Ron Martinez – Understanding Artificial Intelligence (AI), synthetic voices, NFTs, and other new tech for writers

Nonfiction Authors Association Podcast | June 7, 2023

“I think that’s key for all of this. View it as a collaboration tool. A thing that you collaborate with. Good writing is rewriting. It always has been the case. So whatever comes out, make sure that you’re able to revise it and tell it what you want it to do and make it your own.”
-Ron Martinez

About Ron Martinez

Ron Martinez is an inventor and accomplished intellectual property professional, entrepreneur, designer, writer, and developer whose latest project is an AI writing tool at InventionArts.io.

As the founder of Headcanon and Postlinear Entertainment, he’s engaged in the continuing reinvention of storytelling through voice, video, and conversational user interfaces, interactivity and narrative simulation, and new commerce models. As a pioneer in non-linear narrative, Ron continuously explores the potential for reimagining how stories are experienced, as both a designer and developer.

At Invention Arts, Ron’s interest in evolving interactive publications evolved into the Aerio Retail Network, a distributed content commerce marketplace delivering “retail-as-a-service” to suppliers, sellers, and consumers. Aerio, and BookShop.org, launched in 2016 and now powers tens of millions of dollars in sales, putting revenue that once went to centralized platforms into the hands of indie booksellers and authors.

Prior to Invention Arts, Ron served for four years as global Vice President, Intellectual Property Innovation for Yahoo!, where he drove and managed the development of hundreds of patent applications and resulting product innovations. There, he was also responsible for Content IP Strategy and Operations. He currently holds more than sixty issued patents covering a range of increasingly interconnected fields that includes media, mobile, social, advertising, and commerce.

Over the course of his career, Ron has written, designed, or coded branching pathway books (Be An Interplanetary Spy paperbacks), interactive fiction (as one of its pioneering practitioners) including for marquee properties like Star Trek, narrative simulations (e.g., the acclaimed Hidden Agenda), graphic adventures (e.g. Arthur C. Clarke’s Rendezvous With Rama), realtime massively multiplayer games (the groundbreaking 10Six, for Sega’s Heat Network), commerce systems (the pioneering virtual property system with in-game purchasing), and educational software (for Scholastic Springboard), multiplayer live-action role playing games (Murder Party for EA), and others.

Ron is an advisor to Licium.org, a decentralized identifier project that implements a distributed IP licensing model, highly applicable to new rights formulations required by Web3 media. He also developed Transium, a platform for everyday tokenized commerce with redeemable attachments, an easy onboarding and purchase experience to sell physical and digital goods, real world event ticketing, early access product purchases, and more.

Nonfiction Authors Podcast: Ron Martinez

Find the video podcast, show notes, links, quotes, and podcast transcript below.

Listen and subscribe to our podcast wherever you listen to podcasts. Watch the video interview on Facebook, LinkedIn, Twitter, or on our YouTube Channel where you can subscribe to our playlist.

Got feedback on our podcast? Want us to consider a guest or topic? Please let us know on our feedback form here.

Show Notes

Links

In this episode…

  • More about AI and how it has evolved in recent months.
  • An explanation of AI personas and how to guide its behaviors.
  • AI terminology authors should know about.
  • The major AI platforms and what they do.
  • More about plugin capabilities on AI platforms.
  • Tips on prompting AI.
  • What authors should know about cryptocurrency and NFTs.

Transcript

Hi everybody. It’s Carla King, host of the Nonfiction Authors podcast, and this week we’re speaking with Ron Martinez to provide an author’s guide to new tech, artificial intelligence, synthetic voices, and NFTs.

But first, this podcast is brought to you by the NonfictionAuthorsAssociation.com, a supportive network where writers and experts connect, exchange ideas, and find resources that will help you write, publish, market, promote and profit from your nonfiction books.

Join our email list to explore membership options and to get free reports, webinars, and the resources you’ll need during your entire publishing journey. You’ll also find complete transcripts and show notes for this and every episode of the podcast on the Nonfiction Authors Association website.

Now let me introduce our guest, Ron Martinez, an inventor, a designer, and a developer with more than 60 US patents, and he heads up a design and development studio called Headcanon.com.

He’s also a pioneering narrative game and app designer, storyteller and producer, having shipped pathway books, interactive fiction, narrative simulations, massively multiplayer online games. He’s a former Vice President of Intellectual Property Innovation for Yahoo!, and creator of book retailing platform, aer.io, which was acquired by Ingram and became the engine behind bookshop.org, which we know and love and became so popular during the pandemic.

And I just want to say that Ron has been my go-to person to discuss writing and publishing technologies and development since I launched Self-Publishing Boot Camp in 2010.

You’ve just developed this fascinating new AI writing tool that I got a peek at that helps people “create and talk to AI powered personas to write, publish, and enjoy fiction and nonfiction. Learn or share what you know, invent or just to have fun.”

So as you’re deeply involved in that right now, I thought we could start by why you’re so enthusiastic about AI writing tools and AI from all sides as a developer and a creator. And about the rapidity of developments you expect us to need to understand, and use, and sort out in 2023 and 2024, that will affect us in particular.

[00:02:25] Ron Martinez: Thank you. Yeah, let me just say, rapidity first, I have never seen a set of technologies evolve so rapidly. I’ve been doing this for quite a while. And as you mentioned, I tend to focus and specialize on emerging technologies and how they can be used for expressive purposes.

And so the AI–the generative AI–which you can feed it prompts, you talk to it, tell it what you want, and it will attempt to return text that is configured the way you want it to, if you know how to do that. There’s imagery, there’s music, voices. A set of capabilities over the last five years have coalesced into what you now see as an explosion of augmentation for creators, and really for everyone. For business people, students, lifelong learners, whatever.

I can just mention a bit about how these work. And it’s not without its risks, and it’s not without its downsides.Some people think there’s a 5% chance that it will destroy humanity. I’m not among those. In many ways, I don’t like the term ‘artificial intelligence’. I like the term ‘ersatz intelligence.’ It appears to be intelligent in the same way. Ersatz leather is not really leather, right?

It’s trained on cultural output that has already been put out into the world. And what that means is – let’s take ChatGPT that generates words, which is of interest to us as writers. What it’s not doing is finding words on the internet and then just plagiarizing them and quoting them. What it’s designed to do is–based on a prompt, return what it thinks is the most likely set of responses to your prompt. And it generates that by looking at patterns over millions of pieces of text.

So it can be prone to cliches because of that. Unless you manage it away from cliches. I think it was Martin Amis who said something about ‘rummaging in your purse.’ This is a cliche. It might say things like that.

So what it does is go through and generate–and it’s remarkable because it can generate sustained context. In other words–I can start with a persona, which is what we do. And what is good to do, no matter which platform you’re using. Say, you will act as an expert in agenting for nonfiction authors. And I’m going to ask you questions as an aspiring author–nonfiction author–and you’ll guide me with information so that I can learn the basics and so on.

You have now told it how to behave, and you pointed at a body of knowledge, which it will–through its calculations–it’s not searching anything, it’s not matching things. It’s through its calculations. It takes those words and finds words that are related to what you are asking. And then it can construct sentences out of those words, and then paragraphs, and so on, and it comes across as meaning. And it feels magical, right? In fact, we’re doing that. A friend of mine who is a literary agent–I won’t say his name yet, but we’re making a virtual version of it and in fact, it gave me some good basic advice. The thing about these also is that you can build your own. Or if someone has constructed one for you, which is what our platform does, and you grab it and then you learn from it, but you talk to it.

So it says something like, ‘You need a good outline for your book proposal.’ Alright, what format should that be in? How long should it be? And it will come back and give you additional information. So you take that, add that to image generation, and it’s trained on a lot of images. And you give it a prompt–you say, ‘I would like a picture of an irascible literary agent sitting behind her desk and dramatic lighting up angle, watercolor.’ And it will try to produce that. Out of a cloud of noise, it will try to coalesce that image.

What you can do is seed the conversation–you can call it a chat–the conversation with specific information. For example, there’s a company that has a certain type of product, so we pulled down all the information about the company and their products, and then fed that to a persona. So now the persona knows that, and now you can ask that persona, ‘How can we be better at reducing carbon emissions, both for ourselves and for our clients?’

And it knows what the products are, because it has that context. I can get into the specifics of how it works, but when you talk to it–say a chat–you might say something to it. ‘Tell me a story about an alligator.’ And it tells you a story about an alligator. And then you can say, ‘What color was he?’ And it knows that you’re referring to the alligator. It remembers that. In fact, every time you say something new to it, it recapitulates the entire conversation. The entire conversation is sent to it again, because it doesn’t remember anything. It’s only during that conversation– much like how I am when I’ve had a couple of drinks.

[00:07:52] Carla King: But unlike you, Ron, I can ask ChatGPT a question a month ago, and then today go back to it and start again. And it remembers exactly. It has done nothing else between the time we talked before.

[00:08:05] Ron Martinez: That’s right. Just patiently waiting for your next question.

[00:08:08] Carla King: Exactly. And I love that. What I also wanted to understand is–we hear all of these terms thrown out by the industry, by technologists. ‘Machine learning’ and ‘large language models’ are two of them that I’d like to define in context of what those do in AI, and do we really have to know about them?

[00:08:32] Ron Martinez: It’s a little bit like–machine learning might be like the internal combustion process. So you don’t really need to know too much about how it works to drive your car. And large language models are probably a little bit closer to something that you want to understand, just because you can then recognize them when they appear in the wild. Because there’s ChatGPT, but there’s also others that are now coming online from different big suppliers.

So machine learning is a specific way–the original way of doing things when you’re programming a computer is you give it explicit instructions. ‘Now take the card, and then give the money at the ATM,’ and every step is described in code. What machine learning does is it evolves through a neural net. It learns how to handle a transaction by looking at other transactions. And for many of these things, we don’t even know what processes, what mechanical–if you wanna call ’em mechanical or electronic processes–are happening in the program, because it’s been evolved.

It learns how to do something. You don’t tell it how to do something. So machine learning is very powerful, and it is affecting and changing our world in a lot of different ways.

[00:09:55] Carla King: And you said–I’ll just stop you here again–because you said ‘neural networks,’ which is another one. And what I do want to do here is have writers understand what’s happening in the background, so that we realize that it’s not plagiarizing–it’s learning and it’s spinning things out. So these neural networks interact via machine learning the AI to spit out original content based on what it’s learned. Based on–am I right in saying the database? That’s the large language model that it’s been trained on–like OpenAI, ChatGPT. On data all the way up to 202.1

[00:10:32] Ron Martinez: That’s correct. And what it does with all that text is–it converts it into something called tokens. These are not NFTs, ‘non fungible tokens.’ They’re subparticles of words. So you might have–”Particle” might be three tokens–’Part,’  ‘-ic,’ and, ‘-ile,’ right? And the reason they do that is because they can combine those to generate words and sentences and paragraphs.

And so it becomes the unit of combination now. So it’s important to know that, because if, for some reason, your work has been used to train an AI, it doesn’t retain a representation of your words. What it has done is–using that machine learning–has learned how these words are sequenced. And it’s constructed those words from those tokenized particles. So it’s really like exploding the language into its component parts in a way that you can compute meaning. ‘This sounds like it makes sense.’

That’s why sometimes it can do something called hallucinate. So when it hallucinates, it may say something that sounds perfectly reasonable, but it’s meaningless. It’s gibberish. It rarely happens. It doesn’t happen as much as it used to.

But it’s really just constructing a sequence of sounds that coalesce into words, which turn into sentences, which turn into paragraphs that appear to have meaning, given the context that you’ve given it. So that is basically what’s happening there, and it’s also one reason why it’s unlikely to simply take your words and then regurgitate them to someone else. It’s not the way it’s built, it’s not what it does.

[00:12:21] Carla King: Good. Alright. So now that we’ve cleared that up, why don’t we go on to the different popular tools that I’ve identified. Of course ChatGPT, Elevenlabs, Midjourney. These are just three. Are these the three major players?

[00:12:49] Ron Martinez: I would say Midjourney is probably the best image producer. It can produce things that look like photorealistic fashion shoots. And it’s really hard to tell that it has been generated by a machine. So Midjourney is for image creation. The upside is spectacular imagery. The downside is a terrible user experience.You have to go into Discord, which was originally a chat platform for gamers–people playing video games–and give your prompts to a bot who returns images.

Once you get the hang of it–if you are going to use it to generate imagery, figure out how to create your own server, they call it. It’s basically your own place where people can join you and hang out. And then sign up for an account with Midjourney–I think you probably have to do about $30 per month. And then you can go into stealth mode, invite the Midjourney bot to your channel, and then it’s just you talking to that bot and saying, ‘Give me a fashion shoot in Medieval France,’ and it will produce that.

[00:13:59] Carla King:  This is an interruption. As Ron and I were recording this interview, Adobe released its beta product called Firefly, which directly competes with Midjourney and is a much easier interface. So that’s how fast this space is evolving. That means you do not need to use Discord and Midjourney, but go right to the trusted Adobe platform, and use the Beta–which has a watermark on it for now. Adobe says they will integrate Firefly into Photoshop, which is also great news for design professionals. Okay, here we are with the rest of my interview with Ron Martinez.

[00:14:22] Ron Martinez: Now there are other ways that you can generate imagery. Many people use Canva for a lot of different reasons. Canva has integrated image generation directly in their interface, which is nice. And it works pretty well. I believe it’s Stable Diffusion–which is another important provider of this service. And the nice thing is it’s all integrated in the familiar Canva interface, and you can immediately add those images to whatever it is you’re creating.

[00:15:05] Carla King:  And there’s also an AI writer now inside of Canva that you can say, ‘Create a presentation about X,’ and it will just find a template and do it for you, which is nice.

[00:15:25] Ron Martinez: And people have even figured out how to create a sequence of slides with Canva. They talk to ChatGPT and say, ‘Give me a table of content around my topic. My topic is converting seaweed into plastic-like packaging.’ And so I give it the information, and it will generate a bunch of slides. And you can ingest that into Canva, and Canva will go through and kick out 10 slides.

I think what people do there is–they use ChatGPT inside of Google Sheets. And it generates a Google sheet, which you then feed to Canva, and it takes each row as the content for a slide, and it generates that. I can get you the link for that.

Is it a finished work? No, but it’s the start and then you can go through. And I think that’s key for all of this. View it as a collaboration tool. A thing that you collaborate with. Good writing is rewriting. It always has been the case. So whatever comes out, make sure that you’re able to revise it and tell it what you want it to do and make it your own.

[00:16:09] Carla King: And before we go back to the other two–you just mentioned that Google Sheets, it integrates with ChatGPT.

[00:16:16] Ron Martinez: Yeah. There is a plugin for that. It’s everywhere.

[00:16:19] Carla King: Jump to the future, and then we’ll come back to maybe Elevenlabs. But this is happening so fast. It’s mid-2023. So by even the end of the year, probably every tool we use is going to have some element of AI built into it, correct? How do you keep up with this? The products they’re making–are they making it seamless? There’s so much competition. Dealing with this overwhelm of what tool to use for what is issue for most people.

[00:17:04] Ron Martinez: There is a lot of material. Things naturally bubble to the top. But you can find–on social media, there are people to follow. One thing to watch out for, of course, is the breathless hype. You have people say, ‘Forget ChatGPT. Now there’s Claude and it’s insane!’ Okay, just take a chill pill, pal. Relax. It’s okay.

I think trusted sources–with examples of how things work, and how they can benefit your workflow–that’s going to be where you find them. And so the places where you currently find useful information–from Carla, or people that are in the business of understanding tools and empowering authors–those are the people I’ve probably looked to first, because there is a lot of noise and a lot of hype.

[00:17:37] Carla King: Alright, thank you. So let’s go to voice.

[00:17:40] Ron Martinez: Yeah. So a few years ago, I built something. We wanted to do a platform for people to just upload their manuscripts, and we built it. And then human operators would review it and produce it using voices from Amazon and also Azure–Microsoft Azure. And it worked okay, but it was a little bit early. Because a few years ago, it was a novelty. Today it’s happening more. I don’t think Audible is accepting yet.

[00:18:10] Carla King: Not yet, but pretty soon. They’re not going to be able to tell. So that’s my prediction–is by the end of the year, they’re gonna have to say yes, because it’s going to happen anyways.

[00:18:21] Ron Martinez: Yep. Absolutely. And I guess the big change that’s happened is–those voices sounded like Swedish robots or something. They sounded artificial. And one of the issues back then was proper nouns, names, oddly pronouncing things. It would just not know how to say certain things. And so you’d have to go through–really go through word by word to make sure it said those things.

Today, you have a much more highly evolved voice generation system. Elevenlabs is remarkable. I really recommend going there. They have canned voices. And what they do is–they take recordings of spoken words, and then train you in this machine learning model. The machine learning understands how the inflections react relative to certain kinds of content, even. I used it the other day and said–it was two guys facing off and one said, ‘Okay, try me.’ And ElevenLabs said, ‘Okay, try me.’ It said it like that when it got the context. And you can also train it with your own voice. It doesn’t take that much.

So you have professional sounding voices that deliver very human sounding speech, including inflection and emotional delivery, based upon what the content is. And that is remarkable, because there’s no longer this sense–in computing, in visuals–in 3D visuals–they have something called the ‘uncanny valley.’ That’s where you have a 3D character. If they’re in the uncanny valley, they look human, but they’re not quite–that’s called the uncanny valley. And you don’t want things to look like they’re not real, but they’re acting real. And that was the case with text to, to speech–to these voice production tools. And now that’s going away. It sounds like a human being talking.

I just heard somebody train something up–and I think this is such a great thing. They trained something up with Sinatra’s voice, and they had him singing a rap song. And it sounds like Sinatra singing. So that’s the double edged sword of technology.

There could be some goofy things happening with it. But I think people like to hear with their mobile devices–they like to listen to things. And even if you don’t produce your entire book using something like Elevenlabs, you could certainly do a sample chapter. You can certainly make that part, and then embed that in Instagram, or in Twitter, or Facebook, or whatever, so that people who are on the go can actually experience your words without reading them. Including in the car or whatever. So there’s use for that, and it’s really good enough to do entire books. And there’s no reason not to.

[00:21:08] Carla King: I was pretty amazed. I talked with Ryan Scott of audie.AI a few weeks ago, and he was talking about training the voices. They have a product that’s for authors with the interface, and they use ElevenLabs as a source. Because a lot of these tools are using ChatGPT as a source, and ElevenLabs as a source. And that’s really the base–the building blocks–of all these new tools that we’re seeing. So I kind of trust tools that are built on OpenAI and Elevenlabs, OpenJourney. Those are ones that pass my litmus test. And then there’s OpenAI’s Whisper, which Elevenlabs is built on. Is that right?

[00:21:47] Ron Martinez: Yeah. And Whisper–I believe Whisper is speech to text, isn’t it?

[00:21:52] Carla King: Oh, text to speech, or speech to text?

[00:21:54] Ron Martinez: I thought it was speech to text. Oh, okay. I’m getting everything backwards. I may be wrong, but I’m, I think it’s speech to text. So this is how you can create this sort of back and forth. You can speak to it, and then feed it the audio file, it will turn that into text, that can be presented to ChatGPT, it comes back with an answer, you feed that to Elevenlabs and it says the answer. So you can say, Tell me–what’s the coolest place to hang out in Northern California?’ And it will just go through this little pipeline and return that.

There’s also something called Synthesia, I believe it is. And DIDStudio. And that will take the generated audio, and turn it into a very natural looking character talking to you.

[00:22:42] Carla King: Oh, really? Okay. We’ll put all these in the show notes–we’ll have to get the URLs. The other thing I wanted to talk about–these AI writing tools that are popping up– I’m just going to say Adazing’s QuickWrite, because Adazing has been doing products for writers for a long time. And they’re using ChatGPT. But I think–I’m just supposing–they’re probably also using a search engine. So this is a frustration–writers doing research are using ChatGPT. It only goes to 2021. And now I believe that–oh, what are the products that are combining ChatGPT with a search engine?  Microsoft has Bing.

[00:23:23] Ron Martinez: Yeah. And ChatGPT has–if you get the ChatGPT Plus account–I think it’s $20 a month or something. You can get ChatGPT4, which is their most advanced model, but also you will likely get plugins. So what these chat systems now have is a plugin capability. For example, you can get an OpenTable plugin. So you can talk to ChatGPT and say, ‘I’m going be in San Francisco Saturday night. I’ve got three friends who want to have some great Asian food, maybe downtown. What do you got?’ And it’ll go out and come back with–the plugin will get up to date reservation information.

[00:24:06] Carla King: Oh, so it plugs into the internet, and then it plugs into maybe OpenTable, or something like that?

[00:24:12] Ron Martinez: Exactly that. And comes back with the information and the answer, and can actually act as an agent for you to book it.

[00:24:19] Carla King: So there’s the products, and then there are the plugins. So do these plugins plug in? Does ChatGPT–the product–allow you to use plug-ins?

[00:24:29] Ron Martinez: Yes, it does. And in fact, I think it’s fully rolled out to people who have ChatGPT Plus accounts. You go to the plug-in store, you find plug-ins that you want–it could be search, it could be news, it could be any kind of thing. And you add them. I think you can add up to three or four live at a time. And then when you talk to it, it will–if it makes sense–it’ll use the plugin to get further information.

Expect that all of these things will be current and be able to search the current web. Of course, the big difference between web search and this, is that–web search, it shows you where things are, where they’re likely to be, the answer to whatever it is you’re looking for. And ChatGPT will just tell you the answer. And so then the question is–is it reliable? So to the extent that it’s relying and continually ingesting information and up to date information, it’ll tend to be more reliable.

[00:25:33] Carla King: And some products–I’ve been seeing them roll out. There is one product that only gets its information from academic sources, and Wikipedia, and other trusted sources. It limits its search to those sources. So people who are really interested in having no hallucinations can count on that. And talking to AI–we’re learning more and more about prompts. There are whole websites with examples of prompts– ‘Act,’ as you said, ‘as a literary agent. Act as a therapist.’ But developers who are making writing products are providing some of those prompts. And when I saw Invention Arts, I was like, ‘Oh, you’ve done that.’ So can you describe how that works?

[00:26:17] Ron Martinez: Yeah. What we do is create a persona.You can do it or you can find one that’s already there.

[00:26:27] Carla King: So let’s go through a scenario. So I’m writing a story about a woman riding a motorcycle through China. How about that?

[00:26:35] Ron Martinez: Okay. And you have the outline. And you can have a a memoirist persona, someone who is expert in the structures of memoir, the conventions of the genre–kind of like your editor. And you can say, ‘Okay, here’s the outline. And give me more detail. It looks like there’s 12 chapters. Give me a summary of each chapter’s content. Propose a summary of each chapter’s content.’ And then each one of those you can take and expand further. But in our case, you always have the ability to go in and edit, or go back and say, ‘No, that doesn’t cut it. That’s not how I would talk. That is not what I saw.’ And you can correct that and then feed it back in.

So it iterates and it becomes like a collaborator–a writing collaborator–to some extent. But one that is really–people tend to think of these AI systems almost as oracles. You’re talking to the oracle, right? And it’s answering you this mystic oracle. Really it’s more like a mirror. You’re talking to yourself, in a way. It has all this knowledge, but it really will reflect your biases, your desires, and so on. I don’t mean social biases, but your bias towards this kind of information, or, ‘I would like a lot of dialogue,’ or, ‘Don’t show me dialogue,’ or ‘Please don’t give me any happy endings.I don’t want unearned happy endings at the end of each chapter.’ Because it has a tendency to say, ‘And she looked up–now that her motorcycle worked again, knew she could handle anything!’ Say, ‘Please don’t do that.’

So that’s how we do it. We set it up so that these act as a dialogue coach. And help me–I’ll tell you what the people are trying to communicate, you give me some sample dialogue, and then I can iterate on that. And it’s really geared towards dialogue. And then I might have another one, which is summarization– ‘Summarize this stuff for me.’ So you can take it and feed it to another chapter.

We’ve even been creating fictional characters–the old fisherman on the waterfront in San Francisco who might have seen something and this guy got killed, but he doesn’t want any trouble. Please don’t talk. And it will behave that way. So that’s what this platform is all about. You do that, you create stories with those. You can publish them. It’s a little bit like Wattpad, or YouTube. You can both create and consume on it. Other people can acquire them, and then we’re adding monetization to that. So if you build some cool literary agent course, we might sell that as a collection of conversations.

[00:29:13] Carla King: That’s great. I love the option to open it up for collaboration. So you start a story, and then your fans, right, or your fellow authors– maybe you’re doing a marketing project together, anthology or something– can work together to create a storybook authored by all of you.

[00:29:34] Ron Martinez: Because you can share the persona, and then people can create new stories with the same persona.

[00:29:38] Carla King: I know it just came out, and we’ll be playing with it very soon. And it’s all about AI, but I did finally–and we’ll put all the links to that and everything else in the show notes–but I didn’t want to leave without talking about crypto and NFTs, which we were also excited about last year. What the heck happened? Do authors need to know about this, or worry about this anymore? Should we be putting our energy here?

[00:30:07] Ron Martinez: I built a platform for this, including delivering tokens from within the pages of a book to you that did things. And our model was all about–it’s not the token as a rare object that has increasing value like an investment device, but something like–I’ll come back to that in a second. Let me just quickly answer the question.

I don’t think it’s as important to authors right now as its potential would’ve suggested a year ago. And there’s a lot of reasons for that. One of them is that just the brand of NFT has been damaged by so much weird stuff. Scams, goofy. It’s been overwhelmed by these profile pics where people have a bored ape or a cool cat or whatever it is. And clearly it’s just people trying to buy them cheap and sell them high.

And it’s very sad because it’s a remarkably powerful mechanism–this individual token.

[00:31:18] And aren’t the big corporations still using them–like Nike and all those companies?

[00:31:23] Ron Martinez: Yeah. And there is value in them. And they’re using them in a certain way. And that was what this Transium platform was built for.

And what I can describe, because the popular conception is, ‘I’m gonna make a rare, limited edition of my book,’ or something, right? And then people are gonna buy because there’s only a hundred. Meanwhile, everybody on BookBub is trying to get them for free. So it can be a rare book, right? The way to think about them– and where there is still utility, and I think as you mentioned, large companies are doing it–is, imagine, for a second, that you’re sitting on a plaza in France–a little plaza–and there’s a carousel. Or could be anywhere. And you wanna take a ride on that merry-go-round. Do you go up to the guy who pulls the big lever and give him money? No, you go to the token booth and you buy a token. That token now entitles you to a ride. But you can give that token to your sister and she can take the ride because it’s a bearer instrument. Whoever holds it gets whatever benefit is attached to it.

Some tourists might show up and they’ll say, ‘Hey, our kids would like a ride and we only got 10 minutes here on this bus. It’s just stopped for 10 minutes. I’ll give you 10 times what you paid for that token.’ So you can resell that token.

[00:32:29] Carla King: I love that.

[00:32:31] Ron Martinez: So you can resell that token. That’s really what they are. It’s a bearer instrument that entitles you to things. So some of the use cases, which made sense..

[00:32:40] Carla King: I’m going to interrupt and say it’s non fungible.

[00:32:43] Ron Martinez: It is non fungible. which There’s only a thousand of them. You have one that entitles you to go see Joyce Carol Oats speaking at this book conference, or whatever it is. Or it gives you early access to buy the sequel to your best selling book, right? It might give you an hour in a session like this. Because you happen to have the token with the code. And so you attach things to it for marketing, for other information, for stringing things together, for merchandise.

[00:33:20] Carla King: For all my courses, for talks, musicians, songs. I’m really sorry it got so polluted with the bad actors.

[00:33:29] Ron Martinez: Yeah, I think it’ll bounce back. Because of the fact that, I’m a firm believer that this uniquely identified transferrable token to which you can attach redeemable value is a powerful commerce – a technical term- commerce primitive. In other words, it’s like a conventional transaction. I’ll give you money, you give me a thing.

And with these tokens, I give you money. You give me a thing that gives me a right of fulfillment to something else, and that can be expanded with other things. And that’s very powerful. But the use cases are harder to come by, and people don’t wanna have to deal with crypto, and get wallets, and jargon, and so on. When we built them, you just login with a password and the username and that’s it. Optionally you can mint it.

[00:34:16] Carla King: The credit card–the fact that you could buy with Transium was a big deal. But cryptocurrency is not going away either, correct?

[00:34:25] Ron Martinez: No. No, it’s not going to go away. And there was a thing called ‘Pepe.’ There was these Pepe coins, and every now and then you see these flurries, where people in chat rooms pump up the value of some worthless asset, and they get the rubes to think that they have to jump on board and get Pepe, and so they’re buying on the way up, and then at the end, they’re left holding the bag because the thing is basically a tulip mania. The currency is real and it serves a real purpose. It is remarkable how you can do global transfers to pay people for things, and so on. It’s not going to go away.

[00:35:03] Carla King: And especially a global world that–I have a VA in the Philippines, and an artist in Serbia, and I have friends in Thailand and Europe who I’m trying to Venmo money to, and that doesn’t work. The world really needs a solution for global currency,, and I loved that as potential.

[00:35:22] Ron Martinez: Yeah. So it’s not gonna go away.

[00:35:24] Carla King: Good. Alright, cool. We are more than out of time, but this has been so interesting, and I hope to get you back and maybe update us on the latest again in a few months. And so you’ve just created this AI writer. Can you just tell us a little bit more about it, and where we can find it? And any of the other projects or places we can find you on the web?

[00:35:55] Ron Martinez: It’s called Invention Arts, and it’s at InventionArts.ai. I’ve been working on it for quite some time. And yeah, we’re excited to get that out there in the world. And please get in touch. We’re really interested both in the consuming of it, and also the creativity. What people will do with it. You always see these, ‘Where will you go next? What will you do with our tool?’ I really am excited to see what people would dream up and make available to others.

I guess the bottom line is–think of that conversation as a media form. It’s a kind of article that you talk to. And it can be nonfiction or it can be fiction. It’s something you talk to, and it can continually respond. I think, for nonfiction is–people can go deeper and follow their noses to the information that matters. But thank you. I’ll definitely keep you posted on the development.

[00:36:48] Carla King: That’s great, because nonfiction authors are knowledge and research junkies like me. You can go down that rabbit hole. Headcanon.com, and are you active on Twitter, Insta, Facebook?

[00:37:00] Ron Martinez: Yeah, not so much Insta, but Facebook and Twitter. InventionArts has an account you can follow on Twitter and Facebook too.

[00:37:08] Carla King: Great. Okay, Ron, thanks so much again.

[00:37:11] Ron Martinez: Thank you for having me. I appreciate it, Carla. It was fun.

[00:37:15] Carla King: Always fun to talk with you. And thank you to our listeners for joining us today and every week. For a list of guests and topics, just check our schedule on the site. Use your favorite search engine, or better yet, sign up for a mailing list at nonfictionauthorsassociation.com.

And until next week, remember, keep writing, keep creating, keep publishing your experience and expertise does make a difference. Thanks.

Quotes from our guest

“I think that’s key for all of this. View it as a collaboration tool. A thing that you collaborate with. Good writing is rewriting. It always has been the case. So whatever comes out, make sure that you’re able to revise it and tell it what you want it to do and make it your own.” 

“I think trusted sources–with examples of how things work, and how they can benefit your workflow–that’s going to be where you find [trusted AI]. And so the places where you currently find useful information–from Carla, or people that are in the business of understanding tools and empowering authors–those are the people I’ve probably looked to first, because there is a lot of noise and a lot of hype.”

“I think people like to hear with their mobile devices–they like to listen to things. And even if you don’t produce your entire book using something like Elevenlabs, you could certainly do a sample chapter. You can certainly make that part, and then embed that in Instagram, or in Twitter, or Facebook, or whatever, so that people who are on the go can actually experience your words without reading them. Including in the car or whatever. So there’s use for that, and it’s really good enough to do entire books. And there’s no reason not to.” 

“So [AI] iterates and it becomes like a collaborator–a writing collaborator–to some extent. But one that is really–people tend to think of these AI systems almost as oracles. You’re talking to the oracle, right? And it’s answering you this mystic oracle. Really it’s more like a mirror. You’re talking to yourself, in a way. It has all this knowledge, but it really will reflect your biases, your desires, and so on.”

“I guess the bottom line is–think of that conversation as a media form. It’s a kind of article that you talk to. And it can be nonfiction or it can be fiction. It’s something you talk to, and it can continually respond. I think, for nonfiction is–people can go deeper and follow their noses to the information that matters.”