aiLL 9/2/23 - Navigating the Wild Frontier of AI

Description

Chapters

Transcript

0:00 thank you
0:24 welcome welcome welcome
0:27 happy Saturday night yes Saturday night
0:31 September 2nd it is not in my reality
0:35 that it's September but hey what are you
0:36 gonna do togas it's happening
0:39 down in the mud someone from called down
0:42 on the mud is in here I suppose they're
0:44 a burning man right now
0:46 Ted looks at life what's happening sir
0:52 how's life looking
0:54 welcome welcome welcome Saturday Night
0:57 Fever
1:00 oh yeah oh yeah
1:04 I was working on getting uh
1:07 some merch going but printful decided to
1:10 just
1:12 crap the bed
1:15 but what are you gonna do what are you
1:18 gonna do
1:19 um
1:21 I like taters
1:24 potatoes tasty mm-hmm
1:28 all right just missed the end of
1:31 my movie to watch real stuff
1:34 sorry about that
1:39 oh so guys what are we gonna do with you
1:47 all right got some more Us open on
1:51 welcome everybody it's AI learning lab
1:54 we'll wait for some join The Irregular
1:56 Discord very cool
1:59 we'll talk about that
2:03 [Music]
2:10 happy Saturday night welcome everybody
2:12 welcome welcome welcome welcome
2:15 we'll get started here
2:18 we will talk about
2:20 AI
2:23 and
2:25 where to get started and we'll answer
2:28 questions I actually had an idea about
2:30 how to do
2:31 how to balance things a bit more between
2:34 showing stuff and answering questions
2:38 although it's a little tough to show
2:40 stuff clearly but
2:43 I got some ideas a little experiment
2:46 tonight
2:49 see if we can get some more people in
2:50 here if you'd be willing go ahead and
2:52 share the uh share the live
2:56 and we'll see if we can get some more
2:58 folks in here and then we get started
3:03 fantastic Bob tell him what he's won
3:05 well
3:06 you haven't really won anything tonight
3:08 sir you haven't even gotten started you
3:10 didn't play the game
3:12 you got to play the game to win the game
3:14 and you haven't started playing games so
3:15 how are we going to win anything huh
3:18 you know life's not just a cakewalk you
3:20 got to do some work in this world all
3:22 right all right fantastic
3:25 all right
3:27 got 17 folks in here
3:30 Tick Tock is Slow Rolling the Live
3:32 tonight
3:36 um when can I play a whole DND campaign
3:38 with AI pretty sure you can do it right
3:40 now sport cloud
3:43 um
3:44 especially with uh have you tried it
3:47 with Claude claude.ai because club's got
3:50 a 75 000 token context window
3:53 so if you tell Claude to act like a d d
3:55 dungeon master and which
3:57 Edition you wanted to use which rule set
4:02 I don't know if I can do the whole
4:03 campaign but it could probably get
4:05 pretty close because with a 75 000 or
4:07 it's a hundred thousand tokens it's
4:09 about 75 000 words
4:11 you know like a 50 000 word novel you
4:14 can just paste right into Claude and it
4:16 can ingest the whole thing yeah 100K but
4:19 about 75 000 words
4:22 um yeah try that sport Cloud that could
4:24 work and this is check this out I did
4:26 this the other day
4:35 I had chat GPT
4:39 um
4:39 come up with
4:41 different oh wait pictures and locations
4:45 and then had it right so so what was
4:49 cool about this this project so I went
4:52 into Claude
4:54 and
4:56 I had Claude hey ion China I had Claude
5:00 thank you for the rose
5:02 I found an article on sdxl so someone
5:05 had written a like a blog post a
5:07 decently long article on stable
5:09 diffusions sdx sdxl's model and what the
5:13 different you know sort of attributes
5:15 are
5:16 and so I I put the article into Claude
5:19 and started asking it questions about it
5:22 and then I thought oh maybe I could have
5:23 it make me a set of prompts based on
5:25 that and then I said well why don't I
5:27 come up with a category and I had to
5:29 come up with four categories and do some
5:31 prompts for each of the categories and
5:33 one of the categories was fantasy and I
5:35 thought oh that's really cool and so so
5:38 I had it
5:39 um generate you know a name
5:43 let's see let me make this a little bit
5:45 smaller
5:47 how to generate a title like Dwarven
5:50 cleric a name brother ironbeard
5:53 and then a prompt and then optional
5:55 elements and then I literally just and
5:57 with the prompt I had it also give me an
5:59 aspect ratio like depending on what you
6:03 prompt make it either you know uh
6:05 portrait or landscape depending on you
6:08 know what you do and then
6:12 it gave me four different categories so
6:14 these are the characters
6:17 let's see
6:24 and I mean they came out good I came out
6:27 really good
6:29 and then here are the creatures
6:34 nice little unicorn
6:36 mermaid Giant
6:39 baby dragon what's that thing a
6:41 displacer beast that cat looking thing
6:48 little dragon fairy
6:50 and then here's the landscapes
6:54 Serene Springs
6:56 that thing's cool
6:58 so and this was I mean what's cool about
7:01 this what what what took me the longest
7:03 to do
7:04 was
7:06 copy and paste the prompts into and then
7:09 here are the structures
7:11 um copy and paste the prompts into mid
7:14 journey and then curate the best ones
7:17 out right and sometimes I would do like
7:19 you know three or four different renders
7:21 like this ship was hard they were they
7:22 were close but they didn't quite capture
7:24 like the The Prompt says tattered sales
7:28 and this was the one that had the most
7:30 tattered sales it wasn't quite the
7:31 sexiest ship but it it you know looked
7:34 good
7:36 so like look at that thing
7:39 so this was just it's just like a whole
7:41 world of you know
7:44 artwork and things like that so you
7:46 could probably do some really
7:47 interesting things playing with uh with
7:50 that
7:52 all right how many folks are in here got
7:55 a few folks in here welcome everybody my
7:57 name is Kyle Shannon this is the AI
7:59 learning lab let's get started here so
8:03 how I like to start out these little
8:05 evenings happy Saturday night by the way
8:08 is uh
8:11 if if you're new to this stuff and
8:13 you're like where do I start but where
8:15 do I start Kyle I don't know where to
8:17 start
8:18 where to start is this top one
8:23 chat.openai.com so that is the official
8:25 chat CPT website
8:28 um
8:30 what's going on I wonder if my cameras
8:33 are dirty I think my cameras might be
8:35 dirty but whatever
8:37 um that's the official chat GPT website
8:40 these three are kind of the top three
8:42 that I play with I bounce in between
8:43 them
8:44 um last night last night Bard was at the
8:48 Kitty table and I moved Bard up to the
8:50 top three because uh Bing is just really
8:54 annoying with how how much safety stuff
8:55 it's got on it so
8:57 um that's chat EBT Claude is the one I
9:00 just talked about with the DND thing
9:01 where it's got a hundred thousand token
9:03 context window and then Bard is Google's
9:06 entry into this space it's connected to
9:09 the internet they they incorporate
9:11 internet links in a pretty sophisticated
9:13 way and it's not horrible
9:16 and apparently this fall they're they're
9:18 going to be releasing Gemini which is
9:21 supposed to be amazing they claim it's
9:23 going to be five times better than gpt4
9:25 I will believe that when I see it
9:28 but they do have a lot of data so it's
9:31 possible that they did some kick-ass
9:32 training
9:35 so that's that stuff
9:37 um
9:39 okay other things to play with Bing like
9:42 I said it's a bit of a it's safety guard
9:45 rails or a bit of a pain in the ass but
9:47 it is connected to the internet
9:50 you can likely use it within a business
9:52 it can generate images you just say
9:54 create an image of X and it will do it
9:56 it can read images you can upload an
9:58 image to it and tell it to describe it
10:00 to you and it will
10:02 perplexity.ai is a good research tool
10:05 labs.perplexity.ai is them hosting
10:07 meta's llama 2 models for free and their
10:10 code llama model which is which is sort
10:13 of a a fine-tuned version sitting on top
10:16 of it that's really good at coding Pi
10:18 dot AI is fascinating uh we we here at
10:22 the lab have nicknamed it chatty GPT so
10:26 go to Pi dot AI or go get the PI app so
10:30 just look for pi and then personal
10:32 assistant and you'll find the app the
10:34 company is inflection put it into voice
10:37 mode and start talking to it
10:39 if you want to actually get work done
10:41 like you do with chat GPT Pi's not your
10:44 answer but if you want to talk to it to
10:45 one of these models and kind of get a
10:47 sense of what the future is like that's
10:48 what that is and then pogo.com is
10:51 another one to play with that I think is
10:53 is at least worth it at least worth
10:57 exploring good Lord the tongue ain't
11:00 working tonight what is winter tongue
11:02 what is going on okay now you're saying
11:05 but but Kyle all those are just the
11:07 words and I'm not good with words I
11:10 don't like words words words words
11:11 everyone's talking about words all the
11:13 time I like pretty pictures
11:16 where are the pretty pictures Kyle here
11:18 you go here you go right here
11:20 these are the ones I would recommend
11:22 playing with right now the best one is
11:24 mid-journey however
11:27 it doesn't do words ideogram is like
11:30 seven days old eight days old and it
11:32 does words
11:34 really interesting technology in fact
11:39 uh
11:42 bat image and those words were made with
11:45 ideogram
11:47 so pretty cool huh pretty cool well
11:50 that's part of Coco I didn't know it
11:52 could do that oh yeah it can do that
11:57 um what else uh Leonardo and clip drop
11:59 are sitting on top of stable diffusion
12:01 which is another interesting tool if
12:03 you're geeky go play with automatic 11
12:05 11 but these are two non-geeky tools you
12:07 can play with and then prompt book that
12:10 will teach you all of the [ __ ] that
12:12 these things know and they know a lot
12:14 and it'll teach you how to talk to them
12:15 it's designed for stable diffusion but
12:17 it's good for all of them okay now
12:20 uh Luna stick what's happening howdy hey
12:22 hey hey hey hey hey hey
12:25 um openart.a slash prompt book uh I said
12:28 that future PD is a big directory of
12:30 services ai ai software basically so if
12:35 if I don't know something I just
12:36 generally Point people to Future pedia
12:38 and you can go find something there
12:40 all right so the other thing we do here
12:42 is I do my best to answer questions so
12:45 pop the questions down below I can also
12:47 demo stuff what I was thinking I would
12:49 do tonight is I'll start answering some
12:51 questions
12:52 and then at some point if it's just
12:53 getting a little too chatty and we
12:54 haven't done anything I'll just jump to
12:57 the bottom and ask you is there anything
12:58 you want to see and then I'll show you
13:01 if I know how to do it I'll show you
13:02 something uh I'll demo something all
13:05 right
13:06 Indiana and what else what else do we do
13:09 here uh I like to do silly voices
13:11 sometimes I do silly voices it's like
13:13 Saturday Night Live
13:16 without the talent you see what I'm
13:18 saying it's like comedy without the
13:20 laughter very very similar all right so
13:23 we do that
13:25 you know life's tough it's tougher when
13:27 you're stupid anyway lift them so I
13:31 talked um
13:32 and then I talk smart and then I talk
13:34 dumb and then I talk smart and depending
13:36 on when you come in you might think he's
13:37 just a dum-dum and that's that's
13:39 generally pretty accurate and then you
13:42 might come in and you might go it's
13:43 called AI learning lab what am I
13:44 learning here I don't know what do you
13:46 want to learn put it in the comments or
13:48 or you might come in here and say what
13:50 are your qualifications sir
13:52 who are you what is this all about
13:55 I don't have to answer to you I don't
13:57 why aren't you why are you being so
13:58 demanding I'm just an old man on Tick
14:00 Tock she's an old man on tick tock tick
14:03 tock tick tick tick tick tick tock
14:07 see what I'm saying sometimes I cuss if
14:10 you don't like the custom I would go
14:12 some some very
14:15 uptight man
14:17 probably lived in you know
14:19 a nice neighborhood
14:21 told me that my cousin was tacky
14:24 so apparently I'm tacky too I'm tacky
14:27 other people say I look like Rosie
14:29 O'Donnell
14:30 so if you like that if you like Rosie
14:32 O'Donnell here I am for you
14:35 all right
14:36 so we're gonna get started now what
14:40 all right enough comedy jokes
14:43 let's get started
14:46 come on down from the back
14:49 all right
14:54 Ted joined the Discord oh that's the
14:57 other thing
14:58 um if you uh are interested in joining a
15:02 community of like-minded people
15:06 like-minded like hell like dum-dums no
15:08 no people into generative Ai and
15:10 exploring it and getting curious with it
15:12 I created a group last
15:15 December called the AI Salon we meet
15:18 every other week in person here in
15:20 Denver and online and that's a link tree
15:23 that's got a link to kind of our mission
15:25 statement and our values to our Meetup
15:28 where you can sign up to get
15:29 notifications about the meetings and to
15:31 our Discord where you can go
15:34 um you know play with mid journey and
15:36 ask questions and do Show and Tell and
15:38 all sorts of [ __ ] like that so feel free
15:41 to explore that group
15:44 all right folks let's get started let's
15:48 get this show on the road
15:50 all right uh we did the whole AMD and D
15:54 campaign cabruno what's happening Dino
15:56 Dino Dino
15:58 when will ABI used in Minecraft craft
16:01 mods I'm pretty sure it's already being
16:03 used in Minecraft mods isn't it
16:12 Minecraft AI
16:17 AI can now play Minecraft as well as you
16:22 that's not really a Minecraft mod here
16:25 special AI Minecraft mods
16:29 simple AI Minecraft version there's a
16:32 lot of stuff so what I would do Bijan is
16:35 and hello by the way what I would do is
16:37 I would go to Reddit
16:40 look up Minecraft AI mods just basically
16:43 do that I guarantee you there's a
16:46 subreddit where there's I'm guessing
16:48 that there are a pile of uh
16:52 projects
16:54 that are githubs or Google collab
16:57 notebooks that you can play with that
16:58 are doing all sorts of fancy [ __ ] and
17:00 then I know I know
17:02 um
17:04 I'm thinking blockade Labs that's not it
17:07 um Roblox I know Roblox
17:10 um has announced that they're
17:11 incorporating or may have Incorporated
17:14 let's see Roblox
17:19 AI
17:22 Roblox blog in February
17:26 our vision for the future
17:31 yeah but that was February what are they
17:33 doing now I know you have a vision for
17:36 the future have you actually done
17:38 anything
17:44 March 21st Roblox launches first
17:47 generative AI game creation tools yeah
17:49 so Roblox and Minecraft I think there's
17:51 a ton of stuff out there already
17:55 um I'm not deep in that world I don't
17:57 know it but I know it's I know there's
17:59 there's some [ __ ] going on and it gonna
18:02 be it gonna be exciting
18:04 it gonna be exciting I'll tell you that
18:06 it won't be good
18:08 all right
18:13 hello Stevo what's shaking Dr J howdy
18:17 guys hope you're all having a regularly
18:19 irregular weekend by the way the uh if
18:22 you see folks in the comments and
18:24 they're a little odd and just just off
18:27 those are the Irregulars so the folks
18:30 that regularly show up here have coined
18:33 themselves named themselves coined
18:35 themselves
18:37 labeled themselves they're the
18:39 Irregulars and they're awesome and and
18:42 they'll actually the the one of the
18:44 things that's really powerful about this
18:47 community that's forming here is you
18:50 know I'll get answering the question
18:51 I'll sort of run down a rabbit hole for
18:53 15 minutes
18:54 um folks that have been here before will
18:55 answer questions like why isn't Kyle
18:57 answering my question or or you know is
19:01 there a specific tool I need to be using
19:03 so there's kind of a little support
19:05 network with with folks in the uh in the
19:08 comments
19:09 um which which I really appreciate and
19:11 at some point here
19:13 um I'm gonna have some of them be mods
19:15 on the channel and this should get we're
19:17 getting more and more professional as we
19:19 go I you know I'm just like an old dude
19:22 learning to use tick tock only been
19:25 using it for you know like nine months
19:28 or so so it's a little it's a little
19:30 slow you know you know with with older
19:33 people and Technology if you're one of
19:36 them young whippersnappers in here going
19:37 isn't that cute the elderly teaching
19:40 each other how to use the AI
19:43 go [ __ ] yourself all right very nice
19:47 all right sorry my French brain is
19:50 limited so I probably don't understand
19:51 what I may read tokas is awesome so guys
19:56 you can just write in full-on French and
19:58 then I'll just translate it we have the
20:00 we have the technology for that
20:03 um I'll still answer you in English
20:04 though we don't have the technology to
20:07 synthesize my voice yet and and speak it
20:09 in French but you can write in French
20:14 ion China very cool hi this I understand
20:17 I
20:20 like Claude I Like Turtles
20:23 um
20:24 [Music]
20:32 [Music]
20:33 well the traditions we have here at the
20:35 AI learning lab is uh back here live at
20:38 the waterfront village with my friend
20:40 the zombie Jonathan you're looking good
20:42 Jonathan just got an awesome face paint
20:44 job what do you think I like turtles all
20:48 right you're great zombie good times
20:51 here at the waterfront Village open for
20:53 the next 11 days
20:55 I bet that woman talks about that
20:59 many times a week to this day
21:12 uh oh my God Bing's whiny yeah Bing's a
21:16 little [ __ ] lately the other day I said
21:19 this on here but I was doing something
21:21 and it was just like I'm afraid I can't
21:23 answer I'm sorry Dave I'm afraid I can't
21:25 do that I got one of those from Bing and
21:28 and I I didn't I don't think I told it
21:30 to go [ __ ] itself but it was something
21:31 in the neighborhood it was like it was
21:33 like some natchy nasty [ __ ] slap back
21:35 and and um
21:38 and Bing was just like well we're then
21:40 we're just not going to continue this
21:42 conversation are we and it reset the
21:45 conversation I'm like
21:47 it's like come on
21:49 oh let's have fun let's ask Chachi BT
21:52 for training regime an estimated cost
21:54 breakdown for a K2 climb okay we can do
21:56 that that sounds fun curvaruno how's
21:59 your holiday weekend going Kyle it's
22:00 going good Stevo I um
22:02 I worked on putting together some merch
22:06 I chose printful as the as the vendor
22:09 because it can hook into tick tock
22:11 and I started making some stuff and then
22:15 it let me get all the way through the
22:16 design process and I was just like no
22:18 I'm not going to make the product this
22:19 it it you know it failed to make your
22:23 product because of an unknown error
22:25 contact customer support if you want to
22:27 deal with this so that annoyed me but
22:30 you know whatever
22:32 technology is just stupid sometimes
22:37 it will be so much nicer when generative
22:39 AI can do things like actually answer
22:41 questions what is that [ __ ] error Why
22:44 didn't it work
22:45 maybe you could diagnose the bug and
22:47 send that to the development team and
22:49 actually tell me what's going on without
22:50 me having to wait for customer support
22:53 to show up all right coveruno custom and
22:57 strike we'll do this with custom
22:58 instructions too
23:01 all right so here we are at chat gbt
23:06 with a very dirty screen
23:11 it's like it's like my laptop came from
23:13 burning man so
23:15 it's all muddy
23:19 here we are chat gbt
23:22 um if you don't know within chat GPT
23:24 there's a thing called custom
23:26 instructions
23:27 and I did kind of a joke version of this
23:30 last night
23:32 but what custom instructions do is it's
23:34 it's a form of fine-tuning it's what I
23:36 think it's actually fine-tuning where
23:39 you're going to basically tell chat gbt
23:41 either who it is or who you are
23:44 and how you want it to respond and so
23:47 yesterday I I said you're a pompous over
23:50 intellectual academic who fancies
23:52 themselves a theatrical you take great
23:54 pleasure in using large words that
23:58 neither you or your audience understands
24:00 in an attempt to make yourself seem way
24:02 smarter than they are so so I kind of
24:05 did this joke thing and actually this is
24:07 fun I'll turn it on and we'll play so
24:09 basically what I said is no matter what
24:11 anyone asks
24:13 respond with an obscure reference to a
24:16 Broadway play or a piece of overly
24:17 flowery literature and include a quote
24:19 from that piece
24:21 um
24:22 when you answer
24:24 um answer them as if you're Tennessee
24:26 Williams writing a masterpiece so I'm
24:29 going to save that so now no matter what
24:30 I do within Chachi BT I say uh I'm
24:36 thinking about Mountain
24:40 climbing
24:43 um
24:44 how should I prepare
24:47 question mark so this is what converter
24:49 wants to talk about but now that we have
24:51 this theatrical thing in there
24:55 oh ah my dear interlocutor the Endeavor
24:59 of ascending the Towering Peaks peaks of
25:01 Nature's Cathedral is akin to the
25:03 protagonist Ardent journey in Tennessee
25:05 Williams opulent tale The Glass
25:08 Menagerie
25:09 as Laura Wingfield that delicate emblem
25:12 of fragility once mused time is the
25:15 longest distance between two places
25:17 similarly the passage of time shall be
25:20 your steadfast companion so it goes on
25:22 and it's really [ __ ] obnoxious okay
25:23 that's what we told it to do so now what
25:25 we're gonna do
25:27 is we're gonna go get a different custom
25:29 instructions we're going to say
25:33 you are
25:37 an experienced mountain climber
25:40 who has successfully ascended the
25:43 world's most challenging Peaks period
25:52 your expert at preparing
25:56 people
26:02 so that they are prepared
26:06 when they are ready to ascend
26:12 mountains such as Everest and K2 period
26:22 now the the next thing is how do you
26:24 want it to respond and so what we'll say
26:26 here is we'll say
26:30 as a coach
26:35 it is your duty to make sure
26:38 that the people you are talking to are
26:41 prepared for their Journey period
26:46 make sure you ask enough questions to
26:48 understand who they are
26:52 and what their goals are
26:54 and when appropriate
26:57 provide them with
27:05 expertise
27:09 training regimes
27:11 and estimated costs
27:14 for their trip
27:18 okay
27:19 all right so now we have some
27:22 flipping wild ass mountain climber dude
27:24 I'm just going to do this in chat GPT
27:26 3.5 because it's fast
27:28 okay so now we're going to say
27:31 I'm getting ready
27:32 to make a trip
27:34 to ascend K2
27:37 period can you help
27:43 of course I'd be happy to help you
27:44 prepare for your Ascent to begin let's
27:46 gather some information to better
27:48 understand your experience Fitness level
27:50 and goals could you please provide me
27:52 some details of your previous mountain
27:54 climbing experience have you climbed any
27:56 challenging Peaks before this will give
27:58 me a clearer picture of your background
28:00 and help Taylor
28:01 the training accordingly
28:04 I went up a long flight of stairs once
28:09 [Laughter]
28:14 it's great to hear that you have some
28:16 experience climbing stairs however K2 is
28:20 one of the most challenging and
28:22 dangerous mountains to ascend so it's
28:24 important to be well prepared given your
28:27 limited climbing experience we'll need
28:29 to focus on building up your physical
28:30 fitness let's start by uh discussing
28:33 your current Fitness level additionally
28:35 have you received any formal training
28:36 lastly do you have a timeline in mind
28:39 um so let's see
28:42 my fitness level is low or beginner
28:46 comma I don't have any formal training
28:49 period
28:50 and
28:51 I want to Summit K2
28:55 the first week of October
29:01 you tell me you're not allowed to go
29:05 thank you for finding your details
29:08 giving your beginning Fitness level and
29:09 the ambitious goal of summoning K2 the
29:12 first week of October we need to
29:14 approach your preparation strategically
29:17 since we're starting at a low Fitness
29:18 level focus on gradually
29:20 so anyway it's doing it now so now I'm
29:23 just gonna say
29:24 um can you please give me a
29:28 weekly
29:31 Fitness regime
29:35 followed by
29:37 a cost breakdown of the trip period
29:40 format
29:42 both of those using markdown so they are
29:45 easier to read
29:47 period
29:48 so markdown is oops
29:51 markdown is it's kind of like an HTML
29:55 like thing that'll do some formatting
29:57 so notice how it did like a bold a bold
30:00 thing you know bolds and headlines and
30:02 things like that so the weekly Fitness
30:05 regime weeks one to four build Basic
30:07 Fitness three to four days a week of
30:09 brisk walking light jogging or cycling
30:11 for 35 30 to 45 minutes strength
30:14 training
30:15 uh Arab uh aerobic exercise hiking
30:19 strength training hiking enroll in a
30:22 mountaineering course right so it
30:23 recognizes that I'm a clueless newbie
30:26 and so it says go go understand what the
30:29 hell you're getting into cost breakdown
30:31 for the trip flights 1500 2500 yeah yeah
30:36 sure sure
30:39 climbing permit for K2 10 to 20 or 10 to
30:42 12 Grand National Park and conservation
30:44 fees 500 to 1000 guided Expedition
30:47 package 20 to 30 grand
30:50 gear 1500 to 3 Grand mountaineering
30:54 course thousand to two thousand
30:56 so there you have it total range 40 to
30:58 60 Grand excluding personal expenses
31:02 boom
31:04 there you have it so that should help
31:05 with the uh understanding Uh custom
31:08 instructions too where you can sort of
31:10 Bend Bend this thing to your to your
31:12 will
31:13 you could also put examples of your
31:15 writing in there like if you wanted it
31:17 to write
31:18 you know business emails rather than
31:21 just saying hey you're someone good at
31:23 writing business emails you could say
31:24 hey you're the CEO of this company that
31:27 does these things here's an example of
31:29 an email that I've sent that really
31:31 captures my tone and the structure of
31:33 how I like to structure emails
31:36 and then when you say you know whenever
31:38 I ask you to write an email please
31:39 follow the tone of that and follow the
31:41 structure and you know any other
31:43 writings you know should be inspired by
31:45 that as well it'll do that it'll do it
31:47 like that brilliantly
31:51 so yeah yeah there we go there you have
31:55 it yeah well isn't that special going to
31:58 K2 without any training yeah well who
32:01 might be telling you to do that huh
32:03 person that wants you to meet your maker
32:05 could that be Satan
32:09 oh it could
32:13 and perplexity does well for me cool
32:16 Brian Engel thank you for sharing that
32:18 that's cool yeah like I haven't my
32:21 problem with perplexity is just that I
32:22 haven't used it it's one of those things
32:24 where like I keep going I gotta try
32:26 perplexity the handful of times I've
32:28 tried it I didn't really know what it
32:30 was so I kind of went in thinking it was
32:32 Chach EBT and it wasn't and I was like I
32:34 don't I don't want to do this but I just
32:36 haven't played with it so anyway
32:40 all right let's see evening you all have
32:43 to be quiet because I'm on a call with
32:44 work oh Vicky
32:47 you're you're sneaking sneaking some uh
32:51 AI learning lab on the side here don't
32:53 get me in trouble
32:54 they're just gonna be like who's that
32:56 idiot in the background I don't want to
32:57 talk about it right now Vicky
33:00 thank you
33:03 I thought it was on mute
33:05 he's just some idiot
33:08 [Laughter]
33:12 Luna stick what's happening Dr J good
33:15 day good day Nancy's here hello very
33:18 nice to see you
33:20 Nancy Nancy all right guys it's not the
33:24 real you maybe it could be not the real
33:26 me I might be an interloper what if I'm
33:29 the scammer we had a scammer here that
33:32 was all exciting for a while
33:34 Tick Tock finally got rid of them I hope
33:35 they don't come back
33:37 you'll let me know if you find anyone
33:38 but
33:40 um he had duped duped my account and was
33:42 acting like me trying to get people to
33:44 sign up for his stupid crypto [ __ ]
33:49 today
33:53 uh if you have comments or questions pop
33:56 them in the in the
33:58 in there
34:00 words
34:02 some people have a way with words other
34:04 people
34:06 not have way I guess you'd say
34:09 name that comedian
34:12 all right let's see let's go on shall we
34:19 um
34:21 interested in learning how to use a uh
34:23 use chat GPT for questions only okay so
34:27 here is a thing I can offer you
34:30 go
34:32 to Chachi PT
34:34 and then in a separate browser open up
34:36 this document prompts.chat
34:39 that will teach you how to talk to these
34:42 things here's the biggest thing I can
34:46 I can share with you that that really
34:48 opened my eyes
34:50 and and
34:52 just made things better
34:54 if you treat chat GPT like Google
34:59 it's shitty
35:00 so basically if you give it a shitty
35:03 prompt it's going to give you a shitty
35:05 generic answer and it's just it's the
35:07 nature of how these things actually work
35:09 they need a lot of context because
35:11 they've been trained on so much data
35:13 there's there's
35:14 so much data so basically these were
35:17 trained on as much of the public
35:18 internet as openai could get their hands
35:21 on for the gpt4 model and the gpt3 model
35:25 and you know they're doing more and more
35:26 each time but but the fancy one right
35:28 now is gpt4
35:30 so so just greatly simplified
35:34 sitting within gpt4 is everything that's
35:38 put on been put on the public internet
35:40 for the past 35 or 40 years okay so it's
35:44 a lot it knows a lot
35:46 the problem with that is
35:48 what's on the internet for the past 35
35:50 or 40 years is good stuff bad stuff
35:53 inaccurate stuff crappy stuff brilliant
35:56 stuff right and so when you just say
35:59 answer this question for me it will sort
36:02 of look at all of the internet and it
36:04 basically is just a probability engine
36:06 and it goes word by word and says what's
36:08 the next most probable word to to answer
36:10 this question so if it doesn't have a
36:12 lot of context if it doesn't have a lot
36:14 of details about what you want it's just
36:16 pulling from any old wear and you'll
36:18 tend to get really generic [ __ ] or more
36:22 what they call hallucinating or just
36:24 answer is weird and lies it's you know I
36:28 like to call it mansplaining as a
36:29 service like like every good man in your
36:31 life it will confidently answer you even
36:33 if it's wrong
36:35 aren't we awesome
36:37 um
36:38 so that's so so it's like that if you
36:40 give it kind of kind of weak kind of
36:42 Google kind of answers right Google
36:44 taught us just give me just the facts
36:46 ma'am
36:47 um
36:48 so what prompts.chat will teach you
36:50 first of all there's some primers in
36:52 there that will teach you you know talk
36:53 to you about how large language models
36:54 work and how to talk to them
36:56 it will also uh if you were here when I
36:59 just did the custom instructions there's
37:01 a bunch of prompts in here where you can
37:02 tell it to act like a screenwriter or
37:04 act like a
37:06 um a strategist or act like a marketing
37:09 expert or act like a nutritionist or act
37:11 like a mountain climbing dude
37:13 um
37:15 and and that really helps provide the
37:17 context so that's so tip number one is
37:19 just start to understand how they work
37:20 and how they're different than what
37:21 we're used to the second thing is this
37:27 these things get much better if you kind
37:29 of have a rolling conversation with them
37:31 so rather than trying to write the
37:33 perfect single prompt
37:36 start a conversation say hey here's who
37:38 I want you to be I want you to be an
37:40 expert at this kind of thing and here's
37:41 who I am and what's going on thank you
37:43 Nicholas for the hottie hot hot
37:46 um
37:47 and here's what I've got going on and
37:49 here's the problem I'm trying to solve
37:51 do you understand yes I understand and
37:53 based on what you've told me and it'll
37:54 give you a whole bunch of [ __ ] and then
37:56 you're like okay great glad you
37:57 understand
37:58 but what I really want to know is X Y or
38:01 Z and it'll give you an answer and if it
38:02 gives you a crappy answer tell it it
38:04 gave you a crappy answer hey I happen to
38:06 be an expert on this and the point
38:09 number three that you made is actually
38:10 inaccurate oh I'm so sorry you're
38:12 absolutely looks like we lost internet
38:14 do me a favor tap the screen if you're
38:17 um okay
38:18 you can still hear me
38:21 all right
38:23 um
38:30 yeah so try that
38:32 that'll help a lot
38:34 uh what are you drinking look good this
38:36 is just a lovely uh morning coffee
38:39 that's been sitting in the pot all day
38:40 with ice and little leche
38:45 and then this is ice cold Rocky Mountain
38:48 tap water
38:49 it's been filtered don't I'm not a
38:51 heathen
38:52 come on you people
38:54 right out of the fridge door
38:58 how dare you
39:01 even Champy's upset
39:03 all right
39:05 you even make money
39:10 that's oh if you're wondering how do I
39:13 make money with chat gbt step number one
39:16 don't listen to those people
39:19 they're not making money with Chad gbt
39:21 they're making money telling people they
39:24 can make money with chat EBT and get
39:25 them to buy a course
39:29 and you're full of [ __ ]
39:36 all right
39:37 but I can talk about how you make money
39:39 with chat gbt it's just you won't like
39:41 the answer
39:43 uh which one of these is
39:45 is the best at training AI
39:50 I don't quite know what you mean trait
39:52 you mean training you to be able to use
39:54 AI
39:56 um just right now the if you're just
39:59 getting started with this stuff
40:01 go to prompts.chat that will teach you
40:03 how to talk to them and then chat GPT is
40:06 really the best one to start with
40:07 chat.openai.com
40:09 it's just the best
40:11 there's there's reasons to use all these
40:12 other ones and I can go into that if
40:14 people want to know more about that but
40:16 if you're just getting started
40:17 chat.openai.com
40:23 all right your hand on the keyboard
40:25 looks like the thing from The Adams
40:26 Family thank you thing
40:29 actually that reminds me when I was
40:33 five or six
40:37 we went on a vacation to Rehoboth Beach
40:40 Delaware
40:41 and I remember there are other kids
40:43 there
40:44 and
40:46 right before bed one night we watched
40:49 this black and white 50s horror film
40:51 called the hand
40:53 and the hand was like a severed hand
40:55 like severed like halfway down the
40:57 forearm and it would just crawl around
41:01 and it would like crawl up the side of
41:03 dressers and then it would it would just
41:06 choke a person it would just choke them
41:08 out and so it was just terrorizing this
41:12 community choking people out
41:16 I could not sleep with my hands or my
41:20 feet like hanging off the bed
41:24 probably until my late 30s
41:29 [ __ ] me up for most of my life I don't
41:32 know like I must have just had enough
41:34 like you know
41:37 trauma or or uh or therapy
41:41 um that it just sort of faded away and I
41:43 could now just kind of sleep but it was
41:45 like Decades of my life I was I would
41:49 sleep like
41:50 curled up like this with all the covers
41:53 wrapped around because I didn't want it
41:55 to snap off my wrist or pull me under
41:57 the bed
41:59 so funny [ __ ] like that
42:02 uh you have gorgeous hair bro thank you
42:05 very much it's it's
42:06 I don't know gorgeous but thank you I do
42:10 have hair so at my age
42:12 at 58 the fact that I've got some is is
42:15 solid right and I got the little wisdom
42:17 patches well wisdom
42:20 very nice hello
42:24 I'm wise
42:28 um what do you feel is an untapped pain
42:31 point
42:33 for business that an AI solution could
42:35 benefit largely
42:37 it's a really good question an untapped
42:40 pain point
42:46 I think the biggest for for me the
42:49 biggest
42:53 the most universal pain point in in
42:56 business
42:57 and for human beings
43:00 is that the systems that we interact
43:02 with
43:04 are either by Design
43:07 or just by evolution
43:11 way over complicated like have you ever
43:14 used Salesforce it's a [ __ ] travesty
43:17 right of just it's it's just like
43:20 feature vomit and then on top of the
43:23 feature vomit that is the platform
43:25 every sales consultant in the world has
43:27 their own sales strategy so there's all
43:29 these custom versions of Salesforce so
43:31 Salesforce just a [ __ ] nightmare
43:35 everything's like that
43:37 oh we're going to use Trello to do
43:40 project management okay well you got to
43:41 set up Trello and everyone's got a
43:42 different [ __ ] thing and we're going
43:44 to use kanban and then we're going to
43:45 use this and then we're going to use
43:46 that and then we're going to use
43:47 Microsoft Word then we're going to use
43:49 sheets and just everything's [ __ ]
43:50 complicated
43:52 we're gonna do single sign-on but
43:54 there's there's 15 different logins for
43:57 each person because something [ __ ] up
43:59 back in 1987 and it hasn't been fixed
44:01 yet and so there's this ever cascading
44:04 nightmare of complicated [ __ ] and
44:07 um half of the companies on Windows and
44:08 half of them's on Mac and and the ones
44:11 on Windows know how to fix their
44:13 preferences but but most of them are
44:15 don't aren't willing to do it the ones
44:17 on Mac don't know how to do it because
44:18 they've never been taught to do it and
44:20 just everything's [ __ ] complicated
44:24 so
44:26 so I think the thing that is
44:31 the greatest burden is is just
44:33 everything takes longer than it should
44:34 right
44:36 you have a zoom call
44:38 what happens after the zoom call well
44:41 Sally took notes and Jim took notes but
44:44 Sally's really detailed but Jim captures
44:47 the Big Ideas so let's make sure that
44:49 Jim and Sally have a follow-up meeting
44:51 after the call to make sure that they
44:54 get Sally's detailed but Jim's Big Ideas
44:56 and get that into notes and then they
44:58 should collaborate and they should write
45:01 an email back to the client recapping
45:03 the meeting they should capture the
45:06 to-do's that everyone had you want to
45:08 make sure that that Jim because he's got
45:10 a strategic mind gets something to the
45:13 lead account person right so
45:15 everything's [ __ ] complicated
45:20 so now
45:23 as of I think Friday
45:26 it looks like Google has announced
45:28 General availability of duet for the
45:31 Google Workforce products and that will
45:33 also be coming for Microsoft
45:35 office 365. and So within Google meet
45:39 for example it's now going to take notes
45:42 for you and it's now going to pull out
45:44 to do's and it's now going to recap what
45:46 was said
45:48 and you're going to be able to just pull
45:50 that out instantly so I think I think if
45:53 if I were to take a a generic approach
45:57 to how I would think about incorporating
46:00 all of these AI Tools in my business is
46:03 I would just look at where is the [ __ ]
46:06 ridiculously over complicated
46:09 and then I would just try to simplify it
46:11 how can this [ __ ] simplify it maybe you
46:13 do an automation using zapier where
46:15 you're doing calls to open Ai and
46:17 anytime when anytime someone asks a
46:20 question in this slack Channel it
46:22 actually goes you know to the internet
46:24 and searches for news articles and then
46:27 it sends the links of those to open Ai
46:29 and it writes a summary and posts that
46:31 back to slack so you could have a little
46:33 research Channel there I don't know
46:35 maybe that because maybe you have a
46:37 research Channel and it's just clogged
46:39 up with people having side conversations
46:41 and I found this and I found that and I
46:43 found this and I found that and four or
46:46 five hours of every every day are just
46:48 wasted with people screwing around sort
46:51 of re searching but not quite getting
46:53 there and it's because it's over
46:55 complicated
46:56 so I think simplifying the complicated
47:00 is the thing that's going to make these
47:02 things I think it's the killer app for
47:03 for AI the generative part is really
47:06 interesting generate content generate
47:08 all this but in the end it's really just
47:10 about simplifying right you take simple
47:12 inputs you process them through this
47:15 magic machine and you get out remarkable
47:17 answers on the other side if you haven't
47:20 played Brian with um chat GPT is what
47:23 used to be called code interpreter it's
47:25 now called Advanced data analytics
47:30 you can upload files to it like CSV
47:34 files or PDFs and start have it will
47:38 start analyzing that data for you
47:41 and it'll even create charts and graphs
47:43 I did one last night where someone said
47:45 can you
47:47 what's a tool that I can use to map Geo
47:51 data onto a map right and so we went and
47:55 got some fake data we got some Chipotle
47:57 locations and in the data set that we
47:59 got there were all the Chipotle
48:01 locations and then there was Json that
48:03 described the shape of all the states
48:05 and So within I don't know 10 minutes
48:08 this thing analyzed what the data set
48:11 was what was in it and then I said
48:13 create a map of the United States with
48:15 all of the locations of Chipotle on them
48:18 and it did it in it just wrote python
48:21 code executed it so that idea of taking
48:25 things that are overly complicated and
48:27 and then figuring out what AI tools
48:29 might simplify that I think that's the
48:31 biggie where I would start my first time
48:34 here very interesting awesome Kim
48:35 Welcome welcome welcome welcome
48:37 this is we this is what we do here it's
48:39 just if you're like there doesn't seem
48:43 to be any shape to this it seems chaotic
48:45 and random
48:46 a lot of the Irregulars here call it
48:48 chat add
48:49 [Laughter]
48:54 welcome to chat add
48:56 where you can
48:59 return
49:01 [Laughter]
49:06 well there's what we do we like to have
49:08 fun we like we like to mess around a
49:10 little you know we mess around
49:12 we uh
49:14 we joke we joke we joke but then we get
49:17 serious because sometimes it's serious
49:18 you know sometimes really serious I'll
49:20 get up on a soapbox and I'll talk about
49:21 stuff other times not so serious right
49:23 black life
49:25 like life you know what I mean live
49:27 sometimes serious sometimes fun
49:30 a lot of times it sucks you know life's
49:32 just a grind you know we're all just
49:34 hanging in there man
49:36 and then Along Comes AI it's like you
49:38 got to reinvent yourself you're gonna be
49:40 out of a job in a week
49:42 aren't you excited the robots might kill
49:44 us all yay
49:48 help me help me
49:50 oh sweet baby Jesus oh
49:53 yeah
49:55 yeah
49:56 all right and scene AMC we're gonna calm
50:01 it down we're gonna bring it down a
50:02 notch all right and bring it down so
50:04 that's you Kim
50:06 had a little moment I don't even know if
50:08 you're still here but if you're still
50:09 here welcome Silver Fox is here
50:12 all right
50:18 me too when you're on Tick Tock I know
50:20 me too I'm coming up no
50:24 yeah probably about a year I think I did
50:26 my first
50:27 Tick Tock channel the Kyle Shannon
50:29 dreams Channel I started that in I don't
50:32 know January or something like that so
50:36 that was fun
50:38 that was fun like do you ever join a
50:41 platform where you just resent it like
50:43 you just you just resent having to learn
50:45 something new I'm like I don't know how
50:47 to use this [ __ ] Tick Tock I'm an old
50:49 man and I don't care and it's not
50:51 YouTube and it's an [ __ ] you're all
50:52 [ __ ] and why is every why don't the
50:55 [ __ ] tap everything and why is it
50:57 ever why is everything small and why is
50:59 it all hidden behind these [ __ ]
51:00 stupid and how do I do this and how do I
51:03 and how why do I have to watch YouTube
51:05 videos on how to do Tick Tock it just
51:07 doesn't make [ __ ] sense so I was like
51:10 that for like two months
51:15 and that was kind of how it was if you
51:16 were here from the early days of my
51:18 lives that's how I was with the lives
51:20 I'm just like I don't know what all this
51:21 [ __ ] is it's awful but I did figure out
51:24 my little window here which impressed
51:26 people in the early days they're like
51:27 how do you get that dual screen you must
51:30 be some side of tick tock wizard
51:35 um
51:37 all right
51:38 you'll be ex-human too do you think
51:41 neural link will have ai well I think I
51:44 mean neural link is just basically
51:47 um instant instant access to the
51:50 computer and if the computer has AI then
51:52 yes it will neural link will have
51:53 instant access to
51:56 um to the AI I think the gating Factor
51:58 will be the speed of AI processing so
52:00 current AI because it's got you know a
52:03 half a second lag
52:05 um
52:06 is going to be slow but if you look at
52:08 if you pay attention to what um
52:13 what Apple announced with the Vision Pro
52:18 um
52:19 there would there's some
52:22 really fancy [ __ ] technology you know
52:24 there were there are 9 000 patents in
52:28 The Vision Pro goggles nine thousand
52:32 um some crazy ass [ __ ] like that
52:34 um but but one of the things that
52:36 Apple's built into that Hardware is this
52:38 thing called the R1 chip so it's got all
52:40 these cameras and sensors and just all
52:41 this stuff going on
52:43 there there's four cameras you know two
52:47 pointing at each eye
52:48 and one of the things that
52:52 um
52:53 that the Vision Pro does which this is
52:56 mind-blowing
52:58 so it's looking at your eyes right
53:01 it's actually measuring micro
53:03 contractions of your iris to determine
53:06 emotional response
53:09 so the way the interface Works
53:12 apparently I haven't played with it but
53:13 the way the interface works is when you
53:15 look at an icon it'll select it
53:19 what what's actually happening in this
53:22 R1 chip is that
53:25 it's taking the emotional responses from
53:28 from your iris contractions and as you
53:31 look toward an icon it knows if you're
53:34 about to click on it
53:36 and if you're about to click on it it
53:38 starts pre-loading that application in
53:40 the background so that it so that the
53:42 system appears more more instant than it
53:45 actually is so they're solving something
53:48 that neurolink also solves which is how
53:51 do you anticipate how do you sort of
53:52 essentially think faster than humans can
53:54 process
53:56 um so that it's it occurs as magic
54:00 um so Apple's going to be able to do
54:02 that without sticking something in your
54:04 brain
54:06 um and and I think a lot of a lot of
54:07 companies will follow suit but yeah I
54:09 think neuralink's really interesting
54:10 there's a um
54:12 I I mean is it sort of creepy and weird
54:15 yeah a lot of this stuff is creepy and
54:16 weird I just got my Iris skin to get a
54:20 uh blockchain ID that verifies I'm human
54:23 Sam Altman's new company called
54:24 worldcoin
54:26 um you can get it you can get it done in
54:28 New York San Francisco and Miami right
54:30 now I think that's the only place in the
54:31 states although but at this point
54:33 they're probably expanding it but I did
54:35 that why did I do that well I think
54:37 verifying people having a verified human
54:40 on the other end of a transaction is
54:41 going to be really important in a world
54:44 where we got fake [ __ ] everywhere so I
54:46 think it's actually important and I also
54:47 think it's just an interesting
54:49 experiment
54:51 um
54:51 but um so so a lot of this stuff's kind
54:54 of creepy but there's things like there
54:56 was a guy in Italy uh
54:59 um he was paralyzed from a motorcycle
55:01 accident
55:03 for the last 12 years and they use
55:05 generative AI they did brain scans of
55:08 him
55:08 fmri scans when he was thinking about
55:11 moving his legs
55:14 and then they took electrodes
55:17 from uh from they had him think about
55:23 um moving his legs they pressed that
55:25 that signal through AI
55:27 sent a Bluetooth
55:30 signal around the injured part of his
55:33 spine two electrodes in his lower spine
55:35 and now he's walking
55:40 so I know that the initial neural link
55:44 applications are going to be [ __ ] like
55:46 that you know brain injuries spinal
55:50 injuries
55:51 um
55:52 you know other kinds of brain conditions
55:54 where you can essentially route around
55:56 certain parts of the brain or certain
55:58 parts of the nervous system and you can
56:00 do it with instant access with you know
56:02 almost no latency
56:04 so it's interesting we'll see
56:06 let's see you know is everyone going to
56:09 go out and get one of those no but
56:10 there's a lot of people that it could
56:12 make a massive difference for
56:16 and then if you look at what Apple was
56:18 doing with the sort of eye stuff
56:20 it's not going to necessarily be need to
56:23 be invasive at some point you'll
56:25 probably be able to put on a baseball
56:26 hat that's got some sort of wacky ass
56:29 technology in it that effectively does
56:30 what neurolink does without having to
56:32 have surgery
56:34 and by the way if you haven't looked at
56:36 neuralink
56:38 saying what it does so basically they
56:40 came up with this idea
56:42 you drill a hole in someone's skull and
56:44 you can put electrodes into the brain
56:46 right now these little microelectrodes
56:48 and I think there's 60 or 120 little
56:51 wires on a bundle for each little thing
56:53 and I think they put in three or four of
56:55 them
56:56 but what they realized was
56:58 you couldn't there was no way to
57:01 actually install it without building a
57:03 robot first so so they built a robot
57:08 that
57:09 um
57:10 in real time figure figure figures out
57:14 where to position these these 60 or 120
57:18 electrodes
57:19 so that they that they don't
57:23 um put them into any blood vessels so if
57:25 you've got two blood vessels in your
57:27 brain it it sort of does that in real
57:29 time and it it'll Slide the electrode in
57:32 between the blood vessels so it says to
57:34 minimize any sort of bleeding you know
57:37 in the brain it's just this is wild wild
57:41 technology so anyway that was approved
57:44 for human trials about three months ago
57:46 so
57:47 it's possible there are people doing I
57:49 haven't really paid attention to it
57:51 um every time never gets old I don't I
57:53 assume that was uh you can you make
57:56 money in the church oh no it was the
57:58 turtle thing
58:00 um there was an attempt is my favorite
58:02 subreddit oh that's good that's awesome
58:06 um Burning Man Burning Man Burning Man
58:08 is all fire Fest I know it's so true
58:12 they're really trying to outdo firefest
58:14 it's firefest 2.0
58:18 I heard that Chris Rock was was stranded
58:21 in there and then I just saw a tick tock
58:22 video before I came on where he because
58:25 he's Chris Rock he got a a ride out of
58:27 there so Chris Rock's safe
58:33 someone else I said said
58:38 I forget it was it was like
58:41 a bunch of a bunch of rich white women
58:44 with dreads are stranded in the desert
58:47 or something like that it was pretty
58:49 funny hey dang uh didn't get an alert
58:52 was watching some of your prompted
58:54 videos great oh yeah I put up a video
58:57 today
58:58 called the ten thousand hour rule is
59:00 dead
59:02 and so there's a guy named Malcolm
59:04 Gladwell who um wrote a book called
59:06 outliers lots of other books really
59:08 brilliant guy and in the book outliers
59:10 he talks about this idea of
59:13 to attain a level of expertise or
59:16 Mastery at a given topic you you know
59:20 effectively need 10 000 hours and and
59:22 you know if you kind of do the math
59:24 if you spent like eight hours a week for
59:27 you know six to or yeah eight hours a
59:29 day for six days a week
59:31 it's about four years it might be three
59:33 it might be five but it's a significant
59:36 commitment
59:38 and
59:39 what hit me was
59:44 one of the reasons that it's really
59:47 terrifying to think I want to try
59:49 something new is that I think we all
59:51 know instinctively that to get good at
59:54 something it's a major [ __ ]
59:55 commitment right like
59:57 you know I always wanted to be into
59:58 Neuroscience oh well then I would have
1:00:01 to learn about biology and chemistry and
1:00:04 this and that and do I go back to
1:00:05 college or do anything and then and then
1:00:08 you say okay [ __ ] it I'm gonna do that
1:00:09 and you take a certificate course and
1:00:11 then you get into college and you're
1:00:12 like three years into this process and
1:00:14 you realize I don't really like
1:00:16 Neuroscience but at this point you're
1:00:18 [ __ ] three years in and you probably
1:00:19 have student loans and you're right
1:00:23 well with this AI [ __ ]
1:00:26 like like my my breakthrough with chat
1:00:29 gbt was with without knowing python at
1:00:31 all in 90 minutes I wrote a a full
1:00:34 python application in 90 minutes and it
1:00:37 was pretty sophisticated it made API
1:00:39 calls to open Ai and it was insane it
1:00:41 was mind-blowing
1:00:43 never written a line of python in my
1:00:44 life and I thought you know if I wanted
1:00:46 to to write that same basic application
1:00:48 that I wrote it traditionally it'd
1:00:51 probably take me about a week right so
1:00:53 eight hours a day for five days that's
1:00:55 40 hours
1:00:57 I did it in two hours 90 minutes
1:01:00 so it was 20 times faster to do that one
1:01:03 little task
1:01:05 so even if you cut that in half to 10
1:01:07 times faster
1:01:09 so if you can use AI to get good at
1:01:12 something 10 times faster than that 10
1:01:15 000 hours drops to a thousand hours
1:01:17 which is like four months not four years
1:01:22 and then you're like well wait a minute
1:01:26 I could completely reinvent myself three
1:01:29 times a year
1:01:31 with this AI stuff
1:01:34 so I could either get really good at
1:01:36 three things and do all three of them or
1:01:39 I could say [ __ ] it I'm gonna try
1:01:41 neuroscience and then I'm gonna try
1:01:44 um Advanced Bonsai gardening
1:01:47 and then I'm gonna do uh uh 1970s
1:01:52 macrame wall hanging art
1:01:54 like it doesn't [ __ ] matter because
1:01:58 these tools are gonna allow us to
1:02:01 navigate so much more quickly
1:02:04 um so anyway that's that's something
1:02:05 that I've been thinking a lot about it I
1:02:07 talked about it last night and kind of
1:02:09 inspired me to make a video today so I
1:02:11 basically just said what I said in the
1:02:12 video so you don't need to go watch it
1:02:13 but
1:02:15 if all AI were open source how long
1:02:18 until Global super villain Rings form
1:02:20 I'm sure they're forming right now
1:02:21 billionaires
1:02:23 um and AI is open source
1:02:27 metas llama models are open source and
1:02:31 all the weights are out there so are
1:02:33 stable diffusions
1:02:37 and Elon Musk claims his are going to be
1:02:39 too and he just fired up his 10 000 GPU
1:02:42 server Farm
1:02:44 so
1:02:46 um we're gonna find out
1:02:50 I think I think
1:02:52 you know it was it was funny it's like I
1:02:55 don't know it's all such a game but it's
1:02:57 like the stakes are pretty high
1:03:00 you know Elon Musk whenever it was a few
1:03:02 months ago signed on that letter that
1:03:05 anyone doing generative AI should take a
1:03:08 six month pause they shouldn't they
1:03:09 shouldn't there should be no more
1:03:11 advancements in AI for six months
1:03:13 seven days later he ordered 10 000 gpus
1:03:18 so the cynical side of me says oh he
1:03:20 just signed up to get his competitors to
1:03:23 chill out so he could catch up
1:03:26 um
1:03:28 anthropic
1:03:33 the company behind Claude
1:03:36 that thing right there
1:03:39 that that's a bunch of EX open AI people
1:03:42 that
1:03:44 their mission is
1:03:46 alignment like basically making sure
1:03:48 that these tools are aligned with the
1:03:51 goals of humans
1:03:54 um
1:03:55 and Sam Altman talks a lot about it so
1:03:57 those two have different approaches to
1:03:58 it but they're both very keen on it
1:04:00 Google's a little loosey-goosey which is
1:04:03 why Elon Musk says he started open AI
1:04:07 um but now Elon Musk wants to open
1:04:10 source this chat GPT took off and then
1:04:12 Mark Zuckerberg freaked out and said
1:04:14 it's not about the metaverse it's about
1:04:16 AI uh here have one for free
1:04:20 I think that was a little irresponsible
1:04:22 but you know it's not an AGI model it's
1:04:26 not super powerful but it's pretty
1:04:28 [ __ ] powerful and it's open source
1:04:30 and it's out there so so we'll see we'll
1:04:32 see
1:04:34 um
1:04:35 the good news about the is there a super
1:04:38 villain ring forming is that
1:04:41 um
1:04:44 the the white hat
1:04:47 folks also have access to these tools so
1:04:49 the ability to catch those super villain
1:04:52 Rings will be
1:04:53 augmented as well so that that fight
1:04:56 will continue till the end of time I
1:04:59 start as Laura and Glass Menagerie my
1:05:01 freshman year of undergrad that's
1:05:02 awesome that's fantastic
1:05:04 [Laughter]
1:05:10 I know I'll keep I'll keep it between us
1:05:16 99 plus new messages I'm very far behind
1:05:20 if you're new here my name is Kyle
1:05:22 Shannon and this is how I actually talk
1:05:24 not that other way that was just an act
1:05:26 this is how I talk I'm very
1:05:28 sophisticated anyway
1:05:30 one of the things that you might notice
1:05:32 is that I seem to answer questions that
1:05:35 don't seem to be anything anyone's
1:05:36 talking about that's because I'm trying
1:05:39 to answer all the questions and you're
1:05:41 down there and I'm up there
1:05:44 and yes I'm live
1:05:47 I'm sipping water in Denver Colorado and
1:05:50 it's
1:05:52 9 17.
1:05:54 or so
1:05:59 p.m on September 2nd so guys bring in
1:06:02 the fire thank you sir thank you very
1:06:04 much
1:06:05 Silver Fox is here
1:06:11 what was the prompt you told Chachi BT
1:06:14 to Auto bold and bullet instructions
1:06:17 okay so let's go do that one that one's
1:06:20 super cool I like that one a lot it's
1:06:23 actually quite simple
1:06:26 it's just called markdown and so what
1:06:28 you can do is you can just say I'll do a
1:06:31 new chat here
1:06:32 so I'll just say
1:06:34 um
1:06:44 act as a reporter for the New Yorker
1:06:47 and write me an article on the
1:06:50 implications of generative AI on the
1:06:53 workforce
1:06:54 and how
1:06:55 people can reinvent themselves
1:06:59 ten times faster
1:07:01 and that the
1:07:03 Malcolm Gladwell 10 000 hour rule is
1:07:07 Dead period
1:07:10 format the article using markdown to
1:07:13 make it easier to read
1:07:17 and so that format the article using
1:07:19 markdown to make it easier to read it
1:07:22 kind of knows enough to know what that
1:07:24 means and it should give us a title of
1:07:26 the article in subheads when it writes
1:07:28 the article
1:07:29 so there you go there's your title
1:07:31 Jesus GPD 3.5 is so fast it's
1:07:35 disconcerting but but it's like
1:07:37 by your name contributing writer like it
1:07:40 made that italic right
1:07:42 there's your subheads
1:07:45 right
1:07:47 um
1:07:49 please include two bulleted lists and a
1:07:52 table comparing Concepts that you
1:07:55 introduce in the article as well
1:08:00 foreign
1:08:09 there's my table ah
1:08:14 but by the way if you haven't seen
1:08:16 Chachi VT before and you're sitting here
1:08:19 watching this thing write a [ __ ]
1:08:20 article
1:08:22 this fast with a comparison table let's
1:08:25 actually see what's in the table
1:08:26 comparison table traditional learning
1:08:28 versus generative AI powered learning
1:08:30 speed of skill of acquisition
1:08:32 traditional learning gradual requiring
1:08:34 years of practice generative AI rapid
1:08:37 enabling Mastery in months
1:08:39 feedback mechanism delayed feedback from
1:08:42 teachers or mentors instant feedback
1:08:44 through AI systems it's actually quite
1:08:45 good personalization limited
1:08:48 customization of learning materials
1:08:49 tailored content and learning pads
1:08:51 that's actually that one right there is
1:08:53 one of the biggest things for me is
1:08:58 I've been doing this with pie so so two
1:09:01 nights ago my electrical system on my
1:09:04 car went [ __ ] haywire and I'm I'm
1:09:07 calling the roadside assistance people
1:09:09 and I'm and I'm trying to find YouTube
1:09:11 videos and I'm trying to Google [ __ ] and
1:09:13 it's just a [ __ ] nightmare and I
1:09:15 thought you know why don't I try pie and
1:09:19 so pie in in the in the the lives here
1:09:22 we nicknamed pie chatty GPT but you can
1:09:25 basically put this large language model
1:09:28 in in voice mode and just talk to it
1:09:31 and so
1:09:33 it's also connected to the internet
1:09:36 so it can it can look up articles and
1:09:38 things like that so I asked it if it
1:09:40 could help me diagnose the problem
1:09:42 and it actually did a really good job
1:09:45 diagnosing the problem and it and it was
1:09:47 like it was just instant it was it was
1:09:49 like okay let me get some clarity you
1:09:51 know did did this happen when you tried
1:09:53 to start it or did that happen and then
1:09:55 what I told it it said okay you know
1:09:58 here here are two possible things that
1:10:00 happened
1:10:01 um and then it turned out just being a a
1:10:04 shorted battery a dead battery my
1:10:07 battery died
1:10:08 um and so as I was driving back from the
1:10:11 auto parts store with the battery I was
1:10:14 thinking
1:10:15 I kind of remember from some time in my
1:10:18 life that you have to hook up a battery
1:10:21 the battery terminals in a certain order
1:10:23 or bad things can happen and so so I
1:10:27 said to pie you know is there is there
1:10:30 uh
1:10:31 you know what order should you connect
1:10:33 the terminals of a battery for a car and
1:10:37 it said oh I'm very glad you asked that
1:10:39 this is actually really important said
1:10:41 you connect a positive terminal first
1:10:43 because if you don't if you connect the
1:10:45 ground first and then you short out the
1:10:48 the positive you can you know it can it
1:10:50 can blow up your electrical system blow
1:10:52 up your battery or or literally cause an
1:10:54 explosion and I'm like okay good and so
1:10:58 I confirmed that with with the auto
1:11:01 parts store uh but it got it right so I
1:11:04 wanted to make sure it wasn't
1:11:05 hallucinating
1:11:06 um but but yeah this idea that wherever
1:11:09 you are
1:11:11 in your journey of trying to figure
1:11:13 something out
1:11:15 these tools are going to give you
1:11:16 exactly what you need right then the
1:11:18 guidance you need right then and like
1:11:21 just that that think about the amount of
1:11:23 time that you you go back and forth if
1:11:26 you have a mentor or a teacher or you're
1:11:27 taking a class you take the class on a
1:11:30 Tuesday night you get around to your
1:11:32 homework on Saturday night you're
1:11:34 totally [ __ ] confused you reach out
1:11:36 to some friends on Sunday they're not
1:11:37 around Monday comes around you're still
1:11:39 confused you don't get your question
1:11:41 answered until Tuesday so now seven days
1:11:44 has gone by
1:11:45 or you know three from Saturday when you
1:11:48 did the work and now all of a sudden you
1:11:49 can do this but anyway that's markdown
1:11:52 is the name of the
1:11:55 how you get it to do this fancy
1:11:56 formatting but but yeah
1:12:01 the shifting landscape generative Ai and
1:12:03 rap and Rapid self invention let's do
1:12:06 here we'll do something fun
1:12:11 I want you to come up with 20 creative
1:12:14 titles for the article period the idea
1:12:17 that I have is the 10 000 hour rule is
1:12:21 Dead period can you come up with 20 more
1:12:25 that capture the essence of the article
1:12:28 you wrote
1:12:30 are Punchy and memorable and will make
1:12:33 someone want to read the article period
1:12:40 um
1:12:41 here's your titles
1:12:44 uh
1:12:46 let's see beyond 10 000 hours redefining
1:12:49 Mastery in the age of AI breaking
1:12:51 boundaries ai's New Frontier of Rapid
1:12:54 expertise good unshackling potential the
1:12:57 end of the 10 000 hour myth right
1:12:59 because it's not a ten thousand hour
1:13:01 rule it's a myth right and it's it's
1:13:03 even controversial some people are like
1:13:05 well it's way less than that and other
1:13:06 people like well it's way more than that
1:13:08 it's an ongoing exploration of expertise
1:13:10 and no one really Masters anything ever
1:13:14 um
1:13:14 from hours to moments the evolution of
1:13:17 expertise AI Unleashed shattering the
1:13:19 ten thousand hour bet barrier Fast Track
1:13:22 Mastery skill in Silhouettes the demise
1:13:25 of the ten thousand hour standard that's
1:13:27 pretty good
1:13:29 um so it'll do that so so now so here's
1:13:31 so here's a fun thing we can do watch
1:13:33 this so now I'm going to say
1:13:42 I want you to analyze the titles
1:13:48 for their strength
1:13:51 and give them a score from zero to one
1:13:54 hundred
1:13:55 tell me why you gave them that score
1:13:58 and put them in a table
1:14:00 sorted from strongest to weakest
1:14:12 oh good lord is this not blowing
1:14:14 anyone's mind I mean every time it does
1:14:16 this it just [ __ ] kills me so let's
1:14:18 go look at what it did
1:14:21 um
1:14:23 AI Unleashed shattering the 10 000 hour
1:14:26 bet barrier which is really good to give
1:14:27 it a score of 95 strengths and
1:14:29 explanation bold intriguing and directly
1:14:32 addresses the article's theme 93 chasing
1:14:36 skills ai's disruption of the 10 000
1:14:38 hour tradition pretty good 92 ai's
1:14:41 Mastery makeover that's really good
1:14:43 leaving the 10 000 hour rule behind yep
1:14:46 that's good so you know it's it's very
1:14:49 good at doing this kind of thing
1:14:51 um you know brainstorming refining I I I
1:14:54 mean you saw what I just did I mean I
1:14:57 basically gave this no context and no
1:14:59 input and it it wrote something you know
1:15:02 probably in the neighborhood
1:15:04 so imagine you know spending an hour or
1:15:07 two where you actually have you give it
1:15:09 a lot of thoughts and you've put thought
1:15:11 into this and you you start out and have
1:15:14 it give you an outline first and then
1:15:16 you have it flesh out some points of the
1:15:18 outline and if you don't like the way it
1:15:19 does it maybe you give it an example of
1:15:21 your writing and you say okay I want you
1:15:23 to follow my writing and then I want you
1:15:25 to
1:15:26 um okay you did it you did it okay but
1:15:29 you didn't use any metaphors and so this
1:15:31 time I want you to do a password you do
1:15:32 metaphors and this time I want you to do
1:15:34 a pass
1:15:35 where you actually give me quotes from
1:15:39 um you know you know Malcolm Gladwell or
1:15:41 from you know you know uh you know
1:15:43 business compatriots like Seth Godin and
1:15:46 things that you know that that that may
1:15:49 be counter arguments to to all of these
1:15:51 theories whatever
1:15:53 um
1:15:54 the two hours that you would have spent
1:15:56 writing the article you could now have
1:15:59 you know the same length of article but
1:16:02 but you know fleshed out in such a
1:16:04 dramatically more comprehensive way just
1:16:08 because the speed with which you can
1:16:10 process ideas
1:16:17 [Music]
1:16:19 Parappa
1:16:21 hope that helps Silver Fox that was like
1:16:23 17 answers to your one question
1:16:26 that was a good mansplaining that was a
1:16:29 good mansplaining moment
1:16:31 share bear thank you for the gummies I'm
1:16:34 in Denver we we Define gummies a little
1:16:37 differently here if you know what I mean
1:16:44 thank you Winston I I don't know why
1:16:47 George W bush just overtook me but it
1:16:49 happens it's gonna happen it's gonna
1:16:51 happen you know what I mean
1:16:53 iced coffee yeah it's iced coffee oh no
1:16:56 dang it it just jumped me to the bottom
1:16:58 again
1:17:00 oh Tick Tock ah what time are we done
1:17:03 with your Tick Tock what are we gonna do
1:17:07 you duh yeah da
1:17:09 you does have the fan with the ship
1:17:12 again what are we gonna do with your
1:17:14 tick tock
1:17:16 get thee to bed get they to bed I'm
1:17:18 disgusted with you
1:17:22 all right
1:17:24 what'd you do Saturday night well I was
1:17:27 uh I don't I don't really want to talk
1:17:29 about it now what did you do what did
1:17:32 you do I was home watching old reruns of
1:17:34 Matlock how about you what were you
1:17:36 doing I was on that
1:17:38 what did you say
1:17:40 I was on
1:17:43 did you say tick tock yeah I was on tape
1:17:47 what were you doing on tick tock
1:17:50 I don't want to talk about it it's okay
1:17:53 you can trust me
1:17:56 I was learning about AI
1:17:58 AI
1:18:00 you mean artificial you mean the robots
1:18:04 the robots are gonna kill us you know
1:18:05 that right
1:18:07 anyway this is what we do Saturday night
1:18:14 you could be doing something else you've
1:18:16 chosen to be here I appreciate it listen
1:18:18 this is fun we're having a good time
1:18:20 we're learning something
1:18:23 I hope you are a tick tock wizard I am a
1:18:26 wizard
1:18:27 I am the wizard
1:18:31 welcome to my lab this is the AI
1:18:33 learning layer
1:18:35 I am the wizard
1:18:38 Tick Tock is my [ __ ]
1:18:42 all right let's see let's see what we're
1:18:45 gonna do here we are gonna go find a
1:18:48 question
1:18:53 um
1:18:53 let's see all right and we spend all day
1:18:57 having meetings that should have been
1:18:58 emails yes exactly Vicky
1:19:01 I think that's when I was talking about
1:19:03 simplifying the complicated
1:19:06 money with Chachi BT guy reminds me of
1:19:08 the 90s information infomercial Don
1:19:11 lapray yeah exactly placing tiny ads oh
1:19:15 man was he annoying exactly
1:19:19 you can make money with Direct Mail
1:19:22 you can make money with Photoshop you
1:19:24 can make money with Microsoft Word
1:19:26 you can make money with social media you
1:19:28 can make money with Tick Tock you can
1:19:29 make money with chachibite
1:19:32 it's all the same course people it's all
1:19:36 the same course
1:19:42 it makes me heard inside
1:19:49 it's a whole Community my buddy Steve is
1:19:51 in that Community he knows that
1:19:52 Community really well
1:19:55 they all just trade the same form
1:19:57 letters and just swap out words
1:20:04 God I love zapier Yeah zapier is pretty
1:20:07 [ __ ] amazing and their new bubbles
1:20:09 the new um
1:20:10 the new interfaces where you can make a
1:20:12 chat bot and you can make websites it's
1:20:14 pretty slick that's tough it's tougher
1:20:17 if you're stupid
1:20:21 thank you for the Hat I like the Hat I
1:20:23 like the Hat I get sometimes I'll get
1:20:25 dangly lights sometimes I get a little I
1:20:28 don't know what I get the other night I
1:20:29 don't know some sort of oh I get the uh
1:20:31 the
1:20:33 the wizard hands
1:20:36 thank you bubs
1:20:37 yeah zapier is really good
1:20:39 GRC yes simple I see the dangly lights I
1:20:43 love the dangly likes thanks share bear
1:20:44 and pizza
1:20:46 that loves me some pizza which is
1:20:48 evident by my face
1:20:51 ah
1:20:55 simple and it cuts through all the
1:20:56 layers of BS and redundancies hi
1:20:58 everyone welcome jrc good to see you
1:21:01 all right bubbles
1:21:03 um I just started I'm loving zapier
1:21:05 someone's smarter than me should help us
1:21:08 no wrong
1:21:11 you got it figure this is here's the
1:21:15 deal no one is okay
1:21:18 we are at a stage right now so so a
1:21:21 little bit of background here my
1:21:24 my background
1:21:26 I've been an entrepreneur for I don't
1:21:28 know [ __ ] decades at this point I
1:21:29 started 12 or 15 companies
1:21:31 I started at one of the first digital
1:21:33 agencies in the mid 90s when the web
1:21:35 first came out
1:21:37 and
1:21:38 when I started that agency I started an
1:21:41 online magazine I started a group called
1:21:43 The World Wide Web artist Consortium
1:21:45 and the reason I started that group is I
1:21:48 thought there's got to be someone who
1:21:49 knows more about this than I do
1:21:52 it was so early that nobody knew
1:21:55 anything everyone was trying to figure
1:21:56 it out
1:21:58 and so even though
1:22:00 the internet had been around for decades
1:22:05 the World Wide Web was different
1:22:07 the internet up to that point was
1:22:09 research scientists and Librarians and
1:22:12 PhD students and and Uber Ultra Geeks
1:22:17 and it was all command line and you had
1:22:18 to log in and command line and do this
1:22:21 and that and as remarkable as it was you
1:22:23 can kind of see what it was it was a lot
1:22:27 and then Tim berners-lee says what if
1:22:31 this was the interface
1:22:34 click
1:22:35 click
1:22:36 click
1:22:38 I know
1:22:40 I'll make it so you can underline a word
1:22:42 and when someone clicks on it it takes
1:22:45 you to another document that also has
1:22:47 links in it that you can click on
1:22:50 and it was that simple thing that
1:22:54 shifted everything
1:22:55 and that's where we are with this [ __ ]
1:22:57 right now
1:22:59 machine learning and ai's been around
1:23:01 for decades like like it's it's in all
1:23:04 of the software that we already use
1:23:06 right it's in our cars it's in you know
1:23:09 Siri and Alexa and it's when you type in
1:23:12 Google and it finishes your sentence it
1:23:13 for you this is all AI it's all been
1:23:15 there
1:23:17 what's different is
1:23:20 November 30th 2022 last November chat
1:23:24 gbt launches
1:23:26 which is the equivalent of the world
1:23:30 wide web for the internet back then
1:23:32 where we just have a simple Google
1:23:35 search box and you ask it [ __ ]
1:23:37 anything and these things know
1:23:40 everything but
1:23:42 there are no experts right now nobody
1:23:46 knows what to do so if you're if you're
1:23:48 figuring out zapier and you're starting
1:23:50 to figure out oh [ __ ] I can send I can
1:23:53 take input from a Google form or input
1:23:55 from a slack Channel and I can
1:23:57 incorporate that into a prompt and I can
1:24:00 send that to chat GPT and I can generate
1:24:02 an image or I can generate a piece of
1:24:04 social media content or I can I do
1:24:06 whatever and then I can post that back
1:24:08 to a slack Channel holy [ __ ]
1:24:11 yes there are people that probably know
1:24:13 zappy are better than you might know
1:24:15 prompting better than you but you
1:24:18 learning
1:24:19 how to do that and how to start thinking
1:24:21 critically about the kinds of problems
1:24:24 you're trying to solve
1:24:26 [ __ ] invaluable
1:24:28 don't have someone hand it to you don't
1:24:30 have someone hand it to you learn yeah
1:24:32 exactly thank you for the Roses but yes
1:24:35 exactly like you're you're on the right
1:24:37 path and and the only thing I would
1:24:39 augment what you're doing with if you're
1:24:41 not doing it already is talk about this
1:24:43 [ __ ]
1:24:44 talk about it on LinkedIn talk about it
1:24:47 within your company talk about it to
1:24:49 your friends and family even if they
1:24:51 think you're a [ __ ] weirdo I heard
1:24:53 ai's gonna kill us well I don't care
1:24:55 Sally
1:24:56 zapier is really cool let me let me tell
1:24:59 you more about it
1:25:04 because everyone's trying to figure it
1:25:06 out right now and it can only serve you
1:25:08 it can only serve you so
1:25:11 whatever you're doing [ __ ] thumbs up
1:25:15 talk about it if you're not
1:25:18 um
1:25:18 like the opportunities in the next 18
1:25:21 months thank you for the lamo instant
1:25:24 um the opportunities in the next 18
1:25:26 months for anyone
1:25:29 that says you know what [ __ ] it I'm
1:25:31 gonna figure this [ __ ] out because it's
1:25:33 not you can figure it out
1:25:35 this isn't machine learning five years
1:25:38 ago
1:25:40 you know I've said this on the live a
1:25:42 lot this week because it's just it's
1:25:43 really come home for me because I think
1:25:45 as Google just had one of their big
1:25:46 their their tech meeting and and Pate
1:25:49 was in here that works at Google he
1:25:51 works making these models stronger
1:25:54 um
1:25:55 for like the last five years probably
1:25:57 closer to ten the last 10 years I've
1:25:59 been really aware of artificial
1:26:02 intelligence and machine learning and
1:26:04 and every time Google would put out a
1:26:07 press release about tensorflows now
1:26:09 really easy to use I'd be like great I'm
1:26:12 gonna go play with machine learning and
1:26:13 I go in and then I'd realize no no
1:26:17 tensorflow is easy to use if you've got
1:26:19 a PhD from Stamford
1:26:22 you're a master technologist and and you
1:26:26 know you've done data analysis for a
1:26:29 decade or two then it's then it's
1:26:31 [ __ ] simple
1:26:33 but if you have a degree in acting and
1:26:35 you're an entrepreneur and you just want
1:26:36 to figure out how to [ __ ] to use AI in
1:26:38 your business it was too early that
1:26:41 stuff's that stuff is a level two under
1:26:43 the hood
1:26:46 chatgpt is the first thing that says hey
1:26:48 the rest of you
1:26:51 you want to see something cool
1:26:53 and that's why there's no experts right
1:26:55 now because the rest of us we have no
1:27:01 clue whatsoever
1:27:03 what to do how to do how it's going to
1:27:05 impact us how it's going to impact jobs
1:27:08 but for those of us that are curious and
1:27:11 exploring and active
1:27:13 will start to figure it out
1:27:15 and we will actually start to bend it
1:27:19 in a positive direction
1:27:21 so so I think that's why I think it's
1:27:24 important to talk about it this [ __ ]
1:27:26 while it's got the potential to do bad
1:27:29 [ __ ] the more you use it the more you
1:27:32 realize this isn't some scary fictitious
1:27:34 [ __ ] robot that's just out there
1:27:36 aimlessly you know killing humans this
1:27:40 is a tool that we as humans get to
1:27:43 direct
1:27:44 to amplify what we want to do
1:27:48 these are Humanity amplifiers and
1:27:51 Humanity mirrors
1:27:54 they know all of our words because we
1:27:57 put all of our words in them
1:28:00 and so when you ask it a question it's
1:28:02 not the robot telling you [ __ ] it's us
1:28:07 having educated the robot on all of what
1:28:11 we've talked about for four decades just
1:28:13 sitting in there waiting to be tapped
1:28:17 so anyway [ __ ] rock on bubbles and by
1:28:21 the way 9.5 out of 10 on the name it's
1:28:23 hilarious I love it
1:28:25 all right pet cemetery was creepy and
1:28:28 the new version is surprising sorry for
1:28:30 my French brain lagging
1:28:32 that's you don't need to apologize for
1:28:35 for uh your uh English togas we love you
1:28:41 Just As You Are
1:28:43 chatty ADHD exactly uh
1:28:50 13 oh wait 13 000 oh we're up to thirty
1:28:53 thousand okay so Dr J so by the way
1:28:56 um for those of you that are new here oh
1:28:57 up to 95 there's a lot of new people
1:28:59 here welcome everybody
1:29:00 um you see all these Hearts flying up
1:29:03 the side here
1:29:05 this is not because people are loving
1:29:07 this no no
1:29:08 the Irregulars so the people that show
1:29:10 up here regularly who have named
1:29:12 themselves the Irregulars they have ADD
1:29:14 too A lot of them most of them they're
1:29:17 they're all special they're all neuro
1:29:19 spicy as we say
1:29:21 they get bored well
1:29:23 as you would right if you come here a
1:29:26 lot
1:29:27 and so they've come up with a game and
1:29:29 their game is to beat the number of
1:29:31 likes from from the previous sessions
1:29:35 um they still haven't cracked there that
1:29:36 one night they when they first
1:29:38 discovered this game they went just ape
1:29:41 [ __ ] and they broke a hundred thousand
1:29:42 so then they got carpal tunnel and it
1:29:45 hasn't quite been the same sense but
1:29:46 they've been doing some pretty
1:29:47 impressive work so we're like an hour in
1:29:50 and we're at thirty thousand so that's
1:29:51 pretty solid I like it
1:29:55 um cabruno also valuable pain PT is
1:29:58 helping company a understand what
1:30:00 company AI already knows what's pain PT
1:30:04 data analysis
1:30:06 that must have been some sort of uh
1:30:08 spell check weirdness
1:30:10 also valuable pain PT
1:30:17 chat TBT
1:30:19 Maybe
1:30:21 um yeah so so but what you're saying
1:30:23 there is whatever pain PT PT stands for
1:30:26 I assume that I assume that means chat
1:30:28 gbt
1:30:29 um that it's helping companies
1:30:31 understand what the companies have that
1:30:34 like I think that's going to be the
1:30:36 biggest shift for companies
1:30:39 um
1:30:39 so so open AI just announced
1:30:43 um chat GPT Enterprise so it's now
1:30:47 secure and private and they're not
1:30:48 training on your data all that sort of
1:30:50 [ __ ]
1:30:51 um
1:30:53 you have to be spending ten thousand
1:30:54 dollars a month to get the real GPT
1:30:57 Enterprise They are going to be
1:30:59 announcing a self-service version for
1:31:00 the rest of us us plebes
1:31:03 um
1:31:04 but what they also announced is is
1:31:06 coming shortly on the heels of GPT
1:31:08 Enterprise is going to be the ability
1:31:10 they're going to make it very easy for
1:31:12 companies to train on their own data
1:31:14 the minute you can take all of your
1:31:16 company data and basically gptize it
1:31:20 you're going to be able to free
1:31:22 data that's been locked in digital
1:31:25 lockers unseen for decades
1:31:29 and I think just some remarkable
1:31:31 Innovation is going to come when people
1:31:32 are able to sort of pull out the
1:31:35 collective knowledge that's been dumped
1:31:37 into these digital warehouses for for
1:31:40 years and years and years
1:31:43 and say didn't we once do a project
1:31:46 where we tried to combine this approach
1:31:48 with that approach
1:31:51 yep you did and here were here's the
1:31:53 PowerPoint documents that were created
1:31:55 and here's the word documents and here's
1:31:57 a summarization of it
1:31:59 well how would I
1:32:02 you know take that idea and incorporate
1:32:04 it with this new strategy that we've got
1:32:06 oh you might do it like this and by the
1:32:08 way here's pieces of that and right it's
1:32:11 gonna be remarkable now
1:32:14 it's gonna take a while for for uh
1:32:16 companies to get their [ __ ] together and
1:32:19 and you know figure out which data they
1:32:21 want to put in there and what are the
1:32:23 permissions and all that sort of happy
1:32:24 horseship but yeah I actually need to
1:32:26 figure this out I left my job of 20
1:32:28 years
1:32:29 so Palm dog
1:32:32 um
1:32:33 I assume congratulations if you left it
1:32:36 but if you were left I I apologize I
1:32:40 feel for you but if you left it cool
1:32:43 congratulations
1:32:44 um what I would say
1:32:47 is
1:32:49 oh thank you Winston I appreciate the
1:32:52 heart
1:32:53 um the the simplest place to start is
1:32:55 just chat GPT right here
1:32:58 depending on what you did
1:33:00 um
1:33:01 you can there's all sorts there's all
1:33:04 sorts of rabbit holes you can run down
1:33:05 but really just start just really get
1:33:07 your head around how Chachi PT is
1:33:09 different than a Google Search and and
1:33:11 what it makes possible take what you've
1:33:14 done for 20 years the knowledge that you
1:33:16 have and really try to push the
1:33:18 boundaries of this see what it can do
1:33:20 see what it can do for you can it if
1:33:22 you're a project manager does it know
1:33:25 project management it does
1:33:27 if you're a salesperson does it know all
1:33:29 different sales strategies yes why
1:33:32 because the internet was full of sales
1:33:33 strategies from the past hundred years
1:33:36 200 years probably 500 years
1:33:38 so it knows all that [ __ ] right so it's
1:33:40 all in there
1:33:42 so push the boundaries of it see where
1:33:44 the boundaries are and then you can
1:33:46 start to explore okay well here's what I
1:33:48 want to do with my here's what I was
1:33:50 doing with my life here's what I
1:33:51 understand here's what I want to do with
1:33:53 my life have this thing help you bridge
1:33:55 the gap
1:33:56 so I've got these skills over here I
1:33:59 want to have these skills over there huh
1:34:00 how can I get there quickly well you can
1:34:02 get there quickly with this
1:34:05 then you can have it do [ __ ] like if you
1:34:06 want another job write me
1:34:08 a resume rewrite my res you know take my
1:34:11 LinkedIn profile rewrite my resume to
1:34:14 focus on that new thing I want to do
1:34:17 and by the way write me a cover letter
1:34:19 for these six companies that's
1:34:21 customized to them talking about how
1:34:24 awesome awesome I am at the new thing I
1:34:26 want to do
1:34:27 so yeah that could be really cool but
1:34:29 congratulations if Sam corn never shows
1:34:31 up here again it's my fault uh-oh what
1:34:33 happened Winston what'd you do what did
1:34:36 you do to Sam
1:34:37 you like Sam
1:34:39 did you did you did you name out Sam oh
1:34:42 no
1:34:46 I think Sam's a tough cookie
1:34:50 thank you
1:34:52 um
1:34:53 Winston Hardy what do you mean to Sam
1:34:55 and scene
1:34:57 um yes here at 9pm in the mile high city
1:34:59 so that was 45 minutes ago I'm 45
1:35:01 minutes behind okay wait is it weird my
1:35:04 kids want to track me on Life360 what's
1:35:07 Life360
1:35:15 [Music]
1:35:19 Family Locator app and safety membership
1:35:26 hey you know
1:35:31 yeah and it depends why they want to why
1:35:35 they want to track you you can tell them
1:35:36 to [ __ ] off
1:35:39 mama's gonna do what mama's gonna do
1:35:42 I'll track you kids let's let's get one
1:35:45 thing straight here one of us is the
1:35:48 parent and one of us is not
1:35:50 let's see let's do the math let's just
1:35:53 do this with math can you get out your
1:35:54 calculators okay I was born in this year
1:35:58 you were born in that year what's the
1:36:00 what's the year now 2023 okay could you
1:36:02 do the math
1:36:04 which one's bigger me I'm the parent
1:36:07 I will track you you will not track me
1:36:10 family 360 my ass this is family 180.
1:36:15 that's the Gen X approach well that's
1:36:17 actually the Boomer approach
1:36:19 he'll do what we [ __ ] tell you get
1:36:22 outside don't come back till dark
1:36:29 uh your Tick Tock rant had levels of
1:36:32 Lewis Black energy one of my favorites
1:36:34 Lewis black man
1:36:36 I saw him I saw him before he popped off
1:36:40 um when I was doing theater in New York
1:36:42 City
1:36:44 um we were we would do
1:36:46 stuff at the theaters on 42nd Street
1:36:49 and then when our show was done we'd go
1:36:52 over to the West Bath theater and Rusty
1:36:54 McGee would play the piano and Lewis
1:36:56 Black would do his [ __ ] rants it was
1:36:58 like a small little room
1:37:00 Rusty McGee would Place her to this
1:37:02 background piano and Lewis would just
1:37:03 [ __ ] go off on stuff it was awesome
1:37:07 So yeah thank you
1:37:10 that is a compliment
1:37:12 I have a burner phone you can't outsmart
1:37:14 me that's good Kyle's new handle the
1:37:18 wizard I am the wizard is there a way
1:37:21 for Chachi BT to analyze data from
1:37:23 charts and provide a summary for reports
1:37:25 oh Don Marco you should have been here
1:37:26 last night
1:37:28 I babbled at Sam too much in a message
1:37:30 regarding my expertise and ended all
1:37:33 with a question about AI
1:37:36 uh I don't think you chased Sam off
1:37:38 Winston Sims I think Sam's a tough
1:37:40 cookie I get the sense that Sam seems
1:37:43 very very smart and very uh
1:37:47 centered in his confidence
1:37:50 is the way to put it
1:37:54 I like those kind of people that's why I
1:37:56 miss New York New York's full of people
1:37:58 that are just like boom I'm [ __ ] here
1:38:01 you're [ __ ] there are we good
1:38:05 like Lewis Black
1:38:07 [Laughter]
1:38:08 um
1:38:11 all right Don Marco I'm gonna show you
1:38:13 some so here's what I'm gonna do since
1:38:16 I'm 99 Plus messages behind I'm going to
1:38:19 show Don Mark I don't even know if Don
1:38:21 Marco's still here anymore this is 45
1:38:22 minutes ago
1:38:23 I'm going to show you code interpreter
1:38:27 because we did we did it last night I
1:38:29 know what the data set is so I can just
1:38:30 show you really quick we did last night
1:38:33 um
1:38:34 and then um
1:38:36 and then I'm just going to jump to the
1:38:38 bottom and and do do things answer some
1:38:41 questions at the bottom so if you have
1:38:42 if you have questions that I haven't
1:38:43 answered just re-ask them now because
1:38:45 I'm going to skip a bunch of questions
1:38:46 okay all right all right people let's
1:38:50 let's get going let's get going
1:38:53 Microsoft checking for updates we are
1:38:55 not interested
1:38:57 we're gonna go to chat GPT yeah there we
1:39:00 are at chatgpt
1:39:03 and we're gonna go to the code
1:39:05 interpreter which they have renamed
1:39:07 remember if you've been here for a while
1:39:10 when when code interpreter came out what
1:39:13 did I say
1:39:14 [ __ ] fantastic technology horrible
1:39:17 name and then some geek in here went
1:39:20 well it's technically accurate because
1:39:21 blah blah blah blah blah blah blah blah
1:39:24 I'm like okay but it's still a terrible
1:39:26 name so now they've changed the name so
1:39:28 now it's called Advanced Data analysis
1:39:30 which is what it does
1:39:32 that's good good name change so now that
1:39:36 we've picked Advanced Data analysis it
1:39:37 doesn't quite roll off the tongue as
1:39:39 much as code interpreter did but it's
1:39:42 better
1:39:44 so you can upload a file you can
1:39:46 actually upload multiple files
1:39:48 so we're going to upload a file
1:39:50 and I'm going to grab archive.zip I
1:39:53 happen to know right now that this is
1:39:56 um
1:39:57 Chipotle locations
1:40:00 but I'm gonna do what I did yesterday
1:40:05 please open this archive and tell me
1:40:08 what data is in here and what might be
1:40:11 interesting about it
1:40:17 so you see it sort of does a thing it
1:40:21 like pops open if you've ever used
1:40:23 plug-ins it looks like kind of like a
1:40:24 plug-on plug plugin and then it says
1:40:27 this archive contains two files Chipotle
1:40:30 stores CSV and United US states dot Json
1:40:33 and then
1:40:35 if I go open this little box what you
1:40:38 actually notice is that this is python
1:40:42 and you're like well where'd that python
1:40:44 come from it wrote it so what Advanced
1:40:48 Data analysis does when you launch the
1:40:51 chat it spins up a virtual machine
1:40:54 and it's got a pipe it's got a series of
1:40:56 python libraries that it's got access to
1:41:00 and when you upload your data to it it's
1:41:02 basically unpacking all this data in
1:41:04 this temporary virtual machine up in the
1:41:06 sky
1:41:07 and it's
1:41:09 it's prompting itself
1:41:13 to write python programs and execute
1:41:15 them based on the data so that's what's
1:41:18 actually going on here it's it's [ __ ]
1:41:20 mind-blowing so so it says here's what's
1:41:22 in there would you like me to delve into
1:41:24 these files further and explore their
1:41:26 content so just say yes
1:41:38 all right so the first thing it did
1:41:40 let's pop that open
1:41:44 it looked at the CSV file
1:41:52 so the data and the second thing it did
1:41:54 I assume it looked at the Json file yeah
1:41:56 okay so so it looked at so it wrote code
1:41:59 to to unpack what was in the CSV file
1:42:02 grab grab the first few rows so could
1:42:04 understand it
1:42:06 and then it figured out what was in the
1:42:08 Json file and then it says data summary
1:42:11 the CSV the data has columns for State
1:42:15 location address latitude and longitude
1:42:17 of the Chipotle stores
1:42:19 example data for instance there's a
1:42:22 Chipotle store in Auburn Alabama with
1:42:24 specific Latin longitudinal coordinates
1:42:26 the United States dot Json
1:42:30 it contains 50 features likely
1:42:33 corresponding to the 50 states
1:42:38 possible interest
1:42:40 you could use Chipotle data to analyze
1:42:42 the distribution of Chipotle stores
1:42:44 across different states or regions the
1:42:47 Geo Json file could be used to create
1:42:50 geographical visualizations
1:42:53 perhaps even in conjunction with the US
1:42:55 Chipotle stores to show the locations on
1:42:58 the US map
1:43:01 um
1:43:02 um yes
1:43:04 please
1:43:07 visualize the
1:43:10 data
1:43:12 you find most compelling
1:43:18 Force
1:43:20 let's see
1:43:21 four
1:43:24 someone studying the history of the
1:43:26 restaurant chain
1:43:30 we'll see what it does here
1:43:39 so Advanced Data analysis this is full
1:43:41 on like basically you have a full-on
1:43:45 data analyst working for you that you
1:43:48 basically just ask it [ __ ] and it goes
1:43:49 off and does it for you what's
1:43:50 remarkable if you use this for a bit is
1:43:52 it'll try to do something and if it's
1:43:54 got like if it used the wrong ISO
1:43:58 uh
1:43:59 thing oh look what it did here
1:44:01 what did it do
1:44:03 it made a map there's a little teeny
1:44:06 tiny map of the United States Chipotle
1:44:08 stores across the US you got Alaska up
1:44:10 there big giant Alaska with a few
1:44:15 and then you got a little teeny tiny map
1:44:17 here can we make it bigger
1:44:21 that's really funny it made it so tiny
1:44:29 so it drew the state boundaries because
1:44:31 it had the Json data of the state shapes
1:44:34 observation there's a high concentration
1:44:36 of Chipotle stores in the eastern and
1:44:38 western parts of the United States the
1:44:40 sparse areas are the central U.S the
1:44:42 coastal preference many stores seem to
1:44:44 be located near the coast historical
1:44:46 context for someone studying the history
1:44:49 of the restaurant chain this map offers
1:44:51 insights and how Chipotle has expanded
1:44:53 so let me ask
1:44:54 um
1:44:56 does the data contain any date
1:44:59 information
1:45:01 so that we can know
1:45:03 how they progressed from when they
1:45:06 started until now
1:45:14 doesn't contain any data information so
1:45:16 there's no ways to do no ways to do it
1:45:19 so anyway so that's that's code
1:45:21 interpreter or Advanced Data analysis in
1:45:23 chat gbt you have to
1:45:26 um
1:45:30 hang on I'm just waiting for the glitch
1:45:33 to well every time I switch cameras
1:45:34 there's a glitch
1:45:36 um you have to be paying for chat GPT
1:45:38 plus so it's 20 bucks a month and then
1:45:40 you have to enable Advanced Data
1:45:42 analysis and beta features in your
1:45:44 settings you have to turn it on
1:45:45 so you if you pay your 20 bucks and you
1:45:48 don't see that in your list that's
1:45:49 because you haven't enabled it so go
1:45:51 into your settings go to Beta
1:45:53 turn those things on thank you Robert
1:45:54 Rossi I like donuts donuts is tasty
1:46:01 no
1:46:03 I like taters all right
1:46:06 I'm going to scroll to the Bottom now
1:46:08 because I'm way behind there's 93 people
1:46:11 in here and I just missed everything so
1:46:12 I just scrolled all the way to the
1:46:13 bottom and now I'm scrolling up a little
1:46:15 bit to get to some some questions and
1:46:17 then we're going to start off we're just
1:46:18 going to keep it going we're going to
1:46:20 keep it rolling people we're gonna keep
1:46:21 it rolling there's AI learning lab let's
1:46:23 get another let's get another image
1:46:25 that's a pretty one I like that one
1:46:26 pretty good I like taters
1:46:30 yeah yeah time for Wapner yeah I'm an
1:46:33 excellent driver yeah
1:46:35 yeah Charlie battle Charlie Babbitt yeah
1:46:38 all right AR learning lab yeah
1:46:43 cover letter also exactly did I wait am
1:46:46 I alone
1:46:48 experimenting
1:46:50 wait did I
1:46:56 did I am alone experiencing a Time
1:47:00 prompt line wrapping experience oh this
1:47:03 took us okay that's that's why I can't
1:47:06 understand it
1:47:07 um
1:47:09 are you alone in experiments with prompt
1:47:11 the line wrapping
1:47:12 I don't know if if you've figured out
1:47:14 something cool to do with line wrapping
1:47:16 took us let us know I would imagine that
1:47:19 you could tell Chachi BT to do line
1:47:22 wraps at a certain width and it'll do it
1:47:24 you can certainly tell it to give you a
1:47:25 table I've even told it what width to
1:47:28 make the columns relative to one another
1:47:30 and it can do that
1:47:32 1964 here technically a boomer on the
1:47:35 cusp I'm 65 and apparently 65 is the
1:47:38 beginning of Gen X so I am not a boomer
1:47:41 I am I am post Boomer by a year Winston
1:47:44 or Ted
1:47:46 I thought you were a Winston Ted Ted you
1:47:49 Boomer
1:47:50 I'm a gen xer you Boomers screwed up our
1:47:53 lives it's all your fault Ted
1:47:58 we should be having cheeseburger and
1:47:59 margaritas tonight we should
1:48:02 maybe you know you know what we should
1:48:04 do so so if you're in the uh if you're
1:48:07 in the AI salon and you're in the
1:48:09 Irregulars Channel we we should come up
1:48:11 with
1:48:12 um themes for the night for for like
1:48:14 food and drink so that's just geeky but
1:48:18 all right whatever 1962 I win
1:48:23 uh
1:48:24 uh loving tonight thank you bubbles
1:48:26 welcome good to have you here and major
1:48:29 gift fundraising what type of job are
1:48:31 you applying for talking to one another
1:48:33 I love that I love that you're
1:48:35 supporting one another in here
1:48:39 you are RT
1:48:42 I'm real time
1:48:44 Lonnie trying to get into Tech I'm a
1:48:47 substitute teacher now
1:48:49 so here's the deal Lonnie
1:48:51 um it depends what you want to do if you
1:48:53 want to get into Tech if you want to be
1:48:54 an engineer
1:48:55 what I would say is learn to be an
1:48:57 engineer an AI literate engineer so
1:49:01 start learning things like
1:49:03 GitHub co-pilot and you know pair
1:49:06 programming with with AI
1:49:09 um if you just want to get into Tech
1:49:10 like you want to get into more technical
1:49:12 stuff you want to learn this generative
1:49:14 AI stuff you don't necessarily need to
1:49:16 get into Tech right the the large
1:49:18 language models
1:49:20 dang it
1:49:22 the large language models like chat GPT
1:49:24 and Claude and Bard and and the image
1:49:27 generation tools like mid-journey and
1:49:29 Adobe Firefly and things like that
1:49:32 there you're doing incredibly
1:49:34 sophisticated technical things using the
1:49:37 English language or in togasa's case
1:49:39 French can be any language it doesn't
1:49:41 matter you're using your native language
1:49:45 to program
1:49:48 so these things are quite remarkable so
1:49:50 you can you can do incredibly technical
1:49:52 things without getting into Tech right
1:49:54 getting into Tech kind of implies you're
1:49:56 going to learn programming and you're
1:49:58 going to learn I.T and you're going to
1:49:59 learn the technical components it
1:50:01 doesn't hurt to know those
1:50:03 but understand that we're moving to a
1:50:05 future very very quickly where tactical
1:50:09 coding
1:50:10 isn't going to be as required as it is
1:50:14 right now
1:50:15 now understanding how to code and to
1:50:18 think logically and understanding
1:50:21 how that [ __ ] works
1:50:24 that's going to be a requirement for you
1:50:27 know decades to come
1:50:28 but the actual tactical execution of
1:50:31 coding is likely going to be done by by
1:50:34 these tools
1:50:36 and then so what does that mean what's
1:50:38 what's what's most important in in a
1:50:41 world where the repetitive low-level
1:50:44 tactical execution [ __ ] is handled
1:50:48 the most important skill is critical
1:50:50 thinking right so if you're a substitute
1:50:52 teacher and you're already you know if
1:50:54 you can handle [ __ ] dealing with
1:50:56 students you know trying to game you you
1:50:59 know as you come into the system every
1:51:01 time depending you know what grade you
1:51:03 substitute to that we were brutal when I
1:51:06 was in high school with our substitutes
1:51:08 oh man it was it was rough so if you
1:51:11 could handle that you know you're
1:51:13 already good at sort of processing
1:51:15 multiple inputs that's going to be the
1:51:17 critical skill so I would say get AI
1:51:19 literate start to learn what are these
1:51:21 good at what are they not good at learn
1:51:23 this prompts.chat thing if you haven't
1:51:25 really dug into how you interact and
1:51:28 talk with these things that's a good
1:51:29 starting point and then just start
1:51:31 geeking out just start geeking out take
1:51:34 what you know what you're passionate
1:51:35 about I my My Philosophy on this stuff
1:51:39 right now is there's literally nothing
1:51:41 you can do wrong
1:51:44 as long as you're kind of learning how
1:51:46 to do how to do that thing through these
1:51:49 you know with the assistance of these
1:51:52 because even if you're doing something
1:51:54 like I don't know you're super into
1:51:58 um
1:51:58 I don't know tractor maintenance
1:52:02 and and you start using these things to
1:52:06 reconstruct rewrite tractor maintenance
1:52:09 instructions in some way I don't
1:52:11 whatever it doesn't matter
1:52:13 you're gonna then go do some other
1:52:15 project and someone's going to present
1:52:17 you with a problem and you're going to
1:52:19 go oh that's just like the tractor
1:52:21 manual thing I did wait hang on I know
1:52:23 how to do this
1:52:24 and you hop in and you pop open one of
1:52:26 these tools and you go
1:52:30 and then you hand someone a solution
1:52:32 that they thought was going to take a
1:52:34 week and you hand it to them in 20
1:52:36 minutes
1:52:38 and their minds will be [ __ ] blown
1:52:41 and you'll get a good endorphin boost
1:52:45 and they'll be like how did you do that
1:52:47 well let me tell you Mike Rowe reached
1:52:50 level 11 leveling up in here is really
1:52:53 important this is it's just like
1:52:55 Scientology
1:52:57 just like Scientology
1:52:59 you want to reach the Promised Land
1:53:02 gotta level up
1:53:03 life is a competition
1:53:06 I'm a gen xer we have winners we have
1:53:09 losers
1:53:10 not everyone should get a part of
1:53:12 participation certificate okay
1:53:15 they shouldn't
1:53:17 it's not how life Works we're not
1:53:19 preparing the youth for how the [ __ ]
1:53:22 world works
1:53:24 I know they want to feel included I know
1:53:26 they want to feel warm and fuzzy inside
1:53:27 but guess what they're gonna get their
1:53:29 ass handed to them when they show up and
1:53:32 you know Jimmy down in the sales
1:53:34 department decides to wipe their ass
1:53:36 with his
1:53:37 ambition
1:53:41 you know what I mean you know you know
1:53:44 you know what I'm saying yeah no no all
1:53:46 right fine fine
1:53:48 play your Cooperative games
1:53:51 Settlers of katana Katan
1:53:54 do it everybody wins everybody when
1:53:56 you're right everybody wins it's not a
1:53:58 winner take all world you're right
1:54:00 you're right I'm wrong I'm wrong
1:54:02 completely you win
1:54:05 no you don't win uh we all win uh but
1:54:07 but you're correct All Right Moving On
1:54:14 oh yeah the Chipotle thing hop up oh you
1:54:18 were here last night bubbles cool yeah
1:54:20 fantastic
1:54:21 all right
1:54:23 just amazing to me what you could use as
1:54:27 far as real estate or other things like
1:54:28 that any literally anything any data
1:54:31 source
1:54:33 I popped in into into Advanced Data
1:54:36 analysis I rolled three different
1:54:38 screenplays that I'd written into into a
1:54:41 single upload and just started asking
1:54:43 questions
1:54:44 about the stories and characters across
1:54:47 my screen place and it was just like
1:54:49 instant got it did it right it's like
1:54:51 [ __ ]
1:54:53 it's amazing just amazing
1:54:57 um
1:54:58 I hope this live screen is being
1:55:00 recorded hey Molly so these are recorded
1:55:03 I have until September 21st so I'm a
1:55:06 little behind here's the deal we do have
1:55:09 we do have a YouTube channel
1:55:11 so it's AI learning lab Dash TT I have
1:55:15 like
1:55:15 through June
1:55:19 middle of June or end of June up there
1:55:22 and I don't have descriptions written
1:55:24 for him I'm very far behind I got a
1:55:26 whole little shame bucket about how I
1:55:28 haven't anyway and then September 21st
1:55:32 tick tock's gonna stop storing these
1:55:35 things for 90 days so every 90 days
1:55:37 there's this rolling 90-day window where
1:55:39 the old ones are just dropping off so
1:55:41 I'm okay for now but it's gonna go to 30
1:55:43 days in in about two weeks so I gotta
1:55:46 get my [ __ ] together and get this stuff
1:55:48 up there but but yes it is being
1:55:50 recorded and I will get it up to YouTube
1:55:52 but right now not so much
1:55:55 all right
1:55:57 or like I'm not I'm not keeping up to
1:56:00 date maybe I'll do that this weekend
1:56:02 what are some good questions to ask code
1:56:04 interpreter
1:56:07 see here's here's my problem
1:56:10 and is I don't
1:56:13 I know bupkis about data analytics I
1:56:17 like I just don't I don't know anything
1:56:18 about it I'm not good at math I don't I
1:56:21 don't know how to think in mathematical
1:56:22 terms what I would say is
1:56:32 don't
1:56:33 don't start with the tool start with
1:56:35 your business so so say
1:56:39 um
1:56:43 like what are the challenges in my
1:56:45 business or what are the kinds what are
1:56:47 the kinds of things that if I knew about
1:56:49 my business
1:56:50 could help me make decisions better
1:56:52 something something like that
1:56:55 and then say
1:56:58 what data might we have
1:57:02 even if it's
1:57:03 malformed not well organized
1:57:07 in a bunch of PDF files somewhere
1:57:10 that
1:57:12 the information might be in there
1:57:16 and then and then I would say okay so
1:57:18 let me let me figure out where that data
1:57:20 is maybe it's all in PDF files and you
1:57:22 somehow can break that up and get it
1:57:24 into a CSV file or or maybe the the PDFs
1:57:28 are consistently formatted enough that
1:57:31 they have all the same sections whatever
1:57:33 it is
1:57:34 and then I like throw that in a zip file
1:57:36 and upload it to
1:57:38 code interpreter and just start asking
1:57:40 your business questions say you know
1:57:42 take a look at the data there here's
1:57:44 here's what my business is about
1:57:46 here's the kind of insights I'm looking
1:57:48 for here's a pile of crap data
1:57:52 or a good date it doesn't matter here's
1:57:54 here's some data
1:57:58 and then just ask it
1:58:01 ask you ask it to figure it out for you
1:58:04 do you have enough information in this
1:58:06 data to be able to answer my questions
1:58:08 above
1:58:10 and it because it wants to please you it
1:58:13 will probably figure out ways to at
1:58:15 least give you some part of it and then
1:58:17 what you can do is you can say
1:58:19 what data do you need in addition to
1:58:23 what I gave you that could actually help
1:58:25 you give me answers to my questions
1:58:27 that's how I would do it is
1:58:32 because if you just start with I've got
1:58:34 this tool code interpreter and I've got
1:58:36 a pile of data what questions do I ask
1:58:38 it it's like it's it's very similar to
1:58:40 the kind of I I get this sort of white
1:58:44 page syndrome you know you know when you
1:58:46 just stared a piece of paper and you
1:58:48 could write anything and you're like I
1:58:50 don't know what to [ __ ] write I don't
1:58:52 you know I know what I could write
1:58:53 you're a terrible [ __ ] writer you're
1:58:55 a terrible [ __ ] writer you know that
1:58:57 syndrome I get that with the with the
1:58:59 chat GPT box because I'm like okay it
1:59:01 can answer anything
1:59:03 am I worth are you worthy enough to ask
1:59:06 a question
1:59:08 that is worthy of me answering it oh no
1:59:11 I don't know what to ask
1:59:13 so I find it really intimidating but the
1:59:16 minute I go to oh here's my business
1:59:18 here's a problem I'm facing oh I know
1:59:21 what to ask Chad gbt then I go over
1:59:23 there then I know what to ask it right
1:59:24 so I feel like code interpreters very
1:59:27 much the same thing but it's because
1:59:29 code interpreter is more Technical and
1:59:31 it's it's an area that I don't know I
1:59:33 find it even more intimidating so that
1:59:36 was cool last night the person said
1:59:39 if I've got
1:59:41 um geodata
1:59:43 can can is there any tool that can take
1:59:46 geodata and do some interesting analysis
1:59:48 with it I'm like oh yes code interpreter
1:59:51 can do that so I always start with the
1:59:54 problem I just find it much less
1:59:55 intimidating
1:59:58 um
2:00:04 oh yeah from uh Thelonious the USA F the
2:00:09 the Air Force pairs piloted jets with AI
2:00:12 drones yep that happened I think two
2:00:15 weeks ago or three weeks ago that story
2:00:17 came out
2:00:18 how fast it does it is ridiculous isn't
2:00:21 it Mike GPT 3.5 it's funny so when when
2:00:25 you go to um chat GPT if you're if
2:00:27 you're new to this [ __ ]
2:00:29 the free version is is GPT 3.5 and if
2:00:33 you're like me you always want the new
2:00:35 and shiny you're like I want the new one
2:00:37 gpt4 is the new one it's better well
2:00:40 gpt4 is better but it's slow as [ __ ]
2:00:43 molasses GPT 3.5 is just [ __ ] wicked
2:00:47 fast and it's and it's still good right
2:00:50 so if you just need to knock out an
2:00:52 outline or you just need to like do a
2:00:54 table comparing some Concepts or you
2:00:57 know quickly knock out a marketing
2:00:59 strategy framework or something like
2:01:01 that 3.5 is great because like you can
2:01:04 just [ __ ] plow through
2:01:06 iteration after iteration after
2:01:08 iteration and then and then what you can
2:01:10 do is it use 3.5 to get the blocking and
2:01:13 tackling done and then take components
2:01:15 of that over to gpt4 and refine it
2:01:17 because GPT Forge is better
2:01:20 but it's new it's new well relatively
2:01:23 slow it's still way faster than you are
2:01:25 it's like you're still gonna go even at
2:01:29 its slowest this is still pretty [ __ ]
2:01:32 insane that it does this
2:01:34 a data analyst who is available when you
2:01:36 wake up with an idea to research at 4am
2:01:38 yep that's exactly what these tools are
2:01:41 that's a thing on your time what you
2:01:45 need in this moment
2:01:46 that's why I was saying to paint the
2:01:48 other night that
2:01:50 I really hope as Google starts to
2:01:56 as as Google moves toward monetization
2:01:59 of their GPT tools
2:02:03 that they don't pollute them with the
2:02:05 paid [ __ ] like they polluted search
2:02:07 because search at this point like like
2:02:10 the glaring
2:02:15 uselessness of most search results
2:02:19 is made way more apparent when you have
2:02:21 a tool like Chachi BT or Claude or bard
2:02:25 that just [ __ ] answers your question
2:02:28 like yesterday or two days ago when my
2:02:30 car broke down and I was having
2:02:32 electrical problems
2:02:34 like it was a crisis right like I needed
2:02:37 to like I had to figure out what am I
2:02:39 doing at work tomorrow do I need to get
2:02:40 this car to the shop do I need like
2:02:42 you're in crisis right
2:02:44 and ideally you got a buddy that's a
2:02:47 mechanic but I don't right so I couldn't
2:02:49 just call someone up and go what the
2:02:50 [ __ ] do I do
2:02:52 so I got the I got the roadside
2:02:54 assistance well roadside assistance guys
2:02:57 are like I got the battery charger and
2:03:00 my guy was Advanced he had like a fuse
2:03:02 Checker so we could check fuses and we
2:03:05 could charge the battery but he didn't
2:03:06 know anything right so I'm still on my
2:03:09 own
2:03:10 and then when I was trying to use Google
2:03:12 and I was trying to use YouTube and I
2:03:13 was trying to use all these things that
2:03:15 are like here's all these links of
2:03:16 things that could possibly be everything
2:03:18 was [ __ ] littered with you know by
2:03:21 your extended warranty for your Ford
2:03:23 Flex and you're this and you're that and
2:03:25 I'm trying to dig I'm like I just want a
2:03:27 [ __ ] answer and even when I went to
2:03:29 Reddit it was just like there was old
2:03:30 [ __ ] and new [ __ ] and I was asking about
2:03:33 a Ford Flex but why are they talking
2:03:34 about Alfa Romeos and it's just like you
2:03:37 it like you just sometimes you just
2:03:39 [ __ ] need the answer
2:03:42 and so I went to Pi this stupid [ __ ]
2:03:45 thing that talks to me like I you know
2:03:47 you can have like a therapy conversation
2:03:49 with it I'm like hey do you know
2:03:51 anything about electrical systems and
2:03:53 cars of course I do I know everything I
2:03:55 know all of the internet so yes
2:03:58 oh
2:04:00 here's my symptoms here's what happens
2:04:02 oh okay yeah yeah it sounds like you
2:04:05 you've got some things figured out there
2:04:07 let me ask you a couple other questions
2:04:11 so Google better not [ __ ] that up
2:04:15 well they can
2:04:17 but they're gonna lose market share to
2:04:20 people that just answer your questions
2:04:24 how are we doing time wise 10 15. got a
2:04:27 little bit
2:04:29 got like 15 minutes
2:04:30 you know how that goes
2:04:33 chat 80d
2:04:35 she had 80d At Your Service how can I
2:04:37 help you
2:04:41 I was in this channel I think I learned
2:04:44 a lot it was a little chaotic I don't
2:04:46 know what I watched
2:04:52 how long have you watching for
2:04:55 better part of three hours
2:04:59 everything all right with you
2:05:01 sure about your life choices no no I'm
2:05:04 not I don't know what's happening
2:05:06 anymore
2:05:08 um
2:05:08 that's not what I wanted
2:05:10 that's not what I wanted that's what I
2:05:13 wanted
2:05:14 all right
2:05:17 what window was to Do's AI wait oh what
2:05:21 window
2:05:22 was to das
2:05:24 AI is to everything yep exactly and what
2:05:27 the web was to the internet
2:05:29 chat gbt is to machine learning and AI
2:05:32 yep exactly
2:05:35 so I have Coursera cool thank you very
2:05:38 much middle school private
2:05:42 oh for the were you you're the one that
2:05:44 was a substitute teacher
2:05:46 yeah middle school middle private school
2:05:49 those kids must have been [ __ ] brutal
2:05:51 to you as a substitute teacher good Lord
2:05:53 over privileged little [ __ ]
2:05:58 they're probably smart well educated
2:06:00 they're they're hedge fund hedge fund
2:06:03 daddy taught him how to be brutal to the
2:06:06 plebs good Lord that must have been
2:06:08 awful
2:06:10 yeah you should learn this [ __ ] so
2:06:14 you're you can have chat EBT help you
2:06:16 figure out how to get back to those
2:06:17 little Devils well ain't that special
2:06:20 look at look at a little Janie look at a
2:06:23 little Janie driving up in her Mercedes
2:06:25 wagon
2:06:28 oh good lord
2:06:30 it's unbelievably easy it does it next
2:06:33 Generation born with actual technology
2:06:37 missed the difference before and after
2:06:39 yeah exactly that we're so all of us all
2:06:43 of us all of this right here by the way
2:06:46 Irregulars you're up to 51 000. it's
2:06:48 remarkable what you're doing tonight
2:06:51 um
2:06:53 uh what was it gonna say
2:06:55 um
2:06:57 yeah we're we're at an amazing time
2:06:58 right now we're we're all
2:07:02 we all right now don't know what AI is
2:07:05 going to turn into
2:07:07 so right now we're living in a world
2:07:09 without AI we're just starting to be
2:07:11 able to play with these toys and they're
2:07:12 like
2:07:14 um I think this is pretty remarkable
2:07:16 isn't it anyone anyone Bob any Ai No no
2:07:21 okay I'll just I'll be I'll be over here
2:07:25 um I think this changes everything like
2:07:27 that's that's where we are right now
2:07:30 three years from now
2:07:32 we're gonna be like remember when we
2:07:34 weren't sure what this was gonna do to
2:07:36 copywriting
2:07:38 or screenwriting or
2:07:42 your career here
2:07:48 yeah we're an amazing time it's just
2:07:51 like Scientology
2:07:54 um when they ask how you did it don't
2:07:56 admit AI be a secret cyborg yeah so
2:07:59 Ted's referring to this Ethan Malik
2:08:01 article called discovering the secret
2:08:02 cyborg where basically a lot of people
2:08:04 in business are becoming AI literate and
2:08:08 and
2:08:09 dramatically improving their
2:08:11 productivity but not telling the
2:08:12 companies about it
2:08:14 um and you know listen right now we can
2:08:17 do that as these tools get more and more
2:08:19 integrated in what we do you know
2:08:21 businesses are just going to need to
2:08:22 figure out how to do it and they're
2:08:24 going to expect their employees to do it
2:08:25 but
2:08:27 for all of us that say okay I'm in let's
2:08:31 start exploring
2:08:33 when that time comes in your company
2:08:35 where they're like hey I gotta figure
2:08:36 this [ __ ] out does anybody know anything
2:08:37 oh yeah covaruno does she talks about it
2:08:40 all the time she doesn't shut up about
2:08:42 AI it's really annoying okay can I talk
2:08:44 to her right now
2:08:48 yep that gonna be good be that person
2:08:52 don't be the person who's like AI is
2:08:54 stupid am I stupid
2:08:57 I like doing things the way I've always
2:08:58 done them
2:09:02 I'm a turtle
2:09:04 if you're looking to make a complete
2:09:06 career change in future pre proof
2:09:09 where's the best place to learn AI
2:09:15 right now like I said there's there's no
2:09:18 experts right now oh good lord what the
2:09:20 hell did I do to my hair Jesus
2:09:23 thank you sorry
2:09:26 thank you Caroline I love me some donuts
2:09:29 I rub them in my hair that's how I get
2:09:32 the lovely shape
2:09:33 um
2:09:36 it's a great question
2:09:38 oh my God I played said Lucifer the tiny
2:09:42 little people thank you for the beaver
2:09:43 that locks me a good Beaver
2:09:46 beaver's tasty all right
2:09:48 um
2:09:49 seriously though complete career change
2:09:51 you want to Future proof it
2:09:57 the most phallic of the tick tock
2:10:01 gifts thank you thank you every time I
2:10:04 see it it throws me for a loop I'm like
2:10:06 oh good lord
2:10:09 I don't even see the ears
2:10:11 [Laughter]
2:10:14 well I was on a safari
2:10:18 a very well endowed trunk
2:10:23 um
2:10:25 so
2:10:28 I don't have a good answer for you about
2:10:30 like so I don't think any of us can
2:10:33 future proof anything from anything
2:10:37 I think that
2:10:40 that the shifts in
2:10:44 every single position
2:10:48 are going to be fairly profound I don't
2:10:50 know what time scale it's going to
2:10:51 happen but I think they're going to be
2:10:53 fairly profound
2:10:55 um
2:10:57 the only thing I'm confident about is
2:10:59 that
2:11:00 people who are AI literate are going to
2:11:02 have a distinct advantage over people
2:11:04 who are not AI literate so I wouldn't
2:11:08 this for Dave Axelrod I wouldn't put
2:11:12 um
2:11:14 I wouldn't put the expectation on
2:11:16 yourself that you actually have to like
2:11:19 there's there's not a class that's
2:11:20 teaching you what you need to learn
2:11:22 there are there might be certificates
2:11:24 out there there might be things that are
2:11:26 like here's a prompt engineering course
2:11:27 and here's how in AI Ops course and
2:11:29 here's you know like like Rachel Woods
2:11:32 here on Tick Tock she's got a thing
2:11:33 called the AI exchange and she's doing
2:11:35 some really tactical good work around AI
2:11:38 Ops and prompt engineering and things
2:11:40 like that I think those are all good
2:11:42 skills to have
2:11:44 but
2:11:46 anything that you learn in the next
2:11:48 three months
2:11:49 is going to be irrelevant and completely
2:11:53 different to what you're going to need
2:11:54 six months after that
2:11:57 so we're so we're in a rolling series of
2:11:59 Innovations right now where all of these
2:12:01 tools right now are just spit and duct
2:12:04 tape they're just as I call them janky
2:12:06 pieces of [ __ ]
2:12:08 they're janky pieces of [ __ ] sitting on
2:12:11 top of a remarkable infrastructure so
2:12:13 what they do is remarkable
2:12:15 but what skills you need to get the best
2:12:18 out of out of them is going to change
2:12:20 every three months for the next probably
2:12:22 two or three years so the most important
2:12:25 thing you can do is take what you're
2:12:27 good at what you're passionate about
2:12:29 what you want to do
2:12:31 get clear about that first
2:12:34 don't mistake
2:12:37 AI tools for a strategy
2:12:40 this isn't the strategy this is the tool
2:12:43 the strategy is what
2:12:46 what's the stuff that you're good at
2:12:48 that you want to do more of
2:12:51 and then you can also figure out what's
2:12:52 the stuff that you're good at that you
2:12:53 want to do less of
2:12:55 right
2:12:56 because we all have those
2:13:01 so take the stuff that you're good at
2:13:02 that you want to do less of and figure
2:13:05 out how AI can help you automate
2:13:08 or make the stuff you don't enjoy faster
2:13:12 more automated just get the work done
2:13:15 blow it out get the work done blow it
2:13:16 out don't get precious about that [ __ ]
2:13:19 the stuff that you're good at that you
2:13:20 want to do more of or the stuff that
2:13:22 you're not good at and you want to do
2:13:24 more of
2:13:25 really dig into these tools and figure
2:13:27 out how can I get better at that stuff
2:13:30 how can these things amplify what I want
2:13:33 to do who I am who I want to become
2:13:36 and just focus on it as this rolling
2:13:39 learning exercise and don't have
2:13:41 expectations that there's a there's a
2:13:43 very clear destination because there's
2:13:45 not nobody knows what's coming nobody
2:13:49 um Google is launching a thing called
2:13:52 Gemini probably within a month that
2:13:55 they're claiming is five times better
2:13:57 than gpt4 and it's multimodal so it does
2:14:00 pictures and images and sound and all
2:14:02 that sort of [ __ ] if that's true in
2:14:04 any semblance of the case there will be
2:14:07 a whole new set of skills required to
2:14:09 deal with Gemini
2:14:11 and then good old Sam Altman and the
2:14:14 wacky kids over at open AI are going to
2:14:16 respond to that
2:14:19 because Sam has said publicly we're not
2:14:21 training chat gbt5
2:14:26 do you believe that I don't of course
2:14:28 they're [ __ ] training gpt5
2:14:32 this is an arms race
2:14:36 you think Microsoft's sit and back going
2:14:38 man let Google Take the Lead back no
2:14:43 so so this stuff's gonna change
2:14:46 so the biggest thing you can do is just
2:14:48 get clear about what you want and just
2:14:51 learn as much as you can just get into
2:14:52 sponge Motors I call it I'm in Swiffer
2:14:54 mode there's sponge mode where you're
2:14:56 learning a lot Swiffer mode is like
2:14:58 you're trying to learn it all
2:15:04 yeah
2:15:06 the moon Rose
2:15:10 oh my God I played settlers Katan
2:15:12 bubbles came back she likes us hey
2:15:14 bubbles she likes us she really really
2:15:17 likes us 1005 in Denver that was 20
2:15:20 minutes ago
2:15:22 Kyle your CEO delegate I know I need do
2:15:25 need to delegate more Alexa has a GPT
2:15:27 feature I'm adding it right now oh cool
2:15:29 oh that's great that'll be interesting
2:15:31 yeah I mean honest to God Apple
2:15:37 Siri needs a facelift
2:15:40 Siri needs a brain lift
2:15:43 Siri needs to not suck
2:15:47 now here's the good news about Apple
2:15:49 they have been spending the bulk of
2:15:52 their research budget which their annual
2:15:54 research budget this is [ __ ]
2:15:56 mind-blowing Apple's annual research
2:15:59 budget is 24.6 billion dollars
2:16:03 it's more than the market cap of of open
2:16:06 AI
2:16:08 that's their research budget they've
2:16:10 spent the bulk of their research Budget
2:16:12 on generative AI so Apple when they come
2:16:15 out they're going to come out swinging
2:16:18 They're Gonna Come Out Swinging and
2:16:20 everyone else is gonna go like oh that's
2:16:22 what you meant by generative AI these
2:16:25 were just jokes this was just a these
2:16:27 were just like sketches on a napkin from
2:16:29 a restaurant we were at a diner one
2:16:31 night we thought chat GPD could be fun
2:16:33 we put it out 100 million people use it
2:16:36 it was awesome
2:16:37 anyway that Apple stuff's really
2:16:39 exciting isn't it what do we do
2:16:43 I hope they don't disappoint me Apple's
2:16:45 been disappointing me ever since Jobs
2:16:46 died it's just like
2:16:49 is someone over could someone over there
2:16:51 get maniacal about quality again
2:16:54 throw a chair through a window in the
2:16:56 spaceship please
2:16:58 anyone Johnny Ives is like yeah I'm good
2:17:01 I'm out
2:17:05 The Vision Pro stuff was impressive
2:17:08 concession BT pick my fantasy team for
2:17:11 this week absolutely in fact uh code
2:17:13 interpreter go get some NFL data
2:17:17 go find the data because I'm sure it's
2:17:18 on kaggle if you don't know kaggle
2:17:21 k-a-g-g-le.com has data sets I'm sure
2:17:24 there's NFL data in there for whatever
2:17:27 you need it for pop it into code
2:17:29 interpreter or Advanced data analytics
2:17:31 in in chat GPT and start asking away
2:17:36 tell it you've got this many points
2:17:41 you know you need a balanced team you
2:17:43 want to focus on yardage rather than
2:17:45 running or whatever the hell it is
2:17:48 and yeah have it start picking your team
2:17:51 I don't I don't know enough about how
2:17:53 that [ __ ] works
2:17:54 to know how to tell you what to do but
2:17:57 yes it can absolutely do that
2:17:59 really really good all right I'm gonna
2:18:02 do another
2:18:03 few minutes here
2:18:05 chat TBD helped me research my machine
2:18:08 learning for molecular design very cool
2:18:11 molecule design yeah Johnson royalty the
2:18:14 uh the uh deep mind uh
2:18:18 Alpha fold thing
2:18:20 holy [ __ ] 400 000 proteins like in two
2:18:24 days or whatever it was
2:18:26 it's just like the Breakthrough the
2:18:29 breakthroughs so here's the thing about
2:18:31 this Tech
2:18:34 this this generative AI Tech the stuff
2:18:37 that Chachi BT is made of
2:18:40 it's not just exciting to non
2:18:43 machine learning people
2:18:45 that I know that have been in machine
2:18:47 learning for a decade and have built
2:18:50 systems using the current state of the
2:18:52 art over the decade they're now looking
2:18:55 at what this generative AI stuff is
2:18:57 doing and they're like [ __ ]
2:18:59 they're like a little bitter about it
2:19:01 because because it's so much better and
2:19:05 so a lot of the scientists are taking
2:19:08 these Technologies and going back to
2:19:09 their data and and
2:19:11 making big advances so I the the
2:19:14 advances we're going to see in science
2:19:16 and medicine and all that sort of [ __ ]
2:19:19 over the next I don't know two to three
2:19:21 years is just going to be face melting
2:19:24 face meltingly powerful
2:19:26 it's janky or Jackie
2:19:29 gotta hop good night Irregulars good
2:19:31 night Ted yeah I gotta hop too I'm out
2:19:33 of here Share Bear Silver Fox let's see
2:19:36 it's remarkable I learned bash scripting
2:19:39 code explanations good at code
2:19:41 explanation
2:19:43 [Laughter]
2:19:47 Winston I believe this is going to be a
2:19:49 complete Revolution generating enormous
2:19:51 wealth through the AI explosion yeah the
2:19:54 most recent McKinsey
2:19:57 um state of tech 2023
2:19:59 predicts uh 4.4 trillion dollars in new
2:20:05 Revenue because of AI and then I've
2:20:08 heard a couple of others that are north
2:20:09 of that that are in the teens
2:20:11 you know 15 or 16 trillion or 14
2:20:14 trillion something like that
2:20:16 um
2:20:17 the McKinsey predictions from the mid
2:20:19 90s were way under so they tend to be a
2:20:23 little conservative so 4.4 trillion as a
2:20:26 economic impact is probably low so yeah
2:20:29 it's going to be it's going to be pretty
2:20:30 remarkable
2:20:32 anyway listen I gotta go I gotta go to
2:20:36 bitty bye it's Saturday night you owe
2:20:38 you should go out dancing go go you know
2:20:40 Saturday Night Fever you know put on
2:20:42 your disco shoes it's time it's perfect
2:20:44 if you're in New York what time is it
2:20:46 there 10 30 it's 12 30. this is per the
2:20:49 evening is just getting started you'll
2:20:52 go out to those hot clubs you put on the
2:20:54 Spandex and you just go
2:20:56 and I know we're all too old for spandex
2:20:59 but just do it anyway just get get out
2:21:01 go party all right or go to bed
2:21:06 peace everybody all right let's do a
2:21:08 little Romper Room say goodnight to
2:21:10 everybody uh Silver Fox good night
2:21:12 Winston thank you okay farmer Tobias
2:21:16 Andrea 444 night everybody let's see
2:21:20 Stevo peace out brother Vicky
2:21:23 um oh my favorite user
2:21:28 167-389-60-85942 great seeing you come
2:21:31 back change your name
2:21:34 Dr Jay Emilio's wife peace out Kevin
2:21:37 share bear have a good one
2:21:40 Johnson royalty love the name by the way
2:21:42 cabruno good night
2:21:44 micro uh this is not an exit sorry this
2:21:48 is not an exit this is an exit I'm out
2:21:51 jrc peace everybody all right micro
2:21:54 thanks all right I'll see you guys
2:21:56 tomorrow