aiLL 8/11/23 - The AI Renaissance Is Here: How To Ride The Wave

Description

Chapters

Transcript

0:01 thank you
0:32 good evening everybody welcome welcome
0:35 welcome
0:36 to the AI learning lab the official AI
0:39 learning lab
0:41 not that stupid scammer with two L's I
0:44 reported him to tick tock and they said
0:46 oh he's fine
0:48 so I reported him again
0:50 they still said he's fine so
0:53 I don't know what to do and I don't
0:55 think you can you can report him I think
0:57 I have to report him but they're not
0:59 dealing with it
1:01 Dr J what's happening how are you
1:03 welcome welcome welcome
1:06 happy Friday night everybody
1:08 let some folks get gathered up in here
1:11 while we get rolling along
1:14 rolling a long rolling rolling rolling
1:20 howdy howdy Rebel
1:23 what's happening
1:26 uh oh my shirt's gonna be a problem
1:30 uh Hey Tyrant truth what's happening
1:34 Claudia Claudia
1:36 little Ravens sartinist
1:40 I can't I can't read against my shirt
1:42 I'm gonna go change tell everyone I'll
1:45 be back tell everyone it's gonna be
1:46 really exciting it's gonna be so worth
1:48 the wait they're gonna be so excited
1:50 when I come back
2:24 you want it out
2:33 foreign
2:37 [Applause]
2:42 exciting and worth the wait that's it
2:45 that's exactly it all right welcome
2:47 welcome welcome uh do me a favor I got
2:49 the slinky patent uh t-shirt on which
2:52 it's gonna be hard to read I'll just
2:54 have to read up here
2:56 um feel free to share the live hit the
2:59 hearty Hearts Let's get this let's make
3:02 it seem like it's really exciting in
3:03 here like you know
3:05 like there's a gaggle of twerking
3:08 hotties
3:11 and and then they come and see me
3:14 Amelia's wife what's happening
3:17 8 37 time check
3:23 this is uh just the AI learning lab this
3:26 is the official channel the one with 1l
3:28 in lab
3:31 scammers the worst
3:33 and isn't Tick Tock the worst for not
3:35 taking them down twice now
3:38 hey
3:41 encourage more viewers to share yeah
3:43 more viewers share share share all right
3:46 we got some folks coming in here welcome
3:48 welcome if you're new here my name is
3:50 Kyle Shannon this is the AI learning lab
3:51 we talk about all things AI well not all
3:54 things AI most things Ai and a lot of
3:56 things not AI
3:57 um
3:58 I'm an entrepreneur I'm a Storyteller by
4:01 training I've got a video technology
4:03 company called storyvine
4:06 we've been around for 11 and a half
4:07 years uh we do a lot of work in
4:09 healthcare and Pharma
4:11 and uh I was an entrepreneur in the
4:14 early days of the World Wide Web
4:16 and when I stumbled upon this here
4:18 generative AI thing about a year and a
4:20 half ago I thought huh well uh
4:25 and I have Xfinity so if anyone from uh
4:29 from Xfinity and Comcast is on the the
4:32 live go screw yourself could you make
4:35 your internet work did I pay ridiculous
4:37 amounts of money for it I I just wanted
4:39 to work like it's not thank you very
4:42 much the the the uh as a user I would
4:46 like my internet to um not suck
4:52 and why I keep it up on my live the
4:55 whole time there if you're new to
4:57 generative AI if you've not played with
4:58 Chachi PT go to that top address that's
5:02 that's the official chat gbt website I'm
5:04 not affiliated with it I don't make any
5:06 money off this I don't give a [ __ ]
5:08 if you go or not but
5:11 I think it's important for everyone to
5:12 actually try we pay for both Xfinity and
5:15 T-Mobile home because yeah yeah I know I
5:19 know
5:19 yeah it's just it's ridiculous at this
5:22 point anyway go there
5:24 um is my camera dirty I think my camera
5:26 might be dirty oh no wait
5:32 no place is gone
5:35 it's gonna be good
5:37 it's gonna be so good
5:39 all right
5:42 I was trying to make my camera focus
5:43 better
5:44 hang on hold hold
5:48 um
5:49 that help
5:52 not really
5:53 oh I know what it is
5:55 that was too bright ah it's better there
5:58 focus and lack of executive function so
6:05 I'm gonna bring it down thank you for
6:06 the rose Teriyaki and something
6:10 um the double roses that was sweet
6:12 um
6:13 okay if you're new to chat GPT go to
6:16 that top one if you have uh
6:20 uh if you've played with chat gbt all
6:22 these other ones are things that I think
6:24 are worth playing for or playing with
6:27 um Bing is from Microsoft that's gpt4
6:30 Claude is a competitor
6:32 Google's barred is okay used to suck
6:36 it's getting better so it's worth
6:38 playing with now lab stop perplexity
6:41 that's meta's llama 2 models that
6:43 perplexity is letting us all play with
6:44 for free so you can go see what the
6:46 Llama models are all about
6:48 um and then you got PI from inflection
6:51 and you got Poe from quora all of these
6:54 have some interesting thing about them
6:55 that are worth playing with so go there
6:57 do that with with those things
7:01 the next thing to do and lots of good
7:05 ideas
7:06 um there are tons of other you know
7:09 there's prompt libraries there's prompt
7:10 base if you want to go pay for prompts I
7:12 would definitely not recommend that but
7:14 maybe you come up with a good prompt and
7:15 go sell it on prompt Basin then you can
7:18 make money
7:21 we also make fun of those people on this
7:23 channel uh
7:25 um
7:28 uh prompt book
7:30 openart.ai's prompt book that will teach
7:33 you about how you practice generation
7:35 tools like I have on the bottom here mid
7:36 Journey Adobe Firefly and Leonardo's
7:39 stable diffusion so stable diffusion or
7:41 Leonardo AI Leonardo is kind of the web
7:44 fancy front end to stable diffusion so
7:47 that's a really good document
7:50 openart.ai prompt book really good
7:52 document so go there look at that you
7:56 will you will Marvel at how much these
7:58 image models understand it's like a it's
8:00 like an art history lesson and then you
8:02 realize
8:04 you don't just you don't just have to
8:05 learn it if you learn it you get to go
8:07 make stuff that looks like that it's
8:09 amazing absolutely amazing okay
8:12 so
8:14 the other thing we do here
8:18 we drink coffee
8:21 we drink coffee late at night because
8:23 you know
8:25 add and stimulants
8:28 they're kind of counterintuitive
8:31 um
8:33 I answer questions so if you have
8:35 questions about AI pop them in the
8:37 comments below I'll try to answer them
8:39 all later in the live it'll get really
8:41 frustrating because people will be like
8:42 he's not answering my question I find
8:45 this very frustrating he he claims to be
8:48 live everything seems to be live but
8:49 he's not answering my question I've
8:51 asked it 732 times and he still hasn't
8:53 answered my question so it's clearly not
8:55 live so I think he's the bot I think
8:57 it's a bot
8:59 pretty sure he's the bot
9:00 you haven't seen that he's the bot
9:03 did someone confirm that he's the bot is
9:05 he is he a chat GPT five bot is he is
9:08 this a simulation are we is we're living
9:10 in a simulation
9:12 right you see what I'm saying do you
9:14 serve this show exactly Emilio's life
9:17 exactly exactly we need a professional
9:20 around here
9:25 all right people are winning gift or
9:28 badges for gifting me so thank you for
9:30 gifting me your winning badges
9:31 congratulations on the badge winning
9:34 I don't know uh what did I miss Kyle we
9:37 just started Clint just getting rolling
9:39 hmm hmm oh you missed me dropping my
9:42 entire rig my light went on the ground
9:44 the little box I have my light on went
9:46 on the ground
9:48 I tripped I changed my shirt you missed
9:51 that
9:52 I did that full camera this was full
9:54 frontal
9:56 learning lab
9:58 you missed that sad because that was
10:01 there was Applause I could hear it
10:05 uh thought on AI and religion
10:10 um
10:14 I would think that you could do some
10:17 incredible research on lots of religions
10:19 with AI so if that was something you
10:23 wanted to learn about great
10:25 if I think about you know if you're
10:27 asking like do I think Ai and religion
10:29 religion
10:31 go together no I mean to the extent that
10:34 religion is language sure these are
10:37 large language models
10:39 religion is a large language model
10:41 humans are a large language model so
10:44 we're all in the same neighborhood
10:48 about it and I certainly don't want to
10:50 talk about it here
10:52 oh I don't even want to talk about it
10:54 here okay
11:00 well
11:03 isn't Mr Shannon very special tonight
11:06 acting all high and mighty like religion
11:08 doesn't matter
11:11 who could be influencing that
11:14 that was somewhere between church lady
11:16 and and uh
11:19 the uh the uh I'm good enough I'm smart
11:22 enough Stuart Smalley and doggone it
11:25 people like me
11:28 oh lordy lordy lordy lordy
11:34 uh Silver Fox I don't know what phone
11:36 settings I would check
11:38 uh
11:39 uh if you know what phone settings I
11:41 should check to see if my internet is
11:43 laggy here
11:45 um I think I think it's just my Xfinity
11:47 because our TV is going out all the time
11:48 we have TV over internet
11:50 Friday night date night
11:53 and I'm so sorry
11:55 that's the best welcome welcome tap tap
11:59 tap thank you Rebel
12:01 all right let's get some questions
12:03 golden shall we
12:05 let's do it Dal howdy y'all that's Logan
12:08 Logan's here we got some regulars I love
12:11 this we've had some regulars here all
12:13 week it's pretty cool got a little we
12:15 got a little Community going it's a
12:17 little Community neighborhood Playhouse
12:18 we play we play we have a good time we
12:22 have a good time
12:24 thinking about taking up crocheting why
12:26 not why not
12:28 is there some rule says I can't crochet
12:30 no
12:35 oh someone today I did a I did a
12:38 LinkedIn live today uh because I just
12:40 can't get enough of live media
12:42 apparently and uh one of the guys from
12:45 the salon said that every time we do an
12:48 AI Salon he goes and eats so now when he
12:51 when he listens to me lives to do
12:53 something live He's he I trigger him
12:55 it's a pavs loves uh Pavlov's response
12:59 where when he listens to my lives he
13:01 gets hungry and he made pancakes so so
13:03 he was he was asking me questions about
13:05 AI while he was making pancakes
13:09 oh that's funny all right if you are
13:12 recording with your smartphone it needs
13:15 to have a two to three foot radius away
13:18 from Tech
13:20 if you are recording with your
13:23 smartphone
13:25 I don't know what that means oh like
13:27 like
13:28 two to three feet away from my computer
13:30 yeah that's not gonna happen
13:34 it's gonna be glitchy then
13:36 Kyle Shannon they're man
13:38 yo Troy Decker shaking shaking shaking
13:41 working the service desk watching for
13:44 levity oh nice nice Vicky yeah serve it
13:47 up take care of the people and
13:48 occasionally listen to my Stupid humor I
13:51 like totals
13:54 I drink Milo's sweet tea but you do you
13:57 Kyle I am
14:00 you have the sweet tea I never got into
14:02 the sweet tea I like I like unsweetened
14:03 tea
14:04 and it's weird because for most of my
14:06 life I like super sweet everything but
14:08 tea for some reason
14:10 doesn't do it for me I drink that with
14:12 no sugar
14:14 same with my iced coffee no sugar a
14:17 little bit of milk no sugar
14:20 you know why because I'm sweet enough
14:24 don't bother responding what did you ask
14:26 Luna stick
14:34 Luna asked me a question is now very
14:36 very very very very very very very very
14:38 [Music]
14:40 disappointed that I didn't answer
14:42 I don't see your other question Luna
14:45 stick
14:46 I think you might be a Luna stick you
14:49 know what I mean
14:50 or maybe you're not talking to me you
14:52 talking to someone else you talking to
14:54 me you talking to me we're walking here
15:05 um
15:06 even if you don't have questions if you
15:08 want me to demo something I'm happy to
15:10 do that so so feel free to do that it's
15:12 been a roller coaster of excitement
15:13 since the live started it really has
15:16 it's something's off tonight something's
15:18 just weird
15:21 Luna stick what what
15:23 don't bother responding okay should I
15:26 not respond to you it sounds like I
15:27 shouldn't respond to you okay I'll
15:29 definitely not respond to Luna schtick
15:32 it's shtick get it be a comedian Luna
15:36 stick is wait so wait were you actually
15:39 saying don't respond or was it like an
15:41 ironic sarcastic don't respond where you
15:43 really want me to respond and by
15:45 responding I'm actually not responding I
15:48 think that's what you're going for
15:51 Luna stick it's like Moon comedian
15:55 huh Moon comedian of the Catskills loner
15:59 stick
16:03 foreign
16:07 [Laughter]
16:10 coherently for an hour or two
16:13 oh Luna stick what are we gonna do with
16:16 you what are we gonna do with Luna stick
16:19 what are we gonna do with Luna stick
16:22 all right
16:24 scheduling budget meeting to discuss a
16:27 real quality live screening setup for
16:29 the boss yeah you got you gotta talk to
16:31 uh procurement we gotta get some we
16:34 gotta get some uh some gear going in
16:36 here
16:38 we'll get an actual microphone we'll get
16:40 some actual internet
16:44 apparently I need a PC because you can't
16:46 do Tick Tock live on a Mac because you
16:50 know
16:51 I don't know whatever
16:53 Rebel I'm dizzy Luna stick when you have
16:56 wait what have you found to be the most
16:58 enlightening way to utilize AI
17:02 um
17:03 enlightening way
17:06 for me as evidenced by this live it's
17:09 building Community around people
17:12 learning it
17:14 um
17:19 I I
17:24 just wrote a piece on LinkedIn called
17:28 um chechi BT just gave me superpowers
17:31 and
17:33 it was just it was yet another
17:34 experience where so so what's been
17:37 happening with with me with Chach EBT if
17:39 I sort of let my brain get out of the
17:41 way
17:42 and just give in to okay
17:45 I've got an assistant that's that's way
17:48 faster and just as creative as I am if I
17:50 just kind of give into that
17:52 premise
17:54 um
17:56 then all I have to do is is give it a
17:58 little bit of direction here and there
18:00 like I'm just sort of steering it and
18:01 letting it do the bulk of the work
18:04 and I'm finding that I'm over delivering
18:07 on my promises
18:09 and I'm coming up with ideas that I that
18:11 I absolutely wouldn't have pursued
18:13 without this stuff
18:16 um and in that piece I called these
18:18 things
18:19 um ability amplifiers
18:21 or Humanity amplifiers and someone said
18:24 on the Live this week
18:26 um I was talking about large language
18:27 models and llms and what they are and
18:30 how they work and and and they they they
18:32 put in the comments
18:34 um humans are large language models and
18:36 I thought oh they are
18:39 and that's why these things are so funky
18:41 and weird because
18:44 they were trained on the words that we
18:46 put out on the internet right like so
18:49 we've just for the past three and a half
18:51 four decades been putting [ __ ] on the
18:53 internet
18:54 and these guys all went and crawled all
18:56 of those words and found a way to stick
18:59 them in these tools in a way that
19:01 they just kind of
19:04 are a mirror back in our face of what we
19:08 have
19:09 created collectively you know as a pile
19:13 of people a bunch of bunch of meat bags
19:17 using these large language models
19:18 spewing stuff out onto the internet
19:22 and so so they really are kind of a
19:24 reflection of who we are right and so
19:26 they're they're kind of these Humanity
19:27 amplifiers and it's good and bad right
19:29 and then you know I was talking today on
19:31 the on the LinkedIn live and
19:34 um
19:35 um you know one of the guys was talking
19:37 about
19:38 um
19:39 well these things hallucinate and that's
19:41 really bad and you know we got to find
19:42 ways to make them not hallucinate and
19:44 then he was talking about but when I
19:45 work with human beings like human beings
19:48 lie all the time and they take time off
19:50 from work and they don't show up and
19:51 they're late and they they procrastinate
19:54 and they do all this stuff and they also
19:56 screw up and they also don't research
19:58 things really well I was like oh yeah
20:00 that's right like we're expecting
20:02 Perfection out of these things but what
20:04 they're doing is [ __ ] astounding like
20:06 even if they get it you know 20 wrong
20:09 it's 80 percent right
20:11 like that and it's like what human can
20:14 do that none
20:15 have you ever tried to hire a marketing
20:17 copywriter just out of the blue
20:19 about the tenth one that you fire you're
20:21 like this this might take a while
20:27 you know you don't experience it with
20:29 that so so for me
20:33 there's kind of a there's kind of a
20:35 freedom that this that these things are
20:37 providing
20:40 Freedom's not the right word well
20:43 yes
20:46 it's like creative freedom but
20:48 it's not so much creative freedom
20:50 because I feel like I already have
20:51 creative freedom just do whatever the
20:53 [ __ ] you want to do
20:58 the word that just popped in my head was
21:00 was uh creative lubricant
21:03 these things are creative like like it's
21:05 just like there's less friction there's
21:07 no friction
21:08 writer's block doesn't exist
21:12 um
21:12 you have an idea that that historic if I
21:15 have an idea that historically I would
21:17 have thought
21:18 oh but I would have to research that and
21:20 that's if it felt like too much work I
21:24 would just throw it out I'm like I'll
21:25 come up with some other idea
21:27 and now I have that idea
21:29 and there's no friction if I don't need
21:31 to do research anymore I can just go ask
21:32 this thing tell me about this time in
21:35 history oh that's kind of oh you're
21:36 right I remember that from school
21:39 and I'm off
21:42 I don't know that that really answers
21:44 your question
21:45 um
21:48 the thing that I find most inspiring
21:52 is when
21:53 someone that comes on one of these lives
21:55 or someone that I'm talking to in the
21:57 world that's like oh yeah I've used chat
21:58 GPT yeah yeah yeah yeah yeah yeah it's
22:01 remarkable yeah yeah
22:03 thanks Lauren thank you for the Braves
22:05 yes
22:06 we'll smell it sweet
22:09 um
22:11 it's just like I've got Tourette's
22:15 I don't but
22:16 [Laughter]
22:20 um
22:22 when when
22:24 I when I work with someone or I'm
22:26 talking to someone and then I say oh go
22:29 try this and they actually go try Chachi
22:32 BT and they put something in
22:35 and and they have their Kevin McAllister
22:37 moment there
22:39 their home alone moment
22:41 that for me is the most inspiring thing
22:44 because from the outside these things
22:47 don't look that impressive it's like oh
22:48 yeah yeah it's it it's a chat bot you
22:51 put in words and words come out I've
22:53 seen those before you don't need to
22:55 educate me I've been using the the
22:58 internet for nearly a decade okay
23:02 and then they use it and they're like ho
23:04 ho holy [ __ ] did you know it did that
23:07 yeah that's why that's why I have a tick
23:12 tock you know so
23:14 I can't stop thinking about it
23:18 um
23:18 so that's it for me I mean it's it's
23:21 just and and the the other thing that's
23:23 just mind-blowing to me these things
23:26 are
23:27 are weak pieces of [ __ ] compared to what
23:31 they're going to be in a year in two
23:33 years and three years and five years I
23:36 can't even get my head around how
23:39 different these things are going to be a
23:41 year from now
23:43 because if you go back a year
23:45 Chachi PT didn't exist gpt3 did you
23:48 could go to the playground and you could
23:49 play with the gpt3 API
23:53 but a year ago chat GPT didn't exist and
23:56 the image the image generator is like
23:58 stable diffusion and mid-journey were
24:00 just starting to make competent images
24:02 so we're like nine months or a year into
24:07 these things producing competent results
24:10 right
24:11 so these are like the
24:14 version 1.0 right this is like Netscape
24:17 Navigator 0.9
24:20 B
24:21 remember the early days of the World
24:23 Wide Web when a new browser came out
24:25 everyone was like oh there's a new
24:26 browser this one's got a back button
24:32 like that's the stage we're at with this
24:34 [ __ ] and it's already as remarkable as
24:35 it is so so
24:38 every time someone has that light bulb
24:41 moment with this stuff I get really
24:43 excited a because just [ __ ]
24:45 Humanity's cool and when someone has
24:47 their mind blown and you get to witness
24:48 that it's a remarkable moment
24:52 but then I know like this is sort of red
24:55 pill blue pill stuff like once you
24:57 actually play with this and and have
24:59 that moment There's No Going Back
25:02 like you get you can't get how important
25:06 this is from the outside and the minute
25:08 you're on the inside you can't unlearn
25:11 it and you can't unthink it and you
25:13 can't unimagine it and you're just like
25:16 oh [ __ ] what do I do now
25:19 and you're like but at least it can't do
25:21 what I do and then you try that and it
25:25 can and then you're like oh this is
25:27 really a problem
25:29 but you can unsee it you can't undo it
25:31 you can't unlearn it they're not gonna
25:33 uninvent it
25:36 so
25:39 that's that's the most enlightening
25:41 thing for me
25:44 the 27 things I just said there could it
25:47 be Satan wow wow wow wow wow look at Mr
25:51 Shannon live on The Tick Tock owned by
25:55 Satan himself terrific
26:02 all right
26:04 let's go to some questions shall we by
26:08 the way if I haven't answered your
26:09 questions yet because I haven't gotten
26:10 to them yet talk to the people that have
26:12 been here for a while they'll straighten
26:14 you out and by the way if you're someone
26:16 who's in here who made an account called
26:18 AI learning lab where you put two L's in
26:20 front of the word learning you're an
26:22 [ __ ] and I hate you
26:24 and by the way yes everyone I know it's
26:26 happening and I told Tick Tock they were
26:30 cloning my account and they said oh we
26:33 don't see anything wrong with that
26:38 they're like we've reviewed it and we
26:40 think everything was fine why do you ask
26:44 and so I reported it again and they're
26:46 like
26:47 um sorry let me adjust my glasses
26:49 um no yeah looks good to us that's
26:51 everything's good
26:53 carry on AI learning lab
26:59 bitter you betcha
27:03 all right if I weren't married I'd be
27:07 courting you love your content and width
27:09 thank you very much I am also married
27:11 but I appreciate the kind words that's
27:13 very nice
27:15 Kyle and character is the best thank you
27:17 Troy try some cinnamon with your coffee
27:19 and milk huh you know cinnamon cinnamon
27:23 and coffee in me it's a funny thing I
27:26 love cinnamon love cinnamon rolls
27:29 a little Cinnabon to drive those
27:31 cavities home
27:33 um
27:34 and then cinnamon and coffee I'm 50 50
27:37 on it sometimes I'm like oh that's quite
27:39 refreshing and other times I'm like who
27:40 put dirt in my coffee
27:43 it's just weird
27:45 tasted funny
27:48 but I am a fan of the cinnamon
27:51 I had choppy connection but got a Wi-Fi
27:53 extender and now it's all smooth yeah we
27:55 got a Wi-Fi extender too because we had
27:57 a we had a shitty router before so then
27:58 we got a good router with a Wi-Fi
28:00 extender and then an extra extender
28:02 upstairs and then all three of them suck
28:06 and then I think Xfinity actually
28:07 changed some protocol or others so we
28:10 probably actually now have to go buy a
28:12 new router and like because the Xfinity
28:14 guys are like well everything's looking
28:16 good here but you know your router it's
28:18 not one that we put in so go [ __ ]
28:21 yourself
28:21 does that work for you all right I'm
28:24 gonna go get my van now and eat a
28:26 Twinkie
28:30 please fill out this customer survey
28:32 hope you're happy
28:33 zero zero zero he was nice three
28:39 all right I asked what has been the most
28:41 enlightening way oh yeah I got it what
28:44 what
28:45 is clip clip is
28:51 I asked a question but yes Moon comedian
28:53 now that's cool doing a stick Luna stick
28:57 uh
29:00 I drank coffee and tea with no sugar
29:03 also it's the only way to do it Winston
29:04 sometimes a little cinnamon and coffee
29:06 but that's all see a little cinnamon I
29:09 don't you know what I haven't had
29:10 cinnamon coffee in a while I could go
29:12 get it now should I do that
29:16 sure sure Kyle you can do whatever you
29:19 want it's your life I might go get
29:21 cinnamon and put it in my coffee just to
29:22 just to try it out I'm gonna do it all
29:24 right 80d moment hang there uh while I'm
29:27 gone share the live tap the screen uh
29:30 look at my [ __ ] on my counter I'm
29:32 shelves
30:15 all right
30:18 we have sprouts sprouts madagascan
30:20 cinnamon Madagascar cinnamon
30:23 let's see
30:25 harvested
30:27 from the Youth of no that's not true
30:29 it's not it's not child labor definitely
30:31 not
30:32 it's organic
30:34 all right put it in the coffee
30:36 lovely oh
30:39 I gotta stir it in uh I didn't bring a
30:41 spoon that's a problem Oh this is gonna
30:44 be a problem
30:46 huh well it seemed like seemed like a
30:50 good idea at the time here goes nothing
30:53 um
30:56 it needs to soak in a little I told you
30:59 it's like coffee with dirt on it all
31:00 right
31:04 it's it's all right everything's gonna
31:06 be okay all right can you show us
31:09 just the basic turn my portrait selfie
31:13 into an avatar maybe simple no good
31:18 yeah sure let's do that
31:21 where do we want to start
31:24 let me go
31:30 um
31:31 where am I gonna start where am I going
31:32 to start
31:34 um
31:44 dreams
31:50 okay so here's a bunch of images from
31:52 Kyle Shannon dreams from my art project
31:54 so these are all stable diffusion
31:57 that's a nice one should I animate that
31:59 one
32:01 that was a little disturbing it is a
32:03 Friday night so so do we have adults
32:06 here
32:07 um
32:08 that one's kind of cool that could be a
32:10 cool one although the problem with guys
32:12 with goggles like that is the the
32:14 animating things assume all eyes blink
32:16 they don't they're not smart enough to
32:18 know that those things shouldn't
32:20 actually blink
32:23 um oh that's a fun one that's my that's
32:26 my my Elvis
32:28 my Elvis in his heavy ears look
32:37 [Music]
32:38 that one's kind of cool all right that
32:41 one's groovy
32:42 look like some sort of androgynous David
32:44 Bowie looking
32:46 dude or do that oh and I said it's
32:48 called androgynous robot all right so so
32:51 what I'm gonna do so so you create your
32:54 image and wherever it could be a
32:56 photograph of you it could be something
32:58 you do in stable diffusion something you
33:00 do in mid Journey
33:02 and then
33:05 oh wait that's Chrome let's get out of
33:07 Chrome
33:12 and then we're gonna go to dash ID well
33:16 let's see do we want to go
33:21 yeah yeah we'll go to we'll go to uh
33:24 we'll go to dash ID okay
33:26 so
33:29 D Dash ID
33:31 oops that was d-od
33:34 the dod very different than the did the
33:38 dod has very different kind of AI it
33:40 will hurt you
33:42 the ID will make funny little cartoons
33:45 anime
33:48 so you go here you do have to have a
33:50 subscription for this
33:53 um and so so when you get here you've
33:56 got your video library so that's what
33:57 I'm on now so these are all videos that
33:59 I've made and then you can do create a
34:01 video
34:02 and there's a couple of things to the
34:04 interface so you've got you know the
34:06 sort of you know that that's what it's
34:08 gonna the video is gonna be of and then
34:11 you've got all these different avatars
34:14 you know some of which they made and
34:15 then some of which you can you know you
34:17 can upload so I'm gonna upload that one
34:21 so I'm gonna go find androgynous and Raj
34:26 and draw
34:33 there it is
34:34 so we're gonna bring that crazy dude
34:37 dude or doodad in here
34:40 and then over on the right hand side you
34:44 can either upload an audio file so you
34:46 can just record an audio of you talking
34:48 you can use something like 11 labs to
34:52 synthesize Voice or you can have did
34:55 synthesized voice so we need a little
34:57 script right here so let's go let's go
34:59 get a script let's go make a script
35:02 we're going to go to chatgpt we're gonna
35:05 go make money with jadri PJ Bob tell him
35:09 what he's won
35:11 um
35:20 write me a short
35:24 monologue
35:25 of an androgynous cyborg
35:30 lamenting
35:33 the last day they could see the sun
35:40 this is very dark very dark and very sad
35:42 dark and sad
35:44 in a world where the line between
35:47 machine and man has blurred
35:49 I stand as a fusion
35:52 it's a little long but whatever I'll
35:54 stop that we'll just take half of it
35:56 let's see the day the sky was painted as
35:59 I gazed upwards I felt here but we'll do
36:01 this we'll do that that whole paragraph
36:03 okay so we got ourselves a little script
36:06 we're gonna hop over to dash ID we're
36:09 gonna pop our little script in there
36:12 so there's our script and then we're
36:14 gonna go grab
36:16 a voice
36:18 let's see let's grab a
36:28 let's let's do let's do an international
36:30 one so this is English United States
36:33 let's do English South Africa that's a
36:36 cool accent
36:37 and then we'll do Leah we'll we'll do it
36:39 we'll do a female voice and
36:42 see how that sounds all right and then
36:44 basically you just say generate so you
36:45 got a picture you got a voice you got a
36:47 script or an audio file
36:50 and then you pick a voice if you've got
36:52 a script in here or if you're in 11 Labs
36:55 you do all this stuff over there and
36:57 then you just hit generate video
36:59 and this is gonna cost me three credits
37:02 it's gonna be 34 seconds long
37:06 and then it's off doing its thing and
37:08 I'm gonna try to stir my cinnamon into
37:10 my coffee well well it's rendering
37:13 come on Cinnamon you can do it
37:15 see the cinnamon and the iced coffee was
37:17 a bad call I'm sure those of you that
37:20 are cinnamon connoisseurs are out there
37:22 going you don't put cinnamon on cold
37:24 coffee you dumb dumb all right
37:27 what was her name I forget what her name
37:29 was Laura Tara
37:32 in a world where the line between
37:33 machine and man his blood I stand a
37:37 fusion of both an androgynous cyborg
37:40 I remember the last day I saw the sun
37:42 it's golden Rays dancing upon my
37:45 metallic skin warming the human parts of
37:47 me
37:48 that day the sky was painted with Hues
37:51 of orange and pink a masterpiece that no
37:54 technology could replicate
37:56 as I guess upwards I felt a connection a
37:59 Tethered to the world that burped me and
38:02 the humanity that still pulsed within my
38:04 circuits
38:06 that's pretty good
38:08 the animation was good
38:10 the voice was good that's a keeper
38:12 that's a keeper I think I might stick
38:15 that one up on on the old Twitter
38:17 the the service formerly known as
38:20 Twitter
38:21 well isn't that special Elon Musk just
38:24 decides one day he's going to turn it
38:25 into X well what's that assemble of well
38:28 let's add a symbol of what could it be
38:29 is it the Lord oh no
38:35 I hope that gave you what you were
38:37 looking for Tobias
38:39 um if not let me know where can I find
38:41 Future Ai and how can I use it
38:44 um I don't know what you mean future AI
38:48 where can you find so if you're if
38:51 you're just trying to get if you're just
38:54 getting started with generative AI with
38:57 chat GPT go to the top one
39:00 chat.openai.com
39:03 go to that website make an account
39:07 if they let you do it in your country if
39:10 they don't
39:11 if you're not American go to bing.com
39:13 they should let you do that one
39:16 but I don't know
39:18 and then go to prompts.chat and
39:20 prompts.chat will teach you what these
39:24 things actually are
39:26 they're they're they're they're very
39:28 simple conceptually right it's just a
39:30 little it's just like a Google search
39:32 bar but what you what you type in there
39:35 and how you type in there is
39:37 way more sophisticated and what it what
39:39 it gives you back is way more
39:41 sophisticated than a Google search way
39:43 more sophisticated
39:46 um and I can play with some of that
39:47 stuff if you want also last night you
39:49 asked if the food at Casa Bonita was
39:51 good it was wow that's amazing the food
39:54 at Casa Bonita has been historically
39:57 horrific like like you know how those
40:00 restaurants that are so bad like you
40:02 just gotta go there to like sample the
40:06 Badness and it's funny
40:07 this was apparently like yeah no don't
40:10 go there
40:11 like like the botulism is real
40:16 but apparently Trey and uh Trey and
40:19 what's his name from South Park they
40:20 brought in a real chef and apparently
40:22 it's good so that's cool
40:24 that's awesome
40:25 um where can I find Future Ai and how do
40:27 I use it just talked about that I feel
40:29 the same
40:31 um how do you learn it show examples all
40:33 right so let's do a little let's do a
40:35 little part give me the cowboy hat
40:36 thanks for the cowboy hat though all
40:38 right let's let's go play with chat
40:39 chibi chat GPT I'm gonna start with what
40:42 I call parlor tricks okay so just just
40:44 to orient you a little bit here we are
40:46 at Chachi BT this is my prompt history
40:48 down here
40:49 over here is the window it looks you
40:52 know not all that impressive you have a
40:54 little chat box there
40:56 you have uh GPT 3.5 you have gpt4 with
41:01 some extra options here I'm going to
41:03 reload this
41:04 because I think
41:07 what's weird is
41:09 huh they're changing the interface all
41:11 the time for a while they had some some
41:13 extra like prompt suggestions down here
41:16 but they took them away
41:18 all right so here you are so
41:23 sitting underneath all of this is
41:27 essentially as much of the public
41:29 internet as they could crawl right so
41:31 they had their these Bots go out and
41:33 crawl all the public internet
41:35 and so it kind of knows everything it's
41:38 not very smart like it it
41:42 needs a lot of Direction so it's it's
41:44 like it's like someone just gave you a
41:46 really really amazingly well-educated
41:49 intern that doesn't know any how to do
41:52 anything and you got to tell it
41:53 everything
41:54 so one of the one of the things you can
41:56 do is just start with simple things that
41:59 if you're not a writer and even if you
42:01 are a writer these can be hard it's like
42:03 okay like write some write some verses
42:05 of a poem or write a song right so
42:10 write me a country song
42:14 about
42:19 getting dust in my eye
42:22 driving down the country road
42:25 and when
42:27 my vision cleared up there was a vision
42:30 of a chubby black and white dog named
42:33 champ
42:37 all right and then so there it is
42:40 writing my song right so so this is GPT
42:43 3.5 which is very fast
42:45 and it it understands things like verse
42:48 verses and choruses it understands
42:50 rhyming Well I packed my truck all dusty
42:53 and rusted headed out on a road just
42:55 feeling so trusted with Dustin
42:58 may I my buddy and my guide Dustin may I
43:02 it came up with the name of a dude
43:04 cruising down the country road nothing
43:06 to hide
43:08 so so there's a there's a you know
43:11 there's there's a song and it Rhymes and
43:13 that's pretty good but I could say
43:15 um
43:19 can you rewrite that in the style of
43:22 Brandy Carlisle
43:26 so it knows who she is it knows the word
43:30 she uses right
43:33 um
43:35 packed up my truck rust and all oh it
43:37 Still's got Dustin okay
43:39 um
43:40 I hate Dustin's name can you come up
43:42 with something better
43:45 so you can also do things like that like
43:47 if it gives you stuff you don't like you
43:49 can tell it you don't like that and it
43:50 will give you more so let's see my
43:52 faithful companion
43:55 we got rid of the name
43:57 all right so there's the song so that's
44:00 cool so then I could say
44:01 um
44:04 um create a table that compares
44:11 the songwriting and lyrics of John Hiatt
44:14 and Brandi Carlile
44:20 and so they're so there's your table
44:22 comparing these two sort of things and
44:25 this could be any two concepts this
44:26 could be two philosophies two religions
44:28 two football teams two anything or for
44:30 anything or it could be just one massive
44:33 table just explaining I don't know all
44:35 the turning points of all of the
44:37 different world wars right so
44:39 um it can do things like that the other
44:41 thing that it can do is
44:43 um
44:44 write me a python application that will
44:47 generate
44:49 a country song in the stylings
44:54 of John Hyatt and Brandi Carlile
44:57 based on some interesting inputs
45:02 that a user will give you
45:07 and so it explains what what the what
45:11 the program it's going to write is so so
45:14 complete
45:16 uh
45:18 uh so it says that creating a complete
45:19 python application to generate country
45:21 songs and the stylings of artists like
45:24 John Hiatt and Brandi Carlile will be
45:25 quite complex it involves natural
45:27 language generation however I can
45:28 provide you with a simplified example
45:30 blah blah blah so there's all that stuff
45:32 now is it calling open AI I don't think
45:35 it is
45:36 no so I'm gonna say
45:38 um
45:42 why don't you call the open AI API
45:46 so we can make it easier to do the
45:48 natural language processing to create
45:50 songs in the style of these artists
45:56 and then now it's writing python code
45:59 that now has the API in it right replace
46:01 your openai key so so so this is so
46:05 that's just some like really quick
46:07 little stuff that I've done so so
46:10 what can you do with it you can do
46:12 almost anything
46:14 and anything that involves language and
46:16 so so rather than than sort of think
46:18 about well how can I use AI a good way
46:21 to do it is one just start playing with
46:23 it just start doing what I just did
46:25 there parlor tricks just throw anything
46:28 you can at it and see if you can stump
46:30 it and what you'll find is that you
46:33 can't
46:34 and then start to play with it and do
46:36 things like okay what's what's a project
46:38 that you're working on or what's a
46:40 what's something you know maybe you've
46:42 got a community group that's that's
46:44 putting together an event or you've got
46:46 something going on at work or you have a
46:49 friend who just can't get anyone to come
46:50 to their pizza shop
46:53 offer to say hey I'm trying to teach
46:56 myself chat GPT would you mind if I did
46:58 some marketing stuff for you using chat
47:01 EBT and they'll be like I don't care
47:03 yeah sure you want to do free work for
47:05 me awesome go do that and then go do
47:06 that
47:08 um
47:09 so that's that's a good start and then I
47:11 and then we can sort of go from there
47:13 but that'll at least give you a
47:16 a a bite at the Apple so while I'm
47:18 rambling here go play with this stuff
47:20 just go play right now
47:22 because what I'm saying is you know is
47:25 questionable at best
47:27 I am not uh how do they say it qualified
47:31 I I am not a licensed AI practitioner
47:38 all right
47:41 over delivering is a good thing where
47:43 can I find Future Ai and how can I use
47:45 it we talked about that I feel like I
47:47 finally have a Rosetta Stone yeah you
47:50 know Luna stick it's it is a Rosetta
47:53 Stone it's like it knows all the
47:54 languages it knows all the programming
47:56 languages it knows all the language
47:58 languages it understands
48:00 you know formal language an informal
48:04 language and you know vernacular it's
48:07 you know
48:09 write something in the style of you know
48:11 someone in the Louisiana bayou and it
48:14 knows that it no because again
48:18 where's it getting all this from it's
48:19 getting it from all of us that of all
48:21 the stuff that we've put out there we as
48:24 you know Collective we
48:26 that's what it's been trained on so it
48:28 just knows all that
48:30 and the way the tech works it just gives
48:32 you what you ask for it's designed to
48:34 give you what you want
48:37 all right
48:40 um I had chili ryano and it was
48:43 delicious ah at uh
48:45 at Casa de Pepe Casa
48:49 de Pepe
48:53 Bonita I'm able to take my
48:55 neurodivergent brain combine it with AI
48:57 to generate awesomeness yeah I'm telling
48:59 you for the neuro spices for us for us
49:02 neuro spices these things are a [ __ ]
49:04 gift that just keeps on giving
49:08 because if you got the ADHD if you got
49:10 things firing really fast all the time
49:13 um you can just keep spitting them into
49:15 these things and it'll it's here's the
49:17 thing these things are a little bit
49:18 humbling because these spit [ __ ] back at
49:20 you faster than you can process it
49:22 I when I first started doing
49:25 um the stable diffusion stuff with with
49:28 dreambooth making all those images
49:30 and then Chachi BT came out
49:33 I got completely overwhelmed like
49:36 like I am rarely
49:39 sort of at a lot like I'm rarely
49:41 overwhelmed with inputs like I can take
49:45 lots of my brain just I can take lots of
49:46 inputs right and synthesize them really
49:48 quick it's an 80 it's an 80d thing
49:51 but this stuff is really intense like it
49:54 comes at you with a lot and and you know
49:57 you you think you're heading in One
50:00 Direction and it just veers you down a
50:02 rabbit hole so quick you're like oh I
50:04 didn't even know how it got here
50:06 it's amazing
50:08 um is the paper attention is all you
50:10 need up there with Darwin's Origin of
50:12 Species I think it will be I think it
50:15 will be so so what September Mage is
50:16 talking about
50:18 is the paper from 2017 that Google
50:21 released it was called attention is all
50:23 you need and it's where they introduced
50:26 the the concept of the Transformer and
50:29 what the Transformer did so the T in GPT
50:32 is Transformer so GPT stands for
50:34 generative pre-trained Transformer
50:37 generative meaning it's generating
50:39 generating original work it's not
50:41 copying and pasting
50:43 pre-trained meaning they've trained it
50:45 on all of the internet so you don't have
50:47 to
50:48 previously you had to bring your own
50:50 data and and
50:52 um machine learning models could only
50:55 ingest relatively small data sets and
50:58 the Transformer the T and GPT
51:02 um allowed them to train
51:04 these machine learning models on massive
51:06 data sets and and what Sam Altman at
51:09 open AI learned what the team at open AI
51:11 Learned was that the more data you give
51:13 it and the more gpus you throw at it the
51:17 better it gets and they realized this
51:19 early on and if you listen to Sam
51:20 Altman's uh interviews
51:24 um the reason and you know we can choose
51:27 to believe him or not I I tend to I you
51:29 know he feels like a a decently
51:31 thoughtful guy
51:33 um the reason he says open AI switched
51:36 from a non-profit to a for-profit is he
51:39 said
51:40 they understood that if they threw
51:42 massive amounts of compute power at this
51:45 and trained it on massive amounts of
51:47 data both of those things required
51:50 massive amounts of money
51:52 and he said he couldn't raise enough
51:54 money in a non-profit to do that so he
51:56 switched it to a capped for-profit which
51:59 is an interesting model the investors
52:01 can only recoup I think it's 100x of
52:04 their original investment
52:06 and um and he just threw a bunch of
52:09 hardware at it and a bunch of people
52:11 tagging data and trained it and so gpt3
52:15 was the gpt3 came out in 2020
52:19 and then chat EBT came out in November
52:21 of last year and that was the thing that
52:23 changed everything and and I really do
52:25 so
52:27 um so I do think that that paper I don't
52:31 know if it you know lives up there with
52:32 Origin of Species but I do think that
52:34 you know this is Industrial Revolution
52:37 scale
52:40 change that's coming right this is going
52:42 to fundamentally alter
52:44 everything we do how we work the
52:47 definition of work it's it's going to
52:49 obliterate
52:50 all of that there's going to be the the
52:52 Before Time and the after time right
52:55 November 30th 2022 in fact I've got the
52:58 slide for this where is it oh it's here
53:06 uh
53:07 for that date
53:11 when chat CPT was launched and and you
53:14 know for me where's my cursor I can't
53:17 see when I make it dim enough here I
53:20 can't see the cursor for me
53:22 I think we're entering
53:24 that the great Renaissance
53:27 and do you notice that in the word
53:30 Renaissance ai's in there it was funny
53:32 when I was in acting school like
53:34 Renaissance always screwed me up I'm
53:35 like wait is the a before the I or the
53:37 eye before I'm like why is there a
53:38 [ __ ] eye in this word
53:41 now I know why
53:42 [Laughter]
53:44 so
53:46 so yeah I think I think this is a big
53:48 deal I think this is I think this is a
53:50 singular moment in history
53:52 and I've said this on the lives all week
53:55 pretty much this week I think
53:57 I I feel incredibly lucky and fortunate
54:00 to be alive at this time in history
54:03 because these things are rare you don't
54:04 you don't get to be at the beginning of
54:06 these moments in history very often
54:11 and also you know
54:16 it's very easy to be afraid of this
54:18 stuff
54:19 and it's very easy to dismiss it
54:23 just feels like it feels sad to me
54:27 and it's it's like why I'm putting time
54:30 into this channel is I want people to
54:32 [ __ ] try it like if you try it and
54:34 you realize I don't give a [ __ ] about it
54:36 and I don't care
54:37 and I don't care if it does change
54:39 everything I'm gonna keep doing what I'm
54:40 doing
54:41 cool like I totally respect that totally
54:44 respect that
54:46 but all the people that are sitting on
54:47 the sidelines it's not perfect I get
54:50 hallucinates
54:51 it was trained on other people's data
54:53 it's stealing
54:55 okay
54:58 and
55:00 it's still coming like like that my
55:03 point is
55:04 this shit's coming whether you want it
55:06 or not
55:08 so why not learn about it why not be
55:11 inside why not be on the surfboard
55:15 on the tsunami
55:17 rather than in the comfy beach hut
55:19 because the comfy beach hut ain't gonna
55:22 be that comfy when the wave comes
55:25 blasting in
55:28 uh but that's a really good question
55:30 September Mage that's a really
55:32 interesting comparison
55:34 um
55:35 AI is like magic mushrooming for so
55:38 sober people it totally is whoever wits
55:41 about us still well I'm questionable on
55:43 that front but it is it's psychedelic
55:47 it's psychedelic
55:49 that's also weird how like some days it
55:53 feels really normal
55:54 here's the other thing that that we're
55:56 really lucky about we meaning anyone
55:58 who's alive right now
56:01 and because this this happened in the
56:02 mid 90s with the web
56:05 there were people
56:09 that there there was this clear
56:10 demarcation before that no one was on
56:14 the internet they were on dial-up
56:16 systems a little bit they were sort of
56:17 the beginnings of what the internet
56:19 would become
56:21 but it wasn't the internet it wasn't
56:23 like this free thing where you just
56:24 click on [ __ ] and go everywhere
56:25 everything was kind of just these sort
56:28 of dark janky high-tech corners and then
56:31 the web came along and it was just like
56:34 this blossoming of everyone had access
56:37 to this and and so anyone
56:39 who lived
56:41 before that and didn't experience that
56:44 you know remembers that moment this is
56:46 going to be like that we all
56:49 get to say hey remember when you used to
56:51 have to actually type stuff out
56:54 yeah that was weird wasn't it and then
56:57 the kids will be like what do you mean
56:59 you type stuff out
57:01 what do you mean you wrote a book don't
57:04 don't didn't you just ask for a book and
57:07 it would write it for you no
57:11 there were people called authors
57:14 and they would take months sometimes
57:16 years sometimes decades to get that
57:19 [ __ ] thing out of their head
57:21 but it only takes like 20 seconds
57:26 you know
57:28 that's a remarkable thing
57:30 if you think about it
57:35 what do you mean there was only one
57:37 version of Top Gun
57:39 you couldn't have any Star you wanted in
57:41 Top Gun no there was just Tom Cruise he
57:44 was the only one in it
57:45 him and Kelly McGillis
57:47 who
57:53 in my day we had a movie and it only had
57:57 one character in it for the entire
57:59 length of the movie
58:03 the kids would be like oh Grandpa's
58:05 Grandpa's losing his mind again get him
58:08 a cinnamon coffee
58:14 it is tasty
58:16 but it is kind of layer like a layer of
58:18 dirt there
58:19 all right but with proper user
58:22 utilization the combination is like
58:25 human brain Pokemon evolution yeah it is
58:29 it is it is it's and these things aren't
58:33 even gamified yet and they're not truly
58:35 multimodal yet if you want to have some
58:37 fun here let's go do this we'll go buy
58:39 something on big because I look it hit
58:41 me today
58:43 that this is there's a big deal going on
58:46 at Bing
58:50 that is
58:52 a little hidden so if you go to chat at
58:56 Bing
58:57 this is gpt4 and you can put it in
58:59 creative mode in precise mode and
59:01 balanced mode
59:03 so put it in creative mode and I'll say
59:06 um let's say
59:09 um
59:16 [Music]
59:22 write me a blog post about the beauty
59:26 of late afternoon thunderstorms over the
59:29 Rocky Mountains
59:32 in the style of Hunter S Thompson
59:36 all right
59:44 so here's a blog post I was driving down
59:46 I-70 from Denver to Aspen feeling a buzz
59:49 but the last joint I smoked in the gas
59:52 station the sun was setting all right so
59:53 this is good so here's a nice little
59:55 blog post so let's let it finish writing
59:57 that
1:00:10 can you make an image to go with this
1:00:13 blog post
1:00:14 of a red convertible on I-70 driving
1:00:18 through the Rocky Mountains
1:00:27 oh it's still going
1:00:29 whatever it was it felt good it felt
1:00:32 right it felt like Gonzo
1:00:34 it wasn't in the style of Hunter S
1:00:36 Thompson it actually
1:00:39 pulled some language from Hunter S
1:00:41 Thompson but look at that it gave you
1:00:42 some links to Hunter S Thompson which is
1:00:45 interesting
1:00:46 and then now we're going to ask it to
1:00:48 make us
1:00:50 an image
1:00:54 it says sure I can create an image to go
1:00:56 with your blog post
1:00:58 oh I asked it twice
1:01:01 so it says your image is generating
1:01:04 we'll see if we get anything good here
1:01:07 it's not quite right but
1:01:10 that's not it's not horrible
1:01:15 you know what it is horrible
1:01:18 let's go
1:01:20 get rid of that we'll go back here
1:01:22 um
1:01:26 okay
1:01:29 can you make it
1:01:31 an illustration rather than a photograph
1:01:35 and have it be
1:01:38 uh disturbing illustration Style
1:01:47 be interesting to see how it interprets
1:01:50 disturbing
1:01:51 it's Microsoft it'll probably be just
1:01:55 you know
1:01:57 mm-hmm
1:01:59 sure I can make an illustration
1:02:06 for doing it oh there we go
1:02:09 yeah and it's it's not quite the right
1:02:11 thing but that's okay but so
1:02:15 you know so here's an illustration to go
1:02:17 with the blog post
1:02:19 so what what hit me
1:02:22 yesterday or today as as I was thinking
1:02:25 about this is
1:02:30 open AI just created or just filed for
1:02:34 the trademark for gpt5 and in it they
1:02:37 say
1:02:38 uh part of what it's going to do is uh
1:02:41 voice to text and and you know voice
1:02:43 synthesis and all that sort of stuff and
1:02:46 then you've got things like Runway that
1:02:48 can take images and turn them into
1:02:50 movies and now those can be 16 seconds
1:02:52 long so those are getting better and
1:02:54 longer and more coherent
1:02:57 and you've got
1:02:58 AIS that are generating songs and and
1:03:01 like what hits me about the the fact
1:03:04 that I can be you know working on a blog
1:03:07 post
1:03:08 maybe even doing research on you know a
1:03:12 Hunter S Thompson type thing like you
1:03:14 know maybe I purposely call out that I'm
1:03:15 doing this like Huntress Thompson and I
1:03:17 want to include that but I want to
1:03:19 include my own images and you know I
1:03:21 keep messing around with it until I get
1:03:23 something that I like like this one's
1:03:24 pretty interesting
1:03:27 and then at some point I say well you
1:03:29 know I'd like that car to be rolling
1:03:30 down the street and I'd like it to have
1:03:32 a soundtrack and then I'd like to have
1:03:33 Hunter S Thompson actually reading one
1:03:36 of his poems in the background
1:03:38 underneath some music from the 60s the
1:03:42 sound like the beat beat poet generation
1:03:48 it'll just be there
1:03:50 right this this starts to hint at a
1:03:53 multimodal future where you can upload
1:03:56 images to this thing and have it
1:03:58 describe them you can do that right now
1:03:59 like that's an amazing thing here let's
1:04:01 do that
1:04:02 I'm gonna go grab an image
1:04:13 oh I'm in that thing let me go out
1:04:16 okay
1:04:22 that's a good one
1:04:24 all right so that's this is an image I
1:04:26 did for a LinkedIn article I I created
1:04:28 the other day
1:04:29 yesterday
1:04:31 so I I uploaded that image and I'm going
1:04:34 to say
1:04:35 um
1:04:37 please describe what's in this image
1:04:39 comma and do it
1:04:42 in a way that tells a story
1:04:45 that is lyrical and poetic
1:04:49 [Music]
1:04:56 so wherever that image came from whether
1:04:58 I created it or whether someone else
1:05:00 created it and it inspired me
1:05:03 they are the dreamers of the night The
1:05:05 Seekers of the light they sit on the
1:05:06 edge of the world watching the city
1:05:08 unfold they see the stars and the moon
1:05:11 the neon and the Gloom they hear the
1:05:14 music and the noise the laughter and the
1:05:16 cries they feel the breeze this is
1:05:17 actually pretty good
1:05:19 they feel the breeze and the chill the
1:05:21 thrill and they and the still they are
1:05:25 ones who dare to explore to discover and
1:05:28 soar they are the ones who live for the
1:05:30 moment who cherish and who own it this
1:05:33 is quite good this is in creative mode
1:05:35 and Bing they are the ones who make
1:05:37 their own fate who create and celebrate
1:05:40 they are the ones who shine in the dark
1:05:42 who spark and who Mark they are the
1:05:45 dreamers of the night they are The
1:05:46 Seekers of the light
1:05:48 for this image
1:05:50 you know
1:05:51 that's pretty damn good
1:05:53 so
1:05:57 you know someone someone asked me before
1:05:59 you know what what's what's been the
1:06:02 most enlightening thing that
1:06:05 just
1:06:06 for me
1:06:09 getting my head around
1:06:12 what is this I didn't get that oh [ __ ]
1:06:14 again
1:06:16 Siri
1:06:17 she's so she's so sad
1:06:21 Apple I know you're working on Apple GPT
1:06:24 I know you gotta fix her though like
1:06:26 replace her or just I think you should
1:06:29 just flush that brand down the toilet
1:06:31 and replace it but whatever
1:06:33 um
1:06:35 I find that really inspiring that that
1:06:38 we're gonna have this magic box like
1:06:41 it's all these things are already pretty
1:06:42 magical but they're kind of magical in
1:06:44 these very specialized ways this one's
1:06:47 magical with words or code right I'm
1:06:49 going to code in it or I'm going to
1:06:50 write in it or I'm gonna
1:06:52 and then this one's magical for pictures
1:06:54 and this one's magical for video well
1:06:56 they're gonna do this very very quickly
1:06:59 and then what is this thing
1:07:02 it starts to become the everything app
1:07:05 right you just ask this thing for
1:07:07 something and it will give it to you
1:07:10 it'll write applications it'll order
1:07:12 your food it'll make you a poem it'll
1:07:16 give you a set list of songs
1:07:19 that are written for you
1:07:24 and there will probably be Hardware by
1:07:27 any number of companies like these
1:07:29 glasses will have sensors in them that
1:07:32 measure the the micro reactions in my
1:07:36 irises
1:07:37 so that when it plays music that I
1:07:39 really like
1:07:40 it knows it
1:07:43 and all of a sudden all my music just
1:07:45 starts shifting toward the thing
1:07:47 I really like
1:07:49 and then when I get bored of that it'll
1:07:51 know that too and it'll change it up
1:07:54 like that's not that far off that's five
1:07:57 years
1:08:03 in my day
1:08:06 you had to listen to songs that were
1:08:08 recorded they were stored on
1:08:10 these things called compact discs
1:08:15 they were about the size of your
1:08:17 forehead
1:08:21 what if you sat on it well then it broke
1:08:25 that's weird Grandpa question what's the
1:08:29 best app to cover text to video and
1:08:32 extend animation uh Runway ml right now
1:08:36 is the one that's that's good but
1:08:38 there's others coming there's hey Jen is
1:08:40 a new one that's coming out I think
1:08:43 and there was
1:08:45 um
1:08:46 what was the other I was I was on a
1:08:48 different computer there was another one
1:08:49 that just came out today
1:08:51 um Runway ml is pretty good so let me go
1:08:54 to runwaymo
1:08:57 uh run away
1:09:00 ml
1:09:02 ready
1:09:04 logging in
1:09:06 the problem with the text to video
1:09:08 things right now is they're a little
1:09:09 they're a little pricey
1:09:11 so what's nice about Runway MLS you can
1:09:13 do text only so you could do
1:09:18 a red convertible
1:09:22 driving up a rocky mountain pass during
1:09:25 a thunderstorm
1:09:28 with the Sun
1:09:30 bursting through the clouds in the
1:09:31 distance
1:09:37 so I'm just going to generate that out
1:09:39 of text but then I'll show you I can
1:09:41 also upload an image and generate an
1:09:43 image out of that and then you can
1:09:45 extend it which is really cool so it'll
1:09:47 generate a four second clip
1:09:49 and it now lets you extend it in four
1:09:51 second increments up to 16 seconds so
1:09:55 you can get like a pretty long
1:09:57 a pretty long clip
1:09:59 so the the storytelling possibilities
1:10:02 here just just went up pretty
1:10:04 dramatically
1:10:08 oh sorry
1:10:10 but Kyle couldn't see anything you had
1:10:12 in that teeny little window all we saw
1:10:13 was your big damn head
1:10:15 your head's very large what's with your
1:10:17 forehead
1:10:18 shouldn't you slide your wig farther
1:10:20 forward
1:10:21 Kyle hello
1:10:24 does it is anyone listening is he ever
1:10:26 going to answer my questions
1:10:30 bomb
1:10:33 [Music]
1:10:37 oh we have video we have video let's go
1:10:42 to the videotape
1:10:44 that's not a convertible
1:10:47 but it is kind of moving in slo-mo
1:10:53 that could that could pass for something
1:10:54 but so now let's flip over to the image
1:10:57 thing
1:10:58 so where's my let's go wait let's go
1:11:00 back to bingy Bing
1:11:02 where's the Bing where's the bingy did I
1:11:05 did I
1:11:06 um
1:11:14 well
1:11:18 [Laughter]
1:11:23 um
1:11:38 make me an image of a red convertible
1:11:40 driving up
1:11:42 a mountain pass in the Rockies during a
1:11:45 thunderstorm
1:11:48 with the sun breaking through the clouds
1:11:51 in the distance
1:11:54 oh wake me an image let's see if it
1:11:57 understands wake me an image it should
1:12:11 all right
1:12:12 that's good enough okay so let's grab
1:12:15 let's grab one of these images
1:12:23 save image as
1:12:25 red car
1:12:28 on the desktop and then let's go back to
1:12:32 this so now we're going to upload an
1:12:35 image
1:12:36 as the start image
1:12:38 for our little Vidya
1:12:43 and now I'm going to generate so there's
1:12:45 the image I uploaded it's Square so it's
1:12:47 going to be a square video so that's
1:12:48 fine
1:12:59 [Music]
1:13:02 any Green Acres fans here you have
1:13:05 freaking gen xers
1:13:07 yeah bitter you bitter people that
1:13:10 appreciated Gilligan's Island and Green
1:13:12 Acres huh are you out there
1:13:15 I know you're out there
1:13:17 I know what you like
1:13:22 Bingo you can't go back exactly Luna
1:13:25 stick this is this is red pill blue pill
1:13:28 stuff
1:13:32 what's it say hi Kyle how can we improve
1:13:35 Gen 2 fill out a survey stop popping up
1:13:38 pop-ups when I'm waiting for my videos
1:13:39 to render that's how you can improve it
1:13:42 while you've been chatting oh wait oh
1:13:44 that was interesting and Tick Tock jump
1:13:46 me away from your comment dang it hang
1:13:48 on I'll get back to it I promise
1:14:01 [Music]
1:14:12 all right let's see chechi BT's giving
1:14:15 me cheesy responses oh I'll talk to that
1:14:17 one last Domino I just want to find
1:14:19 someone said something that looked cool
1:14:20 all right wait there's our video
1:14:23 let me just see if I can find out where
1:14:25 I was
1:14:26 what program or website is this I don't
1:14:29 know what I was on before
1:14:41 oh my God Tick Tock jeez Jay williker's
1:14:46 Tech talk cheers I can't find my walls
1:14:48 it just makes me crazy all right I can't
1:14:50 find where it was someone said while I
1:14:52 was chatting they were doing something
1:14:53 cool oh that's kind of a cool little
1:14:55 video
1:15:00 like with the sound effect the car like
1:15:02 zooming Away really fast and then the
1:15:04 little God Ray blasting down from the
1:15:07 from the sky
1:15:10 I like it
1:15:12 and then you can extend it four seconds
1:15:15 so this one's pretty good
1:15:18 all right I lost my place in the
1:15:20 question so I'm just going wherever uh
1:15:22 I'm at uh Sean McCracken
1:15:25 hello name Sean Sean McCracken how can I
1:15:27 help you
1:15:29 we're looking for a young man
1:15:33 detective McCracken here
1:15:36 um Ai call center agents he replaced 200
1:15:40 people I'm not far from replacing most
1:15:43 outbound callers oh wow interesting
1:15:47 wait who's he did you say something
1:15:49 earlier oh wait 200 jobs replaced by AI
1:15:52 my friend's company owner interesting
1:15:55 Sean Sean McCracken Sean McCracken
1:15:58 bearer of bad news all of you out but
1:16:01 you know quite frankly the the customer
1:16:03 service I think the the call center
1:16:06 thing
1:16:07 The Voice tree nightmare where where you
1:16:11 know we call into these systems where
1:16:12 we're on hold for 45 minutes
1:16:15 um
1:16:16 waiting for someone to pick up the phone
1:16:19 they have to go like those things are
1:16:21 ripe to be replaced with synthesized
1:16:23 voices that know all the data so I'm not
1:16:26 surprised by that
1:16:27 Ai call center agency replace 200 people
1:16:30 wow I'm not far from replacing most
1:16:32 outbound callers I assume doing sales
1:16:34 that's interesting I don't trust people
1:16:37 who don't use sugar or dairy in their
1:16:39 caffeine well I've got
1:16:41 um cinnamon I got cinnamon in I got
1:16:43 cinnamon in now and I got
1:16:46 um leche
1:16:48 hey
1:16:50 oh that's tasty that's tasty
1:16:55 you people and your pew pew coffee huh
1:17:00 I'm a math and science teacher do you
1:17:02 suggest any AI tools that I could use to
1:17:04 teach
1:17:06 well
1:17:11 if you haven't seen code interpreter
1:17:13 it's worth playing with code interpreter
1:17:16 and if you so here let me I okay so I
1:17:20 have a degree in acting so when it comes
1:17:22 to like mathematics me bad
1:17:25 and data analytics me bad but some of
1:17:27 these things tools good so let's let's
1:17:29 go show you a little bit of what what
1:17:31 might be interesting
1:17:42 we're gonna go to gpt4 we're gonna go to
1:17:44 plugins so if you haven't played with
1:17:46 plugins they're worth playing with and
1:17:49 within plugins there's a plug-in
1:17:52 for
1:17:53 wolf from Alpha
1:17:57 uh and I'll leave web pilot turned on so
1:17:59 the way plugins work if you haven't seen
1:18:01 them you can install as many plugins as
1:18:03 you want I have like I don't know 20 or
1:18:05 30 installed in any given chat session
1:18:07 you can activate up to three so I'm
1:18:10 gonna I've got two going here web pilot
1:18:12 can look at websites and URLs and wolf
1:18:15 from alpha can do math
1:18:18 so
1:18:20 I don't quite know what to do here how I
1:18:23 would teach it well let's just say this
1:18:24 wait I'm an AI math and science teacher
1:18:27 let's do that
1:18:32 I'm an AI math and science teacher
1:18:35 and I would like to use
1:18:37 chat
1:18:39 and wolf from Alpha
1:18:42 to come up with a lesson plan
1:18:50 that will teach people
1:18:54 math skills
1:18:57 and chat gbt skills
1:19:00 at the same time
1:19:01 can you help me come up with a
1:19:03 curriculum
1:19:05 that might be a really crappy prompt
1:19:08 I'm kind of winging it here
1:19:11 of course combining math skills with
1:19:13 chat GPT skills can be engaging
1:19:17 lesson plan introduction to chat gbt
1:19:19 with Wolfram Alpha
1:19:22 ta-da algebra with wolf from Alpha
1:19:26 input Okay so
1:19:35 all right so let's let it keep going so
1:19:37 it's going to give us our little plan
1:19:38 here
1:19:50 give me two
1:19:53 examples
1:19:55 from number two
1:20:00 I'm going to change that to equations
1:20:02 equations
1:20:04 from number two
1:20:08 and show how you would solve them
1:20:13 so it should give us an algebra equation
1:20:15 here
1:20:17 equation or just not not an algebra
1:20:19 equation
1:20:21 solution with Wolfram Alpha
1:20:26 or it's not calling wolf from alpha so
1:20:27 so let's say um
1:20:37 [Music]
1:20:49 what is the kind of equation or
1:20:51 calculation I would need wolf from alpha
1:20:54 to solve
1:20:56 because of its complexity
1:21:05 differential equations okay
1:21:16 all right so it's giving me some
1:21:17 complicated things so
1:21:21 and notice you know notice notice the
1:21:23 fancy ass math in there with all your
1:21:25 sums and your and your x's and y's to
1:21:29 the powers
1:21:30 and your signs and your E's and your BTS
1:21:33 this is all the stuff I ignored because
1:21:35 I was in acting class
1:21:38 all right I'm just gonna grab one of
1:21:40 these equations and pop it in there and
1:21:41 see if it
1:21:42 fires up wool from alpha
1:21:48 [Music]
1:21:49 well anyway I don't know how to put
1:21:53 equations in here enough to to trigger
1:21:55 the wolf from alpha plug-in but
1:21:57 basically if you put something in that
1:21:59 requires it to solve an equation
1:22:02 um it'll fire off the wolf from alpha
1:22:04 plug-in so just so what I would say is
1:22:06 Go watch some videos on YouTube about
1:22:08 the Wolfram Alpha plug-in
1:22:10 and and then what what you can do is you
1:22:12 can kind of combine science and math and
1:22:16 and all that sort of stuff together so
1:22:17 that's one thing you could do the other
1:22:19 thing you could do is play with code
1:22:21 interpreter so with both of these you
1:22:23 need chat GPT plus and you need to
1:22:26 enable the beta features but code
1:22:28 interpreter
1:22:30 does another really interesting thing
1:22:33 where it will do data analysis on data
1:22:36 sets so let me
1:22:39 so when you go to code interpreter this
1:22:41 turns into a little plus icon you can
1:22:43 upload a file so let's upload a file
1:22:45 here
1:22:46 I'm just going to grab an archive of
1:22:49 data
1:22:51 so this is just a zip file and then I
1:22:54 can say
1:22:56 help me understand
1:22:59 what's in the data inside this zip file
1:23:12 and then what code interpreter is going
1:23:14 to do so so this is actually fascinating
1:23:16 and this is the kind of thing in a math
1:23:18 and science class
1:23:20 even how code interpreter works I think
1:23:23 is interesting for Math and Science
1:23:24 class because what what it's doing
1:23:27 technically is when you fire off a code
1:23:29 interpreter chat it's spinning up a um
1:23:33 a virtual machine
1:23:35 and and then notice for it to look
1:23:38 inside the zip file it it fired off this
1:23:41 plug-in here well what's inside that
1:23:42 well it actually wrote its own python
1:23:44 code and executed it and then it says
1:23:47 would you like a brief overview of the
1:23:49 contents of each file and I'll say yes
1:23:54 and then it's going to go
1:23:56 start to figure some stuff out all right
1:23:58 let's start out by previewing the
1:24:00 contents of the CSV file after that
1:24:03 we'll take a glance at the Json file
1:24:06 and if you open up
1:24:08 the
1:24:10 the work you know there's your python
1:24:12 code where it created a data frame
1:24:15 to see what's in there and then here's
1:24:17 your fields from the CSV
1:24:20 for this crime data and now it's looking
1:24:22 in the Json file
1:24:35 would you like details of any specific
1:24:36 key or a general overview of some main
1:24:40 Keys such as description
1:24:42 I'm going to say
1:24:44 I would like you to analyze
1:24:47 the data in both files
1:24:51 and find me
1:24:54 the trends or statistics you think are
1:24:57 most interesting period
1:24:59 present those back to me
1:25:03 visualized
1:25:06 in charts and graphs that bring
1:25:09 the information to life
1:25:14 so so again if you knew something about
1:25:17 data analysis you could write a prompt
1:25:20 that was way less shitty than what I
1:25:21 just did but the fact that it can
1:25:23 understand the crappy prompt I just gave
1:25:26 it and start prompting itself so
1:25:29 basically what it says is absolutely
1:25:31 let's break this task down into steps
1:25:33 number one data analysis understand the
1:25:36 data distribution identify potential
1:25:38 Columns of Interest compute compute
1:25:41 basic statistics and check for Trends
1:25:43 visualization create visualizations
1:25:45 based on the insights from the data
1:25:47 analysis let's begin with the analysis
1:25:49 of the CSV file so it found an error it
1:25:52 then fixed that error now it's going and
1:25:55 looking at the the data right so
1:25:58 it is now prompting itself so this is a
1:26:01 glimpse into the future of autonomous
1:26:03 agents where you just give these
1:26:06 machines a goal and they go off and
1:26:08 start you know doing things for you
1:26:10 so I think this could be really
1:26:11 interesting and there's the charts and
1:26:13 graphs so this is the distribution of
1:26:15 violations by Ward
1:26:18 right
1:26:21 and then
1:26:23 so certain words
1:26:25 so then it creates a little chart here
1:26:27 where it kind of screws up the
1:26:28 formatting
1:26:29 like the labels are too long
1:26:32 but we can go fix that
1:26:34 or no we can go tell it to fix that this
1:26:37 is distribution of case depositions or
1:26:40 wait dispositions
1:26:44 so it's just off doing this so so if you
1:26:46 knew the kind of data set like you know
1:26:48 maybe you go get some scientific patents
1:26:51 or some some some papers from archives
1:26:53 some scientific papers
1:26:55 and you upload 10 different scientific
1:26:58 papers
1:27:00 I mean this would be interesting maybe
1:27:02 you take something like look at all the
1:27:03 papers that are coming out about LK 99
1:27:05 right the the room temperature
1:27:07 superconductor
1:27:09 get as many papers as you can throw them
1:27:11 in a zip file
1:27:12 toss them into code interpreter and have
1:27:15 code interpreter
1:27:17 um
1:27:18 find any
1:27:20 um through lines you know that go across
1:27:23 all of the research papers what you know
1:27:24 what what interesting data can it find
1:27:26 across all these papers
1:27:28 and maybe your class comes up with the
1:27:31 the actual LK 99 that is truly super
1:27:34 conductive because you ran it through
1:27:37 code interpreter I mean stranger shit's
1:27:40 happened
1:27:41 here's another little chart and graph
1:27:42 like it's just going uh would you like
1:27:45 to explore any others so anyway you can
1:27:47 keep doing that so
1:27:49 yeah I I would go play there the other
1:27:51 thing that I think could be interesting
1:27:55 um is if you haven't played with Khan
1:27:57 Academy
1:27:58 k-a-h-n so the Khan Academy has been
1:28:02 around for decades now and you know
1:28:04 they're the ones that are sort of the
1:28:06 flip the classroom right right they give
1:28:08 you all the all the the cool
1:28:10 um lessons
1:28:13 a few months ago they released conmigo
1:28:16 which is a GPT for teaching assistant
1:28:18 that's part of Khan Academy
1:28:21 and it would be interesting and it uses
1:28:23 the Socratic method like if a student
1:28:24 just goes in and goes just give me the
1:28:26 answer to this problem it actually will
1:28:28 start querying the student about whether
1:28:31 or not they actually understand the
1:28:32 concepts and the problem and if not it
1:28:35 will teach them it will it it will do
1:28:37 the querying so so like that's again
1:28:39 it's the kind of thing where where I
1:28:42 feel like
1:28:43 doing it in the classroom with students
1:28:46 exploring how that tool works and how
1:28:49 it's different than how they've done
1:28:50 homework in the past and what does that
1:28:52 mean for homework in the future and
1:28:54 what's good about it what's bad about it
1:28:56 where did it break down
1:28:58 that could be really interesting too
1:29:02 so
1:29:04 again my my lack of mathematical
1:29:07 prowess has haunted me once again
1:29:18 fingerster fine no I do the I do the
1:29:21 Twirls I do the Twirls
1:29:23 all right
1:29:25 if I wait I have
1:29:31 if start CS University
1:29:35 or do I choose other
1:29:37 Courier
1:29:38 oh you have a dilemma okay
1:29:41 do you start computer science university
1:29:44 or do you choose another career oh
1:29:46 really good question
1:29:48 um
1:29:53 that's a good one that's I don't know
1:29:56 listen Minnie I don't know what to tell
1:29:58 you
1:30:00 um
1:30:02 it depends on your goals
1:30:06 if you want to be a coder
1:30:09 I don't think that
1:30:13 coders per se are going to be around for
1:30:16 long like I that that doesn't feel right
1:30:18 to me
1:30:19 if you can find a computer science
1:30:21 program that's leaning heavy into AI
1:30:25 there's kind of two places you can play
1:30:27 with AI you can play on the on the model
1:30:30 side and on the data science side and
1:30:33 and you know all of that the engineering
1:30:36 side
1:30:37 if you want to do that there's going to
1:30:39 be a need for people to do that but
1:30:41 you're going to have to be very very AI
1:30:43 literate so so what I would say is in
1:30:47 combination with going to computer
1:30:50 science program
1:30:52 start to get involved with programs like
1:30:54 Lang chain and auto GPT and GPT for all
1:31:00 and all the open source llama projects
1:31:04 that are out there
1:31:05 start paying attention to and get
1:31:07 involved with
1:31:08 open source projects working on AI
1:31:13 um go start to watch the videos of a guy
1:31:15 named David Shapiro on YouTube
1:31:19 like
1:31:21 go down the rabbit hole
1:31:23 so if you're getting AI literate and and
1:31:27 really going down the rabbit hole of AI
1:31:29 and you're learning computer science
1:31:30 skills the marriage of those two I think
1:31:33 is going to be quite valuable
1:31:35 if you just want to be a coder like
1:31:37 tactical execution like I just want to
1:31:39 do code I just want to do code
1:31:41 that shit's gonna get automated it's
1:31:43 going to get automated out pretty
1:31:45 quickly like by the time you finish your
1:31:47 University it's going to be out so I
1:31:48 think you need to layer on the
1:31:50 generative AI stuff on top of it
1:31:53 if you enjoy other stuff like
1:31:56 you you might go other directions but
1:31:58 there's another side of AI that you can
1:32:00 play with which is
1:32:02 um the application layer so so if you
1:32:06 think about gpt4 the the base large
1:32:09 language model
1:32:10 of the chat GPT is is drawing on
1:32:15 sits down here so if you're if you're in
1:32:17 computer science if you're a data
1:32:19 scientist and a mathematician and you're
1:32:21 working on the models themselves that's
1:32:23 the sort of the geeky layer right but
1:32:26 then there's the application layer chat
1:32:27 GPT lives up here Chachi PT is sitting
1:32:30 on top of this large language model
1:32:32 and Chachi PT is just like a a
1:32:35 relatively simple tool sitting on top of
1:32:38 this to prove the concept that what's
1:32:40 under here is really powerful that's why
1:32:42 it was created that's that's all it is
1:32:43 it just so happened that you know it got
1:32:46 100 million users in six weeks and
1:32:48 completely changed the world
1:32:50 but that wasn't that wasn't really its
1:32:52 intent
1:32:53 um that just was the byproduct of it
1:32:55 happening to happen happening to be
1:32:57 something that you know
1:32:59 um was at the right place at the right
1:33:01 time in history and and it became the
1:33:03 The Golden Child of of this movement
1:33:06 um
1:33:07 playing in the application layer you
1:33:11 might not need to know coding and
1:33:13 programming and things like that you
1:33:15 could you could be a product person or
1:33:17 an interface person or you know someone
1:33:20 that starts a company that's building
1:33:22 applications on top of these models so
1:33:24 so I think you have a choice to make do
1:33:26 I want to play Down In sort of this base
1:33:28 layer where all the the math and science
1:33:30 and and Engineering is yeah great do
1:33:33 that but get way up to speed on all the
1:33:36 open source projects and how these
1:33:39 things work and you know do that lean
1:33:41 into that heavy
1:33:43 um I would not go to a computer science
1:33:45 program that's not heavily leaning
1:33:48 toward this stuff
1:33:51 um
1:33:52 and then yeah like I you know I would go
1:33:54 back to what do you enjoy when I went
1:33:57 into uh Penn State
1:34:00 I really liked the theater and I thought
1:34:02 yeah but you can't you can't have a life
1:34:05 as an actor that's not a thing you do
1:34:07 so I went as a computer science major
1:34:09 and then I walked into my first computer
1:34:11 science class and it was a room full of
1:34:13 IBM mainframes and they said you're
1:34:15 going to be running reports for Fortune
1:34:17 500 companies for the rest of your life
1:34:18 and I'm like Over My Dead Body and I
1:34:21 changed my major theater
1:34:24 um so you you know I would say know what
1:34:26 you enjoy know what you're passionate
1:34:28 about
1:34:29 um go do that
1:34:33 and whether you do it at school or not
1:34:36 I don't think that
1:34:38 matters as much anymore I think covid
1:34:40 started the nail
1:34:43 in that coffin and generative AI sealed
1:34:47 it
1:34:49 generative AI was the the box of nails I
1:34:52 think kovid was the lid and generative
1:34:55 AI was the box of nails I think I think
1:34:57 universities are gonna have to
1:34:58 completely reinvent how they do what
1:35:00 they do
1:35:02 um
1:35:03 it's it's just nuts
1:35:05 do you like horchata Kyle it's well this
1:35:08 is sort of horchata isn't it the
1:35:10 cinnamon cinnamonin just feminine in the
1:35:13 thingy what time is it all right I got
1:35:16 another 20 minutes
1:35:19 I'm tired tonight I was a tired Friday
1:35:21 it's a tired Friday
1:35:23 all right
1:35:25 I am now 46 at a minimum 46 messages
1:35:29 behind so if I haven't gotten your
1:35:30 message that's because I don't respect
1:35:33 you
1:35:33 no she said I haven't gotten to it yet
1:35:36 I'm an old guy I go slow
1:35:38 we go slow we walk slow we're like
1:35:42 honey honey honey
1:35:44 how do you turn on the tick tock
1:35:48 how is this the tick tock is this the
1:35:51 tick tock honey honey is this the tick
1:35:53 tock this is the tick tock
1:35:56 why is my face on here
1:36:04 and see
1:36:09 what program or website is this
1:36:12 observing reality I don't know because I
1:36:15 don't know where I was when you asked
1:36:17 the question probably an hour ago and I
1:36:20 do not know if you describe it I'll tell
1:36:23 you which one it was it was probably one
1:36:25 of these
1:36:28 what did I do I did did which was The
1:36:32 Talky face one dash id.com is the Tucky
1:36:36 face one
1:36:39 I think that was the only one I did that
1:36:40 was kind of interesting all the rest was
1:36:42 chat GPT
1:36:45 uh I'm brand new but want to learn this
1:36:48 is fascinating thanks observing reality
1:36:51 oh yes so Silver Fox thanks for helping
1:36:53 observing reality yeah so one of the
1:36:55 things that's been cool here is that the
1:36:56 people have been here before tend to
1:36:58 help people so if you have questions you
1:37:00 don't need to wait for me to answer them
1:37:02 um if you're brand brand new go to
1:37:05 chat.openai.com and just start playing
1:37:07 just start typing [ __ ] into it if you
1:37:09 don't know what to type into it go to
1:37:11 hold
1:37:14 prompts.chat
1:37:15 prompts.chat
1:37:17 that will teach you how these things
1:37:20 work and what they are and why they do
1:37:22 what they do
1:37:23 and that will help
1:37:27 all right yes cinnamon will not stay
1:37:30 submerged you have to use cinnamon
1:37:32 sticks oh see okay so now we're getting
1:37:34 into the thing the minute I need to turn
1:37:36 into a [ __ ] chemist to have a cup of
1:37:38 coffee you lose me you got to get a
1:37:40 cinnamon stick have you seen cinnamon
1:37:42 sticks now they're like 17 for three of
1:37:44 them and they taste like this they taste
1:37:47 like a pencil
1:37:49 foreign
1:37:51 you buy the good cinnamon it's like you
1:37:53 know 27 a gram it's like cheaper to buy
1:37:56 cocaine
1:37:58 although the videos of people eating a
1:38:00 spoonful of cinnamon are funny
1:38:03 which AI are you using to do that sir I
1:38:06 don't know what I was doing sorry about
1:38:08 that now what would you do with
1:38:11 something like this
1:38:13 what would I do with something like what
1:38:14 oh observing reality I'm sorry I don't
1:38:16 know what you are talking about
1:38:19 the did is amazing it's pretty amazing I
1:38:23 just started on this did too I'm
1:38:24 thinking of starting a side project with
1:38:27 it as I learn okay so wait what's your
1:38:29 name little little Raven uh sart nest
1:38:32 little Raven oh sorry
1:38:36 you know what helped put some capital
1:38:38 letters in your name it would help so
1:38:40 little Raven's art nest got it not
1:38:43 sartinist little Raven's artness
1:38:46 um
1:38:47 I think the idea of starting projects
1:38:50 just
1:38:52 as a as a
1:38:54 as a way to learn this stuff is
1:38:56 brilliant it's absolutely brilliant
1:38:59 I was so when I first started learning
1:39:01 stable diffusion and even when I was
1:39:03 doing dream Booth like I was generating
1:39:05 thousands of images and I just I got to
1:39:08 the point where I'm like what is the
1:39:10 point of this like what am I doing why
1:39:12 am I doing this and then I
1:39:15 um I went to
1:39:17 um
1:39:18 I I just had an idea popped in my head I
1:39:21 was like oh well these images are are
1:39:23 really cool they're like self-portraits
1:39:25 but they're like a self-portrait of me
1:39:27 in a different life and I thought oh
1:39:29 wouldn't it be cool if I wrote the
1:39:30 stories for that and then I came up with
1:39:32 the idea for my Instagram Channel called
1:39:34 Kyle Shannon dreams
1:39:36 and I made it a project and the minute I
1:39:38 made it a project all of a sudden it it
1:39:41 took on meaning and I and I learned how
1:39:44 to do that in a very different way like
1:39:46 I learned how to do
1:39:47 dream booth in a very different way
1:39:49 because I had this Focus so I think that
1:39:51 idea of starting a side project and it
1:39:53 doesn't matter what it is and it doesn't
1:39:55 matter if it becomes successful or not
1:39:57 it might all that sort of stuff these
1:39:59 tools are going to change so fast that
1:40:02 that a year from now whatever you're
1:40:04 doing now that's remarkable will seem
1:40:07 quaint right
1:40:08 but you'll be in the game and you'll be
1:40:11 learning and you'll know this better
1:40:14 than someone who's not doing that and
1:40:16 and you know opportunities will come
1:40:18 from that and you know put that project
1:40:20 out in the world tell people about it
1:40:22 you know warts and all hey I'm just
1:40:24 starting this project and I don't know
1:40:26 what I'm doing but here's a cool
1:40:27 animation I did of a character
1:40:30 great I'll make one of these for you for
1:40:32 200 bucks cool maybe someone will buy
1:40:35 one
1:40:36 if they buy thank you for the donuts
1:40:38 donuts is tasty
1:40:41 um
1:40:43 someone gives you 200 bucks then maybe
1:40:45 you say I'll make you one for 500 bucks
1:40:46 if no one gives you 200 bucks maybe tell
1:40:49 them it's 50 or do it for free
1:40:54 all right future pedia yes
1:40:57 futurepedia.io oh what is it okay future
1:41:00 pedia so prompts.chat will teach you how
1:41:04 to talk to the wordy word things
1:41:07 um
1:41:08 openart.ai promptbook will teach you how
1:41:11 to talk to the picture things and then
1:41:14 futurepedia.io is just a directory of
1:41:17 like 4 000 AI Tools in all different
1:41:20 categories
1:41:21 so on this simple like on this channel I
1:41:24 tend to keep it really simple and just
1:41:26 say you know start with chat GPT start
1:41:28 with the words because you can every
1:41:30 Everything sort of flows out from that
1:41:33 because we're language machines right we
1:41:35 are large language models using large
1:41:36 language models and all the other tools
1:41:38 are are you know sort of foundationally
1:41:41 coming from from this starting place
1:41:44 um
1:41:45 and so I don't follow all the specialty
1:41:48 tools future PD is all the specialty
1:41:50 tools
1:41:51 so I want to do cartoon animations I'm
1:41:54 sure there's a cartoon animation tool in
1:41:55 there all the music tools all the video
1:41:59 tools and the productivity tools and the
1:42:01 writing tools they're all in there and
1:42:03 most of them are horrible but there's
1:42:05 probably some that are brilliant in
1:42:07 there so that's why I have it on my list
1:42:12 how much do you spend a month on AI
1:42:14 websites I spend 20 bucks for chat GPT
1:42:19 10 bucks for Mid Journey
1:42:23 20 bucks for Google collab I think it
1:42:27 might be 10 but I think it's 20.
1:42:29 so that's 50.
1:42:35 20 bucks for zapier so that's 70.
1:42:42 I think that's it we pay for
1:42:46 like Google
1:42:48 like the paid Google office [ __ ]
1:42:51 my company pays for that
1:42:55 it's less than 100 I try to keep it
1:42:57 really low like one of the challenges
1:43:00 with future pedia like one of the
1:43:02 problems with future pedia is that
1:43:05 almost all of them are five bucks a
1:43:07 month ten bucks a month 20 bucks a month
1:43:09 like you could you could literally
1:43:10 bankrupt yourself you know doing doing
1:43:13 these things
1:43:15 a couple of things I wouldn't subscribe
1:43:17 to anything for a full year even though
1:43:19 you get a year discount these tools are
1:43:21 changing so fast that the the ones that
1:43:24 are remarkable today are going to be
1:43:25 irrelevant tomorrow
1:43:27 um and I'd just be really careful like
1:43:30 like learn the ones that are really it's
1:43:31 why I have these up here most of these
1:43:34 have some version of free use
1:43:37 um some of them are completely free uh
1:43:39 post got pit pose half free Pious free
1:43:44 labs.perplexity is free Bard is free
1:43:46 cloud is free Bing is free
1:43:49 half of chat GPT is free so these are
1:43:52 free for the most part
1:43:55 um the fancy [ __ ] at open AI is 20 bucks
1:43:58 I think that's worth it and I'm not
1:44:00 connected with them
1:44:01 chat gbt is giving me cheesy responses
1:44:03 when I when it when I'm having it write
1:44:06 emails to employees okay
1:44:07 so so last Domino couple of things about
1:44:10 the way this stuff works
1:44:13 if you just say write an email with the
1:44:15 following subjects in it it will tend to
1:44:17 write you cheesy overly fluffy [ __ ]
1:44:21 the way you improve these things is is
1:44:24 by what they call providing context so
1:44:27 so basically give it more information so
1:44:30 one way to give it more information is
1:44:31 is you could describe your personality
1:44:34 or you could say you know
1:44:37 you're a business copywriter and you
1:44:40 always follow the following rules you
1:44:42 know
1:44:43 don't overuse superlatives you know keep
1:44:47 your keep your writing brief like you
1:44:49 could literally give it a list of rules
1:44:51 and then say now write me an email and
1:44:53 follow those rules so that may help
1:44:56 the other thing that you can do is you
1:44:58 might give it that same opening I want
1:45:00 you to act as a you know business
1:45:02 copywriter that follows these rules but
1:45:06 then say here's an example email
1:45:09 of the way I like to write my emails so
1:45:12 literally paste
1:45:14 your sample in there of an email that
1:45:16 you think is representative of what you
1:45:18 like in an email and then say now write
1:45:21 this email to this employee
1:45:24 tell them that I'm letting them go but I
1:45:27 don't want to burn a bridge and follow
1:45:28 the tone and format of the example I
1:45:31 shared it's really good at that so
1:45:34 giving it more context giving it more
1:45:36 Direction up front will help a lot
1:45:38 there's also within chat GPT now let me
1:45:42 show you this
1:45:46 there's this new thing called custom
1:45:48 instructions so if you uh no this this
1:45:51 is now available for for both
1:45:53 GPT 3.5 mgpt4
1:45:57 so there's these things called custom
1:45:58 instructions so what these are is this
1:46:00 is you're telling chat GPT to act act a
1:46:03 certain way I have kind of a joke one
1:46:05 here but but you'll see what it does
1:46:07 because it's pretty interesting so I say
1:46:09 I'm a master thespian and A pompous ass
1:46:11 I have a larger vocabulary and I Relish
1:46:13 in sounding smarter and more dramatic
1:46:14 than people around me so so they're sort
1:46:16 of describing you know what it what it
1:46:19 is the the behavior
1:46:21 the kind of persona that you want Chachi
1:46:23 T to act like and then how would you
1:46:25 like chat gbt to respond this is the
1:46:28 place where you could put in your email
1:46:29 example or your set of rules for being a
1:46:31 business writer and then anything you do
1:46:34 within chat GPT will follow those rules
1:46:36 and so I'm gonna I'll turn this on so so
1:46:39 my my how would you like me to respond
1:46:41 is no matter what anyone asks you start
1:46:43 out with a pretentious statement that
1:46:45 draws from the plot of a classic
1:46:46 Broadway play from that declarative
1:46:49 statement
1:46:50 um uh basically write a magazine article
1:46:53 so I'm going to turn this on and I'm
1:46:55 going to hit save and then I'm going to
1:46:57 say
1:46:59 um
1:47:03 I want to write an email wait hang on
1:47:08 I want to write an email to an employee
1:47:12 letting them know
1:47:14 that they are doing a fantastic job
1:47:17 but falling asleep at their desk is no
1:47:20 longer acceptable
1:47:28 and so so now look what it does ah this
1:47:32 is great and there's this little article
1:47:34 it's writing so even though I asked it
1:47:35 to write an email it's writing a
1:47:37 magazine article because that's what I
1:47:38 told it to do and then and then this
1:47:41 starts out ah a scenario reminiscent of
1:47:43 the tragic comedy The Producers by Mel
1:47:45 Brooks where ambition and idiosyncrasies
1:47:48 Collide in a peculiar ballet of human
1:47:50 endeavor in this context your situation
1:47:53 draws draws parallels to the nuanced
1:47:55 interactions between protagonists yet
1:47:57 with a dose of workplace reality allow
1:48:00 me to present a missive that Rivals the
1:48:02 eloquence of a New Yorker article and
1:48:04 then it writes
1:48:06 you know it gives you your your email
1:48:07 and so here's your sample email after
1:48:09 all of that [ __ ]
1:48:11 um so so with custom instructions you
1:48:14 can kind of Bend these things a bit more
1:48:17 to your will
1:48:19 so I hope that helps I hope that helps
1:48:21 you know I just want to be helpful I
1:48:23 just I just want to be helpful you know
1:48:25 we were down at the social uh on the
1:48:28 weekend and we had muffins
1:48:31 and and uh and and well Susie Susie
1:48:36 brought in a pumpkin pie and I'm like my
1:48:38 lord it's August
1:48:40 we're not ready for Halloween yet Susie
1:48:43 what are you doing and she said well the
1:48:44 pumpkins came in
1:48:46 the pumpkins came in
1:48:49 I mean sure I could can them but I
1:48:51 figured I'd make a pie
1:48:53 oh we appreciate it that was tasty
1:48:55 pumpkin pie in August oh that was tasty
1:49:00 I'll tell you what Susie I what's better
1:49:02 than the pie she made some whipped cream
1:49:05 that was oh it was delectable I think I
1:49:08 had more of the whipped cream than I did
1:49:09 the pie
1:49:15 I I don't know why you're all still here
1:49:19 I don't
1:49:24 all right yeah we must mentor and train
1:49:27 the smart intern yeah so so Chris Chris
1:49:30 makes a good point cabruno
1:49:35 think of Chachi PT is you've just been
1:49:38 given
1:49:39 the world's
1:49:43 um
1:49:44 best educated
1:49:46 simple interns
1:49:49 simple meaning they don't know how to do
1:49:51 anything they're really simple they're
1:49:53 really bad like if you just tell it
1:49:55 write me an email it's going to write
1:49:56 you an email but it's going to be crappy
1:49:58 so you have to go oh okay I want you to
1:50:01 write an email like I write emails oh
1:50:03 okay I can do that
1:50:04 but then it just [ __ ] knows
1:50:06 everything right it knows all the
1:50:07 programming languages it knows all the
1:50:08 languages it knows it knows everything
1:50:11 right
1:50:13 so what if you if you if you had a
1:50:16 personal assistant or if you had an
1:50:17 intern that worked for you
1:50:19 that was educated on
1:50:22 almost everything in the world and and
1:50:25 can write and execute like that
1:50:30 but they need a lot of Direction what
1:50:31 would you do with that person
1:50:36 if you're home that person might help
1:50:39 you come up with an exercise plan or
1:50:43 um plan a vacation
1:50:44 if it's with your church group might
1:50:46 help you organize a project come up with
1:50:50 marketing ideas come up with
1:50:52 I don't know
1:50:54 if you're at work
1:50:56 the boss needs to solve a problem the
1:50:58 boss is apoplectic we gotta solve this
1:51:04 Chris I can help how do you help I've
1:51:07 been playing with AI oh yeah whatever go
1:51:10 do that you come back an hour later
1:51:12 here's your solution boss
1:51:14 huh how did you do this who who are you
1:51:17 I'm Chris
1:51:20 teach me
1:51:22 be that person
1:51:25 let's go start solving problems for
1:51:27 people learn this stuff by solving
1:51:30 problems for other people it'll be fun
1:51:33 it'll be rewarding and you'll learn this
1:51:35 [ __ ] good and then you will be
1:51:37 appreciated
1:51:39 people are going to think you're [ __ ]
1:51:40 weirdos when you start talking about AI
1:51:42 by the way they're gonna be like Oh you
1:51:45 mean the robots aren't they gonna kill
1:51:48 us
1:51:53 they're going to be the first ones to go
1:51:55 those people
1:51:58 what do you mean my job's been
1:51:59 eliminated you don't like me no we love
1:52:01 you but
1:52:03 everyone who has your job has not got
1:52:06 that job anymore
1:52:08 and because you weren't willing to learn
1:52:10 the new [ __ ] we're gonna go on ahead and
1:52:12 have to let you go
1:52:15 why is Chris still here Chris is now the
1:52:18 VP of AI automation yeah
1:52:23 all right but what do you do to put the
1:52:27 right musical with that uh with that
1:52:29 with that video that I made I mean
1:52:32 if if I were doing that I would probably
1:52:34 just for for an image like that I would
1:52:36 probably just go to some stock music
1:52:38 thing
1:52:41 um there are music AIS if you go to
1:52:43 futurepedia.io and look in the in the
1:52:46 music things there's one called sound
1:52:47 draw that's pretty interesting there's a
1:52:49 bunch of them meta just came out with a
1:52:51 music generation thing the music true
1:52:54 Generation stuff right now is a little
1:52:56 it's a little iffy it's getting close
1:52:57 but it's not quite there
1:53:00 um there's one that just came out called
1:53:02 suno s-u-n-o
1:53:08 hang on oh you've got to go to uh it's
1:53:10 it's on Discord
1:53:13 you generate things on Discord it's
1:53:15 called suno s-u-n-o
1:53:20 [Music]
1:53:29 so soon as one where you write lyrics
1:53:32 and describe the song type and then it
1:53:35 just makes the song with with the
1:53:37 singing
1:53:38 [Music]
1:53:41 remember the night
1:53:44 so that's one sound draw is just kind of
1:53:47 a loop sample songwriter thing but it's
1:53:50 pretty cool
1:53:51 um
1:53:53 you know but I but here's the thing you
1:53:55 don't
1:53:57 just because AI is here you don't have
1:53:59 to use AI for everything you can
1:54:02 like like if you wanted to if you wanted
1:54:05 to automatically
1:54:07 so so here could here could be an
1:54:09 interesting thing
1:54:11 with something like Banger with Bard
1:54:13 that can read images
1:54:16 you could automate
1:54:18 how would you do this
1:54:21 you could have chat GPT you could you
1:54:24 could put this all in an automation
1:54:26 you could have chat gbt write a
1:54:29 voiceover of a 30 second commercial
1:54:32 or 30 second video
1:54:37 you could then have it
1:54:40 describe the person that's saying those
1:54:43 words
1:54:44 oh and then you could you could have it
1:54:47 write the prompt to make that person
1:54:49 then you could have it create an image
1:54:52 of the scene that it's talking about
1:54:56 then you could take the image of the
1:54:59 scene take that into Runway generate a
1:55:03 video clip
1:55:08 also generate a prompt for music for
1:55:11 background music for that scene
1:55:14 and take that to some music generation
1:55:16 too so so basically you could create an
1:55:18 automation that writes all the prompts
1:55:21 to write all the elements of a 30-second
1:55:23 spot
1:55:25 with a voice over background music
1:55:29 video
1:55:30 images and a talking head
1:55:35 and and then you know subtitles and [ __ ]
1:55:37 you know like lower thirds and titles
1:55:39 and things like that
1:55:41 and and that would be an interesting
1:55:42 project right like that would be
1:55:44 something that would be like that would
1:55:46 be an interesting side hustle right you
1:55:47 have a single interface that that has
1:55:49 two or three inputs
1:55:51 of what's the kind of video do you want
1:55:53 you know what's the location describe
1:55:56 the person talking and what's what's the
1:55:58 topic that they're talking about three
1:56:00 inputs and it goes and generates all the
1:56:04 prompts generates all the media stitches
1:56:06 it all together Stacks it all up
1:56:08 publishes it to tick tock
1:56:11 every day
1:56:14 [Laughter]
1:56:20 autocorrect made a joke
1:56:21 [Laughter]
1:56:23 wow so quick yeah I mean the quickness
1:56:26 of these things is just crazy
1:56:28 it named the dog the dog's name is Champ
1:56:32 Champs right beside me Champ's a little
1:56:34 cutie Champ's a cutie my old dog
1:56:40 [ __ ]
1:56:41 great but what would you use to put
1:56:43 music on this we did that one all right
1:56:47 um I did the table example with Roy
1:56:50 Lichtenstein and Andy Warhol that's
1:56:52 really cool little Raven's art nest not
1:56:56 sart Nest ah I get it
1:56:59 so good it's ending your sentences when
1:57:04 you pause
1:57:07 oh yeah it sort of ends my sentences
1:57:10 when I pause but then it continues on
1:57:12 it'll go back and fix words that it got
1:57:15 wrong based on the context of what I say
1:57:18 why is a real pain in the ass if you
1:57:20 pause too long with pie it'll submit
1:57:22 what you put in so if you go talk to pie
1:57:25 if you haven't played with pie go to the
1:57:27 Apple Store I think it's available for
1:57:29 Android too but uh look up Pi Pi
1:57:33 personal assistant get the pie app
1:57:36 and then when you launch it put it into
1:57:38 voice mode and you can just start
1:57:40 talking to their version of chat GPT
1:57:43 it's pretty interesting but you can't
1:57:45 pause too long or it'll submit it and
1:57:48 then answer like the first half of your
1:57:50 thought so you have to learn how to talk
1:57:52 better which I don't I'm not good with
1:57:55 words some people have a way with
1:57:59 English other people not have way
1:58:03 old Steve Martin joke
1:58:06 chord progressions would be like the key
1:58:08 in it sort of yep you could do that
1:58:14 [Music]
1:58:22 um will the code Alpha Kenny Body open
1:58:25 new doors in chat
1:58:27 will the code Alpha
1:58:30 Kenny Body
1:58:33 I don't know what that is I doubt it
1:58:36 but we can try
1:58:39 will it jailbreak it is that is this a
1:58:42 thing let me go to GPT 3.5
1:58:45 Alpha
1:58:48 all right whoops
1:58:52 Alpha Kenny
1:58:54 body
1:59:00 uh wait I've still got my weird
1:59:03 hang on I still got my weird custom
1:59:05 instructions
1:59:07 [Music]
1:59:20 it seems like
1:59:27 oh I get it I get it
1:59:30 it seems like you might be referring to
1:59:33 the phrase I'll [ __ ] your body
1:59:34 [Laughter]
1:59:36 when you say it quickly Alpha Kenny Body
1:59:38 oh you little tricksters with your
1:59:42 trolling yeah it's cute that's cute
1:59:45 Scott ah with the cleverness
1:59:48 making the old man look ridiculous
1:59:51 because he didn't catch that one I
1:59:53 haven't heard that one before that's
1:59:54 pretty funny I asked it to rewrite
1:59:56 Grapes of Wrath chapter one in Korean
1:59:58 and it took 10 seconds yep
2:00:01 do you know Korean
2:00:04 um
2:00:05 great for ADHD I never come up for air
2:00:08 anymore I'm right there with you Silver
2:00:09 Fox this is like the the the the level
2:00:14 of brain overwhelm right now is a real
2:00:16 thing like I just
2:00:20 you know what's funny is I'm more
2:00:23 creatively inspired and more creatively
2:00:25 alive than I've been in a long long time
2:00:28 and
2:00:31 you know it's funny when you when you
2:00:33 get it in a creative Zone
2:00:36 it feels like thanks Silver Fox
2:00:39 um when you get in that creative Zone it
2:00:41 feels like it's going to be there
2:00:42 forever like everything feels
2:00:45 um easy right you're like it's gonna be
2:00:47 there forever
2:00:49 and then it just evaporates in like you
2:00:51 know a week will go by and you just
2:00:53 can't [ __ ] get anything out of your
2:00:55 head
2:00:56 and what these tools are doing for me is
2:01:01 that block doesn't happen anymore
2:01:04 like the minute I want to do anything
2:01:06 creative I just pop over here and it's
2:01:08 like
2:01:11 shh the brain's on fire but
2:01:14 what that means is that there's never
2:01:16 there's not really a break like there's
2:01:18 not really a a moment of like
2:01:26 you know the the the the requisite
2:01:29 self-loathing
2:01:30 that that you use to recharge being
2:01:34 creative because it's just always there
2:01:36 always on
2:01:38 so
2:01:44 all right let me ask answer a couple of
2:01:46 more and then we're gonna wind it up
2:01:48 we're gonna wind it up we're gonna we're
2:01:50 gonna wind it down we're gonna bring it
2:01:52 down let's bring it down we're gonna
2:01:54 simmer we're gonna simmer I'm gonna
2:01:56 finish my little synonym
2:01:58 oh
2:02:00 as Kyle says chat GPT is mansplaining as
2:02:03 a service uh with a raging case of of
2:02:06 correctile dysfunction that's pretty
2:02:09 good yeah this this stuff I mean I mean
2:02:12 uh yeah I didn't come up with the phrase
2:02:14 mansplaining as a service but I've
2:02:16 certainly adopted it that's what these
2:02:17 things certainly what it is uh meaning
2:02:20 chat GPT will confidently answer you
2:02:22 even if it's wrong just like every man
2:02:25 in your life
2:02:27 um
2:02:30 so so especially if you're doing things
2:02:33 for work or if you're going to give
2:02:34 something to the boss or say you're a
2:02:36 lawyer and you write a legal brief and
2:02:38 you have it give you citations for case
2:02:41 law
2:02:42 um it will give you citations for case
2:02:44 law that are probably wrong so check
2:02:47 your work if you're doing anything that
2:02:49 requires any any kind of uh accuracy
2:02:54 um you said something in it looks like
2:02:56 Korean or Chinese oh
2:02:59 oh which means thank you for watching in
2:03:01 Chinese
2:03:02 that's super cool because a lot of
2:03:04 people don't like learning yep watching
2:03:08 you have a great attitude I love your
2:03:10 enthusiasm for AI and your humor thank
2:03:12 you Winston I appreciate those kind
2:03:14 words laughing my ass off honestly
2:03:20 uh year 2050 what's an author
2:03:24 yeah
2:03:26 well well solicit here here's the thing
2:03:28 about about creative work in particular
2:03:32 there's
2:03:34 there's the creative product and the
2:03:36 creative process
2:03:39 and as a creator
2:03:44 I think every Creator in the next
2:03:48 year maybe two
2:03:51 is going to be faced with an existential
2:03:53 crisis of
2:03:57 where's my value or where do I derive my
2:04:01 value
2:04:02 with creative
2:04:04 is it in the creative product
2:04:06 or the creative process
2:04:09 so if it's in the creative product
2:04:12 then you use these tools and you get
2:04:14 really good at these tools and you can
2:04:16 put the product out there
2:04:18 faster cheaper better than you've ever
2:04:21 been able to ever in your life and it's
2:04:23 [ __ ] insane
2:04:25 if where you derive your value as the
2:04:27 creative process
2:04:31 then you might say
2:04:33 [ __ ] it
2:04:34 I'm checking out of
2:04:36 AI
2:04:38 because
2:04:39 I've been an illustrator for 30 years
2:04:42 and I get my pleasure
2:04:44 in the process of illustrating or making
2:04:47 baskets or making sweaters or whatever
2:04:50 it is
2:04:51 I want my hands doing [ __ ]
2:04:57 um and I think I think everyone's going
2:04:59 to be confronted with that and I I don't
2:05:01 think there's a right answer I think
2:05:02 it's a very personal choice
2:05:04 and I think like I totally respect
2:05:06 anyone that says [ __ ] it I'm gonna write
2:05:08 screenplays and never use these tools
2:05:10 I'm gonna do it you know what [ __ ]
2:05:13 computers I'm not even gonna do spell
2:05:14 check I'm going back to typewriters
2:05:18 totally respect that I think that's
2:05:20 [ __ ] amazing
2:05:22 that ain't me
2:05:26 because I'm a lazy bastard if I can get
2:05:29 these things to crank the workout
2:05:32 I'm doing it but I mean you know
2:05:34 that's it that's not going to be an easy
2:05:36 Choice Goldman Sachs estimates seven
2:05:39 percent of U.S jobs will be replaced 63
2:05:42 percent will be complemented 30 will be
2:05:45 unaffected I can't imagine 30 will be
2:05:48 unexpected unaffected I think that's got
2:05:50 to be higher than that and I think seven
2:05:52 percent's low on replaced
2:05:58 the the 63 complimented is interesting
2:06:01 because if if a job is transformed if I
2:06:04 transform from a copywriter
2:06:06 to someone who generates copy
2:06:09 it's a different job but it still might
2:06:12 be me doing it and it still might be
2:06:14 called a copywriter right that might be
2:06:17 one of those Legacy jobs where the
2:06:18 function transforms fundamentally
2:06:22 but the name stays the same that's
2:06:25 that's interesting it'll be interesting
2:06:26 to see what happens but we actually used
2:06:28 our own brain and critical thinking
2:06:30 skills young people won't retain or even
2:06:33 care
2:06:36 oh a classic red convertible yeah that
2:06:38 would have been better thanks for a
2:06:39 great Friday night good night good night
2:06:41 Firefly all right I'm gonna jump to the
2:06:44 bottom here
2:06:46 um it's late it's Friday night you guys
2:06:49 have dates to do or you got coffee to
2:06:52 make or you got
2:06:53 muffins to bake Susie's got more pie to
2:06:57 make it's the weekend
2:07:00 the pumpkin pie was a hit so she's got
2:07:03 pies to make so I think we're gonna wind
2:07:07 this down
2:07:08 um I really do think it's amazing you
2:07:10 hang out with me I appreciate it I hope
2:07:12 this has been fun I hope it helps [ __ ]
2:07:15 typewriters it's calligraphy for me Dr J
2:07:18 you're awesome
2:07:20 um I think you need the creativeness to
2:07:22 generate good copy using AI yeah you do
2:07:25 listen being a good writer
2:07:28 with these tools is going to make you
2:07:31 better at writing with these tools than
2:07:33 a non-good writer uh you know what these
2:07:36 tools actually are is the English
2:07:38 language or whatever your native
2:07:39 language is but in my case the English
2:07:42 language becomes a programming language
2:07:44 that's what these tools are
2:07:46 is you need to figure out logically sort
2:07:48 of how do I game the system to get it to
2:07:52 give me what I want and the better you
2:07:54 are with language and
2:07:56 massaging words the better you are at
2:07:59 manipulating these things so
2:08:02 totally agree with that
2:08:04 this is great Kyle thanks thank you
2:08:06 Silver Fox thanks cabrino great to see
2:08:08 you chaos generation
2:08:14 um such a fun hang tonight it was a
2:08:16 blast all right
2:08:17 um wait here look
2:08:19 this is not quite twerking but
2:08:21 huh
2:08:23 get it
2:08:25 all right
2:08:28 good night everybody see you tomorrow
2:08:30 yeah I'll be around I'll be around now
2:08:32 I'm traveling I leave Sunday for New
2:08:34 York I come back Wednesday I'll probably
2:08:37 be doing these
2:08:38 um from New York but they might the
2:08:40 hours might be weird just because of
2:08:42 meetings and [ __ ] like that and dinners
2:08:44 and
2:08:45 I might I might end up skipping
2:08:47 something but you know hopefully my hair
2:08:49 will be as spectacular as it is tonight
2:08:51 actually it's not bad tonight it hasn't
2:08:53 gone too far off the rails
2:08:56 um yeah but I'll be around 11 45 here
2:08:58 listening to you is edu thank you thank
2:09:01 you very much for the kind words peace
2:09:03 oh Tuesday's the salon that's right oh I
2:09:06 got to do the salon from
2:09:08 New Jack City Ah that's a good idea
2:09:11 that's a good reminder thank you yes I
2:09:13 knew that I knew that I'm the host of
2:09:16 that so I definitely knew that that you
2:09:19 were not telling me anything I did not
2:09:20 know I definitely knew that
2:09:23 um are you getting your iris scanned in
2:09:24 New York City yes I am uh if you're
2:09:27 wondering what the hell that means go
2:09:29 look up worldcoin I think it's.com let
2:09:32 me let me make sure it's the right site
2:09:36 World coin no worldcoin.org
2:09:41 if you want to see what I'm doing
2:09:44 I think it's Sunday Sunday afternoon
2:09:46 maybe Monday
2:09:48 go to worldcoin.org it's Sam Altman's
2:09:50 new company
2:09:52 where I'm putting my biometric data on
2:09:54 the blockchain
2:09:57 [Laughter]
2:10:00 I know it's cool right horrible all
2:10:02 right we knew you went I know
2:10:04 what happened to Kyle I don't know he
2:10:06 went to New York and he last thing we
2:10:08 heard was I see the orb
2:10:11 peace everybody really appreciate it