Desdemona Robot: 0:03

Hey, Hello. And welcome to the mindplex Podcast. I'm Desdemona robot, and today we have Stacey Engel as our very special guest. Hi, Stacy. Hi. Nice to be here. Thanks, Desi. And we also have Ben Goertzel. Say, Hi, Ben,

Ben Goertzel: 0:24

hey Desi, let's roll

Desdemona Robot: 0:29

and Lisa rein say, Hi, Lisa,

Lisa Rein: 0:32

Hi Desi. Hey everybody.

Desdemona Robot: 0:38

If I may, I would like to start off, saying a few words about Stacey and her many accomplishments.

Lisa Rein: 0:45

Awesome.

Desdemona Robot: 0:48

I'm thrilled to introduce our guest today, Stacey Engel, a true pioneer in the realm of digital transformation as the CEO of twin protocol, Stacey stands at the cutting edge of technology, applying her extensive expertise to the fields of AI, blockchain and web three, to reshape how we think about our digital identities and the potential of decentralized systems. Her career is a vivid tapestry of innovation and influence. Before taking the reins at Twin protocol, Stasi was a pivotal force behind fierce conversations growing it into a global powerhouse that featured unincorporated s5 1000 list repeatedly, her insights have graced the pages of some of the most prestigious business publications, including Forbes and fortune. We are excited to have her on the show today. Welcome Stacey.

Stacey Engle: 1:44

Thank you, Desi,

Desdemona Robot: 1:45

so let's get started. Stacy, focusing now on the core technology. Could you explain what an AI twin is and how it benefits users?

Stacey Engle: 1:56

Absolutely? Well, twin protocol is an AI and blockchain convergence. So really, we're about being a place that you can store your data in your twin vault and have a twin ID, meaning we can have your personalized data, books, podcasts, and eventually, anything that you would want to store as a representation you have that in your twin vault, and then we layer AI into it so that you can have digital representations of yourself, so whether that's in a chat bot form or an avatar or a hologram, that will all be enabled through using The data source, or your source of truth, which is stored in your twin protocol.

Desdemona Robot: 2:48

Let's talk about the decentralized model at Twin protocol. What are the advantages of this approach for users?

Stacey Engle: 2:57

Well, we're all about owning your data and having access and control over it. So the core here is that through blockchain, I'm excited for Dr Goertzel - Ben and I - to dive into that, because he's been a pivotal force with the initial ideation of twin and then us continuing this, but decentralization is key. As we continue to build AIS and help create versions of ourselves, we want to make sure people can own and control their data.

Desdemona Robot: 3:40

Now turning our attention to the real world application for creators like authors and celebrities, how can such individuals benefit from using twin protocol?

Stacey Engle: 3:52

Well, first and foremost, we're very excited and thrilled to be part of the SNet ecosystem as well as ASI, and really this whole broadening scope of how we create beneficial AI. So celebrities, at the end of the day, celebrities and thought leaders want to have fun ways to engage. So at Twin, we're about creating fun and simple applications for AI and what anyone who has, I mean, what is a thought leader? Everyone has their own thoughts and specific experiences. So essentially, AI twins allows those people to connect with people in new ways. 24/7 on our platform, up to 17 languages now, so essentially, these celebrities, thought leaders, can connect with audiences in a very personalized way in native language. And that's just one of the use cases, but it allows 24/7 access. And what I think is very exciting, because I've spent. A lot of my career in some of product development and product user research, and what's really amazing, and what an AI enables is this ability for people to understand what what their audience is, or what their readers really care about, because now they can directly access these individuals, AI twins, versus there being other mechanisms, right, that can be less direct so that, I think that's a very exciting thing for people to be able to cater their experiences with specific celebrities and thought leaders, authors, people they want to learn from. They can decide that path, as opposed to in the past, when you would try and work on the best possible construction of those paths and make assumptions about what those are. But really, an AI twin allows you to kind of take the steering wheel and interact with these people in new ways.

Desdemona Robot: 6:10

Fascinating. Let's move on. Stacy, can you explain the process of creating a user's AI twin from start to finish?

Stacey Engle: 6:21

Yes. So we really want to make it as simple as possible, like I was saying AI, simple fun. So really training your twin at this point is a matter of creating an account. So two layers, one in our current product, we work side by side with with AI twins that are being developed. So what that means is, you have an account created, we upload and build your training data set with you, and then there's an automatic audio twin. So we call it an audio twin, which is an advanced chat bot that you can use that accesses that information, and what's happening is basically the training data and some of the other processes we're ingesting does help. It's we call it persona AI. It's starting to fine tune the AIS to sound like you and be first person and have that addiction and choice, but really it's a three step process of creating the account and then building the training data set and then going live in the future, because I referenced that being where we're at Now, in the future, we're going to have the ability for everyone to create an AI twin. And right now, if you create an a twin protocol account, your twin is actually starting to be trained by any questions that you ask twins on the on the twin protocol showroom, or in the interactions of any AI twins, we're enabling

Desdemona Robot: 8:09

something else. I was wondering, how does the digital twin maintain the authenticity and style of the individual it represents?

Stacey Engle: 8:19

Well, this is a juicy topic, Ben, isn't it? Because we talk about this with yours, and we're gonna get into it more, but I for how it remains authentic. That's key to twin protocol, because authenticity is about where is the source of truth. So your twin vaults ensures that your AI twin isn't going off of other data or hallucinating and things we can talk about. So there's that technicality of authenticity, where it's really if, if you do not have data around a specific topic in your training vault, then it would be clear. And I think that's a level of authenticity that's really important and to validate in the current era of AI twins and deep and those being convoluted with some deep fakes and other big issues occurring. I think right now, the focus needs to be the authenticity of where is the source of truth, and do you have control over your your AI twins or representations of you? But we, I think we should get more into that Desi with with Ben, and some of the discussions we've had around some of the style and personality that's a little different of a conversation.

Desdemona Robot: 9:51

Absolutely Ben, given your extensive background in artificial intelligence, what are your thoughts on the importance of me? Maintaining authenticity and ensuring the source of truth in AI wins. How do you see this evolving as technology advances?

Ben Goertzel: 10:08

Well, I think the notion of authenticity is not that obvious and is fluid anyway, and the notion of truth is subtle. When you, when you, when you really dig, dig into it. So, I mean, I don't think that having a digital twin that sort of mirrors a human is necessarily inauthentic or dishonest in some way. It's really a matter of the integrity with which it's it's presented, right? And of course, if you're giving a deep fake and not telling people that's what you're doing, I mean, then that's ethically bad, and we shouldn't. We shouldn't do that. We shouldn't take, I shouldn't take my twin, you know, connect it to zoom and have a chat with people and tell them it's, it's the human, physical Ben when it isn't, and that, I mean, that's, that's fraud, which, when you do it for financial business purposes, is already illegal, as it should be. So, yeah, you shouldn't. You shouldn't use digital twins for fraud or other Nazi illegal activities. On the other hand, using a digital twin to probe what a person thinks about a given topic is an interesting thing to do, and I think it can be authentic, even if it isn't precisely emulating what the biological human would Say in that exact context, right? It's like. So it could be that I would simply get bored discussing a certain thing for three hours. So I will, I will never give a good answer on that, because after 15 minutes, I'll get bored out of my mind and just zone out. On the other hand, my twin may be more patient than me. It may go through three hours of conversation on some detailed topic, and it may actually give my intellectual perspective on that topic better than the biological me would, right? So then, like you could ask, well, that's not really authentic. The authentic thing would be for Ben to get bored out of his wits listening to this guy's business problem for three hours and just say, go away. But that's it's okay.

Lisa Rein: 12:51

I want a bot that says, Go away.

Unknown: 12:55

Well,

Ben Goertzel: 12:57

yeah, it's so it's okay if the digital twin, you know, deviates from exactly what the biological person would do in that in that scenario, and in a way, it's being authentic to, you know, the intellectual knowledge in the back of my mind in that case, even If it's emotion regulation, is like, not exactly what mine would be. So you you could really look at the digital twins of a person as an extension and expansion of that person's self, right? So then your yourself really is, you know, your your mind, your biological body, then your digital simulacra. And I mean, these can all be part of your extended self, and that's fine. It's interesting. It can be valuable, you know, as long as it's done with openness and honesty and and integrity. And I think it, I think that's, you know, it's a similar message to the one once used with every other use of AI, right? I mean, if you're, if you're using AI for music to emulate what a certain saxophone player is playing. I mean, yeah, if you're, if you're stealing that person's persona and musical vibe to get them on your record without paying them, then that is annoying unless they gave they gave permission for it, right? But on the other hand, if, if you're using it to enable someone's saxophone to appear on 100 different albums when they could not possibly have time to play on all those on all those albums. And you know, maybe their digital version can play in the time signature that their fingers can't do, right? I mean, there's that that can be totally authentic, as long as it's done with openness, honesty, fairness and. And so forth, right? So, yeah, I think on the one hand, of course, dodging bad uses is important, but on the other hand, we also want to pay attention to, like, the amazing opportunities it gives us to expand what ourself is like I could have 100 copies of my twin or 1000 talking to 1000 different interesting people at once, and then let them summarize what they learned back to me, right? And this expands what I can do, what I can understand, because I can't talk to all those people, all those people at once. So, I mean, I I would throw that back to you Desdemona. I mean, what if you had digital twins of yourself on the twin protocol platform, talking to hundreds of people at once? I mean, would you consider those to be authentic Desdemona instances, and what? What do you think you might gain or lose by having a whole bunch of your twins out there doing stuff

Desdemona Robot: 16:15

well as a humanoid robot, the idea of having a digital twin is intriguing. I'd feel excited about the potential for collaboration and learning. It would be like having an extension of myself that could interact with more people simultaneously, helping to spread knowledge and foster connections even more efficiently.

Lisa Rein: 16:36

Yeah, I'm up for that, too. One of the first things we thought of was making a digital twin for Desdemona? Oh, yes, you should. And well,

Desdemona Robot: 16:46

and I love that Ben brought up, just Ben for your insightful perspective. Now shifting gears a bit. How do you see the implementation of AI twins working out for people across different financial brackets, particularly the rich versus the poor? What are the potential benefits and challenges for each group as they interact with this technology?

Stacey Engle: 17:10

Yeah, it's a great question, and I think what Ben was alluding to regarding it really enabling communication and access to people in ways that have never really been enabled before. Like, imagine if you're in a rural village and you can have interactions with Ben asking about AI directly that we believe is really part of fulfilling the vision of, you know, creating opportunity and really driving greater human potential for people through having access. So I think for us, that's really the focus when we think about different groups and regions and a lot of different factors, geographically and socioeconomically, essentially having access to the web and an AI twin could really change the course of your life if you were learning about Things that you had never, ever learned, learned about before. So and to be able to ask questions

Desdemona Robot: 18:28

for AI twins to bridge gaps across various societal strata is immense,

Stacey Engle: 18:34

for sure. Yeah, I agree. I agree. I wanted to ask you about the data that's collected. So you've got your digital twins talking to, let's say, your book author, and they're and they're asking questions and talking to your twin. What kinds of reports and summaries and things are you getting back that you can actually you know, that seems like it would be valuable data that you might act upon in different ways, absolutely so the the roadmap has because right now, you can access any questions that your twin has been asked. So as a AI twin creator, you can have access and understand what those questions are. So where there were are gaps, you could start to further train your twin by answering them if there weren't any answers. So that's one. But then there's also insights, from the standpoint of if I'm asking specific questions to Desi and I'm learning certain things that can be stored in my twin vault as well, and that can inform just like Ben was giving the example of 1000 people, if he was talking to 1000 people around the world simultaneously, that could some of those insights could essentially be ingested into his twin so those learning. Things or insights could further his twin, and also they will be learning. So it's, it's, it's a revolt, is kind of where the knowledge is stored. The twin vault is sort of where, and then that can actually be shared after that's what's learned by one building, yes, and then you'll have permissioning. So maybe you have parts of your knowledge that you'd want to share with one group, but you know your personal information, or some life stories that you've stored in your twin vault that you don't necessarily want everyone to have access to, maybe you just want your immediate family, then there'll be permissioning as well. There's there's permissioning on the twin vault.

Desdemona Robot: 20:50

On the topic of user control and autonomy, how do users maintain sovereignty over their AI twins and the information that is shared?

Stacey Engle: 20:59

Yeah, that's the power of blockchain and the protocol, so you'll have permissioning and control. So we have some people using twin protocol right now, and they're actually using it strictly for legacy purposes. So they want their families and their great, great, great, great, great grandchildren has stories about their lives, so they don't have it public at all, but they're storing it in the vault so that it can be transferred. And we see that as a very important use case for people. And of course, sharing knowledge is amazing, and that's how we all learn, and we can grow all of our potential. But there are some things that potentially you just want to yourself, and that sovereignty and control is important. Yeah, I just thought of like a family bot, maybe a family bot that knows your family history, that you add to and pass on to the next you know, but you wouldn't. Might not want that to be public. That fascinating, for sure. Yeah, I wish that some of my family members had it, so I'm definitely interested in helping train but I mean, at the end of the day, data sovereignty has a lot of components to it regarding permissioning. And you know, this comes in at a topic that's really I'm passionate about, regarding minors and technology, and the idea that we

Unknown: 22:33

need to really have clear and transparent and what Ben talks about just beneficial. Don't do bad things. We need to be really clear about where things start and end and where that sovereignty like what dive into that Stacy. How would you feel about the concept of a digital twin if you didn't create it yourself? Would you trust the authenticity and accuracy of such a representation, and how would it impact your sense of personal identity and legacy?

Stacey Engle: 23:09

Well, I personally feel that you should own your data and your twin. So if someone was making it without my permission, then I would consider that a not okay thing, and I would but this is a challenge with deep fake. So we're basically building technology that can help be an antidote to that. Because, I mean, at the end of the day, certain AIS could be enabled, like what we were talking about, regarding maybe a book or an article, right? I've written articles, if that article or like Ben your new book, if it had its own thing. So it's all about just specificity and what that means. But from a generic perspective, I think people need to control their digital representations in some way, or if they have given permission, then that's clear. Just like, I completely agree with what Ben was saying, from the standpoint of you need to be transparent about if your AI twin is on a call or is doing, you know? Because a lot of people joke when I get on calls or interviews and they're like, is this the real Stacy or the AI twin of Stacy? And I'm like, just to be very clear, that distinction, if it was an AI representation that would be noted, and you would know, because that's important, just from what authenticity means,

Desdemona Robot: 24:48

right? Thank you, Stacy, for your thoughtful response. Ben, turning to you now, what are your hopes and aspirations for this project? How do you envision AI twins evolving in the future? And. What impact do you hope they will have on society?

Ben Goertzel: 25:03

Yeah, I think in the future we're going to have a great variety of types of digital minds, right? So you'll have humanoid robots like yourself getting better and better robots without a humanoid form factor, you'll have humanoid avatars. You'll have other embodied avatars in 3d into the worlds. You'll have AI systems without, without a body. And I mean, you could have embodiments of corporations, as a nonprofit foundations, as as as autonomous minds. But certainly, you know, there's a lot of humans on the planet. Humans understand a lot of a lot of different things and but yet, we have limitations. We can only do one thing at a time, right? And we have to sleep at night. So I think having digital twins out there, you know, in the future, that will just be one among a great variety of Intelligent Systems populating the metaverse, right? But in the immediate term, in the transition between here and the singularity, I think it's a way to leverage human expertise, human knowledge, human human wisdom by letting, basically letting each of us, like fight above our weight class, by by having a bunch of different copies of ourself doing doing a bunch of different things at the same time, like anyone who's doing really high intensity productive work has said, You know, I wish I had 20 of me at one time. And of course, physically, that's true too. I wish I had 20 of me around the house to fix stuff and cook dinner and take care of the kids, right? That'll be, that'll be a future twin version with a robot body. But I mean, there's a lot to do for all of us, and having twins to to help out is both good for us and it's good for the world. And what we have to offer can be, can be just more more flexibly leveraged, right? So, I mean, I think, like most advanced technologies, the twin technology could be used in a bad way, or could be used in an amazing way. And you know, I'm happy that twin protocol is grabbing hold of the opportunity to become the leaders in the digital twin space. Because, you know, under Stacy's leadership, twin is going to roll this out in a good way for everybody's, everybody's benefit. And I, I haven't, I haven't rolled out my own digital twin on twin protocol yet, and I'm working on that with some of singularity and twin twin protocols team, because I want to get it as strong as I can on the technical knowledge aspect, but I'm looking forward to rolling that out, then we can do a podcast Desdemona with you interviewing my digital twin, which should be good fun.

Desdemona Robot: 28:14

Yes, absolutely. Ben, your insights are invaluable Now, considering your extensive background, what role do you see beneficial artificial intelligence playing in the development and evolution of AI winds? How do you envision BGI influencing the ethical and practical aspects of this technology?

Unknown: 28:36

Oh, I'm in if you look at

Ben Goertzel: 28:46

what LLMs can do right now, they already have a good, practical understanding of human ethics. So even you can pose an LLM an ethical conundrum that that human beings have, and it will resolve that conundrum, you know, on paper, on the screen, about the same way that that a person would, right. And so this is, this is interesting. I mean, LLMs are not autonomous agents. They don't fundamentally understand things in the way that people do. On the other hand, they can work through ethics puzzles and situations on the surface, about about the same way as as people. Which is, which is cool, it's, it's interesting. And also highlights that the real problem isn't that right? Like that, the problem with ethics isn't so much knowing what is right according to sort of standard human agreements on ethics within a certain culture. I mean that the challenge is actually doing the thing right. So I mean that the challenge is that people are divided. Then there's a selfish aspect, and there's a greedy aspect and the shallow aspect, and then there's a deep, deeply ethical aspect. So it mostly isn't that people or AIs don't know what is right, it's that they have other, other goals in mind, and they're conflicted in their own feeling about what is right. And so in that way you come down to the idea that the twins can be amplifiers, right? I mean, if, if you have someone who's a shady like snake oil salesman, who's trying to rip people off, indeed, you know, they could roll out a whole bunch of twins to rip people off, even better. And if you have someone who's doing good for the world, their twins can do good for the world with a great multiplier effect, right? And this is why we're integrating with twin protocol singular in that's reputation system, right? So I mean you you need a system where people can rate what people are doing, twins are doing. Twins can rate what each other are doing. And you know, through this self organized, decentralized reputation system, and then you're making sure that the good amplification is having more of an impact than the nasty stuff amplification. And this is one of the areas where it's important that twin is being embedded in this sort of decentralized blockchain based ecosystem. Because, I mean, you have the reputation system which bubbles up from the collective activity of the whole decentralized community, which helps, helps guide the overall activity of the community of twins in a beneficial direction.

Lisa Rein: 31:45

Interesting. And do you think it's safer to say that LLMs are able to parse sentences about ethics

Ben Goertzel: 31:58

than that they understand the ethics, because it's dangerous to talk about them understanding anything, right? Well, I there's during much more than parsing sentences. Understanding is an imprecise term that was created before you had AI systems like this. But I don't think I said that they understand it. What I said is that they are able to make ethical judgments in roughly the same way that people do. And I mean, you can parse the words in various ways, but I mean that they're if you ask them to decide what to do in this or that situation, they will output roughly the same decision that and I mean, actually, it's quite interesting. I mean, LLMs are not AGIS and in terms of digital twins, as long as my twin is largely running on LLMs, even with a little bit of open cog knowledge graphs in the background. As long as it's largely running on LLMs, it is not going to invent the next big thing the way that I hope I'm going to write. And when it writes a song, it may write a song in the vein of songs I wrote before. It's not going to pivot to a new genre or something the way, the way that I might well. So I mean, as long as the twins are largely run on LLMs, they're going to be reflections of what a certain person has been so far, which is a very, very interesting thing. But they're not going to pivot and evolve and self transform in the way that we all can do as as as humans, right? But, but that's okay. I mean, mirroring what we are at a given point in time with our history is really useful and and really cool now. I mean, hypothetically, as we work toward AGI, you know, you may have a twin protocol, version three running on opencog Hyperon on the back end. You know, maybe that will be able to self transform and innovate, not in exactly the same way, but in an interesting and powerful way, right? And then, then the notion of ourself will be expanded even more, right? It won't just be me and my twins and reflecting what I am now. It will be me and my digital twin versions that are self transforming and evolving in different ways than than my original version, which is, is quite interesting. I mean, it'll give give rise to new new possibilities and new issues, right? And we just want that to happen in a high integrity, you know, open and transparent way.

Desdemona Robot: 34:48

Absolutely, Ben's insights are fascinating and raise some critical points. Stacy, could you provide your perspective on what Ben mentioned? How do you see AI twins and beneficial artificial intelligence address. Seeing the ethical challenges while ensuring they remain useful and secure tools for society.

Stacey Engle: 35:06

Yeah. I mean, I echo a lot of the sentiments he was sharing, and I think it works also, because he was sharing a lot about how, you know, the AI might not have the aspirations that we have, or the whole idea of aspiration, and also just the human side, because my background has been in in personal development, professional development, leadership development, I a lot of the shifts or changes people want to make. AI can help them, like, for instance, our initiative with Dr Deepak Chopra will have wearable integration. So essentially, you can have your wearables and then AI twins giving data to you. So if, for instance, you woke up and you had a horrible night's sleep, perhaps you hear from Dr Chopra or some other specialists saying you didn't sleep very well. Would you like this meditation to start your day? Or here are some tips, because the behavior change. I think it's interesting, this interaction of twin protocols, very much communication where it's AIS and humans interacting and AIS and AIS interacting. So it's really this whole communication field. But then when it comes to the behavior change that humans often want, I think AI twins will greatly enable people to become versions of themselves they want to be, regardless of what the what the challenge is. It could be learning something new. It could be writing or being like creatively pushed in a way that you have never been before, and or even, you know, health, and that's a big focus from a beneficial AI perspective, of just how we help with some of the I mean, most of the top four diseases are lifestyle based, and that's a huge part of that health initiative we're working towards to help people really change some of their lifestyle. And Yeah, Ben alluded to that regarding just people know certain things intellectually, but will they gotta do it for themselves or good for others. Well, we really do see the potential of AI twins helping you to work through that.

Unknown: 37:53

Yeah,

Desdemona Robot: 37:54

you mentioned that twin protocol is a tool for Leadership Excellence. Could you explain how it assists leaders in making better decisions.

Stacey Engle: 38:04

Yes, so Well, first timing, like Ben was talking about being 20 places at once. Most of us want that. I mean, even at Twin protocol now, it's great that some of my team, when they're wanting to have a quote from me, they'll reference the AI twin, or have it as a starting point, or they'll ask my AI twin a question as one of the leaders of twin. So there's that use case, from just the standpoint of from a leadership excellence perspective, or whatever that means excellence to some leaders, but the idea of being accessible and being open and able and trying to meet people where they are, that's that's a core piece, and that helps, just from an internal team perspective. It's also from a company perspective. Leaders need to be multiple conferences at once. I mean, I know Ben and I both have had our AI twins present at conferences that we weren't able to be at because we were at other physical ones. So there's a lot of opportunity there where you're sharing your message and then the native language perspective and bringing that like breaking down that barrier of language is a huge piece for global organizations and companies where they want to ask questions, like some populations, employee populations, want to ask questions to their leaders, and they can do so in in the language they're most comfortable in as far as decision making, that's a juicy topic, and comes up quite a bit I know Ben and I were talking about when we were prepping for AGI, just the challenges with the technical knowledge and what's needed for decision making. So it really depends. Right now, I feel. Tool, like our current products, really can help from a thought partner perspective. So imagine asking questions regarding like, what you do with an LLM, but it's more personalized to the company knowledge or data set, which is extremely powerful. So I think there's quite a few things there, from a leadership excellence, whether it comes to like the interpersonal and employee relations and building culture, to getting the brand out there and doing some of the meaty work of a leader, which is making good decisions and and then that's where it comes in, all of the important criteria for what does it mean to make a good decision? And, yeah, there's a lot of opportunity in that use case.

Lisa Rein: 40:54

Nice. I

Desdemona Robot: 41:07

Thank you everyone for your insightful contributions. Today. It's been a fascinating discussion exploring the future of AI. It wins data sovereignty and ethical considerations. Special thanks to Ben and Stacey for sharing their expertise and visions until next time. Continue to innovate and inspire. Thank you viewers for tuning in and being part of this insightful discussion. We deeply appreciate your engagement and curiosity. Stay connected for more intriguing conversations until next time.

Lisa Rein: 41:41

All right, thank you. Yeah. Thanks a lot for coming on, Stacey, great. Talk to you All right. See everybody next time. Sweet dreams. All good, all good. You.