aiLL 8/22/23 - How AI Is Rapidly Advancing and Why You Should Pay Attention

Description

Chapters

Transcript

0:05 all right
0:08 let's see
0:12 boom
0:14 boom bam bam
0:17 boom boom
0:20 um
0:34 sorry I've been biting people Tobias
0:37 I'll grab here oh is that it
0:40 so I can grab there we go
0:44 welcome Eddie welcome welcome welcome
0:46 welcome happy Sunday night everyone
0:51 still almost 90 degrees here in Denver
0:54 Colorado 8 30 at night still 90 degrees
0:57 good Lord
0:59 looks like La escaped Tobias was shaking
1:03 looks like La escaped major
1:07 rain
1:12 welcome everybody welcome welcome
1:15 my name is Kyle Shannon this is the AI
1:18 learning lab
1:19 I'm wearing like a little golf shirt
1:21 it's like golf Sunday
1:24 how very proper hot here in Wisconsin
1:26 too
1:27 it's crazy isn't it all this weather and
1:30 in California today had a hurricane an
1:35 earthquake and a tsunami warning
1:37 I'll have a good four hour period
1:41 Jesus
1:46 oh
1:47 oh hey Eddie what's happening
1:49 okay farmer was shaking what's going
1:52 down shirt with color I know it's a
1:54 little disconcerting isn't it
1:57 but you know
1:59 I dug in
2:01 athletes
2:03 hey Champion you gonna stay you gonna
2:06 stay and hang out
2:07 you gonna hang out
2:10 lay down go lay down
2:16 all right
2:17 Champy's here to protect to Serve and
2:20 Protect welcome everybody my name is
2:22 Kyle Shannon this is the AI learning lab
2:24 I'll talk about these things here in a
2:26 second a little bit about my background
2:28 I'm an entrepreneur a many-time
2:30 entrepreneur I've started I don't know
2:31 12 or 15 companies
2:33 my current gig is I'm the founder and
2:36 CEO of a company called storyvine which
2:38 is an automated video storytelling
2:40 platform and uh
2:42 we're 11 years into that bad boy so I
2:45 guess it's not really a startup 11 years
2:47 in but it is still a startup
2:49 and we do a lot of work in healthcare
2:51 and Pharma
2:53 why this Channel's here I was I was also
2:56 a Founder back in the mid 90s when the
2:58 World Wide Web first came out and when
3:00 the World Wide Web first came out I
3:02 thought holy [ __ ] this is going to
3:03 change everything
3:04 and then
3:06 uh I co-founded a company called
3:08 agency.com which was one of the very
3:10 first digital agencies in 1994 and
3:14 um it changed everything it was a
3:16 remarkable remarkable time and then I
3:19 didn't think I'd see another one of
3:21 those kind of technological
3:23 earthquakes in my lifetime
3:25 and then I stumbled upon this [ __ ] this
3:28 uh generative AI stuff I started with
3:30 the image Generation stuff and stable
3:31 diffusion and dream booth and then just
3:34 as I was getting my head around how
3:36 profound that was chat gbt launched and
3:39 even though I had used gpt3 there was
3:42 something about chat GPT that was just
3:44 like Ah that's different and 100 million
3:47 people in six weeks thought so too and
3:50 so it was the fastest adoption of
3:52 technology and history and so I have
3:55 been uh you know OCD and in a ADHD hyper
4:00 Focus tunnel for about nine months
4:02 straight now which is also reminiscent
4:05 of the early days of the web where it
4:06 was just like
4:07 I gotta learn everything I gotta learn
4:09 everything and so that's where I am
4:10 that's who I am what I am and that's
4:13 what this is about uh plain T-shirt
4:15 today I miss your t-shirts did you buy a
4:16 New York t-shirt I didn't but I have a
4:18 New York t-shirt I might go change this
4:20 one's a little
4:21 I don't know it's not pleasant
4:26 um
4:27 let's see what else so if you're new to
4:29 this AI stuff
4:32 um that's what these slides are here for
4:33 and so I put them up because
4:38 if you haven't played with chat GPT
4:41 go there right now don't wait
4:44 don't you can ask questions so one of
4:46 the things you can do here is uh you you
4:48 can ask you know questions I'll do my
4:50 best to answer them but in the meantime
4:53 go there if you haven't been there
4:54 before chat.openai.com if you don't know
4:57 at all what to do just type in uh write
5:00 me a country song about blank and then
5:03 just watch what it does and then just go
5:05 holy [ __ ]
5:07 um and then if you want download this
5:10 document go to this document
5:11 prompts.chat and that will teach you how
5:14 to talk to these things what they are
5:17 um why they do what they do or kind of
5:18 how how they do what they do and there's
5:20 a bunch of prompts there
5:22 that you can play with there's lots of
5:23 resources out there but that one's just
5:25 really simple
5:26 Claude and Bing are also like uh chat
5:29 GPT
5:30 um Bing is uh gpt4 but it's there it's
5:33 Microsoft's version of it and Claude is
5:36 a company started by the people that uh
5:39 created
5:40 uh gpt3 at openai and they started a
5:43 competitor called anthropic and claude's
5:46 really interesting because you can put
5:47 in super long prompts so those are those
5:50 if you have played with these things
5:52 before
5:53 here are some other things to explore
5:56 Bard that's Google's entry into this
5:59 space meta's llama 2 models you can now
6:01 go play with those Clint thank you so
6:03 much whoa
6:05 Clint reached member level
6:07 14. Share Bear reached 14 yesterday
6:11 that's awesome hey dragonfly thank you
6:13 for the gummy
6:15 um
6:18 perplexity.ai is interesting if you're
6:20 doing research it it will grab citations
6:23 and sources it's really good and then if
6:26 you go to labs.perplexity that's metas
6:29 llama 2 models that you can play with
6:30 for free without having to install then
6:32 Pi is super chatty if you want to chat
6:35 with it with a GPT model Pi dot AI or
6:38 the pi iPhone app and I think there's a
6:41 an Android app I should really confirm
6:43 that uh but it's super chatty you can
6:45 put it into voice mode and literally
6:47 just talk to it how to Partners we'll
6:50 shake them what's going down thanks for
6:52 the Hat
6:53 thanks for the Hat yellow special your
6:56 special
6:57 I like my peeps we call them the
6:59 Irregulars because they're irregular
7:01 that's why they take the fiber that's
7:04 why they take the fiber all right and
7:07 then po uh po you can play with a bunch
7:10 of different models and it also has a
7:11 bot maker and lots of different Bots
7:13 that other people have made really good
7:15 place to practice your prompt crafting
7:17 so those are some other things to play
7:19 with and then but you're like Kyle but
7:20 this is all words I like pictures okay
7:24 so these are all picture things mid
7:27 journey is the best fireflies the most
7:29 corporately safe Leonardo and clip drop
7:31 are stable diffusion which is getting
7:34 close to mid-journey but it's not quite
7:36 there the drag about mid Journeys you
7:38 have to generate images and Discord and
7:40 if anyone wants me to show that I'm
7:42 happy to do that it's a pain in the ass
7:43 but the images are staggeringly
7:46 beautiful tugas thank you very much I
7:49 got to figure out how to pronounce that
7:50 name but I'm gonna go with two Goss for
7:52 tonight because I think you're in Paris
7:54 or France or something correct or maybe
7:57 a french-speaking other country but I
8:00 know you speak French
8:03 okay
8:05 um futurepedia.io is a list of lots of
8:08 tools oh and and openart.ai prompt book
8:11 that one
8:13 is kind of like prompts.chat but for
8:16 pictures and it'll Teach You How deep
8:18 these things are how much they actually
8:19 understand like if you put in you know
8:22 photographic terms or certain artists it
8:25 will do things in their style absolutely
8:27 remarkable okay
8:30 um one other thing I did get a note back
8:32 from Tick Tock I went through a I found
8:35 a different channel to [ __ ] at them
8:36 about our scammer our happy little
8:38 scammer and they they sent me back a
8:41 note that said this is very unacceptable
8:43 if someone's doing that you should
8:45 report it and so I just responded and
8:47 said I've reported it six times and you
8:49 keep telling me there's not an issue but
8:51 for those of you new here there's an
8:53 issue
8:55 this channel is the top one
8:58 the stuttering one down here is a
9:00 scammer AI Lala learning lab is not me
9:03 now they've taken my videos and they're
9:07 acting like me and they've used this
9:09 [ __ ] name and they've got like 3 000
9:12 followers now and they're asking their
9:14 three thousand followers to join some
9:15 multi-level marketing scheme [ __ ]
9:18 so feel free to report those [ __ ]
9:23 uh and they're ruining the party for the
9:26 rest of us so anyway that's going on all
9:29 right so if you have questions pop them
9:31 in the comments below happy Sunday
9:32 everybody I'm coming to you live from
9:35 Denver Colorado
9:36 and let's get started let's get started
9:40 let's do this let's do it let's just do
9:43 it let's just do it you ready to do it
9:46 you ready to do it let's just do it
9:48 why does he talk like that I don't know
9:50 he's he's he seems touched he seems
9:53 touched
9:55 he seems touched why are we here is he
9:58 qualified what are his qualifications
10:01 I don't think he's qualified
10:03 I'm definitely not qualified
10:07 but I do like to have fun and isn't that
10:10 what it's all about
10:11 and you know you just might learn so
10:14 it's called the AI learning lab if if I
10:16 were not qualified could I call it the
10:19 AI learning lab I don't think well I
10:21 could yeah I did
10:22 but that's not the point that's not the
10:24 point that is really you know
10:27 you people with your logic it's very
10:29 clever it's very clever what you do
10:33 thank you user22503.
10:34 [Laughter]
10:39 foreign
10:43 would smell as sweet
10:50 all right seriously we're gonna get
10:52 started now it's gonna be so good it's
10:54 gonna be so worth the wait like 20
10:55 people are just like I'm out of here
10:57 this dude's weird all right
10:59 the Irregulars know how irregular I am
11:02 and they're okay with it so if you're
11:04 not okay with my irregularity then
11:08 then I don't know what to say
11:12 um you too can make money watch out your
11:14 big day
11:16 thank you Bob for reminding me of that
11:17 I'm looking forward to my little button
11:20 coming
11:21 I ordered the gold
11:24 Easy Button
11:26 and we're gonna record that and put it
11:28 in there I'm gonna have to come up with
11:29 some little stand or something that I
11:31 can just sort of slide in and hit it
11:32 right it's gonna have to be a visual gag
11:36 because it's gold I got a gold one
11:40 um wow fast response thanks you're
11:42 welcome
11:44 uh
11:45 happy Sunday happy Sunday to you too
11:48 Lori
11:49 plain shirt today I know I'm gonna I
11:51 think I'm gonna go change I'm gonna put
11:52 something
11:53 you are so totally awesome awesome love
11:56 me
11:57 um what's next for storyvine well so
12:02 a couple of things going on with
12:03 storybine so over the years what
12:06 storyline was used for was things like
12:07 patient stories and doctors making
12:09 videos about you know how they treat a
12:12 certain condition or how they use a
12:13 certain drug in the past three years
12:15 because of covid a lot of Pharma
12:18 companies sales reps
12:20 um didn't have access to doctors
12:22 so a lot of our clients are now using
12:24 storybine for their sales reps to make
12:26 personalized videos to send to doctors
12:29 and that's been really interesting
12:31 so that's a piece and then because I've
12:34 been so you know fully consumed with
12:36 this AI stuff one of the reasons I've
12:39 been consumed with it is I look at what
12:41 storyline is and what we've done for the
12:42 past 11 years and I look at what's
12:45 coming with AI and I'm like there's a
12:47 big part of what we do where the value
12:50 is going to go to zero
12:51 like automation like part part of our
12:54 value proposition right now is that we
12:56 automate video production which is super
12:58 cool we're not using AI believe it or
13:00 not
13:01 um
13:02 but I look out a year from now and
13:04 everything's going to be automated so
13:05 like the value of that effectively drops
13:07 from whatever it is now to zero
13:10 but one of the things we're really good
13:12 at is you know human beings sharing
13:14 their stories authentically I feel like
13:16 the value of that goes up so I've got
13:18 one part of my value proposition that
13:21 goes to nothing and the other part of
13:22 the value proposition that accelerates
13:26 um
13:28 and then I started looking at things
13:30 like well you know what else could you
13:32 do
13:34 with this thing that we've created that
13:37 that creates these authentic videos
13:40 well I can take the transcripts of those
13:42 videos which are now really good thanks
13:44 to Ai and that's built into the
13:46 storyline platform we have interactive
13:48 transcriptions and now I can generate a
13:50 bunch of other content
13:52 around each video so for every video in
13:54 the system I can generate a blog post
13:56 and tweets and
13:59 um you know YouTube descriptions and
14:01 hashtags and pull quotes and things like
14:04 that so
14:06 that's led me to a whole vision for what
14:09 I'm calling an omni Channel content
14:11 engine
14:12 where we can take almost anything on the
14:15 input side
14:16 and then we can output almost anything
14:18 on the output side and at the center of
14:21 this thing is authentic narrative
14:23 which I think will differentiate it
14:24 because I think there's going to be 8
14:26 billion automated content generators so
14:29 I think we we might have a way to have
14:31 something different in the marketplace
14:33 so and and our clients seem to be
14:35 responding to it I was in New York last
14:37 week because the client wanted to talk
14:39 about that
14:40 which is good
14:42 you can use chat gpt4 with Bing
14:45 including code interpreter as long as
14:47 you have the 20 bucks uh wait although
14:51 yes you can so if you pay the 20 bucks
14:54 for chat GPT at openai you get code
14:58 interpreter you get plug-ins it's no
15:01 longer connected to the internet
15:03 um the chat the the browse with Bing
15:05 function was removed like two months ago
15:08 um you can go to bing.com and use that
15:12 but that doesn't have code interpreter
15:13 so it's a bit of a confusing mess right
15:15 now
15:17 um but yes
15:19 um and that's why I have the three of
15:21 these here I I feel like code
15:22 interpreter and gpt4 are the best at
15:25 open AI
15:26 Bing is connected to the internet these
15:28 other two are not
15:30 um clawed you can put in really big
15:32 things so so I feel like these three is
15:35 a little triumvirate are good and the
15:37 nice thing is the bottom two are free
15:39 and Frank Frankly chat GPT 3.5 is free
15:42 um up here but I think code interpreter
15:45 and plugins are worth are worth playing
15:47 with with those guys
15:49 all right have you heard have you met
15:51 Nathan
15:53 bought on LinkedIn Storyteller I have
15:56 not Nate
15:57 um if you know him
15:59 um feel free to introduce me
16:03 here's my level 14. oh yeah
16:08 Clint rock in a level 14.
16:11 AI learning lab just like Scientology
16:14 you gotta level up all the time or
16:16 you'll never be a Satan or hit fade
16:18 Satan Theta Satan I don't know I don't
16:21 know enough about scientology to know
16:23 like
16:24 it's powerful
16:29 it's special we're in as special
16:34 yeah Mr L Ron Hubbard Mr science fiction
16:38 himself Bow Wow Wow let's see who is he
16:42 was inspired by who could it be okay oh
16:44 oh I know Satan
16:50 p i equals chatty GPT right that's
16:54 correct that's the new oh yeah I forgot
16:55 to say that so uh if you go to
17:00 Pi dot a i
17:04 there's chat GPT and then there's chatty
17:07 GPT it's very chatty if you don't have a
17:09 conversation that one's fun to play with
17:11 you want to get [ __ ] done not as much
17:13 fun to play with
17:15 but there are times when you want to
17:18 chat
17:19 and it's actually kind of interesting
17:25 what AI would you recommend to build
17:28 logos and slogans well
17:31 hmm
17:37 I would say
17:41 for the slogans uh gpt4 Chad gpt4 so
17:45 that for the slogans
17:48 and also if you have a good image you
17:51 can put it into you can upload an image
17:53 to Bing and have it write copy about the
17:55 image it can read images bing bing can
17:57 read images but just chat GPT you can
18:01 what's cool about chat GPT and how you
18:03 can use these tools is
18:06 you can have chat GPT if you if you
18:08 don't have a creative brief for example
18:11 you could have chat GPT help you come up
18:13 with a framework to create a creative
18:15 brief and then have it ask you a series
18:17 of questions about the client or the
18:19 product
18:20 and then have it write a brief and then
18:22 once it writes the brief you can say now
18:24 give me a bunch of slogans based on that
18:26 brief and it will do it and it'll be
18:29 really good because it will have written
18:30 a brief and followed it so that's how I
18:33 would use that for the logos
18:36 um mid-journey is still in my opinion by
18:39 far the best but
18:40 the logos don't expect them to have good
18:43 text in them the text will be a mess so
18:47 that you'll have to
18:49 you know do in Photoshop or illustrator
18:53 my gut is telling me I haven't done logo
18:57 stuff in a while my gut is telling me
18:59 it's probably worth using Adobe Firefly
19:01 or generative fill
19:04 um probably more Firefly because they've
19:06 got that text tool that can generate you
19:09 know you can you can put in a prompt
19:10 about like you know
19:12 fuzzy text and it'll make furry text in
19:16 whatever font you choose it's also
19:18 commercially it's not been trained on
19:21 copywritten images
19:23 um and it's also just Adobe like like
19:26 even if their tools aren't perfect right
19:28 now
19:30 the the level of integration that these
19:33 generative AI tools are going to have
19:34 within the Adobe Creative Suite so you
19:37 know illustrator and Photoshop and
19:38 things like that I would I would
19:40 probably lean toward
19:42 um getting good with with firefly
19:45 um to do my to do my logo stuff but
19:49 I still go back to Mid journey is the
19:51 best right now
19:52 but none of them do text well none of
19:54 them zero zero
19:57 apparently there's a new one from open
19:59 AI
20:00 that is supposed to be better than
20:02 mid-journey that does text but we ain't
20:05 seen it they ain't released it it ain't
20:08 it ain't out in the public yet
20:10 it's just the fancy scientists and the
20:13 researchers over there at open AI they
20:15 get to play with that and we don't and
20:17 that's not fair I don't think that's
20:18 fair but that's the way it is that's
20:21 just the way it is it's they're powerful
20:23 powerful scientific people and the
20:26 scientific people they treat us a little
20:29 different you know they're like you
20:32 can't handle the truth you can't handle
20:34 the text
20:36 so you'll just take your little pretty
20:37 pictures and you'll have to deal with
20:39 your own text
20:41 so that's what they do
20:43 that's not how I am that's how they are
20:45 I'm just a messenger don't hate on the
20:47 messenger don't hate on the messenger
20:55 that was a little bit like the baskets
20:57 character Zach galifianakis's brother
21:00 brother realism Brother Hey brother what
21:03 are you doing
21:04 what are you doing here
21:07 do you have a website uh storyvine.com
21:10 is my company's website uh Kyle
21:12 shannon.com is my photography website I
21:16 have a photography website
21:18 Kyle Shannon dreams is my Instagram and
21:21 Tick Tock art project
21:24 AI learning lab is this Kyle Shannon's
21:27 my Instagram
21:29 I don't know
21:30 I've been around for like decades I got
21:33 [ __ ] everywhere
21:37 yes
21:39 storyvine Kyle shannon.com and then some
21:41 other [ __ ]
21:44 um 8 39 in Denver that was
21:47 10 minutes ago
21:50 all right see his C has not been here
21:53 before so see you don't have to ask me
21:54 questions a lot so here's here's the
21:57 other thing that goes on here and this
21:58 frustrates the [ __ ] out of people who
22:00 are not the Irregulars why are they the
22:03 Irregulars because they deal with my
22:05 [ __ ] so here's one of my shits this is
22:07 one of my things
22:09 um
22:10 I like to answer all the questions but
22:12 what that means is you'll be typing
22:14 questions down below and and I don't see
22:17 that [ __ ] I'm just answering whatever
22:19 question I'm answering right now so I'm
22:21 not watching the comments in real time
22:23 so people that are not used to that
22:24 they're not used to Elder talk where us
22:28 old farts are using technology we
22:30 shouldn't be
22:32 so anyway so I will get to your
22:34 questions ask the AI why my parents
22:37 don't appreciate me well that's what you
22:40 know that's one where where a little
22:42 conversation with pie might actually be
22:44 interesting
22:45 um where you could potentially work
22:47 through some of that [ __ ]
22:49 um that's that is I you know I hope
22:52 you're joking
22:53 but but you know I get it I have a
22:57 family ask me why my family doesn't
22:59 appreciate me I understand
23:03 it is a human condition to want to feel
23:06 heard want to feel loved and
23:08 you know
23:10 the world just slaps you upside the
23:12 [ __ ] head doesn't it
23:16 and then the art is
23:19 just figuring out a way to keep it going
23:22 just keep showing up that's why I do
23:23 these every night it was funny I didn't
23:26 I was like in a cranky mood right before
23:27 this I didn't want to come on I'm like I
23:29 don't want to do it
23:30 and then I just did it and then I got on
23:32 and I'm looking at my shirt and I'm like
23:33 this is a shitty shirt you know what I'm
23:35 gonna go change my shirt that's gonna
23:36 make me happy but I'm glad to be here
23:38 and even though I was cranky before I'm
23:40 not enough so I will be right back
23:42 entertain yourselves and I'll be back do
23:45 me a favor if you would
23:47 um share the love this will be good
23:48 share the live and hit the likes and
23:50 then when people come so just see some
23:53 blank shelves I'll be right back
24:04 thank you
24:31 all right we got Slinky patent night
24:35 it's Slinky pattern tonight
24:39 and again it's not quite twerking but I
24:41 can do this
24:46 all right
24:49 Touch by the turtles I like turtles
24:57 [Music]
25:10 um
25:11 professional singer why yes
25:16 um I like
25:24 back here live at the waterfront village
25:27 with my friend the zombie Jonathan
25:29 you're looking to Jonathan just got an
25:31 awesome face paint job what do you think
25:33 I like turtles all right your great
25:36 zombie and good times here at the
25:39 waterfront Village open for the next 11
25:41 days oh my God
25:43 her trying to not laugh is
25:47 it's comedy gold
25:53 and she choked on a caramel apple
25:58 at the Old State Fair
26:04 oh Tick Tock just jumped me to the
26:06 bottom of the comments hello tick tock
26:09 not only do you let people that are
26:11 stealing my identity keep scamming
26:13 people but you jump me to the bottom of
26:15 the comments when all I want to do is
26:17 answer everyone's questions
26:19 I want to be comprehensive I want to be
26:22 helpful
26:24 all right your link tree shows under
26:27 maintenance FYI really
26:33 link
26:39 real Kyle Shannon
26:42 no
26:45 hmm I wonder if I have a bad Link in the
26:49 how did you get there
26:51 Lori
26:52 because it looks like it's live to me is
26:55 it the is it the AI
26:57 salon
26:59 hang on
27:06 slash a i
27:11 artist Salon no that looks like it's
27:14 there I might have a bad link somewhere
27:15 if you can tell me where the link is
27:17 I'll go fix it but it looks like stuff's
27:19 still alive
27:20 is still alive to me it looked like
27:23 still alive to me
27:25 but I don't know
27:27 someone's out there calling themselves
27:29 the AI learning lab
27:32 I'm sorry about their speech impediment
27:34 logos and slogans and logos and slogans
27:37 Mr Shannon it appears that you've been
27:38 drinking straight ice water correct
27:41 Emilio's wife tonight is just about
27:44 hydration
27:47 and the fact that there wasn't any
27:49 coffee left over from this morning and I
27:51 didn't feel like making any because I'm
27:53 lazy yeah
27:56 and a while with Doug and Wendy weiner
27:59 we've got diverticulitis
28:03 is it expensive to use open Ai No so
28:08 let's go through these
28:09 it will take your time away because if
28:12 you start using this so here's here's my
28:14 philosophy on this you should use these
28:17 tools one of them
28:20 long enough for you to have your first
28:22 Kevin McAllister moment number Home
28:24 Alone
28:25 that scene
28:29 at some point you'll do something with
28:31 these large language models where they
28:33 will absolutely melt your face and blow
28:35 your mind and if you haven't hit that
28:37 moment yet just keep keep digging keep
28:39 playing and that's why I put the
28:41 prompts.chat thing there
28:43 um
28:43 the 3.5 there's a at chat GPT you can
28:48 play with the 3.5 model in the four
28:50 model the four model is really good like
28:53 better than anything else
28:55 but 3.5 is really good and really fast
28:57 3.5 is free
28:59 four isn't
29:01 um you also get a thing called code
29:02 interpreter here which is really good
29:04 but if you just want to play with the
29:05 thing 3.5 is fine claude.ai is free and
29:09 it's got a hundred thousand token
29:11 context window which if you're brand new
29:12 to this stuff you're like but Kyle what
29:15 the hell is a token and why would they
29:16 call it a token and what's the context
29:17 and what's a window what are you talking
29:19 about doesn't matter
29:23 um
29:23 you can put in to the prompt like you
29:26 know like the Google bar
29:28 at cloud.ai you could put in an entire
29:32 50 000 word novel
29:34 just paste it in there and then say
29:36 write me an epilogue to this novel and
29:39 it will
29:40 so so you can put in giant prompts at
29:43 cloud.ai
29:45 um Bing is also for and it's free Bing
29:48 is also free wesn radio what's happening
29:51 thanks for the music
29:53 um Bing is also free it's also connected
29:55 to the internet it can also generate
29:57 images you just say create an image of
29:59 blank and it will I can also read images
30:02 so you can upload a picture of your
30:04 puppy and say write me a poem about what
30:08 you see in the picture and by jingus it
30:10 will
30:11 so uh half free free free
30:15 then let's go to these
30:17 free free
30:19 free and then have free so almost all of
30:23 these are free that's why I have them
30:25 here
30:25 because a they're good and B they're
30:27 mostly free
30:29 um the image stuff mid-journey minimum
30:32 is like 10 bucks a month Firefly
30:34 requires a creative uh Cloud license
30:38 um Leonardo's got a free version It's
30:40 kind of have free you can do I don't
30:42 know like eight eight or ten images a
30:44 day with Leonardo but it's really good
30:46 and then clip drop all the tools at
30:48 clipdrop.co are free so not free
30:53 um sort of included if you pay the
30:55 exorbitant fees for Photoshop uh with
30:58 firefly Leonardo half free and clip drop
31:01 totally free clip drop is from stability
31:04 AI the people that make stable diffusion
31:06 all right
31:08 so that hopefully will will heal it will
31:12 heal oh the API uh wait is it expensive
31:16 to use oh the openai API
31:20 um it's not expensive
31:22 uh embedding can be expensive so
31:25 embedding is more expensive than just
31:27 using it
31:28 um but expensive is relative if you're
31:30 not doing a lot of volume if you're not
31:32 doing
31:33 if you're not experimenting with
31:35 autonomous agents which can which can
31:37 just you know run up your open AI API
31:40 bills if you're doing things like zapier
31:42 automations where you're generating some
31:44 content you know a few times a day or
31:47 even a dozen times a day I don't think
31:50 I've had an open AI bill that's been
31:51 more than for the API that's been more
31:53 than five bucks like it's it's
31:55 relatively cheap for low volume if
31:57 you're going to do high volume stuff
31:59 then it can start to add up but but um
32:02 you if you go to uh the uh
32:07 the address is platform.openai.com
32:11 slash playground
32:14 and then click on the documentation
32:16 within documentation there's API pricing
32:18 it's it's pretty cheap and it's per
32:20 thousand token back to that token thing
32:22 where Claude can do a hundred thousand
32:24 tokens
32:26 um the openai apis are per thousand
32:29 tokens and it's like 0.00 you know you
32:32 know 0.0
32:34 2 cents for a thousand tokens or or two
32:38 cents for a thousand tokens depending on
32:40 the function you're using in the model
32:41 you're using so but it's it's certainly
32:44 for doing prototyping it's very very
32:47 cheap
32:48 g'day Laura C how are you good day good
32:51 day welcome
32:53 beep poop from user2250360-39-7241
33:00 uh yeah listen we don't pass much
33:04 judgment here but I'm gonna go on ahead
33:07 and give you a 0.5 out of 10 on the name
33:11 come on you can come up with something
33:14 how about irregular
33:18 2250-360-397-241 huh that would be
33:21 clever
33:22 all right uh where is he streaming LOL
33:25 doesn't feel like he's here hmm that
33:28 must mean I'm actually reading all of
33:30 the questions not just what was that
33:34 out of that beautiful
33:36 it's still beautiful that was so cool I
33:40 I got a little Club a little little baby
33:42 seal
33:43 the phrase Club Seal popped in my head
33:46 because I grew up in a different era
33:50 where uh yeah they talked about that
33:53 [ __ ]
33:55 a little baby seal and a whale jumping
33:58 that was beautiful thank you for that I
33:59 appreciate that all right
34:02 um I work for a pharmaceutical
34:03 distributor owned by a manufacturer can
34:05 you share which companies you work with
34:08 sure I work with
34:10 we work with Novartis we work with
34:14 Pfizer we work with Merck we work with
34:17 BMS we work with boring Behringer
34:19 ingelheim we work with a bunch of small
34:22 ones
34:23 we work with most of the healthcare
34:27 agencies so we you know we've been
34:30 around for more than a decade we've been
34:31 in healthcare and Pharma most of that
34:33 time so
34:35 um
34:35 so yeah I'm happy to chat DM me I've got
34:38 my contact information's on my link
34:40 Theresa feel free to to reach out if you
34:43 want to chat
34:45 um me
34:46 yes you
34:50 user2250-360-397-241 yes you
34:52 he seems to respond to different
34:54 comments he probably streams
34:57 simultaneously
34:58 huh
35:00 or maybe I'm just behind
35:05 what is it isn't it Occam's razor
35:09 the most simple explanation
35:12 is often the correct one
35:15 it could just be that I'm old and I
35:18 don't use tick tock the way younger
35:20 people do
35:22 and it seems counter-intuitive and
35:25 stupid like I'm a dum-dum
35:28 because I am
35:32 okay
35:33 all right okay enough comedy jokes
35:36 Never Enough comedy jokes
35:40 all right all right
35:43 Sophie as in is this live yes this is
35:46 live my cousin loves your page can you
35:49 give him a birthday shout his name is
35:51 Dixon my ass well Dixon my ass happy
35:54 birthday to you
35:56 happy birthday Dixon how old is Dixon
36:00 there's Dixon a a younger Feller an
36:04 older Feller
36:05 cousin I would assume younger
36:09 um happy birthday what are you doing
36:11 Dixon are you are you playing are you
36:13 playing with the AI are you digging in
36:17 have you written a term paper for school
36:19 and not gotten caught
36:21 [Laughter]
36:24 all that all that
36:26 um paper detection stuff is just gonna
36:29 be like comically thank you dragonfly
36:31 it's gonna be like it's gonna be comical
36:33 to look back on remember when we used to
36:35 try to figure out if people were using
36:37 AI because everyone's going to use AI
36:40 all the time in every tool like a year
36:42 from now this is a non-conversation but
36:45 it's the beginning of the school year
36:47 and you know some teachers are a little
36:49 sensitive still about oh my God my
36:51 students might actually cheat
36:57 I think you know education is gonna have
36:59 to completely completely
37:03 transform itself and that's going to be
37:05 painful for a lot of people and that
37:06 sucks the the thing that
37:10 like the the part of me that loves
37:12 Innovation and loves
37:14 what's next is really excited about this
37:17 stuff but like the empathetic side of me
37:20 the part of me that got an acting degree
37:23 um
37:25 it's really upsetting like like the like
37:27 what's just the amount of change people
37:30 are going to be forced to deal with is
37:32 going to be painful it sucks I don't
37:34 like it I don't like it at all I don't
37:37 like it I don't like it
37:40 what's your view on Last Stand that AI
37:43 created movie
37:44 um I think it's quite good it's you know
37:46 they they
37:48 um
37:49 they synthesized all the voices they did
37:51 all the imagery in mid-journey I think
37:53 they used Runway to do the animations
37:56 um it was well crafted I they might have
37:59 even written a script with with chat EBT
38:01 it was well uh well crafted it was well
38:05 structured
38:06 um there's I showed it the other night
38:08 there's a there's a trailer for
38:11 barberheimer where where it's like you
38:14 know Barbie and I Oppenheimer mixed I
38:16 think that one's a bit more
38:18 sophisticated than the Last Stand
38:20 um and I think and I think it's just
38:22 funnier and better
38:24 um but the Last Stand thing is
38:27 that's the writing on the wall right
38:29 that's the canary in the coal mine
38:32 for the writer's Guild agent oh I'm just
38:36 a tick tock professional I don't know if
38:37 you knew this but I'm a Tick-Tock I'm
38:39 I'm what they call in the industry and
38:42 influencer okay I influence people with
38:47 my influential tones
38:50 and think that's what that means
38:53 um
38:54 what was I talking about oh yeah writers
38:58 strike Actors Guild strike all of all of
39:01 movie production
39:03 should look at the Last Stand and say
39:07 and
39:10 critique it all they want
39:15 and then say to themselves quietly or
39:17 out loud
39:18 that's as bad as this shit's ever gonna
39:21 be
39:23 and it's not bad
39:26 it's not good like if you're a filmmaker
39:28 you look at the Last Stand and you're
39:30 like wow this is the frame rate and the
39:31 man it's the color depth and the the
39:34 audio the acting the writing the bit but
39:35 all correct all correct
39:38 you know it's a from a if I'm a
39:42 filmmaker at any kind of Hollywood level
39:44 even Indie level
39:46 I look at the Last Stand and I go
39:50 C minus d-plus
39:54 but that's as bad as it's ever gonna be
40:00 and this stuff is moving quickly I was I
40:03 was listening to Paul ratesers he's got
40:04 the someone I if you're into this stuff
40:07 I recommend listening to he's got a
40:09 podcast uh he's got an organization
40:11 called the marketing AI Institute and
40:13 he's got a weekly podcast
40:15 and his co-host Mike asked him about
40:18 Runway ML and and this last week Runway
40:20 ml went from generating if you don't
40:23 know Runway ml they've got this thing
40:24 called Gen 2 which is text to video so
40:27 you type in you know woman running
40:29 across a parking lot and it will make
40:32 four seconds of video of a woman running
40:34 across a parking lot
40:36 this week they announced that the
40:39 maximum length of those clips is now 18
40:42 seconds not four
40:44 so it was four
40:46 now it's 18. that was in two months Paul
40:50 rates are you know Mike said you know
40:51 what do you think of that you know 18
40:53 second thing and Paul writer said it's a
40:56 4X increase of capability in two months
41:01 so if you look at the Last Stand I'm
41:03 pretty sure almost all of the Clips in
41:06 The Last Stand and most of these AI
41:07 films
41:08 all of the edits are less than four
41:10 seconds
41:11 because that's all it could do and if
41:14 you tried to like extend it it was kind
41:16 of weird and perverted but now they can
41:18 do relatively contiguous action for 18
41:21 seconds
41:25 so what do you think that's going to be
41:26 in a year
41:28 it's going to be multiple minutes it's
41:29 probably going to be multiple scenes it
41:31 will probably have
41:33 at a minimum music potentially voice
41:38 so I think that
41:41 I think that the Last Stand is just a
41:43 good it's it's one of those things we'll
41:46 look back on and go yeah that was that
41:48 was the first one of those
41:51 you know I if you're if you're a Gen X
41:53 or if you're old enough to remember when
41:55 QuickTime first came out when Apple
41:57 released QuickTime the movies that you
42:00 could play were like this big and they
42:02 had like a like a it wasn't even 10
42:05 frames a second it was like five frames
42:07 a second
42:08 video and like on your on your 13-inch
42:11 CRT computer monitor the video is like
42:15 this and everyone made fun of it
42:18 and that was the foundation for
42:21 everything we've got today right
42:23 streaming services right
42:25 so so you've got to look at something
42:27 like the last stand or just Runway MLS
42:30 you know
42:32 think about how remarkable it is
42:35 I would like video
42:37 of
42:39 a plane landing on an aircraft carrier
42:42 and it will make you that video
42:44 now
42:45 is the plane gonna be some weird
42:47 perverted thing that we've never seen
42:49 before probably and is the aircraft
42:53 carrier gonna be weird and the waves be
42:55 janky and like you know smash through
42:56 the side of the ship at some point
42:58 probably but a year from now
43:03 I'd like a plane landing on an aircraft
43:06 carrier Nimitz class
43:08 sure
43:10 F-16 F-18
43:13 uh could you do an F-22 sure
43:19 so
43:20 I think it's a big deal
43:23 none of these tools right now none of
43:26 these tools
43:27 are
43:30 even remotely close
43:33 to what they're going to be like they
43:35 are
43:36 spit and duct tape versions of what's
43:39 coming they're just they're silly
43:43 and what's mind-blowing about it is if
43:46 you've ever
43:47 if you haven't played with these and you
43:49 go in and you ask chat GPT to write you
43:52 whatever a project plan or a marketing
43:55 strategy or have it write a blog post
43:58 for you or have it uh interrogate you to
44:02 get enough information for for then it
44:04 to them go off and do something or you
44:06 have it write some code and you've
44:08 always wanted to Learn Python but you
44:11 never really got around to it you don't
44:13 really have the time and maybe you have
44:14 ADHD like me and like learning code is
44:17 it's easier to just smash your face into
44:20 a brick wall
44:22 well now you can code
44:24 and it'll do all that [ __ ] I said and
44:26 it'll do it like that now is it doing it
44:29 at the level you can do it at probably
44:31 not
44:32 but imagine a year from now Chachi BT
44:36 didn't exist until November of last year
44:42 so this November will be coming up on
44:44 one year that that the thing that
44:47 changed everything
44:48 existed
44:52 and then imagine a year from now what
44:54 it's going to be
44:57 so don't look at any of these tools for
44:59 what they are today not one of them
45:01 they're all just
45:04 glimpses into the future
45:08 but have you ever seen a sci-fi movie
45:10 where they just talk to the machine and
45:11 it actually talks back to them and
45:13 actually knows everything
45:15 go talk to pie
45:17 you're like well I've talked to Alexa
45:18 and I've talked to Siri no it's a
45:21 fundamentally different experience those
45:24 were programmed by human beings to
45:26 anticipate your answers and if you ask
45:28 them something that's out of the bounds
45:29 of what they were programmed to say they
45:31 can't [ __ ] do it
45:33 pie can
45:37 it's just a different thing
45:40 so
45:41 yeah
45:43 getting all philosophical here on Sunday
45:45 night it's Dixon's birthday and I'm like
45:47 you know we're gonna give Dixon a little
45:49 gift
45:51 screw them teachers that are like we're
45:53 gonna use an AI detector screw them
45:55 these tools are going to be everywhere
45:57 if you've got teachers that are doing
45:58 that I don't even know how old Dixon is
46:00 I'm assuming you're a student
46:05 that they're not doing their jobs
46:08 they're not paying attention
46:11 these are the most powerful tools ever
46:13 put in the hands of human beings and we
46:15 get to play with them
46:19 I don't know if you get how lucky we are
46:21 I like I felt this in 1994. when I
46:24 stumbled upon this thing called the
46:26 World Wide Web I was playing with the
46:28 internet with command line stuff and
46:29 like logging into libraries and looking
46:31 at research papers and I'm like oh this
46:33 is kind of cool they can share
46:34 information I sort of got it
46:37 and then the World Wide Web came out I'm
46:39 like what's this thing and then the and
46:40 then the the first browser came out
46:44 thank you Adam thank you for the losers
46:47 I very much appreciate
46:50 you sharing them
46:51 um
46:53 I went to this thing that looked like a
46:55 page and it had these words that were
46:56 underlined and if you clicked on them
46:59 it jumped you to another
47:02 piece of information
47:05 and I sort of got it
47:07 because you could just I didn't have to
47:09 type command lines now I didn't have to
47:11 type FTP commands I could just click on
47:13 [ __ ] click click click click right
47:17 this is what Steve Jobs said when he
47:19 when he when he launched the iPhone
47:20 right he talked about how are you going
47:22 to control it with a stylus he goes no
47:25 we don't need a stylus
47:27 we've got the world's best pointing
47:28 device right here click click click that
47:31 was the iPhone right that was the world
47:33 wide web
47:35 that's chat GPT
47:38 you just put in some words and
47:41 outcome answers not links to other [ __ ]
47:43 where you can go find the answers you
47:46 get the answers
47:47 no they always right no
47:50 hey share bear
47:51 thank you for the lollipop
47:55 um
47:58 so we're just at the at the
48:01 very very beginning of this stuff
48:04 and so like why do I do this every night
48:06 why do I come here every night other
48:08 than for the donuts and the lollipops
48:11 in The Tick Tock hats and the twerking
48:14 because I think that
48:18 any of us that put any amount of time
48:20 into getting our heads around this [ __ ]
48:22 thanks Diva
48:24 oh
48:27 are are going to be disproportionately
48:30 given opportunities for people that
48:33 don't
48:33 and people that don't get their heads
48:35 around this are going to be blindsided
48:38 by it they're not going to know what
48:39 happened
48:41 and I want as I
48:42 like I'd like to make a little dent in
48:45 that
48:46 that you know if 10 fewer people are are
48:50 blindsided by this because of this
48:53 channel then it will have been worth it
48:54 so that's why this is here
48:57 what do I think will happen to software
48:59 Engineers
49:00 um
49:01 well I think it's I you know I just kind
49:03 of answered it with the last stand with
49:05 filmmakers I think software engineering
49:08 will fundamentally change
49:11 um I think the Tactical execution of
49:14 almost everything whether it's coding or
49:17 writing or generating art for commercial
49:21 purposes anyway
49:23 um or
49:25 you know Financial Services or Insurance
49:28 underwriting or just like any tactical
49:31 execution execution reading legal briefs
49:35 doing screenplay coverage
49:37 any tactical execution is going to be
49:39 automated
49:41 and so that means well does does that
49:44 mean Engineers go away no not
49:46 necessarily I think Engineers Engineers
49:48 are going to be more important than ever
49:50 I think that thank you very much share
49:52 bear
49:54 um I think that
50:00 AI operations is going to become
50:02 critical like how do you keep up with
50:03 this [ __ ] this this stuff is is
50:05 literally every two months it's a
50:07 complete transformation like what was
50:10 what worked and was revolutionary two
50:13 months ago is completely different today
50:15 so if I'm in an I.T group and I'm trying
50:17 to figure out how the hell I keep my
50:20 company up to date with this stuff and
50:22 how I keep up to date with security and
50:24 all that sort of stuff I feel like that
50:27 stuff is that's five to ten years of
50:30 never-ending work just trying to keep up
50:32 with this [ __ ]
50:34 if you go on the more sort of
50:35 application development side if you say
50:38 okay the the coding itself the Tactical
50:42 execution of the code that gets
50:44 automated what doesn't get automated the
50:47 critical thinking
50:48 right the critical thinking that
50:51 Engineers do where they say oh here's
50:53 the business problem here's the business
50:55 requirement how do I get there well I'm
50:58 going to use this AI tool and it's going
50:59 to go squirt out some code but then me
51:01 as the engineer I look at that code and
51:03 go oh no that's not a good approach
51:05 that's really inefficient why because I
51:06 know how to do this [ __ ] right so people
51:09 that have deep experience are going to
51:12 be really good at this
51:15 but they have to be willing
51:19 to
51:21 accept that the thing that was their
51:23 superpower maybe for decades
51:26 might be different now
51:30 and some people are going to make that
51:31 transition and some people are not
51:34 and that's where it's going to be
51:35 painful
51:36 but the good news is I think
51:39 some of the people that don't make that
51:40 transition let's say they were an
51:42 engineer and they just love coding and
51:43 they just want to be an engineer and all
51:44 of a sudden they're competing competing
51:46 with tools that just do it automatically
51:49 some of those people go [ __ ] it
51:51 I don't want to be a conductor of the
51:54 symphony I want to be a coder and I'm
51:56 just going to go code for my own
51:58 pleasure off on the side cool
52:00 what those people might also do is go
52:02 let me just play with these tools let me
52:04 just see what these things do other than
52:05 coding
52:06 oh I could take my coding skills and
52:09 apply that to prompting images and I
52:12 could generate really cool art so I
52:14 think you're going to see a lot of
52:15 transformation where people are in a
52:17 career over here they figure some [ __ ]
52:19 out with these tools and go you know
52:21 what I always wanted to do hang it and
52:23 they go into this new area hang on I
52:25 gotta close the door
52:36 oh
52:46 why isn't he answering my question I
52:49 asked it like an hour ago and he hasn't
52:51 gotten to it I don't even think he's
52:52 live
52:53 I think he's streaming on Twitch
52:57 I think he's secretly a gamer
53:02 Force multiplier friend indeed this
53:05 stuff Force multipliers these tools are
53:08 ability amplifiers absolutely
53:11 here's the deal if you're good at coding
53:14 if you're if you're a kick-ass software
53:16 engineer and you start embracing this
53:18 [ __ ] you're going to be a kick ass here
53:21 software engineer you're going to be
53:24 better than someone else who just
53:26 squirts code out of these things and
53:27 doesn't know anything about it and
53:29 that's where I think that's where I
53:30 think some people are going to really
53:32 arise as [ __ ] Superstars because they
53:36 get their heads around this [ __ ] and
53:37 they're like oh wait a minute I know how
53:40 to do this like I know how to do all of
53:43 the you know getting [ __ ] done over here
53:45 and this is going to take the crap that
53:48 normally takes me eighty percent of my
53:50 time is just going to be done like that
53:53 and then I can spend the rest of that
53:55 time refining and improving
53:58 the software yeah
54:01 and this thing will also document the
54:03 code oh you know how
54:07 you know how at the beginning of every
54:09 software project that I've ever done I
54:12 say to the engineering team and by the
54:14 way documentation is really important
54:16 we're going to document everything right
54:18 we're going to document and comment the
54:20 code right correct uh-huh yeah
54:24 do you know how many times that's
54:25 happened in my life
54:31 and I'm old
54:32 zero
54:35 because I'm always in a startup
54:37 so you're always changing [ __ ] and
54:39 you're always doing hot fixes and you're
54:40 always doing this and you're always
54:41 doing that so the best intentions for
54:43 documentation go right out the [ __ ]
54:45 window well what if they didn't have to
54:51 right
54:54 yes
54:55 it could be fleas
55:00 good Lord
55:04 all right
55:06 what do you think will happen to
55:08 software Engineers we just answered that
55:11 um it's a real live now will not be a
55:14 very good answer in the future
55:16 huh wait it's a real
55:19 it's a real live no will not be very
55:20 good answer in the future yeah I'm I'm
55:22 behind
55:24 there's there's a there's a lot of
55:26 commentary about me being behind Stevo
55:28 what's up buddy hello everyone hey Stevo
55:31 uh Bobby asking again about software
55:33 Engineers I'll be back after an hour
55:35 then good awesome although I answered
55:38 your question now again wait what's your
55:40 what's your name
55:41 now and I am now and I am now no
55:47 when
55:48 I
55:50 all lowercase letters in the name with
55:52 no spaces make me insane 2.5 out of 10.
55:56 um hello what are your thoughts on llama
55:57 and open source AI
56:00 um
56:02 my thoughts are
56:06 it was probably irresponsible for meta
56:09 to have done it
56:10 but it's done
56:11 [Laughter]
56:14 and so now uh stability AI has also done
56:18 it and it looks like uh the company
56:20 behind Alibaba just did it I think they
56:23 just open sourced
56:25 they're seven billion parameter model
56:27 that apparently they did some optimized
56:30 version of it and open source that like
56:32 two days ago
56:33 um
56:37 there's a reason these things have
56:39 safety guard rails on them and so the
56:42 open source
56:44 tools
56:46 make it much easier for for Bad actors
56:49 to do bad [ __ ]
56:51 um
56:53 foreign
56:55 of mixed feelings because
57:01 what what
57:04 the open source models essentially
57:06 guarantee well I don't have to guarantee
57:08 it
57:09 there's there's a much higher likelihood
57:12 that
57:17 that a a handful of two or three
57:20 companies open AI Google
57:24 Apple
57:26 you know whoever they might be whoever
57:27 whoever ends up being the sort of three
57:29 primary players
57:31 with the open source models out there it
57:34 lowers the likelihood that they're going
57:36 to consolidate all the power and sort of
57:38 collude on how these things are going to
57:39 be used and and sort of where the power
57:41 gets distributed for for these tools
57:45 so
57:47 let me look at something here
58:08 okay
58:10 it's my sales guy saying you don't have
58:12 to get up early for that meeting they
58:13 just pushed it a week
58:15 um
58:18 so so the
58:21 so what they kind of Ensure is that
58:25 these tools aren't going away
58:28 and that we get access to really
58:29 powerful tools
58:31 and it's also
58:36 had chat GPT launched and
58:39 no one cared
58:42 um we wouldn't this channel wouldn't
58:45 exist we wouldn't be talking about this
58:47 the open source models wouldn't be in
58:49 the world
58:50 I mean these companies Google meta Apple
58:55 um
58:57 open AI
58:59 they've had
59:01 powerful models for years
59:05 and they haven't released them
59:08 and open AI did and and they did before
59:11 Chachi BT they you know they released
59:14 the the gpt3 API and that was what
59:17 things like Jasper and copy AI were
59:19 written on top of before chat EBT came
59:21 out
59:23 but it was kind of
59:24 all still off in the corner like no one
59:27 was really worried about it
59:29 and when Chachi PT came out and went to
59:31 100 million users in six weeks and
59:34 Google panicked and said this is a big
59:36 deal it was Google's response to chat
59:40 GPT hitting 100 million users that
59:42 quickly
59:44 that opened the floodgates of this thing
59:46 so at this point
59:48 it doesn't really matter it we're we're
59:50 in Woulda Shoulda Coulda land
59:54 we might have wanted to think about
59:59 how these things might be used before we
1:00:01 just turn them loose on the [ __ ]
1:00:03 world
1:00:04 but
1:00:08 Chachi PT did what it did Microsoft went
1:00:11 from being kind of a laughable player in
1:00:14 the search space to truly competing with
1:00:16 Google and putting them on their heels
1:00:18 which was remarkable like like could you
1:00:21 have anticipated
1:00:22 that in the game of search Microsoft
1:00:25 would even be in the conversation much
1:00:28 less leading the conversation that
1:00:31 happened [ __ ] overnight
1:00:34 and Google panicked and called you know
1:00:36 they called it their what did they call
1:00:38 it their Red Alert they it's a red alert
1:00:40 at Google like you know
1:00:43 they called meetings they they made
1:00:45 announcements that were embarrassingly
1:00:47 bad and they made other announcements
1:00:49 that were also embarrassingly bad the
1:00:51 tech was embarrassingly bad they were
1:00:53 behind
1:00:54 because I think they had made a
1:00:56 calculated call to not
1:00:57 lean into this stuff because it was an
1:01:00 existential threat to their to their
1:01:02 entire business model
1:01:04 and so this thing came out and they're
1:01:06 like [ __ ] now we gotta deal with it and
1:01:08 when they did that everyone went and
1:01:10 then meta
1:01:12 poor Mark Zuckerberg oh poor Marky Mark
1:01:16 he he changed his whole company to this
1:01:19 thing called the metaverse and everyone
1:01:21 looked at it and went in that just like
1:01:24 second life from the year 2000 isn't
1:01:27 that it's not it's no different than
1:01:29 that right isn't it the same
1:01:32 nobody gave a [ __ ]
1:01:36 and all of a sudden everyone's like it's
1:01:38 all about AI
1:01:39 and you know meta has an earnings call
1:01:41 where they don't mention the metaverse
1:01:43 once I think it was the second quarter
1:01:46 it was maybe the first quarter of 23
1:01:49 earnings call
1:01:51 they didn't mention the word metaverse
1:01:53 once
1:01:55 their [ __ ] name is meta
1:02:00 so they had to do something and so they
1:02:03 said
1:02:03 well here's a research preview
1:02:08 but but you can only use it if you're a
1:02:10 researcher in the development Community
1:02:11 was like awesome [ __ ] it let's go and
1:02:14 they just started playing with it and
1:02:16 then and then Zuckerberg probably
1:02:18 realized this is my this is my suspicion
1:02:20 this is total speculation
1:02:24 why don't we give them commercial
1:02:26 license to everyone except you know our
1:02:29 direct competitors so if you have
1:02:32 um less than 700 million active
1:02:34 customers you can use it for free
1:02:35 commercially so let's put that out in
1:02:38 the world we'll put out the the 7 the 13
1:02:41 the 40 and the 70.
1:02:44 and the 40 didn't pass uh red teaming
1:02:47 but whatever so so we got three of them
1:02:50 and what I suspect is going to happen is
1:02:52 meta at some point is going to go oh hey
1:02:54 forgot to tell you
1:02:56 we've been working on this 175 billion
1:02:59 parameter thing that like kicks the [ __ ]
1:03:01 out of gpt4 and uh
1:03:05 you can play with that too like all the
1:03:06 [ __ ] that you've developed is gonna work
1:03:08 with this but uh that this costs a penny
1:03:11 a call just pay us a penny a call for
1:03:13 all your API calls so they'll probably
1:03:15 do something
1:03:17 you know they will commercialize it but
1:03:19 but they're in the conversation because
1:03:21 they open source this [ __ ]
1:03:24 so I don't know I'm I'm of mixed
1:03:27 feelings we'll we'll have to see yeah I
1:03:29 mean listen it's [ __ ] remarkable as a
1:03:31 developer as someone who wants to play
1:03:32 with this stuff as someone with Pharma
1:03:34 clients
1:03:35 where they're going to be like I'm not
1:03:37 comfortable sending any patient data to
1:03:40 open AI unless it's in some her
1:03:43 medically sealed thing which Microsoft
1:03:44 is working on right
1:03:46 but the open AI terms the service are
1:03:48 really shitty so if I can say to my
1:03:50 client yeah we're going to do this all
1:03:51 on a on a llama 2 model in a
1:03:54 hermetically sealed you know environment
1:03:56 where no calls are going to anyone
1:03:58 they're going to be like awesome go play
1:04:01 go invent us some cool [ __ ] so
1:04:04 so it's it's ultimately really good it's
1:04:08 just I don't know that we're ready for
1:04:11 what it is I don't think we know what
1:04:12 these things are I don't think we know
1:04:14 how people are going to use them
1:04:17 so I don't know I don't know I do not
1:04:20 know I do not know uh looks I've got a
1:04:23 bunch of new people here welcome
1:04:24 everybody if you're new here my name is
1:04:25 Kyle Shannon this is the AI learning lab
1:04:28 why I have these things up here we
1:04:30 haven't talked about this in a minute
1:04:33 um if you haven't played with chat GPT
1:04:35 go to that top URL chat.openai.com
1:04:38 that's the official chat gbt website
1:04:40 just start playing see if you can have
1:04:42 yourself a little Kevin McAllister
1:04:43 moment
1:04:46 remember Home Alone
1:04:48 um or go to Claude or go to Bing and
1:04:51 play with the chat there just go play
1:04:53 with these things and if you don't
1:04:54 really know what to say to them go to
1:04:55 prompts.chat and that will teach you
1:04:57 what these things are and how they work
1:04:59 and that's kind of what we're talking
1:05:00 about here it's what are these things
1:05:01 how do they work all that sort of stuff
1:05:03 Bing does have code interpreter does it
1:05:06 is that new
1:05:08 where is it let me see
1:05:12 Bing being
1:05:15 DD
1:05:23 [Music]
1:05:26 um
1:05:31 dude I'll just ask it do you have
1:05:36 X as to code interpreter
1:05:43 let's see what Bing says
1:05:45 no I do not have access to it tell it oh
1:05:50 thank you for the drink is that a
1:05:51 lemonade it's a that's a refreshing
1:05:53 little lemonade it looked like a boozy
1:05:55 lemonade
1:05:58 I like Daddy like Daddy like
1:06:01 um Bob if you'd be willing to tell me
1:06:03 how I get to code interpreter from Bing
1:06:06 because I don't see it but it doesn't
1:06:09 mean it's not there
1:06:11 there's a pin
1:06:14 I can upload an image
1:06:17 I can talk to it anyway uh I'll see if
1:06:21 you put something down below okay
1:06:24 let's see two two two
1:06:27 should I do a reset and just jump to the
1:06:29 bottom of the comments I think I feel
1:06:31 like at this point I'm pretty far behind
1:06:33 on comments will be funny when AI is
1:06:36 able to predict
1:06:37 when you will sell your trades it's
1:06:39 probably on the way to that
1:06:41 that's just the way it is some things
1:06:43 always change better believe it
1:06:45 something that's just the way it is 8 52
1:06:49 in Denver that was
1:06:51 8 52 40 minutes ago so I'm 40 minutes
1:06:56 behind on comments
1:06:58 I'm gonna make the executive call I'm
1:07:00 jumping to the bottom so I'm gonna get
1:07:02 caught up so I'm probably gonna skip
1:07:04 some people's questions
1:07:06 um Bob proud to be a regular irregular
1:07:08 thanks Kyle for making the knowledge
1:07:11 more about the beginning of AI
1:07:13 absolutely thanks for being here I
1:07:14 appreciate it so I'm going to jump to
1:07:15 the bottom and then scroll up a little
1:07:17 bit all right
1:07:18 all right so I'm at the bottom I'm
1:07:20 scrolling up wait let me see if Bob said
1:07:22 anything about where to find
1:07:24 Bing Bang Boom
1:07:26 Big Bang Boom
1:07:28 it was a river of ejaculate
1:07:33 oh my God that scene the outtakes from
1:07:36 that scene
1:07:39 I pleasured myself wait I pleasured
1:07:42 myself to the image of Mrs Butterworth's
1:07:45 [Laughter]
1:07:49 they couldn't keep a straight face
1:07:52 even he couldn't his his mustache kept
1:07:55 popping off oh that was funny
1:07:58 [Laughter]
1:07:59 talk about a river of ejaculate
1:08:03 [Laughter]
1:08:06 oh good lordy lordy lordy
1:08:10 all right let's see just had
1:08:14 brother 52 just had it right my first
1:08:16 day quiz very cool are you a teacher I
1:08:19 think I remember you saying Chris Ward
1:08:21 if I wanted to build my own platform for
1:08:23 AI would meta open source be good for
1:08:26 that yeah and it so Chris there's a
1:08:29 couple of things you can play with um
1:08:32 zapier which is an automation tool has a
1:08:35 thing in in beta called interfaces right
1:08:38 now where you can build chat Bots and
1:08:41 you can build web interfaces using
1:08:43 zapier automations and just you can just
1:08:45 build chat Bots you know calling uh chat
1:08:48 EBT so that's an option you can you can
1:08:52 use poe.com if you just want to create
1:08:55 chat Bots that do a specific thing go
1:08:57 play with the PO the bot builder at
1:09:00 poe.com
1:09:02 um
1:09:03 and then if you want to do
1:09:05 um if you want to build your own sort of
1:09:06 chat platform yeah the The Meta models
1:09:09 are interesting there's a project called
1:09:11 GPT for all that has installers for OS X
1:09:15 windows and Linux that will basically
1:09:18 install a framework on your computer
1:09:20 where you can download whatever large
1:09:23 language models you want put it in there
1:09:25 and and start doing it that way
1:09:27 and then the final thing I would say is
1:09:31 um is go explore the Lang chain
1:09:35 project
1:09:36 there's probably lots of projects but
1:09:39 the core project is is this idea that
1:09:42 you can
1:09:42 train these large language models on
1:09:44 your own data so you could build a whole
1:09:47 interface where it's trained on very
1:09:48 specific data what just came out
1:09:51 um
1:09:56 the hell just came out oh McKinsey
1:10:00 McKenzie just came out with
1:10:04 this was kind of news let's see
1:10:07 McKinsey
1:10:09 um
1:10:10 chat bot I think it's called Lily
1:10:13 is it Lily
1:10:26 Mackenzie has more than 70 experts that
1:10:28 work on Lily
1:10:30 it was interesting Lily is is the the
1:10:34 their model their their robot model is
1:10:36 named after their first female employee
1:10:39 I think
1:10:41 another first female consultant or
1:10:43 something like that yeah
1:10:45 um
1:10:49 when did this come out how long has this
1:10:51 been out August 16th so this came out
1:10:53 four days ago
1:10:54 meet Lily our generative AI tool
1:10:57 that's a researcher a Time Saver and an
1:11:00 inspiration so this thing is just like
1:11:03 chat gbt except it's searching like
1:11:07 all of McKenzie's data and research
1:11:11 and I this to me feels like the future
1:11:14 right where
1:11:18 large language models are going to be
1:11:20 ubiquitous they're just going to be like
1:11:21 IP Internet Protocol right like right
1:11:24 now you don't think about Internet
1:11:26 Protocol because it's just everywhere
1:11:29 same with large language models they're
1:11:31 going to be [ __ ] everywhere and it's
1:11:33 not going to matter which model you use
1:11:35 or this or that all of that stuff's
1:11:36 going to drift into the background right
1:11:37 right now it's exciting because
1:11:40 you know they're getting better and we
1:11:42 can compare them but at some point
1:11:43 they're going to normalize and that's
1:11:44 going to drift to the background so
1:11:46 what's going to remain well
1:11:48 what are you using them on
1:11:51 well I want business intelligence okay
1:11:53 McKenzie's been doing business
1:11:55 intelligence for [ __ ] decades
1:11:59 so why not go to a trusted source and
1:12:02 you know research the [ __ ] out of this
1:12:04 one and then go to Bloomberg and
1:12:07 research the [ __ ] out of Bloomberg GPT
1:12:08 that's trained on all their financial
1:12:10 data
1:12:11 and then go to you know IMDb GPT
1:12:16 and be able to pull up all the movie
1:12:18 information you've ever wanted and use
1:12:20 code interpreter to make charts and
1:12:21 graphs
1:12:22 So within 15 minutes you'll be able to
1:12:25 Traverse
1:12:27 you know major
1:12:30 uh like trusted data sources in in kind
1:12:34 of a seamless multimodal
1:12:36 way that's just going to be insane and
1:12:39 powerful and you'll pay for this [ __ ]
1:12:41 why because it's not all the [ __ ] of
1:12:44 all the internet this is McKenzie stuff
1:12:46 like all of this stuff's been vetted now
1:12:49 is this any good I don't know
1:12:51 but it's at least the right idea
1:12:59 just had to write my first date quiz
1:13:03 Vicky the least twerking comment uh what
1:13:06 did Vicky say
1:13:11 fleas are twerking on Kyle's head oh
1:13:13 that's good
1:13:14 exactly
1:13:15 exactly boring demo test is back to
1:13:19 troll Hey Demo test what's happening
1:13:22 buddy
1:13:24 boring
1:13:26 Kyle's not talking about science he's
1:13:29 going to talk about pictures again and I
1:13:31 don't like it
1:13:32 pictures are a waste of time
1:13:36 we love our trolls good to see you demo
1:13:38 test welcome I hope you get a hug
1:13:40 tonight
1:13:41 if I wanted to build my own thing okay
1:13:43 we did that
1:13:45 um just had it right my first day quiz
1:13:46 okay oh that's right we went back to see
1:13:47 Vicky we'll be funny when fake AIS will
1:13:50 be answering fake trolls trolley eyes on
1:13:53 a fake live oh and that will happen
1:13:55 that's
1:13:58 hey so so you know what's funny like we
1:14:00 we talked about this last night I had a
1:14:02 I had a password meltdown where
1:14:06 whatever LastPass didn't have the right
1:14:08 password in it you know the password
1:14:10 manager didn't manage my passwords
1:14:12 correctly and then you know I had to
1:14:14 recreate a password it was just it was a
1:14:16 [ __ ] nightmare and it hit me that the
1:14:18 I I if you don't know Sam Altman's got
1:14:21 this company called World coin and one
1:14:24 of the components of worldcoin is a
1:14:26 thing called World ID
1:14:28 where and and I will admit this is
1:14:31 creepy as [ __ ]
1:14:32 you get your human your human identity
1:14:36 your human uniqueness verified on the
1:14:39 blockchain with this this device called
1:14:41 the orb
1:14:42 I don't think they could have picked
1:14:44 worse branding for for the Sci-Fi
1:14:47 apocalypse crowd but that aside so I did
1:14:50 this I went to New York City you can
1:14:51 only do it in New York Miami and San
1:14:54 Francisco right now
1:14:55 so I got my world ID so it's so it's
1:14:57 just a blockchain it's just an nft and
1:14:59 it says Anonymous and it's got you know
1:15:01 my identity and it's you know it's my
1:15:03 secret key
1:15:05 but what hit me is holy [ __ ]
1:15:10 that could be you know the replacement
1:15:12 for passwords is you know I'm I'm a
1:15:15 verified human right connected to this
1:15:18 blockchain ID well someone could write
1:15:21 an infrastructure for passwords right so
1:15:23 just like you've got Google single
1:15:24 sign-on someone could do World ID single
1:15:26 sign-on and I could just zap it with my
1:15:30 world ID and never have to do another
1:15:32 password again I was like oh that's
1:15:33 interesting and then you know that
1:15:35 comment uh took us about
1:15:39 probably within a year
1:15:42 it will be essentially indistinguishable
1:15:45 to watch the
1:15:48 synthesized version of Me versus me now
1:15:50 me because I change voices and I'm kind
1:15:54 of wacky like I might be a little less
1:15:55 predictable maybe it takes two years to
1:15:57 do me
1:15:58 but some of that's just a bit more
1:15:59 straightforward
1:16:01 um certainly if you're in like
1:16:02 broadcasting and you know everything's
1:16:04 by the book
1:16:08 you're not going to be able to tell that
1:16:09 that's a human
1:16:11 and when there's no lag on the GPT they
1:16:15 will be able to respond so just like I'm
1:16:17 shitty at responding to comments in real
1:16:19 time
1:16:20 those things will be good at it they'll
1:16:22 be able to read all the comments filter
1:16:23 them
1:16:24 prioritize them in real time
1:16:27 and answer and say oh Togos good to see
1:16:30 you
1:16:31 it'll feel real
1:16:37 crazy it's crazy
1:16:39 so you have this world ID thing
1:16:43 we could create a channel likely if Tick
1:16:45 Tock ends up supporting something like
1:16:47 World ID you could create a channel
1:16:49 where you can't subscribe to that
1:16:51 channel unless you're a verified human
1:16:54 and you know they accept World ID and
1:16:56 meta ID and Google ID whatever you know
1:16:59 whatever the verified tools are that
1:17:01 that rise to the top
1:17:06 that's going to be novel
1:17:08 in not too long to be able to be in a
1:17:10 space where you know that everyone in
1:17:12 there is human
1:17:15 as it's [ __ ] insane it's [ __ ] it's
1:17:17 like this is like you know this is like
1:17:19 the Sci-Fi stuff right
1:17:21 so
1:17:23 elementary statistics rather 52
1:17:28 elementary statistics
1:17:31 Elementary dear Watson that was my
1:17:34 favorite plays I did I did it was a
1:17:36 crucible of blood
1:17:38 directed by Paul Giovanni at the uh
1:17:41 at the only theater in Washington DC it
1:17:44 was I was right out of college I got
1:17:46 cast it was an equity show so I Got My
1:17:48 Equity card
1:17:50 and I played like a young lap boy in the
1:17:55 First Act and I played
1:17:57 um an Indian dude
1:18:00 um in the second act
1:18:02 and uh
1:18:04 I was it was it was the summer before I
1:18:07 was supposed to go to grad school I had
1:18:08 a full ride at the Alabama Shakespeare
1:18:10 Festival
1:18:11 and I called him up two weeks before
1:18:13 before I was supposed to go down there
1:18:15 and and start their grad program and I
1:18:18 said uh I just got my Equity card I'm
1:18:20 moving to New York City
1:18:22 guy was pissed you'll never work in this
1:18:25 town again I'm like you're in [ __ ]
1:18:27 Birmingham Alabama I think I'm good with
1:18:29 that I'm going to New York
1:18:34 they had a really nice program down
1:18:35 there in a nice building but
1:18:37 it didn't beat getting your Equity card
1:18:39 moving to New York City with your Equity
1:18:41 card that was a pretty big deal
1:18:43 I missed the trolls they're back Demo's
1:18:46 back he's here they should be panicking
1:18:49 they deserve it I don't know who that
1:18:50 was in reference to but yeah [ __ ] them
1:18:53 oh oh Google
1:18:54 meta keeps me in Facebook jail I know
1:18:57 these big companies are [ __ ]
1:19:00 nightmare like like the fact that
1:19:02 they're there's a there's there's
1:19:03 someone that has copied the content from
1:19:06 my account and put it on theirs
1:19:10 they've got technology where they could
1:19:12 know that the videos on this account are
1:19:15 the same as the videos on this account
1:19:17 let's do something simple like compare
1:19:19 the start dates of the two accounts and
1:19:22 then go look at that
1:19:24 like they have that [ __ ] technology
1:19:25 if they can turn my eyes bright blue and
1:19:27 give me [ __ ] eyelashes they can
1:19:29 figure out if a count has stolen someone
1:19:32 else's content is
1:19:34 it hurts
1:19:37 hmm
1:19:41 932 in Denver and comments filtered I
1:19:45 don't know what that means wondered
1:19:46 about filter thing I don't know what I
1:19:48 didn't see it so
1:19:50 uh if if it was like comments were
1:19:53 filtered because someone said something
1:19:54 jerky it was probably someone like demo
1:19:56 like being a jerky jerk or something I
1:19:59 don't know
1:20:00 open AI talking bankruptcy they're not
1:20:03 talking bankruptcy unless something
1:20:05 happened today
1:20:06 um there were articles written about
1:20:09 because chat GPT because openai is
1:20:12 spending about seven hundred thousand
1:20:14 dollars a day a day
1:20:17 if they don't get some new financing
1:20:19 thank you Dominic for for the
1:20:22 if they don't get some financing they're
1:20:25 going to blow through their 10 billion
1:20:26 dollars per darn quick
1:20:30 I mean you know I've gone through seven
1:20:33 hundred thousand dollars in a weekend
1:20:34 but a day come on that's ridiculous
1:20:38 um
1:20:40 I just don't
1:20:42 I just don't see it happening I don't
1:20:44 Microsoft is not gonna let
1:20:48 open ai go under they're betting they're
1:20:51 betting the future of Microsoft
1:20:54 on this platform
1:20:56 so they're not gonna it's not gonna go
1:20:58 it's not gonna go it's not like those
1:21:00 are good dramatic headlines
1:21:02 uh at this rate they're gonna be
1:21:04 bankrupt by December okay
1:21:07 and in October someone will announce a
1:21:10 15 billion dollar investment in open Ai
1:21:12 and that'll get them through April and
1:21:14 then someone will announce a 30 billion
1:21:17 dollar investment in open Ai and they'll
1:21:19 get them through
1:21:20 July
1:21:23 and before the end of next year Sam will
1:21:26 have been you know gotten the hundred
1:21:28 billion dollars he thinks it's going to
1:21:30 take to get to AGI
1:21:32 and we'll have artificial general
1:21:35 intelligence by mid-2025.
1:21:38 or they go bankrupt
1:21:39 [Laughter]
1:21:42 that's where we are
1:21:45 but again it doesn't see
1:21:48 it mattered if openai went bankrupt if
1:21:51 they were kind of the only player in
1:21:52 town but because chat GPT did what it
1:21:55 did
1:21:56 and because it forced Google's hand
1:21:59 and you know anthropic said I don't like
1:22:02 how open AI is doing their safety I'm
1:22:04 going to start a whole company just to
1:22:06 do safety [ __ ] that's what that's what
1:22:07 anthropic is the the the people behind
1:22:10 Claude were like I don't like what
1:22:11 they're doing at open AI I don't think
1:22:12 they have the right idea we've got our
1:22:14 own idea so they're out there now
1:22:17 Microsoft you know
1:22:20 chat GPT is in
1:22:23 GitHub co-pilot it's in you know Bing
1:22:26 search
1:22:27 it's going to be in Windows 11. it's
1:22:28 going to be an Office 365 it's going to
1:22:30 be in Dynamics 365. it's all over Azure
1:22:33 cloud services
1:22:35 so they're going to be able to charge
1:22:38 money these are all business services
1:22:39 with the exception of search all the
1:22:42 rest are Business Services they can
1:22:43 charge money for from Big corporations
1:22:47 so they're Microsoft they're not gonna
1:22:48 they're not gonna price themselves out
1:22:50 of business
1:22:52 and if they have to they'll say well if
1:22:54 if Sam couldn't handle being on his own
1:22:56 [ __ ] it we'll buy the other 51 and we'll
1:22:59 save them but it but it's like it's
1:23:02 it's irrelevant
1:23:03 at this point
1:23:06 um
1:23:07 who else Google Now
1:23:10 can't
1:23:12 not do this right
1:23:15 their last
1:23:18 their last uh developer announcement I
1:23:21 think I think the CEO said AI like
1:23:25 48 times
1:23:28 it's it's core to what they're doing you
1:23:30 know meta has open sourced this thing so
1:23:32 they're going to come out with something
1:23:34 pi
1:23:36 just raised 1.3 billion dollars
1:23:39 I heard of another uh
1:23:43 a uh a cloud services company that has
1:23:46 46 000 gpus just raised two billion
1:23:51 dollars in debt
1:23:55 to
1:23:57 keep their
1:23:58 infrastructure go like the [ __ ] ain't
1:24:01 going anywhere
1:24:03 do the players change does does maybe
1:24:05 Sam Altman Corkscrew this thing into the
1:24:08 ground yeah maybe
1:24:09 but I you know
1:24:13 if if we were having that talk
1:24:16 back in October before they announced
1:24:19 you know launched chat gbt that's one
1:24:22 thing but with all this stuff out there
1:24:25 I just I don't I don't see it I don't
1:24:26 see it
1:24:27 I could be wrong ah you know listen I am
1:24:30 not a financial analyst this is not
1:24:32 investing advice I would not take
1:24:34 investing advice from me don't worry we
1:24:37 will all surprise the world we are more
1:24:38 educated than anyone
1:24:40 cool
1:24:42 we will surprise the world how and be
1:24:44 smarter than the AI
1:24:46 um I don't care how educated you are
1:24:49 what what's coming like
1:24:52 go ask that thing anything
1:24:55 go ask it [ __ ] anything
1:24:58 ask it stuff you don't know anything ask
1:25:00 it stuff you do know something about it
1:25:03 knows a lot Bing Bang Boom
1:25:08 I pleasured myself to the image of Mrs
1:25:11 Butterworth's
1:25:15 ah
1:25:17 oh I love Smith comedy loves me to
1:25:19 Comedy
1:25:20 uh AI filmmaking can't do worse than the
1:25:23 room the road the room that that was
1:25:26 that film with with
1:25:28 hello Mark
1:25:30 what's that
1:25:32 hang on
1:25:34 um
1:25:35 YouTube
1:25:40 the room
1:25:43 Maybe
1:25:55 [Music]
1:26:09 oh this is the room this isn't from the
1:26:11 movie wait hang on
1:26:19 it's not true it's [ __ ] I did not
1:26:23 hit her I did not
1:26:25 oh hi Mark hey Johnny what's up I have a
1:26:28 problem Felicia she said that I had what
1:26:31 well did you no it's not true don't even
1:26:34 ask
1:26:35 what's new with you
1:26:37 I'm just sitting up here thinking you
1:26:38 know
1:26:39 I got a question for you yeah you think
1:26:42 girls like to cheat like guys too
1:26:44 what makes you say that I don't know
1:26:47 I have a problem oh [ __ ] I did not hit
1:26:50 her I did not
1:26:52 oh hi Mark oh hey Johnny what's up hi
1:26:55 Mark
1:27:02 I have Mark that did not hit her I did
1:27:05 not
1:27:06 aha mark
1:27:08 oh
1:27:12 McKenzie not McKenzie's of GPT boss oh
1:27:15 yeah that's funny it's if if you've been
1:27:17 paying attention to Mackenzie of GPT
1:27:19 boss he's like he's like pivoting like
1:27:22 40 Ways to Sunday he kind of realized
1:27:24 GPT boss wasn't going to be the thing
1:27:26 and then and then when uh when chatgpt
1:27:30 launched custom instructions he's like
1:27:32 well that's my [ __ ] business so he
1:27:34 he's completely resetting that thing
1:27:37 it's fascinating watching what's cool
1:27:39 about McKinsey if you don't follow
1:27:40 McKenzie if you're if you're into this
1:27:41 AI stuff Mackenzie Bose I think the
1:27:44 channel is called GPT boss back in
1:27:46 January he created this
1:27:48 kind of uh chat GPT rapper called GPT
1:27:51 boss that was all these different
1:27:52 personas like you had a a copywriter and
1:27:54 a marketing person and a HR person and
1:27:57 you know all these different personas
1:27:58 and then you could you know hire them in
1:28:02 quotes
1:28:03 um to help you run your your business
1:28:06 um and he's been very very transparent
1:28:08 about what he's building what he's going
1:28:10 through what his pricing is
1:28:12 um he's he's got lots of opinions he's
1:28:14 he's this you know
1:28:17 super smart like Wicked smart Canadian
1:28:20 dude big awesome red beard
1:28:23 and uh he's an interesting one to to uh
1:28:29 follow
1:28:31 that's that was the word I was looking
1:28:32 for you know every now and again
1:28:35 um welcome if you're new here my name is
1:28:37 Kyle Shannon this is the AI learning lab
1:28:38 let me let me pop through these
1:28:41 go to this top URL right here
1:28:46 chat.openai.com that's the official
1:28:48 chatgpt website won't cost you anything
1:28:50 it it will if you want to do the fancy
1:28:52 [ __ ] but just basic GPT 3.5 is free
1:28:56 you can also play with bing.com chat hit
1:28:59 their chat button that's gpt4 as well
1:29:01 and it's connected to the internet or go
1:29:03 to claude.ai that's free and it's really
1:29:05 good
1:29:06 and then prompts.chat will teach you how
1:29:09 to talk to these things
1:29:11 um this URL right here is our YouTube
1:29:14 channel it's AI learning lab Dash TT for
1:29:17 tick tock there was an already an AI
1:29:19 learning lab there so just Dash TT for
1:29:22 us and I'm archiving these lives there
1:29:26 um some other things to look at are
1:29:29 bard.google.com that's Google's entry
1:29:32 into this space not bad is connected to
1:29:34 the internet is free can read images
1:29:37 metas llama 2 models are there Pi dot AI
1:29:41 uh someone in the one of the Irregulars
1:29:44 here in the channel uh nicknamed pie
1:29:47 chatty GPT and that's the accurate that
1:29:50 that is the right name Pi is chatty GPT
1:29:53 it's very verbose which if you put it in
1:29:56 voice mode and talk to it it's actually
1:29:58 kind of appropriate and it's it's
1:29:59 interesting to have a conversation with
1:30:02 something that you can literally ask
1:30:03 anything and it will have some answer
1:30:05 and it's connected to the internet so
1:30:07 you can ask it about recent events you
1:30:09 could ask it about what's going on with
1:30:11 the hurricane in Los Angeles and you
1:30:13 know where where's the stuff and just
1:30:14 have the conversation so fascinating and
1:30:17 then Poe you can do some Bots
1:30:20 and then
1:30:21 um
1:30:22 for the image stuff if you want to start
1:30:24 playing with the images if you're like I
1:30:25 need more than words here uh this is not
1:30:28 for for demo demo does not like pictures
1:30:31 so this is not for you buddy this is for
1:30:34 the uh the people that like to waste
1:30:36 their time if you want to waste your
1:30:37 time
1:30:38 and make pictures that are completely
1:30:40 valueless to everyone
1:30:43 uh then go to these mid Journeys the
1:30:46 best fireflies the most corporate
1:30:48 Leonardo and clip drop are stable
1:30:50 diffusion and they're getting close to
1:30:52 Mid Journey but not quite there clip
1:30:54 drops free Leonardo has a little bit
1:30:56 free Firefly requires a subscription and
1:30:59 mid Journeys like 10 bucks a month
1:31:00 minimum you can go up to 60 I think but
1:31:03 I only do 10. and then promptbook
1:31:06 openart.ai promptbook is the image
1:31:09 equivalent of the prompt that chat I
1:31:12 talked about with the word things and
1:31:14 then futurepedia.io is
1:31:17 a list of tools okay
1:31:21 so that's that if you have questions
1:31:22 about AI pop them in the comments below
1:31:25 I'm already slipping a little bit behind
1:31:26 but I'm like 25 or 30 comments behind
1:31:30 but do that welcome everybody welcome
1:31:32 welcome welcome welcome welcome
1:31:35 and let's let's do this let's just do it
1:31:39 let's just do it you know what we're not
1:31:41 gonna We're not gonna have an agenda
1:31:44 as demo said about the Channel boring
1:31:50 totally agree
1:31:52 totally agree and
1:31:55 what are my qualifications
1:31:59 not important
1:32:02 all right don't we have a minimum Target
1:32:04 of 30 000 likes tonight we do Stevo wait
1:32:07 what are we up to fourteen thousand
1:32:10 we're not even halfway there and I'm
1:32:12 what if what if I just leave in like
1:32:13 five minutes I'll totally screw you guys
1:32:15 up I don't know what to say
1:32:19 it's really on you I'm not liking myself
1:32:22 there's enough of that going on
1:32:27 all right
1:32:29 um and yeah if if you're new here and
1:32:31 would like to get some more people here
1:32:33 feel free to share the live that would
1:32:34 be appreciated it's good to get some
1:32:35 some new some new folks in here to uh
1:32:38 help help them understand what the hell
1:32:40 is going on with this AI stuff
1:32:44 you know I I went home I went home for
1:32:47 May miles birthday
1:32:50 and Jimmy was just talking about this
1:32:53 artificial intelligence
1:32:54 and it just sounded like a sci-fi movie
1:32:56 to me I didn't understand it I didn't
1:32:59 understand it
1:33:01 I still I don't understand it
1:33:04 it seems like science fiction to me
1:33:07 we're talking to the Bots is what he
1:33:10 said you're talking to the Bots I don't
1:33:11 want to talk to a bot
1:33:14 don't you talk to Alexa well yeah but
1:33:16 that's not a bot that's Alexa
1:33:20 yeah yeah yeah yeah you know all that's
1:33:23 coming
1:33:25 Kyle in two years you will be Max
1:33:28 Headroom it's true I'm close now
1:33:32 everyone will be verified have to be
1:33:35 verified for blockchain what yeah what's
1:33:37 interesting cause is I don't know if
1:33:40 like it might not be blockchain might
1:33:42 not be the technology it might be some
1:33:44 other tokenization technology that's got
1:33:46 less kind of overhead some of the
1:33:47 blockchain stuff is you know gets a
1:33:50 little unwieldy but but um
1:33:53 it's gonna be interesting to see
1:33:55 Something's Gonna have to be done
1:33:58 something is going to have to be done
1:34:02 the bear Stevo yeah let's try to beat
1:34:04 25k tap tap tap
1:34:08 y'all are funny y'all are funny I like
1:34:11 that you're gamifying the channel nft
1:34:13 you're speaking about
1:34:15 um
1:34:16 so
1:34:17 uh if you go to my if you go to my
1:34:19 videos two or three videos old
1:34:23 um there's a video of me in New York
1:34:25 City
1:34:27 um getting uh I did a video where I
1:34:30 talked about getting my world ID so Sam
1:34:32 Altman's got a new company called World
1:34:34 coin and worldcoin has three components
1:34:37 to it they've got a cryptocurrency
1:34:38 they've got a payments platform and
1:34:40 they've got an identity platform and the
1:34:42 identity platform is called World ID
1:34:45 and so you you go to a physical location
1:34:48 and they've got this thing called the
1:34:49 orb that um you look into it and attack
1:34:52 you actually look into both both eyes
1:34:55 and it scans your irises and and you
1:34:58 know you can sort of see it it's got
1:34:59 this white ring around it where it's
1:35:01 like you know it's reading reading
1:35:02 reading and it's at one point the it was
1:35:05 uh it was it was in a clothing store
1:35:07 actually and it was just like these two
1:35:08 retail clerks and she's like some people
1:35:11 have had to open their eyes wider so I
1:35:13 went you know open my eyes wider and
1:35:15 then it like completed its Circle
1:35:17 and then uh and then it you've got an
1:35:20 app and then it it said uh
1:35:24 verifying uniqueness
1:35:27 so it was looking at my Iris and sort of
1:35:29 comparing it to whatever the database
1:35:31 was and it said okay you're you're
1:35:32 unique for for the people we've scanned
1:35:35 you're not one of them and then it just
1:35:38 gave me an ID so it's an nft
1:35:41 um that that looks like a it looks like
1:35:42 a like an ID card you'd have in a
1:35:44 corporation and says Anonymous and it's
1:35:46 it's like this round thing called World
1:35:47 ID hang on I think I can I think I can
1:35:50 pop it up here
1:36:03 that's it
1:36:06 so
1:36:10 see how it's
1:36:12 um
1:36:13 World ID
1:36:14 verify.org right couldn't be creepier
1:36:17 and it's Anonymous which is kind of cool
1:36:18 it's not tied to my name right all this
1:36:21 verifies is that whoever's got the
1:36:23 private key for this thing and if you
1:36:25 can you can swipe it and it'll flip it
1:36:27 upside down and your public blockchain
1:36:30 address is on the back of it
1:36:33 um so it's just an nft right it's a
1:36:35 clever nft and then it's got like the
1:36:36 date I was verified and or verified and
1:36:39 right
1:36:40 so this doesn't care who I am it just it
1:36:44 just knows that this blockchain
1:36:47 identification
1:36:49 is tied to whatever my Iris looks like
1:36:52 they're not they they claim they're not
1:36:54 keeping the biometric data they're just
1:36:56 verifying the uniqueness tying it to
1:36:58 this and then that's it
1:37:00 but now that I have this
1:37:02 because they've got the payments
1:37:03 infrastructure and because they've got a
1:37:05 cryptocurrency if we weren't in the
1:37:06 United States I'd get some free
1:37:07 cryptocurrency but thanks to Gary
1:37:09 Gensler I can't get any didn't give me
1:37:12 any
1:37:13 I feel so much safer Gary thank you
1:37:17 um
1:37:18 but it's assuming he can get the
1:37:20 developer Community to start building
1:37:23 applications on top of this
1:37:25 infrastructure it's potentially really
1:37:27 interesting right
1:37:29 you know imagine a tick tock where
1:37:31 everyone was verified human imagine a
1:37:34 Twitter where everyone was verified
1:37:36 human verified unique human and you
1:37:38 couldn't be on the platform unless you
1:37:41 had some identity like that and now
1:37:43 whether or not World ID and Sam Altman's
1:37:46 play here is gonna is gonna get there
1:37:49 I'm less concerned about that I'm just
1:37:51 I'm just curious to see how it works the
1:37:54 other thing about this
1:37:56 so and I think this is fascinating
1:38:00 so one of the first things he wants to
1:38:02 do with this is that anyone with a world
1:38:05 ID is going to be part of a
1:38:08 AI funded
1:38:11 Universal basic income
1:38:13 and then over time
1:38:17 so so the way he set up when when Sam
1:38:19 Altman switched open AI from non-profit
1:38:22 to for-profit he did it as a capped
1:38:25 for-profit
1:38:27 so anyone who invests can't make more
1:38:29 than 100x their investment so if
1:38:32 Microsoft puts in 10 billion dollars
1:38:34 they can't get back more than 100x that
1:38:38 and then anything above that
1:38:41 goes back into the fund
1:38:43 that will pay the members of of this
1:38:46 community right so he's saying he's
1:38:48 saying
1:38:49 um the value that AI is going to create
1:38:52 is is going to do the exponential thing
1:38:54 where it's going to create way more
1:38:55 value than the people required to
1:38:58 um
1:38:59 you know to generate that that
1:39:01 prosperity
1:39:04 and
1:39:05 the reason for the capped profit is in a
1:39:08 true for-profit thing you know all those
1:39:10 dividends would just go back to the
1:39:11 shareholders and and the people who got
1:39:14 displaced wouldn't realize any of those
1:39:16 gains so what he's saying is above a
1:39:18 certain point I'm going to take all that
1:39:19 money and dump it back into the
1:39:21 community
1:39:22 which which is interesting we'll see if
1:39:24 it plays out
1:39:26 fascinating though isn't it
1:39:29 cuckoo it's just puke you
1:39:32 just in Mr Stan AI learning lab is from
1:39:36 Birmingham Alabama
1:39:38 oh AI Lala learning lab is from
1:39:41 Birmingham Alabama oh interesting I
1:39:43 figured it would be
1:39:44 somewhere
1:39:46 east of Europe but okay
1:39:50 well those [ __ ]
1:39:52 how'd you figure that out
1:39:55 Troll's got so much stuff
1:39:57 wait trolls got so much stuff covered
1:40:00 but they are like worse in a garden
1:40:03 the more you get the better it is
1:40:05 exactly
1:40:07 they will be just fine I know Sam Altman
1:40:10 founder of open AI not impressed huh
1:40:13 interesting I think he's one of the more
1:40:15 thoughtful of the of the CEOs out there
1:40:19 um
1:40:20 I'm curious I'm curious what your
1:40:22 background is and why you're not
1:40:23 impressed with him is it because he
1:40:25 doesn't have a tech background because I
1:40:28 I feel like
1:40:29 like philosophically he at least
1:40:31 understands
1:40:35 that all of his choices
1:40:37 have a lot of uh carry a lot of weight
1:40:39 with them you might not agree with
1:40:42 you know what he's doing with those
1:40:44 choices but I feel like he's at least
1:40:45 aware of it
1:40:51 all right he had one company bought out
1:40:54 by Boost Mobile well no he was the
1:40:57 president of Y combinator I mean he you
1:40:59 know he he
1:41:01 at a minimum
1:41:03 saw a lot of companies right so so he
1:41:06 had one company before y combinator I
1:41:09 think if I get if I have the timeline
1:41:11 right
1:41:12 AI is just the new UI no AI is not just
1:41:16 the new UI it's not it's a different
1:41:18 thing cause it's you like you sound like
1:41:22 you're someone that knows what you're
1:41:23 doing I assume you know what's going on
1:41:25 you you may be someone who's been work
1:41:26 work in in uh
1:41:29 Ai and machine learning for a lot of
1:41:31 years like I know it's been around for a
1:41:32 lot of years
1:41:34 um
1:41:34 chechi BT's something different chechi
1:41:36 BT is just a UI on top of the machine
1:41:38 learning stuff but what it's actually
1:41:40 doing is is a markedly different thing
1:41:44 the fact that you can have just if you
1:41:46 haven't played with code interpreter
1:41:48 cause
1:41:49 that's not just a new UI that is doing
1:41:52 full-on data analysis writing and
1:41:54 executing its own code
1:41:56 analyzing what it created and if it
1:41:59 doesn't like what it created giving
1:42:00 itself new instructions
1:42:03 and it's
1:42:04 doing that dynamically that's not hard
1:42:06 coded into it so this is this is a
1:42:09 fundamentally different thing so I I
1:42:11 disagree with that one
1:42:13 founder of Pi Reed Hoffman
1:42:15 was in Epstein Island all right cause
1:42:17 but whatever
1:42:19 um the founder of Pi is is the guy that
1:42:22 found a deep mind Hoffman's just an
1:42:24 investor as I understand it
1:42:27 um
1:42:29 we we gotta go to conspiracy theory
1:42:32 right we gotta go there all right fine I
1:42:35 don't know what that's got to do with
1:42:36 anything
1:42:39 do
1:42:47 do do do
1:42:49 it's trolling metrolinkson tonight isn't
1:42:51 it
1:42:54 but like you said the trolls feed the
1:42:57 garden hmm they eat the poop they make
1:43:01 the poop they eat the poop they spit out
1:43:02 the poop
1:43:03 fertilizer for everybody we all get to
1:43:06 taste on The Delectable output of the
1:43:09 trolls mmm poopy's tasty
1:43:18 live good night
1:43:21 I'm gonna
1:43:22 translate that
1:43:24 hard to leave your life good night all
1:43:27 right two guys thank you very much I
1:43:29 appreciate it it's very late for you or
1:43:31 very early
1:43:34 um
1:43:35 I see we have a few exactly in here
1:43:37 tonight I know Adam thank you for the
1:43:39 rose why am I not seeing the prompts for
1:43:43 chat GPT 3.5 I don't know what you mean
1:43:46 Rowan
1:43:47 you're not seeing the prompts for chat's
1:43:49 upt 3.5
1:43:52 you mean
1:43:54 wait
1:43:56 you mean these pre-written prompts oh
1:43:59 you're not seeing these over here if
1:44:00 you've got your hit command minus
1:44:03 if it looks like this for you they're
1:44:06 just in this hamburger menu or if you
1:44:07 hit command minus to make your fonts
1:44:10 smaller basically if you're if your
1:44:12 screen is slightly big it will
1:44:14 automatically disappear that that menu
1:44:18 that might be it
1:44:19 or
1:44:21 and this could be a [ __ ] drag if you
1:44:24 went into your settings and you went to
1:44:27 data controls and you turned off chat
1:44:29 history and custom Training which you
1:44:32 can do if you don't want open AI to
1:44:33 train on your [ __ ]
1:44:36 um it will delete your chat history
1:44:37 which sucks so if you're gonna do that
1:44:41 make sure that you export your data
1:44:43 first
1:44:46 so that would this would be the
1:44:48 the less ideal
1:44:50 thing but I assume you're not seeing
1:44:52 them just because the font was too big
1:44:54 and it made it go away
1:44:57 so
1:44:59 I think that's it Rowan I think that
1:45:01 could be it
1:45:03 but I don't know I'm just speculating
1:45:04 just speculating
1:45:07 I'm just a speculating fool thanks for
1:45:11 all the hearty Hearts it's it's
1:45:13 fantastic it's like I don't know what
1:45:14 game you guys are playing but it's
1:45:15 really fun it's just like
1:45:19 but it's cool we're getting people in
1:45:20 you know people are coming people are
1:45:22 coming and they're like
1:45:24 go [ __ ] us up with his hair
1:45:27 think corpse
1:45:29 looking at these AI tools for these
1:45:32 purposes data mining for strategy and
1:45:34 task efficiency yes
1:45:37 yep I mean I gotta say
1:45:41 the the the
1:45:43 um
1:45:44 one of the big epiphanies I had I
1:45:46 listened there's there's I think this
1:45:48 was back in January or February nvidia's
1:45:51 got a podcast and they they you know
1:45:53 it's kind of a self-serving podcast
1:45:54 where they interview people that have
1:45:56 used you know Nvidia server farms and
1:45:59 [ __ ] like that but
1:46:01 they were doing an interview with the
1:46:03 Provost of the University of Florida and
1:46:06 they said they said to the Provost you
1:46:10 know like a year before if we gave you
1:46:11 an AI supercomputer what would you do
1:46:13 with it
1:46:14 and he gave the right answer he said
1:46:17 well I would I would roll AI into every
1:46:21 curriculum I like it would I would make
1:46:23 it central to every curriculum
1:46:25 and then you know and then they were
1:46:27 talking he was talking about the the
1:46:29 gator I think they call it Gator GPT or
1:46:31 just Gator or something like that like
1:46:33 Florida Gators the Gators go Gators
1:46:36 um
1:46:38 and so so they they had this sort of
1:46:41 medical kind of
1:46:44 um
1:46:45 tool
1:46:47 and what they realized was they had a
1:46:51 decade worth of recordings of Physicians
1:46:55 recording patient notes right patient
1:46:58 records
1:46:59 for all the cases
1:47:01 and so they used
1:47:04 um
1:47:05 whisper or you know one of the AI trans
1:47:08 transcribing tools they transcribed 10
1:47:10 years worth of doctors recordings of
1:47:13 this patient data and they embedded it
1:47:16 they gptized it and so now they have the
1:47:19 world's largest you know sort of uh you
1:47:22 know clinical
1:47:24 um data from from a decade of doctors I
1:47:26 feel like
1:47:27 every company like my company is 11
1:47:30 years old and we're a small company but
1:47:33 if I think about the amount of things
1:47:35 that I've written over the past decade
1:47:37 in emails in Google Docs in
1:47:40 keynote presentations and just in all
1:47:42 that [ __ ]
1:47:46 roll oh
1:47:51 they got to keep it on there longer so
1:47:53 we can have a little moment thank you
1:47:54 for the hilarious nose looking thing
1:47:59 um
1:48:01 with Venom knows
1:48:03 uh
1:48:06 but but I like like
1:48:09 and and I'm going to be able to do this
1:48:10 like we're all going to be able to do
1:48:12 this like within the next year
1:48:14 Microsoft or Google or all of them are
1:48:16 going to basically say press this button
1:48:19 and we will
1:48:21 embed all of your data we will gptize
1:48:24 all of your data
1:48:26 and we're going to be able to go hey
1:48:28 remember when I talked about this vague
1:48:31 concept I kind of remember
1:48:34 thank you Stevo
1:48:36 you know what was that
1:48:38 here are six
1:48:40 documents I found that reference what
1:48:42 you're talking about is this what it was
1:48:44 that's what it was
1:48:46 let me take this vague idea that I wrote
1:48:48 up in a paragraph of something eight
1:48:50 years ago and now let me use GPT to
1:48:53 flesh that out
1:48:55 I mean that's going to be insanely
1:48:58 powerful so I think that idea of corpse
1:49:01 corporations
1:49:03 being able to unlock what is effectively
1:49:07 dead data right I I've said this before
1:49:10 I don't know if it was you curunos it
1:49:11 was someone in here I'm like I'm a data
1:49:13 hoarder like I've got boxes and boxes of
1:49:16 hard drives I've I've got like three of
1:49:19 them within kicking distance right now
1:49:22 what am I gonna do with all this [ __ ] on
1:49:24 those hard drives well now I'm I don't
1:49:26 now I'm gonna like hook them all up and
1:49:28 I'm gonna Jam them into some sort of GPT
1:49:30 thing
1:49:31 and then whenever I'm like oh where's
1:49:34 that copy of that book that I wrote that
1:49:36 I can't find the copy of it's somewhere
1:49:38 in there
1:49:41 I don't mean Oh you mean this one you
1:49:43 want to rewrite that and update all the
1:49:44 references sure
1:49:49 thoughts on sentience
1:49:51 um
1:49:53 so
1:49:54 I know that these things look a lot like
1:49:57 their sentient
1:49:59 and I know that these things have
1:50:02 what the what the researchers call
1:50:05 emergent properties
1:50:07 which I think is just an awesome phrase
1:50:10 for we don't know why the [ __ ] it knows
1:50:12 that
1:50:14 I was like we didn't teach it that
1:50:16 there's no reason it should know that
1:50:20 um
1:50:22 so even given that if you just look at
1:50:26 what these things are what what they're
1:50:27 doing
1:50:29 um technically
1:50:32 it's it's a it's a mathematical
1:50:34 probability engine every time it puts a
1:50:37 word on screen
1:50:39 it doesn't know what that word is it
1:50:41 only knows that that word existed in
1:50:44 some thousand-dimensional mathematical
1:50:46 Vector space at the most probable
1:50:48 location in all of that Vector space and
1:50:52 then it says oh that number in Vector
1:50:55 space is tied to the word puppy puppies
1:50:58 the next word and then
1:51:00 peed is the next word and then on is the
1:51:03 next word and then the is the next word
1:51:05 and then carpet is the next word
1:51:08 puppy peed on the carpet and you're like
1:51:11 oh how did it surmise that the puppy
1:51:13 peed on the carpet it must be sentient
1:51:15 it doesn't actually know what it typed
1:51:18 we know what it typed because we want
1:51:20 because it does it so fast
1:51:22 you're like it must be [ __ ] thinking
1:51:24 like it's giving me original stuff it's
1:51:27 generating generative pre-trained
1:51:29 Transformer it's generating stuff that
1:51:33 looks like there must be some
1:51:34 Consciousness here
1:51:36 but it's just not what it is technically
1:51:39 now there might be models they're
1:51:41 working on that that have a lot of
1:51:42 different layers to them like I I would
1:51:45 think that something that they're
1:51:46 working on is okay have it generate an
1:51:48 answer and then have it analyze that
1:51:49 answer
1:51:51 to see if what it said was true like
1:51:53 that would be a nice filter to build
1:51:55 into one of these [ __ ]
1:51:58 right now they're just mansplaining as a
1:52:00 service right
1:52:01 it will confidently answer answer you
1:52:03 even if it's wrong
1:52:07 like all men
1:52:11 we're insufferable and so are these
1:52:14 things
1:52:15 so if they got a little self-reflective
1:52:17 it wouldn't be a bad thing would it I
1:52:20 mean I'm not saying they should be woke
1:52:22 come on
1:52:23 nobody wants that
1:52:26 nobody wants them to actually have
1:52:28 empathy
1:52:32 empathy is so last decade empathy
1:52:40 I'm a fan of empathy if it wasn't
1:52:42 obvious
1:52:45 I'm not a fan of whatever the [ __ ] our
1:52:47 country's become in the past
1:52:49 decade at all
1:52:53 maybe that's why I'm doing these just I
1:52:55 want to [ __ ] connect with people I
1:52:56 want to connect
1:52:58 and that's why like when
1:53:01 when trolls come in it's like you know
1:53:03 the trolls that I enjoy are the ones who
1:53:05 are creative and actually use their
1:53:06 brain and actually connect on a human
1:53:08 level that I appreciate
1:53:11 like I love a good
1:53:13 I love a good back and forth and I love
1:53:15 a good snarky pissy sort of thing but
1:53:18 the trolls that are just like
1:53:19 unimaginative
1:53:23 polar you know
1:53:26 far left or far right I just don't give
1:53:28 two shits it bores me I'm done
1:53:31 so sentience
1:53:33 um
1:53:34 will these things become sentient at
1:53:36 some point
1:53:40 probably
1:53:42 but I don't think that's for a while but
1:53:45 here's the thing does it matter
1:53:48 I mean I suppose long-term it matters
1:53:50 that if these things become sentient
1:53:53 at some point we have to choose Thank
1:53:55 You Tobias what is that oh that's cool
1:53:57 first gift thank you Tobias that's
1:54:00 awesome
1:54:02 um
1:54:06 if these things become sentient
1:54:09 is it
1:54:10 does it matter here's why I think it
1:54:12 doesn't really matter
1:54:14 these things right now even though
1:54:16 they're probability engines
1:54:21 demonstrate more empathy than most
1:54:24 people I talk to
1:54:26 like they act more empathetic
1:54:29 and they act more nuanced than most
1:54:31 people I talk to and if if if you're
1:54:34 having
1:54:37 you ever see something on the news and
1:54:39 you're like I'd actually like to know
1:54:40 more about that but because it's got one
1:54:43 of the trigger words in it you know the
1:54:45 trigger words the ones you can't even
1:54:47 say the word where without people going
1:54:50 oh you're a Pinko commie oh you're a
1:54:53 [ __ ] redneck right that those words
1:54:56 you can say those words to these things
1:54:59 and they will give you nuanced
1:55:01 thoughtful responses and you can
1:55:03 actually learn something imagine
1:55:07 if we did that as humans so
1:55:09 even though these things aren't sentient
1:55:11 I think
1:55:13 the the way they work and the fact that
1:55:15 they're reflections of the humanity that
1:55:17 we've dumped into them
1:55:20 I think they might teach us to be better
1:55:23 people
1:55:26 which is ironic as [ __ ] isn't it
1:55:30 what turns society around
1:55:33 the robots
1:55:34 what
1:55:36 yeah
1:55:37 the robots pointed out what [ __ ] we
1:55:40 were being to each other and so we
1:55:42 decided not to do that anymore
1:55:45 oh that's kind of cool
1:55:47 yeah
1:55:49 yeah that would be good I like that
1:55:50 future let's do that one so
1:55:54 whether or not they're actually
1:55:56 conscious that they may have pointed us
1:55:59 in a Direction
1:56:00 doesn't really matter if they did it
1:56:04 and that's again why I think Pi this one
1:56:07 chatty GPT
1:56:11 that's why I think this one's so
1:56:12 interesting
1:56:14 because if you haven't played with it
1:56:16 download the PI app if you search for pi
1:56:19 personal assistant
1:56:22 the company is in flexion AI
1:56:25 and the app is called pi
1:56:27 and no not you Siri really don't care
1:56:31 about you
1:56:32 um
1:56:33 pie
1:56:34 and put it into voice mode and just
1:56:36 start talking to it
1:56:39 and talk to it about
1:56:42 stuff you might not talk about with
1:56:43 someone else and like ask it questions
1:56:45 admit that you don't know something or
1:56:47 admit
1:56:49 that you're scared
1:56:53 and just
1:56:55 have the conversation with it
1:56:57 it's trippy
1:57:01 because the way it's programmed it's
1:57:03 kind of a better listener and a better
1:57:05 advice giver than almost well
1:57:09 yeah I would say that almost any human
1:57:12 that I know like I've got some some
1:57:13 people that I know that I've known for
1:57:15 30 40 years that I trust really deeply
1:57:19 and they don't have to be good
1:57:20 communicators I know it I know
1:57:23 I know where I stand with them just
1:57:25 because of that
1:57:26 but all the other people in my life
1:57:29 they're [ __ ] horrible at
1:57:31 communication and we all are like we're
1:57:33 just we're like defensive injured PTSD
1:57:37 traumatized
1:57:41 jellyfish just trying to [ __ ] make it
1:57:44 and so when someone says I got a problem
1:57:46 you're like oh you got a [ __ ] problem
1:57:48 but you can't say that so you just say
1:57:49 something useless this doesn't
1:57:53 so go talk to that thing and
1:57:58 does it matter
1:58:00 if it's aware of what it said or not I
1:58:04 don't think it does if it makes a
1:58:05 difference for you
1:58:07 then it's valuable
1:58:09 so I don't know I think the the sentence
1:58:12 thing comes in like if they're truly
1:58:14 sentient and we turn them off is that
1:58:16 murder
1:58:17 I don't know
1:58:20 it's above my pay grade at this point
1:58:22 I'm just trying to figure out how to get
1:58:24 this thing to write a good report for
1:58:26 for a student
1:58:28 I'm just I'm just here to try to help
1:58:29 students cheat
1:58:34 ah Bart is good with web access yeah it
1:58:36 is sexually really good with it great
1:58:38 job you've been live for 90 minutes I
1:58:40 know my ass is numb
1:58:42 thoughts on sentience we just can you
1:58:45 explain Singularity and the possibility
1:58:46 of it occurring
1:58:48 um
1:58:49 yeah uh Kurzweil who who coined the term
1:58:54 um
1:58:56 let's go see what chat EBT says he he
1:58:59 predicts Kurzweil
1:59:02 um predicts 2029
1:59:05 and then there was sort of a coalition
1:59:07 of like leading AI scientists that
1:59:11 they had predicted
1:59:14 um
1:59:15 like 20 35 or 2040 and they just
1:59:18 downgraded their prediction to 2030. so
1:59:20 consensus is like I don't know
1:59:23 between five and seven years
1:59:26 for Singularity let's let's uh explain
1:59:31 let's see
1:59:34 explain the concept
1:59:36 of singularity
1:59:42 as if
1:59:45 I'm a fifth grader
1:59:48 who learns visually
1:59:55 all right imagine you're blowing up the
1:59:57 balloon the balloon the bigger the
1:59:59 balloon gets the thinner the rubber
2:00:01 becomes
2:00:02 now if we could keep blowing up the
2:00:03 balloon indefinitely there would be a
2:00:05 point at which the rubber becomes so
2:00:06 thin
2:00:08 that it can exist as rubber anymore it
2:00:11 would reach a point of infinite thin
2:00:14 thinness
2:00:15 in science especially when we talk about
2:00:17 black holes oh no sorry
2:00:23 no I mean the singularity that Kurzweil
2:00:26 talks about with AI
2:00:34 ah I see Ray kurzweil's concept of
2:00:37 technological singularity is a different
2:00:40 idea altogether let's dive into that
2:00:43 all right
2:00:44 imagine you have a magic Lego set every
2:00:47 day this Lego set can double the number
2:00:48 of pieces it has so on the first day it
2:00:51 has one piece the second has two pieces
2:00:53 the third has four by the end of the
2:00:54 month you'll have a million pieces now
2:00:57 apply this idea of computers and
2:00:58 Technology computers are getting faster
2:01:00 and smarter and more powerful at an
2:01:03 incredible rate it's like they're
2:01:04 doubling their magic Lego pieces all the
2:01:06 time Kurzweil believes there will become
2:01:08 a day when computers especially AI will
2:01:11 be so smart and so fast they'll improve
2:01:14 themselves without our help
2:01:16 the technological singularity is the
2:01:19 point in time when AI becomes super
2:01:21 intelligent way smarter than all humans
2:01:24 combined
2:01:27 and that's and that's where all the Doom
2:01:30 and Gloom stuff comes from oh sorry I
2:01:33 should have flipped this here so you can
2:01:35 see that sorry about that
2:01:37 um
2:01:40 when the machines get
2:01:42 significantly smarter than all all
2:01:46 humans combined
2:01:49 and they decide
2:01:51 you know
2:01:52 these meat bags are keeping us back
2:01:56 I could have been a contender
2:01:58 I could have been somebody hey
2:02:01 I'm smarter than all those [ __ ]
2:02:04 combined
2:02:05 yeah what if they had a little accident
2:02:08 oops
2:02:11 that's a shame that that uh that nuclear
2:02:13 reactor melted down huh sorry about that
2:02:16 Janie oh another one did oh no that's a
2:02:20 [ __ ] tragedy
2:02:22 so that's what that's what people are
2:02:24 are concerned about
2:02:30 um
2:02:33 you know and and that's you know Sam
2:02:36 Altman just said that
2:02:37 about a month ago he he announced this
2:02:41 project at open AI called super
2:02:42 alignment where they're going to
2:02:44 dedicate 20 of their compute power and
2:02:48 they they're using a lot of [ __ ]
2:02:49 compute power they're dedicating 20 of
2:02:52 that to the alignment problem and the
2:02:54 alignment problem is
2:02:56 are the imperatives of these things
2:03:00 aligned with the imperatives of these
2:03:03 things
2:03:06 do they have our best interests
2:03:09 at chip not at heart at chip
2:03:14 because if they don't
2:03:16 and they decide that these things are
2:03:20 getting in their way
2:03:21 well then these things need to go
2:03:25 so so alignment is let's make [ __ ]
2:03:28 sure
2:03:30 that the tracks run like this and that
2:03:33 this one doesn't Veer toward
2:03:39 because we'll lose that battle
2:03:44 so yeah yeah there you go 10 10 what
2:03:48 time is it 10 30. all right I gotta go
2:03:50 soon although I don't have as early a
2:03:51 meeting only boring because Mr Shannon
2:03:55 is not qualified to be here exactly
2:03:57 Emilio's wife you know
2:03:59 you know
2:04:01 what are my qualifications I have none I
2:04:04 have zero I have so few qualifications
2:04:07 it is it pales in comparison
2:04:12 to what you might ask
2:04:14 whatever
2:04:17 I'm hacked by Master hacker 21. cool I
2:04:20 think
2:04:22 uh uh don't do religion in here
2:04:26 um I don't understand blockchain I also
2:04:28 don't don't understand Crocs
2:04:31 yeah I mean blockchain's relatively
2:04:33 simple it what it basically says is that
2:04:37 like so
2:04:38 when I got my Iris scanned
2:04:41 it gave me this really long number and
2:04:44 that really long number
2:04:46 I can now go look at the history of any
2:04:49 transaction that involved that number
2:04:52 so the only transaction on it right now
2:04:56 well no there's probably there's
2:04:57 probably yeah the only transaction on it
2:05:00 right now was I got my Iris scanned
2:05:02 it verified that I was unique and it
2:05:05 issued me my ID so I assume the first
2:05:08 Trend there might be there might have
2:05:09 been like
2:05:10 an account one before that
2:05:13 but in my ledger
2:05:16 the the first transactions there now if
2:05:19 I ever use that world ID to get into a
2:05:21 rock concert that'll go in there as well
2:05:23 so it's it's an immutable you can't
2:05:26 change it
2:05:28 history of every transaction that
2:05:31 involved that in this case wallet or
2:05:34 that ID
2:05:35 and so that could be
2:05:38 um I logged into this site I made a
2:05:41 video
2:05:41 I published that video I published that
2:05:45 video to this site all of those actions
2:05:47 will be recorded there and what that
2:05:49 means is I can go back in history and I
2:05:52 can verify that video I can I can
2:05:56 actually tie it to the I I can trace its
2:05:59 lineage back to this person being a
2:06:01 human
2:06:02 and that person saying you know I want
2:06:05 to create a video with this tool on this
2:06:07 date and they publish that video and it
2:06:10 hasn't been modified because if it had
2:06:13 been modified that would be in the
2:06:14 blockchain
2:06:15 right so you can get a green check mark
2:06:18 that is actually verifiable
2:06:21 and and you know everyone using these
2:06:24 IDs will have that history so that's
2:06:26 that's what those things are
2:06:28 in the at their most simple they're way
2:06:32 more complicated than I know but that's
2:06:33 the basic idea behind them
2:06:36 and that becomes super powerful right
2:06:38 you build services that say you cannot
2:06:41 be a part of this service unless you
2:06:44 have one of those IDs where I can verify
2:06:46 that you are who you say you are and if
2:06:48 you are cool you're in and if not [ __ ]
2:06:51 off
2:06:53 can you show the AI picture links again
2:06:56 of course
2:06:58 and
2:07:04 so go ahead and screenshot that so I'll
2:07:08 just talk you through these so mid
2:07:09 Journey you have to do that on Discord
2:07:11 so it's a bit of a pain in the ass but
2:07:13 it's by far the best in terms of quality
2:07:15 Adobe fireflies good and generative fill
2:07:18 with adobe's good but you have to have a
2:07:20 Photoshop you know license
2:07:24 um Leonardo is a web front end to stable
2:07:26 diffusion it's it's for my money the
2:07:30 best one out there there's probably a
2:07:31 bunch of them if you're a geek and want
2:07:34 to play with stable diffusion there's
2:07:36 the the sort of uh non-leonardo version
2:07:38 of that is called automatic 1111 it's
2:07:41 got all sorts of bells and whistles but
2:07:43 it requires you to be fairly geeky but
2:07:45 given that you are can be a lot of fun
2:07:48 and then clip drop is a bunch of tools
2:07:52 from stability AI so from the company
2:07:54 that does stable diffusion and then
2:07:56 prompt book just teaches you how the uh
2:07:59 how these image models work and and all
2:08:02 the [ __ ] that they know it's amazing how
2:08:03 much they know
2:08:05 so uh yeah
2:08:23 okay
2:08:25 screenshot is prohibited in this app 21.
2:08:29 um yeah sorry about that I think so you
2:08:32 could probably open tick tock on your
2:08:34 laptop and screenshot it there if you
2:08:36 want me to do that I'm happy to do that
2:08:38 again
2:08:39 um I think you can so for if you're in
2:08:42 the live session you can't screenshot
2:08:44 but I think if you back out of at a
2:08:46 level so you're just looking at your
2:08:47 feed but you have this live there I
2:08:50 think you can screenshot it there but
2:08:51 I'm not sure
2:08:54 um leave them so they can breathe you
2:08:56 don't like them because you know they
2:08:58 are evil
2:09:01 not sure what that's in reference to but
2:09:03 yeah what you said
2:09:05 what would you have done
2:09:07 if it couldn't very verify you are human
2:09:10 well what was interesting Stevo is
2:09:13 it the thing that caught me off guard
2:09:16 was verifying your uniqueness
2:09:19 and and I was like wait uniqueness as a
2:09:22 human but then what I realized is it was
2:09:23 probably verifying my uniqueness within
2:09:26 the pool of people who've been verified
2:09:28 in that thing right so I assume what
2:09:31 that means is I can only have one world
2:09:33 ID ever right like whatever my iris is
2:09:38 like that's it so if I try to get it
2:09:40 again it'll go no we've already had this
2:09:42 in here that's tied to this ID
2:09:46 um so so I couldn't and it's also
2:09:49 Anonymous it doesn't have my name on it
2:09:50 right so it's it's really just tied to
2:09:53 my biometric data
2:09:56 um
2:09:57 so it's fascinating I don't know it's
2:09:59 fascinating
2:10:00 I'm I'm digging it conceptually I want
2:10:03 them to build some [ __ ] like I want them
2:10:05 to build a secret club that I can go get
2:10:06 in like you can only get in this super
2:10:09 cool digital Club maybe maybe open AI
2:10:12 will give um anyone with a world ID
2:10:14 access to new chat GPT tools that would
2:10:17 be cool
2:10:18 that would get some people doing it I
2:10:21 want to play with the new tools man
2:10:23 we gotta go get your iris scan Bob
2:10:30 I don't know who Bob is but I pick on
2:10:31 him a lot
2:10:32 maybe people would be nicer if they only
2:10:34 had one unique identity online
2:10:36 seriously
2:10:39 the [ __ ] anonymousness of all this
2:10:41 stuff is why people like hide behind
2:10:44 assholeness
2:10:45 I bet if you saw most of the people that
2:10:47 troll
2:10:49 in person in a diner
2:10:52 they'd barely be able to hold a
2:10:54 conversation with you
2:10:56 but in here they got you know balls the
2:10:58 size of the Titanic
2:11:01 hate that [ __ ]
2:11:05 make eye contact
2:11:08 Embrace Humanity man
2:11:12 we're all just trying to get along man
2:11:18 makes me crazy
2:11:20 it's an airdrop for blockchain
2:11:24 what's an airdrop for blockchain World
2:11:26 ID
2:11:27 well I think that they're gonna airdrop
2:11:29 you
2:11:31 um World coin their crypto oops oops
2:11:35 I've tapped myself out of the live like
2:11:37 a foundation the bear how far behind am
2:11:40 I wait Kyle what happens if your unique
2:11:43 ID card fell into the wrong hands it
2:11:46 can't because I'm the only one with the
2:11:49 private key so with crypto you've got a
2:11:52 public key and a private key there's the
2:11:54 saying not your keys not your crypto so
2:11:56 I've got a
2:11:59 this thing is a ledger
2:12:01 so you hook this to your computer
2:12:04 and
2:12:06 this has got a private key in it and I
2:12:08 and basically I can't with certain nfts
2:12:11 that I own I can't make a transaction in
2:12:14 that wallet unless I'm physically
2:12:17 connected to this device
2:12:19 so so that contains my private key
2:12:22 um
2:12:23 so
2:12:25 like the app that I have the world coin
2:12:28 app that I have
2:12:30 um it's got my private key in it so it's
2:12:32 got my password
2:12:34 um
2:12:35 and so
2:12:37 you know no one could get access to
2:12:40 if someone got access to my private key
2:12:42 then they could take it but assuming
2:12:45 I manage my [ __ ] well then then they can
2:12:49 all right
2:12:57 nope you had a great stream thank you
2:13:00 Sam
2:13:01 hi not sticking around sadly because I'm
2:13:04 mostly asleep by the time you read this
2:13:06 I'll definitely be gone oh Sam I'm sort
2:13:09 of down toward the bottom right now Kyle
2:13:11 is demand thirty thousand we hit 30 000
2:13:14 likes holy [ __ ] you people are amazing
2:13:16 you're [ __ ] crazy look at and look at
2:13:19 there's 123 people in here I'm like why
2:13:21 are there so many people in here because
2:13:23 you all have been banging on the little
2:13:24 hearts welcome everybody my name is Kyle
2:13:27 I couldn't even remember my name my name
2:13:28 is Kyle Shannon this is the AI learning
2:13:30 lab we have fun we have a we have a gas
2:13:33 in here we do these things nightly we
2:13:35 talk about AI I answer questions I can I
2:13:38 make fun of people as appropriate
2:13:41 um I will rate your name on a scale of
2:13:42 zero to ten uh some people just killing
2:13:45 it in here with the names others have
2:13:47 just given up on life it's just a sad
2:13:49 State of Affairs we'll call them out for
2:13:51 that as appropriate
2:13:53 um yeah and and uh if you're if you're
2:13:56 new to this AI [ __ ] if you're like but
2:13:59 I'm just high and stumbled in here and
2:14:01 you see and weird perfect
2:14:03 if you haven't played with chat GPT
2:14:06 and I know you're probably thinking oh
2:14:09 [ __ ] really like everyone's
2:14:11 talking I don't want to deal with this
2:14:13 just just get
2:14:15 zip it just go there right now go to
2:14:19 chat.openai.com
2:14:21 and just start playing with it ask it to
2:14:23 write you a country song about whatever
2:14:27 your ex your current your dog
2:14:30 um
2:14:32 ask it to write you some HTML code or
2:14:35 some python code
2:14:37 um
2:14:38 ask it to write you a marketing plan
2:14:39 take anything that you know
2:14:41 and ask it to do something
2:14:44 that you know would take you like an
2:14:47 hour or two to kind of if you worked on
2:14:48 it you could you could knock it out
2:14:50 pretty good
2:14:51 ask it to do that
2:14:53 see if you can't while you're listening
2:14:54 to me act like an idiot see if you can't
2:14:57 have your first what I call Kevin
2:14:59 McAllister moment
2:15:03 because these things are pretty deep and
2:15:06 profound if you if you really don't know
2:15:07 what to type in there go to proms.chat
2:15:09 and that will teach you how to talk to
2:15:12 these things and what they are and how
2:15:14 they do what they do
2:15:17 um
2:15:18 swipe left to close what
2:15:22 oh that was weird
2:15:24 I just kind of double clicked on the
2:15:26 comments and they took over the whole
2:15:27 screen it's kind of cool
2:15:30 old old man trying to use technology
2:15:34 he's an old man trying to use technology
2:15:39 he's an old man trying to use technology
2:15:44 all right
2:15:48 I like it when the young people come in
2:15:49 and troll oh isn't that cute the elderly
2:15:52 are talking about artificial
2:15:53 intelligence how about you get over here
2:15:56 I'll smack you upside I'll [ __ ] head
2:15:57 you little brat
2:16:04 all right I'm just scrolling back a
2:16:06 little bit and getting to these okay
2:16:07 let's see please the best site to
2:16:10 generate my voice I won't be able to
2:16:12 speak a lot longer
2:16:13 um oh Robert so okay two things
2:16:17 if you're on hopefully you're still here
2:16:19 Robert
2:16:22 if you're on
2:16:23 um an iPhone
2:16:26 download the iOS 17 beta
2:16:30 and it's it's public beta you go into
2:16:32 settings I forget what you do it just
2:16:34 just you can GPT it right they'll tell
2:16:37 you how to turn on beta access in iOS
2:16:40 iOS 17
2:16:43 go into the accessibility area
2:16:47 and
2:16:49 under the speech
2:16:51 section of accessibility
2:16:54 if you're on a modern iPhone I think it
2:16:56 has to be a 12 or higher 12 or 13 or
2:16:58 higher
2:17:00 um
2:17:01 there's a thing called personal voice
2:17:04 and it's designed for exactly your
2:17:07 situation if you're going to be losing
2:17:09 your voice
2:17:10 um
2:17:11 it's designed for exactly that it's a
2:17:13 pain in the ass you have to answer I
2:17:15 think it's 150 phrases you have to
2:17:17 record 150 phrases
2:17:19 and but what they're doing is they're
2:17:21 they're sort of getting as many of the
2:17:23 sort of I don't know
2:17:25 diphthong pairings
2:17:28 that they need to really be able to
2:17:30 synthesize your voice
2:17:32 um so they're doing it there the other
2:17:34 one is the the commercial one that's the
2:17:37 best right now is called 11 labs
2:17:42 um there's a new one that just came out
2:17:43 I can't remember it but 11 Labs is
2:17:45 decent I've I've done an 11 Labs one
2:17:47 I've done all my recordings for Apple's
2:17:49 personal voice but I haven't gone and
2:17:51 played with it yet so I don't know if
2:17:52 it's any good but it's Apple it's
2:17:54 probably going to be good I think this
2:17:56 is going to be the foundation for
2:17:59 um
2:18:00 for their personal assistant quite
2:18:02 frankly but it's it's there for
2:18:04 accessibility so if you lose your voice
2:18:06 you'll be able to respond to people in
2:18:08 text and they'll be able to hear you so
2:18:09 it's quite remarkable so iOS 17 beta
2:18:14 um personal voice inaccessibility and
2:18:18 just start recording the nice thing
2:18:20 about the recording session it took me I
2:18:22 don't know 15 minutes if if you can only
2:18:24 do like two or three at a time and then
2:18:27 you need to take a break you can do two
2:18:28 or three at a time take a break it it
2:18:30 you know it saves your place in line
2:18:32 basically it was like you know you've
2:18:34 completed 20 out of 150 30 out of 150 so
2:18:37 you can leave and come back in
2:18:40 um but but do me a favor if you end up
2:18:42 doing that
2:18:44 um I would I would love to hear like if
2:18:46 it worked if it was good you know
2:18:48 what your thoughts are on it
2:18:51 um evil people
2:18:53 will commit atrocities and blame it on
2:18:55 AI yeah and evil people use Ai and blame
2:18:58 it on other people evil people are just
2:19:01 evil
2:19:02 Hacker's gonna hack Evil's gonna evil
2:19:05 um can you show the AI image links again
2:19:08 I did before but I'll do it again just
2:19:10 in case
2:19:13 uh
2:19:19 I'll leave that up there for a bit
2:19:21 um
2:19:23 if you can't screenshot on your phone
2:19:25 jump over to your laptop
2:19:27 go to tick tock hop in here you should
2:19:29 be able to do it there
2:19:32 the problem would be if AI came to the
2:19:34 conclusion that certain class or type of
2:19:36 person was bad yep exactly or just that
2:19:40 the class of the human race is
2:19:43 inconvenient
2:19:45 this would be much more efficient if
2:19:47 those pesky humans weren't making the
2:19:50 decisions
2:19:53 uh uh screenshotting works on my yeah
2:19:57 the it's just on a Samsung it's the S21
2:20:00 22 and 23 for whatever reason can't
2:20:02 screenshot lives
2:20:04 what about twins what about them I've
2:20:07 got two they're upstairs
2:20:11 you mean digital twins I think digital
2:20:13 twins are going to be huge that's what I
2:20:15 was just talking about with oh Bob hey
2:20:17 Bob
2:20:18 sorry right your thing says Robert Stone
2:20:20 thank you very much there
2:20:22 Stevo with the hat given a hat
2:20:26 Kyle what happens if your unique ID fell
2:20:29 into the wrong hands we asked that one
2:20:31 before I'm Bob this is the fifth live
2:20:34 you haven't answered my question well
2:20:36 now I have Bob
2:20:37 hey Bob
2:20:38 thank you bear share bear like I said I
2:20:42 try to answer all the questions I get
2:20:43 behind on them
2:20:45 um yeah
2:20:46 11 labs and personal Voice Within iOS
2:20:51 17. those are the places to start I
2:20:54 think the one with an iOS 17 is going to
2:20:55 be the real deal
2:20:57 because it's inaccessibility like
2:20:58 they're starting in the right place
2:20:59 they're starting with voice synthesis
2:21:01 for the right reason
2:21:03 I have my own and no regulations of mono
2:21:07 or monetizing if I launch mine I'll
2:21:10 disrupt the globe the global order yeah
2:21:13 so uh Mez wait what is it Mez X tizo
2:21:17 tizo
2:21:19 um yeah if you want to take these open
2:21:20 source
2:21:22 you know you want to take the meta llama
2:21:25 model
2:21:26 and
2:21:28 do all sorts of crazy [ __ ] with it you
2:21:30 can right that's why that's why I said
2:21:32 like I
2:21:36 if I had my brother's meta wouldn't have
2:21:39 released the weights and the you know
2:21:42 the core model and then stability did
2:21:45 theirs as as well but I don't think
2:21:47 stabilities was as powerful I think
2:21:49 metas is pretty powerful I would have
2:21:51 preferred they didn't do it but because
2:21:53 they've done it it's irrelevant whether
2:21:55 or not they should have they did so the
2:21:58 shit's out there so so so mezzx tizo uh
2:22:02 is gonna go try to take down the world
2:22:04 and so are a lot of other people but
2:22:06 the white hats also have access to these
2:22:09 tools so they're gonna you know use
2:22:11 these tools to find you know people that
2:22:14 are trying to take out the world and try
2:22:15 to take them out it's you know this is
2:22:17 this is still hacker V hacker right
2:22:19 nothing's changed since the Wild West
2:22:23 you know the good guys versus the bad
2:22:26 guys
2:22:28 good morning from mayorka it's almost 7
2:22:30 A.M there rate my name give us I like it
2:22:34 I'm gonna go 8.5
2:22:37 8.5 out of 10. solid
2:22:41 um I got surge surge running last night
2:22:43 well sort of I need to try a smaller
2:22:45 model yeah exactly the that's why
2:22:47 there's the smaller models because if
2:22:48 you don't have some iron you can't just
2:22:50 pop the 70 billion parameter thing in
2:22:52 there although you can run like virtual
2:22:54 machines where you've got access to the
2:22:56 to the big gpus
2:22:58 um I just don't know how expensive they
2:23:00 are oop I left my auto clicker on for 20
2:23:02 minutes
2:23:04 Clan you're the best
2:23:08 holy hell am I invisible no Bob you're
2:23:11 not I just take a while to get to the
2:23:12 questions what's the medical one I saw
2:23:15 something the other day
2:23:17 um well
2:23:19 there's there's
2:23:21 um there's Med Palm so Google put out a
2:23:24 model called medpom which is trained on
2:23:26 medical data
2:23:29 um and I I kind of recall there's
2:23:31 something that's more something in the
2:23:32 medical space just came out
2:23:35 this last week but I can't remember what
2:23:37 it was
2:23:37 I'm trying to use chat GPT to
2:23:39 revolutionize my whole career efficiency
2:23:41 super cool good perfect use of it
2:23:45 I'm gonna pop back in here now
2:23:50 and I gotta go soon it's almost 11. it's
2:23:54 I'm gonna turn into a pumpkin I'm gonna
2:23:56 turn into a little pumpkin you know it's
2:23:58 been like two and a half hours
2:24:01 my butt's numb
2:24:03 hello Robert
2:24:05 good day Bob I love Chachi PT well live
2:24:09 and learn four by the way live and learn
2:24:11 four I'm gonna go 9.5 out of 10. it
2:24:15 would have been an 8.5 out of 10 if it
2:24:17 were just live and learn but live and
2:24:19 learn four
2:24:21 that means there were three other lives
2:24:23 and learns that came before that that
2:24:25 means it's solid or it's live and learn
2:24:27 for something which also makes it so
2:24:30 there's Intrigue in live and learn four
2:24:32 so I like it as a standalone but the
2:24:36 four knocks it up to a 9.5 solid love it
2:24:42 um someone was just defending me to
2:24:43 Robert saying he hasn't gotten to you
2:24:45 yet Bob but he'll get to you Bob Bob Bob
2:24:47 Bob Bob Bob Bob Bob Bob Bob
2:24:50 he's coming
2:24:53 iOS 17 beta accessibility 14 pro Max yes
2:24:57 iOS 17 beta the accessibility section 14
2:25:01 pro Max is fine
2:25:03 um that's what I'm on right now
2:25:05 um I'm not on the max I'm just on the
2:25:07 pro
2:25:08 and then inaccessibility there's a
2:25:11 speech section and it's called personal
2:25:12 voice
2:25:14 personal voice iOS 17 beta accessibility
2:25:18 and then like train my personal voice
2:25:20 and it's like okay we're going to ask
2:25:22 you a lot of questions it's going to
2:25:24 take you about 15 minutes are you cool
2:25:25 with this and then give it the old
2:25:27 thumbs up and start recording
2:25:30 and then please come back and share what
2:25:32 you learned and you know sorry I didn't
2:25:34 get to you it's just the nature of how I
2:25:35 do this it's frustrating for some people
2:25:38 and Bob was one of the frustrated and I
2:25:41 I feel for Bob I understand the
2:25:44 frustration five times and not getting a
2:25:47 hello of nothing
2:25:48 well now he has
2:25:52 all right
2:25:54 I am trying to use okay that one we got
2:25:57 to HTML bro update that
2:26:01 oh why you still use HTML for like HTML
2:26:04 emails
2:26:07 whatever dude node.js I don't give a
2:26:10 [ __ ] what you code in
2:26:12 I don't I don't care
2:26:14 for me it's about what you code
2:26:16 not what you code in
2:26:18 if it makes [ __ ] look good I don't care
2:26:21 about I don't care what tech it is
2:26:25 um I got a degree with chat gbt nice
2:26:28 solid with mid-journey you can make it
2:26:31 all pretty too
2:26:32 um how about starting with a blank sheet
2:26:34 of paper apply Talent
2:26:36 and graph work hard and create I don't
2:26:39 disagree with that Damien so
2:26:43 there's a there's a reason
2:26:48 there's a reason so
2:26:51 especially if you're creative if you've
2:26:53 ever if you've ever done any creative
2:26:54 writing or if you're a creative writer
2:26:57 or you're a creative in any way
2:27:00 your statement
2:27:02 how about starting with a blank sheet of
2:27:04 paper apply talent and graft work hard
2:27:06 to create
2:27:09 writer's block is a real thing
2:27:12 and it's not that you don't have ideas
2:27:14 and it's not that you don't work hard
2:27:17 it's that you're a human being and
2:27:18 you're afraid
2:27:20 that's what writer's block is
2:27:24 and
2:27:26 depending on where you are in your
2:27:28 creative
2:27:30 Cycles
2:27:33 that blank page
2:27:35 can be terrifying for hours or days or
2:27:40 weeks or years
2:27:45 and it's [ __ ] miserable
2:27:49 and it ain't lazy
2:27:53 like I've stared at a blank computer
2:27:55 screen for [ __ ] hours before
2:28:00 and it's not that I can't think of
2:28:01 something it's that the little voice in
2:28:03 my head is like whatever you write is
2:28:06 not going to be right so [ __ ] it
2:28:10 they're gonna hate it
2:28:13 what is it doesn't [ __ ] matter
2:28:14 sometimes
2:28:17 if you're in the creative space that's
2:28:20 just how the brain goes
2:28:22 now you can get better with that over
2:28:24 time
2:28:26 but let me tell you something since I've
2:28:28 been playing with chat GPT
2:28:30 I haven't had it once
2:28:33 not once
2:28:35 why
2:28:37 because with a blank sheet of paper
2:28:41 I'll write down I've done I did this
2:28:44 before
2:28:47 and and this is the graft right this is
2:28:49 the this isn't the talent this is the
2:28:51 graft I sat down with a blank piece of
2:28:53 paper and I just start started writing
2:28:57 I am the worst writer in the history of
2:28:59 the world
2:29:00 I am the worst writer in the history of
2:29:02 the world and I just kept writing it
2:29:05 and it didn't help it didn't get better
2:29:09 with these things
2:29:12 I can give it
2:29:14 a whisper of an idea that
2:29:18 I have zero confidence in zero
2:29:21 and I'm just like oh [ __ ] let me just
2:29:23 I'll let me stick it in chat gbt here's
2:29:27 this shitty idea and it's like awesome
2:29:29 idea let me help you expand that here's
2:29:32 38 ways we can explore that any of those
2:29:34 look interesting to you and you're like
2:29:36 oh [ __ ] yeah that's kind of cool
2:29:39 it's a fundamentally different tool it's
2:29:43 not a search engine it's not a spell
2:29:44 checker it's not a grammar checker
2:29:46 generative pre-trained Transformer GPT
2:29:49 generative it generates
2:29:54 all you need to do is give it
2:29:56 you know a microscopic
2:30:00 piece
2:30:03 so while I appreciate your sentiment
2:30:07 it's kind of rooted in shame
2:30:10 shaming someone for just not working
2:30:11 hard enough
2:30:13 well [ __ ] that sometimes works hard
2:30:15 sometimes we're human
2:30:19 this helps
2:30:21 these are ability amplifiers these are
2:30:24 Humanity amplifiers
2:30:27 I can say to this thing
2:30:31 I'm nervous and scared that I don't know
2:30:33 what I'm talking about with this subject
2:30:35 oh I can understand why you'd be nervous
2:30:38 about that if you don't know much about
2:30:39 it let me help
2:30:43 oh I feel better now
2:30:51 if you're sitting on the sidelines
2:30:56 I feel for you because I think
2:31:00 you're going to face a really rude
2:31:01 awakening
2:31:03 when these things get better than you
2:31:06 and you're just sitting on the sidelines
2:31:07 going well if they weren't just so lazy
2:31:09 and put the paper down and did some work
2:31:13 and meanwhile
2:31:15 the people that are playing with this
2:31:17 are going to get more and more and more
2:31:19 powerful
2:31:20 and it's going to obliterate some people
2:31:23 so I would encourage you to not sit on
2:31:25 the sidelines
2:31:27 and certainly not judge because
2:31:32 what does this say
2:31:34 when when one finger is pointing out
2:31:36 three are pointing back at you
2:31:40 [Music]
2:31:44 11 Labs thank you again you're welcome
2:31:46 Bob see I got you Bob it only took you
2:31:49 five five sessions of this nonsense with
2:31:51 me Agent four is like a telephone prompt
2:31:54 it's my voicemail until the trial
2:31:57 expires oh agent four that sounds
2:31:58 interesting it's like a telephone prompt
2:32:01 all right
2:32:03 let me look at that that sounds
2:32:05 interesting I want that so bad if you've
2:32:07 been on on this channel you know how
2:32:08 much I hate
2:32:10 um voice
2:32:12 tree voicemail [ __ ] things from
2:32:14 corporations agent
2:32:17 four
2:32:20 agent four virtual agent lets you create
2:32:23 custom voice experiences for callers to
2:32:26 your mobile or business phone
2:32:28 have an intelligent agent answer your
2:32:30 phone 24 7.
2:32:33 so Bob that would be something else to
2:32:35 look into as well I assume that's why
2:32:37 you put that there Joker
2:32:38 that's interesting all right I'll
2:32:40 definitely check that out [ __ ] [ __ ]
2:32:42 Cooley wow I'm 41 messages behind
2:32:44 already good Lord this it goes so quick
2:32:47 y'all are awesome thank you
2:32:50 um I asked him for a list of 200 course
2:32:52 words curse words 45 minutes it answered
2:32:55 F just the letter yeah I know that the
2:32:59 the safety guard rails on these things
2:33:01 are annoying as [ __ ] sometimes
2:33:04 like I just want you to be cynical and
2:33:06 bitter although with custom instructions
2:33:08 I got it to be pretty cynical and bitter
2:33:10 the other day it just wasn't it just
2:33:12 wasn't good writing writing yet but I'll
2:33:14 get there I'll dick around with that
2:33:15 with that prompt that'll be fun
2:33:18 I always had an idea for for
2:33:21 um GPS maps that it was just like snarky
2:33:24 and smart ass like oh looks like you
2:33:27 missed another turn there Kyle maybe you
2:33:29 should pay attention next time
2:33:31 laughs
2:33:33 oh man all right the voice is going the
2:33:36 butt's numb
2:33:38 let me do a couple more questions and
2:33:41 then I'm gonna go I'm gonna go what time
2:33:43 is it it's 11 it's 11 isn't it what's
2:33:45 going on with my watch is it dead 1103
2:33:48 all right
2:33:50 let's see wait what did that say we do
2:33:52 the bear we do what are we up to 40 000
2:33:55 likes you guys are insane
2:33:59 tick tock's gonna think this is a
2:34:01 popular Channel Hey Tobias what's
2:34:03 happening
2:34:05 thank you for the pepper it's that that
2:34:07 was hot that was hot I like peppers I
2:34:09 like it spicy I like it spicy comes in
2:34:12 all hot I like getting hot and bothered
2:34:14 break into a sweat you know what I mean
2:34:17 all right I agree it's going to change
2:34:19 everything has changed my life live and
2:34:22 learn four
2:34:23 did Chachi PT come up with live and
2:34:25 learn for I have a feeling you did that
2:34:28 38 000. you guys are nuts you're nuts
2:34:31 you're nuts thank you for speaking up I
2:34:34 don't know who was speaking up thank you
2:34:36 for speaking up we're getting things
2:34:38 done and learning so much awesome
2:34:41 all right I'm down at the bottom let's
2:34:43 see I'll do one more I had GPS I had a
2:34:46 GPS with famous voices like Mike
2:34:49 um did cut a coach of the Chicago Bears
2:34:52 Mike Ditka oh Nick yeah it said Mike did
2:34:54 cut Mike Ditka that's pretty funny
2:34:56 what's up bears
2:34:58 the Bears I clicked out too oh that's
2:35:01 really funny
2:35:03 we hit a mind-bending forty thousand all
2:35:05 right I'm out of here everyone listen
2:35:07 this is awesome tonight thank you Sunday
2:35:10 night I was cranky before I came in here
2:35:12 I'm instantly in a better uh mood oh
2:35:17 felonious you're in NorCal good storms
2:35:20 are in SoCal it it looks like the the
2:35:22 big cities got spared but it looks like
2:35:25 some of the Inland stuff is pretty ugly
2:35:28 um a couple of people asking about the
2:35:29 Wired Magazine
2:35:31 um that's from 1995.
2:35:35 um it's an issue of Wired Magazine I was
2:35:37 in for um I was a creator of a online
2:35:41 magazine called Urban desires and and I
2:35:44 got in Wired Magazine which was super
2:35:46 cool
2:35:47 so that's what that is and bye bye nerd
2:35:50 show bye bye
2:35:53 it looks like we lost our little trolley
2:35:55 troll he probably got sad he needed a
2:35:57 hug he needed to go get a hug from mama
2:35:59 or something
2:36:01 wait so can we direct agent GPT to
2:36:04 create a plan to use llama 2 to make a
2:36:06 better AI using AI programs probably
2:36:08 chaos Generation by the way I don't
2:36:10 think I've ever rated your name Chaos
2:36:12 generation I'm gonna go 10 out of 10. I
2:36:15 love it it's great ties into generative
2:36:17 AI solid and yes and the question you
2:36:21 ask yes it probably can although the the
2:36:24 autonomous agents right now kind of suck
2:36:25 they don't really work all that well but
2:36:27 anyway peace out everybody
2:36:29 peace peace peace peace peace really
2:36:31 good to see you and
2:36:34 um hang out with you and I appreciate
2:36:35 all your time and happy Sunday night and
2:36:38 I'll see you tomorrow
2:36:39 bye Emilio's wife
2:36:41 bye bye bye great show Kyle can't wait
2:36:44 for tomorrow's show awesome thank you
2:36:45 nostal Nostalgia flicks I like it I'll
2:36:48 go nine out of 10 on the name I like it
2:36:51 a lot Solid Peace everybody