The Only People Who Actually Beat AI | Dr Vivienne Ming
Hosted by Mykel Salomon · Featuring Dr Vivienne Ming · Published
Mykel Salomon is the founder of The Human Protocol Group, creator of The Human Protocol Framework, and host of The Human Protocol Podcast. He explores how technological change affects human beings, and what it takes to stay human while building the future.
Guest: Dr Vivienne Ming — The Atlantic once called Dr. Vivienne Ming a “professional mad scientist”. In more mundane terms she is a theoretical neuroscientist, entrepreneur, and author. She is the Founder and Executive Chair of The Human Trust, a philanthropic "mad science" incubator, and the Chief Scientist at Possibility Sciences, where she is pioneering hybrid intelligence systems for the future of innovation. Vivienne was an early pioneer in using AI for human potential. As Chief Scientist of Gild, she built machine learning models trained on 122 million professionals to predict human capability. Today, through The Human Trust, she has developed AI to treat diabetes and bipolar disorder, reunite orphan refugees, and help autistic children learn facial expressions. Her new best-selling book, "Robot-Proof: When Machines Have All The Answers, Build Better People", shares the stories of 30 years of using AI to improve lives.
Most conversations about AI start with the same question: “What can I do that a machine can't?” Dr Vivienne Ming argues that this is the wrong question. There is no single magic skill that will remain permanently “robot-proof.” If humans can clearly define a task and determine what the right answer looks like, machines will increasingly become better, faster, and cheaper at performing it. Instead, we need to focus on the deeper qualities that make people uniquely capable. Drawing on research involving 122 million people, Vivienne explains how qualities such as curiosity, resilience, fluid intelligence, and perspective-taking can be more meaningful predictors of positive life outcomes than traditional credentials or individual technical skills. That insight becomes even more important in the age of AI. Vivienne describes an experiment in which humans were asked to make meaningful predictions alongside AI systems. Some participants simply automated their thinking. Others kept a human in the loop but relied heavily on the AI's reasoning. The strongest performers were a small group Vivienne calls “cyborgs.” These people didn't simply accept AI answers. They combined the machine's capabilities with their own reasoning, curiosity, and judgment. The result was something neither human nor AI could achieve alone: hybrid intelligence. But there is a serious warning. AI can make us dramatically more capable — while simultaneously encouraging us to think less. Chapters • 00:00 AI, human potential, and the future of humanity • 01:45 The wrong question about becoming “robot-proof” • 02:39 Why there is no single magic skill • 04:05 The brutal truth about what AI can already do • 06:44 What 122 million people taught us about human potential • 08:39 The deeper human qualities that predict success • 10:09 Why curiosity may matter more than technical skills • 13:38 Why you can't learn curiosity from a lecture • 16:02 Why we don't really know ourselves • 18:43 The experiment that changed children's curiosity • 20:58 Why the wrong question can be more valuable than the right answer • 21:48 Can AI predict the future better than humans? • 24:11 Hybrid intelligence: humans and machines together • 25:06 The danger of outsourcing your brain to AI • 26:30 The “cyborgs” who beat both humans and AI • 28:29 The human qualities that make AI collaboration work • 30:19 Are brain interfaces the future of hybrid intelligence? • 32:22 Building an actual human-machine cyborg • 33:49 Neuroprosthetics and human augmentation • 34:19 Can we increase working memory? • 37:15 What happens when AI merges with the human brain? • 40:19 Why we should never automate away creativity • 41:25 Will machines eventually control us? • 42:00 AI knows everything — but does it understand? • 45:20 The hidden danger of AI making us think alike • 46:50 Our greatest advantage may be the crazy idea • 49:11 Finding the balance between human and machine intelligence • 51:16 Can AI make people better over time? • 54:12 Socrates: the AI that refuses to give you answers • 55:40 What happens when AI teaches instead of answers? • 57:40 The future of AI benchmarks • 58:43 “I want to build better people” • 59:28 Purpose, meaning, and human potential • 01:03:18 The most human thing Vivienne has ever built • 01:05:10 Using AI to make ourselves more human About the Guest Dr. Vivienne Ming is a theoretical neuroscientist, entrepreneur, AI researcher, and author working at the intersection of artificial intelligence, neuroscience, and human potential. Her work has explored machine learning, human cognition, neuroprosthetics, education, healthcare, workforce technology, and hybrid intelligence. She has built AI systems to address deeply human problems, including developing an AI system for her son's diabetes and working on technology designed to reunite refugee children with their extended families. Her book Robot-Proof explores what humans need to develop as machines become increasingly capable. Her work ultimately comes back to one question: How can technology make us more human — not less? Connect with Vivienne Ming LinkedIn: https://www.linkedin.com/in/vivienneming/ SOCOS Labs: https://www.socos.org/ Thinking about implementing AI in your business? Download Mykel Salomon's free 10-Step AI Strategy & Execution Roadmap to clarify the business problem, identify the right opportunities, decide what should remain human, and move from idea to execution. Clarity first. Tools second. Humans always: https://dub.sh/thp-roadmap-yt Technology is changing quickly. The human questions matter just as much. Subscribe to The Human Protocol newsletter for weekly perspectives on AI, leadership, work, and what it means to stay human while building the future: https://act-3-agency.link/thp-news-yt
Conversation Summary
In this engaging conversation, Dr. Vivienne Ming challenges the traditional notion of seeking a "robot-proof" skill as AI continues to evolve. She argues that no single skill can withstand the capabilities of AI, as tasks clearly defined and measurable are often more efficiently performed by machines. Instead, Ming emphasizes the importance of cultivating deeper human qualities such as curiosity, resilience, and perspective-taking. These traits, she suggests, are more reliable predictors of success and fulfillment in life than any technical skill. Through the lens of her research with AI and human potential, Dr. Ming explores how humans and AI can collaboratively create a hybrid intelligence that harnesses the strengths of both.
“Because most people entire AI anxiety really lives inside one question, what can I do that a machine can't?”
— Mykel Salomon, host of The Human Protocol
The Myth of a Robot-Proof Skill
Dr. Ming states that the idea of one magic skill that is immune to automation is a misconception. If a task is clearly defined, AI will likely perform it better and more cheaply than humans. Instead of focusing on robot-proof skills, individuals and education systems should foster human attributes that machines cannot replicate.
The Importance of Human Qualities
Qualities like curiosity, resilience, and fluid intelligence are key predictors of life success and happiness. These 'foundation skills' are more critical to positive outcomes than traditional academic skills or credentials. To thrive amidst AI advancements, these traits should be integral to personal development and education.
“That's where AI has access to all this vast human data that is out there.”
— Mykel Salomon, host of The Human Protocol
Hybrid Intelligence Outperforms Both Humans and Machines
Ming's research identifies a small group she terms 'cyborgs'—who combine human judgment and machine capabilities—perform better than AI or humans alone. This hybrid intelligence approach leverages the unique strengths of both, leading to superior decisions and predictions.
Educational Shifts Toward Experiential Learning
Educating individuals to develop attributes like curiosity involves experiential learning. Dr. Ming highlights that lectures do not instill such qualities; rather, they are developed through experiences that reshape who we are. This shift in education could prepare future generations for a world where deep human qualities are more valuable.
Defining Purpose as a Key to Human Flourishment
Dr. Ming emphasizes the importance of having a strong sense of purpose. Purpose is not only linked to productivity and longevity but is also a critical factor in achieving happiness and success. Her upcoming book, 'Small Sacrifices,' explores the construct of purpose and its profound impact on human life.
“Are we really becoming better human beings while doing all that?”
— Mykel Salomon, host of The Human Protocol
Final Thoughts
Dr. Ming's exploration into AI and human potential offers a profound shift in how we perceive our roles in a technologically advancing world. By focusing on human qualities that cannot be easily measured or replaced, we can foster a society where technology enables humans to become more of themselves—creative, driven, and, ultimately, human. The path forward is not choosing between humans or AI, but embracing a partnership that harnesses the strengths of both to push the boundaries of what is possible.
Key Takeaways
- The Myth of a Robot-Proof Skill
- The Importance of Human Qualities
- Hybrid Intelligence Outperforms Both Humans and Machines
- Educational Shifts Toward Experiential Learning
- Defining Purpose as a Key to Human Flourishment
Frequently Asked Questions
What are foundation skills according to Dr. Vivienne Ming?
Foundation skills refer to deeper human qualities such as curiosity, resilience, and perspective-taking that are more predictive of success and happiness than technical skills or credentials.
Why does Dr. Ming argue against the concept of a robot-proof skill?
She argues that no single skill is immune from automation; if it can be clearly defined and measured, AI will likely perform it better. Instead, she suggests developing human traits that machines can't easily replicate.
How does hybrid intelligence benefit over pure human or AI efforts?
Hybrid intelligence, a collaboration between human reasoning and AI capabilities, results in better decision-making and predictions than either can achieve alone. Dr. Ming's research found 'cyborgs' outperform both AI and human individuals.
What role does curiosity play in Dr. Ming's perspective on education?
Curiosity is seen as a critical characteristic that should be developed through experiential learning rather than traditional lectures. Educational systems should prioritize fostering curiosity to prepare individuals for an evolving world.
What is the significance of purpose in Dr. Ming's research?
Purpose is a significant predictor of a range of positive life outcomes, including productivity and longevity. Dr. Ming's upcoming work will delve deeper into how a strong sense of purpose can enhance human flourishing.
Full Episode Transcript
This is the full recorded conversation from The Human Protocol Podcast, published by The Human Protocol Group. © 2026 The Human Protocol Group. Please link to and attribute the original episode when referencing this conversation.
THE HUMAN PROTOCOL (00:01.74) My guest built an AI train on a hundred and twenty two million people to predict human potential. Then she turned that same science on her own son's diabetes, on reuniting refugee children with their families, on teaching autistic kids to read a smile, and most people talking about AI had never used it to help a single human being. She has spent thirty years doing almost nothing else, so stay with me. Welcome to the Human Protocol, the podcast about staying human while building the future, where leaders, operators, and humans from every walk of life come to ask and answer one question How do we build the future without losing ourselves? I am your host, Michael Salomon. I have wanted to have this conversation since the moment I saw my guest's latest book titled Dr. Vivienne Min is a theoretic theoret theoret theoretical neuroscientist. Sorry about that, and in her own words, a professional med scientist. She has founded companies, trained AI on a hundred million people, and then use that same science to solve. Some of the most human problems imaginable her son's diabetes, refugee children separated from their families, autistic kids learning to read emotion. Her new and latest book it's called Robot Proof When Machines Have All the Answers Build Better People. Vivienne, it's a pleasure to have you on the show. Thank you. Vivienne Ming (01:45.871) It's wonderful to be here. That was a very flattering introduction. I I try not to cringe when people say nice things about me. it's it's lovely. Thank you. THE HUMAN PROTOCOL (01:56.959) Ha THE HUMAN PROTOCOL (02:01.779) it's it's not only well deserved, it's it's it's something that that is needed for someone as human as you are. And I really want to start with that, right? Because most people entire AI anxiety really lives inside one question, what can I do that a machine can't? And I know that you say over and over again that that's the wrong question to ask. So I would like to start there with you. Why is that the wrong question and what we should be asking or pursuing instead? Vivienne Ming (02:39.354) Well, maybe it's the way people ask that question because the presumed answer and and let's be clear, I for many people thinking about this today, maybe this started in twenty twenty two when Chat GPT got released. So I've been having this conversation with people like the Department of Education, you know, that big nowadays empty building. with NATO, with giant banks around the world since 2008 when I started my first company, let's say. and what they want, what they're really asking is a little bit more specific and and presumptive. What's this one magic skill? Right? If we could teach every kid in America or every kid in the world one magic skill. And for a long time the presumption was we'll teach them all how to program. Or maybe slightly broader, we'll teach them all STEM skills, which as a fancy Smancy scientist, I'm not exactly certain what a STEM skill is. We're teaching everyone statistics. Like so there's this idea. Here's how I often here's genuinely what it feels like to me when I was getting asked these questions. I still when I get asked it today, when I got asked it 15 years ago. THE HUMAN PROTOCOL (03:42.333) Hm. Vivienne Ming (04:05.676) It's like someone saying, Hey, Dr. Ming, I I got all this money. You know, turns out I got a trillion dollars sitting in a bank account somewhere. And I want to know the one single stock. I'm gonna put all of my money in that stock. and then I'm going to not touch it for 30 years. What's the one what an insane question to ask. Right. If you asked any actual financial advisor from Nobel Prize winners to, you know, a day-to-day schlub, they'd say, diversify. What are you what are you talking about? One single stock, that's insane. So here's kind of my brutally honest story. There is no skill that is robot proof, which is to say, if I'm saying, here's a thing humanity has figured out how to do. THE HUMAN PROTOCOL (04:49.512) Sure. Vivienne Ming (05:04.442) i.e., I give you a question related to it, you can give me the right answer, you know, the fancy algorithm or equation or whatever to get the right answer. We've got a better thing than often the best human beings for getting those kinds of answers. And it does it for three cents on the dollar. That's that's pretty hard economically to compete with. But as a lot of students are beginning to figure out, it doesn't even make sense to them. Why are you asking me to do this on a school exam when we all know there's a better answer in my phone and it's basically free? W so if that's your framing, what's the one magic skill? I have bad news for you. There is no magic skill that we are better at. Some, you know, elite performers are usually still better than the best AIs at these things, but you know, elite performers are expensive and there aren't many of them and they tend to be cranky about the world. And if they're anything like me, no one actually wants to spend too much time with them because We're all a little bit like, you know, the character house from the show House. So nobody actually likes having us telling them what to do all the time. And if that's the end of the story, obviously it's pretty dismal. so I want to reframe this. You talked about the project I did where we had 122 million people. So I was the chief scientist of one of the first companies ever using AI in hiring. And so this was way back in THE HUMAN PROTOCOL (06:29.222) True. Vivienne Ming (06:44.443) 2014 and 2015. I mean, we ended up on the cover of the New York Times and the Atlantic. So AI existed 10 years ago and 50 years ago for that matter, but it hasn't really registered like it has today. And my job was to figure out how good you were at a job you've never held. Because our customers were Google and Facebook and Rolls-Royce and Nestle, and they wanted to know who they should hire. And so the the funny thing we found was, wow, none of these skills like you'd list on a resume or LinkedIn or that you know recruiters were searching for on sites, none of them predicted anything. You say you know the Python coding language. What programmer doesn't know Python? You say you know this technique for engineering. Everyone who went to university and got engineering knows that technique. Like none of that was predictive or meaningful. In fact, most university degrees weren't predictive or meaningful, which is probably pretty shocking to people, but it was true. once we already knew someone was a working professional. So again, the story is well, then what's left? If hiring wasn't about these magic skills and being robot. Proof wasn't about these magic skills. And I won't be a jerk and just tease, but I'll only give a partial answer for right now. It's you. It's whatever it is that makes you different than everyone else. And that probably sounds a little bit like a Hallmark card, but what I really mean is there are these deep, deeper human factors. I call it meta-learning. THE HUMAN PROTOCOL (08:23.968) Mm-hmm. Vivienne Ming (08:39.266) I really like the term foundation skills. I do not like the term soft skills because I'm a hard number scientist. I like measuring things. These are measurable. They're just harder to measure. So think resilience or fluid intelligence, perspective taking, your ability to understand other people. Like without that, none of those forget AI. Without that. THE HUMAN PROTOCOL (08:44.94) Mm-hmm. Vivienne Ming (09:05.816) Those other skills don't really predict very much about positive life outcomes, period. so that becomes the real shift. There is something AI can't do that you can do, but it isn't a magic skill. It's these deeper qualities. THE HUMAN PROTOCOL (09:23.657) I love how you frame it that it's not a magical skill because I'm listening to you right now and those are those are skills that are tough to learn, tough to teach. Those are mostly things that you cannot measure. You have to experience, you have to feel it, you have to build it. And that's what makes all this conversation and the type of better people that we need to build in people. Hard to get to because for some people unless they can measure, unless they can identify a path to learn something, it's easy. But for the rest that they have not had the opportunity to experience and going that path, it may be something hard to swallow in a way, or to see themselves there, right? Vivienne Ming (10:09.164) Yeah. But I think what's interesting is again being a hard numbers scientist, a computational scientist, many people like me would steer away. this is gonna be a little foreshadowing because I guarantee we're gonna come full circle here, but we'd steer away from looking at something like curiosity. Feel squishy. How do you measure it? Well, you know. THE HUMAN PROTOCOL (10:31.848) Mm. Vivienne Ming (10:37.174) One of my fancy mancy degrees is psychology, and there are ways the psychologists measure curiosity. And it turns out if you took something like that and then combined it with this wonderful gift to humanity resource, the UK Biobank, where the the the United Kingdom Britain has collected all of this data. From their National Health Service and everything else, and made it available for researchers to study interesting questions. You know, you can imagine around Alzheimer's, around cancer. Well, some amazing researchers went and looked at these supposedly squishy qualities and found what would you like to be true when you're 65? I want to have made some money. I want to have friends. I Don't want to be dead. I don't want to have cancer or diabetes. you know what? I want to be out and able to walk around. you can imagine all these things. I want to be happy. Turns out, this research shows that people that score highly on this, yes, hard to measure quality, curiosity, all of those things are more likely to be true. Now, it isn't like a magic crystal ball. Just because someone's curious when they're younger doesn't mean none of these things, and no bad outcome will ever happen to them. But when you look at a population, a whole country like the UK or the US, things like curiosity and some of these others I mentioned, resilience and perspective taking, fluid intelligence, turn out to be big predictors of positive life outcomes, often much more so than. Do you have a degree in computer science? Or do you know how to factorize a polynomial? THE HUMAN PROTOCOL (12:35.921) Not any hard skill. Mm-hmm. Vivienne Ming (12:40.306) so you know, taking these into account, what I'm trying to do is open it up. This isn't wishful thinking or just sort of some weird, I don't know, progressive ideal about people. In a hard number sense, if I had to choose, every kid takes an algebra algebra class or every one a kid takes a class that nobody is currently taking. A course on curiosity. We'll come back to whether that's impossible to do. But a course on curiosity, 100%. I would say, why has no one ever taken a course on curiosity? Why isn't that part of every school system? Because it predicts so much more. And to just close the little tie it off here, you said earlier, it's hard to learn. Yeah, this is one of the differences. I can't give you a lecture about curiosity. I mean, I can lecture THE HUMAN PROTOCOL (13:34.989) Uh-huh. Vivienne Ming (13:38.555) PhD psychology students about the construct and how you measure it and these hard skills, you might say. I a lecture will never make you more curious. It will never make you more resilient. We call this experiential learning. You have to go through experiences over time that in some sense change who you are, right? Going from being less to more curious or less to more resilient. THE HUMAN PROTOCOL (13:49.563) Yeah. THE HUMAN PROTOCOL (14:03.226) Mm-hmm. Vivienne Ming (14:07.758) You're kind of a different person. And, you know, getting comfortable with that idea. Our goal is not to deliver a bunch of facts and magical algorithms. Our goal is to make you a different person, the best person you could be. And I know best is a loaded term, but you know, you're gonna live longer, you're gonna be happier, you're gonna have a bigger positive impact on the people around you. Hey, you're gonna earn more money. Best. And to me. Curiosity is a absolutely required curriculum. THE HUMAN PROTOCOL (14:45.306) It's like life itself. What you're talking right now, it's like you are painting a path forward on how we should experience life itself, right? And the experiences that comes because you are curious enough to try things and fail. Because you go through maybe hard time and you adapt and you adjust to it. You create as you go, right? Those are the things that yeah, I totally agree with you. Those are the things that life It's asking us to be more and more, which is hard because you cannot learn it, you have to experience and it would also I think Vivian require our educational system to to have this shift on on on teaching our kids and even us adults on these type of skills of e of life experiences that will get us to a better place as humans. And I think that's the part of the conversation that I also feel missing sometimes in our society. Who we are, right? Do we really know ourselves? Are we doing a lot of work on us to discover the things that we like and we don't like, right? Vivienne Ming (16:02.02) Well, let me make that very true statement you said that we often don't have a real full grasp on on who we are internally. I mean, I appreciate psychologists that use survey methods like questionnaires. That's real research. It has value, but never never fall into the trap of believing who someone says they are is who they are. It's who they believe they are. often or who they want you to think they are, depending on how transparent the test is like when you study Facebook data, what it becomes very clear, this is a little bit from a while ago when there was a richer, more natural data on Facebook, but what became very clear was people's Facebook profiles are who they wanted people to think they were. it wasn't so they'd like things That seemed like the right thing to like for their friends and followers, that's still information. I can still use that for research, but that's very different than a deep insight, as opposed to, you know, the the contrast in a nerdy researcher like me is, you know, what did what did you actually buy? what did you do when it cost you something? When you had a choice and there were costs associated with each, which one? Versus when it was just free and very public. So these are some interesting things. Can you find those moments? But thinking of, for example, a course on curiosity and how you really transform someone, there was actually a a a recent paper in Nature Human Behavior. I really liked this paper, although it was kind of small, needs a replication, all the classic caveats for me as a scientist talking to the public. THE HUMAN PROTOCOL (17:55.873) Yeah. Vivienne Ming (17:58.116) But in some ways it was so simple and spoke to how to change a person in a positive way. You know, we are, I was saying earlier, we are these, we're these much messier things. We're not just, in fact, just one person. We're kind of like, here's a really nerdy, geeky metaphor, we're kind of like a quantum mechanical waveform. We're all these different selves in superposition. And any given context, you become, you're not a totally different person, but you're a different version of yourself. that has some big impacts on, say, research into ethics, but let's just stay focused on curiosity. In this one paper, they ran a controlled experiment. THE HUMAN PROTOCOL (18:30.066) Yeah. N and one. Mm-hmm. THE HUMAN PROTOCOL (18:43.133) Okay. Vivienne Ming (18:57.4) So there were these eight to 10-year-olds that got a normal classroom experience and then randomly assigned. There were eight to 10-year-olds where their teachers were trained to instead of praising right answers, they would praise interesting questions. And then they let the class run. Not for a day or a week. They let it run for weeks and weeks and weeks. But already by six weeks, THE HUMAN PROTOCOL (19:14.353) Vivienne Ming (19:26.358) The treatment group, the ones with the interesting questions, started asking many more questions. And in standardized assessments of curiosity, the kids were scoring higher for curiosity. And when the teachers brought new science curriculum into the course, the question praise kids dove into it more deeply without any additional prompting than. completely random population of kids that we're still getting praised for right answers. And it's not that there's a fundamental divide here, but it's sometimes, you know, when I'm talking in a business concept, I say, how many leaders inside your company are praising interesting questions instead of just right answers? Like, who do you want the people around you to be? Who do you want your kids to be? Your peers, your employees? THE HUMAN PROTOCOL (20:14.482) Mm. Vivienne Ming (20:24.343) We might almost say your bosses also, like where are the opportunities for celebrating a leader taking a chance, even if it didn't quite work out. so you know, part of this is this idea of how do you change a person? If we're this messy mashup of cells, you start by changing context. You change the environment that they're in. We we have All of the right answers for free in our pocket. THE HUMAN PROTOCOL (20:55.461) Yeah. Yeah. Vivienne Ming (20:58.049) An interesting question, even if maybe it doesn't take us anywhere, an interesting wrong answer that maybe leads us to a new way of thinking about a problem nowadays is vastly more valuable than the same old boring right answer that anyone can get for free. THE HUMAN PROTOCOL (21:19.248) 'Cause the machine at the end of the day will continue to have all the answers available, right? That's where AI has access to all this vast human data that is out there. Who's able to get the more interesting answer from it by asking better questions and that's where curiosity will come into play? We'll have a leg up. We'll be the one that is probably gonna have better answers and better outputs to work with. Vivienne Ming (21:48.143) So let me actually give you some hard information here. So I ran an experiment last summer, and then had the chance very flatteringly to write about it in a Wall Street Journal article and and some other places. I had some very smart UC Berkeley students trying to predict the future. What will the price of oil be? Will there be a war in various places? Questions of actual importance. for anyone that cares about this sort of stuff, we actually took the questions off a website called Polymarket. So this is sort of a notorious prediction market. People bet on the future. Unfortunately, most of what they bet on, let's be clear, I mean millions, tens of millions of dollars is like which Kardashian is getting married next, or how many times will Elon Musk tweet in the next week? Lots of sports betting. I don't care about any of that. No offense to anyone, but the world doesn't need a prediction market on these sorts of things. but if you search around, you can find these actual meaningful questions. And I asked these very bright, you know, 20 year olds to make these predictions and they were terrible. Cause, you know, let's be honest, even a smart 20-year-old, can they tell me the price of a barrel of oil today? much less six months from now or whatever the time frame is. Well no, they were they can't. It's not I mean, they're not completely guessing. They don't, they don't, it's not like they say it's a dollar, but so they were terrible. We had four different little agentic models, Google Gemini, GPT, a couple of small open source models. The worst model on its worst question was So much better than the very best human being. So again, if that's the whole story, can you provide the right answer to what I'm going to call a well-posed question? A question we understand we know what the right answer would be or how to get there. Wow, the there is our big answer, which is there's no magic skill or knowledge that's robot-proof. But that's not what interests me. Vivienne Ming (24:11.19) I've been working since 1999 when I started work on my undergraduate honors thesis. I have been working on what I'm now calling hybrid intelligence. What happens when you bring the best of what machines can do and the best of what people can do together? and what we find is people are messy, so the answer's messy, which is to say, Even amongst my very bright Berkeley students, most of them just said, Hey, GPT, come with the answer. And then they just cut and pasted it and they were done. we called them automators, and somewhat scarily they kind of turned their brains off when they were answering these questions. Like I'm I'm kind of worried about that. And I see that a lot in students, I see it in the workforce, people just going from THE HUMAN PROTOCOL (24:46.539) Mm-hmm. Vivienne Ming (25:06.734) Brain activity that suggests they're working hard on a math problem or a great conversation to their zoning out in front of the TV. And the difference is they're asking an AI to do the actual cognitive work. and that's like 60 to 70% of the people in my experiment, which is, you know, it's alarming for a variety of reasons, including the long-term implications. the next group is something a lot of companies call for: help the Pope. Called for it. The next group we might call human in the loop. In my paper, I called them validators. In other words, the human says, hey, AI, analyze this question for me. Tell me what you think. I'll decide. The problem is they almost never did what the AI suggested unless they already believed that to be true. So in other words, They did they didn't understand the problem because the AI did all the actual hard work. The AI understood the problem. They didn't, but they were making the decision. Here's the interesting thing. They thought hard, they participated, but often human in the loop is worse than no AI at all. So right now, AI's the best. then the automators, because they're basically just doing what the AI is telling them. THE HUMAN PROTOCOL (26:06.858) Mm-hmm. Vivienne Ming (26:30.9) Then the humans by themselves, then human in the loop was the worst performing on average. So again, if that's the end of the story, we should all be concerned, you know, dystopia in our future. The smallest group was the far and away the most hopeful. It's the one that I'm really pushing for. I don't know if it'll naturally be the winner, you might say. but importantly, People do this already. The question isn't is it possible? It's how many people we can pull across into this. because I'm a bit of a science fiction nerd. I called them cyborgs. the reason I call them cyborgs is when you analyze their decision process, you can't tell who made the decision. The human and the machine made a unique decision together. Hybrid intelligence. Not artificial, not natural. Yeah. THE HUMAN PROTOCOL (27:06.291) How? THE HUMAN PROTOCOL (27:26.183) That's a hybrid intelligence. Got it. Vivienne Ming (27:29.946) And they were better than the best human. They were better than the best AI. In fact, all of my cyborgs were nearly as good as the people, the experts betting millions of dollars on polymarket that had time and access to data on the individual questions. So that's really helpful, right? There is a role for us in this future. AI makes us better. We make AI better, the question is how. And here's where I get to go full circle. It turns out all the standard AI benchmarks, how well can Claude and GPT answer questions all by themselves? And all these questions have absolutely well-posed known answers, right? That's why I had people predict the future. There's no right answer until the future happens. So you can't know the right answer. It's ill posed rather than well posed. Well, it turns out the benchmarks didn't predict anything. when the AIs are working autonomously, Gemini at the time had the highest benchmarks and it preferred the best. My little model performed the worst. but in hybrid intelligence, most of that went away. What predicted? It's gonna be a big shock to anyone who's been listening the whole time. Curiosity, fluid intelligence. Intellectual humility, I haven't mentioned that one yet, and perspective taking. If the AI gave you a good enough answer, did you keep looking for a better one? Curiosity. If the AI told you you were wrong, did you just collapse and do whatever it said? Did you completely ignore it and do whatever you wanted? Or do you say, okay, now I understand the problem better? What about this? And you keep exploring. THE HUMAN PROTOCOL (29:10.401) Mm. Mm. Vivienne Ming (29:28.929) intellectual humility. And perspective taking, if you've never heard it before, turns out I'm not the first person to find human beings that score higher in perspective taking do better with agentic AI. But essentially it's also related to theory of mind. Do you understand why other people are doing what they're doing? And it turns out the better you are at understanding other humans, the better you are at getting good work out of an AI. So these deeply human qualities, by the way, the same thing I found predicted career outcomes in 122 million people 10 years ago also predicts success and hybrid intelligence using AI. THE HUMAN PROTOCOL (30:17.43) A lot to unpack there. Vivian I'm curious because now when I'm when we're talking about this some sort of hybrid intelligence, right? Human beings plus AI machines. I'm curious to see where do you see brain interfaces playing a role in this future of these cyborgs, you know, that you are just mentioning because then Vivienne Ming (30:19.715) yeah. THE HUMAN PROTOCOL (30:46.35) Things will change as as we get more hybrid interfaces in human beings. Vivienne Ming (30:54.371) We are looking, you know, in this work, obviously these agentic things, you're interacting in a chat prompt or via coding APIs. it took me a while when I sort of moved out of my academic work and my first company was in educational technology, using AI to help kids learn, to really start thinking, wow, the interaction of the AI and the student, is a kind of hybrid intelligence. They are a cyborg, if only metaphorically. But my original interest as an academic, and one of my continuing domains of of purely academic research, is what are called neuroprosthetics. So you know maybe the most famous and I've I got to work in this space are cochlear implants. You have damage to the cells In your inner ear, call your cochlea, as it's called. And this is essentially a hearing aid that literally wires in to a part of your brain, the eighth cranial nerve called the auditory nerve. So if you're not familiar with it, it can actually restore hearing to people that have deafness in their ear, which is kind of amazing. And it's been around for a while and they've gotten progressively better. So part of my dissertation research was. THE HUMAN PROTOCOL (32:00.77) Mm. Vivienne Ming (32:22.349) Building an actual cyborg, which is to say, I got to build an AI, or as I would put it, a machine learning model that learned how to hear, given the constraints of a cochlear implant. And it turned out what it learned was very different, you know, than the way people were building cochlear implants, but it also meant an AI and a brain were actually interacting together directly. So no one correct me if I'm wrong. I just like to tell myself I got to build the world's first cyborgs back then. So it's that's just a very concrete example. It's a thing most, though not all, people can do. They've lost it due to age, let's say, and now we're able to restore the hearing in someone who has lost it later in life. Or maybe they had a stroke and they're partially paralyzed, and we want to restore a limb movement. THE HUMAN PROTOCOL (32:56.875) Cyborg. Yeah. Vivienne Ming (33:20.313) well, there's prosthetics for that or for damage to your eye. There's an artificial retina. My son has type 1 diabetes. We're building artificial pancreases nowadays to directly con interact with the body and release insulin just when you need it automatically, but with an AI inside of it, which coolly enough works better than an actual pancreas. I mean THE HUMAN PROTOCOL (33:44.391) Wow. Vivienne Ming (33:49.198) The reason why it's not better than actual pancreas is it's still a device that runs out of power, you know, you have to have it implanted. There's a lot of reasons why you wouldn't want it. But when it's working, it seems like it might actually work better than the biological thing in certain ways. but here's my real passion area: cognitive neuroprosthetics. Kids riding a bike down the street, out of nowhere, car hits them, traumatic brain injury. THE HUMAN PROTOCOL (34:01.266) It works. Vivienne Ming (34:19.469) The entire rest of the course of their life has changed. particularly something called working memory. how many things can you keep track of at any given moment? I I don't know if you remember the old when I was a kid, there was a a a it's crummy little game called Simon, where there'd be like lights and sounds, and you'd have to beep, boop, boop, and you'd have to play it back. and it'd get longer and longer. In psychology, we call that the Kosi experiment. THE HUMAN PROTOCOL (34:41.325) Mm. Yes. Vivienne Ming (34:48.513) And it's one example of working memory. How long of a sequence can you remember before it starts to get really hard to reproduce? And turns out for the average human being, that thing that doesn't exist, an average human being, that number is about seven. So this famous paper, the magic number seven, plus or minus two. I think that was published in like 49 or something, a long time ago, but super famous. Turns out working memory span predicts all sorts of things. about life outcomes. The bigger it is, again, doesn't predict as specific life outcomes. There are people driving Ubers with enormous working memory spans. But again, across a whole population, bigger working memory span predicts more higher income, more wealth accumulation, a longer life. Turns out people with larger working memories have better diets. 'cause it's also like the richness of an argument you can follow. The the THE HUMAN PROTOCOL (35:39.76) Interesting. THE HUMAN PROTOCOL (35:47.74) Mm. Vivienne Ming (35:48.76) The richness of your imagination, like all these are at least related neurocognitive systems. So on my website, way back when I was a grad student, it said, what if the magic number was 20 plus or minus two? So if you're curious, if you met someone who was a 10, you'd notice it. If you spent time on major R1 research university campuses, THE HUMAN PROTOCOL (36:11.791) Wow. Vivienne Ming (36:17.965) Maybe you meet a low teens person. 20 is off scale. but what if we could build a neuroprosthetic that would fundamentally change what it means to be human? I meant it as a provocation, because there's who gets it? Probably rich kids. Like there are huge ethical considerations here. But if you're looking at people with progressive neurodegeneration like Alzheimer's. Or kids that had traumatic brain injury. Maybe what we're simply talking about is can we give them back the life they already had? We're add a couple, they they were a seven. Now that damage means they're a five. A five doesn't mean you can't do amazing things with your life. It's just saying you're really gonna have to work harder in other areas. Boy, I hope you're resilient. I really hope you're curious because some of that raw cognition. THE HUMAN PROTOCOL (37:05.155) Mm. Vivienne Ming (37:15.189) Isn't going to be there to help, but what if you could give that back? so it turns out in some of the companies I've worked on, and one of my current ones, we look specifically at neuroprosthetics that can actually upregulate in certain different ways cognitive performance that produce measurable increases in working memory span while you're being stimulated. And That current company specifically, for example, works in Alzheimer's. we're also looking at stroke recovery and major depression recovery. So imagine these things that have this profound ethical c implication. What does it mean to be human in a world where this is readily available? I don't think many people are interested in elective brain surgery right now. THE HUMAN PROTOCOL (37:53.657) Fascinating. THE HUMAN PROTOCOL (38:00.633) Right. Right. Vivienne Ming (38:11.597) But if you have ALS or Alzheimer's or traumatic brain injury, what an amazing thing to bring this into the world. And then having said that, well, now we live in a world where there is AI that, on some measures, is smarter than we are for those sort of well-posed question problems. It knows more than we do. What does it mean as we start to merge hybrid intelligence? out of a keyboard and a prompt and into direct interaction with your brain. And I think the most important thing for anyone who's not building these things themselves is just, I'm no longer talking science fiction. No, this isn't happening in the world right now, but why do you think Elon Musk has Neuralink? Why do you think Brad Johnson has kernel? all of these people Google has a project in this space. Facebook has a project in this space. A lot of in my recent newsletter, I shared a whole bunch of research in IBCI. I stands for intracranial. So the actual electrodes recording activity are inside your skull. but the amazing part, to just add a little punchline, is is in one of those recent papers, this was in Nature. THE HUMAN PROTOCOL (39:29.803) Wow. Vivienne Ming (39:39.459) There's a guy with an IBCI who has profound paralysis. It generates speech, it generates typing, it controls a mouse cursor. He can interact on his own, not in a research clinic, in his home, totally transforming his life. And that's exactly the sort of thing. How is technology making us better? Not I don't care what it can do on its own. I mean, there are automate a way. Things humans shouldn't do. Automate away being bent over in a field picking lettuce 16 hours a day. I I don't have any problem with that if we have something for the for everyone who is feeding their family with that job to do. But don't automate automate away art and creativity. And guess what? Every job is creativity. Once you take away all the well-posed parts and it's all THE HUMAN PROTOCOL (40:19.348) Mm-hmm. Vivienne Ming (40:46.133) ill posed, every job is a creative job. THE HUMAN PROTOCOL (40:52.659) And what do you think would happen? 'Cause if if we are looking and you know, listening to everything that you just say right now, and all this integration, humans plus machines, and machines getting better and maybe smarter than we human beings, because I think that's probably what will happen. Are are you concerned or afraid that at some point we may be controlled by the machine and potentially lose some of what make us really us humans if we continue this interaction. Vivienne Ming (41:25.795) So Vivienne Ming (41:29.293) I it's almost like the nerd in me wants to break out different dimensions. Like there's more and less control, but there's it's not even more and less intelligence. Here's the thing I sometimes say. My first exposure to a large language model, LLM, was barred. This is one of Google's early attempts at LLM. OpenAI was already working on GPT, but GPT, Chat GPT hadn't come out yet. So you there were things like again before ChatGPT, there were versions of this you could interact with. I played around with Bard because as a scientist, there was something about it that interested me more than what OpenAI was doing. And the more I interact with it, the more I thought, wow, this is just feels a lot like my graduate students. So I would ask it really challenging questions. And it would give me interesting, seemingly right answers. Except it made my spidey sense tingle. Like, that seems like the right answer, but it also seems like maybe it's answering something I didn't ask. Like there was a miscommunication. So when I work with my graduate students who are brilliant, who are spending literally years researching one subject, they are genuinely a world ant. expert on that subject. And like once or twice a month I'm meeting with them. I mean hopefully more than that, but and I'm advising them why. They know more about it than I do. And if they didn't know, they could go to the library, you know, that thing people used to do. Or nowadays they could just talk on their phone. Why am I there? Because my graduate students know everything and they understand nothing. And I think a little known understood issue, maybe even among a lot of scientists, is our jobs aren't about teaching the knowing, it's about teaching the understanding. Our jobs with our PhD students. When you give lectures to undergrads, it's the knowing. Here are a bunch of science facts. but when we give when we work with our graduate students, we're teaching understanding because our whole job Vivienne Ming (43:57.899) Is what happens when there are no more right answers? When the where the research has stopped, there's no more facts that we've collected. How do you discover the next ones? knowing helps a lot, but it it doesn't allow you to explore the unknown, it turns out. So when I started interacting with BARD, and then nowadays with all of these existing models, wow. They know everything and they understand nothing. And for this current class of models, I will say this strongly, it doesn't matter how much we how much bigger we make them, they will never understand. That isn't taking away with what they can do. I use these models all the time. I use it in products inside my companies. I use it in my research, and I use it in my I use broadly speaking machine learning models in my philanthropic work all of the time. But I never mistake their astonishing knowledge and fluency for understanding. And so what I'm looking for is where, what did they produce that is useful to me? And what is a failure to understand? Which includes, if this is their answer to me, it's going to be a virtually identical answer to anyone else asking a similar question, which makes it. THE HUMAN PROTOCOL (45:10.288) Mm-hmm. THE HUMAN PROTOCOL (45:20.752) Two ask. Vivienne Ming (45:24.484) Not super valuable. I mean, it's valuable to me maybe in the moment, just because of what I'm going to do with its answer, with its knowledge. Now I need to add understanding. Where can we take this in a direction no one else in the world would take it? And that's, I think, part of what makes me worried about AI. If we talk about like not understanding dimension and controlling dimension, it turns out. An AI that understands nothing can still be very controlling. a paper literally just came out today in Nature Human Behavior, showing that the diversity of language and thought produced by humans using AIs is going down over time. because it's driving I mean, if most people are automators or human in the loop, it's driving everyone towards a s the same sort of not just answers, but the same description, the same words and the same kind of thoughts and metaphors. and it f turns out, again, going back to my experiment, but also a new paper I'm working on, our superpower isn't the right answer. THE HUMAN PROTOCOL (46:27.938) The same. Vivienne Ming (46:50.262) Our superpower is the crazy out there answer that probably isn't right, but might be. And if it is, it changes everything. That's where we are uniquely good against machines. And I don't, I'm a computational neuroscientist. On some level, I have to believe in theory, AIs can do everything we can do. In practice, no existing AI is in that space yet. And they they will never be just because we make them bigger or faster. They're still so useful. I wouldn't have based 30 years of my career on machine learning if I didn't find them incredibly useful. But it's never gonna wake up someday and say, you know what? I'm just a big messy guy, so I'm gonna root for, you know, Miami now. And that's just how it's gonna be. And no one's ever gonna get me to support the Seattle Sounders. What a horrible b foggy Midwest or Northwest. Why would I ever do that? So the you could ask it to pretend to care, but it's never truly gonna care. because it's just not part of its capability. Again, in theory, I think AIs are in theory capable of all this. What can they do today? And what are our risks? Our risks are outsourcing our unique creativity and then becoming borally the same as everyone else. Our risks is outsourcing our cognition. And this is speculative, but there is evidence. Outsourcing our cognition and actually increasing things like dementia rates and cognitive decline. And there are some other risks, like wouldn't it be annoying if you went to get a second diagnosis from a doctor, but it turned out they just used the exact same AI model, and so effectively you got the exact same diagnosis. so how do we make certain that that sometime frustrating variability of human experience that is actually a positive thing overall? Vivienne Ming (49:11.326) is still richly preserved. And to me, it's well, we need hybrid intelligence, we need cyborgs, we need humans to inject that amazing creative madness and the machines to catch all the times where we're making obvious foolish mistakes, and and add their value of sort of being right and that finding that healthy balance between them. THE HUMAN PROTOCOL (49:37.171) Mm. Vivienne Ming (49:40.395) And I'll even say this one of the challenges though is that humans are different. So we're gonna find different value, we're gonna interact with these models differently. So whether I'm building it directly into your brain, and yes, of course I'm a sci-fi nerd and I want to do that, or it's still something we just chat with on a phone. How do I build a model that gives you what you need and not just what you want? THE HUMAN PROTOCOL (49:51.666) Mm-hmm. THE HUMAN PROTOCOL (50:04.776) Yeah. And that would be the key to all this, right? And how we're gonna build that model where we still keep being this curious creature species that we are, right? That is not satisfied with the first answer that we get. 'Cause 'cause you know what also Vivian, what I'm seeing is a lot of organizations and people deploying and trying to adopt AI. It's just looking for output and output and increased capacity without even knowing if that output it's really, you know, understanding that is the right answer that you need for you. And and at in in the middle of it, we human beings are in the in this pursuit of doing more with AI that we may be risking becoming smaller without noticing that. And we're thinking that hey, we're getting all this output out and we are increasing our capacity. Are we really becoming better human beings while doing all that? Or just becoming smaller because we are relying more on AI? Vivienne Ming (51:12.439) Yeah. Vivienne Ming (51:16.386) So I you know the the the thing that I, if I may s put it this way, I like about my research is we look at these sort of different levers that are available to us. So I talked about the human factors. Now those are slow. You know, there are things we can do with people out in the workforce today that would help, but I don't I'm not I'm not a complete dreamer. I don't imagine somehow magically we'll pull THE HUMAN PROTOCOL (51:43.057) Mm. Vivienne Ming (51:46.227) Every working age adult into this creative economy and everyone becomes a cyborg. It just isn't likely to be true. But we can help. We we can make changes on the human side. And with kids, we could make huge changes that could pay off a generation from now, both in value to those kids, but also to society at large. There's the technology side. I referred to it abstractly, but what if you had a question for a model? Let's say what is this whole why is America and Iran? You know, for me, I was a kid during the hostage crisis, you know, in sort of 79 and 80. so this has like been my the scope of my whole life. But If you're going up now, maybe it just why do we seem to hate each other so much? Well, and and you go, well, it just gives you the answer. Well, here's the history, here's why. okay, I'm done. And also you it tricks your brain. I read it, so I know it. And your brain tr genuinely responds for most of us as though somehow we this thing that was given to us was in fact earned. But It isn't. So, what if, in asking that question, the model did two things. I want to give this person a useful answer, but I've also gotten to know my human being. And I also want to give an answer that makes them more resilient or lifts their curiosity. How can I phrase this question in a way which changes them over time. And interesting enough, what we found in our research is it ends up being a very different answer for everybody. you know the the curiosity one needs to encourage, you gotta stay away from that, I know it because it said it, right? You need to somehow say, what an interesting question. Vivienne Ming (54:12.152) Tell me what you know about this. So we actually ran this in my experiment. We took that worst performing model that I mentioned, the little open source model, and we trained it to never give answers. And we'll be clear everyone hated using this model. We called it Socrates, right? Because it only asked questions or gave context. Everyone that interacted with Socrates hated it. But it had the highest THE HUMAN PROTOCOL (54:34.592) Mm-hmm. Vivienne Ming (54:40.984) Highest average hybrid intelligence. Not because it made the existing cyborgs better, it more than doubled the number of people. So instead of like 5 ish percent, 20% of participants in my experiment switched into cyborg mode. Now, one out of five people is still a distinct minority, but if somehow I will say this, having started a lot of tech companies, there's a real sense in like the VC. World, the venture capital world of you got it or you don't. You're just magically good at this stuff, or you're not. Well, here's a case where, yeah, there's this group of exceptional people, one in 20, that went deep, did hard things, said, AI, challenge my idea. AI, I disagree with your idea, let's work it out. But it could have been one in five had the AI been truly. THE HUMAN PROTOCOL (55:30.742) Mm-hmm. Vivienne Ming (55:40.28) Trying to bring people along for the ride. and I bet it could be even higher had we started earlier with younger kids and built these capabilities into them, this foundational skills. So you can change the technology. It doesn't just give you what you need. Well, it turns out it can't just give you what you need because then people hate using it. Because it feels right, it's annoying. THE HUMAN PROTOCOL (56:02.912) That's that's that that's what I was hoping you get to it, because that's it that's probably the answer why the technology behaves the way it behaves today, right? Vivienne Ming (56:12.789) Why is Facebook the way it is, or Instagram, or these things where we God, I'm just I'm watching all these short videos and I know it's not, but I just keep doing it. Well, it's giving you what you want. And you have to be honest with yourself. You're clicking on it. You hate the politicians say terrible things online. Guess why they're saying terrible things online? I'm not letting them off the hook. I'm not letting the the tech companies off the hook. It's THE HUMAN PROTOCOL (56:14.975) Yeah. THE HUMAN PROTOCOL (56:25.076) Yeah. Vivienne Ming (56:41.975) it's an ecosystem. THE HUMAN PROTOCOL (56:42.982) It's a society conver exact conversation that needs to happen. Vivienne Ming (56:45.731) so the main area where you can make a change is what if we changed that ecosystem? You know, we changed the economics of it. Like we changed the economics in the classroom with the little kid, praising interesting questions instead of right answers. So now your AI, whatever your favorite one is, let's say instead of a benchmark for how good it can answer questions completely on its own, you know, well-posed, factually answered questions. THE HUMAN PROTOCOL (56:56.191) Yeah. Vivienne Ming (57:15.637) We do a hybrid intelligence benchmark. How well does it do making you smarter? And in this case, no such benchmark exists. There are some interesting tantalizing work, and my lab is starting to work on this as well. But there is no such benchmark. But wouldn't it be amazing to be able to say, our model? THE HUMAN PROTOCOL (57:21.596) You better. THE HUMAN PROTOCOL (57:38.024) Would be awesome. Vivienne Ming (57:40.665) Has the lowest all-cause mortality of any user base. Our people literally live longer, or our people are better at solving complex creative problem-solving tasks, or our humans have more friends. I don't know, whatever it is, but something meaningful about people rather than about the technology. Now you still have to have pretty amazing technology to do that, but what it means is. It needs a trade-off. So now I'm taking Socrates and I'm starting a visiting professorship and collaboration with a group at the in SIP business school at Paris, and we're trying to build a version of Socrates that actually builds a model of you first, and then essentially says, What is the best answer I can give that both makes you better short term, i.e., it's a correct answer, but also makes you better. long term. You have grown as result. So THE HUMAN PROTOCOL (58:40.764) Music to my ears, Vivian. Vivienne Ming (58:43.682) But it's turning out to be really challenging, which is, you know, again, not to let anyone off the hook. But why is no one building these models right now? Cause it's hard and the models are already expensive to build. And so no one wants to put in the extra work. But like I said up front, I just don't care about I don't want to build better robots. I want to build better people. THE HUMAN PROTOCOL (59:08.291) And it would be amazing, Vivian, if we build that AI that build better people and while building those better people, we build people that are able to find meaning, purpose in their lives, right? Really, as part of this conversation, to to discover all that. It would be amazing if we can get there. Vivienne Ming (59:28.702) I, if you don't mind a completely indulgent plug, yes, thank you for mentioning my book, Robot Proof, which you can go order right now. My genuine point of pride here. It topped Amazon's chart for psychology and humor, which really tickles me. I loved the humor part, but you have set up my next book that I'm working on. There's a poster behind me, it's called Small Sacrifices. And THE HUMAN PROTOCOL (59:33.089) Yeah. THE HUMAN PROTOCOL (59:45.815) And humor. Okay. THE HUMAN PROTOCOL (59:55.979) Mm. Vivienne Ming (59:58.777) It is a hard numbers computational scientists take on purpose. THE HUMAN PROTOCOL (01:00:07.033) Oof when is coming out, Vivian? Working on it Vivienne Ming (01:00:08.31) I talked about why I'm still figuring out the details. I it's been an an interesting thing interacting with the traditional publishing world when you also live in a world where you push a button and things just happen. publishing is an old business and AI is not. But in this case, this is an old thing. Life meaningfulness. you know, the nerds in psychology call it construct. THE HUMAN PROTOCOL (01:00:21.241) I see. Vivienne Ming (01:00:37.264) Strength of purpose. So it's not necessarily the thing, the thing you have a sense of purpose in, but your strength of commitment to it. And you know, I mentioned a lot of my research on life outcomes is looking at these big, really important life outcomes. And I said no single construct predicts at the level of an individual. This is also true of purpose, but if I had to pick one, you know. Resilience is a huge predictor of life outcomes. Working memory span is another big predictor of life outcomes. Interestingly, not as much happiness in that case, but uncorrelated with happiness. Purpose. There's a whole paper, a wonderful paper on life meaningfulness and life outcomes. Think of anything you would want to be true about you when you're 65. So much more likely to be true if you have a strong sense of purpose in your life. And so I decided to write a whole book about it, my research exploring it, just because it's another one of those things that feels so squishy and so undefinable. Well, how would a mad scientist like me try to give it a little definition, some measurement, and then at the scale of how Hundreds of thousands of people look at the differences between when our we are our worst self and when we are our best self and how purpose plays a role in that and some really concrete things like the fact that it predicts higher productivity at work and it predicts that you'll live longer. what an it's a it's an amazing thing. It just it I I and looking at all these different factors that predict life outcomes. I just sort of stumbled across it in my work. And after a little while it became so clear it just needed its own book. THE HUMAN PROTOCOL (01:02:33.398) Mm-hmm. THE HUMAN PROTOCOL (01:02:42.249) Love it. I'll be looking forward to the updates and hopefully getting a copy of that in the near future for sure. Viviane, I want to close the the conversation today. Yes, one hour went by very quick. And sorry for us being a little bit over of our schedule time. But after thirty years on all these working on all these inventions. And and all the exciting stuff that you you have worked on. What is the most human thing that you have ever built? Vivienne Ming (01:03:18.646) Listen, you mentioned two, and in some ways they're so obvious, but I I think they're the most human for very different reasons. Maybe the most obvious is I got to build the first ever AI for diabetes for my own son. But here's the important takeaway twist to that one. After I built it for him, I just gave it away. I went and gave talks, no patents, no licenses. I went and gave talks to every major d device maker and diabetes and I told them how to build it. I let them know I also told their biggest competitors how to build it too. and so when my son is having a hard day and he also has autism, so he has hard days, you know, we get to know that because this happened to our family, millions of people might be alive someday. so that's it's hard numbers. It's very human. and the other you also mentioned THE HUMAN PROTOCOL (01:04:13.641) Yeah. Vivienne Ming (01:04:15.714) But for a very different reason. I got to build an AI to reunite orphan refugees with their extended family members. I'm incredibly proud of that project. I write a whole chapter about it in my book. Let's just say the chapter ends with a twist, which is less happy and it's a little darker, but it was a huge lesson learned in my life about how to make choices and what really matters to me. Let's say it's in some way a preamble to my book about purpose. And so those, you know, even though I I never knew any of the kids that that was supposed to help, obviously I was incredibly proud of working on that, even though it has this less happy twist ending. because again, no one thinks about AI in this context. AI is just, how do we make businesses more efficient? THE HUMAN PROTOCOL (01:05:00.914) Mm. Vivienne Ming (01:05:10.232) you know, how do we make it cheaper to do complicated things? How do we use AI to make ourselves more human in the best way? That's what I spent 30 years working on. THE HUMAN PROTOCOL (01:05:27.901) Amazing. Thanks for sharing all that and all your work. it's been truly an honor. If people want to reach out to you, Vivian, what would be the the best way to connect with you? Vivienne Ming (01:05:39.702) I try to be super easy to find online. So if you're the kind of boring ass grown-up that's on LinkedIn, feel free to reach out. But if you want to drop a more direct line, visit socos.org, S-O-C-O-S, for SOCOS Labs, one of my labs, and you'll find my newsletter there. you can read part of it for free. If you're insane, you can also pay for virtually the same thing. But let's just say you're being a nice person who supports our research at SOCOS Labs. but you can also find links to drop me a direct message. I run a nonprofit called the Human Trust. We get to help people for free. And so if it's about people, please be honest with yourself. It's if it's about people and it's a real problem you would like to see me and my team work on, let me know. I pay for everything, whatever we invent, we give away. THE HUMAN PROTOCOL (01:06:10.255) Yeah. THE HUMAN PROTOCOL (01:06:39.138) Awesome. And don't forget to grab a copy of Robot Proof. I'm sure we all gonna learn a little bit about it. Vivienne Ming (01:06:45.834) I think you are now legally obligated to buy a copy of Robot Proof. I maybe I don't fully understand the law, but let's THE HUMAN PROTOCOL (01:06:50.01) Yeah, absolut I There is no sho there there is no choice left. Thank you one more time, Vivian. It's been an honor to have you on the show. Appreciate you, okay? Vivienne Ming (01:06:56.064) Yes. Vivienne Ming (01:07:03.104) A real pleasure. Thank you. THE HUMAN PROTOCOL (01:07:04.774) Thank you. Take care. Talk to you all next week.