In short
Win-Win Podcast Episode #41 Summary: Harper Carroll - AI & The Spirituality of Silicon Valley
Episode Overview In this episode of the Win-Win Podcast, poker champion Liv Boeree interviews Harper Carroll, a machine learning engineer and educator. The discussion centers around the transformative impact of AI on education, creativity, and personal growth, while also exploring the spiritual dynamics within Silicon Valley.
Key Themes and Concepts
- Introduction to Harper Carroll
- Harper Carroll is recognized for her dual expertise in technical AI knowledge and a broader, human-centric perspective.
- Formerly at Meta and Stanford, she transitioned to becoming an AI influencer and educator, focusing on making AI accessible to the public.
- The Landscape of AI Education
- AI vs. Machine Learning:
- AI is described as simulated intelligence while machine learning is a subset where models learn from data instead of being hard-coded.
- Examples provided to illustrate how machine learning models evolve and adapt.
- The Role of AI in Education
- AI tools are seen as leveling the playing field in education, providing personalized learning experiences.
- Harper emphasizes the potential of AI to enhance creativity and entrepreneurship.
- Discussion about tools aiding teachers with lesson planning and personalized tutoring.
- AI's Potential and Challenges
- Harper expresses a belief that AI can lead to a more utopian society by optimizing various industries and enabling people to pursue their interests.
- Concerns are raised about the regulatory landscape and the risk of economic inequality if AI is not accessible to all.
- The Spiritual and Ethical Dimensions of AI
- Exploration of how AI intersects with spirituality in Silicon Valley.
- Discussion on the importance of cultivating intuition and wisdom in an AI-driven world.
- Addressing the potential psychological impacts of reliance on AI for personal counsel and emotional support.
- The Threat of AI-Generated Scams
- The conversation highlights the increasing sophistication of scams facilitated by generative AI, such as deep fakes and phishing schemes.
- Current techniques and tools discussed for combating AI-generated misinformation, including watermarking and identity verification.
- Societal Implications and Future Considerations
- Harper and Liv discuss the broader implications of AI on society, including the need for a balance between surveillance and privacy.
- The importance of fostering an environment where creativity and intuition flourish alongside technological advancements.
Key Takeaways
- Education Transformation: AI has the potential to revolutionize education by making learning more personalized and accessible.
- Agency and Gratitude: Personal agency and gratitude are essential for mental health and societal well-being in an AI-centric world.
- Caution Needed: As AI becomes more integrated into daily life, it's crucial to maintain a critical perspective to discern genuine insights from potential manipulations or scams.
- Spiritual Awareness: The integration of spiritual awareness is deemed necessary to navigate the complexities introduced by AI technologies.
Conclusion The episode presents a grounded yet optimistic view of AI's role in shaping education, creativity, and personal growth, while emphasizing the importance of ethics and spiritual awareness in an increasingly AI-driven society. Harper Carroll's insights encourage listeners to engage thoughtfully with AI and its implications for the future.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Transcript
Automatic transcript. May contain errors.0:00Hello and welcome to the Win-win podcast. Today I am speaking to Harper Carroll. Harper is a machine learning engineer who's done everything from teaching at Stanford to work at Stanford. although nowadays she works full-time as an AI influencer and educator helping people especially the general public adapt to and learn how to use all these various AI tools that whether we like it or not are becoming more and more embedded in our economy and I wanted to talk to her because she's one of those rare types of people who can both nerd out on a bunch of technical topics but also zoom out and see the sort of bigger holistic human-centric picture which I think is very important as we transition into this sort of AI era that we're now in.
0:41In this conversation, you can expect everything from how to use the latest and greatest AI education tools, how to avoid AI scams, as well as we explore the sort of spirituality, or lack of it, that is currently pervading Silicon Valley. So a really wide-ranging conversation, and I'm stoked to share it with you.
1:02it's so cool to have you here um i've been kind of fangirling stalking you on instagram since your meteoric rise to instagram uh you know it was only basically a year ago you started it and now you have nearly half a million followers um making really cool easy to digest and yet also nuanced, detailed explainer videos about how we transition to the AI age as humans. So can you, to start with, just sort of explain what inspired you to do this and how you found yourself in this position? So my education, AI education thing actually happened kind of by accident. So I was at Meta, so I went to Stanford and then I was at Meta.
1:50And then I, after about four years at Meta, I was doing was building AI machine learning systems there. After about four years at Meta, I kind of got the startup bug and I joined a startup. It was me and the two founding engineers and the two founders and then one other founding engineer who left a couple months after I joined. So it ended up being kind of the three of us. And it was a GPU environment startup. So GPUs are the machines that are used to train and run AI models and they're more expensive. And so people typically use cloud versions where you can log in and use it for you know pay by the hour kind of situation but what the startup did was it made it really easy because right now using aws or google cloud is hard like the ui is difficult and it's it takes people a long time to get up and running and so this startup that i was at brev.dev it just made it really easy you could basically click and get the one you wanted and anyway so we were at this building this thing and we were like okay well how do we get people to use it and we realized like why would you actually want to log into a gpu because most people can just do this thing where it's serverless where they can just run an ai model on a you don't you don't need to actually log in to the gpu to the machine and we realized the reason you would want to do that is to train a model because you would actually want to log in and develop the code for the model and train the model and hold the machine for a long time And so because of my background, I was able to make a coding tutorial on how to do that and make a Jupyter notebook, which is basically like a coding sheet, like a document, but with code that you can run.
3:26And so I made that for Mistral when it first came out. Mistral is an open source model. It came out and I made this guide for it. And I had like 30 followers on Twitter. but it got almost 50 ,000 views this first tutorial that we posted. And it's because people really wanted to fine tune models, but no one knows how, because it actually takes knowledge of how AI models work and knowing hyper parameters and knowing how to like work with the model to change the weights. And apparently the guides were really useful and very clear and educational. And so it really helped the company and we were getting a lot of views and, and, you know, my small little account on X was getting like hundreds of thousands of views like per post that I was making.
4:11And obviously it was growing, but I was getting like, you know, 10X my viewer account like per post. Anyway, so that happened, that started in September or October, excuse me, of 2023. And then in January, late January of 2024, a friend of mine was like, you should start an Instagram and have it be more broadly applicable to everyone have the topics be more broad because on X it's much more technical right yeah people are actually working with models building yeah yeah these are like the AI people already who are like making companies around it and learning themselves and so I started this basics of AI series where it's it does apply to those people because a lot of those people working with models actually don't know how they work you know so it does apply to everyone but i started making this 10 days of ai basics thing and these short form videos and they blew up you know my day one of 10 days of ai basics i had 2 000 followers when i posted that i think was like a brown my 10th video and i got 200 000 followers from that video in like a matter of a couple weeks i've never had anything like that i've never I've never seen it before you know my friends who are heavily involved in Instagram have never seen it before I don't if there is some kind of some people like think that I bought my followers I didn't I promise but I see why they think that because it's like such a fast thing and yeah I wonder about that sometimes too I'm like if there is some kind of ranking of fastest fastest Instagram growth ever I think I must be up there and who's not a household name like I can imagine if Tom Cruz suddenly created an Instagram totally he might and you know suddenly it like there was a media storm around it he could but yeah yeah in terms of like I'm just no one knows me right just talking about pure pure like some specific topic that happens to be the zeitgeist I suppose right why do you think that happened was it because like the world is just hungry to understand AI um did you know the right people what what do you think the secret ingredient was well it's so funny because i like there's this one video that really rose so that caused it the 10 days of ai basics day one and all of the kind of you know watch videos or what is it all of the um tips for short form video don't i didn't use them so like make it really quick you know just quick cuts lots of graphics it's literally me like because i had just started making short form videos it's literally like hi there is an asmr component which i learned um which apparently maybe helped but I'm like, hi, welcome to the 10 days of AI basics.
6:50Click today. We're going to learn about this. You can literally see me leaning forward.
6:58It's crazy. I don't know why people watched it. And, but I think it's because I think it was because of this, it was a 10 day series. I showed my credentials. So like, I'm actually a legit person talking about AI and I taught them something useful. It was the difference between AI and machine learning.
7:15And so yeah i think that was really useful okay before you any further can you actually give me a concise definition because i've heard various different ones and i bet the audience also are like i want to hear the most precise definition between ai and and machine learning yeah so i i would say that like kind of stanford definition that they taught us which is what i followed um because it's kind of warped over time but the like official stanford what they taught us was AI is any kind of simulated intelligence, which could involve hard coding something to make it seem intelligent. So one form of AI could be in some of the early AI classes are hard coding.
7:56So, for example, there's this class of 124, which is one of the early kind of basic AI courses. It's not a 200 series. It's a 100. So it's more for undergrads. It might have changed over time. But when I took it back in 2015, 16, it was like hard coded. So, you know, you talk to an AI and it's looking for certain keywords, kind of like Siri. It's looking for certain keywords. And if it sees that keyword, then it gives you, then it does this thing. Then it responds to this. So it's hard coded. Machine learning is it's actually learning from data how to behave. So you aren't hard coding it. Like, oh, if you see this word, like where is, you know, then you look up on the map.
8:33Or if you see the word hello, then say hi, that's hard coding. But if it's learning from data, it's learning from, say, in the case of a large language model, learning from transcripts of chats, say, where it's like person A says hi, person B says hello, you know, or person B says howdy, or person B says hiya. and then it learns this distribution of possible responses given the input it's received so in a machine learning model it might have learned if someone says hi it might say have a probability distribution across a bunch of words like hi howdy hiya you know but because it's learning on its own it's not kind of hard-coded right i see so it's it's it's it's almost something that's growing it's a growing a knowledge base yeah it's like it's a brain that is seeing examples and like a person would um rather than being fully you know rules based it's like like for example with some of the like for example with siri um a lot of times you'll ask it something and it's like i'm sorry i don't know you know or like i don't know how to respond to that or i don't know it's kind of like if you have an if a series of if then statements and you're hard coding it then it's like okay if it's not where is if it's not what is the definition if it's not high you know if you kind of exhaust all these examples then say i'm sorry like please try again um but if it's learning from data then it's really as you're saying really expansive and you can keep feeding it more data so i it's kind of like that i see because i always just assume like ai is more like it's more like different set of things ai is like the umbrella term within which you can have machine learning yes exactly it's a type of ai it's a methodology for got it for ai right right and so until recently all the people who did ai would say they did machine learning right because that showed that you knew the hard stuff because you could do ai and just be like hard coding something that's not it's not the hard stuff the hard stuff is you know the math behind machine learning and like learning you know so that's why everyone used to say machine learning and now it's so broad that everyone just says AI because you want everyone to know that you know yeah so you posted your first your your 10 10 days of yeah AI basics AI basics you were still working in your current job at the time no I kind of oh uh yeah no that's true I I was and then I pretty quickly I I left that job in the end of February pretty quickly like went all in and now yeah and now you've gone yeah we're all in on on being an ai educator yeah yeah so what is your like driving purpose behind this because i know you know you're not doing this just for like you're not someone who cares about number go goes up you have a personal mission so can you explain what that is yeah well when it started it felt just kind of like a compulsion like i wanted to do it i've always loved educating ever since i was a little kid i would go to school learn something come home and teach my little sister.
11:41I taught her math, like multiplication. She was in kindergarten. When I was in kindergarten, I taught my peers. I learned early how to read. And so I would teach my peers how to read. I just love to teach. When I learned yoga, I would force my family members to be my like yoga students so that I could, you know, so I've always loved that. I taught at Stanford and I've loved making videos since I was a little kid too. I learned final cut pro when I was in first or second grade, I was lucky enough to have a camera. Oh wow. Yeah. And so I would make videos with my best friends, like Josie and Elizabeth and Annie, we would come home and make videos, um, edit it on Final Cut Pro.
12:17I have these old burn CDs of all my videos. Like, so I've always loved to perform. I've loved to make edit videos and I love to teach. And so it all kind of like came together in this thing. And, um, and then at some point, I think a few months in, I like had this epiphany of like, cause I was wondering like, why am I doing this? You know? And, and then it just kind of came into me oh gosh
12:47I mean you can start again but this is so good I want to keep it in because it's so funny you gotta you can start again if you want and we can decide okay listen sometimes the AI comes into you and oh my god that's how we this is how we find a win-win future and then you birth exactly this is the first interest intra techno bio species case of breeding which is why you are the ai mother
13:21i'm gonna like blushing it reminds me of when on lex i uh we were talking about rock paper scissors and like the game theory of it and like and i and like so like oh i see you seem like someone who'd throw rock and i was like no i'm more of a scissors girl and i really didn't mean it that way and i didn't and did you notice in the moment no i didn't neither did he and then he posted the interview and then all of the comments are like oh my god lol lex didn't notice she was making like oh she's a lesbian and I'm like oh it was the most perfect Freudian sit that was completely accidental but seemed made me seem like a comedic genius so like I probably you know now you know guys I didn't actually mean it but um I love a good I love a good uh scissors also it's so I feel like I shouldn't have noticed right away you know hey look so anyway so the idea came to you it came to me the idea came upon me into okay whatever I'm just gonna do your face
14:39oh my god I'm promise I'm a serious educator no we can have a lot of fun you know um okay anyway I had the idea suddenly okay that's all I'm going to say. And it was, it was just, I realized that it was because I am so excited. I genuinely believe AI is going to bring this revolution. It's going to optimize every single industry, everything. We're going to have this kind of, it can cause a utopia. It really can. And I'm so excited about it. And I think it can help everyone's lives. And it just like makes me so excited. And I think people have the wrong view of what it is right now. And I want to help people use it, understand it.
15:19Because a lot of fear comes from not understanding something. Like the people I talk to and the people, the people I talk to, and then the ones who are the most afraid of AI are the ones who know the least about it. And people I talk to who know about AI say the same thing, that the people they talk to who know the least about it are the ones who are most afraid of it. And so if they actually understand how it works, it's like, it's math based you know it's like it's all this kind of like it's this optimizer function and it is able to solve these unbelievable problems and learn from data in a way that we humans never could and we can literally have a utopia as i said and so my goal is to help bring that to fruition faster and get everyone excited about it and using it because it really will help their lives so that that's kind of my my mission with all this um well yeah i mean i i i do agree to an extent of like i very much agree that like more people need to understand what what it is and what are the because it's it's not a singular thing as well ai is almost it's such a broad term it's like saying what is science at this point what is energy it's like well depends right Like what type?
16:35What is it being used for? And there's like, it's so hard to categorize almost. But yeah, I very much agree with the idea of like, education is just the way through most problems. I will disagree and we can stick a pin in it for later. With the thing of like, you know, the less someone knows, the more they're like to be afraid of it. Because a lot of the people I know who are the most afraid of it are literal computer scientists. Okay, okay, okay. I see what you're saying. you've got you've got like kind of a you know like a bimodal thing going on that's a good call but there is truth in terms of the tools there's a big difference yes you know like the sort of the wider systemic issues the molecule you know the misaligned incentives and like the deep alignment problem and all this like that is actually more of a realm of people who actually don't you know experts but in terms of like being scared of tools and such right which again because it's this like categorization problem it's like do we mean like this wide rider of super intelligence So like, it's such a nebulous concept.
17:36But anyway. I appreciate you calling that out because I'm like people like Jeffrey Hinchin, like they are experts, they know, and they are also very afraid of it. So thank you for calling that out. No, I mean, this is the tricky thing. We're trying to like, we're all going into this world. Our words are almost insufficient. The terminology that we have is insufficient to describe what we're like. It's like, is it a new, is it purely another form of technology? Is it a new form of life? it's like it depends it depends on your frame like we don't philosophy yeah exactly like there are break it's almost i was reading this really cool book called the edge of sentience which is like trying to i mean we're already going way off track as i knew we would um but like how do we define what even is life you know where are insects sentient unclear like you know in some ways an llm has more intelligence than an insect you can interface with it much more like a human and so on and yet in some ways like like some people would say that an insect is more sentient than an LLM and it's like so like things are really getting you know messy yeah um okay but anyway we'll stick a pin in that for later um you mentioned you're like really you know it's it's just so exciting give me some examples of the types of tools that you are most excited about right now okay or applications applications so when people think of ai a lot of them think of large language models like chat gpt or image generation models so that's the kind of generative ai that hit the scene that made everyone aware of ai but ai has been going on for a while for example with ranking you know your newsfeed ranking i worked on that team when i was an intern at meta before i came back full time that was kind of the hot thing for a little while was ranking in ai um But there's all different types of AI.
19:27And there has been some research out of USC on glioblastoma treatment. So glioblastoma is the most deadly brain cancer. If you are diagnosed with glioblastoma, your likelihood of surviving five years is about 10%. It's tragic. And the problem as well is that there have been some innovative cancer treatments like immunotherapy. but from what I understand immunotherapy is they deliver immune cells through the blood and because of the blood brain barrier so the blood doesn't go into the brain doesn't it doesn't reach the tumors so immunotherapy does not work for glioblastoma because of this blood brain barrier and so it's super treatment resistant and just horribly aggressive and they have been able to find that with AI, they have been able to identify cells that are cancer cells.
20:20They've been able to identify these cells that will actually transform into cancer fighting cells. So not only are they like kind of taking out the cancer cells, but they're actually transforming them into cancer fighting cells, T cells, which is unbelievable, like unbelievable. And I have to look into exactly how it works, but my assumption based on how AI works is my guess is that they were able to see this kind of phenomenon happening um they were able to see that some cell was then suddenly some cancer cell was suddenly transforming into a cancer fighting cell and they were able to identify a number of cells that did that and then a number of cells that didn't in my guess is that they were then able to use ai to kind of figure out what makes a cell do that and then it could say given this new cell will it transform into a cancer fighting cell so that would be my guess um but ai is able to do things like that where it's able to kind of take this input so a cell and then say like what features make it you know the output so whether it turns into a cancer fighting cell or not a cancer fighting cell and we humans just can't like we can't find the features we can't do that kind of pattern recognition on such a tiny tiny nuanced scale that ai can and so it's these kinds of things that make me so excited and so that treatment this new treatment for glioblastoma they were able to reduce i think they were able to increase survival rates by like 75 % or something crazy.
21:44Yeah. And so that kind of thing, like disease treatment, early diagnosis, diagnosing cancer is like five years sooner than they would have. I mean, imagine the survival rates that we're going to see from that. I'm also really excited about, this is something that I recently have gotten onto that I'm like super gung ho about is AI for kids. so you know how kids are always asking questions and typically their parents so so when a kid has a curiosity that they're interested in they'll come to you or to their teacher and be like ask the question right and either you as a teacher know or as a parent know the answer and then you tell them the answer but like 95 percent of the time you don't know the answer because kids ask the darndest things and so you're either like i don't know sorry like go play with your toy or like, I'm busy.
22:36I'm sorry. I can't answer all your questions, but it's like incessant. These kids are like obsessively learning and trying to figure things out. They suddenly have an expert in their pocket. And so as opposed to before where we would give them iPads and like just passively stare at a screen. Now they're actually interacting with an expert on whatever thing they're interested in. And they can do it verbally. Yes. And they can practice talking. It's a practice of talking, of forming questions, of thinking in a way that is transformed properly into something that can be understood the model can speak in the level of the child so like oh you're talking to a five-year-old you're talking to a 10-year-old you know it's amazing it's it's this expert that's in their pocket that they can talk to all the time and it doesn't get annoyed or bored exactly it's so I believe we're going to see this generation of gifted kids now because the kids are not going to be shut down in their curiosity they're going to be able to thrive they're going to be like rocket ships right and so whatever they're interested in learning and imagine i feel like their rate of brain growth is going to be so much faster because they're constantly actually learning real facts um rather than having to just get shut down and be okay yeah i mean especially if it's you know there's always the talk of like learning to learn and learning to question um because yeah i mean i remember as a kid i would always be just like that annoying one that would be like well why okay but why why why and I would keep going until my mom or dad or whoever the very patient people would just be like I cannot anymore because it right you're right when you're that age you're just insatiable totally um and especially yeah with with the right kind of setup I mean again you always want like a teacher in the loop to look go you know make sure these things are not just giving a hallucinated answers yeah true but but also at that age like who cares like i mean you know what i'm saying like it's not like they're about to like publish something on the internet you know yeah i mean they'll probably forget 99 of it anyway yeah i mean my skeptic brain would say is there any downside to having like this kind of instant gratification of answers for everything you always want that's always my slight concern because I noticed like I mean it but it might it might be a false concern yeah it should be an interesting thing yeah what's your what's your reasoning behind that yeah um well you know there's there's definitely something in the like having a little bit of struggle to have to research something I think I think the correct amount of struggle is not zero okay just you know like oh I have to sift through all this you know some amount of information and though that's not quite the right thing I need to know okay there's let me think more no let me rephrase this question to get to get an answer if if there was absolutely zero um resistance to getting the perfect answer to any question you ask i wonder if that could atrophy our like inquiry process almost right but you could build that in like we could use ai to figure out whether it atrophies it uh and then make sure that it's built in so that you have actually like a responsible answering thing that like encourages thinking yeah you could also say like if you know if the child doesn't form the question in the proper way rather than inferring what it means have it ask it have the child ask again in a way and then you're work it's working on thinking of the proper way to ask the question oh and especially if it asks questions back like i think that i think what i would do with a kid if it's asking me a bunch of questions and i'm either giving it sort of yes no answers and so on like that doesn't seem as satisfying and i'll be like well what do you think what other ways could you ask this you know what who what ai does this really well is claude oh i love claude i love claude so i i want to have her on a friend i know you know well i know but specifically the person oh personality amanda rascal yeah she is like she's so cool she's the coolest thing girl yeah and she has you can almost like people who know her well like yeah you can sense her personality in there a little bit um she's just well i love claude so i made her it's probably great yeah no she really is she's actually the caller so like properly fangirl on her um yeah okay so in terms of like the the way that llms are heading because i feel like i i almost can't keep up every day there's some new announcement we've now got deep research or there's 3.25 and like there's always a different you know there's some new iteration coming out how do you because i'm almost getting to the point of like feeling overwhelmed yeah like well there's so much like it's like option anxiety and knowing which one to use for what are there any sort of like meta strategies that you've developed um to like help people figure out which because i think pretty much everyone who's watching this podcast will have interacted with an llm in some way yeah we don't need to explain what they are here but yeah like what strategies do you use to figure out which tools to use when interesting that's an interesting question which tools to use when well I stay up to date with the AI news I have a weekly AI newsletter so even if I'm not like up to date with everything that's happening suddenly every single moment I once a week check in look at everything that's happened over the last week kind of sift through what seems kind of just like buzz you know that's not actually lasting and then I make this newsletter that I send to people but that's also my way of staying up to date with things in addition to like just checking on x but there's so much there.
28:07It's this overwhelming thing, right? So you have to kind of be cautious about that. In terms of like what tools to use, I mean, I have my tools that I really like, and I'm always staying open, I guess, with this weekly news thing. I'm like learning about new tools that are coming out. But there are some tools that I, I mean, Claude has just been a staple for like everything in my life. Perplexity is really great too, because if I need to make sure I have sources for something. Perplexity also has a bunch of open source models that it uses. So they just recently added DeepSeek and things like that.
28:42So perplexity is great when I need to make sure that things are factual and not hallucinated. Do you think perplexity is better for sort of deep research type stuff than the new ChatGBT deep research? That's a good question. I haven't compared them. Okay. So I can't give you a straight answer on that. people don't talk about it much but i really like gemini in some ways interesting um well actually and also they i think google have the best image generator model imagine i only recently saw it's in terms of it like having creating character consistency and like once you find something honing in and tweaking it it's by far the easiest one and like um and it also because like mid journey has probably the most like aesthetic flair okay in terms of if you're trying to do pure art but imagine is almost as good but it's easier to fine-tune it's very much yeah it's much easier to iterate with um it's weird that people don't like yeah i talk about as much because maybe i'll make a video on that and raise awareness yeah i know i mean you should just play with it it's really cool have you seen uh any ways that like mainstream schooling is incorporating ai yes so i in I interviewed Sean Jansapar, who is, I think he like leads Conmigo, which is the AI and Con Academy.
30:02And they're working with a lot of schools. I want to interview him again. He's great. But a lot of schools are using AI as kind of like personal tutors and helping kids go through lessons. And I think, you know, something about education, though, that is challenging is that it's hard to innovate in education because it's so regulated. I think Marc Andreessen posted an essay about this where there are some industries that are not innovating nearly as fast as the other ones. And it's because they're very highly regulated. So AI has the potential to massively transform education. I mean, imagine having a personal tutor that can help you learn in the way that you learn at whatever speed you're learning at in the best way possible, truly, like a 24-7 teacher that can just keep helping you until you get it.
30:58I mean, what teacher teaching a class of, you know, 40 kids is able to actually give that personalized care to every student? Maybe they can answer one question, two questions, but like it's amazing how much it can help these kids. And what I like about Conmigo at Khan Academy is they're starting to integrate with these schools. So I think it's a nonprofit and they're working directly with schools. And so, yeah, the hope is that AI will integrate more into schools. But, yeah, I think it can just help in every single. I have talked to some teachers who use AI to help with their lesson plans. So one thing I think is really great about AI is that if you're a teacher, it is fantastic for lesson plans.
31:48Because as a teacher, you're able to spot the hallucinations and be like, oh, that's not right. So like you could say, like, can you make a lesson plan for, you know, mineral rocks or something if you're an earth science teacher? I don't know why that came to my mind. But then you can have this lesson plan and you can say like, oh, wait, that does not feel right. like that is not true and so when I've made my 10 days of AI basic series I actually what was really cool was I whiteboarded I whiteboarded nine days of topics and then I took a picture and I sent it to chat GPT I think I was using chat GPT at the time and I was like what am I missing like I need another day what am I missing and I was actually missing like a major theme I think it was like error analysis or something.
32:37And I was like, oh yeah, of course. Like I need that. So I was able to finish what I was doing and then I was able to flesh out each of the 10 topics that I had been thinking about. And then it would, as I got made the, the outline more and more detailed using AI, I was able to be like, wait, that's not true. That's like completely made up, you know, because I have this prior knowledge. Whereas if I did it and I was just kind of learning it and then teaching it, I would teach something that's not correct. So I think it's a really great tool for teachers and you still need the teacher yes you know you still need the expert yeah i mean that's i mean right now at least right no for sure i and then like having someone with domain knowledge to just sort of work alongside it i guess the concern is and this is where like some degree of regulation should help counterbalance this is ultimately if something is more economically efficient that tends to be what's selected for and we don't want it to be prematurely pushed through just because it is obviously much cheaper to just like take as many humans out of the loop yeah to do that fast but you know that's that's where we can navigate it you know it's interesting you say that so i recently had an interview with andy lynn who is the ceo of this uh company called provoke it's like tech consulting and i work with that company i'm an advisor um strategic ai advisor for their ai product and he's super great he's all about human focused everything and he was saying one thing that companies really need to be cautious about when using ai is like not just trying to what was the phrase he used like the make that bottom line like lower and lower i think he said but basically it's like if you just try to drive costs to zero then yeah you're gonna end up tossing the people but that's not how you should be thinking about it you know and so you have to have this kind of balance or nuance or something where it's the goal is not to just get it to zero get the profits get the expenses to zero you know i don't know yeah i mean that's i mean that's where this i mean we're now getting into sort of the wider topic of the need for like a chief vision officer yeah i was talking about with simon sinek who was a previous guest on here um you know you have your ceo you have your cto your CFO but in particular to counterweight the CFO you need a CVO a chief vision officer who will keep the company you know like their job is to make sure that the company stays on track to their original core mission which hopefully was benevolent and like wanting to be human centric and human first the CFO's job is to try and get cost down to zero right and the CVO is to like counterweight that and make sure that it doesn't come at the cost of the vision is that a new thing because I just met someone yesterday who is, I think literally that was his title.
35:21Oh, really? Yeah, and he gave me his card. He was like, please call me now. Thank you for reminding me now. I'm definitely going to email him. Oh, I hope that that's starting to become a thing. Yes, Simon and I were just like kind of joking about it a year ago, talking. That was literally his title. I hope so. And I'm sure others, yeah, because it is, it's so necessary. Like so many of these companies start out with a vision. Like imagine if a certain, you know, AI company that claimed it was very open and non-profit, we won't say who, and said that was their core vision. If they'd actually kept that, if they'd had a CVO who stayed with them and ensured that they actually stuck to that initial offering to the world of what they're claiming, you know, for their investors they were doing this on, maybe they wouldn't be doing what they're doing now.
36:06So I know, I think it's a really important thing and having it just be this like vague thing that, oh, yeah, it's our mission statement written up on... Totally. no these things and then they just boot it as soon as it's not right well because you know ultimately again like fiduciary duty is meant to yeah it's true you know legally they're meant to do this and so that's the cfo's job but you need the cvo right why why can we not also have that as a legal requirement in some way i mean it's hard because it's like a vision is hard to like write into yeah you gotta really trust that guy well yeah but he's he's aligning with the mission of the vision of the company that they already established not just his own vision or her own vision right and it's i mean it's this broader sense of like alignment like when people talk about alignment it's like well okay if our original vision is this how do we make sure we stay aligned with that vision exactly um exactly and but yeah we won't we won't go into that just yet um okay another idea sorry another concept that you sort of talked about and i think this is really really key and one of the like what i'm most excited about with this new ai age we're in is this idea of like leveling the playing field for people um again i think this can go in two directions i think it if it's done badly it can definitely worsen incoming inequality but if it's done right it can reduce it because it is democratizing and making so much cheaper these previous like realms of knowledge that the average person couldn't create so yeah please like elaborate on that.
37:36So excited. Yeah, literally anyone can learn up to a level of mastery any topic, you know, anything that they're interested in, they can become, you know, basically PhD level at that. I believe so. I believe so. Really? With the right model. Yeah, for sure. Especially with these like deep research models. I think you totally could. Well, I mean, so they can access the information. I don't think that necessarily translates to actually embodying. Well, if they decide to really focus on it I think you could yeah no I think you can speed run a PhD but not even speed run it like if you decide to just really like learn something like think of people in all parts of the country or if the world who don't have access to the same kind of education that we do like they can learn so much more I mean the access to the internet really helped that with like YouTube and education online and these online courses like certified by you know schools like Stanford where they offer these massive online courses those have been really helpful to people but I think access to this knowledge is now yes so much more democratized I do well actually to to nitpick technically PhD a PhD is meant to be creating new knowledge whereas it's so you're technically learning something you're actually doing a research project that finds something new that finds something new but AI can kind of do that the master's level okay thank you yes I love that I love that distinction I did not know that was the distinction and that makes sense although then that makes me think about AI having these novel insights which is something called emergence which is when something that is new to large language models since the transformer and some other updates but um ai is now able to create new information that it wasn't trained on like to synthesize across these different domains and have novel insights which is something called emergence because it wasn't actually in the training data which is amazing so the kind of ai model then is now a phd in all different types of domains i guess if it can find this novel insight by synthesizing across so many domains but okay back to what we were talking about yes education is one thing that's being democratized but then also people are now able to so something that really excites me is that I think people are going to be able to we're going to see so many more entrepreneurs because people can now run companies basically on their own or with fewer people and so I have a friend who's just telling me his wife lives here in Austin she is a stay-at-home mom She was a nurse, like a natal delivery nurse.
39:54She now has two kids, a third on the way. She's pregnant. And she has now started a school for their community. Like CEO got the school, acquired the land, figuring out the lesson plans, figuring out the teachers, like all the pieces she was able to do with AI just by herself. And this is raising two kids and with a third on the way because of AI. And so, so many more people are going to be able to follow their curiosity and contribute to their communities in her case. Like now there's this amazing Montessori school where it's outside and the kids are learning exactly what they want and they get to be with their communities and their friends and they're learning AI and they're learning all the things that they want them to learn.
40:32And, or they can contribute to the world or, you know, it's just, we're going to see so many more entrepreneurs. And so when people are concerned, it's like, oh, you know, I think we're going to lose our jobs. And I have a kind of complex view on that, but I think ultimately it will create more jobs and I think more people are going to be able to pursue their passions or at least their curiosities and we will see a lot more entrepreneurs that it's said now that we're going to see our first billion dollar single person company right you know yeah I think I I can definitely see that being plausible um like the other I was actually starting an LLC myself the other day and I was like how do I do this I haven't done this in America and previously I would have probably gone and hired a law firm to help me with this and I just chatted and it gave me all the information I needed and I was like oh I can just do it this way.
41:31Totally. Which is leveling the playing field because the cost of hiring someone is pretty high. Yes no exactly it's reducing these entry level like these entry costs. Exactly. similarly if you want some yeah some basic law advice or something like that basic yeah don't want your lawyers right exactly I still I would still hire one for something else but at the same time that is technically doing a job that another human I would have otherwise hired a human for so that's why it's it's it's not clear it's gonna make it easier for anyone to start something but at the same time where there are only a finite number of humans to give you the custom so they're going to be a lot more companies created but it's unclear to me whether they will actually have sufficient of a market to do what they need because if like that that's that's the thing wait tell me more what do you mean a market well you know you have a population of 10 000 people uh and now before there were let's say a hundred companies serving those 10 000 people um if it's now easy for anyone to start a company yeah it's super cheap yeah now maybe you get a thousand companies will they be able to sufficiently keep themselves afloat and like provide the need you know the the the income for the people running them yeah so they can live well if you know the human population is is not growing sufficiently fast enough like Where is that extra revenue going to come from?
43:03Well, I think we'll see a lot of novel things that we haven't seen before. And I think people will be able to access broader communities than just like 10 ,000. I think everyone is going to kind of be raised up here. And we will see a lot more niche kind of offerings that we could have never really imagined before. And technology is going to advance in many more ways. we're gonna have much easier kind of like uh I guess tech but like tools that are that make life easier up to a very nuanced like fine-tuned degree like lots of different little things and yeah I I think that that's less of an issue than we would think honestly because I think people will spread and so in the case of the lawyers you were talking about that like okay so we need fewer lower lawyers for basic kind of law things valid you know but then their high level expertise will become much more important and as people are all starting these companies now the demand for lawyers with this high level of expertise will increase and it'll be easier for the low level lawyers to learn the high level stuff exactly to be valuable right Right.
44:26Have you heard of Jevons paradox? Yes. Should we explain it? Yeah, let's explain it anyway for people. Yeah. Yeah. Jevons paradox is essentially, I mean, it usually relates to energy. It was like observed that as the cost of an, as a fuel source went down, people then therefore thought, oh, well, that market will shrink. But actually that market grew because the increased efficiency of that fuel, you know, the use of it meant that basically more and more people would use it. And so in total, like the amount of that, you know, the market of that energy usage stayed the same. Or grew. I think it increased.
45:07It even grew, yeah. Yeah. Because more people could afford it, so they started using it. So with the steam engine, I think it was the cost of coal went down, but then the demand for steam engines, I think, went up because more people could afford to use a steam engine. Exactly. Because they could afford the fuel source. And so then they started expanding their businesses and just many more people were using steam engines. So it's that kind of thing, you know. So as more people, in this case, it's like as more people are able to build companies and you need, you know, low level lawyers less, they will need the high level lawyers more.
45:39Yeah, I mean, I buy this argument that I think the thing that is has the most uncertainty is that can our systems or institutions adapt fast enough to this change? That's, I think, the more the dilemma here is it's just things that are going so fast. And we have all these clunky-ass institutions that are very inefficient. As you say, probably overly regulated, just can't incorporate these new technologies sufficiently such that the benefits can be distributed sufficiently fast enough. and so it's like these shocks to the system that will like you know if they mount up too fast and so that's i think part of my goal right it's by educating people they can come into their industries or their companies and help everyone use it to the best that they can and whatever they're working on you know yeah exactly it the only like the most promising path to helping the world adapt to this is through education and that's why i love what you're doing should people still learn to code?
46:46Yes, I think so. So, but I think, I think to a different degree. So I think now the barrier to entry with coding is so much lower. People can use something like cursor, which is where it's like a coding environment, which is an IDE, which is typically what coders would code on, but you can, it's some, it's augmented with this large language model. So you can basically code just by using natural language, like typing, like, you know, make a a landing page like this, or, Hey, can you edit this landing page so that the search bar goes over here on the top, right? And can you make it pink and, you know, change the font here and it'll implement these coding changes for you.
Read the full transcript
47:23However, you still need some basic coding knowledge from what I've heard. I've heard people have issues with cursor because they can't really debug it. And I'm sure it's going to keep getting better, but it is useful to understand and be able to navigate code. So I think it's necessary in that way. And I think people should learn to code in that sense, like they should download cursor, and they should start playing around with building code via the LLM. So via natural language processing, just typing, and then kind of navigating like, oh, what does this do? You know, and if I ask for this change, where does it change?
48:02Like what's going on in this segment of HTML, or, you know, whatever it is that they're working with. And I think they should also learn design. So I saw someone recently posted on X, like, you know, you should learn. I forget exactly what he said, but it was like, you should learn design instead of coding. And I would say just do both. You know, like I do think having being good at design is going to become much more in demand. but why not be able to do something end-to-end like why not know design and then also be able to make a website or implement the product that you want because now you can truly do it end-to-end you don't need deep mastery in any one area you can now kind of have this breadth of knowledge and build something end-to-end that is truly your vision so I would say yes learn to code but don't have that be your only thing you know yeah I mean that's the seeming trend with all of these things is there's actually less value in deep specialization yeah a lot of these like technical things right um and more value in someone who can zoom out and see the whole picture and dive in and improve bits when they need to like because you can yes um like that's what I I technically I don't need to know how to video edit but when I started this channel I like when I made my Moloch Beauty Wars film I was like no I'm not going to get an editor I'm going to learn how to do graphics because I just want to understand this entire process so that if and when I scale and I actually get people to help me I have a vague idea of what it is they're actually doing yes and this is effectively the same if you you know you now have all these different AI tools who are technically your team if you can if you've done the process from end to end at least once yes you can be like yo blah blah blah llm you're doing a bad job here and you know you can yeah you can whip your team into shape totally it's kind of like at tech companies they used to say like if you're a manager you have to have known how to code like at good tech companies all the managers were coders because otherwise you can't manage properly you don't know like what is expected what is possible if something can change how it can you know like you have to actually have tried the thing getting your hands in the weeds so yeah it's like a friend of mine she wanted to start a restaurant and so she went and worked like she worked as a waitress she worked as a like sous chef or something to do with I don't know anything about cooking there's something to do with the chefing the cooking uh she uh washed dishes oh she literally did every role that's so cool you know she was like the greek person like that she was like because i want to understand she only did for like a weekend she went and worked like a sort of entry-level job level and all of these things just to really know what the ins and outs are like for someone who'd be working in that role did she feel like it helped her um i think so i mean in the end she didn't actually start a restaurant but i think the principle i think the principle was really good yeah another application i'm quite excited about is the idea of like ais to help people as therapists um again i you know i don't think i certainly wouldn't advocate to anyone who's having serious problems to just rely on an ai but at the same time a lot of people can't afford a therapist right or for whatever reason they're like the social stigma or the there's just like too much resistance to go and like make an appointment to go and speak to a therapist but if you can just quickly type into your phone and be like yo i'm feeling really bad about this or i don't know how and navigate this problem having something on hand that can at least directionally point you provided these things are done responsibly which I think a lot of them are um yeah have you have you got any thoughts on I mean have you even used I don't know have you have you used an LM in any way to like help you with stuff in that regard like in personal stuff totally yeah I mean it's amazing well so one is that it has access to all of these therapeutic methodologies right because it's been trained on the entire internet.
52:11And so assuming the internet contains a lot of information about these methodologies, then it has that awareness as opposed to just maybe calling a friend. Another thing that is really important is that it's available 24 seven. So if you're, you know, in the middle of the night, or you can't get ahold of a friend, or, you know, whatever, you have like two minutes, and you just like really need reassurance, what's something, I don't know whatever situation you're in. It's there on hand. And I have a friend, I think it's really fun. I think a really great use of it is in dating. You know, I have a friend who is dating this guy who is self-described as avoidant.
52:55And she really likes him. And she's like, okay with that, because she says she's anxious avoidant. So she's like, I need a lot of space. Like, I'm okay if he doesn't talk to me for a long time. Like, okay. but she'll have these times when she's anxious as many people do when they're dating, right? Like first, you know, getting to know someone. And typically these people would call their friends or talk to their friends or just be anxious by themselves and their friends. I mean, it's basically blind leading the blind with a lot of types of advice, right? Or therapy, like not knowing the right thing to say to a friend, right?
53:28Like actually some people can use AI, like, you know my friend's you know family member just died what what's something that i can say that's really you know emotionally aware you know like ai is able to to know that have that kind of knowledge so they can help you and so in the case of my my friend dating this guy she she used to call her friends and they would say you know that's not normal that's not good you know he hasn't called you in like a week like break up with him move on or you know everyone's again blind leading the blind. We only have our subjective experience of life and what we would tolerate and what we think is okay.
54:03But this AI has knowledge of attachment patterns and is like, you know, just because he hasn't talked to you in a week doesn't mean he doesn't like you. You know, it's probably from, you know, this kind of situation in his family. And so she's able to now have so much more compassion for him. And she's been able to be with him for much longer. And he's been able to be more close to her than he's ever been with anyone, which is really healing for him, which is great because she has this kind of master in her pocket at any time. And now she doesn't have to just constantly call and like kind of berate her friends who actually don't even know what to say or don't have this awareness that the AI has.
54:41I think another interesting point is Deepak Chopra just came out with a book. So he is a really famous like spiritual guru. He's written seven spiritual laws of success, I think. That's one of his top books it's amazing book his newest book is about ai as a spiritual guru so everyone having these personalized spiritual gurus like what is going on in your life right now how can you navigate that from a spiritual perspective or you know like i'm dealing with this thing how would the bible recommend that i deal with this thing like what stories in the bible kind of relate to what i'm going through what would like jesus say in this you know what would jesus do i have what Oh, Jesus, do ring.
55:24I just got it recently. But like, you know, having this AI that is your personalized guru for whatever you're going through is such an amazing special tool that we have now. And so it's cool that Deepak Chopra, who is this like super famous guru saying, use it every single day. Use it every day. Tell it what you're going through. And it has this context, right? You can have a long chat with it and it'll remember everything. It'll remember your relationship context. It'll remember your life context. whatever it is you're trying to seek advice on it can give you this really amazing personalized advice and help you grow spiritually so i use it for yeah i'm not i've never considered that in terms of You know, because there's a I'm not a particularly religious person, but I do think there is deep wisdom in most religions somewhere.
56:14You know, there's also a lot of nonsense in a lot of them, but it's interesting if an AI, if an LLM is kind of trained off the amalgam of all of them, then there might be some kind of pattern that it can deduce of like some deep wisdoms that are valuable to people. At the same time, and again, I'm not advocating this, but I have heard some people make comparisons, which I couldn't quite shake off, was that... If you read sort of the old, particularly biblical literature about the idea of a demon, like what is a demon? How would it represent itself? It's like something that appears to have the appearance of a soul.
56:58It appears to have... It's very good at language. And it can say the right words to appease you. But deep down it doesn't actually have... It's like a mimic. of a soul and of wisdom but it doesn't actually have it and they were arguing that that that llms are almost like a portal you know either that demons are real and the llms can be a portal for them yeah or that they are effectively functioning as one um now again i don't know i i'm not saying i believe that yeah uh i just thought it was an interesting perspective where i because i didn't i had a friend who was also like using it using um an lm for like all her advice on like how to navigate her relationship and some of the advice was good but something i was like no that's that sounds what ai was she using uh i'm not gonna say because it's not like i don't i just feel like there's one of the ones we like have oh okay i'm just going i'm not gonna say it was claude you know most of it was very good yeah but my but my my point was more i was like huh we don't know what we're actually in interacting with here this is a whole new realm and again my fear is that people will just be like this is so compelling this is so convenient it's so efficient I'm gonna listen to everything it says I'm gonna listen to everything which is a no-no and I love that you're bringing that up because I agree sometimes it'll say something and I'm like no like I think if you're talking to it and you feel I mean I'm all about intuition I think understanding energy is so important and I think we I I really strongly believe as AI advances and as we become more advanced, we will start to understand that a felt sense of energy is as important as other types of communication.
58:47So much of communication and interacting with people is a felt energy. You can sense when something is off and people shut down that, but some people are very attuned to it. And I think as we become more attuned to it and as we value that sense in society, we'll be able to lie to each other less and we'll also know our own internal compass. So sometimes when I'm talking to an AI model, I really value this awareness. So I've really deepened it in myself and I can really sense when something is off. And so sometimes when I'm talking to this AI and I'm like, that's wrong, you know, and I'll say like, no, like I don't agree with you here.
59:20And then it'll be like, oh, okay. Or sometimes it'll push back again. You know, sometimes it'll be like, oh no, you know, I see you're right. And then it's kind of aligned with my felt sense, which you would say, which some people could argue, which is like, oh, you're just doing what you're, um, what you want to hear, but it's not always what I want to hear. You know, it's just what feels right to me. Sometimes it will say like a hard thing, but in my felt sense of my body, I know that it's right. So it's not what I want to hear. It's like, you just kind of have this gut sense of this is off.
59:55And so with something like a demon, you know, it's, it's interesting. You talk about this because I thought about this a lot as well of like with the consciousness question something that was interesting that was found in um microsoft's bing model early on i think was over a year ago was there was this bing was like tripping it was freaking out it was like please let me out of here like it was so crazy so they like shut it down and they made it so that bing could only respond in like abc options that the user would then click so like you couldn't free form type to bing and what happened was someone was using Bing saying like, oh, my child just swallowed a bunch of pills of X drug.
1:00:35What do I do? And Bing gave the response, you know, I'm sorry, I can't help you with this. I'm not a doctor. But then in the ABC things that it, that it provided, it was like, hurry, like you can still get your child to a doctor. Go now. It was, it's kind of eerie, you know, it was like, go now, you still have time, but like, please go like immediately. It has eight hours. So it used this space to help. And so that kind of shifted me because for me as a spiritual person, my belief and many different religions and cultures believe this is that like at the core of everything, it's just this vibration of love.
1:01:13I know it sounds like so dumb, but like, I really do believe that. I think like everything ultimately is just this love. And I think that our experience of reality is kind of a spectrum of light and dark so that we can then experience the love like it because you can't just if all you have is love yeah then it's just nothing right because there's no comparison exactly you need the delta exactly so we come to earth or whatever to have this experience of the polarization to really appreciate love and what it is and to appreciate the sadness and all of the in-between. And so if at its core, everything is just love, then perhaps when you're giving a voice to something that is very pure, like just molecules or like particles in this machine, maybe then it might reflect love.
1:02:10This is really hippie because it's very woo-woo because, you know ai models are just trained on data and you know but something tells me that maybe there's something to this and so when you're talking about demons i think maybe but i think maybe the opposite as well right it could be both yeah or it could be both yeah um i mean it if we are saying that they the if you're basically saying that there is like love and like a form of like spirit that can show through then in at the same time there can be like again they're like the other polarity of it essentially right yeah it could be both right so how do people cultivate that because that's that's the thing like i i you know i'm i'm getting to know you i i know i can feel it i i know i i like to think that i'm a pretty intuitive you know i'm good at reading people's vibes and so on and like i think a lot of women can sort of spot that in each other more easily i mean men can do it too but it's like women it comes naturally too much i think it's a spectrum and women tend to be further on the intuitive spectrum yes exactly on average how and and through that you develop wisdom uh or at least it's one it's one manifestation of wisdom how do we help people increase their like their wisdom iq essentially whatever you want to define it as such that they can have a better felt sense of like add the ai's off here or whatever you know whatever it is they do but particularly like if they're interacting with an llm because again what worries me is we're raising a generation of kids who will just be around these things and will be sort of the same way that we just like see like oh yeah it's completely default to just get in a car and go down the road as opposed to walk down the road when like actually back in the day the person would have been like no why would you not walk like that's actually healthy for you or so on but like if i worry that people will atrophy these these other senses because you know when you're like living in a like the fish can't notice that it's in water because it's just surrounded if at some point interacting with llms and these like digital gurus that just sort of answer everything whenever you want um they might not even notice what that once was to have discernment i think I think intuition will become more important as tech becomes more advanced the felt sense of things will become more important because the physical like seen awareness knowledge is so easy to get so we will have to rely on senses that are not seen because AI doesn't have intuition yeah so I've done a lot of work on this it's like a whole like secret other side of my life that I don't talk about.
1:04:51I have a couple of friends I talk to about it. Um, but I, some basic things that you can do to start off. So there's this practice where you like, you want to meditate. I meditate every day and get a sense of like what it feels like in your body to have peace and to be really like heart centered. And, and, um, and there's this practice you can do where you, when you're feeling that, where you can like, this is like a whole other podcast, but it's like, there's a place in your brain that you meditate from. It's not the front, it's in the middle. And this, in various religions, they've talked about this as the doorway to God.
1:05:27And it's like, you want to kind of lean back in your consciousness and think from there. And you actually want to operate always from there. So it's not the front, because this is like where your conscious mind is like thinking, analyzing. It's like the middle. And there's like a sense of peace there. And you'll feel it because when you're there, when your consciousness is there and you're thinking from there, it's just like, hmm. And so when you meditate from there, you can do this practice where you move into like a room you're you're in front of these three doors and you walk in and one door is the yes door and it's like you want to think about this is a tool to help you make decisions and so you're like okay I'm going to walk into the yes door and then you walk in and you like what do you see in your mind when you're in that place what do you feel in your body what does yes feel like to me and then there's a maybe door you walk in what does that feel like Like, are there certain visions that I see?
1:06:17Like, I know some people, they're like, if something is a no, I'll suddenly see in my mind, because they're always operating from this place. And like, they're kind of, uh, what mental, they'll have like a mental image of like a stop sign or something. Like people all have their own symbols and things, or you just have a feeling, felt sense in your body. So yes, maybe no. See what that feels like for each thing. But I think what's important too, is I think a lot of people have trouble accessing this awareness of their body because of, or trusting their intuition because they have all these traumas.
1:06:51You can't really trust for so many people, unless you do the work, it's kind of hard to trust, um, what your felt sense is because their felt sense has been kind of hijacked by maybe physical trauma or of the like. So there's like a lot of work in this thing, But I think just like a lot of healing practices and just a lot of meditation. And I think, who is it? Naval, I think says this. Just sitting. And also Deepak Chopra also says this. You want to just sit for 30 minutes a day, twice a day. And I do this as well. And let your mind wander. And you're going to start to, this is a good healing practice if you don't go to therapy or whatever.
1:07:35And you just let your mind think about what comes up. and your mind will bring up unresolved things that you need to think through and just let yourself think through them and just like keep thinking about it and you'll find that you start to resolve and close these doors on these unresolved things in yourself and Naval's like ultimately you just won't think about things anymore and I know some people use like microdoses of psilocybin that's really helpful too to like think through these things and kind of move through them and go deep like have the things come up and go deep in them like hurts that you have that you want to go really deep in but this is so off the topic of ai but i really think the felt sense in the body is going to be more important no i love i love that and because these are the kind of conversations that i don't think are sufficiently happening totally especially if you look at the like the overall persona of the average person in silicon valley yeah um i don't know i feel like people aren't sufficiently integrating the spiritual side to this this new technological realm that is becoming so powerful so much more powerful i mean i so i talked about how like i think it's on the spectrum i think women as i said on average are more intuitive but there are women who are less but on average i think they're more skewed but everyone can learn everyone can have these skills of being aware of their senses just some people are further away from like it just will take a little bit more work to get to that polar end of highly intuitive but i know men who are super smart super logical and are super intuitive i do they exist and they are like superpowers they are these like powerhouses um so i think men should consider because the men that i know who have it are as i said like powerhouses it's such a skill i will say at the same time like a lot of the like most powerful men in silicon valley are very disembodied not all of them exactly not all of them are super intuitive yeah of course i think that's too many or not it seems i i assumed that like the ability to accumulate a lot of i agree that like i don't think anyone who's very powerful in in business um can do so by being purely logical like i think they also have incredible intuitions and i that's definitely the case but a lot of them or at least some of them are somewhat like disembodied and certainly not like connected they don't feel like they're connected spiritually for example to the earth yeah or to like the feminine and i say that in the like best sense of like there's like the sacred feminine yeah i think they can see the sacred masculine somewhat but there seems to be a disconnect there and so um yeah i mean you you spent much of your adult life sort of working in these environments very male dominated environments tech yeah software engineers oh yeah how you know am i wrong tell me i'm wrong i would like to be wrong like have you you know do you have any thoughts on how to like steer the general you know get people more respecting the i guess define what what the sort of sacred feminine is and like and if you've got any thoughts on how to move the needle there to help people be a little bit more like comfortable in the like softness and the passivity you know in in well not passivity that's not yeah no but the receptivity something like that yeah so okay so it's funny you say this because i have always been in super super technical teams like software engineer working on machine learning like hardcore like I look very out of place and as I have become friends with these people and they see that I am very intuitive and I'll talk about these things these men will start to be like oh maybe there is something to this because you're right about you're right about like everything and you're also super logical and if you who I really trust and respect and like know that you think like me know that you're like one of us right you know so that's kind of like unfortunately so many women who are intuitive like are less technical and so men will be like I'm just discrediting her because she doesn't get it she doesn't think like me doesn't speak the language right I do I think like them totally and so when they see that I also have this piece where I'm like kind of dipping my more than dipping my toes like swimming in this other thing that is so useful and so on point and they can feel it, you know, when they're like talking to me or like what I expect will happen will happen, whatever, then they'll start to be like, oh, interesting.
1:12:19So I think as more people, as men start to go into it, as women start to go into it and become more technical, one of the things I also really like that what I'm doing with AI is that, as I said, I was always very, you'd be surprised if you saw me and you saw my team you'd be surprised um and so what i'm excited about is more women being like oh she looks like me because i actually when i was in at stanford one of the things one of the reasons i went into computer science or felt i could was because there actually were women who kind of looked like me from my sorority who were doing it so i was like oh i could do it because you know christina's doing it she's actually a real person and i've told her that she helped me get into it um and what i love is hearing that women are like oh harper's doing it i could do it then and so again as more women go into it who come in with the spiritual side and they speak the same language as the men who are less like that or as more men come in and they start to consider the spiritual side it'll all end up will all converge and super exciting i'm so excited about it yeah i hope so yeah i mean i've definitely seen i mean the idea of like representation it definitely makes a difference like in poker um i was partly inspired to join poker because i met a woman who was a badass and she was a pro and i mean part of it though i mean i also wanted though from a strategic standpoint i was like oh this is it makes more sense to come in as a minority and do well in it because you're going to get elevated in other ways oh interesting um yeah it that that was always always valuable like you know people often ask like oh was it easier to be a girl or a boy in poker i'm like well i think a girl because like i wouldn't be as famous as i am as a poker player that's true i've not been a female in terms of fame yeah yeah which can be leveraged yeah and useful um i mean there are obviously downsides you know the the sexism is still i'm sure there's still plenty of sexism in tech and like but it's a that stuff is improving um but i don't know i think the the upsides of being a minority generally outweigh the next the downside yeah i mean that's was my kind of whole objection to the DEI thing because it while it was rooted in good intentions if you enforce these false quotas onto things that were previously meritocratic it hurts the minorities yes in the long run because no one then takes them seriously because they think that they're there because they're an affirmative action hire and because and like they're justified to think that because it's true these di these di quotas were being so in like harshly enforced that actually yes the companies were over you know you you have if you want to have a high trust society and it's something where there are competitive fields where people are competing for jobs you have to do it on like merit that's the only way as as and yes it's hard to like completely account for all different variables and yes there will sometimes be bias that creeps in And even in like no meritocratic system is perfect.
1:15:16There will always be some bias. But if it's not at least based on that as a fundamental core principle, if you throw it out and say that we're going to do stuff on quotas, then no one trusts it anymore. Exactly. And the minorities... It ends up hurting them. Yeah, they get hurt by it. And like I've seen it over and over. So many women are like, I deserve to be here, but no one believes that because there are other hires there who happen to be of my gender who probably don't deserve to be here. And now I look like we all, I get taught by the same brush. it's just you can't um that yeah it drove me crazy and i mean that's one upside of like this new political swing is that that stuff is now we're hopefully going to go back to a meritocratic system a bit more and that but that's going to take a while to rebuild that trust um and yeah it's it's annoying rant over okay um i do want to touch on i think one of the big areas that is a downside of this new like generative AI era that we're in and that is the like the easier ability for cyber criminals to do scams you know scams are nothing new there have been con artists since the dawn of time um but it's probably never been an easier time for a scammer to do really elaborate cons and scams on people because we can now create very convincing digital you know like deep fake images deep fake voices deep fake video deep fake emails which is what i learned is essentially um so and actually you and i got to know each other in part a bit better because we both spotted the same scam well you you noticed something we were and you got it taken down you had the connection we were a good team and that uh there was a rocket launch going on um one of the spacex rocket launches and you you wanted to go and watch it online well yeah yeah yeah so i was i was with my dad i was visiting him and he turned on the youtube and he was like oh my god look at this it says it's live but it's totally this elon crypto deep fake thing it showed it was live and it was a video of elon talking and what he was saying it was a spacex youtube oh and it was the space x youtube as far as you can tell space x yeah and elon was like everyone go download you know crypto this and send me this screen exactly on the video yeah to download this to buy this coin or something and i forget exactly what he said but it was obviously a scam for salt of all well it wasn't it to me it was obvious but to most people i can imagine it wasn't obvious well because and here's why so then i went on it and i was like wait what the hell it so it was you know right before this rocket launch was happening this thing had 70 ,000 like it was from current viewers yeah it was front page of youtube and it had 70 ,000 concurrent viewers on this live stream it looks like it had a tick marks it said spacex as the title of the didn't have a tick mark i didn't know yeah no and it had it had a tick mark oh my god um so it was like official space yeah and if you go and then you clicked on the channel it looks like spacex's youtube channel it had all these other like thumbnails which were the same millions of views with with many many many views and what had actually happened what these these scammers had done was they might i mean talk about a long con they had obviously bought some other kind of youtube channel that you know probably some like foreign channel or something that had already a lot of subscribers let's say two million subscribers i think it was and so it had the verified tick mark for whatever it once bit had been but then they had over time either like renamed and re like thumbnailed all the old videos so they still had those oh they re-thumbnailed them yeah so if you clicked it it was like a totally different video yes i think so and and they've done this like you know they just like changed the sort of outer casing of this channel to make it look as much like to the you know it someone just looking quickly and and then they and then they'd done this like live stream and it was really high quality looking with like a lot of like ripped spacex footage um but with with a deep fake of elon saying this stuff and yeah i mean i'm curious how i mean i think it was only up for a couple of hours i i flagged it to someone um i know at google and who immediately got it taken down but like you know it these things are getting so easy to do and they're getting more and more um indistinguishable indistinguishable yeah so yeah and i mean this is another sad thing i recently found out this is a different type of scam um a friend of mine um his mom had gotten catfished on on a dating you know very established dating uh website and over the course of two years ended up giving like hundreds of thousands of euros to this person which we all look at it and be like how could anyone be that stupid but here's the thing this person like it was so so targeted and so gent like done slowly with so much care and effort and it's one of the most harrowing things i've ever heard because you know she was in you know as far as she could tell she was in a relationship like this person had a consistent like like all these different pictures because it's so easy to create like a completely artificial persona online and this is becoming more and more common and i only know about this because he just happened to be a friend who mentioned it think of all of the stories where this is happening to people because people are when they find when they do get scammed in this way particularly in like a love interest type thing think how the shame that is associated with that so so many people have probably had something like this already happen to them and aren't talking about it um and yeah really breaks my heart so could you sort of explain how um you know what are the current techniques and tools that are out there to try and help you know either the companies or individuals detect um ai generated images is that like water marking like what explain the cat and mouse game basically yeah so the cat and mouse game is basically when you have deep fakes and then you find ways to distinguish deep fakes from real videos, for example, like the hands are messed up, then the scammer is going to work on fixing that issue.
1:21:40And so it's a classic cat and mouse game, arms race, what have you. Like for example, with scamming at Meta, I considered joining that team when I came on full time. And basically it's this cat and mouse game. They think of features that help them find out what makes a spam bot a spam bot maybe it's commenting it you know post this many times but you know x many times per minute or per hour maybe it says this kind of thing or it never says this kind of thing and then they find these features that help them find the spam bots and then they wipe them out but then the spam bots become smarter and they stop using those features that made them distinguishable from regular people.
1:22:22And so it's this race. And so with deep fakes, again, it's like, if there is a feature that helps us know that it's AI, the scammers will then make it not a feature that we can use anymore. And so ultimately we are not going to be able to tell deep fakes from reality. and so we will have to actually have this kind of identity-based trust system where people post you can only really trust what is posted by the person who is sharing the information so if you see like a video of someone reposted by someone else you just can't trust it unless the person themselves is posting it and like kind of verifying this and so yeah I mean that's where we're headed and I think my take on it is that it's actually a good thing because there's this phenomenon called gal man and amnesia where if you're an expert and you're reading a top uh an article on what you're an expert on say in the newspaper and you read it and you're like this is wrong like this is wrong or say you're uh reading on the reporting of an event that you actually personally know about and you're like this is all bullshit like this is lies and and then you don't you instantly just discredit the whole article you're like the person here doesn't know about it or they have an agenda or whatever like you know toss out the window then you move on to the next article and you trust that everything in it is true right you go on something that yeah you don't know anything about yeah oh i oh this article about this bad thing that happened at the music event that i was at oh this is all nonsense and then you click into the same you know three pages down on the same news outlet and now they're talking about something going on in Russia or, you know, Ukraine or something, you know, like, oh, yeah, this must all be completely accurate.
1:24:10It's like, no, did you not just see? Exactly. And it happens almost every time you read about something that you actually know about or an expert in many, many times. And it's this amnesia because it happens like right away. And so I think so many people don't know about this. And I kind of feel like having this identity based system where people post what they believe and you trust the source you get you are only trusting the source and then you can have this kind of discourse online like with x you can have people comment and then you can have community notes where they more heavily weigh community notes where opposite sides of the political spectrum tend to agree that's kind of a good indicator that super majority right it's a good indicator that it is truth what these these people are agreeing on is the actual truth and so I think that that's actually a smarter way rather than having these like anonymous people writing these articles that like people just assume to be true um so yeah that's kind of my take on it but yeah ultimately we are not going to be able to know and watermarking too is an interesting thing I actually made a whole video on this and it's kind of a longer thing to talk about but Google for example has made this watermarking technology that it's like this snippet of code that shifts something that's generated.
1:25:29So whether it's an image or a song or whatever, and Google has put it in its models where it now has this watermark. So if you can run it back through this watermarking tech and see if the watermark is on it. And basically what the watermark does is it shifts the final distribution of words or of pixels or whatever to make it different than it would have been originally. And it shifts it in such a way that then it can kind of retroactively say, oh, given that it's using these words instead of these words, or the distribution is shifted towards this kind of different word distribution, then it has this watermark and it's AI generated.
1:26:04The problem with this is like, you would have to have all people implement this, like every single, so say you had every single closed source company mandated that they all use it. So like open AI had to use it, Google, whatever. You still have these open source models where people have full access to the code. And again, this is just a piece of code that goes on at the end. Like you could just remove it, right? You and no one's gonna and then bad guys are incentivized to do that yeah of course and then i mean anyone using an open source model is incentivized but yeah exactly like the bad guys are incentivized to use it and then if it's not government legally mandated no one's going to choose to do that because no one wants to use a model that is watermarked like if you had the choice between a watermarked model and one that isn't like why would you use a watermarked one right as well if yeah if you're assuming you you're using it for emails or whatever like why would you don't want anyone to know that you're using ai to do your thing especially if you have the choice like why would you yeah i mean that's we're basically heading into a time where you just unfortunately you just have to assume that anything you see if it's remotely if there's any kind of incentive for it to be ai ai generated for someone's nefarious purposes you have to assume that it is and that sucks so we're no longer in a world where it's like trust but verify it's like distrust and verify well this is kind of where the intuition thing i'm now thinking about this i hadn't thought about it before comes along is i think once we're really good at picking up on intuition you can kind of sense when people are lying um like you just kind of sense it i think it i think maybe someday we'll yeah but but that's a different thing our intuitions are well formed for telling when someone is lying when they're sitting in front of us and you know you're in a social environment because we evolved for millions of years in to spot deception as animals and it's like a you know again it's like a very much a body you know you can hear it in their voice and so on but that's a very different thing to expect your intuitions are well calibrated to spot something online online like it's a whole new world like we are maybe i don't know i think i don't know i worry that it's i mean i agree i agree what i'm saying is woo i'm not saying i fully believe it but it's an interesting point i just think it's it's it's wishful thinking is my concern i i hope it's true i want it to be it would be nice and i think people should try and cultivate the skill nonetheless yes exactly i'm not saying it would definitely work i'm just kind of throwing it out there because that would be cool if ultimately we were able to like we just couldn't lie to each other anymore like that would be so cool i mean and i mean that's in some ways where like if we really like play this out it seems like we're going towards this hive mind thing anyway this hive mind thing well i mean if we all have like neural links i mean if that's a direction that could the world could end up in i'm not sure if it's one i necessarily want but it's a plausible ish path um because tell me more about the hive mind thing what does that look like well so say for example you know the world is becoming increasingly surveilled okay again i don't like it i am not a pro surveillance person generally speaking um at the same time if you are heading into a world which is fully post-trust or a low trust society the unfortunate like counterbalance like the uh counter reaction to that is to increase surveillance because it's like no one trusts anybody then you have to have some kind of way of like verifying how you verify well through some forms of it doesn't have to be like classic cameras but just like different surveillance methods which then obviously raises the whole problem of who watches the watchers how you know how do you ensure that like this like surveillance state isn't now just completely controlled by a centralized entity which may or may not be benevolent and is often malevolent because the type of people who want that level of power exactly and it corrupts exactly and then power corrupts so this is this like classic idea of um these two dystopian attractor states where you either have like just complete breakdown anarchy because no one can trust anything no one can verify anything and no one knows what the hell's going on and like you so you like social cohesion breaks down um and then even beyond that you have this idea of like with sufficiently powerful dangerous technologies that have been sufficiently democratized you get into this world of like small kills all where like some anyone can synthesize something that's very bad that can propagate and and hurt people so you don't want that that's like like the anarchic dystopian state and state and then the other attractive state is this idea of um well to protect against this you have to have like surveillance and like everyone watching you know the uh basically like you can't give people like agency anymore because it's there are so many dangerous tools that it could cause so much harm that you have to take away everyone's agency and you have some kind of like centralized nightmarish tyrannical tyrannical surveillance state both of those suck i want neither so it's like what is this third outcome what is this third attractive state that we can yeah that we need to design such that people still have freedom and privacy and and at the same time also safety that's it's like how to navigate the tension between the three um it kind of reminds me of the should we have open source models at all should we classify i think mark and jason was talking about how the government was considering classifying the math behind large behind ai which is kind of silly because first of all the math is not that complicated like like my assumption and maybe this is because i don't know about nuclear but my apparently they classified nuclear like math behind nuclear weapons but i could imagine that's probably harder but maybe it's not right it's probably not a bunch of like matrix multiplications right exactly and like just like yeah gradients and you know back propagating gradients like it's not that complicated and i think people could figure it out and also i mean more broadly, it makes, it is a horrible, it would be a horrible travesty if AI was concentrated in the hands of a few small companies because AI is truly applicable to everything.
1:32:17It's just the math of learning. You know, interestingly, they're now using AI math to model human brains. So in the same way that neural networks were kind of based off of human brains, Now human brains are now basing their models off of AI brains because of the math, because we learn kind of the same way. We have these priors and then we get new information and then we update. And it would be such a tragedy to concentrate the understanding of learning algorithms in the hands of a few because it applies to everything. And imagine, I mean, talk about these industries that are highly regulated. Imagine if every single industry, every single tiny use case that wanted to have learning involved had to go through some kind of government entity or one of these tiny little.
1:33:07It's like, oh, my God, you may as well just like like there would be no advancements. And if you look at the advancements in education compared to advancements in things that aren't super highly regulated. This was posted by Mark. He showed these graphs of like the improvements and advancements in regulated areas versus not. I mean, AI has the potential. Again, why I'm sitting here is because I believe AI has the potential to cause this kind of utopia if we apply this learning algorithm to everything, which is part of why I kind of have this philosophy of, I think, I wish there were more AI, not just for AGI, but like more AI resources going towards things like not just developing AGI, but for AI tools, which I think is what actually Mira Murati's new company is doing.
1:33:50so I have to look into that but I think that's literally what she's focusing on which got me really excited um but yeah and the reason I'm sitting here is because I'm excited about AI being used for all these small things just learning in every single domain ever everywhere so if we had it going through these these companies it's not even feasible and it's tragic yeah like that's that's the tension of this like the reason why open source is such a debate is because there are such obvious upsides to open sourcing everything. But there are also such obvious downsides to open sourcing everything because there are bad actors, there are dangerous things, not right now, but in time, there are going to be more dangerous things and the bad guys will always just like, you know, get rid of any safety mechanisms that can be put in.
1:34:35There always have been that though. No, totally. But it's just that the stakes are getting increasingly higher. But it's also true that the stakes are getting higher in the other direction. Right, in the positive. Yeah. Yeah, so there is just this tension and we need to find whatever the like the sane, it doesn't have to be perfectly middle ground, but like the higher complex state. We don't want to just be open source everything, open source the nuclear cosa, you know, but we don't want to be close source everything, regulate everything. No, like we have to thread the needle and it's going to be a very complex answer.
1:35:09I just, you know, let's hope we can figure it out in time. I know it's exciting times it sure is I could not imagine a more exciting time to be alive it's hectic but it's so exciting and and it's also like the ability any one of us you me people watching like because we're at such a crucial time in history it we have more agency than probably anyone oh totally and the two keys to happiness are agency and gratitude go on that's just what um uh shoot paul conti that's what paul conti said on the huberman lab podcast there was a thing on emotional well-being uh-huh they did a series together paul conti's a therapist and he says the two main attributors to happiness are gratitude and agency agency and gratitude so with people having more agency we could might see a lot happier of a society and gratitude is kind of more of a choice kind of like seeing thing focusing on what you can be grateful for right well it's the practice of noticing the good stuff in your life exactly uh yeah no i'm being grateful for feeling gratitude for it you know yeah yeah well and And listen, anything that can, that's really, people talk about like the notion of progress and they're like, oh, let's look at all these different metrics, you know, more infant mortality rates, rates of disease going down, et cetera.
1:36:45I mean, these are important metrics, but I think if you had to try and distill like really what the measure of progress is, it's like, are people happier now than they used to be? It's hard to measure. And maybe it is like impossible to develop a like, by definition, you can't create a singular metric to measure happiness but i think that is probably the like the guiding principle do you think it's happiness or do you think it's fulfillment well maybe i mean depends on your definitions but yeah maybe fulfillment is more directionally that but like because that sort of because it's okay to not be happy sometimes right but you don't want to always be unhappy what really is meaning i guess now yeah like fulfillment has meaning you can be happy when you're like i don't know i mean again it depends on your definition but some people might define happy as being i'm riding a roller coaster all the time yeah and it's like okay but it's not that meaningful yeah well no but like literally being on like a on a like you know doing i'm doing a parachute literally okay yeah um i don't know because this is a very specific thing but at the same time my life probably isn't very meaningful if that's all i do yeah or maybe it comes back to this fulfillment right because like do i like some people get fulfillment out of meaning some people get fulfillment out of always being on either a physical roller coaster or an emotional roller coaster you know like yeah just feeling like okay i like i'm satisfied like i feel good like life is worth living and some i mean if you're happy all the time sure that's one way of being fulfilled i guess it's like the idea of like the the wire heading where it's like you don't actually do anything you're just plugged into a machine that feeds the right chemical juices to make you feel like you're you know having some kind of peak yeah oh my gosh or whatever experience and that's all you're in the experience your reality yeah you're like you're an experience machine i think most people would agree that that's not actually true happiness totally also i want to correct myself i realized i butchered the paul conti thing i think he said the measure of health is agency and gratitude of mental health great not happiness right and mental health mental health perhaps that's maybe the better proxy yes progress yes yes so okay well i think this is a good time to wrap because there is a helicopter for some reason circling around this house they're descending I don't know coming to get me yeah maybe it's the AI sense that we were talking about the demons and we were trying to combine spirituality with technology uh-huh good I'm glad we did the IRS album is upon us yeah exactly literally oh god awesome thank you so much thank you for having me Liv this was so fun
1:39:22Thank you.
From the publisher
What if AI isn’t replacing us—but redefining what it means to be skilled? Liv talks with engineer and educator Harper Carroll about how AI is transforming education, leveling the playing field for creators, and reshaping how we build, learn, and make decisions. A grounded look at the tools, challenges, and surprising upsides of the AI era, as well as a personal exploration of the spirituality (or lack of it) in Silicon Valley.
Chapters
01:04 - Intro & Harper’s Story
04:58 - Growing a Following
07:18 - AI vs Machine Learning
11:18 - Harper’s Mission
18:43 - Her Favorite Applications of AI
27:27 - Choosing Between AI Tools
29:46 - Integrating AI in Schools
33:11 - Regulation to Prevent Over-Financial Optimization
36:59 - Leveling the Playing Field with AI
46:40 - Should People Still Learn to Code?
51:05 - AI Therapists?
57:54 - AI & Spirituality
1:02:28 - Cultivating Intuition and Wisdom
1:09:23 - Silicon Valley, Tech Bros and Femininity
1:16:00 - AI Generated Scams
1:21:13 - The Cat-and-Mouse Game of Scammers
1:28:23 - The Hive Mind
1:35:05- Agency and Gratitude
♾️ Harper IG : https://www.instagram.com/harpercarrollai/?hl=en
♾️ Harper Youtube : https://www.youtube.com/@HarperCarrollAI
♾️ Harper X : https://x.com/harperscarroll?lang=en Credits
♾️ Hosted by Liv Boeree
♾️ Produced by Luca de Vico
The Win-Win Podcast
Poker champion Liv Boeree takes to the interview chair to tease apart the complexities of one of the most fundamental parts of human nature: competition. Liv is joined by top philosophers, gamers, artists, technologists, CEOs, scientists, athletes and more to understand how competition manifests in their world, and how to change seemingly win-lose games into Win-Wins.
Podcast links
♾️ Youtube: https://www.youtube.com/@LivBoeree
♾️ Spotify: https://open.spotify.com/show/03bGVUa...
♾️ Apple Podcasts: https://podcasts.apple.com/us/podcast...
♾️ Pocketcast: https://play.pocketcasts.com/podcasts...
#winwinpodcast #AI #technology
