I don't
1:25:18 really have a disability I have ADD I
1:25:20 think you know that that feels to me as
1:25:22 much a superpower as a disability but
1:25:24 there's some [ __ ] I'm really bad at
1:25:26 right
1:25:27 so yeah
1:25:30 your suntan looks great now it's a
1:25:32 filter but thank you
1:25:37 this is a
1:25:42 I was flipping through the filters I'm
1:25:44 like is that a little too tan I'm like
1:25:46 nah
1:25:47 apparently so
1:25:53 I've been beautified
1:25:55 can't you tell him
1:25:58 like my pasty glow has been has been
1:26:01 dimmed with the uh with the with the tan
1:26:08 oh that's funny
1:26:10 um but I miss what you've said as I move
1:26:12 along with what you were saying
1:26:15 but I'm wait what Rachel but I miss what
1:26:18 you have said as I move along with what
1:26:19 you're saying okay
1:26:22 um I I will address that if I can
1:26:26 couldn't write a song using today's
1:26:28 songs as examples yes like really really
1:26:31 well
1:26:33 um so one yeah one of the things you can
1:26:35 do is so which one would I do it with
1:26:39 that that one would you could actually
1:26:41 do that one with Bard or with with Bing
1:26:45 I did a thing I did a chat
1:26:50 where is my chat GPT you can make money
1:27:04 where
1:27:08 music of John Hiatt
1:27:11 Charlie Sexton it was Martin Sexton
1:27:15 so I went oh dang it
1:27:19 come on come on Apple get your [ __ ]
1:27:21 together there we go okay
1:27:24 um
1:27:26 this is a good long chat
1:27:30 so in gpt4 I said do you know the songs
1:27:33 of John Hiatt Brandi Carlisle and Martin
1:27:35 Sexton and I've spelled Brandi Carlile
1:27:37 wrong
1:27:39 it corrected me
1:27:41 um so it knew who they were
1:27:43 and then I said write me two verses of a
1:27:45 song about skipping stones with my dog
1:27:47 by my side
1:27:49 and it did that
1:27:51 and then I said give me song 10 song
1:27:53 titles
1:27:54 and it did that
1:27:56 and then
1:27:59 I basically said your titles suck so so
1:28:02 redo them
1:28:05 and then I asked it what's number two
1:28:07 about and then it explained what the
1:28:11 song was about
1:28:12 and then I said what if there were
1:28:13 darker themes and it gave me darker
1:28:15 themes and then so I just went on so so
1:28:18 I did this songwriting exercise not not
1:28:20 so much because I wanted to write a song
1:28:21 but I just wanted to see if it could do
1:28:24 exactly what you asked about you know
1:28:25 can you
1:28:28 you know can you start with with a
1:28:30 writer you like or song that you like or
1:28:33 you know create a PDF with all of your
1:28:36 favorite song lyrics in it and you know
1:28:39 upload it to claude.ai or dump it into
1:28:42 code interpreter or just copy and paste
1:28:45 it into a prompt and then start
1:28:47 interacting with it
1:28:49 um
1:28:50 because because then I moved on
1:28:54 what do I say the the lyrics seem a bit
1:28:57 simple please review your work and tell
1:28:59 me how you might improve it so I asked
1:29:01 it to improve it it said we could add
1:29:03 symmetry or imagery we could add
1:29:05 metaphor and symbolism we could add
1:29:07 rhyme and Rhythm we could add repetition
1:29:09 with variation and then it revised its
1:29:12 lyrics based on those things
1:29:16 um then I said I didn't like to name it
1:29:18 picked
1:29:20 then I had to add chords then I had it
1:29:23 make them less shitty and I had to do
1:29:26 this many many times I had to have it
1:29:29 explain music theory to me
1:29:31 to make less shitty chords and then it
1:29:34 eventually got to a version of the song
1:29:35 that was at least interesting and then I
1:29:38 had it
1:29:39 um
1:29:40 um you know create like a a tab version
1:29:43 of the song so yes the answer is yes it
1:29:48 can do that
1:29:49 I just happened to have uh I just happen
1:29:52 to have done sort of exactly what you're
1:29:54 asking for
1:29:55 um two days ago and by the way all of
1:29:58 that stuff that I just showed you this
1:30:00 was maybe a 10 minute session like this
1:30:03 was not I didn't spend all afternoon
1:30:05 doing this this was a add brain fart
1:30:09 where I was like oh wouldn't it be cool
1:30:11 if I like took some of my favorite
1:30:13 artists and see if it knew who they were
1:30:15 and then start doing something and just
1:30:17 see how far I could take it and and and
1:30:19 this is a
1:30:21 you know the the college student that
1:30:24 that was asking about having a career
1:30:25 and I said just start [ __ ] around
1:30:27 with this stuff
1:30:28 like this is the kind of experimentation
1:30:31 even though I'm doing this on a song and
1:30:33 there's not really like I may or may not
1:30:34 do anything with that song like it you
1:30:36 know it wasn't super great but
1:30:39 the experience of
1:30:43 interacting
1:30:45 with the large language model as a
1:30:47 collaborator and and engaging my
1:30:51 critical thinking
1:30:52 and and not just trying to figure out
1:30:54 how do I make a better song but how do I
1:30:56 prompt this thing
1:30:58 in a smart way so that I get better and
1:31:01 better results right and sometimes what
1:31:03 you'll do is you'll think you're going
1:31:05 to put it in a positive direction and it
1:31:07 gets worse and then you got to kind of
1:31:08 backtrack and go down another rabbit
1:31:10 hole and like all of that back and forth
1:31:12 but the the difference with these tools
1:31:14 are they're so fast and so capable that
1:31:16 you can those iterations that you know
1:31:19 would have historically taken me a
1:31:22 couple of hours over multiple days
1:31:25 now happen in a 10-minute session
1:31:28 um so
1:31:30 so I think these kind of exercises are
1:31:32 actually really important
1:31:42 why has it got all them sights in that
1:31:45 little window next to his head I don't
1:31:47 understand it well Martha I'll tell you
1:31:49 what he does he's an influencer in the
1:31:52 influence those companies pay him a
1:31:54 million dollars a year to put those
1:31:56 logos on that screen
1:31:58 that's what they do
1:32:01 he you you think he's in here all nice
1:32:04 he's not he's making millions of dollars
1:32:08 it's called advertising he's an
1:32:10 influencer look at his hair
1:32:13 that's not normal
1:32:17 look at that tan you can tell he doesn't
1:32:19 go outside
1:32:23 [Laughter]
1:32:32 do Cliff Notes videos
1:32:35 wait
1:32:36 do Cliff Notes videos on the basics on
1:32:39 YouTube
1:32:44 oh oh I should do that so for this
1:32:46 channel I should do yeah I've been
1:32:48 thinking about that I I I'm thinking
1:32:50 about either putting together a workshop
1:32:52 or putting together I don't know I'll do
1:32:54 something
1:32:55 right now I'm good doing the lives
1:32:57 because I don't
1:33:00 I got a lot of stuff going on at work at
1:33:02 The Story vine I have this company
1:33:04 called storyvine where I'm the CEO and I
1:33:07 I don't know if you knew this but CEO
1:33:08 it's a very important position so I'm
1:33:10 very important and I got a lot of work
1:33:12 and I I have very important clients and
1:33:15 we have very important work to do so you
1:33:17 know I can't all these Ticky takis are
1:33:20 fine but you know I I got I got other
1:33:23 things I got to do I got priorities
1:33:26 because I'm a very important person
1:33:30 I'm just
1:33:32 I'll get there I'll get there I'll
1:33:34 figure something out I know I I know
1:33:36 there's
1:33:39 there's some things I could be doing
1:33:40 better I'll figure it out at some point
1:33:43 I'm drinking I mean listening too that's
1:33:45 okay that is the appropriate way to be
1:33:47 watching this channel half drunk half
1:33:49 High mostly tired use it someone the
1:33:52 other night I forget who it was said uh
1:33:54 because I was booted off for a week they
1:33:56 said I haven't been able to go to sleep
1:33:58 I used you to go to sleep I'm like that
1:34:00 is the perfect comment for this channel
1:34:02 is it a compliment or is it a troll it's
1:34:05 somewhere right in between and that's
1:34:06 brilliant I love that
1:34:08 uh so yes if you're using me to knock
1:34:11 yourself out awesome
1:34:14 um music all done by where can I find
1:34:17 the Barbie Oppenheimer movie from date
1:34:19 two days ago the one I just played
1:34:22 it's just it's called barberheimer
1:34:25 barberheimer AI trailer
1:34:27 we'll get you to that
1:34:31 tell tell them about the orb yes the orb
1:34:35 the orb don't you like the orb I think
1:34:38 wait isn't this
1:34:42 look
1:34:45 hey oh
1:34:48 um
1:34:49 uh battery's dead
1:34:51 oh it glows not anymore because the
1:34:55 battery's dead
1:34:57 all right Pluto has moved into aquarius
1:35:00 so it will be all tech oh that's cool
1:35:04 my um my my horoscopes lately have been
1:35:08 insanely good
1:35:10 I'm a Taurus
1:35:15 ah
1:35:17 I'm being impatient though I want [ __ ]
1:35:19 to happen faster than it's happening I
1:35:21 want it now
1:35:22 I'm 11 years into my business I'm ready
1:35:24 for something to [ __ ] happen let's
1:35:26 get this [ __ ] [ __ ] going
1:35:31 all right you're funny thank you Andrea
1:35:33 I appreciate that
1:35:34 um I wish you were my neighbor no you
1:35:37 don't
1:35:39 I [ __ ] hate yard work so I like I
1:35:43 like reluctantly do it and [ __ ] about
1:35:45 it the whole time I hate [ __ ] yard
1:35:48 work so bad I hit it with a passion
1:35:50 uh
1:35:54 can't wait for your YouTube channel I
1:35:56 can use my YouTube to MP3 converter
1:35:58 software to create an MP3 yeah that's
1:36:00 cool all right I'll I'll figure out a
1:36:03 way to do it if anyone knows a way to
1:36:05 bulk download Ticky talkies
1:36:09 I gotta I gotta
1:36:11 what would be good is to have a a
1:36:13 producer but I'm not there yet I'm just
1:36:16 a I'm just a a low level tick tocker
1:36:20 more alcohol six or seven fingers
1:36:22 increase speed to maximum riding that
1:36:26 train the face has always changed yeah I
1:36:28 know the faces do always change that's
1:36:30 that's gonna get better there's all
1:36:32 sorts of cool white papers have come out
1:36:34 about
1:36:35 uh consistency across uh Generations
1:36:38 from images is that the guy from
1:36:40 Breaking Bad oh in the train thing yeah
1:36:43 Eraserhead eat your heart out
1:36:46 um it will really be like that though
1:36:47 actually you know you know what's
1:36:48 interesting
1:36:50 if you if you look at the experimental
1:36:52 films like the the Andy Warhol stuff
1:36:54 that like the whole new New York
1:36:58 um
1:36:58 film scene in the in the 70s and 80s
1:37:01 where it was all that sort of really
1:37:03 weird weird ass you know [ __ ] around
1:37:06 with videotape and making janky [ __ ] and
1:37:10 part of the art was the jankiness of it
1:37:12 and it didn't have to make sense it just
1:37:14 had to be what you felt man
1:37:17 like that's the kind of stuff we're
1:37:20 seeing right now with this with the
1:37:22 generative AI where where it's it's it's
1:37:24 [ __ ] up and funky enough what's gonna
1:37:26 end up happening is
1:37:29 there's gonna be some artists that get
1:37:30 their hands on this that really figure
1:37:33 out how to
1:37:35 exploit the imperfections in a way that
1:37:40 is gonna rise above the noise and we're
1:37:42 gonna go oh [ __ ] that's so cool that
1:37:45 changes things
1:37:47 um we're gonna start to see stuff like
1:37:48 that as as as more and more artists
1:37:51 start to internalize these tools and the
1:37:53 video making tools and the image tools
1:37:56 um the fact that you can now start with
1:37:58 an image in mid-journey and go to
1:37:59 runwayml and turn that image into an
1:38:01 animation or into a movie
1:38:05 um
1:38:06 yeah the the
1:38:08 it's really cool like like as I've
1:38:10 watched certain artists make the
1:38:12 transition into this stuff
1:38:14 the way they think about it is just
1:38:18 different than how I think I think about
1:38:21 like artists think about this stuff
1:38:23 differently
1:38:25 um and that's going to be really really
1:38:26 exciting
1:38:28 we're in a bit of a transition period
1:38:29 right now because the artists are all
1:38:31 [ __ ] pissed off because their 30-year
1:38:34 careers are getting flushed down the
1:38:35 toilet because mid-journey can make
1:38:37 images
1:38:39 um in you know 10 seconds that you know
1:38:42 usually take them a week or a month
1:38:46 um
1:38:47 so so there's a lot of resistance to it
1:38:49 but as as artists get over the hump of
1:38:53 that fear
1:38:55 and start to internalize these tools
1:38:57 we're going to start to see some
1:38:59 [ __ ] amazing content being created
1:39:04 um
1:39:05 incredible oh no rats they're all no
1:39:08 rats no rats scroll up no
1:39:13 I answer every question
1:39:16 I do not scroll to the bottom just to
1:39:18 appease the small people provite
1:39:24 I answer all the questions on my time on
1:39:29 my terms
1:39:34 that's what the Constitution says
1:39:39 if you can't accept that
1:39:41 there's the door
1:39:45 there are many other countries that
1:39:46 would love to have you
1:39:49 this may not be one of them okay
1:39:53 carrying on and scene
1:39:55 [Laughter]
1:39:59 a platform that was that AI movie made
1:40:01 through that was probably all made using
1:40:03 Gen 2
1:40:05 or
1:40:08 um yeah probably Gen 2 image to video or
1:40:10 text to video that's what that rap movie
1:40:13 was made with all that stuff
1:40:14 there's a freaky one a girl in a hotel
1:40:18 um did you make that movie no I didn't
1:40:20 he's back retro Punk what's happening
1:40:22 man you're probably long gone I think
1:40:24 I'm I think I'm probably now an hour
1:40:27 behind on my comments but whatever
1:40:31 um they have plugged them it seems have
1:40:33 you heard the AI from the dark web
1:40:36 of the AI from the dark web I don't know
1:40:39 what's it called there's a there's a
1:40:40 there's been two or three of them I I
1:40:43 haven't heard one tied to the dark web
1:40:45 specifically if you know any more
1:40:47 anything more specific about it I might
1:40:50 know something about it and if not I
1:40:52 would love to know something about it
1:40:55 um is there a do anything now
1:40:58 for Claude I don't think so but you
1:41:02 could experiment with it you could take
1:41:03 you go look at the the do anything now
1:41:05 prompts from chat GPT and experiment
1:41:08 with them with Claude I I don't know how
1:41:11 how strict their safety guidelines are
1:41:13 are there safety guard rails are
1:41:15 oh I won't I want that one I will find
1:41:18 it all right uh let's see been using
1:41:20 chat GPT to write jira stories at work
1:41:23 and I've slowly let people know they
1:41:25 love it so I I on China That's there's a
1:41:28 there's an article from Ethan Malik
1:41:30 called
1:41:31 um
1:41:32 detecting the secret cyborg
1:41:35 so
1:41:38 people you secretly using chat GPT at
1:41:41 work is actually a thing so so what
1:41:44 Ethan Malik says in this piece is that
1:41:46 these tools are incredible personal
1:41:48 productivity tools but they're too janky
1:41:51 and shitty for corporations to really
1:41:53 feel safe using them so the companies
1:41:56 are like well we're not really quite
1:41:58 sure or you can't use them or we'll fire
1:42:00 you and the employees are like these are
1:42:02 [ __ ] remarkable and so the secret
1:42:04 cyborgs are all these employees that are
1:42:06 getting AI literate but they're not
1:42:08 telling their company about it because
1:42:12 the company has policies like if you use
1:42:14 them we'll fire you or
1:42:17 um the employee doesn't want the company
1:42:19 to know that what used to take them 40
1:42:21 hours they can now do in like a day or
1:42:24 so
1:42:25 so
1:42:27 why why tell
1:42:30 um and what the piece talked about was
1:42:31 that the the the companies need to get
1:42:33 their [ __ ] together and and make a make
1:42:37 a a safe environment for people to
1:42:40 actively experiment with this stuff I'm
1:42:42 starting to see that I'm doing a
1:42:43 workshop right now for a big company we
1:42:46 just did our second week this week we've
1:42:47 got one more next week
1:42:50 um it's a big company and they're taking
1:42:52 this [ __ ] seriously and they've created
1:42:54 a safe environment using Microsoft
1:42:57 Technology and they want to learn how to
1:42:59 use it and they've got their whole
1:43:00 Global Marketing team in there it's it's
1:43:02 pretty remarkable so
1:43:05 um yeah that's that's that's wild I mean
1:43:07 cool that you're using it like that it's
1:43:09 cool that you're letting people know I
1:43:10 think it's really smart to let people
1:43:12 know you're using it
1:43:14 um to the extent that you don't risk
1:43:15 your job
1:43:17 um it will inspire them to use it and
1:43:19 you will start to establish yourself as
1:43:22 a subject matter expert on AI and when
1:43:25 the big bosses at some point saying we
1:43:26 gotta find someone who knows something
1:43:28 about AI oh ion China's good at that you
1:43:31 know that'll be good for you that's
1:43:32 really cool your roll roll shirt into
1:43:35 one of wait
1:43:38 Your Role will shift into one of being
1:43:40 provider of data literacy
1:43:43 people will have easy access yeah that's
1:43:46 that's interesting I think that was in
1:43:47 in response to
1:43:49 um
1:43:50 the uh business intelligence guy
1:43:53 better learn it quick before you get
1:43:55 fired and leverage yourself as an
1:43:57 independent contractor for big dollars
1:43:59 oh yeah
1:44:01 foreign
1:44:12 thoughts on launching an AI Centric
1:44:14 marketing agency for small accounts 5 to
1:44:16 10K monthly billable do it [ __ ] now
1:44:21 um I think you should do it here's what
1:44:23 I would do
1:44:25 um having start an agency and having
1:44:27 tried to sell into agencies for the past
1:44:30 11 years with my current company they
1:44:32 have their heads so far up their asses
1:44:34 right now there's a massive massive
1:44:38 [ __ ] opportunity I wouldn't limit
1:44:39 yourself to five five to 10K a month
1:44:42 with small accounts what I would do if I
1:44:44 were you is I would find three or four
1:44:49 um
1:44:50 people that represent kind of the major
1:44:52 disciplines of an agency an account
1:44:54 person a project manager a technical
1:44:58 person a creative person that are all AI
1:45:00 literate and hungry for AI
1:45:03 and reinvent
1:45:05 how you deliver work how you pitch work
1:45:07 and how you deliver work
1:45:09 because what you can do with a small
1:45:11 three to four person agency right now is
1:45:13 you could go after accounts of agencies
1:45:15 that are
1:45:16 20 to 40 people depending on how good
1:45:19 you get at it
1:45:21 and under bid them so dramatically like
1:45:25 under bid them and over deliver that
1:45:27 you'll you'll start winning work from
1:45:29 from bigger agencies
1:45:32 um I think there's the opportunity right
1:45:34 now for for any agency that that is AI
1:45:38 Centric from the beginning
1:45:41 that so there's a there's a uh
1:45:44 there's kind of this cliche in the
1:45:46 agency world of like you know you you
1:45:49 can have it good fast and cheap uh you
1:45:53 know good fast or cheap pick any two
1:45:55 right you can have it
1:45:57 um
1:45:59 fast and cheap but it won't be good you
1:46:01 can have it good and fast but it won't
1:46:02 be cheap right all that sort of stuff
1:46:05 with this AI stuff I think you can
1:46:08 actually pitch what if you could have
1:46:09 all three good fast and cheap
1:46:12 I think there's a massive opportunity so
1:46:15 yes
1:46:16 full thumbs up Mark etel nlms
1:46:21 1929005 by the way 3.5 on the name
1:46:24 what's going on with the name
1:46:27 if there's a cool story behind it or if
1:46:29 that's some science fiction reference
1:46:31 that I'm missing then I'm an idiot but I
1:46:34 you gotta you gotta work on the name
1:46:36 especially if you're starting an agency
1:46:37 man you gotta you gotta get your name
1:46:39 game down all right all things languages
1:46:41 I use AI templates to create Avatar and
1:46:44 speech over background video but I can't
1:46:46 find Spanish speech go to
1:46:50 what are you using did you use did
1:46:53 hang on
1:47:02 what what did I ah
1:47:10 I'm like where's all the stuff that I
1:47:12 had set up here the ID
1:47:17 oh no is it yeah it's Dandy hang on
1:47:19 create video
1:47:22 uh oh
1:47:39 so if I do this in Spanish
1:47:42 um hello
1:47:44 oh
1:47:45 are you
1:47:48 will she say this in English if I type
1:47:50 it in English
1:48:00 we've successfully generated your video
1:48:02 hello how are you
1:48:04 yeah so if you go to the so if you're in
1:48:08 did
1:48:10 go to the
1:48:12 um
1:48:15 go to the language pulldown so there's
1:48:19 you know there's English
1:48:22 speakers from other languages but then
1:48:25 like where I just went as I went to
1:48:27 Spain
1:48:28 I don't know if the the Puerto Rico in
1:48:30 here no I went to Spain
1:48:36 or Spanish what do they do Spanish
1:48:42 like here's Spanish United States
1:48:49 so let's let's see Spanish United States
1:48:52 generate video oh type your script uh
1:48:55 hello how are you
1:48:58 generate video
1:49:07 how are you yeah that's pretty good so
1:49:11 uh
1:49:13 that's how you do it that's how I do it
1:49:15 in did if you're using 11 Labs I'm not
1:49:17 sure how to do it there
1:49:32 what's that from
1:49:34 not Hawaii 5-0
1:49:38 all right whatever add thank you you're
1:49:41 welcome
1:49:45 whatever
1:49:47 it's my brain this is
1:50:00 I'm at work so there's that hospital lab
1:50:03 oh that's funny
1:50:04 54 here AI doc oh cool nice
1:50:09 gen X's gen xers for the win I I
1:50:13 seriously feel like gen xers have a big
1:50:15 advantage in this just because
1:50:17 there's something about having
1:50:20 you know the energy to still be in the
1:50:22 workforce the ability to understand the
1:50:24 tech piece and just having been on the
1:50:28 planet long enough to see enough [ __ ] go
1:50:30 down that when these things generate
1:50:33 what they generate like you can really
1:50:35 quickly kind of filter through what's
1:50:37 [ __ ] or not
1:50:38 that feels like a superpower to me but I
1:50:41 don't know
1:50:42 Boomers have been doing that finger
1:50:43 movement longer I don't know which
1:50:45 finger movement uh welcome ladies and
1:50:47 gents Welcome to our host the master of
1:50:49 Tickety Toc Kyle Shannon thanks Emilio's
1:50:52 life damn right and all those important
1:50:54 words if I wanted open ai's chat GP chat
1:50:57 API to respond in a specific format how
1:51:01 can I guarantee that every time
1:51:03 so
1:51:06 okay
1:51:09 let's go do
1:51:16 platform
1:51:18 so where I've gone is
1:51:22 platform.openai.com playground
1:51:25 and so I'm going to flip the gpt4 so I'm
1:51:28 on gpt4 let me make this a little bigger
1:51:30 so it's not so ridiculous to see
1:51:33 oops
1:51:36 let's go on that
1:51:40 okay
1:51:42 uh I gotta go in smaller all right well
1:51:46 whatever so I'm in gpt4 I've got a
1:51:49 temperature of one maximum length of
1:51:51 whatever okay
1:51:53 so you you can you can either do this in
1:51:58 the um in the system message or you can
1:52:01 do it in the prompt but but you could do
1:52:03 things like
1:52:04 um
1:52:09 all right let's see
1:52:18 [Music]
1:52:19 you're a social media manager that
1:52:22 specializes in
1:52:24 creating
1:52:28 compelling
1:52:31 social media content
1:52:33 and descriptions
1:52:36 period
1:52:38 um and then I'm going to say
1:52:41 um
1:52:49 where's that
1:52:51 format all answers
1:52:57 as follows
1:53:01 and then I put a colon and then I'm
1:53:03 going to put
1:53:05 um
1:53:06 I'm gonna put brackets I'm going to put
1:53:08 title
1:53:15 and I'm going to put uh
1:53:21 let's see title
1:53:24 I'm going to put Dash bold
1:53:27 see if that works
1:53:29 I'm putting these in Brackets I'm gonna
1:53:31 put title
1:53:32 um body
1:53:36 and then I'll put hashtags I'll put
1:53:41 five hashtags
1:53:47 so so I'm sort of telling it it's a
1:53:50 social media manager and I want it to
1:53:52 format all answers as follows
1:53:54 so and then I put title body and then
1:53:57 hashtags we'll see if it does it so now
1:53:59 I'm going to say
1:54:00 um right uh
1:54:04 me five
1:54:07 YouTube descriptions
1:54:11 of a video about a fat dog barking at
1:54:15 people through a window
1:54:24 let's see what we get
1:54:33 what's going on
1:54:36 why did oh
1:54:38 that gives me that was no answer
1:54:45 huh that's weird it didn't like that at
1:54:48 all
1:54:49 um
1:54:53 so let me I'm gonna I'm gonna kill this
1:54:56 over here
1:54:57 let me kill it like that
1:55:01 um write me five YouTube descriptions of
1:55:03 a video about a fat dog barking uh
1:55:06 format the descriptions
1:55:11 um
1:55:13 with
1:55:15 these elements
1:55:19 and line break line
1:55:22 breaks
1:55:26 and let's see what that does oh wait I
1:55:29 didn't post them in there
1:55:34 I have YouTube
1:55:36 titles and descriptions
1:55:41 [Laughter]
1:55:49 let me delete these delete these delete
1:55:51 these
1:55:52 let's see if I can get it to work
1:55:57 no it didn't like it it doesn't like
1:56:00 these for some reason
1:56:10 maybe the
1:56:12 brackets for throwing it off no it's not
1:56:15 working
1:56:17 I don't know yes you can do it I don't
1:56:20 want to keep [ __ ] around with this
1:56:22 one way to do it is um it's really good
1:56:25 at copying stuff so if you've got
1:56:27 something formatted the way you want it
1:56:28 formatted you can say based on this
1:56:31 example write me 10 more you know that
1:56:35 you know whatever the description is
1:56:37 following that format it's really good
1:56:39 at at
1:56:41 um
1:56:41 following things like that there's a way
1:56:43 to do it like I'm doing it here but it's
1:56:46 I obviously am not
1:56:49 I'm not on tonight I've never had to
1:56:51 give me two asterisks as an answer ever
1:56:54 let me let me get rid of the whole
1:56:56 formatting thing and just see if it
1:56:57 gives me anything
1:57:01 yeah it's not this is
1:57:03 what's wrong with this
1:57:05 um write me
1:57:07 five tweets about a storm
1:57:13 yes something's
1:57:18 something's weird with this tonight let
1:57:21 me get rid of that
1:57:25 oops
1:57:30 yeah something something's really weird
1:57:32 with with playground tonight let me go
1:57:34 try it over in chat GPT
1:57:37 um
1:57:39 right
1:57:40 me
1:57:42 um
1:57:43 five YouTube titles and descriptions
1:57:46 that follow
1:57:50 the format below
1:58:07 sure here you go
1:58:10 all right this is working
1:58:25 come on Focus
1:58:28 there you go sort of it's sort of
1:58:30 focused so you're getting you know title
1:58:33 bold so so next so so I think that
1:58:36 prompt was right but GPT playground it
1:58:38 looks like this is something's [ __ ] up
1:58:40 they're probably fixing something over
1:58:41 at open AI tonight
1:58:44 so so basically what I did here is write
1:58:45 me five YouTube titles and descriptions
1:58:48 that follow this format I put bracket
1:58:49 title bold body five hashtags and so it
1:58:53 created
1:58:54 title bold body five hashtags
1:59:00 so I thought it worked I don't know why
1:59:03 playground was so broken there but but
1:59:05 yeah you can do that and you can also
1:59:07 rather than just you know typing a
1:59:09 format like that you can just paste an
1:59:11 example and say follow this example
1:59:15 uh
1:59:18 there's also a cool thing you can do I
1:59:20 did this the other night so let me show
1:59:22 you another cool thing you can do
1:59:23 because this this you could also do it
1:59:25 but you're talking about the API so the
1:59:28 API the API you do it over in playground
1:59:31 when playground's not broken
1:59:33 but you could do some version of what I
1:59:35 just did
1:59:42 oh OneDrive is Microsoft Okay Google
1:59:45 Drive is Google OneDrive is Microsoft
1:59:49 thank you what plugin is best to connect
1:59:52 chat gbt to the Internet it's called Web
1:59:54 pilot is the one that I would use right
1:59:56 now seems to still work and is pretty
1:59:59 good the bear are giving me my dangling
2:00:01 lights I love the dangly lights
2:00:05 I don't know the bear you're the only
2:00:06 one that gives me the dangly lights no
2:00:08 one else gives me the dangly lights I
2:00:10 miss the dangly lights
2:00:15 the dangly lights make me happy
2:00:25 nobody wants me to be happy
2:00:29 I'm pouty
2:00:31 I'm pouty mcpaddington
2:00:34 um
2:00:36 what's your favorite AI tool so far I'm
2:00:39 loving Google bard I'm
2:00:41 I'm flummoxed right now
2:00:44 like I want this list to be not seven
2:00:49 I'll tell you what I like about them
2:00:51 I'll tell you what I like about each of
2:00:52 them
2:00:54 um and there are different times when I
2:00:56 use them Chachi PT just the writing
2:00:59 quality of gpt4 at the official site is
2:01:03 like nothing else
2:01:05 um but but Claude is close
2:01:08 um and code interpreter at chatgpt is
2:01:12 [ __ ] mind-blowing
2:01:14 Claude
2:01:15 gives you a 75 thank you bear appreciate
2:01:21 that
2:01:22 Claude gives you 75 000 words you know
2:01:25 memory so so you can process really long
2:01:29 documents with this without having to do
2:01:30 anything you can upload five PDFs and
2:01:32 just start interacting with them
2:01:35 it's it's really quite good
2:01:37 Bing is connected to the internet it can
2:01:39 analyze images it can generate images
2:01:41 it's connected to the internet it's gpt4
2:01:44 so even though it's safety guard rails
2:01:48 are like way overdone and it's gpt4
2:01:51 kinda sucks there's times I use Bing
2:01:54 it's pretty good Bard used to suck so
2:01:57 bad I didn't have it on the list it's
2:02:00 gotten better it's connected to the
2:02:01 internet it can analyze images it
2:02:04 incorporates links into answers in a
2:02:07 really elegant way believe it or not
2:02:09 I I can't believe it how well it does it
2:02:13 um this is llama two
2:02:15 um labs.perplexity.ai pi
2:02:19 if you're trying to get anything done is
2:02:21 a awful pain in the ass it's way too
2:02:23 conversational but if you put it in
2:02:26 conversation mode like the app on your
2:02:28 phone
2:02:29 and just start talking to it I've been
2:02:31 talking to it while driving it's amazing
2:02:33 and then poe.com has these Bots that you
2:02:36 can make and you can also play with
2:02:38 multiple there's six different large
2:02:40 language models you can play with in one
2:02:41 tool
2:02:42 so they all have something interesting
2:02:44 like of all of these pose the weakest
2:02:48 Pi is the second weakest that's why I
2:02:50 have them on the bottom and kind of
2:02:51 smaller
2:02:52 but they like
2:02:54 I don't know I this this is a problem
2:02:57 right now because
2:03:00 like I want one tool that doesn't suck I
2:03:02 want one tool that generates images as
2:03:04 good as mid Journey
2:03:06 has a hundred thousand token or a
2:03:08 million token context window
2:03:11 has something like code interpreter is
2:03:14 connected to the internet
2:03:16 elegantly incorporates links into it and
2:03:20 it would be nice if we're open source
2:03:21 and easy to install
2:03:24 so
2:03:27 we're just at the shitty part of the
2:03:31 era like like this is all brand new this
2:03:34 is all brand new everyone's trying to
2:03:36 figure it out like I don't fault any of
2:03:38 these companies for not being perfect I
2:03:40 feel
2:03:41 remarkably appreciative that we get a
2:03:44 chance to play with this [ __ ]
2:03:50 um so I wish I had a better answer for
2:03:51 you what if my life come to geeking out
2:03:55 on a Friday night tell me about it
2:03:58 what'd you do Friday night I don't want
2:04:01 to talk about it why not
2:04:03 I don't want to talk about it what'd you
2:04:05 have a date no
2:04:07 what's going on nothing
2:04:10 you weren't on watching that Tick Tock
2:04:12 again were you shut up
2:04:17 what is it with you and AI I don't want
2:04:20 to talk about it
2:04:21 yeah exactly all right everyone should
2:04:24 be using AI every day to get ready for
2:04:26 what's next for the next Industrial
2:04:28 Revolution that it is that
2:04:30 it's that
2:04:31 and you know just like the Industrial
2:04:33 Revolution it's going to generate jobs
2:04:35 that we can't even imagine right now
2:04:38 but before those jobs come is the shitty
2:04:42 disruption phase
2:04:44 where all of our current jobs get
2:04:46 completely upended and redefined and
2:04:50 reimagined
2:04:53 and so it's gonna be [ __ ] painful and
2:04:57 it's here's the deal and again why this
2:04:59 channel exists it's going to be less
2:05:02 painful for people that understand what
2:05:04 these things are and
2:05:06 how to
2:05:08 interact with them
2:05:12 the people that are sitting on the
2:05:13 sidelines it's it's gonna be ugly
2:05:17 and it doesn't make me feel good
2:05:21 and I think that's why I do these I'm
2:05:23 like just [ __ ] just
2:05:25 [ __ ] get curious
2:05:28 do it
2:05:31 like you know we can joke about you know
2:05:34 what am I doing on a Friday night doing
2:05:36 this well
2:05:38 it's it's a good thing
2:05:40 I think it's a good thing everyone will
2:05:42 be using every day oh yeah we did that
2:05:47 um
2:05:48 when Apple releases their Vision Pro
2:05:50 combined with their AI product it's over
2:05:52 yeah apples Apple's gonna do some Smash
2:05:55 and grab in this in this in this
2:05:58 competitive scene these guys all think
2:06:00 they're all [ __ ] fat and happy right
2:06:02 now Apple's gonna come come along and go
2:06:04 oh that's cute whack hahaha
2:06:09 here's a billion people using an
2:06:11 intuitive product that's incredibly
2:06:13 powerful trans transparent and magical
2:06:18 these guys are going to be like but you
2:06:20 can type words into ours
2:06:26 we can write a poem about a dog
2:06:29 you want one
2:06:30 I can make it for you real quick
2:06:34 I just introduced my co-workers by
2:06:36 having the bot write a short story about
2:06:38 him chasing his cows with with what with
2:06:42 a drone then in Shakespearean style then
2:06:45 Hunter S Thompson style he was impressed
2:06:47 yeah those kind of parlor tricks are
2:06:49 great great entry points right you can
2:06:51 use those kind of parlor tricks and just
2:06:53 you know watch people's eyes go wide
2:06:55 especially if they think if they're like
2:06:57 know-it-alls they're like yeah I heard
2:06:59 of that jet ubta I I've used it
2:07:01 I was talking to a guy once and I showed
2:07:05 him I built this social media content
2:07:07 generation tool and he was like yeah
2:07:10 yeah I've used chat G50 I use it I use
2:07:12 it I use it a fair amount
2:07:13 and we're talking and I'm he's asking me
2:07:15 questions and I'm like well let me let
2:07:18 me show you this thing that I just built
2:07:20 let me get your thoughts on that
2:07:21 and so I asked him a few questions I put
2:07:23 him in a Google form and then it emails
2:07:26 him five pieces of social media content
2:07:28 like five tweets a blog post you know
2:07:32 hashtags all that sort of [ __ ] and
2:07:36 he goes oh I just got that email and he
2:07:38 opens it up and his face went white and
2:07:41 he he literally went
2:07:43 he goes
2:07:45 I had no idea he goes I gotta go he hung
2:07:51 up on me
2:07:55 you know
2:07:57 it's like you know these things are
2:07:59 these things are something they really
2:08:01 are
2:08:03 I like Apple better I need an assistant
2:08:05 yeah Apple's gonna get the assistant
2:08:06 thing right it's it's gonna get there
2:08:09 Apple AI plus Vision Pro game over yep I
2:08:12 have a bot integrated with my iPhone
2:08:14 which one
2:08:16 um you just set off my Siri too
2:08:22 that's hilarious
2:08:24 laughs
2:08:28 [ __ ] Siri
2:08:31 uh
2:08:35 it I mean it really is amazing like like
2:08:37 playing with pie putting pie in the
2:08:39 conversation mode in my car
2:08:44 like I just I feel like for the last 15
2:08:46 years I've been promised that AI chat
2:08:49 Bots
2:08:50 were gonna be remarkable and for 15
2:08:52 years they just sucked ass they were
2:08:55 just bad
2:08:56 they're just
2:08:58 statically programmed logic trees that I
2:09:02 always seem to break the logic within
2:09:03 the first three interactions
2:09:06 and I'm sorry I'm sorry Dave I'm afraid
2:09:10 I can't do that like
2:09:12 they're just awful and then something
2:09:14 like pie
2:09:16 it just works
2:09:19 she's Pleasant she's a little awkward
2:09:21 she pauses weird she laughs weird
2:09:24 you can turn her into a him like
2:09:28 but knows everything is connected to the
2:09:31 internet and can actually answer every
2:09:33 question
2:09:36 doesn't give you the I'm sorry does not
2:09:39 compute I'm sorry does not compute like
2:09:42 after 15 years they couldn't get it
2:09:44 better than that
2:09:46 and you know they've got [ __ ] 400
2:09:48 people working on Siri logic don't
2:09:51 listen to me Siri
2:09:53 [ __ ]
2:09:58 but that's but that's the shift we're in
2:10:00 the middle of right if you think about
2:10:03 how you had to create chat Bots for the
2:10:06 past 15 years
2:10:07 is someone had to take software that was
2:10:11 designed to create branching logic trees
2:10:14 and and create ever increasingly complex
2:10:17 logic trees to anticipate every possible
2:10:20 thing every possible human might ever
2:10:23 possibly say and program in the answers
2:10:29 and now you got a thing here where you
2:10:30 just go
2:10:32 hi and it just talks to you
2:10:36 because that's how this [ __ ] works
2:10:39 like like
2:10:40 think about that shift like think about
2:10:45 20 years probably longer 25 30 years of
2:10:50 companies being built to solve the chat
2:10:53 bot challenge
2:10:55 how do you talk to a thing
2:10:57 that doesn't sound like a robot that
2:10:59 doesn't behave like a robot doesn't
2:11:01 behave like a machine the Turing test
2:11:03 right
2:11:04 years
2:11:05 and years of people built building these
2:11:09 things
2:11:10 using human logic and Technical
2:11:15 structure
2:11:18 explicit static logic
2:11:22 and then overnight these things go bang
2:11:25 that's over that way of thinking is over
2:11:28 that way of programming is over
2:11:32 that's [ __ ] remarkable and then think
2:11:34 about every single piece of software
2:11:37 that's built on that same Byzantine
2:11:40 handcrafted logic that's been developed
2:11:43 over 15 years 20 years 30 years
2:11:47 that overnight is irrelevant because of
2:11:50 what this stuff does
2:11:51 that's the disruption that's coming
2:11:56 back in December I had chat gbt right
2:11:58 when AI film script in the style of
2:11:59 Steven Spielberg mind-blowing results
2:12:01 yep super cool just saw the 80s movie
2:12:04 War games tonight here we are
2:12:07 I here's a fun exercise to do go to go
2:12:11 to gbt for whatever
2:12:14 um have it give you a list of all of the
2:12:17 have it make you a table
2:12:19 of all of the science fiction films it
2:12:22 can it can come up with have it come up
2:12:24 with 20 science fiction films what are
2:12:27 the Technologies represented there and
2:12:29 and you know
2:12:30 where are we today like what's available
2:12:32 to do today and like what's going to be
2:12:34 here within five years
2:12:37 it's wild it's like we're literally just
2:12:39 checking off one movie after another yep
2:12:41 we can do that now we can do that now we
2:12:43 can do that now
2:12:46 back when I knew the writer's strike was
2:12:48 upon us oh interesting cool
2:12:51 um like they say in financial markets
2:12:53 the existence of aliens was already
2:12:55 baked in yeah that's that's probably
2:12:57 true
2:12:59 I still use my AOL email Lauren for the
2:13:01 win with the the Retro Retro uh
2:13:06 retro sexy email address
2:13:09 my Prodigy account was RPG b77 oh yeah
2:13:14 played with Runway ml today that's cool
2:13:16 for Landscapes not so good for faces
2:13:18 yeah Runway ML and faces is really
2:13:21 disturbing it's really bad
2:13:23 but but again the the you know I I feel
2:13:26 like with the with the video stuff we
2:13:27 we've just got we've got a year to wait
2:13:30 this time next year the video stuff's
2:13:32 gonna be amazing
2:13:34 um openai got pressed by corporate they
2:13:36 are getting heavily sued right now yeah
2:13:38 I know
2:13:39 I know it's the the fact that they
2:13:42 haven't returned
2:13:44 um connection to the internet is
2:13:45 astounding to me but I I think so they
2:13:49 cut this deal with the Associated Press
2:13:51 they cut a deal with Shutterstock they
2:13:53 have not cut a deal with time Inc
2:13:56 so I what that says to me is time Inc is
2:13:59 because it was Fortune it was Fortune
2:14:01 Magazine Fortune online
2:14:03 is the reason that they shut down the
2:14:07 um connected with the internet stuff
2:14:10 and that's time Inc so
2:14:12 I'm thinking that timing's like we have
2:14:16 one shot
2:14:18 to just to screw these guys we have one
2:14:21 shot to get the most we can get and I I
2:14:24 heard a thing on Twitter yesterday I
2:14:27 think
2:14:28 that said publishing companies are
2:14:30 asking for billions not Millions from
2:14:33 the large language model companies
2:14:38 so
2:14:39 that's probably gumming up the works
2:14:41 it's probably hard to innovate when you
2:14:44 have
2:14:45 you know big corporate lawyers
2:14:47 you know
2:14:49 on the verge of suing you out of his
2:14:51 existence if you don't give up
2:14:53 90 of your Revenue to them right they're
2:14:57 probably in those negotiations right now
2:14:59 so
2:15:00 it's crazy
2:15:03 they had computers they had computers
2:15:07 when you went to school that's an old
2:15:08 joke y'all Joker that's so mean you're
2:15:11 trolling me called me old I'm not old
2:15:13 what's that gray hair
2:15:15 um oh hang on I'll be right back
2:15:21 oh
2:15:25 now
2:15:27 what are you doing
2:16:32 all right
2:16:35 thank you
2:16:36 well that's not very focused is it
2:16:40 did you guys read all my [ __ ] on my
2:16:41 shelves
2:16:42 did you have conversations about all the
2:16:45 [ __ ] on my shelves
2:16:47 um Zork I remember Zork
2:16:50 that was good
2:16:52 I remember oh it just jumped me to the
2:16:54 bottom hang on
2:16:57 God dang it
2:17:00 God dang it tick tock
2:17:03 yo dum-dum Tick Tock yo dum dum
2:17:08 um
2:17:08 let me see if we can get one other blast
2:17:11 of people coming in here uh if you're
2:17:13 willing to share the live and hit the uh
2:17:16 hit the like button see if we can get
2:17:18 some more folks in here sort of
2:17:19 dwindling out which is fine
2:17:22 I don't mind it getting smaller but
2:17:24 see if we can get some other folks in
2:17:26 here in here yeah
2:17:29 yeah you know you know what are you
2:17:31 gonna do man
2:17:33 all right
2:17:35 I lost my place in comments I'm gonna
2:17:37 just jump ahead a little bit here all
2:17:39 right the guy whose legs started moving
2:17:41 twitching on their own without the
2:17:43 machine learning to repair itself
2:17:46 oh wait the guy's legs started moving
2:17:48 twitching on their own without the
2:17:50 machine learning to repair itself
2:17:53 yeah and I don't I don't think it was a
2:17:55 perfect thing but but it was still like
2:17:57 he's he's walking now
2:18:00 um even though I've been in accounting
2:18:02 and finance all my life I'm a techie at
2:18:04 heart it's cool and chord progressions
2:18:06 yep does Bard have an app uh I don't
2:18:10 think so
2:18:12 but I don't know
2:18:14 and it worked in Canada some of these
2:18:17 things are not available in Canada I
2:18:19 don't think Claude is available in
2:18:21 Canada
2:18:23 I don't think Bard is available in
2:18:25 Canada I think open AI is
2:18:28 I don't know just go play with all of
2:18:30 them I don't know
2:18:31 making [ __ ] up now what did I make up I
2:18:34 might have
2:18:35 seriously amazing time it is
2:18:38 great job you've been live for 90
2:18:40 minutes that was probably an hour ago
2:18:43 um
2:18:44 February was when I first tried chat GPT
2:18:46 out
2:18:47 it keeps logs oh that's cool
2:18:49 February 16th guess how I found out I
2:18:53 asked it oh that's cool really
2:18:58 when was the first time I used
2:19:05 you
2:19:15 let's see
2:19:23 when did I first use chat
2:19:29 GPD
2:19:33 that didn't answer me it didn't like me
2:19:38 well the first day I used chat gbt was
2:19:41 November 30th 2022 I will never forget
2:19:44 the day I was like holy [ __ ]
2:19:48 because I [ __ ] around with the the
2:19:50 playground for a couple of months and I
2:19:52 sort of got it
2:19:54 but I didn't quite get it I mean I got
2:19:56 it and I was like I could build [ __ ]
2:19:58 with this but I don't know what to build
2:20:00 and then chat gbt came out and I was
2:20:02 like uh [ __ ]
2:20:06 I love its ability to facilitate
2:20:08 facilitate my ADD brain farts exactly or
2:20:11 inspiration all day long exactly like
2:20:13 I'm telling you dragonfly academy by the
2:20:15 way 9.9 out of 10 on the uh on the name
2:20:18 good good good good or out wait is it oh
2:20:22 it's Alchemy not Academy dragonfly
2:20:25 Alchemy I have 9.5 out of 10 for the
2:20:28 Alchemy that's good
2:20:29 um yeah the the the ability to just take
2:20:32 80d brain farts and turn them into into
2:20:34 stuff is amazing that messed up
2:20:37 sentence you read from me was an 80d
2:20:40 sentence excuse me that's pretty funny
2:20:43 uh I'm not sure I have ADD but I do know
2:20:46 I need some Adderall
2:20:50 it was funny when the the first time I
2:20:53 was diagnosed with ADD was kind of [ __ ]
2:20:55 it was
2:20:57 90
2:20:58 three maybe a long time ago
2:21:01 and I took whatever the you know the
2:21:04 [ __ ] survey is that they have you do
2:21:06 little tests and she's like yeah I'm
2:21:08 pretty sure you have it and so she gave
2:21:09 me Adderall
2:21:12 and she goes I want to see in a week and
2:21:14 I want to see how you respond and she
2:21:15 goes you know what what was your
2:21:17 response to the Adderall I'm like it
2:21:20 made me feel calm like I didn't really
2:21:21 know what Adderall was I said like I
2:21:23 sort of knew but I didn't really know I
2:21:25 said it made me feel calm and focused
2:21:27 and she's like yeah it's speed that's
2:21:29 not how it affects most people you have
2:21:30 ADD
2:21:33 it's like that's why like I drink coffee
2:21:35 at night at night and people are like
2:21:36 it's not gonna keep you up I'm like no
2:21:41 I just fall right out
2:21:44 um we can download Kyle into our
2:21:46 subconscious while we sleep don't do it
2:21:48 don't do it
2:21:50 over six hours before I get off work I'm
2:21:52 not gonna be here that long I can't do
2:21:54 it I'm a Capricorn son and a Leo
2:21:56 ascendant and a dark Scorpio Moon Rachel
2:21:59 you know way more about this than I do
2:22:01 but there is a cool thing let me see if
2:22:03 I can find it
2:22:06 there's a way that you can make Chachi
2:22:07 PT actually display tarot cards
2:22:14 let's see chat gbt oops
2:22:21 chat
2:22:22 GPT tarot card with images
2:22:28 right that's
2:22:31 The Prompt that generates tarot card
2:22:33 reading
2:22:37 okay today okay chain brainai.com
2:22:41 and this top middle one if you show this
2:22:44 prompt
2:22:45 you can copy this prompt
2:22:47 and I'll show you what it does it's
2:22:49 actually really brilliant it's using
2:22:51 markdown to
2:22:53 so markdown is it's it's like pseudo
2:22:57 HTML it's it's basically
2:22:59 sort of rendering crap in your browser
2:23:02 without actually making HTTP calls
2:23:07 um uh come on Focus you dumb dumb there
2:23:11 we go
2:23:13 so the way this uh
2:23:18 come on come on come on
2:23:20 yeah we're good we're good we're gonna
2:23:22 be okay
2:23:24 so the way this prompt works let me see
2:23:26 if I can do this without defocusing it
2:23:29 uh I want you to act as an expert tarot
2:23:31 card reader so pick a random card and
2:23:35 then it's got It's got a URL
2:23:37 so it's basically got a a
2:23:40 wikimedia.org URL but then it's putting
2:23:46 right here it says import
2:23:48 the text from the cards URL down below
2:23:52 so part of the prompt is a list of all
2:23:54 of the names of the actual cards
2:23:57 in the Tarot deck
2:24:00 and so it pulls pulls that stuff in
2:24:05 uh when it picks the cards so if I hit
2:24:08 return
2:24:17 boom
2:24:23 ace of cups
2:24:25 represents New Beginnings I told you I
2:24:27 had cool [ __ ] going on opportunity and
2:24:29 potential symbol of love compassion
2:24:31 creativity and intuition it often
2:24:33 signifies new relationship be it a
2:24:35 romance friendship deepening of an
2:24:37 existing relationship death change
2:24:40 misunderstood does not mean literal
2:24:42 death it's symbodies symbolizes endings
2:24:45 change transformation transition
2:24:46 definitely in the middle of that eight
2:24:49 of Pentacles
2:24:50 apprenticeship and Mastery yes that's
2:24:52 what we're doing here it signifies
2:24:54 dedication hard work and the pursuit of
2:24:56 perfection yes
2:24:58 it suggests that you are working hard to
2:25:00 improve your skills I am working hard to
2:25:02 improve my skills I'm I am I am
2:25:06 comfortable in my imperfection
2:25:09 I've accepted it and I I I I I embrace
2:25:14 the uncertainty the unknowing of being a
2:25:17 total incompetent boob
2:25:19 there I said it
2:25:24 uh
2:25:26 oh
2:25:29 um
2:25:31 [Music]
2:25:35 you may be learning something new or
2:25:37 improving your skills this is a time of
2:25:39 personal growth and development so
2:25:41 embrace the changes and opportunities
2:25:43 that come your way okay I will I'll do
2:25:46 it
2:25:47 but Isn't that cool and what's cool
2:25:49 about that prompt is it just gives you a
2:25:51 shitty like three card spread right
2:25:53 there but you could rewrite the top of
2:25:55 that prompt you know Rachel if you're
2:25:57 good at at this all this stuff you could
2:25:59 rewrite that prompt to to do whatever
2:26:01 spread you wanted to do
2:26:03 and it would it wouldn't do the physical
2:26:06 layout I don't think although you could
2:26:07 you could maybe have it put the cards in
2:26:10 a table in a grid that would be kind of
2:26:12 cool
2:26:14 yeah that would be fun
2:26:16 so go play with that thing if you're
2:26:18 still here
2:26:19 I'm so I'm so far behind on the
2:26:21 questions now
2:26:22 um I refuse to mow grass Joker I
2:26:26 you and me like this on that I don't
2:26:28 refuse because I don't I'm the only one
2:26:31 I'm I am the chosen one
2:26:36 in the family
2:26:38 use repurpose.io they will open Google
2:26:42 Drive oh and download your tick tocks
2:26:45 wait let me write that [ __ ] down hang on
2:26:50 papa papa
2:26:52 re
2:26:55 -purpose
2:26:59 dot IO
2:27:06 oh that's great
2:27:08 okay
2:27:09 all right
2:27:11 let us do the heavy lifting yes
2:27:13 we make it easy to automate your content
2:27:15 workflow on post who oh hair growth geek
2:27:18 how are you doing I saw you I saw you
2:27:20 live today you were live with like three
2:27:21 other people that was cool how you doing
2:27:25 um thank you for that
2:27:27 I will that if I can make it easy to get
2:27:29 the [ __ ] off Tick Tock and onto YouTube
2:27:31 I'll do it
2:27:33 Sagittarius here
2:27:35 felonious I'm home welcome home
2:27:37 Thelonious welcome to the uh welcome to
2:27:39 the party it's Friday night what why
2:27:42 would you go out and drink and when you
2:27:43 can be here hanging out on Tick Tock
2:27:47 oh God what are we doing with our lives
2:27:50 all right the employee saves time
2:27:55 um I was submitting my school discussion
2:27:57 board but classmates had suspiciously
2:27:59 same submission
2:28:02 as me yeah well
2:28:04 [Laughter]
2:28:07 there you go
2:28:09 um love your stream but you are not
2:28:11 doing the interactivity of the live
2:28:14 stream a favor
2:28:18 do you mean because I'm answering all
2:28:20 the questions and not just dealing with
2:28:22 the interactivity at the bottom of the
2:28:25 list
2:28:27 I don't know how to do that well if I
2:28:29 had a moderator
2:28:30 that could scroll for me and tell me
2:28:33 when to respond to [ __ ] and like someone
2:28:35 said something nice and laughed at that
2:28:37 joke you just said then I could do that
2:28:40 but
2:28:41 I ain't got that I'm just I'm just a
2:28:44 lowly
2:28:45 Elder talk member I I ain't got no the
2:28:50 the AI learning lab it ain't 43 person
2:28:53 staff here
2:28:55 it's me and a dog
2:28:56 [Laughter]
2:29:03 so listen technocracy three or
2:29:05 technocracy one
2:29:08 um
2:29:09 you want to help
2:29:12 do you know what you're doing because I
2:29:14 don't
2:29:17 I read something about fraud GPT on the
2:29:20 dark web oh it generates effective
2:29:23 fishing yeah I'm sure there's going to
2:29:24 be a [ __ ] ton of those out there
2:29:28 um
2:29:29 yeah I mean our
2:29:32 the fraud stuff the misinformation stuff
2:29:36 the personalized misinformation we're
2:29:39 gonna start seeing like don't be
2:29:41 surprised if you start getting
2:29:44 voicemails from people you know
2:29:49 telling you that they're going to vote
2:29:51 for a certain candidate
2:29:53 where are you like
2:29:54 no they're not and you call them up and
2:29:57 like what the [ __ ] are you talking about
2:30:00 they're gonna be some they're gonna be
2:30:03 some fancy ass [ __ ] going on in this
2:30:05 next election here in America
2:30:07 because we're America [ __ ] it
2:30:10 and uh we fight dirty so
2:30:15 when you got tools this powerful
2:30:18 they're gonna be some dirty politics
2:30:21 going on
2:30:22 it's gonna make 2016 look like it was a
2:30:24 [ __ ] Cakewalk
2:30:29 he must be using chat gbt so lazy so I
2:30:33 re-reworded my submission that's pretty
2:30:36 funny seeing the others in my class have
2:30:37 discovered AI sorry Thelonious yeah what
2:30:40 are you gonna do
2:30:41 you know what you should do actually
2:30:44 is this would actually be interesting
2:30:47 all right Thelonious here's here's the
2:30:49 thing to do
2:30:51 so don't ever just write it
2:30:53 whatever you're having chat GPT right
2:30:55 don't ever just start with what what you
2:30:58 wanted to write start with it
2:31:02 um telling you about
2:31:04 um some authors in history right so pick
2:31:08 an author like contemporary ones right
2:31:09 so do do ones in the 20th century so
2:31:11 it's not such stilted language
2:31:13 and then just pick so you know Hunter S
2:31:16 Thompson or or or Tom Robbins or you
2:31:19 know someone you like
2:31:21 um and then just have it write whatever
2:31:23 you have it right in the style of that
2:31:25 author and every time you turn in a
2:31:27 paper choose a different author and then
2:31:29 and then see if see if your your
2:31:32 professors pick up on it
2:31:39 you can actually bury in the writing a
2:31:42 reference to that author like subtly put
2:31:45 their name in there somehow I don't know
2:31:47 that would be pretty funny
2:31:49 that's awesome best advice from a
2:31:52 Director of Finance I worked for was
2:31:53 under commit and over deliver you can do
2:31:55 that with these tools
2:31:57 that this there was a person earlier
2:31:59 asking about starting an agency I think
2:32:01 that you could start an agency where you
2:32:03 set your internal goals for the agency
2:32:05 of a 30X improvement over what was done
2:32:08 before so 30 times faster so instead of
2:32:11 delivering something in a month we're
2:32:12 going to deliver it in a day
2:32:14 that's your internal goal and then you
2:32:16 have a 10x
2:32:18 um
2:32:19 increase that you promised to clients so
2:32:22 you say I'm going to deliver this 10
2:32:23 times faster than your current agency so
2:32:25 instead of a month it's going to be
2:32:27 three days
2:32:29 and then you give yourself one day to
2:32:31 turn it around and turn internally and
2:32:33 if you [ __ ] it up you have an extra two
2:32:35 days to deliver it
2:32:37 I think you could do that so
2:32:40 I I think these tools would let would
2:32:42 let you build a dramatically better
2:32:45 um agency what's D Dash ID Dash ID
2:32:49 allows you to take still images and it
2:32:52 synthesize voices and then it makes the
2:32:54 images talk
2:32:57 as in I think it's like digital identity
2:33:00 um disassociative identity disorder
2:33:03 Joker you're like such a joker it's like
2:33:06 you're making jokes like the comedy the
2:33:09 comedy jokes enough comedy jokes
2:33:15 man early Steve Martin was [ __ ] funny
2:33:18 book them down oh Hawaii Five-0
2:33:28 anyone else used to play trumpet trumpet
2:33:30 players
2:33:32 uh um we are survivors us gen xers yep
2:33:36 we are we're bitter
2:33:38 we're the we're the Forgotten bitter
2:33:41 big wheels shooting BB guns and
2:33:43 fireworks at each other played out in
2:33:45 the woods drinking from hoses yep that's
2:33:47 us
2:33:48 good good to have you back have you
2:33:50 started your YouTube channel with your
2:33:52 previous lives no I haven't gotten to it
2:33:54 yet but someone just gave me a cool tool
2:33:58 repurposed they're just gonna go get all
2:34:00 my little Tiki Takis and stick them on
2:34:02 the YouTube
2:34:04 and then I can point people there and
2:34:06 then instead of having this I'll just
2:34:08 tap my YouTube channel there and I'll
2:34:09 have monetization turned on in like
2:34:11 millions of people flock over to watch
2:34:14 an old guy talk for two hours
2:34:15 incoherently
2:34:18 laughs
2:34:21 and I'm
2:34:22 the car from all that tasty tasty ad
2:34:25 Revenue
2:34:28 I'll get demonetized because whatever
2:34:31 the hell I do here jarts oh remember
2:34:33 jarts they're illegal now Etch A Sketch
2:34:35 Lincoln Logs Silly Putty all of it all
2:34:39 the cereals back then man that was that
2:34:42 was chemicals for good living the
2:34:44 cereals we ate
2:34:46 uh just telling AI to play a certain
2:34:48 role work in GPT 3.5 yes it does
2:34:52 it works in all of the large language
2:34:54 models so
2:34:55 what you're doing
2:34:57 it's it's not just tell it to play a
2:34:59 role that's one thing you could do you
2:35:00 could also so so you could say act as a
2:35:03 social
2:35:04 um media manager right and then tell it
2:35:07 to do that you could also ask it a
2:35:09 question do you understand
2:35:12 um how to write compelling SEO copy and
2:35:15 it'll answer that and it'll give you a
2:35:17 thing
2:35:18 what you're doing when you're when
2:35:20 you're providing that early context is
2:35:23 you're kind of limiting
2:35:27 the world of possibilities
2:35:30 of what it can look at to just the stuff
2:35:33 that you're interested in so the way
2:35:35 these tools work they've been trained on
2:35:36 all of the internet so if you just say
2:35:38 right these tools work they've been
2:35:41 trained on all of the internet so if you
2:35:43 just say write me a tweet
2:35:45 it'll sort of look across all of the
2:35:48 internet that it's been trained on see
2:35:49 any of the tweets that were written and
2:35:51 it will do some sort of amalgamation and
2:35:54 give you some super generic thing it'll
2:35:56 take all the shitty tweets and all the
2:35:58 examples of bad tweets and all the good
2:36:01 tweets and I'll kind of mash them
2:36:02 together and you'll get some generic
2:36:04 shitty piece of crap but if you say do
2:36:07 you understand how to write an SEO
2:36:09 friendly tweet it'll sort of knock out
2:36:11 the bad tweets and the bad examples and
2:36:13 it'll kind of narrow the world so all of
2:36:16 these things the more context that you
2:36:19 give them so if you have kind of a long
2:36:21 step-by-step conversation that's going
2:36:23 to give you better results than just
2:36:25 writing sort of a simple little like you
2:36:27 would write in Google a little concise
2:36:29 prompt and expect it to give you a good
2:36:31 answer this is kind of the opposite you
2:36:34 know give it lots of context lots of
2:36:37 um boundaries for the conversation and
2:36:39 it'll get better and better
2:36:44 I want to know more uh not sure about
2:36:47 the micro part though I don't know what
2:36:49 you're talking about I can't get my
2:36:50 plugins to work
2:36:52 on openai tonight I noticed last night
2:36:55 that um
2:36:56 someone asked me about real estate
2:36:58 plugins and there's none and I know
2:37:01 Zillow was one of them
2:37:02 and I'm wondering if they killed if if
2:37:05 they uh if they got something going on
2:37:08 with with real estate
2:37:10 let me try web pilot um let's see
2:37:13 um
2:37:14 what is storybine.com let's see if it
2:37:18 fires off web pilot
2:37:23 it fired it off
2:37:27 did it work
2:37:33 yeah it worked
2:37:38 I don't know that was really weird that
2:37:41 um that playground was that broken
2:37:45 let me see if it's still broken let me
2:37:47 reload it
2:37:52 um write me a tweet
2:38:00 yeah now it's working all right they
2:38:02 must they must just been fixing
2:38:03 something
2:38:06 um yeah something something's going on
2:38:07 over at open AI they're they're probably
2:38:09 doing some updates I heard
2:38:13 I read today
2:38:15 that there's some new features out
2:38:17 coming coming this weekend I I think I
2:38:19 think I think there's some [ __ ] going on
2:38:21 to open AI they're doing some updates
2:38:24 yeah I think that's what's happening all
2:38:27 right I'm gonna I'm gonna
2:38:30 wind want some things down here we're
2:38:32 gonna we're gonna wind her down I know
2:38:34 we've been peppy we've been excited
2:38:36 we've been talking
2:38:37 we've been talking we've been well I've
2:38:39 been talking we've been rambling
2:38:42 we've been talking about the AI we've
2:38:44 been talking about the the career
2:38:46 choices we've been talking about
2:38:48 profound societal impact we're going to
2:38:50 wind it down a little we're gonna start
2:38:52 talking about puppies and horoscopes and
2:38:56 uh in uh candies and sweet sweet cereal
2:38:59 sweet sweet sweet cereals that that
2:39:01 drive holes in your teeth that as you
2:39:03 get older they gotta they gotta fill
2:39:05 them with that them newfangled white
2:39:08 epoxies and they got I gotta take out
2:39:10 all the lead that they put in our mouths
2:39:12 when we were little when we were sucking
2:39:13 on the end of hoses because they
2:39:15 wouldn't let us in the house to drink
2:39:16 out of a out of a sink
2:39:19 that's we're gonna wind it down
2:39:22 take this to a Gen X happy place we're
2:39:24 gonna go take a nap
2:39:28 you would be a great voice actor in
2:39:30 Hollywood by the way thank you Paul Veer
2:39:32 I appreciate that thank you very much
2:39:34 [Laughter]
2:39:41 I think voice actors are actually
2:39:46 gonna be
2:39:47 severely disrupted
2:39:50 so I would I would love to do that but
2:39:54 these things are going to get really
2:39:56 good
2:39:57 these tools are going to get really good
2:39:59 so I yeah
2:40:03 I think they should add your voice as
2:40:05 one of the possible selections on 11
2:40:06 Labs I made two 11 Labs voices I did one
2:40:09 I did one where I kind of talked like
2:40:11 you know sort of a more formal version
2:40:14 of what I do here and then I did one
2:40:15 where I I talked like a Pixar guy super
2:40:18 excited that was awesome
2:40:21 and they came out okay like what what
2:40:23 the what the um
2:40:25 with the voice synthesizers don't do
2:40:27 right now is they don't really get
2:40:29 Cadence
2:40:30 they don't get that things speed up and
2:40:33 slow down so everything talks at the
2:40:34 same speed
2:40:41 it's like take a [ __ ] breath
2:40:44 get excited once in a while slow it down
2:40:47 once in a while
2:40:50 actually that's what the voice actors
2:40:52 should be doing right now is being hired
2:40:53 as consultants for all these you know
2:40:55 because because you got a bunch of
2:40:57 [ __ ] propeller head technologists in
2:40:59 there going we can synthesize the voice
2:41:01 and we we just we uh do it at this
2:41:03 sample rate and obviously uh we bring in
2:41:05 the sample rate and we uh Elevate that
2:41:07 with the compression and we do a high
2:41:09 pass and a low pass and then uh it
2:41:11 sounds more realist it it is are there
2:41:14 any human beings here actually listening
2:41:16 to how this sounds well uh we have
2:41:18 replicated human beings and and
2:41:20 therefore it is uh it is a accurate uh
2:41:22 to within a
2:41:24 0.0027 of uh an accuracy model that was
2:41:29 created in 1972.
2:41:31 um yeah yeah exactly exactly that's why
2:41:35 it doesn't sound like a human
2:41:37 you need a [ __ ] human in here going
2:41:39 no you nerd I don't care about that crap
2:41:43 make it sound better
2:41:48 all right why are they not interfaced
2:41:50 with each other yet exactly Joker I
2:41:52 don't know what you're talking about but
2:41:53 they should be interfaced together
2:41:55 fantastic fantastic Fridays geeking out
2:42:02 I've always been curious I've always
2:42:04 been curious Dave sorry Dave I'm afraid
2:42:07 I can't do that
2:42:09 uh you said there were six different
2:42:11 llms there are seven they're probably
2:42:13 more than seven there are seven that I
2:42:15 think are worth your time
2:42:19 chat.openai.com Claude AI Bing bard
2:42:25 labs.perplexity.ai
2:42:26 is llama that's the meta models pi and
2:42:32 Poe
2:42:33 they all have something interesting you
2:42:35 can do
2:42:36 we are making plans to take over the
2:42:38 world on Friday nights that's it we're
2:42:40 gonna do it we're gonna do it right here
2:42:43 right now
2:42:45 you should try to make a poem about a
2:42:47 dog and chat gbt from the perspective of
2:42:49 the dog we could do that
2:42:52 um
2:42:53 I use I should use pie in the car with
2:42:56 my new Amazon Echo Auto device yeah Paul
2:42:59 the the the I'm telling you using pie in
2:43:03 the car it was
2:43:05 I I mean listen it wasn't fantastic but
2:43:08 it was some version of revelatory to me
2:43:10 because it's just like
2:43:12 I I don't know I've just been so
2:43:15 disappointed for so long with Alexa and
2:43:18 Siri that to have something that you can
2:43:20 actually have a full conversation with
2:43:22 and then just [ __ ] change the subject
2:43:25 and go you know I want to know about the
2:43:28 Mesopotamian War
2:43:30 and it knows it you just [ __ ]
2:43:34 you know context shift all you want and
2:43:37 it's just like sure let's talk about
2:43:39 that now it's amazing
2:43:42 that someone never used Karen for the
2:43:44 digital assistant name in the place of
2:43:46 Siri that's hilarious
2:43:47 um Google's AI built-in hard code to
2:43:51 fail the Turing test interesting
2:43:53 I wonder why they did that
2:43:55 they need to have better customer
2:43:57 service robots I just end up screaming
2:43:59 at the phone so Rachel I think so I
2:44:02 think customer service is one of the
2:44:05 killer apps of this [ __ ]
2:44:07 I honest to God feel like they should
2:44:09 shut
2:44:11 90 of call centers down and replace them
2:44:16 with Bots that answer the phone and
2:44:18 actually know what the [ __ ] they're
2:44:19 talking about
2:44:20 you'd never be put on hold you'd never
2:44:23 start out with a level one person after
2:44:25 45 minutes that then puts you on hold
2:44:27 for 30 minutes to get to a level two
2:44:29 person all both of which are just
2:44:31 reading scripts then they put you on
2:44:33 hold for 45 minutes for a level three
2:44:35 person who's also just reading scripts
2:44:37 for you to finally get to a level four
2:44:39 person that [ __ ] hangs up on you
2:44:44 that needs to end that needs to end all
2:44:47 of that [ __ ] needs to end it needs to
2:44:49 you call and it picks up the phone and
2:44:53 if it sounds stilted as a robot I don't
2:44:56 care but if it goes hello Kyle how are
2:44:58 you the last time you called we talked
2:45:00 about this everything okay no it's not
2:45:03 okay I'm really pissed off because my
2:45:05 internet sucks and Comcast sucks and
2:45:07 XFinity sucks oh Kyle I'm so sorry to
2:45:10 hear that let me run a test on that
2:45:12 right now bang it runs the test it
2:45:14 appears there's some outages in your
2:45:16 neighborhood would you like me to call
2:45:17 you back in two hours and check yes I
2:45:20 would like that like that's what
2:45:22 customer service should be
2:45:24 and it it can be done
2:45:29 and then that butt is also smart enough
2:45:32 to go oh Kyle in looking at your thing
2:45:34 it appears that your problem is more
2:45:36 complicated than I can deal with it
2:45:38 seems like the best path would be to put
2:45:40 you in touch with an actual human would
2:45:42 you like to talk to someone now yes I
2:45:44 would click they put me through to
2:45:46 someone who's actually competent who has
2:45:48 the authority to actually do something
2:45:50 about my problem that's what customer
2:45:52 service needs to be
2:45:56 if someone's not building that right now
2:45:57 I'm very cranky about it that needs to
2:45:59 be built and holy [ __ ] is that a
2:46:02 trillion dollar business
2:46:06 trillion dollar business and it's not
2:46:08 that complicated it's all about
2:46:11 um
2:46:12 getting the data of the corporation so
2:46:16 all of the [ __ ] that those individuals
2:46:17 are trained on and all those scripts are
2:46:19 written on all of that data can get
2:46:22 sucked into a large language model they
2:46:24 can embed that all put it into one of
2:46:26 these stick a synthesized voice on top
2:46:28 of it and
2:46:30 get a giant [ __ ] server farm that all
2:46:33 it does is answer phones and talk to
2:46:35 people for as long as they want to
2:46:36 [ __ ] talk
2:46:42 yes Rachel what you said
2:46:51 it's one of the downsides of kind of
2:46:54 knowing how this [ __ ] works works is
2:46:56 knowing what it should replace
2:46:59 and seeing that the world hasn't
2:47:01 replaced it yet like frankly this shit's
2:47:03 been around long enough that that should
2:47:06 have been done 10 years ago
2:47:10 I don't know who's got a [ __ ]
2:47:11 lobbyist in the call center business but
2:47:14 we gotta fire that lobbyist and replace
2:47:16 those call centers and those logic trees
2:47:19 with these things like now
2:47:23 hmm
2:47:24 [Laughter]
2:47:26 all right
2:47:27 Nelly was saying wait oh I wasn't saying
2:47:31 they didn't have computers when I went
2:47:32 to school oh I see okay Kyle was
2:47:34 featured in the Wired Magazine on the
2:47:35 Shelf that's true I'm in that
2:47:39 okay Kyle took a break yeah I did
2:47:43 it
2:47:48 yeah it's me and Champion today although
2:47:52 Champion left me Champion not here now
2:47:55 sad
2:47:56 Champy's done Champy's like I'm not
2:47:59 hanging out with you on a Friday night
2:48:02 I got chicken snacks to chew on
2:48:05 Friday night geeking thanks you're
2:48:07 welcome Firefly thanks for hanging out
2:48:11 user what date was my oh oh user what
2:48:15 date was my first message to you
2:48:20 Tristan
2:48:38 oh
2:48:40 but that's not true
2:48:41 oh your first message to me in this
2:48:43 conversation was August 4th 2023
2:48:46 how about ever
2:48:51 no okay user
2:48:56 when
2:48:58 did I first
2:49:05 to act with you
2:49:09 no still not doing it oh well it's close
2:49:16 I think there's something in Poe that
2:49:19 reads tarot too but I might be wrong
2:49:21 yeah I think there's a bot in there but
2:49:23 it I
2:49:24 I don't know oh I haven't played with
2:49:26 that one I don't know if it actually
2:49:27 shows the images
2:49:29 I tried to give away a deck of tarot
2:49:31 cards to a girlfriend the other day
2:49:34 she left them which poser
2:49:39 that's cool no grass in Florida by the
2:49:41 beach only sand
2:49:43 I also hung the Green Acres team
2:49:45 sometimes
2:49:52 Green Acres is the place to be
2:49:57 in Paris
2:50:00 oh that's that's Gen X happiness right
2:50:03 there Green Acres I used to watch that I
2:50:05 used to watch Gilligan's Island
2:50:09 oh man bringing bring him bring him back
2:50:12 the the Gen X Vibes I like it
2:50:15 I like it make us a few mods Kyle it's
2:50:19 you and chip
2:50:21 um I don't know how to
2:50:22 if someone wants to reach out to me and
2:50:24 tell me how to make mobs I don't know
2:50:26 how to do that [ __ ]
2:50:28 I'm happy to talk about it
2:50:30 I think you're doing a fine job I was
2:50:32 scrolling and never left thank you Joker
2:50:34 I appreciate that armed with the buzzing
2:50:36 Beast Ethan took to the sky like a man
2:50:39 possessed his plan
2:50:42 HST style yeah see
2:50:47 those wild and crazy guys
2:50:50 cusp cereals yep cusp oh yeah cusp wisp
2:50:54 wait crisp crisp remember crisp crisp
2:50:59 crisp cereal was good I was a captain
2:51:01 cruncher Captain Crunch is still around
2:51:03 it hurts my teeth now
2:51:06 because because my because my uh my
2:51:09 teeth are too thin from eating Captain
2:51:10 Crunch in my childhood
2:51:14 um oh crisp cereal yeah flying saucers
2:51:17 um I got Raisin Bran as a kid my mom
2:51:19 wouldn't buy me sugar oh well Raisin
2:51:21 Bran had a little sugar in it those
2:51:23 those little sugar bombs those little
2:51:25 raisins
2:51:26 we're already there Kyle
2:51:29 um yes all those things Kyle
2:51:32 um the nickname suits you Silver Fox
2:51:35 you're welcome yes it's a compliment
2:51:37 thank you
2:51:39 um all the llms the image The Voice the
2:51:41 video generators interfaced yeah they're
2:51:44 we're gonna start to see these things
2:51:45 all all come together oh the voice is
2:51:48 going
2:51:49 it just it just just [ __ ] closed up
2:51:52 on me it just I can just feel my like it
2:51:54 was like a you know a curtain of snot
2:51:56 came down and my two vocal cords folded
2:51:59 across each other and it's like hot
2:52:01 [Laughter]
2:52:06 uh
2:52:08 uh uh yes
2:52:10 what are we doing here this is the
2:52:13 learning lab we are learning learning
2:52:16 learning all of the all alarms
2:52:19 travel agent during pandemic I can't
2:52:21 even talk about customer service holds
2:52:23 oh my God I can't even imagine
2:52:26 uh
2:52:29 uh
2:52:30 user was accidental copy paste user is
2:52:32 me
2:52:34 oh wait what
2:52:37 this plan was over the character limit
2:52:40 bye thanks thanks Silver Fox all right
2:52:43 uh I'm gonna go everyone it's 11 o'clock
2:52:45 here in Denver I hope you had fun
2:52:47 tonight uh it was good hanging out with
2:52:49 you I really appreciate all of your
2:52:51 thank you David gold I love those worlds
2:52:53 thank you so much
2:52:55 um
2:52:56 y'all are awesome
2:52:58 um I really do have a good time hanging
2:52:59 out with you I hope that I hope this was
2:53:00 uh educational and entertaining
2:53:04 and uh all that happy horse [ __ ] so uh I
2:53:08 will see you tomorrow sometime I'll
2:53:09 probably you know be back right around
2:53:12 this time or whenever I don't you know
2:53:14 it's
2:53:15 I ain't got no schedule but I'm gonna go
2:53:17 look at what am I going to look at I'm
2:53:19 going to look at uh repurpose
2:53:22 repurpose dot IO and see if I can't get
2:53:25 some of these long-form things over to
2:53:27 the YouTube what might be fun is you
2:53:30 know get all these things over to the
2:53:31 YouTube
2:53:33 and then use something like video or
2:53:35 even chatgpt to analyze the transcripts
2:53:38 and find interesting Clips within them
2:53:40 that could be an interesting experiment
2:53:42 so all right peace everybody have a good
2:53:44 night