Revamping Siri: Apple's Bold Leap into Generative A.I.

[00:00:06] All right. Happy to have you along for another edition of Tech Watch Radio. We keep an eye

[00:00:24] on tech so you don't have to. I am Sam Bushman. I've got Jay Harrison with me. Welcome, sir.

[00:00:29] Hey Sam, how's it going? I'm doing fantastic. Although there are storms all over the Southeast

[00:00:36] to the Midwest, just about everywhere. Huh? Oh yeah. We had a big one yesterday that came

[00:00:41] through and you know it wasn't a hurricane but it was like a hurricane but it's very

[00:00:45] short duration so it didn't last like all day. It was just maybe a half hour to an

[00:00:50] hour. One of the strongest ones I've seen on radar. You know when you see the radar

[00:00:53] image that the Doppler radar locally or whatever you'll see red is bad weather

[00:00:59] maroon is really bad and I saw hot pink and like blue grays. I don't recall ever seeing

[00:01:07] intensities like that in this cell that came through and it really tore up a lot of stuff.

[00:01:11] There's local tornadoes, lots of places have been out of power. We're still out

[00:01:14] of power, but out of power for more than 24 hours. But we're trucking along. We're

[00:01:18] having fun with tech. We're having fun with tech. We are broadcasting live on

[00:01:23] generator. I just thought I'd let everybody know too. That's something kind of fascinating.

[00:01:27] I've done a lot of experiments with broadcasting live on generators and broadcasting live on

[00:01:32] solar generators which is really nothing but a battery backup. They call them solar

[00:01:36] generators because you can hook a solar panel to them and that kind of stuff. But I've done

[00:01:39] a lot of testing with goal zeros and broadcasting live without power and all

[00:01:44] this kind of stuff. You know you got to have power somewhere. You don't have the

[00:01:47] internet right? But you can locally record if you need to and then you can upload it when

[00:01:50] it's you know powers available or whatever. But as long as we have internet even if we

[00:01:54] don't have power we can broadcast and that's what we're doing now.

[00:01:57] You got to have internet. That's kind of cool.

[00:01:58] Everything's got to have the whole world runs on internet. Business especially. I mean you

[00:02:02] don't have internet, you don't have phones, you can't process credit cards,

[00:02:04] you can't go anywhere do anything half the time your servers might be in

[00:02:07] the cloud. I mean you just without internet you're out of business. You

[00:02:11] might as well you know be offline. Your building maybe gone. It's just like

[00:02:15] it's internet's the most important thing. Well and that's changed over the

[00:02:20] last I don't know 10, 20, 25 whatever years you want to you know

[00:02:23] attribute it to for most people. I mean it's more than that if you're someone

[00:02:27] like me because I you know I played with the internet like in the early

[00:02:30] 90s. Yeah it was a novelty then. It was cool. You could do some email and you

[00:02:34] could do you know look things up and have different but it was like an

[00:02:37] addition. It was something good to have. It wasn't essential. Now it is.

[00:02:43] Now it's just critically essential. I remember back in the it was even DOS that I was getting on the

[00:02:46] internet. I used to remote control my radio station with two DOS

[00:02:49] computers and I developed a remote control solution to work with this

[00:02:52] and you know implemented this and that. I used to use the Lynx browser which was

[00:02:57] just a text browser and I remember learning all the commands to connect to

[00:03:01] email by hand and we created a utility called the net because we just felt like

[00:03:04] hey man we got the net absolute control of the internet with this thing and it

[00:03:08] would let us you know have a command line and and get prompts and

[00:03:12] interact with servers by hand and you know I remember those old days kind

[00:03:15] of fondly actually. Sam was a pioneer in the wild west of the internet.

[00:03:20] Yeah. See if you guys can hear the generator in the background just a little bit.

[00:03:24] See if you can hear it. Absolutely. I'm trying to help with that a little bit

[00:03:28] and keep it out of the audio. I'm intentionally kind of draw attention to

[00:03:32] it because we're tech guys. It's kind of a cool tech thing to just bust

[00:03:36] out and hook up your generator and get it done. You know I kind of think

[00:03:38] that's pretty cool. Yeah by the way if you guys want to learn more about

[00:03:41] what we're doing and PI Tech guys dot com is the website for the podcast

[00:03:45] and the show it's on radio AM and FM radio stations first and then it is a

[00:03:49] podcast later. We do two half hours every weekend so it's an hour radio

[00:03:54] show but then it's separated by half hour that way you can have two

[00:03:58] podcasts one comes out shortly after it airs live on the radio stations

[00:04:02] and then the second one comes out midweek like Wednesday or so. So you

[00:04:05] get two podcasts a week if you want to listen to it podcast style or if

[00:04:09] you want to you know call your radio station say hey you got to take

[00:04:12] TechWatch radio a sandwich. Yeah I want to I want to hear that on my radio

[00:04:14] station. So anyway that's just something for you. Network providers

[00:04:17] Inc. Dot com is the website for the company that I work for. Jay works

[00:04:21] for other people too which we won't discuss but we're both tech

[00:04:24] professionals. We've been in the tech industry for a heck of a long

[00:04:27] time between the two of us. What Jay we have over 60 years. Oh at

[00:04:31] least more than that really because I think you know I'm probably

[00:04:34] at thirty five and you're at forty five or something. I mean we've

[00:04:39] been doing it since we were kids literally. It's just like forever

[00:04:43] again a whole life. So anyway I'm just saying 60 years of quote

[00:04:46] professionalism in the at least field at least between the two of us.

[00:04:49] Yeah New York Times bring up this story Jay and I hate to do this

[00:04:53] because your phone is going to go off but hey Siri you remember

[00:04:57] the age old words right. My phone can't hear you Sam. So it's

[00:05:01] only my headphones nine times out of 10 when you say hey Siri it

[00:05:05] ends with a disappointed frustrated event. Never everybody listening

[00:05:09] to this broadcast phone may be going off but yeah but you know

[00:05:14] you say those words and you just expect it to kind of go south

[00:05:17] but you're hoping it'll do a good job at everything else. It's

[00:05:20] a plea even in the darkness. This time they're saying this Jay

[00:05:23] you ready. Hey Siri get ready for a eye. Yeah. Is your phone

[00:05:30] triggering with this. I upgrade no mine's not because it's not

[00:05:33] the room. Oh OK. I left it out of the room anyway they're saying

[00:05:37] this series going to get an A.I. upgrade. Listen to this.

[00:05:40] Apple plans to announce that it will bring generative A.I.

[00:05:45] to the iPhones. They say it'll be the most significant

[00:05:48] update to the iPhone in a decade. Apple's top software

[00:05:53] executives decided early last year that the company's virtual

[00:05:59] assistant needed a listen to this a brain transplant Jay.

[00:06:02] Jay. Yeah it definitely does.

[00:06:05] Siri I just don't even know if it has a brain at least

[00:06:08] compared to what we're seeing in A.I. today. I mean Siri's

[00:06:11] like the worst one on the block. It seemed like in fact

[00:06:13] anytime you ask it to anything that's off script other than

[00:06:16] like the temperature or some kind of pre prepared thing.

[00:06:20] Here's what I found on the web. And if you're driving

[00:06:22] that's not helpful at all. I mean you might as well.

[00:06:26] You've got to look down and swipe the site and click on

[00:06:28] what you want or whatever. It's just not helpful at

[00:06:30] all. The other thing it does oftentimes is it gets very

[00:06:33] confused in loops that are very frustrating. So I'll say

[00:06:35] hey Siri find the nearest pizza place and it says OK

[00:06:38] here's you know number one number two. I think great

[00:06:41] number three I like that call that one because it

[00:06:45] says I can get directions or call. OK call that one.

[00:06:47] Then it says call who. Well let down on.

[00:06:54] So it's very frustrating but the decision came after

[00:06:56] two executives Greg Federer Federer Rinki and John

[00:07:01] Jenembrea. I guess they spend weeks testing open A.I.

[00:07:05] is. I mean they call it the new chat bot chat GPT

[00:07:09] it's not really new to us now but anyway they say

[00:07:13] the products use generative artificial intelligence

[00:07:16] which can write poetry create computer code and

[00:07:20] answer complex questions even conversational style is

[00:07:22] what they're working on now. It made Siri look

[00:07:25] absolutely antiquated said a couple of people that

[00:07:27] were well they weren't able to speak publicly about

[00:07:30] this but they kind of gave us the down low

[00:07:32] introduced in 2011 as the original virtual assistant

[00:07:37] in every iPhone. You know if you had a single

[00:07:41] request it could handle a lot of them and very

[00:07:42] basic things but think about that. Did you know

[00:07:44] that Siri came out in 2011 when I had forgot

[00:07:47] about the time that that seems early to me

[00:07:49] because then the first I thought the first iPhone

[00:07:51] came out in 2012. So no I don't think that's

[00:07:55] right. I think the first iPhones were around in

[00:07:57] like 2008 and all that kind of stuff. I'll have

[00:07:59] to look that up but they say it's never been

[00:08:01] able to follow a conversation at all. It often

[00:08:04] misunderstood questions. Chat GPT on the other

[00:08:08] hand was kind of smart if you asked for the

[00:08:10] weather in San Francisco and then you said

[00:08:11] hey what about New York or Boston. It

[00:08:14] understood that you're still talking about the

[00:08:15] weather that you just want to know about a

[00:08:17] different city so it respond. You know even

[00:08:19] in the chat you be forecast that way right.

[00:08:22] Yeah even in their app now they have an

[00:08:24] interactive voice thing that you can if you're

[00:08:26] driving you can just talk back and forth to it

[00:08:28] and it'll kind of switch prompts to let you

[00:08:29] talk and then it'll go to talking and type out

[00:08:33] everything. It's pretty interesting. Yeah and

[00:08:35] it's not as agile as you would think it would

[00:08:38] be but you know what they're working on that

[00:08:40] and it will get agile quite quickly. I'm

[00:08:42] convinced. And this is what Apple wants to

[00:08:44] tap into. Yeah this is what they want to be

[00:08:48] to be able to do all the Siri things

[00:08:50] interact with your phone launch applications

[00:08:52] do whatever but also to be able to kind of

[00:08:54] have a brain like this article saying yeah and

[00:08:58] when I find the interesting thing about it

[00:08:59] for example I can say Siri set an appointment

[00:09:01] to call Jay at three o'clock and it will say

[00:09:04] OK I'm sending appointment to call Jay at

[00:09:05] three. Then I say oh man I don't need to

[00:09:08] talk to Jay he just called me we're good. So

[00:09:09] then I basically say Siri cancel that

[00:09:11] appointment. I can't do that. Are you

[00:09:15] kidding me now. So hopefully it will get a

[00:09:18] lot smarter and a lot more capable. And I

[00:09:20] get the hesitancy to want to delete things

[00:09:23] because if you you know right change your

[00:09:26] mind and be like no no no no then it can't

[00:09:28] undo a lot of that kind of stuff. Hopefully

[00:09:30] they'll have like undo commands where it's

[00:09:32] like we'll keep five or three or I don't

[00:09:34] know what you know renditions of something

[00:09:36] and you can undo back to two or three

[00:09:38] and then it kind of falls off. Hopefully

[00:09:39] it'll get to that level. But they're saying

[00:09:42] hey this technology is going to help Siri

[00:09:45] leapfrog. They say that there's going to

[00:09:48] be the most significant reorganization of

[00:09:50] the iPhone and Siri in over a decade. They

[00:09:54] say they're determined to catch up in the

[00:09:56] quote a race Jay. What do you think of

[00:10:01] that. I think that they will. I mean it

[00:10:03] won't take much to make it smarter. I

[00:10:06] don't know though there'll be exactly on

[00:10:08] par with GPT because they've got such a

[00:10:09] head start. But I've seen a lot of a

[00:10:13] lot of people that are doing a lot of

[00:10:15] different things with it. I mean there's

[00:10:17] a new search engine called perplexity for

[00:10:18] example that's trying to replace Google

[00:10:20] and it's pretty good. I mean you know I

[00:10:23] don't know if I'm ready to replace

[00:10:24] Google with it but I thought it was

[00:10:25] pretty interesting and I you know I

[00:10:27] don't I don't even know what they're

[00:10:29] based on. They could even be based on

[00:10:30] GPT. But I know there's a lot of people

[00:10:32] working on these language models. Well

[00:10:35] and there's so many models out there

[00:10:37] now too. People are starting to say hey

[00:10:39] let's create a closed model if for some

[00:10:41] reason it gets beyond our closed

[00:10:43] model we can add a general model in

[00:10:46] other words go to the go to the

[00:10:47] bigger models. But we'll let you know

[00:10:49] when we're switching to the bigger

[00:10:50] model so you realize that the

[00:10:52] information may not be as reliable or

[00:10:54] as customized to what you need. They say

[00:10:55] they're going to have to increase the

[00:10:56] memory to the iPhones big time to make

[00:10:58] this happen Jay. I believe that takes

[00:11:00] a lot of hardware to run these

[00:11:01] things. There's a lot of training

[00:11:03] data and that's not the problem

[00:11:05] really but to crunch all that and

[00:11:08] you know especially the larger the

[00:11:09] input is token wise in to make that

[00:11:12] happen it just takes a lot of

[00:11:14] processing power. Now this is

[00:11:16] interesting to say the least. Not only

[00:11:18] is it a memory hog comparatively and

[00:11:20] takes some serious computing power

[00:11:22] they wanted to be as much of it

[00:11:23] locally as they can because they

[00:11:25] don't want to go back and forth. One

[00:11:26] of the big problems with Siri is that

[00:11:28] you know if it's super cloudy outside

[00:11:30] or rainy or this and that then it

[00:11:32] can't communicate with its thing or

[00:11:33] whatever and so they want to do as

[00:11:35] much on the phone as they can. Now

[00:11:36] listen to this part because I found

[00:11:37] this fascinating and it's like you're

[00:11:39] thinking to Sam why are you spending

[00:11:40] so much time on this because the

[00:11:41] iPhone is one of the most popular

[00:11:42] phones in the world and this

[00:11:44] upgrade is absolutely game changing if

[00:11:48] it's good and intelligent and if they

[00:11:50] really build it right. You know even

[00:11:52] the first release of this will be

[00:11:53] child's play compared to two or three

[00:11:55] or four or whatever five years down

[00:11:56] the road but they even say this

[00:11:59] Apple executives worry that the new

[00:12:03] AI technology threatens the

[00:12:06] company's dominance of the smart

[00:12:08] phone industry because it has the

[00:12:11] potential to become the primary

[00:12:13] operating system replacing the whole

[00:12:15] operating system Jay. That would be

[00:12:19] impressive. They say it can create a

[00:12:24] new ecosystems of apps and these

[00:12:26] apps are basically going to be called

[00:12:28] agents Jay think about that an

[00:12:34] ecosystem of apps called agents

[00:12:36] their AI apps called agents. That's

[00:12:38] kind of interesting to me because

[00:12:41] that instead of an app an agent

[00:12:43] kind of suggests that it's got its

[00:12:44] own innate abilities or intelligence

[00:12:47] to itself. Yeah so you might have a

[00:12:50] a travel agent or a yeah and it can

[00:12:53] order Uber make cal or point calendar

[00:12:57] appointments etc. And they say it'll

[00:12:58] flat out undermine Apple's app store.

[00:13:02] So we're talking about you know if

[00:13:03] this kicks off they're worried about

[00:13:05] affecting the operating system they

[00:13:06] may have to just you know replace

[00:13:07] the whole operating system. They say

[00:13:09] that it could replace the whole app

[00:13:11] store because it can go and get all

[00:13:12] kinds of apps Jay not just the ones

[00:13:14] approved by Apple and they may not be

[00:13:16] able to control the engine enough to

[00:13:19] stop it over that sounds too ambitious

[00:13:22] engine and well yeah I don't know I

[00:13:24] don't think they're going to do that.

[00:13:25] Apple doesn't want to relinquish that

[00:13:26] kind of control. They don't want to

[00:13:27] relinquish the revenue either that

[00:13:29] the app store brings them 30 percent

[00:13:31] of everything 24 billion in annual

[00:13:34] sales by the way this app store

[00:13:37] yeah they get a big chunk of

[00:13:38] revenue from it so they're very

[00:13:39] worried and this is where I say

[00:13:42] that they're going to create

[00:13:42] stumbling blocks to the advancement

[00:13:44] of the technology because they're

[00:13:45] going to say well hold on until we

[00:13:46] get an answer to this a feasible

[00:13:48] whatever to this we don't want to

[00:13:49] you know proceed right right.

[00:13:56] They say this they don't know who

[00:13:58] will want to use this new

[00:13:59] technology. Well I can answer that

[00:14:02] for him you ready Jay. Yeah.

[00:14:05] Depends on how good it is.

[00:14:06] Yeah if it's great everybody wants

[00:14:07] to use it. Exactly right. If it's

[00:14:11] killer good and it responds and it

[00:14:12] really answers questions and does

[00:14:14] things. And they say if you're not

[00:14:17] careful the iPhone could become a

[00:14:19] dumb brick compared with other

[00:14:21] technologies. So they're going to

[00:14:23] be forced to do something about

[00:14:24] this. That is also what they're

[00:14:26] worried about it. They're

[00:14:27] worried about Gemini coming out

[00:14:28] on Android and just being the

[00:14:31] killer app for a cell phone you

[00:14:32] know I think it's going to

[00:14:34] happen Jay. I do too. I do too

[00:14:36] and I think they're trying to get

[00:14:36] ahead of it. And I think they're

[00:14:38] smart to try to get ahead of it.

[00:14:40] Now they say they don't know

[00:14:40] how many people regularly use Siri

[00:14:42] right now. Well I'll tell you

[00:14:44] this most people use Siri only when

[00:14:46] they're forced to otherwise it

[00:14:47] sucks and you use something else

[00:14:48] the best you can. Listen Siri in

[00:14:50] a pinch right. If Apple doesn't

[00:14:51] know who all uses Siri who does

[00:14:53] I mean who on the planet would

[00:14:54] know besides Apple. That doesn't

[00:14:57] make sense to me. Well they say

[00:15:01] it's unclear how many people

[00:15:02] regularly use Siri. I just don't

[00:15:08] understand how to even respond

[00:15:10] to that. Do you regularly use

[00:15:11] Siri Sam. Yes but in a very

[00:15:15] limited way so I'll ask at the

[00:15:16] time I'll ask at the weather

[00:15:18] I'll ask it very basic things

[00:15:20] but I've learned if I want to

[00:15:20] ask you know any real

[00:15:22] question believe it or not I go

[00:15:23] to Alexa. Yeah the number one

[00:15:26] thing I use for perfect. It's

[00:15:27] not a great brain like AI but

[00:15:29] it's way better than Siri at

[00:15:30] least it can answer questions.

[00:15:32] Well at least it goes out to

[00:15:33] Wikipedia and we'll read you

[00:15:34] the article at least right.

[00:15:36] Instead of being like here's

[00:15:36] what I found on the web. Yeah

[00:15:38] I use Siri all the time but

[00:15:39] it's mostly for you know set

[00:15:40] a timer for this or set an

[00:15:41] alarm for that. Maybe remind

[00:15:44] me when I get to this

[00:15:44] location but I don't yeah I

[00:15:46] don't ask it really in depth

[00:15:47] questions. What's the meaning

[00:15:48] of life Siri. I don't do

[00:15:49] that you know. Yeah I ask it

[00:15:52] weird questions to see how it

[00:15:53] will answer questions like

[00:15:54] sometimes related to politics

[00:15:56] to see if it's going to be

[00:15:56] all you know one sided or the

[00:15:58] other and stuff like that.

[00:16:00] And you know so far it isn't

[00:16:02] yeah. But so far the AI

[00:16:04] systems have been pretty fair

[00:16:05] but I expect that to change

[00:16:07] in a negative way over time

[00:16:08] as well Jay. Yeah it's

[00:16:09] surprising I think how fair

[00:16:10] they are. I know that some

[00:16:13] people concerned about them

[00:16:14] being too woke or some of

[00:16:15] them you know just concerned

[00:16:16] in general about them having a

[00:16:18] bias but I think they do it's

[00:16:20] good a better job than I

[00:16:22] would have thought being

[00:16:23] trying to be neutral on a lot

[00:16:24] of topics. So I'm going to

[00:16:26] predict this. They say this

[00:16:30] they say that it'll apply it

[00:16:31] will deal with a lot of the

[00:16:33] requests directly on the

[00:16:34] iPhone not needing to go out

[00:16:35] to the cloud. They say this

[00:16:37] strategy will also save money

[00:16:38] now I'm going to be very

[00:16:39] clear about this. I think it

[00:16:40] will save money for the

[00:16:41] company. Yes but I think it's

[00:16:43] going to mean if they're

[00:16:43] going to put a bunch more

[00:16:44] memory in it a bunch more

[00:16:44] capabilities and store a lot

[00:16:46] more on the iPhone they're

[00:16:47] going to have to start

[00:16:48] cranking up the cost of the

[00:16:49] iPhone. So they're really

[00:16:50] going to move the money cost

[00:16:51] to the purchaser of the

[00:16:52] hardware anyway. I say I

[00:16:56] think the iPhones already 12

[00:16:57] cents for about a thousand

[00:16:58] words Jay when they use chat

[00:17:01] GPT or whatever else. Twelve

[00:17:04] cents for about a thousand

[00:17:05] words. They say that will

[00:17:08] help save a lot of cloud

[00:17:09] computing costs. Yeah I think

[00:17:12] that the iPhone itself is

[00:17:13] underutilized. I mean I

[00:17:14] think half the time your

[00:17:15] processor is doing nothing

[00:17:16] and you could certainly run a

[00:17:19] language model on this. I

[00:17:20] see people running them on

[00:17:21] Raspberry Pi's for example

[00:17:23] and how much more powerful

[00:17:25] is a modern day iPhone 15 or

[00:17:27] something compared to a

[00:17:28] Raspberry Pi. Yeah you're

[00:17:32] right about that. Now they

[00:17:33] also say this is kind of

[00:17:34] interesting the New York

[00:17:35] Times that's where this

[00:17:36] article comes from by the

[00:17:37] way but the New York Times

[00:17:38] sued open AI and its

[00:17:40] partner Microsoft in December

[00:17:43] for copyright infringement of

[00:17:44] its stories and so part of

[00:17:46] the problem is these you

[00:17:46] know bots or these what

[00:17:48] are they called not apps

[00:17:49] they're going to call them

[00:17:50] agents. These agents go out

[00:17:51] and collect and gather and

[00:17:52] collate and whatever

[00:17:53] information put it in your

[00:17:55] hands but then hey now

[00:17:56] they're saying there's

[00:17:56] copyright infringement

[00:17:57] issues. So I bring this up

[00:18:00] because we've got a lot

[00:18:01] going on this is going to

[00:18:02] change the world. It's

[00:18:04] going to change your

[00:18:04] iPhone it might change the

[00:18:05] operating system it might

[00:18:07] change the stores and how

[00:18:08] they work. It's going to

[00:18:09] replace apps with agents

[00:18:10] which are much more

[00:18:11] intelligent than apps. It's

[00:18:12] going to increase the

[00:18:13] cost of the phone because

[00:18:14] you're going to have to

[00:18:14] have a ton more hardware. I

[00:18:15] mean there's going to

[00:18:16] bring a lot of changes J.

[00:18:18] That's true. Definitely.

[00:18:22] Now they say that if they

[00:18:24] push everything to the

[00:18:25] phone they're going to

[00:18:25] have to use a smaller

[00:18:26] model for its core a

[00:18:29] smaller AI system housed

[00:18:30] on the iPhone rather than

[00:18:32] some big model system in

[00:18:33] the cloud and they say

[00:18:36] you know hey if you're

[00:18:37] not careful at the

[00:18:38] smaller the form factor of

[00:18:39] the I shouldn't say form

[00:18:42] factor the smaller the

[00:18:43] model is the more likely

[00:18:46] it is to hallucinate or

[00:18:47] get wrong answers. I

[00:18:48] think that's true to a

[00:18:48] point J. But I think

[00:18:49] it's also true that the

[00:18:50] smaller it is the more

[00:18:51] you can make sure that

[00:18:51] it gets the right

[00:18:52] answers. So there's a

[00:18:53] pro and a con there

[00:18:54] based on how the

[00:18:54] filters and how it can

[00:18:56] intelligently respond

[00:18:57] right. Yeah but I also

[00:18:58] wonder how much time do

[00:19:00] you want to spend

[00:19:00] training it and telling

[00:19:01] it whether the answer

[00:19:02] is right or wrong. Most

[00:19:04] people once they get you

[00:19:05] know researchers and guys

[00:19:07] are into it tech guys

[00:19:08] they don't mind doing

[00:19:09] some of that stuff. But

[00:19:10] most users don't ever

[00:19:12] want a wrong answer they

[00:19:13] don't want to tell it

[00:19:13] whether it got the

[00:19:14] answer right or wrong

[00:19:15] that we want to bother

[00:19:15] with that they just want

[00:19:16] something that already

[00:19:17] works out of the box. I

[00:19:19] agree 100 percent but I

[00:19:20] will say it depends on

[00:19:21] how easy it is for me

[00:19:22] to tell when it's wrong.

[00:19:24] If they have an easy

[00:19:24] way for me not to get

[00:19:25] into the weeds and

[00:19:26] answer 20 questions and

[00:19:27] get all crazy side

[00:19:28] tracked. But if I could

[00:19:29] just quickly say hey

[00:19:30] this thing's wrong to

[00:19:33] where they can get you

[00:19:33] know they could if they

[00:19:34] cloud source that

[00:19:36] ability they could get

[00:19:37] it smarter and making

[00:19:38] right choices and right

[00:19:39] answers very quickly.

[00:19:41] The problem is humans on

[00:19:42] the back end of this

[00:19:43] J to say if I say that

[00:19:44] it was wrong and I hit a

[00:19:45] button or do something

[00:19:46] claiming that it's

[00:19:46] wrong. So we're going to

[00:19:47] have to look at that

[00:19:48] collate all those answers

[00:19:50] and come back with

[00:19:50] correcting you know can't

[00:19:52] trust the human right.

[00:19:53] You can't trust the

[00:19:53] humans. That's what's

[00:19:54] happened with things

[00:19:55] like being able to

[00:19:57] thumbs down videos on

[00:19:58] YouTube. You know they

[00:19:59] kind of took that away.

[00:20:00] Yes you can still

[00:20:00] thumbs down but people

[00:20:01] can't see it and it

[00:20:02] doesn't affect the

[00:20:03] algorithm as much

[00:20:04] anymore because you had

[00:20:05] these groups of people

[00:20:06] and they would get

[00:20:07] together on Discord and

[00:20:08] Reddit and stuff and go

[00:20:09] out and politically go

[00:20:11] thumbs down and try to

[00:20:12] cancel a video. They

[00:20:13] were all getting up or

[00:20:14] on a creator and just

[00:20:15] hit all their videos and

[00:20:16] stuff. So you won't even

[00:20:17] go to see what I'm

[00:20:18] saying is you'll have

[00:20:19] people who want to

[00:20:21] influence or bias the

[00:20:22] AI and they will tell

[00:20:24] it it's wrong when it's

[00:20:24] not and vice versa.

[00:20:28] Yeah they say it's also

[00:20:29] a blind spot for Apple

[00:20:30] because Apple has kind

[00:20:31] of separated their

[00:20:32] software on their

[00:20:32] hardware divisions and

[00:20:33] they don't communicate

[00:20:34] well enough. They say

[00:20:34] that's going to have

[00:20:35] to change to that blind

[00:20:36] spots going to have

[00:20:36] to go away. They're

[00:20:37] going to have to have

[00:20:37] hardware and software

[00:20:38] people work hand in

[00:20:39] glove to make this

[00:20:40] stuff seamless and you

[00:20:41] know respond timely

[00:20:43] and appropriately. That's

[00:20:45] another interesting

[00:20:45] twist. The bigger the

[00:20:46] company the more that

[00:20:47] happens. Unfortunately

[00:20:51] true story anyway I

[00:20:53] know we're spending a

[00:20:53] whole episode on this

[00:20:54] but I find it

[00:20:55] fascinating and I find

[00:20:56] it really game

[00:20:58] changing how quick a

[00:20:59] lot of this stuff will

[00:21:00] be here is a question

[00:21:03] and how disruptive it

[00:21:05] will be is really the

[00:21:06] real question and I

[00:21:07] think the answer to

[00:21:08] that those questions

[00:21:09] not only have to do

[00:21:10] with you know how quick

[00:21:11] Apple will roll out the

[00:21:12] increased hardware to

[00:21:13] make it happen and roll

[00:21:14] out the tech behind

[00:21:15] it. But I think we're

[00:21:16] going to have to

[00:21:16] answer a lot of social

[00:21:17] questions too in terms

[00:21:19] of you know people

[00:21:20] suing about copyright

[00:21:22] infringement or it

[00:21:23] knowing too much about

[00:21:25] you and gaining too

[00:21:25] much information. How

[00:21:26] much information does

[00:21:27] it collect when it

[00:21:27] has a conversation with

[00:21:28] you Jay or me. You

[00:21:30] know how detailed is

[00:21:31] this going to be how

[00:21:34] breaching of your

[00:21:35] privacy will it get

[00:21:37] you know they already

[00:21:38] talked about Amazon

[00:21:43] you know on their

[00:21:43] devices listening all

[00:21:45] the time and gathering

[00:21:46] too much information

[00:21:47] is going to have to be

[00:21:47] a lot worked out from

[00:21:49] the social side of this

[00:21:50] thing too. It's like

[00:21:51] driverless cars. It's

[00:21:52] one thing to get a car

[00:21:53] that's you know safe

[00:21:54] and driverless that

[00:21:55] towards you know AI

[00:21:57] driven vehicles but

[00:21:58] it's another to get

[00:21:59] through all the social

[00:22:00] issues about hey I'm

[00:22:01] just not comfortable

[00:22:02] when I pull up at a

[00:22:02] stop sign and you

[00:22:04] know three cars run a

[00:22:06] red light and there's

[00:22:06] nobody in them or

[00:22:08] what. What. There's

[00:22:09] going to be a lot of

[00:22:10] discussions about that

[00:22:11] right. Oh yeah. And

[00:22:12] I think we're still

[00:22:13] we're still right in

[00:22:14] the very beginning of

[00:22:15] this wave of AI not

[00:22:16] only did it just come

[00:22:17] out the November

[00:22:19] before last but we're

[00:22:20] still seeing brand new

[00:22:22] stuff. I shared a video

[00:22:22] with Sam this week

[00:22:24] about AI and music

[00:22:26] and the way it can do

[00:22:27] lyrics and the poetry

[00:22:30] and the baselines

[00:22:32] and all of this stuff

[00:22:32] that put everything

[00:22:33] together for music and

[00:22:34] it is disruptive. I

[00:22:35] mean some of the

[00:22:36] videos that I've seen

[00:22:37] from music producers

[00:22:38] and guys who are just

[00:22:39] candidly saying look

[00:22:40] man this is something

[00:22:42] I would have worked

[00:22:43] days on and this

[00:22:44] thing just produced it

[00:22:45] in seconds. And yeah

[00:22:47] it's still got some

[00:22:48] rough spot. Very little

[00:22:48] guidance too I might

[00:22:49] add. Yeah but it's

[00:22:51] just incredible and

[00:22:51] and you know people

[00:22:53] are saying I can't

[00:22:54] even I don't even

[00:22:55] know where it's going

[00:22:55] to be in days or

[00:22:56] weeks or months from

[00:22:57] now because it's

[00:22:58] it's moving so fast

[00:23:00] and it's becoming so

[00:23:01] good at things things

[00:23:02] that we never even

[00:23:02] thought that it would

[00:23:03] really be able to

[00:23:04] do like you know we

[00:23:05] love the idea of

[00:23:06] the Star Trek

[00:23:07] computer you can ask

[00:23:07] it any question and

[00:23:08] it just knows the

[00:23:09] answer and tells you

[00:23:10] whatever. But the

[00:23:11] idea that it can be

[00:23:12] creative and make

[00:23:14] things not just poetry

[00:23:15] but music and

[00:23:16] instruments and

[00:23:17] chorus and riffs

[00:23:19] and everything else to

[00:23:19] go with that.

[00:23:20] It's just rhythms

[00:23:21] and all kinds. Yeah

[00:23:22] it's it's it's

[00:23:23] amazing and we're

[00:23:23] seeing this kind of

[00:23:24] stuff every single

[00:23:25] day even now.

[00:23:27] Yeah we're going to

[00:23:27] try to in an

[00:23:29] episode here soon

[00:23:30] we're going to place

[00:23:30] some of that back

[00:23:31] for you and stuff

[00:23:31] like that because

[00:23:32] some of it is

[00:23:32] really shocking in

[00:23:33] terms of of I

[00:23:36] don't know if you

[00:23:37] want to say poetic

[00:23:37] license is that the

[00:23:38] word to use. Yeah

[00:23:39] it would be a fun

[00:23:40] episode anyway the

[00:23:41] poetic license of what

[00:23:43] it kind of can do

[00:23:44] and what it kind of

[00:23:44] can and it can it

[00:23:45] can it used to

[00:23:48] be when he used to

[00:23:48] do an auto

[00:23:49] generated kind of

[00:23:50] music thing it

[00:23:50] would sound like it

[00:23:51] was kind of

[00:23:52] tinker toyish.

[00:23:53] Yeah right.

[00:23:55] Nowadays though I'm

[00:23:56] telling you right

[00:23:56] now it is better

[00:23:57] than a human one

[00:23:58] in many ways.

[00:24:00] A lot of people are

[00:24:01] are I don't know

[00:24:02] frightened is the

[00:24:03] right word but

[00:24:04] really concerned.

[00:24:05] Yeah really

[00:24:06] concerned both

[00:24:07] artists especially

[00:24:08] recording in music

[00:24:10] groups and things

[00:24:11] like that they're

[00:24:11] just petrified at

[00:24:13] some of this stuff.

[00:24:13] I mean it's it's

[00:24:14] right around the

[00:24:15] corner and it's

[00:24:15] already happening

[00:24:16] and I think it's

[00:24:17] I think it's cool.

[00:24:18] I mean I like

[00:24:19] the tech.

[00:24:19] I understand the

[00:24:20] concerns but we

[00:24:22] live in interesting

[00:24:23] time Sam.

[00:24:24] This is the

[00:24:25] broadcast where we

[00:24:26] keep an eye on

[00:24:26] tech so you don't

[00:24:27] have to.

[00:24:28] Thanks for listening

[00:24:28] Liberty.

[00:24:30] I'm sorry.

[00:24:31] Network providers

[00:24:32] Inc.com for our

[00:24:32] company's website

[00:24:33] and PITech guys

[00:24:34] dot com for the

[00:24:35] podcast page.

[00:24:36] Thanks to all of

[00:24:37] you for listening

[00:24:37] spread the word

[00:24:38] show the love will

[00:24:39] you please

[00:24:40] and make it a

[00:24:41] great tech day

[00:24:42] William.

[00:24:42] Thanks.