00:00.80 archpodnet Welcome back to the here Mark Podcast episode 2 6 2 the all robot edition and because none of us are real. This is all just a simulation and we were were in the matrix we're in the truman show call it what you will, but this is all fake. 00:08.24 Andrew I. 00:16.22 archpodnet But Doug you were about to have a comment at the end of the last segment go ahead. 00:19.94 Doug Yeah, it was ah so I don't know like I'm gonna be like heavily down on this whole like machine learning stuff like the um, the whole chat bot and whatnot I I Honestly think it's I've seen so many things and so much hype about it. Um, but I honestly think one that anything good out of it is going to be offset easily by all the bad. Um, and it's it's going to be ah, a more of a negative. Ah ah a net negative than anything else but also like so like back a cup. 00:52.90 archpodnet You could have said that about the internet. 00:56.93 Doug But I didn't say that about the internet though. 00:57.60 Bill White Yeah I he's right. 01:05.20 Doug Ah, yeah, no, that's not we're still, we're still the the jury's still out on that but it might be a net negative there? Um, but ah where is going with this is like um like so. 01:05.52 archpodnet Ah, just saying. 01:15.79 archpodnet Nah. 01:20.97 Doug I know people are saying things like oh you can write all this code and stuff but you actually can't like um you if you try to do it. The 1 thing. Um, it can't do and this is because of how it's actually designed like amazingly they actually put out a paper on how they designed everything um and how how the. How it works and it's amazing like 1 it's sorry this could be a couple of points I'll try to make them quick. 1 is like we shouldn't call it ai because it's not Ai. It's not artificial intelligence. Um, it's machine learning. Um, and by calling it artificial intelligence. We're making it sound a lot smarter and a lot more magical than it really is um and that's it's why it fails so horribly when it actually writes like essays because it makes up um all the references and that's by design like it. It can never not do that like it. There's no way it's never not going to be able to make up random um references which is why it it usually fails and it's the same thing with code is you can't you can maybe get it to write like a small function and then you could string together a series of small functions. Um, and that. 02:22.24 archpodnet Um. 02:33.22 Doug That can help but like you cannot it cannot write actual code. It cannot write a program. It cannot write an app. Um, unless you actually already know what you want it to do and you have to give it tons upon tons of detailed instructions and the thing about code is. 02:41.87 archpodnet Ah. 02:50.31 Doug Code is not just about writing code. It's about testing code. It's about fixing all the problems like I got 98% of coding is about literally fixing errors and it can't do that. It can't figure out what mistake has been made. And it can't adapt code. Um, so it's it's pretty much actually pretty useless I think ah so there's a bunch of different ones out there and Microsoft githubs their whole um extra code helper wherever they call it at the end of the day I think they found it basically. 7 % improvement like that was it in terms of um, you know all these great things for coders. It sped up their work by 7 % and that paper didn't even take into account of like how many errors and like. That's going to cause and how many people are going to make mistakes that are going to cost companies I don't know how many billions trillions and stuff like that. So like I am very much down on it and like I think there's a lot hype saying things like 90% of they could do like 90% of coding no like. 03:55.70 archpodnet Ah. 04:05.46 Doug Field testing it slightly helps people who already know how to code but like an average person is not going to be able to write an app by um, using any of these programs. 04:15.59 archpodnet All right doug time for comments on that because I'm not sure well I agree with you right now, right? But what I think what we're really talking about here and the whole point of machine learning is that it learns faster than anything else can right? because it takes in so many different sources. It learns when when we tell it nope that's wrong, try this one and then you get it right? Then it learns that right it learns preferences it learns that that human element that that it needs to have to be to be real to be to do good uiux design right? and it needs to do that stuff. But there's you know when you're talking about coding I mean most developers right? They don't even they they might code they code those little special things and they and they put the big blocks together. But I know plenty of developers that just have a code library. They say I need to do this thing so I'm going to take this chunk and I'm going to drop it right here and then modify it for my own purposes right? So they're already doing a sort of plug and play sort of thing. For the most part until it gets really technical and then they're doing the fine, the fine details and stuff like that. But there's no reason why an ai or or some sort of intelligence or um, our or automatic automated program in 1 way or another can't do that. We we have there's there's programs out there right now that if you want an app for your event. Or something like that. It's plug and play you just plug in the details and all of a sudden you have an app that you can actually download from the app store for your event and you didn't code anything right? And that's just a human doing that if I can just tell a computer here's my event here's the kind of things I want to do here's the schedule build me an app that I think we're there. 05:47.12 archpodnet Right? It wouldn't be super awesome. It's not going to win app of the year but it's there but the point is in 5 years six seven 8 years 10 years it is going to be there and you're going to be able to say hey I want an app to be able to this is that possible and the thing's going to say nope it's not possible right now or yes it is here. It is done. Bam there you go nobody's going to be selling and creating apps anymore at a certain point in time I fully believe that I'm not like all for the robot revolution. You know the Android revolution but I'm just a realist things get smarter and faster exponentially and the more data we give it. And we're giving it the data of the world. You know all these things are connected to the internet we're giving it. Everyone's data. Everyone's thoughts. Everyone's feelings. Everyone's Facebook posts and in tweets and all that stuff it might end up being a total moron if it's reading Twitter but that being said, you know it's it's smart as what we feed it and we're feeding it literally everything. So I don't know. 06:44.16 Doug Ah, Chris I think we need to stop for Bill Bill's having problems. 06:45.80 archpodnet Oh this bill shed. 06:47.81 Andrew And yeah, he got he got like knocked out.