Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Ai as Podcast Producer To Cut The Work In Half
153 Plays1 year ago
Welcome to the latest episode of "Ai Microskills" with your host, Dan Sanchez. In this episode, Dan interviews Susan Diaz, an expert in leveraging AI for coaching, consulting, and time management. They delve into the world of AI and its application to enhance podcasts, discussing the potential of ChatGPT for guest sourcing, scheduling, and post-production distribution. Throughout the conversation, they explore the use of AI in processing and analyzing podcast clips, and anticipate future advancements in AI for podcasting, such as automated production and distribution. Susan shares valuable insights into her preferred tech stack for recording and editing, while Dan expresses interest in implementing AI into his podcast process. Stay tuned as they discuss the transforming landscape of AI in podcasting and the potential of custom GPTs for podcast research. This episode is packed with actionable insights for marketers looking to harness the power of AI in their podcasting journey.
Timestamps:
04:01 Seeking human connections for AI and marketing.
08:44 Injecting fun and controversial content into interviews.
10:12 Improving communication and planning for interactions.
12:56 Preferring cloud for processing larger volumes efficiently.
18:08 Comparing different tools for transcribing audio content.
19:26 Automate scheduling and posting podcast content efficiently.
23:01 Seeking efficient time management through ChatGPT.
26:26 Custom GPT generates content based on research.
30:07 Refinement and innovation in editing technology expected.
34:16 Exciting future with improved AI video analysis.
36:16 Justin Feinberg and team provide valuable automation resources.
Love these show notes? They're made by Ai with CastMagic. Find out how to automate your show notes and more here: https://aimicroskills.com/castmagic
Recommended
Transcript
Introduction to AI Micro Skills Podcast
00:00:00
Speaker
Welcome to the AI Micro Skills Podcast, where I'm curating the best lessons on AI for marketers who want to leverage the best of what's available right now.
Enhancing Podcasts with AI Tools
00:00:09
Speaker
Today, I'm talking to Susan Diaz about enhancing your podcast with AI. So it's kind of fun to kick it off on a podcast to talk about podcasts. So we're getting a little meta today. So I'm looking forward to this conversation. Susan, welcome to the show.
00:00:24
Speaker
Thank you for having me, Dan. I'm really excited to chat. Like you said there, podcasting XAI, it's like, oh my God, totally, totally here for it.
00:00:33
Speaker
So this is, I can't say it's a new topic for me, but I will say that I've only been diving into the deep end, watching everything, reading everything, trying to really decipher what the heck's going on with AI right now. And there's so much to learn. There's
Key Skills and Chat GPT in Podcasting
00:00:47
Speaker
so much to learn. So I'm excited about what you're doing with podcasting because podcasting has been something near and dear to my heart for a couple of years now, having come out of Sweetfish, still running multiple podcasts today, including this show. So let's start with the basics for people.
00:01:02
Speaker
I'm looking to find the best micro skills. What are the minimum viable things that people can be doing with what's available on AI, specifically for this first question, chat GPT? What are some of the best workflows that you were using just with chat GPT? Not an external tool, just with that one tool in order to improve your podcast.
00:01:22
Speaker
Okay, cool. So broadly, I think in order to answer that, I would think about the workflows that I use. It's also in my world for podcasts. It's three sections. The first part is like the sourcing of the guests and the, and the scheduling and that part of it. And I feel like in that section, what it helps with is sort of processing information super quickly. Like once I've say, okay, I want to interview Dan Sanchez, and then I've obviously gone and read some material to do with you and then
00:01:51
Speaker
I feed some of that into chat GPT and then it provides me with like outlines and patterns and things like way quicker. Like my podcast is 205 episodes in, we started in 2018. And if I look at how long it took us to prepare in 2018 for a podcast of good quality and how long it takes me to prepare now for an interview, it's like, I would say 25% of the time.
00:02:16
Speaker
Wow.
Maintaining Guest Relationships with AI
00:02:17
Speaker
And so that's, that preparation, it aggregates is the way I would look at it. It consolidates info, pulls in things as long as I'm feeding it the right sources. So that's the first part. And then there's the middle part. Sorry, go on. Yeah. I was going to say like, put a pin in that I'm going to swing back around. So I have a very specific questions now, cause that's really good. Give me the next two really quickly. And then I'm going to go one by one to unpack what you're actually doing there.
00:02:41
Speaker
Okay, cool. So the second part is the is the is the conversation and like sort of the interaction in the relationship. I'm not sure that it does a whole lot in that space, except for the keeping in touch piece. So I can set it up within chat GPT.
00:02:57
Speaker
that you know give me an eight step way that i'm gonna stay in touch with dan after the interview in a non sleazy way right and so it'll give me some ideas for how to keep in touch with you and then the third part of it and this is where the maximum you know like leverage exists is
00:03:14
Speaker
is what I would call the post-production or the distribution piece of that one big anchor. You've created this one big anchor piece in the form of a podcast and too often people are releasing it four or five times and then that's it, right? And so I think it's really pulling out additional insights from what's been recorded, the ability to decide how many angles, all of that kind of stuff.
00:03:40
Speaker
Yeah.
Sourcing and Preparing for Guests Using LinkedIn and AI
00:03:41
Speaker
So that's amazing. Let's we swing back around to the first one, which I think was one big one that I'm like, Ooh, I haven't done that. I haven't even thought about that. So like, let's unpack this first one of like.
00:03:52
Speaker
Yeah. Guest sourcing and scheduling. Are you, what are you doing for sourcing? Like how are you using? Initially, I think it's all human. Like I decide right now who I want to talk to. Right. And so I'm sure there's a time that will come where I can set up parameters, but that time is not here yet in the interest of practicality. So I'm going on places like LinkedIn and the communities that I'm a part of, and I'm having an idea about like
00:04:20
Speaker
Oh, I want to speak to more folks about this aspect of AI and marketing or I want to dive deeper into like GTM and revenue with AI and stuff like that. And then I will pull out a couple of names and then I will literally give it the LinkedIn profile. Like I'll copy it out. And this is just minor details because you can actually provide a LinkedIn URL and
00:04:43
Speaker
depending on the day, you may be lucky enough to have it read it. But more often than not, it's not going to fully read it. And instead, what it'll do is it'll make assumptions based on, you know, some of the details. Yeah. So I recommend just copy it out. Like most people are pretty good with describing themselves and then putting a summary. So if you copy that out and provide it to the thing, it will come back and it will pick up information, give you an outline and be like, these are topics you could be discussing with this person.
00:05:11
Speaker
What are you wanting it with? I mean, I would, I would say this is, I do have a thread. I have a paid version and because of a paid version, I have very complex custom instructions in there as well. So custom instructions for those who don't know. You made a custom GPT for it.
00:05:27
Speaker
Not no,
Effective Prompting and Outputs with Chat GPT
00:05:28
Speaker
it's just my regular GPT. I haven't yet started playing with as much building as I should. But what I have done is within your settings in a paid account, you have a section called custom instructions.
00:05:41
Speaker
And in there, most people would be tempted to like provide it with like a resume style thing and be like, this is who I am. And this is blah, blah, blah. And sure, you want to do that. But, uh, I would say go deeper, like, you know, give it information. Like don't make analysis, don't make assumptions. Give me analysis is one of the things that I've told my, um, my GPT. And I've said to it, if I give you something to edit, don't make the edits. Tell me what edits you're going to be making.
00:06:10
Speaker
And then don't speak in second person or whatever, speak in first person. There are some of those instructions which I put in. And so on the basis of that, it will already know who I am. It knows in that thread that I'm working on the podcast. And so I don't have to keep explaining to it over and over again.
00:06:34
Speaker
what I'm doing. So I will give it these pieces and say it's for the Forever Report podcast. I want to be speaking with Dan. Dan is a big proponent of audience building and he does it by doing ABC. You have to get deeply granular. People think you can give it one line prompts. You can't. You have to give it texture.
00:06:53
Speaker
Yeah, you can. It's just, it gets so much better once you give it more specific feedback, as it is with all delegation for anything. Even if
Domain Knowledge and AI Tool Use
00:07:02
Speaker
you just have an intern from college and they need to do something, the more granular you can get in your instruction, the better the output you're usually going to get.
00:07:10
Speaker
Right. So works the same. In fact, I, I just talked about this in the last interview I did on the show, but like, I feel like assigning things to chat GPT to do is a lot like giving a very genius college student who's maybe somewhat like doesn't lack some common sense. So think of like a genius freshman intern.
00:07:32
Speaker
who lacks common sense at chat GPT. It'll wander off. It's really smart, knows a lot, but dang, it can miscalculate really critical things. Which is why domain expertise is non-negotiable in my opinion, right? I mean, I think if you start to use chat GPT for things that are outside of your wheelhouse,
00:07:55
Speaker
it will get hard for you to spot where it's like rubbish and where it's really intelligent, right? Because you can have the concepts, right? And get the execution deeply wrong, whether it's in marketing or any other field, for that matter. So your point about the genius student is definitely one that I talk about a lot as well. I'm like, it's a leadership thing. Just like with all other things, you're going to have to teach it.
00:08:24
Speaker
You got to be a master delegator. So we've talked about how you're using it to do research on guests ahead of time. What else are you, once you kind of get the initial thing, like give me some talking points for so-and-so who I want to have on my show. Are there anything else you ask it? Like what kind of information are you trying to get out of that research? What are they, the outputs of that? Some of it is like trying to be, um,
00:08:48
Speaker
trying to get some, what shall I say, like controversial stuff out of there as well, right? Others even have rather a bland conversation. So sometimes I'll feed it a post from the person and be like, this person posted this, this thing that was really funny, because like, I've written a book called Unboring, Take Your Content Marketing from Blah to Brilliant. And so I like to focus on, on like the fun angles and the sort of edutainment stuff. So I'll feed it some of that.
00:09:14
Speaker
Sometimes I'll ask about industry things and say, give me some stuff in this industry that are important that I would want to discuss. Give me
Strategic Communication with Guests Using AI
00:09:26
Speaker
some perspectives, especially if it's going out of my comfort zone. One of my two podcasts is ABCDEI, where we talk about
00:09:34
Speaker
inclusion and diversity. And obviously, in the process of interviewing some of the folks there, I'm deeply out of my comfort zone, like I'm deeply out of my comfort zone. And in those spaces, I feel like just a little bit better if I prepare. And so that's one of the things that I do.
00:09:52
Speaker
It's huge. You said you're using it to create plans to keep in touch with guests, which is another one that I'm like, huh, I actually do this a lot. And I do it a lot for other companies. So tell me, what are you prompting it with? What kind of things are you looking for? And what are some of the best pieces of advice it's given you so far for just staying in touch with guests? So I think one of the things that it's helped me do is just, and I will go into this over and over again a few times, it's helped me get out of my own head.
00:10:22
Speaker
And so if I say to it, like, I would like to be in touch with this person, like, I don't know, eight times over the next six months. And I would like to share with them some of the information that might be of interest to them. And I would like to whatever, like, it gives me a bit of an outline. It's like, you know, touch one. We know this in a manual way. But it's interesting to see it, like, laid out for you in a way that you can't avoid, you know, like,
00:10:46
Speaker
For me, for example, if someone tells me these eight actions need to be done, I'm a little bit of a front venture. And if someone says it needs to be done, it's going to get done. So like that's what it does. It takes me out of my head. And so when it's some of the interesting things is like how short it thinks.
00:11:04
Speaker
conversations should be. And I know from a sales and revenue perspective, this is something people have preached forever. Like, keep it super short. Like, there's no need for long things. So some of the drafts, like, if I was like, Oh, hey, I want to send Dan this article on how, how this company used AI to grow by, you know, 100,000. What it thinks I should add to that, that note is very, very small. It's like, Hey, Dan, this made me think of you. It's awesome because and that's it.
00:11:33
Speaker
If I wrote that, I would have written like 400 words on why I thought it was awesome. I would have overthought it. I would have been like, I wonder whether this is the right piece to send, blah, blah. So it really does get
AI Tools for Post-production and Content Enrichment
00:11:46
Speaker
me out of my head. Interesting. So moving on to post-production. I do a lot here.
00:11:55
Speaker
We're, what are you doing? Let's, I'm going to go onto automation. So there's probably some things you're doing to automate post-production. What are you doing in post-production regards to just manual, like working with chat GPT? The creation, like, so if you feed the transcript, first of all, the transcript is like a source of gold, right? From the podcast. Yep. Yep. And within once you feed it the transcript, which you now can, it's only been in the last few months that you've been able to do this.
00:12:24
Speaker
I will get it to give me sorry. Sorry. Like I know chapter two used to have a limit as far as to how many characters you could upload to it. So they've extended that now. You can do a 40 minute transcript.
00:12:36
Speaker
You can. Oh, that's awesome. Cause I started using magic maybe over the summer because of that limit screwed me up because obviously you can't, you wouldn't, you just couldn't do a 40 minute transcript was too long and it'd be like, ah, too long, too much. So that's cool. Now you can, now you can on the paid version with the chat GPT four on the paid version. And I think one of the reasons that they brought that in was because, you know, cloud from anthropic.
00:13:02
Speaker
the other one, which by the way, I'm a huge fan of as well. And I will interchange from one to the next, depending on what's on my plate. And yeah, Cloud was doing that all along. It was processing a much larger volume. And so I think chat TPT has had to catch up. So in terms of manually, I will feed the transcript in, and I will ask it to give me outlines for show notes,
00:13:31
Speaker
you can ask it to, like from a creation perspective, you can ask you to come up with a few angles for articles from it. Because I'm not a fan of, I'm not a fan of just being like, Oh, this podcast happened. And so the article is exactly what the podcast is. That's kind of like, you know, just the start, you need to, yeah, you need to get a little deeper. Yeah. So you're asking it for angles.
00:13:55
Speaker
I'm asking you for
Automation and Video Content Creation with AI Tools
00:13:56
Speaker
angles based on the transcript. So I would, for example, be like, okay, in this conversation, we talked about what AI is usable right now and what use cases right now. So tell me three or four perspectives from which I can write it. And then maybe it'll come in with a few things and then I'd say to it, I don't love this. How about we talk about it from this angle? Like yesterday, I was working on one thing.
00:14:20
Speaker
which was to do with hiring and how to like, as entrepreneurs, like what to look for and resume resumes and how, how do we like know if a resume is AI generated and that kind of thing. And so after going back and forth with it a few times, I've come up with how to hire your next CMO.
00:14:39
Speaker
using AI. And so that angle would not have come directly from my head alone and it would not have come directly from chat GPT alone, right? And so that mashup is where we came to the point from. Nice. So moving on from chat GPT over into the automation side, what AI tools are you using to kind of automate the process? So you're not having to manually prompt chat GPT, but every time X happens, you get a Y output now. If this, then that, yeah.
00:15:09
Speaker
So primarily in this space, Zapier has been our friend for the longest time. And we've used Zapier even before chat GPT became such a phenomenon with a couple of other tools. Like there was a Descript, there is a Descript and a few other things.
00:15:26
Speaker
So I would say the most simplified version of it, and we don't always automate, I have to say, because in the interest of being what's, what's happening now, the simplest version of it is like, as soon as a podcast is recorded, it's uploaded into like a drive. And then there's some sort of, um, zap set up in a way that it will go to Otter. Otter.ai is where we, um,
00:15:51
Speaker
we transcribe. And then the next piece is that that author thing will be emailed to a couple of addresses, like the writer and one more person. And then from that email, the manual step comes in where you will have to put it into chat GPT. You can now there is a plugin from Zapier, you can directly put it into chat GPT. But honestly, I don't always use chat GPT for
00:16:18
Speaker
for this. What is another good tool is a tool called Decipher. It's not spelled with the E at the end. Decipher is a good one because it understands podcasting as an industry. It will give you a more solid set of show notes. It understands time stamping. It understands an intro and a guest thing and
00:16:44
Speaker
like call to actions and such like. So that's one. Another one is CapShow, which these are specific to podcasts. And so I think they do a better job of specific outputs like that. When you want articles, you still want to go back to chat GPT or cloud. But yeah, so those are broadly it. And then the other part other piece is the video. You've got the ability to
00:17:09
Speaker
take the link, and again, this is manual, this is not automated. You take the link and you put it into a tool called Opus Clip. It's opus.pro. And I think you've probably seen some videos in the short form space out there. It sort of mimics that Alex Hormozi style of like vertical video with like chunky, you know, captions on the bottom. So yeah, Opus Pro is what I've been using for that. So you can feed it directly in over there. And then it spits out a
00:17:39
Speaker
a bunch of clips. Let me tell you that none of these are perfect. So you do have to factor in like a bit of time to go over all of it and then becomes the question, who's going to do it? Is it the host who's going to do it? Is there like somebody on your team whose job it is to do it? Is this something that regardless of what it is, it's still something you're delegating to an agency with the understanding that the agency has these tools. I think that's what we're unpacking right now.
00:18:09
Speaker
Yep. I, uh, it's funny, these, you have a different tool set than me and I'm, I'm like need to dig into decipher a little bit and cap show because I'm, I'm familiar with those. I used Opus over the summer, but these tools are changing so fast that I probably need to check it out again.
00:18:23
Speaker
Right now I'm using Cast Magic, which I think is similar to Decipher. It essentially takes, you upload the audio or the video, it transcribes it, creates an AI transcript, and then goes through and essentially has pre-built prompts in it to get all the stuff. And you can even make your own custom prompts. So every time you upload it, it prompts it with the prompts that you want, and then you just get the content.
00:18:44
Speaker
I'm hoping they automate it more. Cause I'm like, Hey guys, wouldn't it be cool if I could just feed it the RSS feed and then it runs it through cast magic and like magic just swaps out the show notes with the right, with your show notes rather than whatever I put in there. And then it fed it out to all the players. I talked to them about it. Hopefully they do it. Cause then you wouldn't have to do anything. You would literally, I mean, it'd be automated and you wouldn't be able to check it before it went live, but I still think it'd be better. Okay.
00:19:13
Speaker
I would be okay with popping back in and making some changes in retrospect. That's my level of comfort with automation, because at least it's getting done. Done is better than perfect.
Comparing AI Tools for Podcasting Efficiency
00:19:26
Speaker
Imagine scheduling it and then within like five minutes of it being scheduled that then hits something like cast magic. Cast magic injects the written part, probably still lets it be host wherever your host is, but then injects all the written content and then maybe shoots it out with all the best stuff. And so you have amazing show notes and you didn't even have to think about it every single time. I'm like.
00:19:47
Speaker
Because we all know like with podcasting, you got to schedule it. But then there's like this in between of like you scheduling it right. And then you, after it's posted, you got to make sure it's posted on your website and then posted all the social. There's still a lot of steps manually that you have to do with each show. And it's getting closer and closer to being able to automate it. Zen classter, this tool we're recording on right now has its own clip tool that I love now, because since the recordings are here natively in Zen caster, I could just be like port to clips and the clips will be done in 30 minutes.
00:20:17
Speaker
But Opus is pretty fast and similar. Riverside is doing that as well, I think. A couple of our clients are on Riverside, but honestly, I continue to use Zoom. Zoom's improved as well a whole lot, purely because it doesn't glitch nearly as much as the others, and that becomes a big factor, especially if you're talking to a lot of strangers, right? Yeah, yeah. If you have a lot of people on the show, too, it's a major thing.
00:20:45
Speaker
So I've already, are there any other AI tools that you're using for podcasting specifically? No, I think I've basically spilled my guts on all the ones, Dan, that's it. Those ones, I think, uh, the script is definitely one that I would say continues to be, um, quite cool. The script is, uh, where you can edit with the, with the text, right? And so it is a little fussy, especially if you want to change the format, like if you have a horizontal.
00:21:12
Speaker
recording and you want it in, in, you know, taller form, it gets a little fussy, but I think they picked up their game quite a bit. I would say if I had to, if I had to pick my stack, it would go like this. I would use zoom to record because it's got AI capabilities. I would have something like auto running in the background to make sure that I have all of the transcribe. And then I would have, uh, it, I mean, Zapier would somehow get it off to, um, to chat GPT or cloud.
00:21:42
Speaker
And then I would have Opus Pro to do the video because it's honestly one of the simplest ones I've found. And then this from someone who has a live actual human editor on the team. But still, I think there's a quality aspect that I can now be like sifted out and be like, this is basic. I'll use AI. This is much more detailed and creative. And so I use the producer. And so there's that. And then for stuff like scheduling and things, I would stick with Buffer.
00:22:13
Speaker
Yep. Yep. That makes sense. Are you using AI at all to help you code like as using AI as a coach or consultant or to help teach you things regarding podcasting? Oh God. Yes. Yeah. Heck yeah. I think not just podcasting a few things really, like I have a business coach thread going.
00:22:34
Speaker
So within that business coach thread, I've asked a lot of questions in terms of, you know, the kind of brainstorming that I would have done. I mean, it's not to say that it's going to dismiss the need for a coach for sure. Definitely not. But like anyone who's had high level coaching. Yeah. Anyone who's had the high level coaching knows that you have specific times and then in between talks come up and stuff. So I definitely go back
AI in Productivity and Creativity Enhancement
00:22:55
Speaker
and forth with it, asking it questions about specific models, like streamlining things. I even ask it what I should be asking.
00:23:04
Speaker
it. Like, I'm like, I want you to do this for me. So tell me what I, you know, tell me what you need from me. So that question and answer thing. I have one section, which is sort of like an executive, like an executive assistant of sorts, where I try to help myself manage my time a little bit better.
00:23:23
Speaker
So I've been one of those people who called myself, I'm a creative and I will not do time blocking. And even after so many years of working, time blocking was a challenge for me. But this past year, I got that down. Now I have time blocked. And I would say that's chat GPT in a large part. I'm like,
00:23:40
Speaker
how do I break up my day, given that I need breaks in between, given that I cannot focus for more than this time, you know, telling you things like ADHD and whatever, it's incredible that like those kind of things. I feel like it also takes, it takes away some of your what shall I say, again, back to that overthinking piece, right? It takes away some of your maybe shame or vulnerability as well, because you're not talking to another human being, you're talking to a machine. And so it's easier to
00:24:08
Speaker
you know, type out what you're thinking. So I definitely use it for that. I create small learning plans because, you know, I think it's important innovation. I mean, I read, I read audio books more, more than, you know, visually, but I've asked it to give me specifics on like how to organize my reading. I've asked it to give me, um, summaries of things, you know,
00:24:30
Speaker
So yeah, definitely using it as a teacher, as a coach, as a consultant, not to replace the ones that are a part of my network instead to supplement it.
00:24:41
Speaker
It's interesting. I find we have a limited amount of decisions, right? Or they call it decision fatigue. And in using, outsourcing some of those decisions to AI lets you spend more time thinking about the decisions that are more important, right? So like time blocking, you give it all the factors and maybe you give it the factors once and then the time blocks for you and you can have it time block over and over again.
00:25:05
Speaker
You've put all the decisions in once and now you've essentially automated decision making for the future so that you can free up some of those decisions because we only have so much decision making ability within a day before we need a reset and start over. So I mean, that's the same reason why people have even decided sometimes to do uniforming, right? They just wear the same thing every day. One, one less decision. I wear the same black t-shirt for like copied it from all the other tech bros, but I wear it all the time.
00:25:31
Speaker
Yeah, I guess I'm one of those tech bros because I pretty much wear the same black v-neck almost every day. Because I'm just like, I just want one. I'm one of those people. Less decisions, right? So have you seen anywhere in AI that can use you can use AI to create original value for your podcast? Have you ever tried to let AI take over a show? Have you ever let AI create content for any of your content or by itself? Not just repurposing?
Future of AI in Podcasting and Job Sectors
00:25:58
Speaker
Okay. Not yet. I'm not there yet. I haven't seen anybody doing it yet. Yeah. I think I want to eventually, but it would depend on what and how and yeah, definitely on my plan for the next few months. Yep. And as AI gets better, I'm sure it'll be easier and easier for AI to run whole parts of the show or maybe even segments of the show would be kind of interesting, right? Yeah, I'd love to try. Maybe we can join up and do something cool later with some folks. We'll try. Yeah.
00:26:27
Speaker
Like I'm already seeing it based on even what you've told me, what you're doing with the guest research. I'm like, Oh, if you made a custom GPT that you fed instructions and some background information to and uploaded some docs, and then you could, after doing the research phase, so you understand the guest and have some questions outlined and all that, there's probably a whole nother step of instructions. You can instruct it based on all that information. It's like, Hey, execute this for this little five minute segment.
00:26:53
Speaker
right? That's relevant to this so that those two match up. It might be like, I don't know, like three tips on something related to the subject that we just researched, right? And that would be kind of an interesting little segment that it's generating the content for and then you're just speaking up as a host. I'm like, there's probably a bunch of little things that you can add as segments to a show currently with what's available right now.
00:27:17
Speaker
It's true. But it would be possible because of the research you did ahead of time. Because if you just asked it right now with the prompt, it probably wouldn't be that good. But if it was based on all the work you'd done previously in a thread, I'm like, ah.
00:27:31
Speaker
I think so. Yeah, maybe you're right. Maybe it's closer than I think. And I'm taking some time off over the holidays. And after that, I'll come back and give it a shot. Yeah. Try the custom GPTs. It's way easier than I thought. Just do a YouTube search for how to build a custom GPT, watch somebody do it once, and then just start fiddling with it. It's way more intuitive than I thought. And once you start looking at it, the applications will start unlocking your mind for what you could be doing and how it even automates what you're currently doing with just the standard GPT.
00:28:01
Speaker
You'll start offloading your instructions that you're putting in your master GPT instructions. And you're going to start putting them into GPTs into separate ones, right? I don't think like what you're doing, adding instructions for the whole GPT. Like as soon as you said it, I was like, like, I think what you'll find is you'll start having wanting to have specific instructions for different GPTs. And I think that's why they did that. So they can all be able to have different bends, you know?
00:28:27
Speaker
Correct. Because as you scale it, you start to think of more users, right? And then you have to split it up. Yeah, you're right. That's it. Because like, so it's personal uses, I use it to write store bedtime stories for my kids, and I don't want it to be instructed on my business stuff. I'm saying it's, it's useful for so many things. Where do you see this all going over the next year, max two years? Like, what do you see coming in the near future that you're excited about?
00:28:54
Speaker
I think it will become non-negotiable. I think it will become like background required reading for everyone, much like when, you know, calculators came about or computers came about or any of those kinds of things that will just be running in the background. I think in the short term, we will all overdo it. I think everyone's going to jump far too far into it. I think there's going to be, you know, people talking about it's not going to take jobs. It's not true. It is already taking jobs.
00:29:23
Speaker
And I feel like there will be a correction on that. Like after deciding, hey, we can do without writers and we can do without Bobo and we can do without whatever, people will realize that it has to go back to a place where it's man plus machine or woman plus machine and not so much one or the other. So I think in the next one year, a lot of people will learn how to use it and then eventually settle down.
00:29:53
Speaker
Yeah, I think that's definitely a possibility where it can get overhyped and then come back. Like you said, I think, what do you think is coming around for like podcasting and AI? I think there will be refinement on editing because that's the piece that I'm like, we've talked for a while now and we haven't really talked about any AI that's able to fully cut up your recordings. There are tools. They may not be as low cost as some of the other ones that we've mentioned.
00:30:24
Speaker
But I think that will come where you're able to take just a raw recording and clip it into different things. I think the video part will get more and more refined. More tools like Decipher and stuff will probably start coming up because the temptation with new tech is that everyone wants to build a better piece of new tech. And I think the real gains are for the people who use the tech in different ways.
00:30:54
Speaker
I think you'll find a lot of players building new tech. Yep. My, my thinking, this is what I'm trying to push these companies to do. I literally call up the founders and I'm like, Hey, wouldn't it be cool if so this is, this is my prediction of what's coming for podcasting and AI. And I predict that it'll be within the next six months.
00:31:12
Speaker
I think it'll take six months for them to finish this, but it's so close. Like Zencast are what I'm using right now to record this. I'm like, Oh, it's right there. But imagine like after having this initial conversation, I could just push play maybe and put in a title. And like I was talking about with cast magic, everything's handled down the line. Like right now on your way in, it captured your name. Cause I could see your name even on your video screen, kind of like zoom does. And then it displays your name.
00:31:36
Speaker
If it also captured your email, what then is possible? Imagine if I just took the show and said, hey, produce it this way, according to the template I've already set up, produce.
00:31:49
Speaker
Everything else happened automatically. So what's going to happen when I scheduled it? So on the day it goes live You get an email an email what to the links to the full podcast to all the places where it's posted To all your clips that are automatically already made For you to post and link up to your own social networks It's already starting to trickle out on all the social networks I have associated with it and I haven't died to do anything. It's just happening It all has custom captions. It all has custom text
00:32:17
Speaker
customized for the platform. So tick tock is a little bit different from Instagram. Right. And I didn't have to do any of it. It all happened based on the start date that I said go live. Right. And the guests follow up is somewhat automated, though, of course, like there's probably some personalization you do with that. But at least at least the one that's like, Hey, your show went live, here's all the stuff, celebratory confetti and gifts and all that kind of stuff that email, and maybe the email afterwards with some stats, that can all be automated with AI.
00:32:47
Speaker
All of this, I mean, some of this doesn't even take AI to do. They just haven't degraded it in. But I think within the next few months, you'll see full platform automation from recording to pushing publish. And then you don't even have to think about it. It's on your website. It's on the social platforms. It's on across all the Apple podcast places within the next six months. At least one platform will do all of that.
00:33:09
Speaker
I cannot wait. Honestly, what will happen though is that with the exception of people like you and I and a bunch of other folks inside the bubble, a lot of people will be uncomfortable with that. I don't think people will be comfortable with that thing of pushing it out. It's probably because everyone likes to reserve that idea of, oh, it can be edited in the future. And I'm like, when's the last time? I haven't edited a podcast that we've done in the longest time, unless someone actively says, can you please take this out?
00:33:38
Speaker
Yeah, so why bother and there might be a place where like like right now I still review the clips because AI currently it does a pretty good job at like rank stacks clips According to what it thinks is best. That's not bad, but I usually will Listen, I'm less and less am I having to go and clean it up?
00:33:57
Speaker
Like it's actually getting much better at finding the sound bite from beginning of thought to end of thought. I rarely have to clean that up now or less and less. So most of the time I'm like approved, kill, approved, approved, kill, kill, approved, approved, done. Maybe clean. I probably edit one clip out of a batch now.
00:34:14
Speaker
It's getting good. Yeah. So it is getting that's exciting. Like I might still do that in the future, but even in the future, it's probably going to get better and better. I mean, imagine using AI on one show and it's getting better and better at like processing the clips and seeing the performance of those clips. Like it's going to be, it's just going to get better and analyzing what clips you want. So you're not like pretty soon you'll just be able to put it on full autopilot.
00:34:37
Speaker
It may be a year or two from now, it can actually even analyze how well those are producing. So it gets better at clipping for you and starts getting used to it. Wouldn't
Resources for Mastering AI in Podcasting
00:34:47
Speaker
that be awesome? Wouldn't it be awesome if it comes back to the front and says, based on the last blah, blah, blah, you should be doing this over here. And then that's it. We can all just fire ourselves and build the GPTs.
00:34:59
Speaker
Yeah, I think that's probably two years out, my prediction, but I think within the next six months, just full process automation because AI now, I was big into marketing automation before and there was many, there was a lot of things you couldn't do because of just the extent that you could have automated AI fills in so many gaps now that we're going to see a lot of breakthroughs in all the tech companies with AI filling in all the missing gaps that automation itself couldn't hit. So I'm excited for that.
00:35:25
Speaker
So to wrap this up, what are some of your favorite resources, either people you're learning from, blogs, articles, videos, things that have helped you in your AI journey?
00:35:38
Speaker
Not intuitively, TikTok has been the source from which I've learned a whole lot. And probably the reason for it is because I jumped right into it. Like I've used a bunch of tools before ChatTPT came around like Jasper and copy.ai and things like that. But I think ChatTPT was like so far ahead of the others that I was looking for creators. They weren't a lot. And then on TikTok, I found some. One, I'll give you a couple of names because they continue to create content. One is
00:36:08
Speaker
Rachel Woods. Another one is Natalie Chopra-Sert, and she calls herself Brand Nat on TikTok. And there's Justin Feinberg. And all of these folks, of course, they do their part about, you know, what's new and what's coming to the market and, you know, the new shiny tool, but they also go deeper into the automation aspects and
00:36:30
Speaker
and give you use cases of how they're putting it into action for them and their clients so I think those are excellent resources. A definite plug for myself and the fact that we do the 4am report podcast as well as the Unboring AI shift which is a weekly newsletter.
00:36:47
Speaker
I learn by writing. And so writing that newsletter, we've been doing it for seven months now on a weekly basis. That in itself has created some mastery in my mind. You know what I'm saying? Learn by doing. Tell me more. Where can people find it? So I will give you a link. It's CP.digital is our website and it's right there. It's called AI slash shift.
00:37:14
Speaker
And yeah, subscribe to it. It comes out on LinkedIn a couple of weeks later, but the fresh one that I do is on the website. So cp.digital. Yes, that's our website. Yeah.
00:37:31
Speaker
Well, fantastic. Susan, this has been a great conversation. I've learned a ton. I'm looking forward to actually implementing a lot of my process, especially on the pre-pre-interview side. I'm like, I can't wait to set up some of these and build a custom GPT for it. I'm already thinking about how I'm going to build this GPT now. And maybe later we'll circle back up. If we both experiment with it and do another episode on what custom GPTs can do for research for your podcast before interviews, it'd be fun. Love it.
00:38:02
Speaker
It's been awesome. Thank you for having me. I can talk for hours about this and being able to chat with power users of podcasting like you is making my day. Fantastic. Thanks again for joining me.