Let's make things that matter. Together.
(Note: In keeping with our theme of artificial intelligence, this episode is published with the help of AI in several areas. Even though it’s super helpful, it still has a long way to go before it’s capable of replacing a human for our workflow.)
In our previous episode, I spoke with Srinivas Rao about all things AI, specifically why we shouldn’t be afraid of AI taking over our careers (and the world). But let’s face it, as time goes on technical jobs are bound to be replaced at some point by our ever-advancing technologies. And that’s the threat that my guest, Michael Kammes, and I are here to talk about today. Michael is the Senior Director of Innovation at Shift Media with over 20 years of post-production experience and extensive knowledge on trends, products, production workflows and post technology.
In our conversation, Michael and I talk about the speed at which our current AI has been changing the world – and why that really means. We talk about the actual existential threats AI brings to us and what kinds of jobs AI could potentially replace in the foreseeable future. But don’t worry, it’s not all doom and gloom! As we dig deeper into the topic, we also shed light on what it takes to not only be irreplaceable in this age of AI, but how to thrive in it.
As Michael put it, AI is not going away and we cannot fight this fact (or it). Whether you’re feeling as if your career is threatened by AI, or you are excited about the potential career advancement it can assist you with, you’re going to love this conversation as much as I did. Michael’s advice is not only helpful to people in the entertainment world, but to anyone across any industry.
Michael Kammes, Senior Director of Innovation at Shift Media, helps shape their suite of solutions in the marketplace, speaking at industry events across the country. Michael has over 20 years of post-production experience, with extensive knowledge of trends, products, production workflows and post technology through roles at BeBop Technology and Key Code Media. His podcast, “5 THINGS,” answers his audience’s most important questions in media creation, drawing an audience of industry insiders worldwide.
Facebook, Follow our guest on Instagram, Linkedin, Click to visit our guest’s website
Michael Kammes: We don't get many opportunities to have. And I'm trying to avoid all the cliches like sea change and paradigm shift a game changer. Yeah, we rarely have things that impact the industry. Totally. And completely. You know, we had the tsunami in 2011. Right, that that was a big one. But since then, or maybe the transition to UHC, there's been nothing that's been this widespread in decades. So it's an important and fun time to be around, despite all the stress that it's causing people.
Zack Arnold: Well, I can't imagine a better way to just seamlessly transition and actually recording our podcast interview together than that, because you have just let us off into the stratosphere. Before we even get started. I am here today with the Michael comas, who is the Senior Director of Innovation at shift media. You're also the host of the five things web series, as well as the star wipes and Alan Smith, the tech podcasts. And I will also say unequivocally that I consider you a friend. So Michael Thomas, it is an absolute pleasure to have you back on the show for I'm can't even keep count, but it's at least the third time. And you're on a very short list of people that have been here more than twice. So it's a pleasure to have you back. Once again,
Michael Kammes: I'm always honored because I love trading tech barbs with you. I love what you've been doing over the past several years. So thank you so much for having me back on.
Zack Arnold: Absolutely. You're more than welcome. And the conversation that we're having today, as you know, is the conversation that everybody is having right now, which is what the hell do we do about artificial intelligence? And the place that I actually want to start the conversation? Is that I noticed you don't consider yourself an AI expert. What are your thoughts on AI experts? Michael?
Michael Kammes: What's funny is AI has been around for for quite a while. And there's been folks who specialize in how to create models, and what the math is behind that and the workflow and chain behind crafting a model. But how it's used practically, is almost anyone's guess. And so there's a lot of side hustles going on right now where people proclaim to be aI experts. And I'm sure they know a lot. I'm sure they know more than I do. But for folks who say I know what's going to happen in a year in five years. It's craps. In fact, I saw study not too long ago that I think two years ago, they asked AI experts, how long will it take until we can train a model for like under $300, or something like that. And AI experts and published papers and studies said, Oh, it'll be years, we, we blew that open and under under 12 months. So it just shows that the experts right now are so off that it's it's to me like a roller coaster, where you just have to hold on, and hope that that seat belt holds you in for the entire ride.
Zack Arnold: Yeah, and I'm sure that something you can, you can dig into even further. And we don't necessarily need to go in that direction. But kind of following it on a superficial level. Those that are literally the godfathers of this technology that are the CEOs and founders of companies that are developing AI are coming out and saying, I have no idea what's going to happen. So I'm just I'm kind of putting a you know, putting a real emphasis on anybody out there right now that are claims to be some form of expert, or they not only understand all the technology and how it's working, or more importantly, I am confident I know where it's going next, you're probably completely totally full of shit.
Michael Kammes: It's the concept of applied AI. And and I know it sounds like that's a you know, college one on one course. But there's building the tool. And then how will that tool be used in an enterprising or a some men just want to watch the world burn kind of way. And I think that's where the differences is that it's easy to say, I want the AI to do this, this and this. But once it's out in the world, how will people use it? I liken it to music, right? A lot of time when do artists create music, they say it's my message. But when I put it out in the world, it takes on a new meaning because it's what you believe that is and how you interpret that. And that's exactly how AI tools will be used above and beyond what their core purpose is for.
Zack Arnold: Yeah, so the there's a whole lot that we can go into in this conversation. And I think that kind of one of the general foundations that I just want to lay for it is that I have no interest in talking about, well, what's the best plugin for you know, having a do my editing for me and adding my cuts or I want somebody to do an EQ pass on my audio bike. I you already know me very well. I could care less about the actual tech and the plugins. And as you alluded to in a recent presentation at last epoch which I've actually I'll link to the video for that again, because you you go you go more into some of the specific tech, but one of the things that you were very smart about in your presentation was saying six one So now every single one of these companies could be at a business. So it's very, it's moving so fast that it's hard to say this is now the definitive industry leader in image generation, or in you know, doing social media, whatever it is, everybody's trying to figure it out. And it's kind of this mad gold rush. It's kind of like the.com, boom, you know. So in that sense, I just want I want to set the table that this isn't really about very specific tech questions or applications. It's about asking the bigger questions about what does my livelihood look like, as a creative professional? And where are we going in the future? And really, the biggest question that I think is on so many people's minds that I know, I don't have the answer to, I'm going to assume you don't have the answer. We're just going to talk through it like Rational adults is are we all going to be replaced. So that's that's kind of the the bigger picture conversations that I wanted to have is not so much about the tech, but just digging into the implications of how artificial intelligence is going to change the world for creatives. And frankly, for everybody, it's clear that it's no longer if it's going to, it's how it's going to. But I want to, I want to step a little bit backwards before we get very, very existential. And I want to go a little bit deeper into something that you said already, which is, there's already been AI around for a long time. So just super, super basic, I want to talk a little bit about how AI has been around in the creative space and in tech for way longer than we might realize, and talk about what's different. Because a year ago, there was a ton of AI and all of our programs, but nobody was talking about it and there wasn't this existential threat. And now all of a sudden, everything has changed. So just give us a little bit of background of what we're talking about with some basic AI. And now why the conversation has changed so much.
Michael Kammes: A lot of the effects artists, I talked to chuckle over this, because they're like, we've had AI tools for years, I mean, rotoscoping, right, a lot of that is based on looking at adjacent pixels, and saying, This is what it should look like here, if I were to remove this object. So just in recent history, that's been there. Right. And so because of that, there's already there's already been a little bit of foundational knowledge. around that. We've also had things like script sync, right? If you're utilizing like Media Composer, you're familiar with phrase find and script sync, which is being able to analyze the audio and then match it to text in a script. And while that's on kind of the lower, I don't say lower end of the AI gene pool, it definitely is I'm kind of on the lower end, but gave us a glimpse of of what's possible today.
Zack Arnold: Yeah, I mean, I would one that I can think of that has really kind of taken the industry by storm, especially unscripted in documentary for years has been automatic transcription. I remember when I did my documentary film, I don't long time ago now at least 10 years ago, now, I remember having to pay 1000s of dollars, to pay human individuals to sit down and go through all of my transcript or go through all of my interviews, and type everything by hand word for word, and I had this giant binder of all of these transcripts. And now you just take an audio file, drag and drop, at least for me, it's into otter, right, boom, done, and it's not perfect. But to me, those are the kinds of things that have been in the category of oh, this is really useful, and it saves me some time and certainly saves me money. But then you think about the person that I hired 1015 years ago who had an entire team of transcriptionist, I would just assume they've been out of business for a long time now. But those are the kinds of things that again, have already been in our lives. Siri is a form of artificial intelligence. So many things happening behind the scenes, whether it's a smart fridge or whatever it is. But what I'm really curious about is why is the conversation changed? And why is it changed so fast? Help us understand the difference now?
Michael Kammes: Well, the difference is billions of dollars are being poured into it, which doesn't hurt anyone. And what we saw is the mass adoption, because of what they call MLMs, or large language models. That's the kind of chat GPT that we're most of us are familiar with. And that's the the feeding into a self training computer, question and answer. So when someone asks it or typed in a question, it can then respond and then mathematically determine what is probably the next word going to be in context here. And once that became part of our, I don't say experience, but when that started looking like it's gonna be reality, that's something that can be done. The modularization of the components that made that up, have now become commercially viable, where you can train a model without having to spend hundreds of 1000s of dollars it can be done for just hundreds or 1000s of dollars, and being able to then take that methodology, and how these machines are learning and use that for different tasks, whether it be creating a fantasy story, or creating a fantasy image These purpose built models. That's what's exploding now. Because now you can tailor these models into any way you see fit, and then put a.com at the end or.ai And now you have a job,
Zack Arnold: Right. And it seems to me that if we're if I know that you love terms like inflection point, and turning point and all these other things, but I think that if we're going to look back on this, with some hindsight, whether it's in six months or five years, what you already mentioned, chat GPT, that seemed to be when the conversation changed, because everybody knew what artificial intelligence was. And we'd all seen the Terminator movies and we knew we had some you know, AI going on and you know, our transcription programs or whatever it might be in our NLCS or you know, whatever field you're in. But it was when you had this model, where you could say, brainstorm, for me, think about these ideas for me, or synthesize this massive amount of information. That was the first time that I had the experience of Whoa, this is a totally different level of anything we've ever seen as a tool or can comprehend. And it's very clear that everything is going to change. So I definitely had that aha moment with Chad GPT. And I had it, it was right as at the very beginning of the conversation, when pretty much everybody that I talked to hadn't heard of it, but a couple of people that were insiders in that world that heard of it, and pass it along to me, and I would start mentioning it like I haven't mentioned it to my wife, and he's like, chat, what am I you know, just regular, everyday discourse, and that it only was maybe weeks where all of a sudden, it's my father in law. Have you heard of this chat? BTG thing? Like, what is? Whoa, this is happening really fast. And I don't know the exact statistics. But you can correct me if I'm wrong, but isn't chat GPT the fastest growing user base of any program in history by like, some ridiculous exponential amount,
Michael Kammes: If I'm not mistaken, is the fastest adopted consumer technology ever. And just like you, I had the same experience about having sushi with my partner, and I mentioned something about chat GPT. And she said, What showed her on my phone and the the real, I would say Genesis, but the real turning point was that the the the models have been around and the concepts been around. But we lacked that one crucial tidbit, which was how does a human interact with it? Right? When when it became conversational, when that piece of the puzzle was was completed. That's what opened the floodgates, because now it was accessible to anyone that that had physical access or technical access. But it was that the the ability to understand what you and I would type in or even speak, as opposed to feeding it, some kind of code or script.
Zack Arnold: Yeah, and I think that the the other thing, too, that we've become so used to imagine, it wouldn't be that long ago, certainly in our lifetime as adults. But imagine what we can just do with a Google search. Now, that would have been jaw dropping, give or take 20 to 25 years ago, I can't remember when you know, those searches really became in the mainstream. But the very first version of searches through AltaVista or Netscape or whatever, were very, very rudimentary, where it literally was just You got something spit back to you. That was essentially what you asked for, then all of a sudden, you have a Google search where you it doesn't have to be the perfect wording or the perfect Boolean characters, and you can kind of rough it and you get all these other things that are relevant, but not exactly what you typed as, like, oh, that's, that's pretty cool that it does that. But for me, it was the idea that you're generating ideas. It's not a matter of here's the search and I want you to feed back this information that you're finding it's solve this problem generate this thought that was the thing for me that really was jaw dropping, is give me five solutions to this problem or generate pros like I've I've tested chat GPT to see if it can write something for me in my voice. Because I'm certainly one of those people that based on whatever you know, crawling these bots are doing and models they're learning from I would assume that they've you know, done some crawling of my own voice on the podcast on my own blog, because I'm connected to the the matrix like everybody else. I've asked it just experiment with write a newsletter about this. I've never actually published anything like that. But I wanted to see can I get this to do it? And there were moments when I'm reading it, I'm like, Holy shit, that's better than when I read a written here. That's where I got scared. I was like, Whoa, like, how much longer am I even going to be relevant, whether it's as an editor as a newsletter writer, you know, as a member of the thought leadership space, that was the the game changing moment of like, this thing knows how to think like us. So to me, that was when things started to change, at least for me personally,
Michael Kammes: I will take a little bit exception of with the word think.
Zack Arnold: Yeah. And I was hoping you what I
Michael Kammes: It's repeating patterns that have a higher probability of coming from something that you'd write and I've done the same thing called vanity. Call it Hey, Want to see how I can augment my marketing strategy? But yes, I type things in and say make it sound like me make it sound like the five thing series. And it works really well. And for the average person out there that that doesn't have a body of public work, you can do that almost free. Throw in some technical jargon out there. One of the words you may hear with AI are embeddings. Some people may call it vector embeddings. And that's being able to take your own content, right, whether it's a book you like, or whether it's Warren Piece, right? Whether it's the personal readings you've done, but I've never put on the internet, being able to have that indexed and added to an embedding, that AI can then talk to means you can then ask a chat GBT like application, tell me something or answer a question using the data in this embedding. And I'll tell you that it's it's phenomenal. When you when you want to make it sound more like you, or index data that may be proprietary. It's a, it's a great way to get a leg up with information that isn't publicly available.
Zack Arnold: Yeah, and what I wanted to get a little bit deeper into for those that may not really understand all the differences in the technology, because I'm very much a Luddite when it comes to most of this. But I realized I have to at least do some basic education because this is a serious game changer. And the reason I wanted to bring up this idea of quote unquote, thinking for me is that I know better understand that if I put something whether it's in the chat GPT, or let's say that there's some closed model that scanned every word that I've ever written and listened to 1000s of hours of my coaching calls, I know that it's not actually thinking for me, the only way that it can generate ideas is based on the information that you feed it. So it really is just a matter of a recompiling and regurgitating of the existing information that it has. Is that a layman's description that's fairly accurate? Or am I way off for understanding how it actually works?
Michael Kammes: No, I think that'll serve our purposes for this call. I would not expect LLM as it stands right now. It's not going to create something out of nothing. It's not going to create something that's not intrinsically based on something that it's scanned in multiple formats for multiple people. So that I want to make sure that that we're making that completely clear. Yeah,
Zack Arnold: Yeah, so the fact that it's derivative, I think, is a really important part of this conversation. And the reason I think it's an important part of this conversation, is that something that I've been saying for a while and I know that when you and I were on the stump the gurus panel a couple of weeks ago, last epoch, I brought up this idea, and I know you second it, it of those that are highly specialized with one specific technological craft, those are the people that are going to be in trouble, because it's going to be much, much easier for technology to replace that one specific thing that you specialize in. And I firmly believe, as do a lot of other experts in this space, that this is the rise of the generalist, where it's our ability to problem solve across a multitude of complex challenges, or it's our ability to understand human emotion and human nuance. It's those people that I believe are going to survive relatively unscathed. And there's still going to be a learning curve in a relationship with the technology. But I was hoping you could go a little bit deeper from your much more educated perspective about this idea of will the specialists become extinct while the generalists become extinct? Well, nobody will everybody. And I'm not talking in an existential as people nature, but we're just talking from, you know, making a living as a creative. I'm curious what your thoughts are. And again, you don't know the answers, nor does anybody else. But you spend a lot more time in the space than the average bear. So what are your thoughts about the existential threat that so many people are afraid of, as far as I've got this one specific craft? Versus I'm a generalist or otherwise? Well, what do you think about all this?
Michael Kammes: There's a, there's a bunch of data points on that. And although nothing is an exact one to one with what we're experiencing now there are some things we can look at in the past and use that as an approximation. I'm drawing a blank on his name. I believe he works for Microsoft. And he used the term hackers. He said the the the people that are that are going to thrive utilizing AI are hackers. And what he meant by hackers was people who are willing to use AI, along with their traditional methodologies and being able to blend the two and kind of hack the two together. Those are the folks that are going to progress. The founder of tools called Philibert Liberland, if I'm pronouncing his last name, right, had a great line which was AI will replace humans humans who use AI will and I think that's goes to goes right to your point. We've also seen in our industry, if we take AI off the table, we used to have people who all they did was edit, didn't do motion graphics didn't do sound, they were a picture editor, we used to have a job of a compression test, right. So when you encoded stuff for DVD or into Flash for the web, right, and we've seen over the years that those jobs have morphed to encompass more responsibilities, except in the top 1% of the 1%. Editors, now we're expected to do some level of motion graphics, or even a rough color pass or competence, some graphics, right. So obviously, those jobs are just being blown up with, here's what you need to do, which falls right into what you said, which was being a generalist doing more. To go back to another example You brought up a few minutes ago, you talked about the transcriptions you used to do for your documentary, and that you know how many 1000s of dollars you you paid? Well, that is a very, not to quote the movie, a very specific set of skills, right. And as you pointed out, those are the ones that are going to get replaced first, because it's easier to build a model to replicate a very narrow lane than it is to create a model that can be everything to everyone.
Zack Arnold: And this is what I want to go into a little bit more. Next is this idea of what different specialties are already not necessarily in danger, but they need to be aware of the things that are changing, versus those that are like, this is something I need to pay attention to, but I don't need to worry about as much. And again, I don't I'm not interested in getting into the specific companies or the tech itself. But if I'm somebody that's in a creative field, and we can make the assumption that, you know, probably the majority of people that are listening to this are in post production are in editing or in writing, some entertainment driven or you know, content creation media. So just kind of give me a general overview of what are some of the crazy things that are already completely and totally possible with artificial intelligence that weren't just a year ago?
Michael Kammes: It's it's funny, because what's possible with AI and generative AI, broadly speaking, now, and for the foreseeable future, is going to be stuff you don't want to do anyway. Right? Unless you're hiring someone to transcribe something. Is there anyone listening to this podcast that says, I can't wait to transcribe something? No, no one wants to go in and organize clips, and find every time someone's face pops up. No one wants to do that kind of organization, right? Heck, that's why at least have had some of that stuff built in for years. So it's the things that we don't like to do. We then start moving up to things like versioning, right? So I need to put a different audio track on this file before it goes out and swap out the captioning files. Okay, great. Now we need to maybe morph the mouth. So it matches the language that's been overdub, so it works better for that localized deployment. So as we move up the food chain, there will be some of those jobs that are replaced. There's no way around it jobs have been morphing for years, especially in our industry. But it's it's not going to be the story telling parts, it's going to be the supporting skills that are needed that are going to make that easier.
Zack Arnold: Well, I know that just a few of the things that you demonstrated, which were it was the tip of the iceberg of all the jaw dropping things that are not quite ready for primetime. But we can see that with the speed, everything is moving, it's not going to take long. But the idea that we can just create an image, we can just say, make an image of this person in this background with 100 extra rows behind them. Like the even though it's not perfect yet, we can see the word we're down, we're not that far down the road from I'm just going to enter a prompt and it's going to spit out a five minute short film or a 15 minute short film, or maybe we're not that far from I can actually create an entire feature film, another one that you demonstrated, and it was one of those where I was thinking oh my God, I need to get this for me, which it makes my life easier. But if I were an actor or a speaker or wooden, I don't remember the name of the technology itself. But it was basically you can do you can correct me on the specifics but you know, shoot yourself talking in front of a camera just talking right to the lens for a couple of minutes. It models that now you can enter whatever text you want. And it's just you performing and you can say raise my eyebrow at this moment or have me scratch my nose or whatever. And as soon as I saw this person talking and I saw their eyebrow raise of inflection based on the way you prompted it, I was like this completely completely changes the game and psi wanna because I hate being in front of the camera. So. So there's this whole other level beyond the efficiencies of actually creating content or replicating humans that I don't think Most people don't even realize exist yet.
Michael Kammes: And I'm torn because there's two sides to this as a as someone who maybe you were working on a web video, Zack, or maybe you're working on a documentary, where you need B roll, right? We've all had to go through the slog of, I'm gonna go to this stock site and see what I can find, right? If there's no budget to hire someone for that, or, man, I'm going to spend all day just hitting preview on these audio files to see what works, right. That stuff that is, is bothers bothersome isn't the right word is not the pinnacle of the creative experience, let's put it that way. And be able to generate that on the fly, or at least tailor something that's more to what you want. And what the story demands is fantastic. But then again, we look across the aisle, and we see our friends who are making money shooting aerials, right? And helicopters. Oh, wait, there jobs are being taken by drones. But now you know. So there's this this food chain of jobs morphing and beach things being taken over. What I'm excited for is the amount of web video that I consume. And the terrible B roll that is used. I'm looking forward to a time where in our Anneli we can say I want the B roll to be this. No, I want to pan right to left left to right. I want it to be nighttime, not daytime. And I think from a creative storytelling standpoint, that's going to be great. The downside? Is is for everything that that makes it great. There is the fallacy of well, I'll have more time to create because these tools are doing it for me. No, that's going to compress the time you have to create, which means you have to have greater throughput. So there's always a consequence for that, unfortunately,
Zack Arnold: Yeah, the this easily could be a three part episode about the one thing you just said, which is the black hole of productivity, this idea that will this thing is going to make my life easier or save me time. Think of all the extra time I'll have in my life to get more sleep and go to the gym and eat better and socialize with my friends. Well, you know what, I don't beat my clothes on a rock outside anymore. I've got a washer and dryer, ask me how that's doing for my quality of life and all of the extra time that I have to spend with friends and family and on my health, right. And this is always the promise, it's always the promise and it drives me crazy. Think of all the time you're gonna save in your life. Because you don't have to spend hours scouring stock footage, all we do is we continue to move the goalposts so that today's miracle meeting this insane deadline on an insane budget then becomes tomorrow's expectation. So the fear, for me is the exact opposite of the excitement that most people have man, think of all the time, I'm gonna have to be creative looking through stock footage, I don't have to do that anymore. There was another feature. And I apologize to anybody that uses Adobe, I can't remember the name of it. But it was another one of those jaw dropping moments where they had a queue of music. And all they did was shorten it. And it edited itself. You didn't have to put edits in it or say, Well, I have to nudge this by a frame or two and cross dissolve it was this was a minute 30 I want to make it a minute and it recomposed itself on the fly. It's like well, there go the music editors. But again, it didn't just save me an hour of time to do something else. It moves the goalposts because now the expectation has changed. So the reason this scares me so much is the speed at which this is moving. This is this is no different than all the advent of all the other modern technology that I feel that we've been immersed in over the last 20 or 30 years. But it's moved at a slow enough pace that we kind of sorta could adjust. Like I don't think we really noticed the transition from I'm driving around with a Thomas guide to now I'm using Google Maps. If I look back at it, yeah, it happened quickly. But it wasn't like one morning, you went in the Thomas guide, and the next day like Oh, my God, my Thomas guide has disappeared. And I don't know what to do, and I'm overwhelmed. And how does Google Maps work? It was a gradual progression. My fear is that that's not going to happen that way with AI. And the goalposts is going to move so fast. We have no room to evolve with it. What are your thoughts about that?
Michael Kammes: I would certainly agree, because now we have AI that can iterate on itself, right? Whereas we didn't have our first garden after the guide and Garmin said, Well, I can build myself better, and then I can sell myself for more. No, so we're gonna see this happen even faster. And it's scary. And I've talked to creatives who have said a word for my job and they're like, Michael, you don't have to worry about it. No, I do too. Because part of what I do is take difficult technology and summarize it and convey it to people. You can type in any technical encyclopedia into a LLM or Chachi T and it will do the same thing. So the thing I can recommend is a You just need to stay up on it and you need to adopt it, you can't fight it, it's not going away. And the more you know about it, the more you can incorporate in your workflow. What I find really interesting, Zach, to take this to the nth degree is, at some point, we will be not too distant future, we'll be having AI, generative AI, create content you mentioned a few minutes ago, creating a five minute film, but there's gonna be so much content that we don't have time to curate it. So we're gonna rely on AI to curate the content that's being created by AI. And that's what I'm gonna find very interesting how we differentiate what is quality content versus AI, and you know what that quality content is going to be. It's going to be people like you editing and not relying on algorithmic machine learning to generate these stories. And that's why I think it's so important for people not to give up hope, and to realize that now more than ever, we need that voice, because that's what's going to stick out amongst the regurgitation of AI.
Zack Arnold: Yeah, and I want to go a little bit deeper in a second into this idea of the necessity for the understanding of human emotion and nuance. And I really believe it's our ability to bring those things to the table that will differentiate ourselves. But what I want to point out, it may seem obvious to many, but for those that is not obvious to, I need to, I want to emphasize a point that you made even further, which is it's not that AI is going to start making content decisions for us. It's been telling us what to watch for years. For years. Now, we've gotten to the point, there was a joke that I made years ago, where I said there's going to be a disorder someday that's going to be called Netflix Q anxiety. I was I wasn't too far off. Because we're now at the point where there's so much content to consume, and I experienced it myself, I go to my watch list and like, no, just tell me what to watch next. Right. And that seems convenient. But I think that from from a thought leader perspective, or from a cultural perspective, that can also be really, really dangerous. And the reason that I bring this up is in learning more about social media and how social media algorithms work. And by the way, if anybody wonders, because I'm asked this, why don't you spend much time like why don't I have a lot of content on social media, that's a whole other conversation in and of itself. But the reason it applies to this is they've actually done studies, where when you look at some of the most successful influencers, the reason they're successful is they know how to play the algorithm. They know how to say the right things, or be controversial or be crazy or funny, or they follow the trends. But then what they found is that psychologically, these people are becoming the person they had to be to feed the algorithm, and they can no longer differentiate, this is me. And this is the version of me that's playing an extreme version of myself to please the algorithm, they become the same person. And from a cultural perspective, to me, that's very, very scary. And that's a larger version of what you just said, where AI will tell us what to watch. And I'm trying to explain to people, it's already telling everybody what we're supposed to watch what we're supposed to like, where it's now matter of all, just give me my 10 choices, Netflix, oh, thank you, you're telling me the top 10 movies this week. But the part that I've found so interesting, that I don't quite understand yet, but just again, from a cultural and a thought leader perspective, is both fascinating and terrifying. It'll say the number one movie of the week is this, and it's a movie that's 15 years old. I'm like, how does that happen? And why are you telling me to watch this old movie that I'd never heard of before. So I just I wanted to go a little bit deeper into this idea that a lot of these things that we think are going to happen, have already been happening underneath the surface for years, and we don't realize how it's changing us.
Michael Kammes: And there's something else that I don't think a lot of people know, or have thought about is that a lot of us, you know, many, many people in the industry and around the world. We're not fans of censorship, right? We, we were happy to have censorship in our house, right? Because I don't want to deal with that. I want to see that. But when it comes to freedom of speech, that, you know, that's something that that we're, especially as Americans are huge into, when you're using LLM. And using Chet GPT. Those are what they call clean models. Someone has gone in there and said, Look, you're not going to talk about balm making, you're not going to use this term or that term. And we're, I don't say okay with that, but it's not something we think about because we don't see how that sausage is made. And there's a large movement out there to say, look, they called Dirty models. We want these dirty models. We want people to have access to this because it may be not peih Gee, but it's part of the creative process to use things maybe even in jest. Obviously, with proper decorum. There are things that that are your cleaning out of there that we could use for stories Telling or, you know if you remember the movie Seven, right? Remember when they there was an index of Banned Books in library, right? And they kept a record of people who check these out. And I believe Brad Pitt said something to the extent of what we're going to catch some kid who's doing a report on World War Two or something. And so now there's going to be people who are handicaps. I'm getting this information, because it's not available, because someone else said, this should not be in here. So that's another issue we have to deal with, with these large language models, because we don't get to see what made the cut. And what didn't.
Zack Arnold: I'm going to do our best to curb this without without it becoming hugely political and cultural. But I think it's, it's very challenging to talk about the threat of AI, or just how do we manage or embrace AI, or make sure that we can save our careers without at least talking about this briefly, but then I'm going to rein it back in kind of down to earth and the practical, but I think it's at least important to note that for years, if we look at the political discourse, and the de evolution of our political discourse, it's largely been because we now have algorithms that can say, this is what we want everybody to see. And this other thought idea is an idea that we don't want people to see. So it's not blatant censorship, but it's kind of sort of censorship. And it doesn't matter which aisle or the political spectrum you're on, it doesn't matter if your far right or far left, we all just know that what we're looking at on our Facebook feeds or on your Google search, or search or anything, it's not just here's the information available, and you choose what you want to consume. It is manipulated in such a way that we're seeing specific things to please the algorithm and make sure that people stay engaged. But now the difference is that that was just choosing human generated content. But now we're amplifying the exponential growth of artificially generated content that's then been artificially curated by artificial intelligence. So it just it's, again, it's this exponential growth, speed of which us understanding how this is going to work or evolve. As I've said before, it's like asking a dog to do calculus. Like, just we have no comprehension of how all of this is going to compound because it's going to move so fast.
Michael Kammes: The there's always going to be bad actors with any technology. And we can, we can, and I'll put an asterisk next to this, we can rely on larger news, quote unquote, news organizations to be to keep their editorial, I don't say indifference, but have a kind of moral high ground of, we're not going to Frank and bite this, we're not going to generate someone's voice. And there's no journalistic standards. But we've seen for years now more and more people are getting their content online where folks who generate that content or report on that content aren't keeping those same journalistic standards. And that's what worries me. And not to get off on a tangent. We right now do not have any ability to trace content from when it was acquired through distribution, right? Blockchain was supposed to be that holy grail. And then the crypto world blew that up. And solid it much like peer to peer sharing in the early aughts got shelved for a good decade, because they had that Napster Limewire kind of piracy vibe to it. So that's what worries me is that we don't have a way of tracking content to verify that this has not been altered. And I think during the next election cycle, and especially overseas, where there may not be the ability to have a political discourse, as rampant as we do, that is what really scares me.
Zack Arnold: Yeah. And the thing, I think, again, I really want to try and not make this political or existential, but it's really hard to not do that with this conversation. But the the other thing, too, and this really, this kind of dovetails us back into the creative process. But it's not just a matter of we have words that are written by AI, like you can say, chat GPT, create 10 posts a day about this topic. But it can now be a matter of like for one of the examples I brought up at that last epoch meeting, and I would guess the technology is already way better than it is now. But if you put in Joe Rogan AI podcast, I played an episode not the whole thing, but I played parts of it, where Joe Rogan interviews Dwayne The Rock Johnson, and if somebody hadn't told me that it was 100% ai generated, I would have thought, it's the two of them. It just sounds like they haven't had their coffee yet. Like it blew my mind. And this is in the nascent early stages of the technology. So one of my fears is why the hell do I even need a podcast anymore? Couldn't somebody just feed a system with all of my past podcasts and decide they want to have the Zack Arnold out Optimize Yourself podcasts. And there's a video of me talking just like this. And it knows my mannerisms. And it knows all of my ideas and my framework, like what
Michael Kammes: But see see that it doesn't know all of your ideas. That was everything I've put out there. Exactly. That's the difference. And that goes right back to what we talked about earlier, which is, if someone wants to replicate what you've already done, okay, you're still behind. Because you're the one who's coming up with new ideas. You're the one that knows how to use AI, to brainstorm, to come up with other ideas to create content, that the someone replicating or trying to emulate you can't. And I think you can't minimize how much that ability to become more than yourself in this imaginative way, and and deviate, which may not have been something you'd ever have done before in prior episodes, or prior times. That is the secret and the the folks who have kind of a fatalist doomsday. My job is over, diversify and be your authentic storytelling self. And you'll be okay.
Zack Arnold: Yeah. And the the example that I want to go back to that I used before, but I want to frame it in a different context that, by the way, at our presentation completely made up while I was there on the spot, I'm like, Oh, I'm actually going to use this so much. So that after my presentation, you know what I did? You made T shirts, I bought the 2002 version of the Thomas guide. Okay, because I, this is the verge out, because I looked around for my version that has all the posters in the notes, and it's not drunk, like no, I didn't, I'm really upset that I didn't keep it. So I actually bought another one. But just for sentimental reasons, I bought the 2002 version, because oh my god, that's how old I am. When I moved out to Los Angeles. The point is that I want to physically help people understand that if your job for a living is to manufacture Thomas guides, and as 2002, your business is about to disappear. However, if your job is to help people navigate from point A to point B, you just need to learn different systems and technology to solve the same problem. And I feel like as creatives, whether it's as editors, or writers, any kind of storytelling that we do, whether it's with visual audio mediums, if it's a matter of I provide this one service with this one technology, you're screwed, you're going to be extinct. But if it's about I want to help you achieve your vision, and I just need to embrace a different process and different technologies to make that happen. That's where I feel like you said if we diversify, we've got nothing to worry about whatsoever. Now that the AGI conversation that can be different when Skynet becomes fully self aware, but at least for now, I feel like again, if it's a matter of I was a Thomas guide salesman, now I'm going to sell Garmins. And now I'm going to sell my own version of Google Maps or whatever it is, I'm solving the same problem with different tech. So now how do we apply that general mindset, specifically to being a storyteller? Because the thing that I'm the most afraid of me, personally, is what you said, which is that you have to learn it, and you have to keep up with it. How? It's moving so fast.
And there's so many things to learn that I'm still trying to figure out. When do I want to adapt the new version of avid, because I really liked 2018. And I don't want to learn the new colors and the new windows and the new menus, and the thought of learning AI and adopting it and move so fast. I'm I'm terrified. I'm going to waste my time learning something that's going to disappear and change in three months.
Michael Kammes: I read a good article this morning, which was a somebody who's going to college. And they were saying do I even bother trying to to learn AI and ML because it's changing so rapidly, right? It's like the people who were learning punchcard computers in the 70s. Before the the evolution in the late 70s, early 80s. And as all my information gone know, the fundamentals are always there, right? Zack, you've been editing for years, whether you're cutting on a flatbed or whether you're cutting ditch digitally, the same underlying fundamental nomenclature and procedures still there. So that isn't wasted. I think if you look at teachers, right, who have had to adapt to how do I teach? And how do I have students not cheat? Well, okay, we'll do in person interviews, right? So being able to mold how you're conveying this information, and and utilize it, I think, is the most important thing.
Zack Arnold: So I agree with you, but I'm gonna play the devil's advocate for a second. Sure. And I'm actually going to blatantly steal from a quote from one of our fellow colleagues, Sven, who runs the go to editor and This Guy Edits YouTube channel and you know, he's got all kinds of amazing resources if you weren't want to learn how to edit. And I'm going to paraphrase a little bit because I don't know the quote in front of me. But he's essentially said something to the nature of AI is going to replace all of the editors. He's he's he said that he's looked at some of the technology and he's like, most I think he says along the lines of like 80% of editors are about to be replaced, which is the total opposite end of the spectrum of we're totally safe, but we know to adopt the tech technology. So what are your thoughts on I just I want to I'm not calling him out specifically. But those that are the the doomsday errs that are like, we're done like they're the it's going to figure out how to do the match edits, what's there already training AI and how to do match edits and how to choose coverage. And so what how do we approach those that are like it's over?
Michael Kammes: Well, there's a couple of things there. It's not binary. It's not our jobs are done tomorrow. It's not. It's it's a long tail as this evolves. Zack, I wouldn't consider you an actor. But I bet. If you were to watch a someone, a poor actor on television, you'd be like, Oh, that person can act. Right. So we may not know how to do it well, but we can recognize when it's bad. And right now a lot of AI is bad. Right? You mentioned the whole Joe Rogan podcast, I think someone said it sounded like him in the rock before their cup of coffee. Right? The generative AI that's being created now has an aesthetic to it. Right? You can look at it like oh, that's that's fake once you look at it for a bit. So there's, there's there's a ways to go before AI can start replacing things. Heck, if we look at some of the most basic things AI is used for like speech to text. I don't know anyone that gets a transcript back. It doesn't have to go in there because it missed someone's name, or it missed an acronym that that industry knows. Right. So there's a ways to go before it's replacing. And I think Sven is looking at it at more of a tribalism, type of field where it's content that needs to be generated without a lot of soul. I could see that being automated to some extent. But no one's that no one who's gotten into media entertainment, got into it, because they liked organizing media. No one got into this industry because I can't wait to do a 32nd montage of B roll right that people got in because they wanted to tell a story. And that, again, is what's going to stick out through this whole AI process.
Zack Arnold: Yeah, and let's go back to this Joe Rogan thing for a second. Because when it comes to this as actually generating something from scratch, this is very much related to I think that the work that editors and storytellers are going to do. And if like you said, there's this long tail and I love your analogy of I'm not an actor, but I can immediately point out a great acting versus bad acting right? And I think that for for a long time, if not forever, we would have known all right, this thing that I'm listening to this was generated by a computer. However, if you listen to that one specific episode, I was thinking to myself, if I didn't know this was AI, I don't think I would know that it was. And I feel like that is such a giant shift in such a short period of time that I'm wondering how far away are we? And again, I'm not expecting you to have the answer. We're just kind of speculating. But how far are we are we from and I'm not being facetious. I'm being totally realistic. Let's say that I am the editor of the Joe Rogan AI podcast. And I put together a version I listened to it. I'm like, You know what, in the next version, please inject 90 milligrams of caffeine into Joe Rogan. And all of a sudden, the inflections change in the energy in the voice goes up 10% Louder. Because there's a difference with me before caffeine and after caffeine. It seems to me that even though there's there's a lot of iteration that needs to happen, the speed at which it happens, I feel like we're months away from not knowing the difference, as opposed to 10 years away,
Michael Kammes: In the next 12 months, there will be the ability to localize this on your machine, meaning you don't have to go to a cloud, right? There will be the ability to do that, it will be great, but you'll be able to, there already are text to speech generators that you can say, you just get what's called the temperature, right, which is usually the amount of deviation that it's applying to and to have it enter in an or to be to have more inflection. So that's already there. What I'm looking forward to is, and I'm sure you've done this with speed ramps with video, right? You're changing at what point it goes faster, it goes slower, you'll have that point inside your timeline to be able to pitch up right at the end of a sentence and change that inflection without much issue. And you're already able to in some of these text to speech to put markers in their interface and say, happier here or louder here. So that's, that's very achievable. And it's already being done to some extent right now.
Zack Arnold: Yeah, and the there's there's two angles from which I wanted to bring this up, the first of which is just I really believe that very, very soon it's going to be almost impossible to differentiate between Is this real or is this generated by AI. But the second part that I think is much more optimistic. And the reason that I wanted to have this part of the conversation is that there's still a human being that saying, I want to change the temperature of this slightly. And I want to change this inflection, you know what, I've got this, you know, digital version of myself on the screen. But I still need to tell a human being, you know what, let's add an eyebrow raise after that sentence. Or maybe maybe the pitch should go up at the end of this sentence, rather than going down. That's still a human being. That's understanding the nuance of human interaction, that's being a storyteller. That to me is the optimistic part of this is it's very different from the tools we have now. But it's still a human interacting with a machine to tell a story with nuance correct?
Michael Kammes: It's evolution, it's evolution of the job of the Creator, to adopt new tools, and use those tools based on the experience of what they've done prior. So is it a bigger leap than going from SD to HD, HD to 4k, or black and white to color, definitely. But it's still the same concept, which is applying what you've learned, and utilizing new tools to exploit or express that.
Zack Arnold: Yeah, and this is going to be 30 seconds of my own shameless self promotion. But I know that I personally am at a place where I don't need to adopt or learn the tools as it's related specifically to editors or storytellers in the Hollywood content space, because I have now transitioned out of that as far as making a living, and I'm pretty sure we're not gonna have aI generative content on the final season of Cobra Kai. However, looking at it from the perspective of me being a content creator and a podcaster, and a writer and creating newsletters and all these other things, I have to learn different forms of AI to make sure that I'm keeping up. But the part of it that self promotion, and this is something that's a bigger picture concept I've been talking about even before AI is that our the current educational system that we have, is woefully under preparing us to manage the volume of information and responsibilities that we have in the 21st century. And if we don't prioritize our ability to learn how to prioritize and manage our time, we're gonna get so lost in spinning our wheels, just trying to keep up that we never keep up. So that that's one of the reasons that I think it's so important for people to not just say, I'm going to specialize in one specific craft, and I'm going to become an expert on Avid or on premiere on After Effects, or whatever it is, you're now at the point where you have to develop the meta skill of learning multiple skills, because the speed at which this moves is so fast. And the information overwhelmed now is so great. I mean, I have multiple students and clients now that come to me with the sole purpose of can you just tell me what to do next? There are so many things I could be doing next. I don't know what to do next. This is a skill that we're not taught. So what I want you to help me understand specific to AI is how do I start to break this down? If I now accept that AI is a part of the future, one of the things that I've written about is that it's that we essentially have two choices, it's either we're going to manage the AI, or the AI is going to manage us a world create a future without AI is not happening, we're not going to stop it from evolving. But if I want to break it down to something that's manageable, and I do want to start learning, but I don't want to massively be overwhelmed. I don't want to spin my wheels, I don't want to constantly learn the wrong apps that go out of business in three months, helped me break this down into manageable bite sized pieces.
Michael Kammes: You know, I don't like saying I don't know. But there is no universal. Let me go to this forum. And it will show me the top apps that are used for these things. There are so many Griffes right now going on. And that's where there is why I didn't talk about a lot of the the names of the software at LACPUG because there's there's 10, that are all doing the exact same thing. So I don't have an answer for you. There are several places that you can that have that are like aI clearing houses, and people can upvote this one works. This one does it. But how do you know that's not bots? Right. So the only thing that I've been able to do reliably and it's not easy, is consult multiple sources, just like how you consume news. I hope everyone out there again, not to get political, but I hope everyone out there realizes that everything is going to have a slant, try and find what what is common, the common thread between all of them. So when you're looking for text to speech, look up. When you find a product, look it up online in various places and see if other people are talking about it. That's the best thing I can offer. Or, like you and I do we go to User Groups, we talked to the community, because if there's one thing our community likes to do is geek out over tools. So being part of the community, whether it be virtual, the social network or in person, that those are Other resources for you to say, Have you heard about this tool? Or I've used this tool? And it's great.
Zack Arnold: Yeah. So I think that this idea of I'm going to consult multiple sources to get a sense of what is the current consensus on this might be the app or the tool that people seem to be learning and adopting, I'm going to bring it up a little bit higher level. Because even though I know next to nothing about AI tools, one of the most common questions that I get is I need a better to do list app, what's your favorite to do list app, and my head literally explodes? Anytime somebody asked me this, because I tell them, It's not about the app, it doesn't matter what you choose. It's about the system. It's about the meta skill of knowing how to prioritize and break things down, then you can choose whatever app you want based on the color choices, or it has a labeling feature versus a tag feature versus automations. Like, I don't care, right? So if we're gonna bring this up to one level higher, and it's a little bit more about meta skills, it's not well, I need to learn mind journey, or I need to learn jasper.ai. Like it's, I can't even bring myself to even start learning the specific tech parts of it. But it seems to me that if I'm going to start to learn meta skills, for example, if it were a matter of I wanted to become a better editor, I wouldn't start day one diving into Premiere, or Avid or anything else, I would start by better understanding how do I analyze performance? And how do I sort through and break down raw dailies, I can do that on a flatbed. I can do that with a movie Ola. I can do with Premiere, I can do it with brand new iPad version of Final Cut. But it's the meta skill of I'm looking at a bunch of raw dailies. How do I find the best moments? That's a meta skill that's technology independent? What are some of the skills that we're going to have to adopt and be good at regardless of whatever app might ever end up taking over?
Michael Kammes: Well, the term we tend to use is is a prompt jockey. Right? Which is, how can I ask, How can I take what what I hear in my head my internal monologue and put that in text in a way that the model can interpret what I'm what I'm typing, and give me back what I'm looking for. And there's a couple of ways to approach that there's, you become a better communicator, or it's people engineering, the LLM to to understand that, when someone is in this age group in this part of the country, they tend to use this phrasing for this and adjust to you. And I think that's what we're going to see more of, unfortunately, I think it should be the other way, I think we should learn to be better communicators. But I think more often than not, we're going to see more and more what they call datasets to train these models to understand poorly constructed commands. And I wish it was more optimistic than that. But I think that's we're gonna see more of,
Zack Arnold: Well, I mean, we've seen the same thing with the de-evolution of the media as well. It's not a matter of let's educate the public so they can better understand the sophistication of politics. It's let's dumb down the headlines and the topics that are getting the most engagement based on the lack of education and society. I mean, it's the same kind of backwards thinking. Again, I'm not going to turn this into a political conversation. But you brought up exactly if this had been in reverse. And you were interviewing me, I would have said the exact same thing. If there's a skill that I want to learn right now that I'm looking into, is how do I become a better prompter, because of I learned how to communicate better. And frankly, what I do now for a living is I communicate and I teach people how to communicate better. I don't believe that there's any more valuable meta skill just to survive and thrive as a human being than your ability to communicate your ideas. And what I want to learn is, how do I communicate with AI, it doesn't matter if it's Chat GPT, or all the million and a half of other AI's, it seems to me that the one constant, and this might even change, but at least for now, the one constant is I have to prompt it to do something, I don't have a bluetooth chip in my brain, where I just think about it. And it does that I have to communicate words, and my ability to communicate the proper words, the most concisely to get the best image or the best piece of music or the best spreadsheet, or whatever it is. I differentiate myself from the other editors, writers, accountants, whomever it is using AI, I now compete with them, because I'm a better prompter and a better communicator.
Michael Kammes: To augment that a little bit. For so long. The software tools that we've used in the industry have been built for people who can work in a certain way. And our software tools and craft tools really need to be reworked to accommodate folks who are neural divergent, right, who think differently. There are some people you've heard of folks that do math, you know, very fast that are ahead. And sometimes they say, well, that number has a shape or a color, right? They process those things differently. And so our AI tools that are being developed need to and they're getting there to say, well, it's the prompt the text, but also here's the image that kind of conjures up in my mind. And here's kind of the beat of what I'm thinking of, and AI being able to synthesize these different forms of input to more closely aligned with you. And that will eventually give way, if you've seen the movie her, right where the OS is, is tailored around that one person, we will have our own personal AI assistants that will start to synthesize things in the way we do. So it really is a personal assistant and not just a general purpose assistant.
Zack Arnold: Yeah, now we're getting into truly terrifying territory for sure. But I really liked this idea of there the one of the positive things or the value in this as it is going to be much more broader and more open to those that very much think differently or process information differently. And I hope that it lends to a lot more accessibility for a lot of people. But just to really triple down on the point that at least for now, the what I'm focused on is not the tools, I'm focused on the skills, what are the meta skills that I need to evolve with, so I can become a part of this new direction that we're going rather than saying, Well, I love my Rand McNally Atlas, and I'm sticking with my Thomas guide, like, I could technically I can navigate LA with this right now. Right? I could do that. So why why would I,
Michael Kammes: For those who are listening to this who are in the LA New York Feature Film Television realm, you need to have a stiff upper lip, because there are companies and facilities that make their money doing things a certain way using certain amount of gear and using certain tools and charging a certain amount for that. And this AI threatens how their business works. Just like Final Cut seven threatened the horror found that three or four, three or four threaten the industry, when that came out and undercut avid right changed everything. So there's going to be a lot of pushback, there's going to be a lot of pushback, and I don't want anyone out there to be dissuaded from it. I don't want someone people to be afraid of it. It's something that will augment how you create. And if you did get in indeed getting this industry to tell a story. This is just going to give you more ways to do it.
Zack Arnold: So now I'm going to piggyback off of that to take an alternate approach. What if I'm not a creator? What if I'm not here to be the storyteller? What if what I'm really good at and the service I provide is that I'm really good at organizing metadata. And I'm really good at organizing bids for my assistant editor. And I am a wizard thinking footage. And I love the technical aspects of my job. I'm in a different situation right now than those that are the complex problem solvers. And I know that many of those people that feel that way, are also wondering what the heck is going to happen to my job and my career path. So let's talk now not to those that are solving the complex, creative challenges and understanding the nuances of human emotion. What about those that love the technical aspect of the job? Now what?
Michael Kammes: Well, again, it's not binary, right? We've been promised for years, hey, this, this relink tool works great. And yet, we still spend hours upon hours relinking media, or we do our online conforming to something and it doesn't quite line up. So there's always going to going to be that that percentage that needs to be checked and double checked. And aside from that, you need to learn more skills. I hate to put it that way. But whether it's I'm going to take this organization methodology, and apply it to transcoding or apply it to this other part of the industry. Or maybe I'm gonna do VFX as well. In terms of prep, I'm not just an editor and all the effects editor as well. You can't hang your hat on this one thing is going to keep me through. I hate to keep saying it, but you got to diversify and learn more.
Zack Arnold: Yeah. And there was a really important conversation that that happened and really opened my eyes. And I kind of had a little bit of an aha moment. Again, when we were we were on the panel giving that presentation. It was this idea of if we're talking about the technical aspect of the job, let's use assistant editors for a second. And it's a matter of well, is AI going to replace the assistant editor because now AI is going to know will all these scenes go into this bin. And this is the way that these scenes need to be organized. I'm an auto sync them. I mean, we've already had things like Pluraleyes for years that are doing a lot of this work for us. And was that proof it was that foolproof and that not only was it not foolproof, but here here's the bigger thing is that if I'm looking at from the editor perspective, I don't want to deal with any of this crap. I want an assistant editor that knows how to deal with all this technology because I want to have nothing to do with it. I don't want to have to learn how to communicate within prompt a system that's going to organize my dailies for me. I want an assistant editor that knows how to work with the technology, they can get the prompts to organize my dailies. So to me, it doesn't replace the person that would be in that chair at all. It would just change the way that they do their job and the way that they interact with it and also the expectations that I have about how quickly would take to do something but never in a minute million years am I going to say, oh, with all of this AI, I don't need to hire an assistant. If anything, I'm thinking I want one even more. Because I don't even want to learn this crap. I want somebody that's taken the time to learn it to learn to write the prompts, so I can focus on my portion of the job. So when somebody had brought that up, that was a real eye opening moment, that the people that hire, you don't want to learn this stuff and replace you. They just want you to better manage it yourself. Yeah, I can't disagree with any of that. So what are we missing? What is the part of AI that you feel you talk about the most, that's the most important part of this conversation that we have not dug into yet?
Michael Kammes: Oh, well, that add that last part that we have not dug into, I think we've kind of beat home the point that storytelling is not going to go and that's what's going to set you apart. The what I would like to share is that everyone out there needs to be skeptical. AI is not perfect stuff that isn't AI isn't perfect. And we have to be skeptical, skeptical of not only the media that we're consuming, but also the promises of this AI will eliminate this requirement that you have. It's AI is not foolproof, it gets most of the way there at best. So just be skeptical of these claims and be skeptical of media that is showing you something that seems too far fetched.
Zack Arnold: So the first answer to this question I can already give myself and then I want you to break it down further. But if I feel totally overwhelmed that I just want to get started by learning some basic resources, I would first answer the question myself and say, well, whatever Michael commerce is creating and putting out into the world is the resource that you should be using, you are now my guru, when it comes to I want to know what the hell's going on with AI. I'm just going to search you first. But beyond that, if there are just some some basic things, if I'm coming in here as a total neophyte, and I'm a Luddite, I'm like, Alright, I understand that I'm going to have to start adopting this technology. But it's not really my thing. And I'm not super adept at it. What are some basic basic resources that are at least out there now, knowing by the time we released this in a week or two could already be out of business? But in general, how can you help me start to curate where I can at least get started?
Michael Kammes: Oddly enough, it's it's self referential, go to Open AI, get a chat GPT account and ask it. I know that sounds scares me. I know that sounds. I know that sounds completely completely dumb. But it it unlike your tech friend who fixes your network when it goes down. You can keep iterating and say I don't understand this. Explain it to me like I'm five. Explain to me, coming from this perspective. Ask it. In fact, that's kind of a power move that I share with folks all the time is, if you use chat GPT and you finally get a response that you like, or the answer your question, asked Chat GPT, what could I have asked you at the beginning, that would have gotten this answer. And it's, it's totally simple. But that will then train you. It will condition you on how to prompt better, because it's told you how to prompt itself for the answer.
Zack Arnold: Yeah, so in that instance, I say, who is training whom, in that scenario?
Michael Kammes: That's it? Well, we talked about the clean models and dirty models. So that obviously has something to play on there. But you're right, you're definitely
Zack Arnold: the one other thing that I'll point out. And I have no education about why this is whatsoever. But it's something that I've definitely learned beyond the shadow of a doubt, you can maybe talk about it a little further. But in the conversations that I've had with Chat GPT, it is very clear that it does not have access to accurate data. Because I've accent I've tried to trick it. And I've asked questions about me where I know what's true, and what isn't. For example, I asked him to write a bio about me, like, you know, write credits or bio list shows I've never even worked on, I'm like, What the hell. So I think when it comes to the the, the ability for it to, you know, compile specific information or have these conversations, that's one thing. But if you're using it as a more sophisticated extension to a Google search, you gotta be really careful about what it spits out, because it spits out a lot of bullshit right now,
Michael Kammes: One of the things and I brought this up earlier, again, not to get too technical, but a popular route is to use embeddings. Right? So you can certainly let's say, index, a manual on something, right? And then as chat GPT, who's indexed that manual will because you gave it that manual, summarize this because I don't understand it. Right. So there there is the ability to not rely on what it's indexed when someone else trained it, and instead, augment what it what it knows, quote, unquote, with your own content, so that's, that's a very inexpensive way to do things and a much more accurate way to do things.
Zack Arnold: Yeah, that that is one thing that I found to be tremendously useful where it was one of those, oh, wow, actually have more time to focus on the creative is I had several articles that were like two or 3000 words. And I said, Take this article, and I want you to break it down into three keynote slides with bullet points. 30 seconds later done. I'm like, Oh, my God, you just saved me an hour. But it was my own thoughts and ideas. So it was all accurate. It wasn't a matter of I don't believe that, or I never said that. It was taken my words, summarize into three keynote slides done, it took me five minutes rather than an hour. But when I was putting this exact same presentation together, I had asked him What are five quotes from the world's thought leaders, and that this is very relevant to what we're talking about. Now, give me five quotes from thought leaders about the value of being a generalist or a specialist. And it gave me the most brilliant, gorgeous quotes that were perfect for a quote card, then I put those into Google, none of them existed. None of those people said any of those things, and even provided me links. And I even said, you realize that none of these quotes exist. Oh, I'm sorry, you're right. Actually, these are not accurate. Try these instead, with hyperlinks, you click on the hyperlink, and it goes to a 404 error. And I'm like, what is happening?
Michael Kammes: This goes back to the that AI, no pun intended, isn't binary, it's not going to be perfect. So there is this long tail of there's the promise, but it's going to take a long time to get there. If you're interested. This is a great example of this is that the recently was a lawyer. In fact, if you go to the YouTube page, legal eagle is a great summary of this. But a lawyer filed a motion and cited different cases to back up his is clients viewpoint. And none of them were real. They cited cases that didn't exist. And the case is under roof. It's pretty hairy right now. But to be even because of that, we need to be cognizant of what information we're asking for. And as I mentioned, double check it. But that again, goes back to the point of this isn't going to fix the world tomorrow, it's going to be a long time, which means we're still employed if we're doing it the old way.
Zack Arnold: Right. So I've done my best up until this point, to make this an evergreen conversation, which frankly, with this topic is nearly impossible, because there is no such thing as evergreen when talking about AI. However, there's one last thing that I want to talk about very, very briefly, that's extremely at least for now it's time, it's very timely in it. Frankly, I don't know if this conversation is going to not be timely anytime soon. But at least as of recording this, we're in the middle of a writer's strike, and the writer's strike could potentially lead to a SAG AFTRA strike. DGA has just potentially come to an agreement, but AI is at the forefront of this for everybody. So the other part of this is not just how do we as creatives make sure that we can, you know, not be replaced by the technology or by the software? But what about when it actually comes to the studios or the corporations that are thinking, why are we dealing with all these below the line people that keep striking and keep unionizing when we can just do it ourselves with these tools. So how do we protect ourselves in that scenario,
Michael Kammes: It's going to come to a point where your likeness and datasets are private, and they can't be used without your consent, which people will then spit into jobs, right, I'm gonna make a model of myself and voice and all my likeness and then rent it out to the highest bidder. But there is as much as I don't want to see it, there's gonna have to be some legislation surrounding that, and how you how that data can be used. I think what the what the guilds are doing now with the strikes, I think, is very forward thinking and saying you can't replace us. And my hope is that if those angles are agreed to that, when it does come up next time, that that language will not only be there, but will be even more fleshed out because we'll have a better idea of what AI and generative AI can do that we just don't know today.
Zack Arnold: Yeah, and what I want to add to this, and both add to it, but also kind of take it off into a different bit of attention before we wrap up. And this is something that I actually wrote about recently in a piece that I did for the Hollywood Reporter. And I'm actually I'll link to that in the show notes. But I think that if the if the fundamental argument is you can't replace us, that's faulty because it's saying we're replaceable, as opposed to you can't replace us because there is no capability of you replacing us and here's why. So what we need to do is have a discussion about what can or can't be done. But if the premises were replaceable, but you're not allowed to replace us evolution is going to show otherwise, it's it's like saying, I'm a buggy whip operator or I'm a buggy whip creator, and I create horse drawn carriages for a living and we can't be replaced by automobiles. There's no legislation in the world that would have stopped that from happening right. So for me, I feel that what's what's really important with the line Good job this argument is it's not you're not allowed to replace us. It's we are not replaceable. And here's why that That, to me is the part of the discourse that I think we have to be very, very careful about right now. But I agree that is, as much as you know, we don't want legislation and regulation and the government to get into it. Even the biggest names for the biggest tech companies and CEOs that are creating this technology are saying, You guys really need to regulate us. You guys need to come in here right now. That's that's a big part of this discussion. And I think that, as far as the strikes are concerned, to think that this is something we tabled to see how it works out, is an absolute existential threat to what all of us do, no matter what our craft is. So having said that, going a little bit into insider baseball on behind the scenes, I wasn't gonna bring this up in the beginning. But I have to say that I'm really impressed that you and I made it through a 90 minute conversation given the amount of COVID brain fog that both you and I are dealing with right now. So we made it my friend, we survived. And I hope that we were at least somewhat coherent.
Michael Kammes: I hope so too. I hope that doesn't sound like our voices were generated by AI. I hope it sounds, it sounds like we've had coffee and not that we haven't. And I'm glad you're on the mend. I apologize for forgetting you sick.
Zack Arnold: I wasn't even going to bring that part of it.
Michael Kammes: There, I'm sure there's people who were at the same event, we were, I don't know if you want to share it. Or I can.
Zack Arnold: Well, essentially the the panel that you and I discussed talking about AI that we've referenced several times, it was also the super spreader panel where everybody walked away and got COVID, myself included yourself included a bunch of our other colleagues, which I wasn't even gonna go there. But I will admit that I've had the thought more than once, since the combination of that presentation and me having COVID was, man, it would be really awesome. Right now, if I already had this technology in place, we can just do the podcast for me. So the combination and the confluence of these two events simultaneously, like I should really look into this. So the next time I'm down, or I have a vacation, I can just keep the train rolling.
Michael Kammes: Slow, be careful, because it's it's the you know, the first their first hit will hook you.
Zack Arnold: Right? Well, at least to our knowledge, you and I are not artificial intelligence. And we are real human beings. And that would be a part two podcast to talk about, are we connected to the matrix and we're not going there now. But at least for now, I can assure you that both you and I are both caffeinated and we're both human. But for now for anybody that has either been inspired, terrified, or combination of the two by this conversation, and they want to interact with you learn more about you go to the all the various resources that you both offer now. And I think all of the amazing things you're going to be doing in the near future, what's the easiest way to find and connect with you
Michael Kammes: At Michaelkammes the majority of the socials or you can also go to Michaelkammes.com or 5thingsseries.com And the next episode because five things has been gone for a bit will be on AI. I'm in the process of finalizing that script right now.
Zack Arnold: Well, I'm looking forward to seeing that and learning more from you. And you've been a very valued friend throughout many, many years, I'm going to make sure that we link to the other past conversations that we've had in the podcast. They're completely and totally different than what we talked about today. But they're still equally valuable and relevant. And can't thank you enough for you taking the time especially given present circumstances and how you feel for still prioritizing this and chatting today. So great to see you again. Thank you so much.
Michael Kammes: Always a pleasure, Zack, thank you so much.
This episode was edited by Curtis Fritsch, and the show notes were prepared by Debby Germino and published by Glen McNiel.
The original music in the opening and closing of the show is courtesy of Joe Trapanese (who is quite possibly one of the most talented composers on the face of the planet).
Zack Arnold, ACE and workflow expert Michael Kammes discuss how to set up your remote post-production workflow and make sure you don't lose your mind working from home.
In our conversation, Srinivas and I go deep into the rabbit hole of all things AI. We examine the specific reasons AI couldn’t actually replace humans and more importantly, how we can use AI as leverage to further advance our careers.
Chase Jarvis and I discuss key strategies you can use to intentionally use your creative calling, as well as how to surround yourself with the right people that will support your creativity and goals along with it.
The advice Lesley has to share with you about how to push through any obstacle between you and your own finish line is as inspiring as her own incredible story, and I can’t wait for you to listen.
In our conversation, we discuss what creativity really is, how it works, and the ways in which anyone can learn these essential concepts and unleash their own creativity (yes, you can LEARN how to be creative).
As you’ll hear in our conversation, not only will you understand how to better write your own stories, but you’ll learn how to apply Will’s unconventional approaches to rewrite the story of your own life.
Learn how writer/director Paul Del Vecchio overcame the anxieties and fears that caused him to always say ‘Yes’ to other people’s projects (even when they were the wrong fit) so he could focus on developing his own projects.
Award-winning editor Michelle Tesoro discusses how to strategically apply the chess mentality to take the right steps in your career (instead of jumping from one gig to the next playing a game of checkers).
Michael Kammes is a production & post-production workflow master and a business developer. He has also fought a lifelong battle with depression.
