AI Learning Lab

aiLL 11/25/23 - Navigating the Wilderness of Generative AI

k3z4Or3IfDY
Video2023-12-162:27:199 views

Description

Kyle Shannon provides an in-depth tour of the current state of generative AI in late 2023, including an overview of major players like ChatGPT, DALL-E, and more. He shares his perspectives on the opportunities and challenges presented by this rapidly evolving technology. Join in the exploration and discussion around using AI as a tool for creativity, business innovation, and beyond. Channel: https://tiktok.com/@aiLearningLab #ai #generativeai #chatgpt 0:37 - Intro singing 5:07 - Speculating on the OpenAI / Sam Altman drama 8:20 - Philosophizing on the impact of AI 15:29 - Reflecting on his web pioneer days 18:39 - Defining artificial general intelligence (AGI) 23:19 - Advocating playing with AI to gain experience 31:13 - Demoing avatar and voice synthesis 35:49 - Advising getting involved with AI open source projects 40:00 - Celebrating the 1-year anniversary of his AI Salon community 44:46 - Encouraging joining his AI Salon community 53:50 - Lamenting his live demo struggles 58:54 - Marveling at ChatGPT's analysis capabilities 1:24:21 - Lamenting the state of being able to find openly available ebooks

Chapters

Transcript

0:08 [Music]
0:20 [Music]
0:27 [Music]
0:29 oh
0:37 [Music]
0:48 Jeffy is singing hey
0:51 [Music]
0:59 everybody
1:03 every time I see you now get that
1:07 looking
1:09 my every time see your
1:12 mouth hear that
1:17 smile early Misty morning l i the
1:24 engine
1:28 outside this is why it's hard to sing
1:30 around
1:32 here you
1:35 [Music]
1:40 me no rapping no rapping you convince
1:44 me again
1:47 today you're leaving this H to looking
1:51 for someone else's Golden
1:54 [Music]
1:57 Ring should say so long
2:03 [Music]
2:06 suzan H now don't you
2:10 cry oh that's Out Of
2:13 [Music]
2:25 Tune can you see the
2:28 doggo um
2:31 yeah hang
2:38 on doggo
2:41 [Laughter]
2:45 [Music]
2:58 Champion champy they want to hear he you
3:00 sing come
3:01 [Music]
3:09 on well I heard was a secret call David
3:15 played and it pleas the Lord you don't
3:19 really care for music do
3:21 [Music]
3:23 you it goes like this the fourth the f m
3:28 fall and major Li the Baff King compos
3:32 it
3:35 hallelujah
3:40 hallelujah hallelujah all right enough
3:43 singing enough
3:47 dogging hey
3:49 everybody Welcome to the AI learning lab
3:52 the AI singing
3:55 lab the AI howling
3:59 lab
4:02 good singing champy good singing nice
4:06 tone nice tone can't keep a beat for
4:09 [ __ ] but he can
4:13 sing Oh Champion has the voice of an
4:16 Angel
4:17 right
4:19 [Laughter]
4:22 champua oh man all right let me get
4:24 reset back up here I got all my all my
4:27 all my gear in a Kur fluffle
4:30 [Laughter]
4:35 howdy howdy howdy happy Saturday night
4:38 happy AI learning lab Happy Thanksgiving
4:42 weekend happy open AI still a company
4:49 week got a terrible drinking
4:53 problem do you believe the rumors of Q
4:55 and super intelligence I don't know
4:59 um
5:00 it's all such
5:02 speculation um I find it hard to
5:08 believe it it feels to me like it's it's
5:10 a a series of events that
5:14 all kind of added up but I think
5:18 something had to trigger I I think
5:20 something other than you know Sam Alman
5:23 getting pissed off
5:25 cuz that one woman you know wrote
5:28 something in a white paper that he
5:29 didn't like
5:30 it just it does that
5:34 feels I don't know that feels Petty like
5:38 this does not feel like a petty like it
5:40 would the board absolutely had to know
5:44 that what they were going to do was
5:45 going to blow [ __ ] up and it was going
5:47 to get ugly so I don't think you do that
5:50 lightly like you know even even if they
5:52 don't have business sense which they
5:54 they clearly didn't in my
5:56 opinion
5:58 um you don't do that [ __ ] lightly so I
6:01 so it it seems to me something happened
6:04 that precipitated it and and that's
6:06 plausible if you don't know uh what the
6:08 question was
6:11 there's there there was apparently a a
6:14 letter sent to the board about a week
6:17 before all this happened uh that
6:20 basically said they had made a
6:21 breakthrough and Ilia sua or
6:25 Suk I don't people pronounce it all
6:27 sorts of different ways I don't know how
6:28 he pronounces it but um he apparently
6:32 been working on this thing called
6:34 qar which is a combination of from what
6:38 I can
6:39 gather again this is more like chat TMZ
6:43 than a news
6:44 organization so take everything here
6:47 with a giant grain of
6:50 salt um there was a there's a thing
6:53 called Q learning and there was a thing
6:55 called AAR and apparently qar might be
6:58 some sort of merging of of these two
7:00 techniques that basically basically what
7:03 they allow this is a model that from
7:05 what I can understand it sits on top of
7:07 the G the large language model and it
7:10 basically says Hey in any given State
7:13 you've got some actions you can take
7:16 take the action that has the highest
7:18 probability of success and then we'll
7:21 score you on that and then the Q score
7:24 is a quality score either goes up or
7:27 goes down on each action it takes and so
7:28 it's basically a way it's it's Game
7:30 Theory kind of stuff it's a way for it
7:33 to navigate
7:35 tasks um such that it it learns if it's
7:39 doing it right basically and so it it is
7:42 effectively reasoning so that's the
7:44 that's the uh that's the skinny on what
7:46 qar the results of qar are and then they
7:50 did it on some like lowlevel like grade
7:53 school level math but apparently they
7:57 got they got it to get
8:00 the math problems 100% right and they've
8:03 never gotten close to that and so the
8:06 minute you're there now you're into now
8:07 you're into these large language models
8:09 reasoning and and according to David
8:12 Shapiro they kind of stopped it before
8:13 it started
8:15 self-improving so I don't know I do not
8:18 know I don't know people all right let's
8:20 let's get started here let's get started
8:23 here you know you know we got we we like
8:26 talking about the artificial
8:28 intelligence here because I'm smart you
8:31 know I got a lot of brains I'm going
8:32 looking at all of it that's this is all
8:35 cranial capacity you know what I
8:40 mean um so do me a favor share the live
8:44 so we can get word out to get some some
8:45 folks in here it's a Saturday night PE
8:49 people out having a life having a date
8:51 on a Saturday night no we've got AI to
8:54 do people let's get some people in here
8:56 do me a favor hit the uh tap the screen
8:58 get the likes going and uh and we'll get
9:02 going we we'll get started we'll we'll I
9:05 don't know I'll give you a list of [ __ ]
9:06 you can go play with I can answer
9:09 questions if you have questions about AI
9:12 happy to talk about my background I'm an
9:13 entrepreneur a many time entrepreneur I
9:15 don't know 12 or 15 companies some
9:17 ridiculous amount I am officially
9:19 unhirable thank you thank you thankful
9:22 Joy I think is your name first gift
9:24 thank you for the rose I appreciate that
9:27 um if you have questions about a pop
9:29 them in the comments below I'll do my
9:31 best to answer them we should have some
9:32 mods in here that'll help flag the the
9:35 questions that I can
9:37 answer um at least I can go out at night
9:42 good old Aussie time oh right Drew yeah
9:45 you and steo look at that Tick Tock
9:53 merch where' champy go champ left the
9:55 minute the guitar left champy was like
9:57 I'm getting on the comfy bed
10:01 all
10:03 right um oh so I was saying so I'm an
10:06 entrepreneur my my current gig is I'm
10:08 the co-founder and CEO of a company
10:10 called
10:11 storyvine which is an automated uh video
10:15 storytelling platform we do a lot of
10:16 work in healthcare and Pharma I know
10:18 some of the Irregulars in here work in
10:20 the healthcare space and the Pharma
10:21 space so hello to
10:23 you um and uh we've been around for 11
10:29 and A2 years so that's good um it's been
10:33 a weird year business-wise I think for
10:37 everyone um and I've added a whole bunch
10:40 of AI stuff to the product road map
10:42 thanks for cluing people in on using AI
10:45 to put their creativity to purpose and
10:49 hopefully monetize it absolutely tanker
10:52 a kush yeah the uh the uh the cliche you
10:56 know early on there there's a lot less
10:57 of this or I see a lot less of this on
10:59 the on the Tik Tok these days but there
11:01 were a lot of these channels that were
11:02 like you can make money
11:04 with so we we made enough fun of them
11:06 that I got myself a gold button to say
11:10 it um and and the thing that
11:13 was annoying about that kind of content
11:16 was it it mistakes the tool for a
11:19 strategy right it's not about making
11:21 money with chat
11:23 gbt how you make money is you solve
11:25 problems and chat GPT just happens to be
11:28 a tool in an arsenal of a bunch of them
11:32 that are really powerful tools and so
11:33 that's kind of what we talk about here
11:34 is how do we use them things like that
11:37 any news on when the GPT store will go
11:39 live I have no idea um one of the rumors
11:42 about all the tension at at uh open AI
11:48 was that the Sam prematurely announced
11:51 the store I that does not feel like
11:54 credible to me at this point but
11:55 whatever who
11:57 knows um
11:59 I would be
12:01 surprised if the Kur fuffle last week
12:04 Kur fluffle is it Kur fuffle or k
12:07 fluffle I need I need my English
12:10 language homies to help me out with it I
12:13 think it's
12:15 kerfuffle I think I would be surprised
12:17 if the Kur fuffle that's what I'm going
12:19 with right now didn't delay things so my
12:23 my guess is that the GPT store will
12:28 launch
12:30 well okay two ideas one is it launches
12:33 on November 30th they said it was going
12:34 to be the end of the month so November
12:37 30th is the one-year anniversary of chat
12:41 GPT I still haven't decided what I'm
12:43 doing on the 30th or the 31st or what
12:46 it's is it the 31st no it's not it's the
12:49 1st of December um I might do a
12:53 24hour uh marathon thing so so for the
12:58 anniversary but
13:00 um
13:04 um so it could be the 30th right one
13:08 year anniversary here's what's what's
13:10 effectively the app store for
13:14 gpts um launching then my my suspicion
13:17 is it'll be mid December or early
13:21 [Music]
13:22 [Laughter]
13:24 January how soon on the horizon do you
13:26 think image generation will get better
13:28 issues with tech Etc well it we should
13:32 have a holiday party yeah it's a good
13:33 idea we should oh that's that's actually
13:35 a really good idea yeah we we we'll do a
13:38 holiday party that might be
13:40 fun we can we can bring a bunch of
13:42 people up on stage and and just chitter
13:44 chatter and drink eggnog and [ __ ] I like
13:47 it call it a
13:49 Schism but they've they've Des Schism it
13:52 cuno it's now now it was just the Kur
13:55 fluffle Kur fluffle Kur fluffle get your
13:57 smart pants on never I will never wear
14:00 smart
14:02 pants usually sir if we work for British
14:05 Airways we wear smart
14:12 pants that was how I heard her say
14:20 it she she slapped me around in the most
14:23 British way
14:26 possible oh man
14:32 um what else we talking about here
14:36 um if you're new to this AI stuff one of
14:38 the one of
14:40 my one of my
14:42 historical um
14:45 experiences is that I was an
14:46 entrepreneur in the early days of the
14:48 worldwide web I started the the
14:50 smartpants joke was uh I started an
14:53 agency in 1994 that was one of the first
14:56 web agencies a digital agency and
14:59 um British Airways we we won the British
15:01 Airways account and they were one of our
15:03 clients and I got yelled at by a flight
15:05 attendant for not wearing smart
15:08 pants it's a good story but it was
15:11 humiliating at the
15:14 time but but anyway the um one of the
15:18 reasons I'm so bullish on this AI stuff
15:22 generative AI stuff essentially starting
15:26 November 30th 2022 is when I'm that
15:29 that's the inle inflection point for
15:32 me um that chat GPT represents for
15:36 machine learning and AI what the
15:38 worldwide web represented to the
15:40 internet right so the internet used to
15:42 be command line and for researchers and
15:45 scientists and Geeks only Librarians and
15:48 then the worldwide web came along and
15:50 you could click on things to get to
15:52 other things and all of a sudden the
15:54 rest of us could not only navigate the
15:57 internet but we could put content on the
16:00 internet and that changed everything and
16:02 so chat GPT is similarly
16:06 simplifying um some incredibly powerful
16:08 technology that's got I don't know three
16:11 or four Decades of research and
16:14 development and infrastructure sitting
16:16 underneath it and so
16:19 we whoever the hell we are um get to
16:23 play with this [ __ ] and so that's what
16:24 this channel is about so um so yeah so
16:28 if if you're new here here
16:31 welcome welcome I don't know what's
16:33 going on with my voice it's getting a
16:35 little crazy it's getting a little
16:36 crunchy scoob scoob what are we going to
16:43 do sometimes I got you know dlid tones
16:47 and other times I got Scooby
16:53 voice and yes I drink coffee at night
16:57 um
17:01 do you think Auto GPT has vastly
17:03 improved um I haven't played with auto
17:06 GPT recently
17:13 um there there are there are little
17:16 hints of a uh autonomous action within
17:20 chat GPT now if you look at how it does
17:23 image prompting and if you look at how
17:26 it does um Cod interpreter it's doing
17:30 some autonomous prompting so I would say
17:32 there's there's little hints of it
17:35 inside um inside chat GPT right now um
17:39 but Auto GPT on its own if you don't
17:40 know what Auto GPT is it's a um it's an
17:43 autonomous agent so basically you give
17:46 it a goal and it takes actions on its
17:48 own it'll go do a Google Search and
17:50 it'll go do I don't know book you plane
17:53 tickets and this and that all that
17:55 theoretically it sort of you give it
17:57 tasks to or you give it a goal and it
17:59 figures out what tasks it needs to do
18:01 and it goes off and does them
18:04 um the the early things were they demoed
18:08 well but were pretty horrible and so I
18:11 don't know if it's gotten any
18:13 better but this qar thing um if the qar
18:18 thing is
18:20 true what it effectively does is it
18:23 course corrects if it takes wrong
18:25 actions and I think I think that would
18:28 move it
18:29 Beyond vastly improving
18:33 to something close to AGI if not AGI and
18:37 AGI if you don't know is artificial
18:39 general
18:40 intelligence um open AI defines AGI as
18:44 the point at which uh the the the
18:48 machine the chat GPT kind of thing um is
18:52 an autonomous system that
18:56 performs
18:57 um
18:59 how do they put it
19:02 valuable valuable tasks better than
19:05 humans most valuable tasks most
19:10 valuable I can go look it up but um but
19:14 yeah it's basically when when when chat
19:17 PT gets better than us at doing our
19:19 jobs so think any knowledge work any
19:22 creative work shoot am I late uh you
19:25 missed champ singing We we put champ on
19:27 camera for a little bit bit he did a
19:29 little little solo
19:31 action um I was trying to sing but you
19:35 know when you got a dog singing next to
19:36 you it's it's hard to carry a tune he
19:38 sings better than I
19:43 do
19:45 um all right so let's let's do this if
19:47 you're like but what do we do here are
19:49 we going to do anything here I thought
19:50 this was the learning lab are we going
19:51 to learn is does he ever teach I thought
19:53 this was a learning lab is he some sort
19:55 of Professor why is he wearing a t-shirt
19:57 what are his qualifications
19:59 does he have
20:00 qualifications
20:08 no
20:10 analytics okay so I'm gonna I'm
20:14 purposefully covering up that thing on
20:15 the right if you're brand new to AI if
20:18 you've not really played with it if
20:20 you're just like I don't know home for
20:23 the weekend
20:25 or some of your children are home for
20:28 the weekend they like Nana you really
20:30 got to get into
20:32 Ai and you stumbled on here cuz you
20:35 can't figure out any other way to do it
20:38 perfect you've come to the right place
20:41 go there and just start playing with
20:43 chat GPT if you've never played with it
20:46 it doesn't hurt okay here's one of the
20:48 misconceptions about AI I need to be a
20:50 mathematician from Stanford right nope
20:53 you don't you just can can you make
20:56 words sort of I'm bad with words perfect
20:59 it'll it's good with words it'll it'll
21:02 fix your typos and your Grammar get it a
21:06 a a poorly formed idea and it will
21:08 understand it um go ask it to write you
21:11 a poem go ask it to do all sorts of
21:13 stuff like that all the bullet points
21:14 below it are just this is just new [ __ ]
21:16 that they just announced two weeks ago
21:19 um so we can we can demo some chat GPT
21:21 stuff the one little caveat about chat
21:24 GPT right now is the really powerful one
21:27 chat GPT 4 is they're not taking new
21:31 submissions right now or new new
21:33 accounts for chat gb4 get your ass on
21:36 the wait list just get your ass on the
21:38 wait list it is the best uh it is the
21:41 sort of highest value 20 bucks a month
21:43 you can buy um even without the 20 bucks
21:47 a month the 3.5 chat GPT is pretty good
21:49 but there's a lot of other options okay
21:52 all right now if you're a little bit
21:54 more geeky that thing on the right is
21:56 the playground that's where Developers
21:58 AG go that want to use open AI
22:00 technology in their
22:02 apps um and we can talk about that the
22:06 other thing about chat GPT we can talk
22:07 about is they just announced these
22:09 things called gpts which are
22:12 personalized um customized versions of
22:15 GPT that you can create and share very
22:18 very easily so we can do that but get
22:20 ready get ready to do some
22:22 screenshotting okay I'm I'm going to
22:24 walk you through a series of of screens
22:27 this is all technology I think
22:30 if oh let me give a little bit more
22:33 Foundation here um I think the most
22:36 important thing you can do right now
22:39 with regards to AI is just be curious
22:41 about it get relentlessly curious about
22:44 it follow news about it play with it try
22:49 different tools don't make it don't take
22:52 anything too seriously just start
22:54 playing so all of the um all of the uh
23:00 sites that I'm about to show you are all
23:02 just things that I think are worth your
23:04 time um to to explore and so this this
23:08 list changes fairly regularly because
23:11 everything's changing so much but but
23:13 anyway get ready to screenshot okay so
23:14 screenshot that one so these are kind of
23:17 the big four sitting beneath chat GPT
23:19 you've got Claude which is the big
23:21 competitor to chat GPT Dario who used to
23:25 work at open AI left and started
23:27 anthropic which creates Claude they just
23:30 rolled out Claude
23:32 2.1 it's got a 200,000 token context
23:35 window which is basically means you can
23:37 put long ass prompts in there um grock
23:40 is
23:42 reportedly going live next week inside
23:47 of Twitter inside of X um as a feature
23:50 inside of X for any
23:54 premium account holders of of X of
23:57 Twitter and so that's basically if
23:59 you're paying 16 bucks a month you're
24:01 going to get grock next week Bard is
24:04 Google's entry into this space and
24:05 co-pilot is Microsoft's and co-pilot is
24:09 built on top of gp4 so I told you before
24:13 go to chat GPT but you can't have gp4
24:17 yet well you actually kind of can go to
24:19 co-pilot
24:21 microsoft.com that is ostensibly gp4
24:25 okay now it's not chat GPT it's co-pilot
24:28 but they're different they're they're
24:29 kind of both applications built on top
24:32 of the
24:33 gp4 large language model
24:38 um but but you can generate images with
24:41 it with dolly3 it can read images it can
24:43 do all the stuff that that chat GPT can
24:46 do but it's it's just a little I don't
24:48 like it as much CU their safety guard
24:50 rails are too high in my opinion all
24:53 right screenshot this bad boy so here's
24:55 the other four that I think are worth
24:57 learning
24:58 Pi perplexity llama 2 which is actually
25:01 a whole world of Open Source but llama 2
25:05 is the is the the um the model that meta
25:09 released for free open- source with
25:13 commercial license and then PO is a
25:16 really
25:17 interesting uh tool that that lets you
25:20 play with lots of different models now
25:23 all nine of these so chbt those four and
25:28 those four are all chat GPT like things
25:34 okay why would you need nine of them
25:36 Kyle tell us which one's the
25:39 best they keep changing a lot like Po
25:41 was about to drop off the list and then
25:44 po sort of up their game with their Bots
25:47 you can make a bot at Po and now they
25:49 just added the ability to add knowledge
25:51 to their Bots and you can sell your Bots
25:54 and commercialize them you can make
25:55 money you can make money with I
25:58 with
26:01 po uh so Po got really interesting um Pi
26:05 is really interesting they're releasing
26:06 a new version in fact yesterday I
26:08 noticed that in their conversational so
26:12 Pi's really good at conversational chat
26:15 gping they basically added a therapy
26:18 version of it like if you want to talk
26:22 to someone about some
26:24 issue um they've got a mode for that now
26:28 which I think is absolutely fascinating
26:30 so so learn get curious about this stuff
26:33 go play with all these tools and and
26:35 figure out which ones are good what
26:37 you're going to find like perplexity is
26:39 really good at research it does
26:40 citations and and it it hallucinates
26:43 less because it's doing everything based
26:45 on very focused searches so so
26:48 perplexity is really interesting but
26:50 it's it like I don't use it a lot
26:52 because I'm doing more creative writing
26:54 and I want the sort of creativity that
26:56 gp4 gives me um but if you're doing
26:59 research perplexity might be the tool if
27:01 you want to talk to to a thing like like
27:04 conversational then Pi's probably it
27:06 even though chat GPT Now can do that as
27:08 well um so go play with all these things
27:11 okay um here's the image ones screenshot
27:15 this tell us the meaning of life I mean
27:18 AI please exactly pi equals real time
27:22 internet data oh that's right tck yeah
27:24 so is po free so PO is
27:29 free except you can't play with the
27:33 fancy Claude models or the fancy open AI
27:36 models so if you want to use gp4 and
27:39 Claude 2 on PO you have to pay for it
27:45 but you can play with a lot of other
27:47 stuff there and it's um you can make
27:50 your own bot for free and you can
27:53 publish it for free and now you can
27:55 monetize it for free so so yeah so po po
27:59 is a really interesting one these are
28:00 the image creation tools I think are
28:02 worth understanding the boundaries of I
28:05 I think that's what I'm going to say now
28:07 more than anything is like I don't feel
28:09 like you have to master any of these
28:11 tools because these tools change on a
28:13 monthly basis and sometimes weekly and
28:16 sometimes
28:16 daily so just kind of learn which ones
28:19 are right for you and if if one if you
28:23 hate it then don't use it again but at
28:24 least you know so Dolly 3 is what lives
28:27 in inside chat GPT and it also lives
28:29 inside co-pilot mid journey is a
28:31 standalone independent company they've
28:33 never raised any money which I find
28:36 absolutely amazing they're doing like2
28:38 200 $300 million a year um just with a
28:41 Discord bot that makes images why cuz
28:44 mid Journey makes the best quality
28:46 images of all the image generation tools
28:48 period hands down unless you want text
28:51 in it and then it's
28:52 dolly3 um Leonardo is sort of the web
28:56 front end for a whole
28:58 um ecosystem of things on stable
29:00 diffusion um that that you if you want
29:03 to geek out Leonardo is a good one to
29:05 play with and then Firefly from Adobe is
29:08 good if you work at a company at a
29:10 corporation because it has not been
29:11 trained on any
29:13 copywritten data
29:15 images um so so that one's
29:18 interesting um here's video creation
29:21 this world is getting insane right now
29:23 so one thing I need to add to this in
29:25 fact let me do it now
29:29 I'm going to grab this thing
29:33 here you can't just change a slide right
29:36 in the middle of a presentation that's
29:38 against the rules one of the things
29:40 you'll learn about this channel is me
29:43 and rules not so good
29:45 along it's it's we like to call this
29:48 channel chat
29:49 add all right so we're going to put down
29:53 here
29:56 Runway Dot
30:01 TV
30:03 um so Runway ml is a textto video
30:07 creation tool as as is Moon Valley and
30:10 pabs and Runway just launched runway. TV
30:13 which is a think of it like Netflix but
30:16 for AI generated
30:18 content you're going to go there and if
30:20 you're a real filmmaker most of what
30:23 you're going to say is this stuff looks
30:24 like [ __ ] this is
30:26 dumb
30:30 um and what's going to happen is real
30:32 filmmakers and real Studios and real
30:35 distribution networks like Netflix are
30:37 going to laugh at that thing they're
30:38 going to laugh at Runway TV until they
30:42 don't because these tools are getting
30:45 better fast um ha Jen and did are are
30:49 like Avatar creation tools ha Jen in
30:52 particular their Avatar right now you
30:55 can upload a video of yourself like a
30:57 two 2 minute video and it
31:00 will synthesize you and then you can
31:03 just type a script and it will say
31:07 it and you're like but C what will it
31:10 say it'll say whatever the hell you type
31:12 and it'll sound sort of like you like
31:13 enough that it's it weirds people out
31:16 well what does that look like Kyle KY
31:18 Kyle can you show us KY hey KY Kyle c c
31:26 c
31:31 D AI
31:34 learnning lab and you can go see
31:38 this without me the site can't be
31:40 reached
31:42 [Music]
31:44 because I called it tick
31:49 tick all right you can't call Tick talk
31:52 tick tick it does not work see you
31:55 learned that
31:56 tonight
31:58 there we
32:00 go all right let's
32:03 see all right watch this hey it's Kyle
32:05 Shannon with the AI learning lab you
32:07 know there are some days I just don't
32:09 want to turn on the camera days where I
32:11 don't even want to get out of bed well
32:13 good news we now live in a post sci-fi
32:16 world and I can just type up this script
32:18 and select my avatar of choice this is
32:21 one from a video I recorded about a year
32:23 ago you might be so so it kept my face
32:27 and like I was wearing and my facial
32:29 expressions but it synthesized my voice
32:31 and so that's that one and then so that
32:34 one's just I typed in text and it just
32:36 synthesized it and then this next one is
32:38 I recorded the audio and just uploaded
32:41 the audio and so that's this one hey
32:43 it's Kyle Shannon with the AI learning
32:45 lab so the previous video on my timeline
32:49 is an avatar generated with hey Jen
32:52 where I didn't say the words I just
32:54 typed in a script and it synthesized my
32:56 voice and and like the fact that it's
32:58 matching the hand movements to the
33:00 Cadence of the video and how I'm talking
33:03 is insanity to me now does it look
33:06 exactly like me it's got some weirdness
33:08 to it but it's pretty [ __ ] close it
33:11 was decent it didn't really sound like
33:13 me it didn't really sound like me be
33:15 recording myself I'm just recording
33:17 audio and I'm uploading it to hey Jen as
33:19 well and it's matching this audio to the
33:22 Avatar's looking Fe so that's hey
33:25 Jen um
33:29 and then Visa is like for making like
33:31 corporate videos basically but but does
33:34 it really quick and and in a kind of
33:36 interesting way okay and then what else
33:38 do we got here here's music and audio 11
33:41 Labs uh does the voice synthesis so the
33:44 haen stuff I just showed the voice
33:46 synthesis engine that they use is 11
33:49 Labs um sunno and RI Fusion are both
33:54 they'll sing You songs they'll write You
33:55 lyrics and sing You songs and write the
33:57 songs and play the
33:59 instruments it's [ __ ] Insanity um
34:03 both of them sound janky right now if
34:04 you're if you're an actual musician
34:06 you'll kind of laugh at it uh but if
34:09 you're an actual musician you'll also
34:11 realize holy [ __ ] this it's not that far
34:14 off uh and then sound draw is like a
34:17 loop composer that uses AI to to mix
34:20 Loops together really really quickly so
34:22 if you if you need to knock out some
34:24 quick background music for a video sound
34:26 draw is really
34:28 good um here's 3D worlds this is if if
34:32 you're in game development these are
34:34 things worth playing with um although
34:37 Wonder Dynamics will replace a human
34:39 being in a video with a robot uh that
34:41 you choose a robot from their library
34:44 and then give it a couple of hours to
34:45 process and it basically rotos Scopes
34:48 out the human and rotos Scopes in a
34:50 robot it's pretty amazing and then these
34:53 people are all people that you should
34:56 follow um Ethan mullik is a Wharton
34:59 business school Professor one useful
35:01 thing is his blog I'd follow him on
35:04 Twitter and on LinkedIn as
35:08 well and um David Shapiro is YouTube
35:14 YouTuber um that really knows about this
35:16 AGI stuff he's he's the one that has
35:19 largely explained all this qar stuff to
35:22 me and what it's not only how it works
35:25 but what its significance is is he
35:28 understands it he's written five books
35:29 on AGI he's got a cognitive architecture
35:32 open source project so if you're a geek
35:35 like if you're a geek and you're not
35:37 following David Shapiro and you're not
35:39 involved in his open source project
35:41 right now I'm like what are you doing
35:43 with your life
35:44 man get get on it get on it I A lot of
35:49 times people come in here and they ask
35:50 oh my son just started computer science
35:52 is that a good
35:53 idea I I think if you're in any way
35:57 technical um you should be diving into
36:01 open- Source projects open source large
36:03 language models um you should be diving
36:06 into projects like Lang chain like Auto
36:09 GPT um like uh like David Shapiro's
36:14 cognitive architectures David Shapiro is
36:16 also doing agent swarms in an open-
36:19 Source way I would just dive into that
36:22 [ __ ] if I were you whether whether you
36:25 know they're teaching you that in
36:27 college or not because they're not going
36:29 to be teaching you it in college unless
36:31 you have some you know
36:33 remarkable um professors if they are
36:36 [ __ ] stay there learn everything you
36:37 can from those people hang out with them
36:40 if they're not teaching you that [ __ ]
36:42 Jump On It Go do it on your own be get
36:44 entrepreneurial we're all going to have
36:46 to uh Paul rater and Rachel Woods are
36:49 both good if you're in business Rachel
36:51 um is a good counter to this channel so
36:54 my channel is kind of Loosey Goosey chat
36:56 add infotainment kind
36:59 of Ed
37:01 edutainment uh content um hers is much
37:04 more buttoned up and if you want to get
37:05 [ __ ] done at work follow her and then
37:07 Paul rer from the marketing AI
37:09 institute's really good Be not Afraid
37:10 thank you so much I appreciate that
37:13 um yeah so that's that um if you want to
37:15 support the channel you can subscribe to
37:17 it um you'll get access to the to the
37:20 Irregulars channel on Mighty networks on
37:22 the AI Salon which I'll talk about um
37:25 you can also um buy Series in the lower
37:28 leftand corner there's a series there
37:30 called underhyped which is a series of
37:32 videos that I recorded over the past
37:34 year um that that are are really my take
37:37 on why I think AI is such a big deal how
37:40 it actually works how you use it how you
37:42 prompt it how not to get overwhelmed by
37:44 it how to have an impact so that's down
37:46 there you can support the channel that
37:48 way and then and then also um if you
37:51 subscribe to the channel or or just
37:53 follow follow the channel that would
37:55 help if you share this live so other
37:57 people can discover this that would help
38:00 um what else and then we've also got a
38:02 YouTube channel so I put the recordings
38:05 of these lives I go live every night for
38:08 two to three hours seven nights a week
38:11 although I did take Thanksgiving off
38:12 because I know how to
38:16 live and you're like but Kyle that's
38:18 insane why would you do that I don't
38:20 know but it feels important to me so I
38:22 do so so we go live every night I've
38:25 been doing it since April and um so if
38:28 you go to YouTube AI learninglab dtt is
38:32 the YouTube channel And subscribe to
38:33 YouTube AI learninglab dtt thank you
38:36 sherbear I appreciate that um thank you
38:40 digital Gods you're a title gifter I
38:42 don't know what that means but that
38:44 sounds important I that that must mean
38:46 you give me lots of roses so I
38:48 appreciate that you're a title gifter
38:51 you've earned a
38:53 title do I have to call you sir digital
38:55 Gods at this point point I think sir is
38:58 below God so I don't know what to do
39:01 here thank you
39:02 Kevin um all right and then the last
39:05 thing that I'm going to talk about and
39:06 then we can move on to Q&A and demoing
39:08 [ __ ] and whatever you want to do is is I
39:11 want to talk to you about the AI salon
39:12 so the AI Salon is a community it's an
39:16 online community we meet every other
39:18 week we meet in person here in Denver
39:21 and we meet online um December 5th the
39:25 December 5th meeting is our one year
39:27 anniversary we launched the salon the
39:29 week after chat GPT launched when chat
39:31 GPT launched I lost my [ __ ] mind I
39:34 started I started a newsletter I started
39:37 a Tik Tok Channel I I I just I lost my
39:41 mind I started the salon um and and
39:46 probably I think I said yes to I don't
39:49 know five or six different projects
39:51 within within a month um because I've
39:54 been here before I saw this in the early
39:56 days days of the worldwide web thank you
39:58 Sherry D um and thank you whoever just
40:01 sent the uh the celebration thing it
40:02 might have been you drew um but thank
40:04 you for that
40:06 um but but when I saw chat GPT I kind of
40:09 had a similar Epiphany that I had in '94
40:12 with the worldwide web where I just
40:14 realized holy [ __ ] this is one of those
40:16 and so this community the AI Salon we
40:18 have a new logo that's our new logo um
40:21 and we have a new um Community Home
40:24 which is mighty Network so if you go to
40:26 AI salon
40:27 mn. Co um that's the that's the home and
40:32 then the you know the description of who
40:34 we are is where a ragtag group of
40:37 Misfits Nero Wells and miscreant who've
40:39 gathered to explore the Uncharted
40:41 Wilderness of generative AI where
40:43 everyone from entrepreneurs to Educators
40:45 from healthcare workers to creatives and
40:46 everyone in between
40:49 um does that sound nerdy that's okay we
40:52 also one of the things I do want you to
40:54 check out if you decide you want to join
40:55 the salon that's cool
41:02 um hang on let's see AI I changed
41:06 browsers
41:15 so I want you to check
41:25 out
41:30 except all the cookies just accept all
41:32 the cookies so the the way Mighty
41:34 networks works it's kind of like there's
41:36 a feed just kind of like Facebook like
41:38 here's the latest stuff people have have
41:40 uh contributed and then there's a lot of
41:43 different channels but the first channel
41:44 under Salon info is about the salon and
41:47 that's got that image that I just showed
41:49 you and it talks about who we are why
41:51 we're here and then it's got our values
41:54 and the values are really important to
41:55 me so I want this community to grow but
41:58 make sure that you resonate with who we
41:59 are like I want Curious people in here I
42:02 don't want cynics like you can be
42:03 cynical if you're funny but I want you
42:05 to be curious at heart I want you to be
42:08 generous so so as you learn stuff share
42:12 stuff share what you're learning um I
42:14 want you to be adventurous exploration
42:16 is one of the values collaboration is
42:18 one of the values like team up with
42:19 other people in the community inclusion
42:22 there's going to be people coming from
42:23 all walks of life everyone is new to
42:27 this everyone there are no experts right
42:29 now everyone's trying to figure it out
42:30 thank you so much Earth rocker um so
42:33 there's going to be people from weird
42:35 Industries and there's you you might
42:38 have 30 years in Ai and you're going to
42:40 be getting insights from someone who's
42:43 been in it for 30 days and like just
42:45 know that that's how it's going to be
42:47 and embrace that and then the last one
42:48 is empathy um this shit's intense like
42:53 as as fun and exciting as this is and as
42:55 enthusiastic as I am about AI I also get
42:59 that this is profoundly powerful stuff
43:02 that's going to change a lot for a lot
43:04 of us we're all going to be confronted
43:06 with um a lot of change and so there's a
43:09 lot of fear there's a lot of anger
43:10 there's a lot of resistance people are
43:13 going to think you're [ __ ] weirdos a
43:15 good way to think about the salon is
43:16 like an AI support group like hi I'm
43:19 Kyle I'm in AI hi Kyle it's been 10
43:23 minutes since I made something in Dolly
43:25 oh yeah I'm making one right now been
43:30 there you know that's the group and
43:33 we're choosing to be on this adventure
43:35 you know and that's that's what that's
43:37 about all right enough comedy
43:42 jokes I'm also a comedy fan he I
43:47 improv uh he looked at me wanted my
43:50 table sorry Ryan I don't know what
43:52 that's about
43:53 tonius sha that is my favorite of of all
43:56 of the things the dangly lights I do
43:58 love the dangly
44:00 lights they make me happy it just feels
44:03 like I'm like out on the back porch
44:05 there's a little River Running
44:07 by uh yes but Rhino is far away yeah we
44:11 we uh it is in Rhino Joe mama um the the
44:15 the salon meetings but we also do them
44:17 online so you can join virtually you do
44:19 not need to be here in Denver but if
44:21 you're here in Denver we do an
44:23 Afterparty we go we go to this place
44:25 called Zeppelin station we hang out and
44:27 you know what we talk
44:28 about AI
44:33 [ __ ] it's pretty
44:37 nerdy all right we got some people
44:39 wanting to join let me just go let
44:41 people in as they're
44:47 joining I don't know if Vicky's on
44:50 tonight but I'll get it all
44:54 right fre fantastic what were you doing
44:58 before AI um so I'm I'm an entrepreneur
45:01 so I I'm the co-founder and CEO of a
45:04 company called storyvine so um it's a an
45:08 automated video storytelling platform
45:10 it's an 11 and a half year old company
45:12 so I've been doing that for a while
45:14 that's still going strong um we're
45:17 looking to to sell that but but I've got
45:20 a whole AI road map that that right now
45:22 our clients are really excited about so
45:24 I'm building AI into the the platform uh
45:27 and then we're probably going to sell it
45:28 hopefully in
45:30 2024 and then
45:32 um and then prior to that just i' I've
45:35 just been an entrepreneur I've started
45:36 12 or 15 companies I I started
45:38 agency.com which was one of the early
45:41 digital agencies in the mid
45:43 90s and so I've been through one of
45:45 these Cycles before um if we're just
45:48 irregular how do we prove that you just
45:50 show up here a lot you well so you can
45:53 subscribe to the channel jenu uh um you
45:57 know speak up volunteer if you want to
45:59 mod this channel you can you know let me
46:02 let me know and we can make you a mod on
46:03 the channel the the
46:06 Irregulars the Irregulars was not just a
46:08 formal thing was not a formal thing it
46:10 was just what what ended up happening
46:13 was I started showing up here nightly
46:15 cuz I thought hey you know people are
46:17 going to want to know about this AI [ __ ]
46:19 and what ended up happening is that a
46:22 lot of other people decided to hang out
46:24 here too and we're learning to together
46:27 and we're exploring together and what's
46:29 really cool Vicky I'm here okay great
46:31 Vicky good to see you um what what's
46:34 happening
46:35 is it's forming this really tight
46:38 community of people who are excited
46:43 about life again are excited
46:45 professionally again some of them have
46:47 retired and they're like [ __ ] it this is
46:48 fun let me go do this um they're
46:55 Reinventing
46:56 what their skill set is in ways that
46:59 they couldn't have imagined right
47:01 there's a guy um Roger in in the AI
47:04 Salon is like a retired software
47:06 engineer he's super software engineering
47:09 and he's discovered his approach to
47:14 coding is really good for prompt
47:17 engineering for Mid journey and he's
47:19 gotten really good at Mid Journey art
47:21 and he actually just got accepted into a
47:24 boulder AI art showing um like two weeks
47:28 ago and and it's like just watching
47:30 people reinvent thems is is um super
47:35 inspiring so so how you becom an
47:38 irregular is you just regularly show up
47:41 and then you you be a little [ __ ]
47:43 weird you be neuroatypical you get
47:46 curious you start helping others you
47:49 just do all the cool stuff and and
47:52 you're in the club like it's not an
47:54 official thing although if you subscribe
47:56 to the lives then then um I'll give you
47:59 access to the to the
48:01 channel um ethical questions do come up
48:04 yes but embrace it I feel the
48:06 restrictions will come um yeah well I
48:10 don't know judging
48:13 how radically incompetent our government
48:16 is right now um you know there there
48:19 might be there might be regulations to
48:23 to try to slow it down but I it's it's
48:26 not going to slow
48:29 down I think if you look at President
48:31 Biden's executive order in fact there's
48:35 a
48:37 um I can share this you know what I'm
48:39 going to
48:40 do I created
48:44 a
48:48 um I created
48:51 this which is the AI executive order
48:54 Explorer so this is a GPT where you can
48:57 actually ask about the the executive
49:00 order and interact with it and like have
49:02 it explain it to you so I'm going to
49:04 actually go share this in the AI
49:08 Salon in the show Andel channel so under
49:13 Community there's a show and tell
49:16 Channel and so I'm just going to paste
49:19 that in
49:22 there and something weird is set up with
49:25 Mighty networks where it's doesn't show
49:26 the preview that's okay we'll fix that
49:29 at some point we're brand new to Mighty
49:33 networks and I'm going to add the hasht
49:36 gpts then I'm going to post it
49:39 boom uh I don't need to notify everyone
49:42 about the post but there is I should
49:46 probably can I edit
49:55 this
49:57 oh whatever it's there if you want to go
49:59 play with it go play with it
50:02 um I think that's the most we're going
50:04 to get for a while on the uh on the uh
50:09 regulatory side because Congress is
50:11 Congress is a [ __ ] show right now
50:12 they're not going to do
50:15 anything
50:19 unless what could change the regulatory
50:22 thing is if if open AI indeed has hit
50:27 AGI it's it's very possible that some
50:30 [ __ ]
50:31 accelerates so so we'll
50:34 see it's going to be a wild 2024 I can
50:37 tell you
50:39 that we need a regulars t-shirts yeah I
50:42 know I'm gonna I'm GNA um I'm gonna get
50:45 it together I'm gonna get it together
50:47 Kevin I gotta I just got to get it
50:49 together I got to get it
50:52 going let's do a battle tonight no
50:54 battles this is a battle Battle free
50:57 zone this is we we are no longer
50:59 battling Dar uh Dar when when she uh
51:04 obliterated me on her comeback uh I I
51:07 realized I am just not man enough uh to
51:10 battle Dar and it just felt like the
51:14 sleaziest weirdest I don't know it's
51:17 like engagement farming in the worst
51:20 possible way it just does not feel like
51:22 this channel sometimes stuff can't be
51:24 slowed down yeah exactly like that well
51:27 that's that's a lot of the reason why
51:29 this channel exists is
51:32 because you can judge this stuff all you
51:35 want but it's still coming every time
51:39 someone falsely claims AGI a puppy
51:41 somewhere gets punted I didn't claim it
51:43 tonight I said if it happens if it
51:45 happens Pate I am not claiming it but I
51:48 but I will absolutely admit um I think
51:50 pecky 2623 wants to battle no battles no
51:54 battles no battles pecky
51:58 no I'm not battling I'm not doing it I I
52:02 it just feels I I felt
52:05 slimed I'll help you okay cool Kevin
52:08 that's great uh DM me let's let's let's
52:11 figure some [ __ ] out because we're I'm
52:12 also Tobias is your T-shirt dude he's
52:15 very creative yeah I know so I've got
52:18 Tobias on the list I've also got Bradley
52:20 who helped me set up mighty networks for
52:21 the salon is working on Salon merch so I
52:24 think I'll just piggy back on whatever
52:26 vendors he selects I if if you've been
52:29 here for a while you know that I did
52:31 start a t-shirt store inside Tik Tock
52:35 and then Tik Tock [ __ ] it
52:39 up and so it it was an absolute
52:42 nightmare I spent a month of my life
52:44 trying to get that thing untangled and
52:46 and I so I just shut it down and so so
52:49 I'm a little stingy I got I got
52:54 slapped
52:58 um every time someone falsely claims AGI
53:02 a puffy gets punted champy champy Pate
53:05 wants to kick you don't let Pate kick
53:08 you it's not
53:11 nice fairness and
53:14 non-discrimination oh man classic
53:17 government sure Dar will overhear us I
53:20 know Dar is the star of Tik Tok as as we
53:23 learned still traumatized me too Sherry
53:26 D every time I think about that I kind
53:28 of flush with
53:31 embarrassment not that I lost that I did
53:34 it in the first
53:37 place and and then she just wouldn't go
53:40 away it was so bad I was just like I
53:43 didn't know how to turn her off I'm like
53:45 I don't know what to do here oh it was
53:47 it was ugly it was not
53:50 good look I'm I'm for CLP
53:54 now
54:00 oh man alive man alive man alive all
54:04 right everybody all right all right all
54:06 right fantastic Bob tell him what he's
54:09 won what do we want to do Dar won't be
54:14 back
54:15 ah punting pump Pieces by gemin on
54:21 hold I want a t-shirt with d we actually
54:23 have some artart in the uh in the
54:26 Discord channels there's there's some
54:28 stuff in
54:30 there with DAR on
54:33 it oh good some people are introducing
54:36 themselves in the salon that's awesome
54:38 yeah please introduce yourself take that
54:40 value of generosity and and tell us who
54:42 you are and what you're what you're up
54:43 to and what are you doing in this AI
54:45 stuff the goat Eric former am I late
54:49 everyone's late we're all late we're all
54:54 early
54:56 [Music]
55:02 this channel is about the love it is
55:04 this channel is
55:06 about us all trying to figure it out and
55:10 recognizing that we
55:12 can't and as much as we want there to be
55:15 an expert there aren't
55:18 any will there be in 5 years probably
55:22 but at that point the AI is likely going
55:25 to be the expert what are we going to do
55:28 I don't
55:29 know I think we might be able to do what
55:33 we want to do more than do what we have
55:35 to do I think that's a possibility my
55:38 name is Pedro I'm 47 new to AI but I'm
55:40 trying to learn it and start a company
55:42 awesome Pedro welcome keep hanging out
55:46 here um the salon where Discord uh so
55:49 the
55:53 salon is here
56:00 there you go AIS salon. mn. Co It's
56:05 mighty networks so we're transitioning
56:07 off of Discord to Mighty networks for
56:09 many many reasons um and uh go there and
56:14 do me a favor read the when you get to
56:17 the salon go to
56:21 the what's it called oh about the salon
56:26 so one of the first links is about the
56:28 salon and it tells you about um who we
56:32 are why we're doing what we're doing and
56:34 what our values are so go check out the
56:36 values and make sure they resonate with
56:40 you if not that's totally cool but this
56:43 is probably not the group for you like
56:45 if you want a bunch of cynical people
56:47 that are trying to figure out all the
56:49 ways that AI is going to kill us it's
56:51 probably not this group if you're
56:53 curious and want to figure out how do
56:55 you deal with this stuff and how do you
56:57 find people to do interesting projects
57:00 with and how do you learn and how do you
57:01 explore then this is the group all right
57:04 that's really important because listen
57:08 if those values don't resonate with you
57:10 you'll be miserable too you just won't
57:13 fit
57:14 um I'm trying to get chat or Claude to
57:17 summarize a couple of books for
57:19 a Blog article is this kosher yes it
57:24 is
57:26 um and you can do it with either of them
57:28 now
57:32 so
57:36 um it's one of the it's one of the
57:38 things it's really good at and depending
57:40 on how old the book is it might know it
57:43 anyway
57:47 um so if you're using it to summarize it
57:50 that's kosher if you're using it to take
57:51 the book and just rewrite it and then
57:53 resell it on your own that's not kosher
57:55 like you know I would say use common
57:57 sense when it comes to the copyright and
58:00 trademark
58:01 stuff but yeah you can you can use
58:04 it it's perfect for
58:08 that chat gbt now a now has 125 token
58:13 125,000 token context window which is
58:16 about 96,000 words and
58:19 Claude um just up their token window to
58:24 200 ,000 tokens which is about 150,000
58:28 words and to put that in context The
58:30 Great Gatsby is 50,000 words so so yeah
58:35 upload a PDF of a book and just say
58:38 summarize this give me bullet points who
58:40 are the main characters I've uploaded a
58:42 bunch of my screenplays and said you
58:44 know give me character breakdowns give
58:45 me summaries give me Hollywood style log
58:48 lines and it does them like that and it
58:50 does them better than frankly I would
58:51 have done them especially with [ __ ] like
58:54 that where if you got like old work that
58:56 you've got you've got an old book that
58:57 you wrote but you never did anything
58:59 with it throw it in chat GPT and have it
59:01 do an editing pass on
59:03 it Chris Ramsey I made a qar shirt and
59:07 it's posted yeah I saw that where did
59:08 did you put that in Show and Tell Chris
59:10 where did you put it or in water cool
59:14 cooler time is an illusion like love
59:18 thank you Joker this this is also a
59:21 philosophy a philosophy class oh Ted
59:24 that's beautiful here's some of the
59:26 here's some of the images people are
59:27 showing off in the in the mighty Network
59:30 show and tell
59:31 Channel be not afraid that one's [ __ ]
59:34 stunning good
59:37 Lord somewhere between paper cut art and
59:40 illustration that's
59:42 beautiful Claire Jacobs killing it with
59:46 mid journey and dalali is that mid
59:49 Journey yeah that's mid Journey look at
59:52 see you notice mid journey is just kind
59:54 of of a different quality level it's
59:57 just of a different
59:59 level um hang on I got some new people
1:00:02 joining the salon the
1:00:04 salon why do you call it a salon do you
1:00:08 get your hair done there is this place
1:00:10 where you get your nails done is it that
1:00:12 kind of
1:00:13 Salon no ask chat GPT about salons of
1:00:17 the 1800s if you don't know what a salon
1:00:20 is it's basically just we get together
1:00:23 and talk about [ __ ]
1:00:26 wow Be not Afraid another killer one
1:00:30 that's
1:00:31 beautiful where's the qar
1:00:34 shirt um Chris
1:00:37 Ramsay I know I saw it
1:00:40 today did you put it oh maybe you put it
1:00:43 in where'd you put
1:00:45 it water
1:00:47 cooler oh I know where I can find it I
1:00:49 can find it at the feed at the feed feed
1:00:52 fade feed feed feed feed fade
1:01:17 oh cool Vicki is actually indexing her
1:01:21 own gpts which is a super cool idea
1:01:24 actually
1:01:36 water cooler all right I had a feeling
1:01:39 it was in water
1:01:41 cooler water
1:01:43 cooler new
1:01:53 posts
1:02:07 I can't find it
1:02:11 Chris I'm an old
1:02:14 guy all right uh just me and the trees
1:02:17 I'm using it for research right now it's
1:02:19 a total Game Changer yeah it's crazy for
1:02:21 research just me and the trees use uh
1:02:23 check out perplexity uh do is it.com I
1:02:27 think it's check out perplexity if
1:02:30 you're doing research and then also Bard
1:02:32 is pretty good at uh
1:02:35 citations um co-pilot does them but I
1:02:39 don't like I don't like how co-pilot's
1:02:41 doing what it's doing perplexity
1:02:48 doai
1:02:50 uh starting a business to help relieve
1:02:53 caregiver stress at the end end of life
1:02:55 care it's funny thankful Jones
1:02:59 um there was someone I was talking
1:03:02 to three months ago at a at a like a CEO
1:03:06 dinner and she's got a business to do
1:03:08 just that so I think that's it's it's a
1:03:11 really interesting business you should
1:03:13 check out
1:03:14 Pi if you haven't check out
1:03:18 Pi so this is my iPad so so this is just
1:03:22 a regular pie chat so that's what it
1:03:25 looks like when you first log in this
1:03:27 little button right here puts it into
1:03:30 conversation mode which I'll show you in
1:03:31 a second but there's also different
1:03:33 models here so notice in the wait why is
1:03:37 that so blown out hang
1:03:39 on see where it says Pi inside the
1:03:42 little chat box there at the bottom I'm
1:03:45 going to click on
1:03:49 that and there's two different models
1:03:51 there there's PI what's it called
1:03:55 just
1:03:56 pi and then there's support pi and so if
1:03:59 you click on support
1:04:02 Pi it changes the little icon to a heart
1:04:07 it's basically been programmed to be a
1:04:09 therapist or to be a counselor right to
1:04:12 be someone who is compassionate about
1:04:15 things so you can put it into compassion
1:04:18 mode is how I'm thinking about it and
1:04:20 then put it into
1:04:23 conversation
1:04:26 hey I'm really struggling over the
1:04:28 holidays I I don't feel like going back
1:04:30 to work on Monday and it's only Saturday
1:04:33 what can I
1:04:42 do of course she ignores
1:04:47 me first of all I want to let you know
1:04:50 it's okay to feel this way and you're
1:04:52 not alone in experiencing these feelings
1:04:55 the holidays can be a difficult time for
1:04:57 many people so I want to assure you it's
1:04:59 normal to feel this way can you tell me
1:05:01 more so that's really cool I mean that's
1:05:04 that's a thing to explore the other
1:05:06 thing to explore
1:05:07 is um the chat GPT
1:05:11 app um this is the chat GPT app also can
1:05:15 do um gpts so you can you you can't
1:05:19 create them in the in the app but you
1:05:22 can use
1:05:23 them
1:05:25 so I've got one here that I did called
1:05:27 Dream Weaver which is a a dream analysis
1:05:30 kind of thing so you can also put these
1:05:33 into conversation
1:05:38 mode wait I think I have to type
1:05:40 something first
1:05:44 hi all
1:05:47 right so I click on the little
1:05:53 headphones
1:05:58 I had a funny dream last night that I
1:06:00 was riding a llama and we tripped over a
1:06:03 turtle and fell on a
1:06:09 hedgehog that's an intriguing dream
1:06:11 let's start by exploring it a bit more
1:06:14 when you were riding the Llama what was
1:06:16 the setting like around you for example
1:06:19 were you in a familiar place somewhere
1:06:21 Outdoors or somewhere completely
1:06:23 imaginary
1:06:25 it was exotic like a valley in the
1:06:27 mountains of
1:06:32 Peru that sounds like a beautiful
1:06:34 setting now about the Llama how did it
1:06:37 look or behave in your dream was it calm
1:06:39 playful or something
1:06:42 else it was like in a story
1:06:47 book that's interesting so it had a
1:06:51 story book- like almost Whimsical
1:06:53 quality could you describe the turtle
1:06:55 you tripped over was it large small or
1:06:58 did it have any distinct
1:07:00 features it was small and
1:07:06 cute got it a small cute turtle and what
1:07:09 about the Hedgehog you fell onto how did
1:07:12 you perceive it in your dream for
1:07:14 instance was it a surprise or did it
1:07:16 seem to be a part of the dream's
1:07:19 logic he was not too happy to be landed
1:07:23 on
1:07:24 [Laughter]
1:07:28 understandable a hedgehog so anyway so
1:07:30 that goes on and what what dream anal
1:07:33 it's called Dream Weaver that's my GPT
1:07:35 called Dream Weaver that one will then
1:07:38 at some point after it ask me how I feel
1:07:40 about the dream it'll visualize the
1:07:41 dream so it actually create that image
1:07:43 and then it'll give me yian analysis of
1:07:46 the
1:07:46 dream so if you're doing this thing you
1:07:49 could very very quickly um prototype up
1:07:53 some different kind of interactions that
1:07:55 could be really interesting Pi tells me
1:07:58 I'm afraid voice search is not supported
1:08:01 she doesn't want to
1:08:03 go into
1:08:05 conversation voice
1:08:08 search so in
1:08:11 pi you have to be using the app not the
1:08:15 website for
1:08:17 one and then it's it's this little phone
1:08:21 icon right here not this microphone icon
1:08:23 the little phonec icon so you click that
1:08:26 and it turns
1:08:31 green can you please tell me if there's
1:08:34 any live music on going tonight in
1:08:40 [Music]
1:08:47 Denver absolutely I can help you find
1:08:49 out about live music events in Denver
1:08:51 can I ask what kind of music you're
1:08:53 interested in and what kind of venue
1:08:54 you're looking for are you looking for a
1:08:56 small intimate venue or something bigger
1:08:59 and more high energy so anyway that's
1:09:01 that so it's it's probably just a you're
1:09:04 probably using the wrong version or
1:09:05 something Pi just so the the uh print
1:09:09 hello world just said Pi changed its
1:09:11 voices um Pi just updated their model um
1:09:16 three days ago so when they updated the
1:09:19 model it probably reset all your
1:09:20 settings that's that's probably what's
1:09:22 going on there
1:09:27 let's get a different image up
1:09:29 here we need some
1:09:32 variation we need
1:09:39 some hey
1:09:41 hey what's shaking I got some AI for you
1:09:45 you want to buy some
1:09:51 AI I asked perplexity how do we use this
1:09:54 oh by the way Nancy I saw your what you
1:09:56 posted earlier perplexity can do
1:09:58 300,000 word uploads awesome that's
1:10:03 great I asked perplexity how do I use
1:10:05 this it showed me how to use the word
1:10:12 this that's
1:10:15 great that's really funny okay so
1:10:18 perplexity is blowing my mind okay
1:10:21 awesome awesome just me and the trees I
1:10:23 thought so like
1:10:24 that's everyone that I've talked to that
1:10:26 does any sort of research they're like
1:10:28 oh my God perplexity has changed my life
1:10:30 I can't get it to work because I don't
1:10:32 know how to research apparently so I
1:10:35 haven't figured that out but I'm glad it
1:10:37 is blowing your mind so so one of the
1:10:40 things we do here is is it is my
1:10:44 contention that it's in everyone's every
1:10:47 human being's best interest to use these
1:10:50 AI tools long enough that you have your
1:10:54 Kevin McAllister moment which is this
1:10:58 moment the Home Alone
1:11:00 moment and you will like at some point
1:11:03 you'll be like ah it's just like Google
1:11:05 it's just like a Google search you know
1:11:07 I don't know if you knew this but you
1:11:08 just put things in a search box and you
1:11:10 hit the thing and it returns an answer
1:11:11 like big whoop I've been doing that for
1:11:13 20
1:11:14 years and then at some point you're like
1:11:17 oh maybe I'll do something a little but
1:11:18 it can't do what I do cuz I'm special
1:11:20 and then you type it
1:11:22 in make me project plan for putting
1:11:25 together plants
1:11:26 and oh my
1:11:29 God it did what I do it did it better it
1:11:32 did it really fast is this what it does
1:11:35 oh my
1:11:37 God you want to have that moment if you
1:11:41 haven't had that moment yet you're not
1:11:42 doing it
1:11:44 right is there an app for perplexity I
1:11:47 think there is I think there
1:11:50 is let's go to the App
1:11:52 Store oh
1:11:54 perplexity hello are you
1:11:58 perplexity let's see per plexity
1:12:05 AI perplexity ask
1:12:08 anything is
1:12:11 this
1:12:14 from inflection it's
1:12:22 not
1:12:30 um I think the perplexity app is a faker
1:12:34 app but I could be
1:12:43 wrong drop by our
1:12:47 Discord Let's go ask
1:12:50 Pi Pi might
1:12:52 know
1:12:57 [Music]
1:13:02 I'm trying to learn something there's a
1:13:04 site called perplexity
1:13:09 doai oh
1:13:12 wait oh no I was thinking of Pi no that
1:13:15 could be
1:13:16 it that could the the perplexity app
1:13:18 could be it I I was thinking perplexity
1:13:21 was from inflection Pi is from
1:13:22 inflection never mind ignore
1:13:24 me just ask her pie you have to be on a
1:13:28 tablet or a phone not your
1:13:32 desktop yeah exactly that's what I was
1:13:34 saying it's the for for pi the the
1:13:36 conversation mode is in
1:13:41 the IOS
1:13:44 app it does look like there's a
1:13:47 perplexity app
1:13:49 yes sorry for the confusion I got
1:13:52 confused I got discombobulate this is
1:13:54 what happens see when when you give old
1:13:57 people technology and a platform just
1:14:00 this this this is the kind of [ __ ] that
1:14:02 happens all right so just you're going
1:14:03 to have to deal with it it's
1:14:05 like this channel is jankier than all of
1:14:09 the AI tools put together okay so just
1:14:12 accept that that's we have a singing
1:14:15 dog you got you got an old guy with the
1:14:18 with figured out how to use dual camera
1:14:20 on Tik Tok and that's pretty much it so
1:14:23 you get all right mods at least pin that
1:14:26 Elon found the research paper for qar
1:14:30 Hodge remains Scher classes for qar hjge
1:14:35 oh hjj Raymond sh
1:14:40 classes he found it and what does that
1:14:43 mean he found it
1:14:46 and elon's been a little trolly lately
1:14:49 of the open AI thing so I he's about to
1:14:53 launch his own competitor to chat gbt so
1:14:56 I'm taking anything elon's saying right
1:14:58 now with a little bit of grain of salt
1:15:00 cuz uh yeah
1:15:02 now this is the guy that called for a
1:15:04 six-month pause on AI and then a week
1:15:07 later ordered 10,000
1:15:10 gpus
1:15:13 so I'm not saying that there's a oh wait
1:15:18 x.com I'm not saying there's ulterior
1:15:21 motives here but let's go
1:15:29 [Music]
1:15:34 look Elon Musk hello my name is Elon
1:15:38 Musk I make big rockets that go boom
1:15:41 boom in his
1:15:42 [Music]
1:15:52 face
1:16:04 dun and
1:16:17 Krueger yeah I just I just see him
1:16:19 talking about how open AI not being
1:16:23 being biased is is a major problem so
1:16:27 he's he's been he's been [ __ ] all
1:16:29 over open Ai and chat GPT for the for
1:16:32 the past month or so but like a lot in
1:16:34 the past
1:16:36 week all right let's say
1:16:44 um d a DW Kyle is legit I think it's
1:16:50 only released a few months ago but it's
1:16:51 pretty well put together
1:16:55 a DW Kyle it's legit can I call for a
1:17:00 six-month pause on mentioning
1:17:07 Elon
1:17:08 absolutely
1:17:11 absolutely he's GNA he's going to be
1:17:13 talked about you know in this channel a
1:17:17 lot in the next
1:17:19 uh couple of weeks when when um grock
1:17:22 comes out
1:17:24 Soul
1:17:31 good Kevin I don't know what you're
1:17:33 talking about aha DW Kyle it's
1:17:38 legit I think it's only released a few
1:17:40 months ago but it's pretty well put
1:17:42 together oh that's the oh Elon found the
1:17:44 paper on qar there's an actual paper on
1:17:47 it ah got it okay yeah I'll go check
1:17:49 that out I I would be surprised let's
1:17:51 see if David Shapiro's already put a
1:17:53 video out on it that's if you don't
1:17:56 follow David Shapiro the man's a
1:17:57 flipping machine he's also neuros spicy
1:18:01 which which you know he gets he gets
1:18:03 points in my book for for being a
1:18:08 weirdo
1:18:10 [Music]
1:18:12 uh
1:18:14 Dave David
1:18:18 Shapiro now I think his last one
1:18:22 is bigger than you
1:18:25 think by the way this video I'm hope bur
1:18:29 with clear.com
1:18:31 hope
1:18:34 sh
1:18:39 um this video bigger than you think that
1:18:42 talks about
1:18:43 qar is is a really good video If there's
1:18:47 indeed a paper about qar it's not in
1:18:49 this video but there's a bunch of other
1:18:51 papers that are and David Shapiro talks
1:18:54 about um the implications of if open AI
1:18:58 has actually hit AGI and then the the
1:19:00 video before it is what is qar so these
1:19:04 two videos from David Shapiro are worth
1:19:06 are worth looking
1:19:08 at
1:19:14 uh
1:19:16 hey hey it's KL Shannon with air
1:19:19 learning L old people with
1:19:22 ADHD
1:19:25 hello is the all this weird no sometimes
1:19:29 he's
1:19:35 weirder what research paper and stop
1:19:38 spamming yeah please it so we've got
1:19:39 mods in here so you don't need to uh put
1:19:42 the same thing in a bunch of times
1:19:43 people will see
1:19:45 it um this is why we need AI it's like
1:19:47 having an external drive to remember
1:19:49 stuff this Mary Mary that's one of the
1:19:52 things I'm actually most most excited
1:19:53 about with AI is
1:19:56 that if you think
1:19:58 about all of the data that you've
1:20:01 created in your
1:20:03 lifetime or all of the data that you've
1:20:06 created at a company in some amount of
1:20:10 years most of that data is
1:20:13 just off to The
1:20:15 Ether a there's a lot of good thinking
1:20:18 in there there's probably presentations
1:20:20 you wrote there's probably documents you
1:20:22 wrote there's probably
1:20:24 products that you came up with support
1:20:26 documents you came up with that just end
1:20:28 up not getting into whatever system
1:20:30 became the
1:20:31 system and it doesn't mean that that
1:20:34 content that data doesn't have value it
1:20:36 likely has value but historically you've
1:20:40 had to spin up a whole software
1:20:43 development project to even be able to
1:20:46 find the data much less do anything with
1:20:49 it and now we've got these tools where
1:20:51 you can just like ah [ __ ] it here's a a
1:20:53 pile of
1:20:55 data go go just dump it into a GPT and
1:20:59 see if there's any value in there oh
1:21:01 [ __ ] that was pretty
1:21:03 good we yeah and and I I I just think
1:21:07 about the the enormous amount of value
1:21:09 that's sitting inside companies
1:21:12 everywhere globally that is just unused
1:21:16 untapped
1:21:18 potential value that's just going to be
1:21:22 blossomed out
1:21:24 and then that's going to lead to new
1:21:26 ideas and that's going to lead to new
1:21:27 ideas and then it's going to accelerate
1:21:29 because it's AI it's
1:21:37 crazy if you can figure out Tik Tok you
1:21:40 can figure out anything Earth rocker
1:21:42 this interface is it is the worst
1:21:45 interface in the history of interfaces
1:21:48 and it's it is you talk about age bias
1:21:51 it's like I don't think there's a font
1:21:53 on here that's bigger than you know a
1:21:56 Sixpoint font most of it is is like
1:21:59 reading
1:22:00 microfilm it's like
1:22:04 what but if we're going to play with the
1:22:06 young kids we got to get our eyes worked
1:22:09 on got to make sure you got your glasses
1:22:11 dialed up right turn it
1:22:22 up qar is the
1:22:25 source what do you mean qar is the
1:22:28 source is that a q reference Elon is
1:22:33 jelly someone said Thank You mods again
1:22:35 I agree thank you mods I put a half
1:22:38 finished script into chat GPT and asked
1:22:40 it if I should finish it gave me five
1:22:43 reasons wow that's great and then yeah
1:22:46 like like
1:22:47 listen it gave you five reasons why you
1:22:49 should finish and then it probably said
1:22:51 you could say uh well then finish the
1:22:53 outline for me and it'll give you the
1:22:56 outline and you collaborate with it and
1:22:57 get the outline dialed in and then you
1:22:59 have a give given outline for each of
1:23:01 the scenes it created and then you just
1:23:04 go write them Tik Tok is why I started
1:23:07 wearing reading
1:23:11 classes see there's a whole sub economy
1:23:14 that Tik Tok is
1:23:16 generating therapy glasses Tik Tok makes
1:23:20 Facebook look look like Shakespeare it's
1:23:23 so
1:23:26 true oh man um I'm sorry I meant
1:23:29 perplexity IOS app was legit and from
1:23:32 the devs yeah no you're right Kevin I
1:23:34 for some reason I had it in my head that
1:23:36 uh perplexity was from inflection and
1:23:37 it's not Pi is from inflection so thank
1:23:40 you for clarifying that Kevin Kevin
1:23:42 Vasquez we know things are bad worse
1:23:44 than bad they're
1:23:46 crazy crazy
1:23:48 Yeti then again I'm not looking at it
1:23:51 could two books in a transcript be
1:23:52 compared by AI to find similarities in
1:23:55 missing pieces absolutely Firefly
1:23:57 absolutely just yes go do that um you
1:24:02 could probably figure out which one was
1:24:04 the original like if the
1:24:06 transcript if it's a transcript was the
1:24:09 original and then you have two updated
1:24:10 books I'm sure it can and and what you
1:24:14 can also do is you could say find the
1:24:17 differences and document them in a
1:24:19 spreadsheet and turn that spreadsheet
1:24:21 into a downloadable like a CSV file or
1:24:24 an Excel
1:24:25 file um because it's got code
1:24:28 interpreter in it it can it can write
1:24:30 files I don't know if you knew that just
1:24:33 tell it to give you a downloadable Excel
1:24:36 file of all of the changes that it finds
1:24:39 across the three
1:24:41 documents now will it get them right I
1:24:44 don't know it hallucinates like a
1:24:46 [ __ ] so probably there'll
1:24:48 probably be some some Jank in there but
1:24:51 it'll probably it'll probably get you 85
1:24:54 90% of the way there and if qar is real
1:24:58 then it's going to get you 100% of the
1:24:59 way there and we don't need to edit
1:25:01 anything anymore because everything will
1:25:02 be perfect every
1:25:05 time
1:25:08 yikes
1:25:10 oops
1:25:12 oops the thing Shapiro talks about in
1:25:15 that video about qar the implications of
1:25:18 it is there's a there was a a a redacted
1:25:22 letter post to deor Chan um that he says
1:25:26 I don't know if this is real but if it's
1:25:27 real it's terrifying that basically says
1:25:30 that um that uh cryptography is
1:25:34 essentially
1:25:36 dead that that whatever this thing is
1:25:39 can decrypt 192-bit
1:25:42 cryptography um I put a okay yeah that's
1:25:45 great we talked about that let's mix a
1:25:47 1996 AOL chat room with video that's Ted
1:25:51 looks at life that is totally what Tik
1:25:53 Tok is this is this is an AOL chat room
1:25:57 with
1:25:58 video that's [ __ ]
1:26:02 awesome oh man and it and it's also got
1:26:04 the same sort of like parental controls
1:26:07 where if it if you flash a booby it'll
1:26:09 ban you for life don't hate me because
1:26:12 I'm janky
1:26:16 exactly I started wearing contacts and
1:26:18 I'm considering lasic due to Call of
1:26:21 Duty yep
1:26:24 it just the eyes don't get better that's
1:26:26 all I know um JC brain is on wind down I
1:26:29 was doing for memory got to look it up
1:26:31 on my laptop output Excel with
1:26:34 comparisons wow yeah give it a shot
1:26:35 Firefly I mean seriously just go go
1:26:38 upload the two books in the transcript
1:26:40 right now and just start in
1:26:42 interrogating it and see if it's getting
1:26:45 it anywhere close and then yeah and then
1:26:47 just tell it to um create an Excel file
1:26:51 that tracks all the changes it
1:26:54 finds and you and then you can go look
1:26:56 at them see if it got it right it's
1:27:03 amazing yeah gp4 is
1:27:06 nuts I should probably do a uh what time
1:27:09 is
1:27:10 it I should probably do like a gp4 just
1:27:13 a walkthrough of all the [ __ ] it can
1:27:15 do You've got mail
1:27:21 exactly can you specify sites that
1:27:25 require authentication for GPT
1:27:30 conversion
1:27:31 um I don't know I assume if you can
1:27:38 write if the site's got API access and
1:27:41 you've got some way to authenticate
1:27:44 maybe but when when you're starting to
1:27:46 get into using authenticated sites as
1:27:50 Source documents for gpts
1:27:53 you're now starting to get into
1:27:55 application development I'd probably
1:27:56 just go over to the playground side with
1:27:58 assistance and just just build an
1:28:03 application how do you find books to
1:28:05 load wait how do you find the books to
1:28:08 load so so we can go do let's let's go
1:28:10 let's go play Let's go try
1:28:14 something
1:28:17 um one of the one of the challenges I
1:28:19 have in finding [ __ ] to load is that
1:28:23 Google search is so useless now that
1:28:28 um it's it's really hard to find stuff
1:28:31 but let's see
1:28:32 free public domain
1:28:38 books
1:28:40 PDF because every because oh Project
1:28:43 Gutenberg okay does it have PDF
1:28:48 downloads search of
1:28:51 bookshelves
1:28:53 frequently
1:29:13 downloaded let's see I want to find um
1:29:21 selfhelp
1:29:32 emergency
1:29:35 childbirth we we can make an emergency
1:29:37 child birth GPT hang on honey I just got
1:29:41 to get a gp4 subscription oh [ __ ]
1:29:43 they're not taking new subscribers oh
1:29:46 hang on honey I don't know what to do
1:29:47 what are I going to
1:29:50 do uh oh this is not good all
1:29:54 right let's just go find a something the
1:29:58 brother brothers caramat
1:30:02 off let's see where do I have
1:30:07 PDF sorry I can't help with that hey
1:30:11 Siri mhm go
1:30:14 away
1:30:16 okay one of my favorite things to do is
1:30:19 to tell Siri to go away
1:30:25 plain text well I guess we could grab a
1:30:26 plain
1:30:29 text online
1:30:41 web save page
1:30:44 as
1:30:46 brothers caramat
1:30:48 off so this will be interesting so
1:30:51 there's Brothers caramat
1:30:54 off let me go a
1:30:59 desktop so I think this is a text
1:31:18 file the profit by khil gabron
1:31:24 other Kindles why does it not have a PDF
1:31:27 it must have one that I'm just not
1:31:29 seeing cuz I'm
1:31:45 old we'll grab an UB
1:31:49 version okay so now let's go to chat
1:31:54 J so now I've got a text file and I've
1:31:58 got an Epub
1:32:04 version so
1:32:08 downloads there's my khil gabron
1:32:13 book and
1:32:18 then there's my brother's katsof text
1:32:21 files so we now have these two things so
1:32:23 now I'm going to say is
1:32:26 um please Summarize each of
1:32:35 these reading
1:32:37 documents as Kyle said takes longer to
1:32:39 find the data than get the GPT
1:32:41 configured it totally does finding the
1:32:45 data is if you want data data like CSV
1:32:48 files go to kaggle.com
1:32:50 called self-help
1:32:52 equals transformational per bars oh
1:32:55 Barnes &
1:32:57 Noble transformational okay yeah
1:33:01 anyway not doesi please that's
1:33:06 good reading Brothers karamat off.
1:33:12 text so I guess it read the other one
1:33:17 already oh Kyle Now find some Snappy
1:33:20 titles on death and dying
1:33:22 oh yeah exactly all right let's well
1:33:25 that's well that's thinking we can go do
1:33:27 that let's see we'll go over
1:33:29 here we'll do a quick search
1:33:33 death and
1:33:51 dying
1:34:03 there's nothing on nothing in Kindred
1:34:07 the
1:34:08 amateurs I don't know I'm I don't know
1:34:11 how to search this site I don't this is
1:34:13 Project
1:34:14 Gutenberg seemed like a good idea at the
1:34:16 time but I just don't get
1:34:18 it it looks like this
1:34:20 hung
1:34:23 I'm going to stop it and say try
1:34:32 again message in conversation not found
1:34:36 regenerate H good
1:34:40 Lord this is why we don't demo things
1:34:43 live people this is why we can't have
1:34:45 nice things let me actually see if I got
1:34:48 Brothers caramat off yeah okay so the
1:34:50 book is there
1:34:53 fine so we're going to upload that now
1:34:55 we're going to upload our
1:34:58 little The
1:35:01 Profit
1:35:04 oh hang
1:35:08 on I wonder if
1:35:14 this I wonder if it's a wrong file
1:35:16 format if I choked it it it shouldn't
1:35:19 let me upload it if it doesn't
1:35:20 understand it
1:35:42 where are
1:35:44 you what's that going to open it with
1:35:47 welcome to Apple
1:35:49 Books get started The
1:35:53 Profit all
1:36:07 right um that file format might be
1:36:11 [ __ ] chat GPT up but let's
1:36:14 see hang
1:36:16 on
1:36:20 um can you to create a table that
1:36:25 compares these
1:36:28 two
1:36:32 documents figure out the
1:36:36 categories for
1:36:39 comparison and then make the
1:36:44 table take a deep
1:36:48 breath and
1:36:50 think
1:36:54 step by
1:36:57 step or stop by step now the think step
1:37:02 by take a deep breath and think step by
1:37:05 step that is a prompt that chat GPT came
1:37:08 up with on its own um in some sort of
1:37:11 experiment that Ethan mik did um that
1:37:13 apparently generated better results than
1:37:16 any human prompts so I sometimes will
1:37:19 put it in there just to mess with it
1:37:22 let's see if we can get past the katof
1:37:25 text
1:37:27 reading and if you're thinking like oh
1:37:29 that's too long it can't that's why it's
1:37:31 taking so long it's because it's too
1:37:33 long no no it this reads really long
1:37:35 stuff essentially
1:37:38 instantly very very
1:37:40 quickly being a nothingness by
1:37:44 satra waiting for you to go okay so
1:37:47 let's
1:37:49 see here's a comparison
1:37:53 it appears that I'm unable to access the
1:37:55 contents of the second
1:37:59 document
1:38:05 anyway so I'm going to say
1:38:08 um okay give me
1:38:16 a list of characters in each chapter and
1:38:23 what
1:38:26 happens to them in that chapter as
1:38:32 a nicely
1:38:37 formatted Excel
1:38:41 file message and conversation not found
1:38:45 what is going
1:38:47 on why is chat GPT breaking people
1:38:52 [Laughter]
1:38:58 shy what' you do Saturday night we
1:39:00 watched some guy try to use chat
1:39:03 GPT why would you do that we don't
1:39:08 know it's not working but anyway it it
1:39:12 works it's just not now this is this
1:39:15 channel is like I'm like one of those
1:39:17 guys like no no it's going to be really
1:39:19 cool come here just wait I just got I
1:39:21 just got to go find the thing it's hang
1:39:23 on no it's going to be so good it's just
1:39:25 if you get and then like an hour later
1:39:26 you're like could can I leave can could
1:39:29 should we order a
1:39:31 pizza waiting for gdau to load it's
1:39:34 totally waiting for gdau to load
1:39:36 manybooks.net sign in with your Google
1:39:38 account all right let's go try to find
1:39:40 some other books that was very
1:39:41 disappointing maybe the UB broke chat
1:39:45 gbt let's give it the benefit of the
1:39:48 doubt and go to manybooks.net
1:39:52 manybooks.net I just want one or
1:40:00 two you may have
1:40:02 visited several potentially infected
1:40:05 sites that oh it took me the wait
1:40:10 manybooks.net
1:40:11 did I spell it wrong I did
1:40:18 okay so let's
1:40:20 go death and
1:40:28 dying on death and dying by Edith CER
1:40:32 Ross stages and Analysis let's grab that
1:40:35 can we grab
1:40:37 that for students study
1:40:49 with no this is
1:40:53 see everything's a [ __ ] sales funnel
1:40:55 to get you to sign up for some other
1:40:58 damn
1:40:59 service it makes me
1:41:02 hurt H The Time Machine
1:41:07 download okay continue with Google oh
1:41:10 this is
1:41:14 good
1:41:20 what
1:41:25 oh no that just added oh
1:41:27 godamn everything a sales funnel I
1:41:31 clicked on the wrong thing people
1:41:34 [Laughter]
1:41:37 no no stop stop forcing me to buy your
1:41:41 crap I don't want to buy your crap I
1:41:43 don't want your stupid [ __ ]
1:41:49 extension free download time machine
1:41:51 free
1:41:52 download continue with
1:41:59 Google
1:42:00 uh free download is it really a free
1:42:05 download is it really a free
1:42:10 download
1:42:13 pdf save
1:42:16 selection no
1:42:19 thanks download now
1:42:22 all right we got the time machine okay
1:42:26 I I think I think we're getting there
1:42:28 all right let's go get let's do the time
1:42:30 machine and see if we can get a book
1:42:32 about death and dying what wait
1:42:38 what how many people are here there's
1:42:40 100 people watching me unsuccessfully
1:42:43 navigate a stupid [ __ ]
1:42:47 website sorry about this people death
1:42:50 and
1:42:51 uh dying uh Talk Amongst yourself I'm
1:42:54 sure you are be nice to each other don't
1:42:56 be
1:43:00 [ __ ]
1:43:03 ah let's see we have
1:43:07 Dracula healing her heart let's
1:43:14 see oh those are sponsored links that's
1:43:17 why that was such a
1:43:20 nightmare
1:43:26 oh I think these are all there's not
1:43:27 non-fiction books here that's why I'm
1:43:29 having a tough
1:43:31 time ah here we go accidental death
1:43:35 death and
1:43:48 after tired of death the flying death
1:43:51 death let's go Edgar Allen Poe all right
1:43:55 we'll do Edgar Allen
1:43:56 Poe The
1:44:03 Raven Okay so we've got the Raven and
1:44:06 time machine as PDFs now let's go try to
1:44:09 chat JPay
1:44:12 T upload upload it's gonna be so good
1:44:16 people I would rather watch my
1:44:17 grandmother attempt wait what was that
1:44:19 Pate that looks good
1:44:21 I would rather watch my my grandmother
1:44:26 something we we'll be dead before we
1:44:29 find the book on death and dying I know
1:44:32 this is really bad this is so bad it's
1:44:35 okay it's well it's okay for
1:44:39 me you're you're the ones that are still
1:44:42 here I put that on you this is not this
1:44:45 is not a me problem this is a you
1:44:50 problem I would rather watch my
1:44:52 grandmother's attempt to Tor into movie
1:44:55 exactly okay we got two PDFs in here so
1:44:58 now let's let's ask old chat
1:45:02 GPT please summarize these two
1:45:06 books and
1:45:08 identify five themes that
1:45:15 overlap all right reading documents
1:45:18 let's see let's say say
1:45:25 say chat am I
1:45:27 dumb am I the dumb one or is Gutenberg
1:45:30 not the go-to Gutenberg was a good idea
1:45:33 at the time and then like many op Source
1:45:37 projects it didn't get the best talent
1:45:40 because the best talent always starts
1:45:41 out there and then goes off and does
1:45:43 other
1:45:44 stuff
1:45:47 um still mowing keep going Kyle in
1:45:51 another hour of
1:45:54 mowing it's find it's hard to find a
1:45:57 Snappy little title in Hospice Care all
1:46:00 right summary analysis of the time
1:46:01 machine and three stories by Edgar alen
1:46:04 Poe all right so it's summarizing them
1:46:07 for us
1:46:09 great the raven a narrative poem about a
1:46:12 man lamenting the loss of his love Lenor
1:46:16 a raven visits him repeating the word
1:46:18 never more symbolizing a descent into
1:46:20 mad and
1:46:22 despair o five overlapping themes the
1:46:25 nature of time and its
1:46:29 effects mortality and the inevitability
1:46:32 of death social class and equality
1:46:34 revenge and retribution isolation and
1:46:36 loneliness all right so now let's try
1:46:43 this
1:46:47 terrific what's going on why come on
1:46:51 seriously
1:46:53 seriously you're making me
1:46:57 crazy what what is
1:47:04 happening oh it restarted it I guess it
1:47:07 was not
1:47:08 done
1:47:11 ah oh my God this is really bad jokes on
1:47:14 us guys nobody is left actually we've
1:47:16 got more people here 120 folks hey
1:47:18 everybody my name is Kyle Shannon and
1:47:20 this is uh a real-time grass growing
1:47:23 game where uh watching grass grow is
1:47:26 more exciting than this channel I am
1:47:29 attempting no I have attempted and
1:47:31 successfully found two PDFs of
1:47:36 books semi semi- related to death and
1:47:39 dying for someone who was trying to
1:47:41 figure something out for that and yeah
1:47:45 so anyway all right let's let's
1:47:47 let chat GPT do its thing here
1:47:53 so I asked it to summarize so we
1:47:55 uploaded the time machine and uh three
1:47:58 books by Edgar Allen Poe or three short
1:48:01 stories by Edgar Allen Poe including the
1:48:03 Raven and so now we're having chat GPT
1:48:07 summarize said books and then identify
1:48:11 five
1:48:13 themes that cross the
1:48:18 books The Mask of the Red Death
1:48:21 is a short story set in fictional
1:48:24 land ravaged by a deadly plague known as
1:48:27 the Red
1:48:29 Death all right five themes that overlap
1:48:33 human psyche and
1:48:34 Madness the inevitability of death class
1:48:38 Division and social
1:48:40 inequality Gothic elements in
1:48:43 atmosphere continued generating Ah
1:48:45 that's what what was wrong before I had
1:48:48 to continued generating that I missed
1:48:52 uhoh did it die no come on come on chat
1:48:56 GPT you can do it Sam alman's back
1:48:59 you're going to be
1:49:01 okay you've got qar coming all right am
1:49:05 I am I good
1:49:08 now that's awesome I love those five
1:49:11 themes can you now create an Excel
1:49:15 file with a different worksheet per
1:49:19 theme and a list of matching quotes from
1:49:25 each
1:49:29 book that bring that theme to life and
1:49:33 show the parallels in the
1:49:36 books I have no idea if this is even
1:49:39 possible and if it can figure this [ __ ]
1:49:40 out but it says analyzing so analyzing
1:49:43 means it's using Code interpreter and if
1:49:45 I open that up here's the python that
1:49:48 it's written if you don't know how code
1:49:49 interpreter works
1:49:51 it it's actually writing python code so
1:49:55 so it's basically saying Define the
1:49:57 themes and
1:49:58 initialize an empty dictionary to hold
1:50:00 the quotes so there's the five
1:50:03 themes then it's it's so it's it's
1:50:05 basically writing all this
1:50:07 um it's only added one quote it looks
1:50:11 like now here's an Excel writer so it's
1:50:14 it's creating an Excel an Excel document
1:50:17 so we'll see what it does and if it's
1:50:18 any good error analyzing
1:50:21 and oh so it's trying again it didn't
1:50:22 like what it did before so it's trying
1:50:28 again an Excel file has been created
1:50:31 with a different worksheet per theme
1:50:33 each containing relevant quotes from The
1:50:34 Time Machine and Edgar Allen po works
1:50:37 you can download the file here so we
1:50:39 downloaded the file we'll go look at
1:50:42 that there
1:50:44 file and then we will open that there
1:50:46 file we're going to open her up is that
1:50:49 that's Excel
1:50:51 open with here we
1:50:55 go is this going to be on the test
1:51:02 absolutely so it only gave us one quote
1:51:06 per
1:51:07 but let's
1:51:10 see it did do it the time machine I
1:51:14 grieved to think how brief the dream of
1:51:17 the human intellect had been it had
1:51:19 committed SU
1:51:21 suicide and the silken sad
1:51:25 uncertain rustling of each purple
1:51:28 curtain thrilled me filled me with
1:51:30 fantastic Terrors never felt before
1:51:34 interesting so yeah so it's got this is
1:51:36 an Excel thing with five sheets in it
1:51:38 with one quote each I mean you could
1:51:40 tell it to do things so yeah that [ __ ]
1:51:44 Works crazy crazy crazy crazy like what
1:51:48 does that make possible I don't know
1:51:51 go figure that [ __ ] out this is what I
1:51:53 mean about playing with this stuff like
1:51:55 just play with it as as painful as that
1:51:57 was sorry about that is it all python
1:52:00 internally yeah so the way the way code
1:52:03 interpreter works is it creates what's
1:52:05 called a container So within your chat
1:52:07 session it for lack of a better term
1:52:10 it's creating like a little mini
1:52:13 instance of of a computer that can run
1:52:15 code and then it's given access to
1:52:17 certain python libraries and so when you
1:52:19 ask it to do stuff it spins up this
1:52:23 container writes and executes the python
1:52:25 code and then writes to like little you
1:52:28 know memory files and and you know sort
1:52:31 of I I guess little mini databases
1:52:34 within its within that instance of your
1:52:36 chat so everything's happening
1:52:38 internally within within the chat
1:52:40 session and Pate I know you're in here
1:52:42 so if I'm if I'm saying that wrong
1:52:44 please and I know you won't be shy smack
1:52:47 me back if I'm describing that
1:52:49 incorrectly all
1:52:51 right that was that was painful sorry
1:52:54 about that listen
1:52:56 normally that was that was a half an
1:52:58 hour that was a half an hour of watching
1:53:01 an old man try to download a PDF you're
1:53:04 welcome you're welcome this is the AI
1:53:06 learning lab welcome to the AI learning
1:53:08 lab where all of your greatest dreams
1:53:10 come true
1:53:13 and you want to hear a dog
1:53:17 sing I can do that well I can I can I
1:53:22 can invoke
1:53:26 that that was painful be not afraid I do
1:53:29 not deserve that Rose take that
1:53:32 back thank you Apple
1:53:35 user you weirdos taking notes in
1:53:48 Python Jenna a I has breathed new life
1:53:51 into python which had sort of taken a
1:53:53 backseat to other languages it sure has
1:53:55 well I mean code interpreter in
1:53:57 particular but but also just all the
1:53:59 data analysis stuff I mean python was
1:54:01 always there for that so I think I think
1:54:04 code interpreter really really kind of
1:54:05 woke woke things
1:54:10 up
1:54:12 but I ain't a programmer so I don't
1:54:17 know that was most helpful oh good two
1:54:19 block time because I thought that was
1:54:22 insufferably awful so if that was
1:54:24 helpful at least one person got some
1:54:26 value out of the last half hour the rest
1:54:29 of you you have my deepest sincere
1:54:34 apologies Kyle I saw your early magazine
1:54:38 cover from agency.com says on the way
1:54:40 back machine it's true it's
1:54:43 true um which one the one back here
1:54:47 silicon alley
1:54:49 reporter
1:54:52 yeah that's Jason kenis is silicon alley
1:54:57 reporter from this was from 96 I
1:55:00 think was it 97 November
1:55:04 97 that was my partner Chan and I from
1:55:07 agency.com
1:55:14 we should find and download a bunch of
1:55:17 PDFs and share on Mighty Network and
1:55:19 then build stuff
1:55:20 in real time that's actually not a bad
1:55:23 bad idea Joe mama I mean and that's one
1:55:25 of the so listen that's one of the
1:55:28 um so a couple of things that's one of
1:55:31 the purposes of the salon and and that's
1:55:33 a kind of collaborative thing the other
1:55:35 thing about Mighty networks and we've
1:55:38 we've just launched it it's like three
1:55:39 days old so we're still learning it but
1:55:43 one of the things you can do on Mighty
1:55:44 networks is you can go live so you could
1:55:47 set up a time and you can also do
1:55:49 calendaring within Mighty networks So
1:55:52 within you know like the in either in
1:55:55 one of the guilds or or even in the
1:55:57 water cooler you could set up a meeting
1:56:00 um with people to do that you know
1:56:03 download a bunch of PDFs and then fire
1:56:05 up a live session and everyone could be
1:56:07 on it working together so yeah sorry for
1:56:10 encouraging but I needed that thank you
1:56:12 guys no Firefly thank you for thank you
1:56:15 for asking about it cuz it's like every
1:56:17 time I I say one of these things like
1:56:20 I'm pretty sure it can do that I I
1:56:22 always doubt myself because it still
1:56:24 blows my
1:56:25 mind
1:56:27 that I
1:56:29 mean if we go back
1:56:32 to what this actually did it you take
1:56:37 the time machine and three short stories
1:56:40 by Edgar Allen Poe right no one would
1:56:44 ever in their right mind compare those
1:56:46 from a literary
1:56:47 standpoint and so just as a lark we said
1:56:50 hey compare them from a literary
1:56:51 standpoint and it did like it summarized
1:56:54 them beautifully and then it came up
1:56:56 with themes and then like like what chat
1:57:00 GPT is actually capable of of doing is
1:57:03 astounding I the the thing that
1:57:07 um the generative stuff where it's like
1:57:09 write me a poem about quantum physics
1:57:11 and the style of Eminem Like that's
1:57:13 impressive because it's sort of
1:57:14 generating this thing out of nothing the
1:57:16 thing that blows me
1:57:18 away again and again is how good it is
1:57:21 at analysis how good it is as taking a
1:57:24 document or taking a piece of computer
1:57:26 code like take this piece of computer
1:57:28 code and tell me what it
1:57:30 is I did that with with my company's XML
1:57:33 files from how we assemble our videos
1:57:37 and I just said what is this and it it
1:57:39 like it got it
1:57:41 flawlessly and then I said I want to I
1:57:44 want to
1:57:46 recreate the XML schema to to follow a
1:57:50 different format and it rewrote the XML
1:57:53 and and Incorporated all of the
1:57:55 variables like it it got it
1:57:58 right and then you know this where it's
1:58:00 saying you know here here are the five
1:58:02 themes that that are present in both of
1:58:04 these books and it's ability to pull out
1:58:06 quotes that that bring those themes to
1:58:08 life what and it did that in like 30
1:58:12 seconds you as was accurately pointed
1:58:14 out it took way longer to find the PDFs
1:58:18 and be able to download them than did
1:58:20 for chat GPT to be able to do that
1:58:24 analysis it's just it's [ __ ] crazy
1:58:27 it's [ __ ] crazy just hit my usage cap
1:58:30 dang it yeah I think what's the usage
1:58:33 cap now 30 4 40 things every three hour
1:58:36 every three hours I think it is yeah 40
1:58:39 messages per three hours that sucks
1:58:42 sorry about that I talk so much here
1:58:44 that I don't usually hit my my usage
1:58:47 cap because I'm just blabbing
1:58:50 oh we know Kyle we're well aware of
1:58:54 that yes generally if we're flying
1:58:57 British Airways we like to wear smart
1:59:03 [Music]
1:59:10 pants you're in need of Dr tar and
1:59:13 Professor
1:59:15 something Professor
1:59:18 feather
1:59:20 uh P modder it's the
1:59:27 best I have my a book my mom wrote but
1:59:30 want to have ai read the book and
1:59:32 improve on as a professional editor so
1:59:35 Tik Tock
1:59:36 dude so so what to do there if you're if
1:59:39 you're new to chat
1:59:41 GPT um there's a there's a document
1:59:45 called prompt. chat and what prompt.
1:59:49 chat is is is it's kind of a primer on
1:59:52 prompting but it's a whole list of
1:59:54 prompts where you can tell chat GPT to
1:59:57 act like a certain thing and I'm sure
1:59:59 somewhere within there is a prompt
2:00:01 that'll tell it to act like a
2:00:02 professional book editor um and if not
2:00:05 you can take something that's close in
2:00:06 there and rewrite it or even have chat
2:00:08 GPT rewrite it and so start out by
2:00:11 telling chat GPT it is a professional
2:00:13 editor and then say you know I've got
2:00:15 this book and here's give it as much
2:00:17 context as you can so if you're if
2:00:19 you're new to chat
2:00:21 GPT unlike a Google search where the
2:00:25 goal is to sort of give it as little
2:00:26 information as possible I just need you
2:00:28 know these three keywords and then it'll
2:00:31 go find everything chat gbt is kind of
2:00:33 the opposite where it needs context so
2:00:37 tell it to be a professional editor tell
2:00:39 it you know what everything you know
2:00:41 about the book and you know that your
2:00:44 mother wrote it and you want to take it
2:00:46 from the level it's at now to some more
2:00:48 professional level and and you want it
2:00:50 to explain to you how it's going to
2:00:52 approach this process and have it tell
2:00:55 you that before it does the work say I'm
2:00:56 going to upload this book before but
2:00:58 before I do that tell me how you're
2:00:59 going to analyze this and it will do
2:01:02 that it's uh it's pretty pretty
2:01:06 remarkable in how it do what it do let
2:01:09 me go check one thing here all right
2:01:12 everything's good there what do we got
2:01:16 here that's that hold please
2:01:20 hold
2:01:25 hold death and
2:01:28 dying Ted looks at life this is what I
2:01:31 did to the
2:01:40 class this is good is this oh this is
2:01:43 steo steo gave us a good one wait hang
2:01:46 on loading we're loading we're loading
2:01:50 Siri go
2:01:52 away
2:01:57 okay oh my
2:02:00 God someone was asking earlier about
2:02:02 autog GPT hang
2:02:05 on there's that's how Auto GPT is doing
2:02:08 good it's doing
2:02:13 good that's hilarious where's the lawn
2:02:16 mower meanwhile I almost completed my
2:02:18 Sunday new times cross word puzzle
2:02:21 listen people don't you get SN Don't You
2:02:23 Get Saucy with
2:02:26 me I'll make champ
2:02:28 [Music]
2:02:42 [Music]
2:02:48 sing
2:02:50 what you doing making champy
2:02:53 sing Irregulars present top of the
2:02:57 hour don't get spicy the dogs how it's
2:03:00 true sing it
2:03:05 champ I what I figur what I've also
2:03:07 figured out about this channel is you're
2:03:08 not here for me you're here for champ
2:03:10 and I'm totally down with that it's he's
2:03:12 he's the reason to be
2:03:14 [Laughter]
2:03:18 here
2:03:21 all right let's see very cute things
2:03:24 over in the AI
2:03:27 [Laughter]
2:03:31 Salon all right we got some intro so let
2:03:34 me so looks like there's some new people
2:03:36 here let me go share some I want to
2:03:39 share a couple of things with you
2:03:42 oops I'm going to do this very very
2:03:44 quickly get ready to do some
2:03:45 screenshotting I'm going to go through
2:03:47 this really quick because I want to wrap
2:03:49 it up up soon if you're new to chat GPT
2:03:51 ignore the thing on the right for now if
2:03:54 you're if you have not played with chat
2:03:55 GPT before just go there right now like
2:03:58 right now like now I'll get to you
2:04:01 tomorrow Kyle I'm not very technical I
2:04:03 don't like all this fancy AI [ __ ] no no
2:04:05 no go there right now chat.
2:04:09 open.com and just start playing with it
2:04:11 so as you're bored as I'm trying to
2:04:13 figure out how to download PDFs just
2:04:15 start telling it to do stuff for you
2:04:17 have it write you poems songs recipes
2:04:20 just just try to stump it with some
2:04:23 obscure knowledge you know just go play
2:04:24 with that um and just keep playing with
2:04:26 it until it blows your mind to the point
2:04:28 that you have your first Kevin
2:04:29 McAllister moment which is that
2:04:33 moment um and then if you're a geek go
2:04:35 to the thing on the right that's the API
2:04:37 access to the back end of chat GPT if
2:04:39 you've not played with that there's some
2:04:41 fun things you can do back there you
2:04:43 know fun in the sense of if you're a
2:04:46 geek if you're a normal human being
2:04:48 you'll be like what are all these
2:04:50 numbers mean what why do I have to pick
2:04:52 different models what this this is
2:04:54 making my head hurt this is what I was
2:04:56 afraid AI was going to be that's what
2:04:58 the right side is the left side is more
2:05:00 fun okay um screenshot this so these are
2:05:04 the kind of the top four competitors to
2:05:06 chat GPT Claude grock which is not out
2:05:09 yet but is coming next week from a from
2:05:13 X Ethan or what's his name Elon Musk and
2:05:17 then Bard is Googles and pilot is chat
2:05:21 GPT 4 but it's Microsoft's flavor of it
2:05:25 and then prompt. chat I just talked
2:05:27 about um these are the other four that I
2:05:29 think are worth playing with if you're
2:05:31 in doing research and things perplexity
2:05:33 is really good if you want to talk to
2:05:35 one of these things like have a
2:05:36 conversation piie is really good it's
2:05:39 why Siri seems so ridiculous at this
2:05:41 point uh and and uh Alexa or whatever
2:05:45 the hell her name is
2:05:47 uh and then you got llama and po both of
2:05:49 those are interesting as well here's
2:05:52 four image tools that are kind of of
2:05:54 different flavors that I think are worth
2:05:56 you knowing about and exploring and then
2:05:59 so screenshot this and then that that
2:06:01 thing at the bottom The Prompt book will
2:06:03 explain to you how
2:06:07 insane these image generation tools are
2:06:10 like what they can actually understand
2:06:13 and create is
2:06:16 wild um then you've got video creation
2:06:19 stuff this is crazy so screenshot
2:06:23 that all
2:06:28 right this is music and audio 11 Labs
2:06:31 will synthesize your voice sunno and
2:06:34 refusion will make original songs that
2:06:36 are
2:06:37 sung it's amazing uh they're janky but
2:06:41 they're fun and then sound draw will do
2:06:43 music uh composition with loops really
2:06:46 really
2:06:47 quickly um there there's 3D worlds and
2:06:51 objects worth playing with the fourth
2:06:54 one I just don't know yet I haven't hit
2:06:55 one that really turned me on yet I I
2:06:58 have a feeling um Unreal Engine 5 in
2:07:01 2024 is going to be super AI
2:07:05 Dynamic you're going to be able to talk
2:07:07 worlds into existence that are like AAA
2:07:11 quality
2:07:12 worlds and then probably by 2025 you'll
2:07:16 be able to talk AAA quality games into
2:07:20 existence it's like yes and I think
2:07:23 that's coming and I think that's
2:07:24 actually quite
2:07:26 realistic um well what happens to the
2:07:29 triaa game studios I don't know man they
2:07:31 got to get their [ __ ] together um these
2:07:34 are people you should follow Ethan
2:07:35 mullik David Shapiro Paul rater and
2:07:37 Rachel woods so screenshot
2:07:41 that and then um and then this so we've
2:07:45 been we've been talking on here a little
2:07:47 bit about Mighty Network works and
2:07:49 people sharing stuff um where they're
2:07:51 sharing it is at the AI salon so this
2:07:53 this channel is really kind of an
2:07:55 extension of this community the AI Salon
2:07:57 which is a group I started a year ago
2:08:00 next Tuesday December 5th is our
2:08:02 one-year anniversary and um if you want
2:08:06 to check out the community it's free um
2:08:09 AIS salon. mn. so mighty
2:08:14 network.co um is the is the
2:08:17 um is the community site and and please
2:08:20 check out about the salon what we're
2:08:23 about and what our values are make sure
2:08:25 the values resonate with you if they
2:08:27 don't that's cool but it's probably not
2:08:28 the group for
2:08:30 you um and I mean they're not scary but
2:08:33 they're things like you know curiosity
2:08:36 and empathy and inclusion and generosity
2:08:42 like I want people that are like not
2:08:47 only curious to learn [ __ ] like I want
2:08:49 people that want to learn but then I
2:08:50 want people that are excited about
2:08:53 sharing what they
2:08:54 learned and collaborating with other
2:08:57 people and being adventurous and being
2:09:00 inclusive because people are coming from
2:09:02 everywhere and like having empathy for
2:09:05 the fact that people might be scared and
2:09:07 angry and freaked out about this stuff
2:09:09 and that's okay too so anyway that's
2:09:11 what that's all about enough soap box so
2:09:16 here you
2:09:17 go that's I'm just trying to that's a
2:09:19 little pallet cleanser to get the the
2:09:21 taste of finding PDFs for half an hour
2:09:24 out of our
2:09:25 mouth so then you recommend chat GPT for
2:09:28 my mom's book then uh I would say chat
2:09:32 GPT or Claude cuz Claude can do take her
2:09:37 whole book in as well
2:09:40 um and then someone just told me that
2:09:43 perplexity doai also
2:09:46 does 300,000 words
2:09:50 so so I would say perplexity doai
2:09:54 claw. chat. open.com for your mom's book
2:09:58 and just pop it in there and just start
2:09:59 asking it questions and see which one
2:10:01 you like best because they're all
2:10:03 different my my experience is that
2:10:07 nothing is anywhere close to chat gp4
2:10:10 thank you Joe mama that's a cool one
2:10:12 I've never seen that
2:10:13 one
2:10:15 um I think just in terms of writing and
2:10:18 and language Mastery and conceptual [ __ ]
2:10:21 chat GPT 4 is there's nothing there's
2:10:25 nothing close right
2:10:27 now so but but there might be a a way
2:10:32 that perplexity does what it does that
2:10:34 you're like oh I like that a lot better
2:10:35 or the way Claude does what it does you
2:10:37 like it a lot better I mean there's
2:10:39 there's people in here that swear by
2:10:40 Claude but that swear by perplexity
2:10:42 swear by Bard some people love
2:10:46 Bard um hours do you with chat GPT is
2:10:52 that a question I keep accidentally
2:10:55 swiping into other
2:11:01 lives that's really funny
2:11:07 Becky oh that's hilarious uh how many
2:11:10 hours do I put in on chat upt I don't
2:11:12 know I mean I'm using it daily at this
2:11:14 point a lot of it's here can you tell me
2:11:18 how to become an AI engineer so let's
2:11:20 see food loves me for who I am okay I
2:11:24 love the name for one thing 9.5 out of
2:11:26 10 I like the
2:11:28 name um I I think AI Engineers way way
2:11:33 way way way way way too broad um I don't
2:11:36 know your background I don't know what
2:11:37 you want to do with your life I would
2:11:40 say um do you want to be an AI
2:11:42 operations do you want to figure out
2:11:44 like what are the tools and how do you
2:11:45 install them and how would they be using
2:11:47 them do you want to be AI
2:11:50 security are you talking about being a
2:11:52 prompt engineer in that case I'd say
2:11:54 just start playing with the tools I mean
2:11:56 in general what you want to do is start
2:12:00 understanding the
2:12:01 landscape so if you want to become an AI
2:12:04 engineer and you're good at coding then
2:12:07 I would start you know joining open
2:12:10 source projects I'd start hanging out on
2:12:12 Reddit um find find people that are
2:12:15 doing that you can join the AI Salon P's
2:12:17 got a AI mechanics Guild in there where
2:12:19 you can talk about more geeky kind of
2:12:21 stuff back in the 90s I used to go to
2:12:24 the Unix Expos PC Expos networking Expos
2:12:27 Etc I want to go to a geni Expo um they
2:12:31 they don't really they're Sam Francisco
2:12:35 right now is having four or five AI
2:12:38 events a night um I just spoke at one
2:12:42 here in Denver that was like four months
2:12:45 ago like here in Denver they're like
2:12:46 once every six months but I do the salon
2:12:49 every two weeks so I guess that counts
2:12:54 um I don't know that there's I I mean I
2:12:56 would
2:12:57 say it it would be in your interest I
2:13:01 don't are you in I on China are you in
2:13:03 New York are you in a big city if you're
2:13:05 in a big city like New York I'd [ __ ]
2:13:07 start
2:13:09 one right or find out who's got one
2:13:11 going I'm sure I'm sure there's some out
2:13:13 there but if not why not start one why
2:13:14 not be like Alan meckler you know with
2:13:16 the early days of the web and and start
2:13:19 you know start that conference
2:13:22 business um Monique told me to ask you
2:13:26 about my Memoirs and
2:13:29 Claude
2:13:33 um New Jersey New York City yeah yeah
2:13:36 yeah if you're there like I'm in [ __ ]
2:13:38 Denver Denver is not New York and Denver
2:13:40 is not San
2:13:44 Francisco I miss New York uh if I were
2:13:47 in New York the the the salon would be
2:13:50 in a I think it would be in a very
2:13:51 different
2:13:53 place oh food food loves me for you're
2:13:56 in La yeah la is another one where La is
2:13:58 an interesting one because you've got
2:14:01 with LA You've Got The Writers Guild and
2:14:03 the SAG strike that just happen so
2:14:04 there's a hyper awareness that AI is
2:14:07 [ __ ] evil and so you're going to have
2:14:09 an uphill battle in La I think but in
2:14:11 New York they're like
2:14:13 like you could you could do this in New
2:14:16 York um Becky yeah so so with your
2:14:19 Memoirs I don't know I don't know quite
2:14:22 what what Monique was referring to other
2:14:24 than you can put Memoirs in there and
2:14:27 tighten them up if you want to if you
2:14:29 want to make them shorter or if you want
2:14:30 to expand them or if you want to finish
2:14:32 them or if you want to do editing on
2:14:34 them you can just upload your entire
2:14:37 Memoir and start interacting with it um
2:14:39 you could also turn your Memoir Becky
2:14:42 into
2:14:44 um into a GPT right you could upload
2:14:48 your Memoir and make your Memoir an
2:14:51 interactive GPT right what lessons were
2:14:54 learned what was what's the basic story
2:14:55 what happened here what happened there
2:14:58 um there could be something interesting
2:14:59 to do there
2:15:01 so yeah I don't know might be something
2:15:04 might be fun something fun to
2:15:07 explore um come join the guild and we
2:15:10 can talk about it more great so there
2:15:11 you go so Pate who runs the the uh the
2:15:15 mechanics Guild in the AI Salon that I
2:15:17 just gave the address for which is here
2:15:20 I'll just put it up on screen
2:15:30 again can you see
2:15:33 that there you go AIS salon. mn. so go
2:15:38 there join introduce yourself be
2:15:40 generous go check out the about the
2:15:42 salon check out our values make sure
2:15:44 you're cool with them and then join P's
2:15:47 um
2:15:48 mechanics Guild and yeah start talking
2:15:50 about doing some stuff Denver just
2:15:52 hosted a supercomputing conference I
2:15:54 used to attend oh super oh yeah I heard
2:15:56 about that yeah yeah yeah that's that's
2:15:58 more on the on the technical side of
2:16:00 things the Gen stuff would would be a
2:16:02 different kind of conference it would be
2:16:04 more like web conferences but yeah but
2:16:06 it's like the time to start everything
2:16:08 is now I live near San Francisco where
2:16:10 can I find these AI meetups Go to
2:16:12 meetup.com there's probably a lot of
2:16:14 them listed there uh Andrea and then I
2:16:17 would
2:16:21 say I would start trolling Twitter um
2:16:25 for like follow Robert
2:16:27 scoble
2:16:29 um Robert sco's got a whole bunch of um
2:16:34 lists of people he follows on Twitter a
2:16:38 lot of them are are Valley based I would
2:16:40 just start following people and just
2:16:42 just start asking just start developing
2:16:44 Community out there can you create a GPT
2:16:47 with all the company question mark used
2:16:49 by all employees you can absolutely do
2:16:51 that the thing that I would caution you
2:16:53 on is
2:16:56 um you should assume that anything you
2:16:59 put in a GPT could potentially get
2:17:02 discovered by people outside so I
2:17:04 wouldn't put anything super proprietary
2:17:06 in there but if you want like HR
2:17:08 documentation or [ __ ] like that you can
2:17:10 absolutely do
2:17:12 that Suzanne message me about San
2:17:14 Francisco events I have resources to
2:17:16 share so suzan my request to you would
2:17:19 be this um where's the best place to put
2:17:22 them join the salon if you haven't and
2:17:25 you could either share in the water
2:17:27 cooler in the water cooler there's a
2:17:29 there's a a tab that says news and
2:17:33 posts
2:17:42 um and then there's also an AI tools and
2:17:47 resources
2:17:49 section so if you've got if you've got
2:17:52 resources for San Francisco events join
2:17:55 the salon and go to the AI tools and
2:17:58 resources I'm still trying I'm I'm still
2:18:01 trying to learn what we've built and and
2:18:03 where's the appropriate place to put
2:18:05 stuff I think that's probably the best
2:18:06 place to put stuff if you'd share stuff
2:18:08 there then all of us can have access to
2:18:10 that that would be great maybe check
2:18:12 Reddit too yeah I think reddit's a
2:18:14 reddit's a thing a place to go for sure
2:18:16 scobleizer yep
2:18:18 exactly Robert
2:18:21 skoble I know Robert from the early days
2:18:23 of the web he's been doing this [ __ ]
2:18:33 forever Suzanne got it thank you Suzanne
2:18:35 that's awesome that's that that's a
2:18:38 really cool
2:18:39 contribution because yeah apparently
2:18:42 like San Francisco right now is just off
2:18:44 the hook with with all the AI
2:18:46 stuff I'm I I rarely have like I miss
2:18:49 New York a
2:18:51 lot but I know what New York was like in
2:18:55 the mid99s
2:18:56 and I think San Francisco is like that
2:18:59 right now San Francisco was sort of like
2:19:01 that in the 90s but New York was where
2:19:03 it was at in the 90s with the web stuff
2:19:05 because it was more media
2:19:07 based and San Francisco right now is
2:19:10 just crazy how careful should you be
2:19:12 using chat GPT to create a logo for your
2:19:15 business from an IP standpoint I don't I
2:19:17 wouldn't worry about that I don't I I
2:19:20 wouldn't that I wouldn't worry
2:19:23 about unless you've got unless in
2:19:25 describing your business like what I
2:19:27 would do
2:19:31 is I wouldn't worry about it like like
2:19:34 it I would say if it's stuff like you
2:19:36 know performance numbers or you know
2:19:40 proprietary product information things
2:19:42 like that that I would say be really
2:19:44 careful careful about or you know
2:19:45 personally identifiable information of
2:19:47 employees that I wouldn't touch but if
2:19:49 it's something like doing your logo on
2:19:51 there doing positioning papers that kind
2:19:53 of [ __ ] I'm doing all the
2:19:56 time but you should just basically you
2:19:58 should basically assume from a security
2:20:01 standpoint right now if you're using
2:20:03 public chat
2:20:05 GPT
2:20:07 that it's not secure you should just
2:20:10 assume that it's not
2:20:12 secure are they promising that they're
2:20:14 not going to use your data if you ask
2:20:16 them not to yeah they're promising that
2:20:18 is that actually the case I don't
2:20:19 [ __ ] know like I would assume not if
2:20:22 you really want it to be secure spin up
2:20:25 an open source locally hosted large
2:20:27 language model and then you can do all
2:20:29 this stuff on your own but they're not
2:20:31 chat gp4 right can you flash the slides
2:20:35 one last time for screenshotting please
2:20:38 absolutely all right so here I'll just
2:20:40 do this this is the first one so this is
2:20:42 the AI Salon the AI
2:20:46 salon
2:20:49 so go ahead and screenshot that bad
2:20:52 boy all
2:20:54 right and then what we're going to do is
2:20:56 we're going to go down to the money shot
2:20:58 money
2:21:03 slides okay screenshot number one and
2:21:07 this one you don't really need to
2:21:09 screenshot unless you don't know the
2:21:10 playground
2:21:12 URL um the bullet points there are just
2:21:15 that was the stuff that was announced
2:21:16 two weeks ago at the dev conference all
2:21:18 right there's your big
2:21:20 four after chat
2:21:23 GPT I know champ it's really I know I
2:21:27 know I I know I'm annoying I know it
2:21:30 took me too long to find the
2:21:32 PDFs I know people have people have gone
2:21:35 to sleep all
2:21:37 right um here's the next four screenshot
2:21:42 that again Pi you can talk to perplexity
2:21:46 you can research with po you can make
2:21:47 Bots with llamas the whole entry into
2:21:51 the open source
2:21:52 world there's your image
2:21:55 creation top four and again top four in
2:21:58 my nonh humble opinion these are the
2:22:01 ones that I think are worth you playing
2:22:03 with are there others that you should
2:22:05 play with probably these are just me
2:22:08 this is like Kyle's opinions and I'm not
2:22:10 affiliated with any of these companies I
2:22:12 get no dollars from any of them I don't
2:22:15 give a
2:22:16 [ __ ] I mean I give a [ __ ] for you I
2:22:18 don't really give a [ __ ] who gets
2:22:20 recommended or not this is sort of just
2:22:21 based on my experience um here's your
2:22:23 video creation stuff the three on the
2:22:25 left are all kind of the same thing
2:22:27 Runway ml just launched Runway TV that
2:22:30 to me feels like 1981 MTV launch when
2:22:35 MTV launched in 1981 I Want My
2:22:38 MTV
2:22:40 um nobody got it everyone made fun of it
2:22:44 they're like this is the biggest waste
2:22:45 of time like who would ever watch
2:22:48 why would you watch
2:22:50 music and it changed the game I I have a
2:22:52 feeling if it's not Runway TV it's going
2:22:55 to be something like that this
2:22:56 generative stuff is going to be big
2:22:59 there's music and audio screenshot that
2:23:01 bad boy you got your 11 Labs you got
2:23:04 your suo which will write you a song and
2:23:06 sing it for
2:23:08 you there's your 3D objects and worlds
2:23:10 blockade Labs will create a 3D World for
2:23:12 you out of a prompt Wonder Dynamics will
2:23:14 make robots out of your videos with you
2:23:18 and it'll replace you with a robot Luma
2:23:22 I forget what they do but it was cool
2:23:23 enough that I put it on there and then
2:23:26 here are people you should follow if you
2:23:27 want to get curious about this stuff
2:23:29 start following smart people here's
2:23:32 here's a trick on Twitter also
2:23:35 X if you find someone that you like like
2:23:38 Ethan
2:23:39 mullik go to their profile and you can
2:23:42 click on the the the you can click on
2:23:44 who they're following it just says
2:23:46 following click on on the word following
2:23:49 and that'll be a list of people that
2:23:50 Ethan mullik files
2:23:53 follows so if someone that you respect
2:23:56 follows someone else chances are good
2:23:58 that they're following smart people um
2:24:00 Paul rater and uh Rachel Woods is here
2:24:03 on Tik to Paul rer as the marketing AI
2:24:05 Institute David Shapiro's a certified
2:24:07 genius neuros
2:24:09 spicy
2:24:11 AGI Master open- Source dude really
2:24:14 interesting guy
2:24:16 um that's this you can subscribe here if
2:24:19 you want there's also in the lower
2:24:21 leftand corner of this screen is a
2:24:23 series that I put out you can support
2:24:25 the channel by buying a series it's a
2:24:27 series of videos on why I think that AI
2:24:29 is actually underhyped right now as
2:24:32 stupid as that sounds um and then this
2:24:35 is just some some other URLs so the the
2:24:39 this Tik Tok Channel's YouTube channel
2:24:42 has Archives of these lives I'm like two
2:24:45 weeks behind on them I know I know I'll
2:24:47 get there um the salon. a that's kind of
2:24:50 an out-of-date URL now and then uh
2:24:53 storyvine is my company and this tiny
2:24:55 URL one is a Google sheet that Theon has
2:24:59 put together every time I mention a
2:25:01 company or a person he puts it in that
2:25:03 spreadsheet so there might be cool stuff
2:25:05 in there that I don't even know about
2:25:07 all right so that's that all right
2:25:10 people listen it's getting to be that
2:25:11 time of night again I'm going to get me
2:25:13 on out of
2:25:15 here uh
2:25:20 let's see what do we want what image do
2:25:22 we want oh that's a good
2:25:25 one that's a cool
2:25:27 one Silver Fox thank you for those Roses
2:25:30 that means a lot I appreciate that um
2:25:34 Christopher Ramsey Center Rose thank you
2:25:36 sir appreciate that um to my mods thank
2:25:40 you so much for everything you do um
2:25:43 please join the salon if you resonate
2:25:45 with the values if you don't it's not
2:25:47 going to be your peeps you're not going
2:25:49 to dig it if the values are like oh that
2:25:52 sounds good great come on in be generous
2:25:56 be
2:25:58 empathetic don't be a douche Sherry d s
2:26:02 fire thank you Sherry
2:26:04 D all right let's get on out of here Be
2:26:08 not Afraid great to see you Firefly
2:26:10 peace out Sher D thank you steo good to
2:26:13 see you as always forever hooked Tobias
2:26:15 Tobias figured out how to get his name
2:26:17 came
2:26:18 back thank you thank you Lauren
2:26:20 appreciate that thanks again for all you
2:26:23 do Kyle thank you be not afraid I
2:26:25 appreciate that Silver Fox good to see
2:26:27 you I am China steo Earth rocker Joe
2:26:30 mama Mary Mary it's it ASU it's so nice
2:26:34 we've got we've got some newer regulars
2:26:36 in here that are showing up you know
2:26:38 night after night it's really good to
2:26:39 see you all I I do appreciate this like
2:26:41 more than you know cuz cuz for me Tik
2:26:44 Tok has turned from like a thing where I
2:26:47 process my thoughts out to this
2:26:49 collaborative community and it's really
2:26:51 really amazing so I appreciate all of
2:26:53 you all right enjoy the rest of your
2:26:56 Saturday night or your Sunday morning um
2:26:59 watch the F1 race tomorrow it's in uh
2:27:03 where is it Dubai I think it's in Dubai
2:27:06 it's the last race of the Season spoiler
2:27:08 alert Max ver Steppin on
2:27:11 poll and he's probably going to win
2:27:13 again so all right everybody uh have a
2:27:17 good good night I'll talk to you