You have to order pizza right now. Right now we're Hello and welcome to episode number 103 of Unrelenting. I am Darren O'Neal. He is dean of journalism. And these these numbers just keep getting bigger. It's funny how math works, isn't it? Is that what it is? Math. Should we have started going in reverse or using multiples? Maybe it's a creative idea, This is podcast number, show number 22. Now we're getting back down to zero. Once we hit zero, we are done. Whew.
Well, you know, I have to think about this because one of the other fields to fill in in a buzz sprout for each episode is the the the episode number which you know, that makes sense, right? But also the I guess it's the sad season. It's something like season. It's like one of the bigger denominators.
Yeah, well, the some of them are started that with seasons now which I don't like her podcast this isn't a television show we don't go from we don't do like 6 to 9 months and then take a break and come back for the next season, don't we. We should. It's not a bad idea to take like three months off. Yeah. Okay. I'm going to go into edit in real time. So this is. Yeah, it is season. So I've been using for season the year. So like last year all the episodes were season 2023.
Wow, we've done a lot of seasons and yeah, season 2023 and episodes, number one through 100. This is for surging Speaks growth. And then this year the season is 2024. And I thought, well shit, does that mean I go back to episode one, right? You would. And I said, Nah, screw that. I'm going to continue starting with episode 101. So now through this year, it's season 2024 episodes one on one through 107, which makes total sense to me.
I don't know if that's going to confuse some of the players out there. It may, and I think a lot of this stuff is becoming irrelevant now as far as numbering, it's a thing that's always been done in podcasting, but it's never been done in any other sort of content creation. If you're doing a nightly news show on CNN, you don't come out on the first night and be like, This is episode one, and then the next day, Well, I'll tell you what it effects.
It affects the order of the episodes in the RSS feed. Well, that would also just drive them by them. Yeah, but. But it does. That's why I'm saying I tested this. I can create an episode tomorrow with a lower episode number and it will be placed further back in the RSS. I wonder if that's RSS or due to BuzzFeed. I don't know. We have to try it on.
Yeah, you can try on your receive, just like skip skipping the episode number and then like the episode after that episode, try creating an older episode number and see if it sticks it in according to episodes or according to Dave because it actually makes sense that it would be according to episode number because you may publish things out of order but want them to be in the RSS in order.
True. And using the date of publication means you would have to publish and republish everything any time you made a change in the past. That makes sense because otherwise your older episode will always show up at the top of the list, even if it's you're just like replacing an older episode because of something. And let's be honest, most people want as a podcast want the episode number because it's like, Hey, we've been around a long time, give us more money.
You know, somebody asked me that recently when they were like, Yeah, I mean, this is gonna have, I don't know, 107 and 109 or something. You're like, so way more than 20 was. Well, yeah, way more than 20. What? I guess everybody knows that most podcasts never go past 20, so they kind of use that as a So if you're, if you're on episode 18, yeah, you're probably not going to survive.
Yeah. And they had a quick I mean the, the rock and roll pre-show, not really a podcast, but I think we're over 400 now episodes of that, which is just crazy to think about. Yeah, well you're twice a week, so. That's right. 200 episodes per year, right? Well, it's 104 episodes a year. All right, Well, I took the math. It's early in the morning. That's my co-host on the other show. That was a math guy. So I was the. A math guy. Yeah, that was his. That was things that.
Yeah, that's where he got his degree in mathematics. That's his claim to fame. Yeah. I don't know, man. I, you know, I, I think math is important and I can do a lot of math things that people can't, like, actually do that in my head. But at the same time, I mean, that math was never a fun topic for me. And I know it was always work math. And then the the science. I hated any of the time I failed calculus twice before I finally passed it. But you know what?
I laughed three times when I got into the computer programing that flipped for whatever reason. yeah, there was no math involved. Just computer could just do everything. Yeah, you just had to know how to get it to do what you wanted to do and learn the syntax and all of that. Yeah. The beauty of the computer is it's basically just logic. There's it's all about rules. And as long as you can memorize the rules, you're set, you're good to go.
And then C came out and those fuckers just got rid of all the rules and said, You just do it any way you want. And I really disliked C for a long time. I'm getting into the weeds being geekier, but back in the early nineties I had made a prediction that said that with object oriented C, we will get to a point where not a single developer working on a project will ever know how the program works. And I think I was vindicated. I think I'm actually right.
I don't think there is a single developer anymore in any company that truly understands how any piece of software that is written in Object C actually works. It's the magic of it now that you can just write it for you. Yeah, but the problem there is the same as Chad. GP will write it in the way that most people have written it because that's how it learns. Which is not to say it's the best way to do something.
It's certainly not the shortest way to do anything, and it's going to incorporate all the additional linked libraries that people traditionally throw at it, which is why software now is in gigabytes instead of megabytes or even kilobytes the way it used to be. You know, Microsoft Word used to fit on an 800 K floppy and a three and a half inch floppy with plenty of room left over.
Yeah, well, the best example of this is for me has always been Adobe Acrobat, which was originally huge as it was originally svelte. It was great. And now it's turned into the bloated bloat from bloat land. So that runs like 20 different processors on your computer, 24 seven for no apparent reason. It's just running Adobe ship. In case you want to look at a PDF at some point. Yeah, at some point during the month you may want to open a PDF file.
So we need all these processes just running in the background all the time, slowing down your computer. It is scary when you go look at a process monitor and see what's running. yeah, if you know what you're, you know, you can prune that down. I bought, I paid for, as you know, because we talked about it Adobe Suite back during Black Friday. So it would have been November, right?
Yes. I have to redo my I've not installed it because I dread installing it but with all the crap that it adds on there. So now I'm just I've got like this is a stupidest thing in the world. I bought it because it was on sale, but I haven't installed it because I really dislike it. If you decide not to use it, just send that activation code over to me. I'd be glad to add an extra year. Would you like another copy? Yeah, I. I have to go in and buy.
And this is getting harder as well because I've made this told this story before. But I have a friend who's a teacher, doesn't like any of the adobe stuff, so it's like you had no problem using her account. And every year you have to renew. And if you don't, smart friend and if you let Adobe do it, they're going to charge you way more. Yeah. Per month than what you can go by. The used to be very easy to get the activation code. Now Amazon no longer sells the activation codes.
The only thing Amazon has, which is where I do a majority of my shopping, all they have is while you could be a part of the Amazon subscription thing, which I tried once and it was like, No, fuck that. So I've been ordering the cards the last couple of years from Best Buy who still has them. Okay, the teacher Student edition. But this year I went in to get that and they no longer ship them. So I actually have to go into one of the stores which allegedly have them in now, and that's fine.
But it's like, why won't you ship the little credit card size thing that you scratched? A little too many people lie and cheat and steal probably. I noticed though, that BNA, which is a very reputable company, you know, going around for a long time, the activation code through email. So that's probably where I'll end up buying. So what don't, don't they need to confirm your teacher email address or something? They did that originally and they just keep every renewing.
They don't, they don't. That's shocking. I would have expected they they I used to have any to email years ago because I found that there was a college in Arizona that would issue you a student I.D. immediately upon you signing up for a class, whether you paid for it or when you cancel the class, you kept your I.D. because they would assume you're going to redo it next quarter or whatever. Right.
And so it was a to just like purely online with zero cost obtain a fully legitimate student ID but they figured out that they had like a million students and only were having, you know, 25 classes that they were running. So I think they figured out, yeah, this is people are just signing up just to get the damn student I.D. or because it is very profitable to have get discounts on not just Adobe. There's a lot of places. Yeah yeah.
And I, I was an idiot for letting go my original student I.D. back at the University of Minnesota because I have the absolute best email. The email is no. Tell me now. Let me tell you, the chairman just said to you, I'm in Edu Really? Chairman Chairman they let you pass. Username was this. I was the chairman of the Student Services Committee at the University of Minnesota, which is kind of like a student government thing.
And they let me pick the email when they got in that position and I was like, Chairman. And it wasn't like the email wasn't tied to the position. Every new person each year picked their own email that they wanted for that, but they couldn't get. Chairman Because you had it. Exactly. And so I kept chairman for years. And then eventually it said something like, Please verify your student registration, blah, blah, blah. And I was like, okay, I guess we're done.
But in retrospect, like having and maintaining it, I should have signed up for a class just to keep it or something, you know? Yeah, absolutely. I was just looking at the Be an Age is selling it for 179 which makes it right at $15 a month where if you don't do that and if you it with Adobe charge you the student rate I think it's over double that I think it's like 33 bucks.
But I think the normal thing to try I'm paying I'm paying I think either 30 or 33 bucks a month as the discounted photographer. Right. So you can't be 15 bucks a month for the whole suite. If you make ten bucks a month is the cheapest it is.
And I do think it's even if you go for literally cheaper and, you know Adobe has their own recording software and out to well that's what we record the show on I mean the as far I mean like one that replaces clean feet there's a website that you can go to the record shows and Adobe now I did not know that I have to look for that. Yeah, yeah. You should do some research. It's what is it called. It's like podcast on Adobe or something. Right.
Because I know they have these simple like will clean up the audio which is basically just gonna clean up background noises and some of the reverb. Yeah. But they do have it's, but there's a bunch of them that, that's what I'm saying. So clean feed is bizarre that they're jacking their prices up instead of down because they're more and more competitors. When they first came out, they were kind of the only game in town. They were the best done.
Yeah. And right now, like I said, the the script, which is the editing software I use for audio, they bought a company that is a clean feed competitor and now that's free to anybody that pays for the editing for clean feed and you can do. I asked them like, is there a limit to the number of shows, number of minutes, number recordings or anything? They're like, Nope, it's still unlimited right now. So yeah. And they have it's exactly like clean feed.
It's all web browser based, but with the addition of video, if you want it. But honestly, I've been doing Zoom now for about a year and Zoom has added more professional features like original sound for musicians, high fidelity music mode, where they realize mode that one of the big things that happened during the pandemic a lot of musicians wanted to play together online.
Yes. Yeah. And they stepped it up because when they originally would, we went to clean, feed, Zoom as an option was horrible. It was horrible. I think in a lot of ways the pandemic was the best thing that ever happened to Zoom because they got a huge inflow of money that they then spent on development because it went from just being a screen sharing app with really crappy video and, you know, a lot of compressed audio to being really the like the lowest latency app for doing meetings.
Now there are other apps that do meetings, but even if you compare the Google one like Google Meet or whatever to zoom, first of all, Zoom doesn't run on the browser. It's got a native app for the phone's, got a native app for the computers, obviously, and it comes out with updates almost every week. I mean, it's as bad as Microsoft. Like constantly doing up their Skype is like that too. It has been for a decade plus. I'm like, What the hell do you keep up there? Nobody can use a Skype anymore.
I know Skype was the thing back in the day. You remember No agenda escape. yeah. And it had some issues. Occasionally you you'd have the the squeaky sound, the robotic voice, the robotic voice or chipmunk voice as it was. Got this synchronized. Yes, but I don't know, man. I at this point for podcasts specifically, I'm a big fan of Zoom because my only real big complaint with Zoom is it's a Chinese company, right? But the podcast just is not secret internal meeting conversation to begin with.
It's literally made to be distributed. Yeah. So we're literally by using Zoom, picking up some extra listeners in China. And that way there's more eyes listening and it's great. Exactly. I'm like, This is great. I don't mind this shit at all because it's, it's being used for creating an excerpt. Now we are in clean feed today. Yes, but I am kind of trying to nudge Darren along into doing a I don't mind test of Zoom. I don't mind trying zoom. That's fine. Yeah.
And I know since we last did a show, is this possible that both you and I have had our plus conversations with the great comic strip blogger? Yes. Yeah. I was on the phone for like, an hour. The guy he talks, he thought he should be a podcaster because the TV man was. I know at 125% speed. Yes. Yes. I mean, that is the ironic thing is like I encourage people to listen that I had 25%. But you got to remember, the reason I encourage that is because I don't speak all that fast.
I speak what I consider a normal, comfortable base. CSB speaks at 125% speed, and I think you speak a little quicker on this show just to keep up than you do a lot Surging speaks. You have a little bit more of a relaxed, effortless, surging speed. So I'm usually multitasking. So I'm like surfing the web while I'm doing the show recording or I'm doing something else.
The thing that I found most interesting because what CSB is doing over at a domain that I own, so I'm kind of helping them out here where I can. I see how that works. No agenda that Tom, is that to make these little custom. TS You're feeding it the transcripts. And one thing I don't do on transcripts, but I know no agenda does, is try to separate and mark when each individual is speaking, it has that tag. So it'll say seriously, you don't do that. No, I never have. No, I never have.
But the interesting thing is the way he figures out who is who, even on a text transcript, which is very interesting to me because I figured, well, what this is going to do is it's going to be totally confused on what, you know, Dean said and what I said and what Ryan said and what I said, no, it seems to be fairly good at knowing, even from a because all the air systems getting here is the transcripts. So we can't hear the voice. It's not like you're feeding in the audio.
I tried doing that and you can do that, but only if you're using the APIs, not using the little idiot version that the stupid. Yes, the creator. Right. Easy. Go in and show you very simply how to do it. Great that people are getting intuition. Yeah, right. Well, for people who have never done it, it's like this is a fun way to get you're any idiot with the podcast that has transcripts can go in without any knowledge of anything. Yeah. And get this up and running, which is cool.
Now if you know what you're doing, you can no further. The one thing that I did ask CSB he wasn't sure about is the eyes are woke right We know very certainly as well so it's just have the podcast this morning that woke right but is that morning all of the woke stuff in the prompts that's going on on open air which is very interesting like if I do find the link, I'm sure it's not all the work stuff, but a lot.
But so the question I have is, okay, so what is it going to censor the podcast when you see the script, when it's doing answers, it probably will. It'll, it would be an interesting test. And as you know, or maybe you don't know answer Jean, I did an interview with a guy who's whose name is Nick. Terrific energy. How do you spell that? Yeah, exactly how you'd think he'd spell it with an asterisks. So I spoke with asterisk. He doesn't because, you know, he's allowed to.
And now I am allowed to say it. I'm just not allowed to spell it. But I suspect that the that GPT will just censor that completely. That wouldn't surprise me because again, it doesn't take things in context. And since it's a level, is there a context for that? I think there is. I think it is a comedy for one which. Okay, there. Okay. Okay. Yeah. Yeah. If it's so, if you label something as comedic, it should just turn off the woke filter. Well, it's probably not going to turn it off.
But again, this concept that any use of that word in particular, what word? The word that you don't want to say, that you want to use the astronaut. I don't even know what word you're talking about. Don't give it my transcripts. Do not want it there. But there was a high school teacher at an all girls Catholic high school here in Chicago that got fired because she made the comparison between the Washington cannons basketball players.
But it was the saying, hey, if saying Redskins to somebody of Native American is the same as saying this word to a black person, and it was done in a academic setting and I see this is where it all breaks down for me, is that if you can't teach your children, well, this is the word that's bad. If you can't even say it in the academic setting and it gets you fired, then we're fucked because we are fucked. That's exactly right. That's the problem.
And when you start doing this, the I these large language models, the and the image creation models, they can do anything. It is the fact that you had somebody adding another layer that's like, well, what if somebody is asking for this don't get they literally have turned off Biden and and Hitler but they kept Trump well of course. So you could make caricatures of Trump all you want, but you can't make them of Biden.
And you can't have it create anything that references Hitler because it thinks you're going to glorify Hitler. It's it's the weirdest thing. And it's clearly humans that are making these rules. It's not coming up with them on its own right. And no, because I can make Genghis Khan all I want, even though he killed more people, let's remember, which I was that that was released.
Was it the early jet that was immediately like pulled because people were having it say the most racist stuff in the world? I mean, there was one of these that was the the original iteration was like, whoa, whoa, whoa, whoa, whoa. Because it immediately went all Nazi. Are people now Really? Yeah. I don't remember that. Yeah, it was it was interesting because it's like left to its own devices, even the even the machines are good. They go all racist.
Well, clearly, that machine was trained on conversations from inside of the Fed of ours. Probably just because that's where the majority of them are Nazi related. Well, and this is the intriguing thing, is that once you have A.I. material feeding A.I. for which it's learning from, to then create more.
the amount of inaccuracies you would assume would continue to blow up unless somebody figures out a way to have a neural network or something that can actually fact check, which I don't think we're anywhere close to that, because this is not a as you said, this is not a thinking thing. It is a I'm looking at the data that's out there. Yeah, it's a statistic. It's a statistical based system. So, I mean, it's a lot of fun that it to get things.
Yeah. When I put up this the bought the URL no agenda trip.com. I'm like, you know what? I need to have something in order for the chat. Get folks to confirm that I own it. So it's like, okay, I need a basic web page is coming soon. Page sitting there and I'm like, Do I, do I load up Dreamweaver? Do I go in and just code something by hand? I'm like, No, fuck that.
I mean. Jeffrey Beatty So I said, Jackie Beatty I need a basic HD HTML page that says, Coming soon for this domain, press the button and beat it to bits. The code I want to read out now. I'm like, That's fantastic. I mean, it wasn't pretty. It wasn't no, definitely not pretty, but it was usable and it was immediate. So for those things, I applaud it. But this is because it goes, hey, well, I know how to write code. I know all about the different HDL.
Yeah, it does know how to write code, just not the best code. That's all I know. We'll see. But this is where the money will come in is having. And this is what a lot of these custom ads, which is what comic strip bloggers working in when he's uploading all of these transcripts, there are what are you allowed to say that? Is he giving you the the permission to discuss this topic? Yes. Okay. Yes, do that because it's an exciting little foray.
And I think just like Apple all of a sudden is doing transcripts which whether people like them or not, now that Apple's providing transcripts, there will be a lot more podcasts who would have never done that on their own, who would have never gone through the extra step because it is a pain in the ass. Apple's going to spell nigger, if correctly. I would hope so, because otherwise they are very racist.
But the more transcripts that are out there, the more data that there is to feed into the air. And I see at some point podcast hosts that will be offering this as a part of the deal, which is, hey, we will update your custom chat GPT. So anybody that has questions about your podcast can just go in, query them here, but we'll make make sure it's updated with your latest transcripts and whatever else they're feeding into it.
I, I won't say that the current version of GPT does not do a very good job of that. So I played around with the stuff that is created. I also made my own and it it gives extreme generic answers with, with fairly specific questions, which is not really what people would be looking for if they go to the trouble of using a GPT that is specifically tied to your product. So I think they've got a ways to go. Yes, to make that customization better.
The other thing is I think that the purely self running like not in the cloud, not using Chad's CBT for language models will become more freely available as time goes on and you'll be able to just run a fairly generic pre populated, you know, GPT for lack of a better word agent that you have provided the custom links for it to learn from. Right. But you'll be able to do it without using Chad's CBT proper, without using Amazon, without using any of these things.
You can just literally run out many holes because. Right. My understanding is that could be incorrect. But my understanding is that the thing that takes a lot of processing power that they're using GPT shoes to do is the initial training. It's not the actual running of the queries, it's the training of the system because it has to go through millions and millions of iterations. Right. I believe that. Get trained. Yeah.
So you could train it and then when you're done training it's offloaded on a system that doesn't have the GPUs running, that doesn't have all the horsepower that you originally needed while you were setting it up. And I think that's where, if I'm correct, I think that's why they'll become a lot more prevalent because you'll just be able to get a copy of a Pre-Trained one in the whatever language and whatever political leaning you want.
Somebody will have a liberal version trained, somebody will have a conservative version trained without all the crap, and then you can just install that to use as you want. Now, you know, I'm assuming all this shit is going to happen fairly soon, but I don't know. It may happen three months from now and may happen three years from now. Right. And what these custom things are doing basically is telling the brain, whatever that may be, look at this data first.
So if I take all of the transcripts from unrelated to get upload them and then say, well, use that first. So look to this data for the answer to if he can't find it, then it'll go to outside of that. Right. Which is fine. Which, which is great. Yeah. That's what you wanted. But also depending on who's doing your transcript, you may have a lot of errors.
There's certain companies like, for example, this script which I mentioned earlier, you know, they've been doing voice to text for over a decade now. They were one of the pioneers in the field. They're extremely accurate, but even they still occasionally will have the wrong word compared to what was said, because a person has a weird accent or, you know, for whatever. It's even beyond that. One thing I've noticed doing the transcription, I use the Adobe software for it.
Yeah. Is in every show, especially for random thoughts. And I understand it's never going to understand random thoughts. I get that. But yeah, on every episode at the beginning and at the end I said my name Darren O'Neil, and quite often with Darnell. Right? Darnell I'll throw Daryl. But the funny thing is it quite often spells Darren and O'Neil differently at the beginning and at the end of the show. That's bizarre.
And because this is where you get into problems with words and then if you're looking for a name to associate something with, well, D.A.R.E. or Ta da, Brian and O'Neil could be spelled with one L or two LS or a alphabetically. Exactly. All of those variants because the Irish don't know how to spell. They're all correct because they're all different variants and I guess they don't know how to spell it. Probably they drink a lot.
Yeah, it depends on the state of drinking that you're spelling changes. Yeah, but you're slurring. And when you say don't all them. Yes. The more L's you have in name and the more drinking was experienced prior to saying it. But I'm curious how that affects the ease as to when people are asking, does it does it see a Darren with a1l or two L's and he was a different person than Darren O and a L when they show up, or is it smart enough to go, wait, this is probably the same person?
Yeah, but I'm not sure if it does or not. I trained this script on my voice years ago, like three years ago, and so it is actually been smart enough to know who I am in listening to the podcast recordings since then automatically. But it's also now figured out who benefits automatically, which is pretty cool. I bet he's going to have his name on the list. All right. He sounded the list ever since we started doing a podcast.
Hey, guess what you get to do now you get to talk about National Beagle Day while I go downstairs and grab my bagel that's just been delivered. my goodness. It's National Bagel Day. Well, that I didn't know. We got we got notified of that actually, like an hour ago. An ex, both of us that you went to. And I was like, shit, I need to order a bagel because I'm out of bagels and it's National Bagel Day and I'm not going to be not eating a bagel bagel day.
So you talked about that while I run downstairs and grab it. I don't know what what is one have to say about National Bagel Day except wow, what a holiday we have here And now, you know, if you need to move a lot of product with Jean, just tell him it's National Something Day. Like, yeah, it's National Iced Tea Day. And he'll be like, I have to buy more. He's always got the HD, but if you're looking to schlep some stuff off, it's exactly what you tell him. National Bagel Day. And you know what?
There's not just going to be bagels, there's going to be bagels. They're going to be lox. Although I don't know what to Austin Bagels tastes like. I'm pretty sure that the only real bagels come out of New York. And if he wasn't aware that it was National Bagel Day, he didn't have his bagel guy in New York pack up some bagels last night and overnight, that would have been the only way to do it.
The only way to do it well would have been to go to your bagel guy in New York and get those bagels sent to you, get some fresh lox, have yourself a delicious breakfast. What if Gene ever actually goes out of the house? Everybody just comes and brings the stuff directly to them. Like, got your bagels. we hear the chair. Let the bake.