Latest / SPEAK LIKE A CEO / 295: How to Decode People: What Humans and AI Get Right – and Wrong. Blake Eastman, Founder of Nonverbal Group
Transcript
- 0:00Let's say you were a kind of person that's not really good at reading
- 0:02on understanding behavior. Our software can be overlaid in an interaction
- 0:06and give you real-time feedback, catch your blind spots,
- 0:08it can coach you individually on specific movements,
- 0:11language patterns, tonality aspects. I don't know exactly what direction
- 0:16I'm going to take this. Blake, welcome back. Yeah,
- 0:19thanks for having me. Exciting. Likewise, so we talked recently,
- 0:23and he said you were currently working on a book. And the book is a new kind
- 0:27of how to win friends and influence people. Now, why doesn't it need an update?
- 0:31Yeah, so social skills have been taught to be the simple things and they are,
- 0:37but social dynamics are just so complex. And I don't think there's been any books
- 0:43that sort of explains the why behind a tactic working. So for example,
- 0:50somebody gives you a tip online, you're like, oh, that's really good,
- 0:53and you use it, but you don't understand the underlying robust nature
- 0:57on why it works. So I wanted to write a book that's going to sort of explore
- 1:02those mechanics, so give tools, but also provide an underlying science and
- 1:06foundation for everything that's sort of going on. And it's kind of like,
- 1:09I've been working on this book for like three years now,
- 1:13because it's writing my materials in my programs. And now just
- 1:17stilling, it's a confronting thing to distill like all this complexity into something
- 1:24that is under 90,000 words. And really useful. Yeah.
- 1:29Can you give us something example of something you're shining a light on that maybe we haven't
- 1:33come across in a book yet? So I think there's a bunch of different,
- 1:37sort of unique takes, but probably the most foundational one is the world of the invisible.
- 1:43So we talk a lot about behavior and body language and what people say and how they say
- 1:48thing, but we don't talk about the invisible pervasive themes that sort of bind human interaction
- 1:54together, things like status and hierarchy, things like safety and survival, things like
- 1:59value orientation. And in order to understand behavior, you need to understand what's not there,
- 2:07which is this world of expectations. And I'm obsessed with cross disciplinary science,
- 2:13like looking at a wide range of scientific fields and seeing where they're all saying the same
- 2:19thing, but differently. And I sort of have this thesis that comes from the pervasive themes.
- 2:25Like, why do we organize ourselves into status structures and hierarchical structures,
- 2:31right? And then why do some companies that try to do this can sometimes work with most of it fail?
- 2:36I've gone into companies and I'm sure you've worked with people like, oh, we have a very flat
- 2:40organization. Yeah, totally flat. We're all equal. And in five minutes, you're like,
- 2:45this is not flat at all. You're just saying it's flat. So there's something about the mechanics
- 2:50of psychology and our neurochemistry and biology that does that. So it's important to understand
- 2:55those themes if we really want to deeply understand people. That's super interesting because a lot
- 3:00of books, especially for business people, stay at the surface level, because they do this,
- 3:05say that. And you know, this value in that, but it's super interesting to go a level deeper,
- 3:09but that's super hard work as well, right? Yeah. So that's kind of what I'd like my book to be in
- 3:13a lot of ways is like this foundational precursor to any social skills advice, because there's
- 3:20sometimes like, I'm in an audience and someone I'm watching on a stage, somebody says something
- 3:24around social interactions. And I'm like, oh, that's like really good. Like, that's a great tip.
- 3:29But I could explain the robust nature of why and the problem is when people apply social skills,
- 3:37they take a tip, but because it's not designed for them, they fail in the execution of that tip.
- 3:43So it's sort of creating a lens where you could pull in information that's useful.
- 3:48Actually, honestly, your question, I just thank you because you just connected two things that
- 3:54were disconnected while I'm writing. So I appreciate that. You're welcome. You're welcome. Yeah.
- 3:59Do you have a title yet? Not yet. We have a working title. It's one of these things that I'm still
- 4:05working with. Listen, I'll be honest, like as a creator, I love that I can change things.
- 4:15So my course, I'm just refilling the whole thing because it's like flawed in certain ways.
- 4:19With a book though, it's hard because the book is the book, right? So I'm spending a little bit
- 4:24more time thinking through like the structure, the architecture. It's taking me way longer than I
- 4:31thought it would. And partially because you have to distill and really critically look at what's
- 4:35important and what's not. But the book is kind of like a summary of my life's work for the last 20
- 4:42years before I start doing other stuff. And you've done amazing things we talked about last time
- 4:47when you won the podcast. The poker pro, your psychology professor, obviously, you
- 4:53build the nonverbal group and you call yourself a behavior list. What does that mean?
- 4:59I mean, it was just, it was kind of a cool thing. So I remember when I was in the grad school,
- 5:06I think undergrad, like learning about like BF Skinner, like the first behaviorist and stuff like
- 5:11that. And a lot of his stuff is completely like wrong. I just thought it was super cool.
- 5:16Like so I like the identity of like this Sherlockian, like sort of like behavior list.
- 5:21For the outside observing. Exactly.
- 5:25Exactly. But the reason why I love behavior is because humans are these prediction machines
- 5:34that are constantly sort of predicting their reality. And there is no such thing as reality. But
- 5:39video captures a raw reality that I can sort of look at. And you know, a video is essentially like
- 5:47raw pixels and raw data and it provides a somewhat level of truth in a world with so much complexity.
- 5:54So I've always been like fascinating. Like you can't dispute somebody's like smile timing
- 5:58or smiling morphology. You can dispute what it meant. You could dispute if it was intended.
- 6:03You could dispute if it was a nice smile or a bad smile. But the raw data is the raw data,
- 6:07which is why I like the concept of behavior. At the nonverbal group, you help executives and
- 6:13businesses make sense of social interactions of communication. I guess in reality like my
- 6:18job is to help people navigate a relatively complex world that gets more and more complex.
- 6:23What do people ask you for? What kind of advice do people need right now?
- 6:28It's really interesting. So I get pulled in for like human process optimization. So like a company
- 6:36is like I have a project with an organization right now that has a ton of sales calls.
- 6:41And I'll analyze like 26,000 sales calls that they've had already and provide feedback on sort of
- 6:47where they're missing the human element in these conversations and how to optimize the workflow
- 6:51and optimize the process. Or I'll work with like an executive team on having a more collaborative
- 6:58environment. So I usually go in with one approach. But then I see so many other things on the videos
- 7:04that it starts going to something else. So it's like and I think my longer term clients are just like
- 7:12Blake reduces the friction of our team unlike anything else. So like let's say a watch a call
- 7:18and now our software catches a lot of this. But like let's say you said something
- 7:23and your CTO was like taken back by it. I'll catch it. Show it to the CEO and be like,
- 7:30hey, like what happened here? Then put you to on a call and get it cleaned up. In minutes as
- 7:35opposed to, oh, I can't believe he said this. I can't believe he said that. So think about this.
- 7:39Like all the back back-minded sort of conversations you have about people in your organization.
- 7:45I'm constantly going in and like clearing it out and fixing it. But the truth is like I've
- 7:50just done so many weird things over the years. But don't call you for the standard things, right?
- 7:55Yeah, it's rare that it's just like, hey, I want to be a better speaker in my organization.
- 8:03Like those things are more rare now. But occasionally old clients would be like, hey,
- 8:07talk to Blake, talk to this. But I'm doing that amount of work is becoming less and less of my
- 8:13overall sort of time I dedicate. Yeah, make sense. And you have this interesting sentence that
- 8:20you mentioned that you're teaching people at machines to read human behavior, right? You also
- 8:26mentioned that there's a lot of data to crunch. So what did the machines come in?
- 8:30So the truth about reading behavior is reading is a really bad term. And it's hard because I need to
- 8:37use it as like a marketer. Right? Like, oh, like learn how to read people. But the truth is it's not
- 8:43possible. Like you can never read someone. You could only predict. And your predictions have
- 8:48accuracies that are either high or low. So the reading people, it's a dangerous term because it
- 8:56infers that you can encapsulate exactly with that person's experience. And you can't like, I'm sure
- 9:01there's times where you look at someone and you go, they're really happy and deep down inside,
- 9:04they want to get out of that meeting is more than anybody else. And there's other times where
- 9:08you think someone's really depressed. And it's just their facial construction, right? So it's that
- 9:13but the same way that let's say like you had a friend for 30 years and you sit down for dinner
- 9:20with them. And in the first two minutes, you're just like, sub things up, right? Like you just
- 9:26get the sense to sub things up. And you're like, Hey, is everything okay? It's like everything's
- 9:29not okay. Actually, you know, I just found out I have cancer. I'm a little bit worried,
- 9:34but whatever it is. How did you come up with that spider sense? And the reason and the mechanism
- 9:40was you processed a large amount of pattern based data with that person. And you sort of made this
- 9:46educated guess about what the reason on why they were doing it. So if you operate from first
- 9:53principles, there's no reason a machine can't do this with higher accuracy than people. And that's
- 9:57kind of what we're building. But it's a very complicated process. And I can explain why, like it's
- 10:06just very understanding. So you're taking a baseline, right? So with you in the example of your
- 10:11friend, you know your friend's baseline behavior because you've met so many times. So you know
- 10:17something isn't right, but you don't know what it is. Hence you press him to say, come on,
- 10:22what's what's going on? It's fair. But the whole notion of like body language, body language
- 10:29baselines, it's just so oversimplified. It's like kind of comical. So like a lot of body language
- 10:36experts will be like, Oh, you got to figure out their baseline and notice when there's a change.
- 10:40They like kind of use that as like a catch all to like fix everything. And the truth is
- 10:46baseline is the most biased process ever. Machines can hyper baseline. Like we have a system
- 10:52called Cheshire that could sort of measure your smiles and figure out the bell curve distribution
- 10:56of your smiles and where they're all aligned. But people do this at such a rapid intuitive sense
- 11:01that it's very hard to reverse engineer and articulate exactly what it is. So the baseline is not
- 11:08relative to that person. The baselines relative to you, your current state and all the interactions
- 11:13you've had in your entire life, right? So it's part of it is not understanding baselines. It's
- 11:20understanding what I call perceptual bell curves. So it's basically like this. I say that
- 11:27there is no way that we can read behavior. But if you gave me a video of somebody,
- 11:34you give me a video of three people. I could predict with a high accuracy which person is perceived
- 11:40to be the most confident in that video. And I can do that not because of anything except mapping
- 11:47the behavioral distribution of what society, insurance cultures, proceeds as confidence.
- 11:53So we can predict that as opposed to predicting exactly what it is. So like I work with teams and
- 11:59I can predict who's who's probably going to cause social frictions based on behavior.
- 12:05And what I'm looking for is those perceptual bell curves. So it's a little bit
- 12:10more robust when you start to build sort of pipelines that do this. But yeah, it's what humans do
- 12:18with less, with actually it's kind of cool. It's what humans do. People suddenly says
- 12:25nothing to me. He was like, well, all people are biased. I was like, well, what's cool about
- 12:29our systems is they're way more biased than people, but they're way more versions of them.
- 12:37So like we have a system called consensus that kind of like argues with each of the sort of
- 12:42logic models to determine, well, you're coming at this way, this one in this way. And what's
- 12:46really fascinating is like you can build out perceptual systems that mimic human behavior.
- 12:53So like I can build a very negative leaning perceptual model that takes all your behavior and
- 12:59be like, well, he was dismissive there and he was this. And then I can build the entire opposite
- 13:03of somebody just looking at all behavior and being like, wow, they're such a great person.
- 13:08They're so alive. They're so attentive. But the real interesting thing is where society agrees
- 13:14on what is average. That's where it's helpful. Interesting. And I suppose that the algorithm
- 13:19having biases, you said more biases, but they probably don't get emotionally attached to them.
- 13:24So then we're learning it to exactly. Yeah, there's no, there's just predictions. It's not like,
- 13:29oh, my God, this person said this. So it's less bias at the processing level. And also there,
- 13:35there are stages that are not biased. Like so when we take the human face and break it down,
- 13:40like, have you ever, I'll tell you the first step. Have you ever seen like those emotional
- 13:44recognition software? Yes. Like where you see someone smile and it says they're happy. Like,
- 13:49they're just complete nonsense. It's just like it's just silly apps, right? They're silly apps.
- 13:53And what they're based on is like open CV or media pipe where they put these facial landmarks on
- 13:58your face and they detect when you smile and they infer that you're happy. Well, you go like,
- 14:03you smile a lot not as an indication of happiness, but like social coordination and all that stuff.
- 14:09So like the first real step is a very complex lift in computer vision to massively improve those
- 14:16systems because the ones that are right now don't capture all the nuance of human behavior. So
- 14:21for example, but when you're doing like facial coding using something like facts or something like
- 14:26that, like you're using like wrinkles to figure out a lot of muscle movements. Those systems don't
- 14:30pick up on wrinkles. So we have a system that sort of like parses your face and firings out where
- 14:35there's wrinkles to help define that. And the goal is to sort of create this mathematical,
- 14:41not mathematical, but like a visual model where it's like me being processed by a machine.
- 14:47And you could understand right now like is Blake interested? Is he not? What are aspects of his
- 14:52face where this is at a high level of fidelity? But the problem, I mean, I could talk about this
- 14:57hour. The problem is compute like if we were to take this conversation and run all of our models
- 15:03on it, it's like a 500 megabyte file. It couldn't do that at real time. The more and more I built this
- 15:08stuff, we are so cool and special and unique as humans. Like I just have more of a respect for us
- 15:17and our biology. The more and more I do this because how do we do all this? This is crazy.
- 15:21It's crazy how we process it and the processing be compute we have between our ears. It's just
- 15:25crazy. It's incredible. The question comes up. Why do you build it? What's the reason for this to
- 15:32exist? It's tough. So I'll be honest. All the stuff that I've done in my life has just been
- 15:40focused on helping people be more effective communicators and sort of the objective function of
- 15:45everything we teach is to increase the amount of value you add and extract from social
- 15:49interactions and minimize the suffering. So a lot of people are suffering and I don't believe
- 15:55they need to and there's tools to sort of reduce the suffering. The technology and the stuff that I'm
- 16:01building, it's like it's hard. Am I actually increasing suffering because this will replace so
- 16:13many jobs? Like it will. But it will reduce suffering when you have a robot in your house that's
- 16:20able to sort of identify and pick up on your little themes and patterns of behavior and coach you
- 16:25at a highest fidelity. It will reduce suffering by getting people to do the jobs that they don't
- 16:29necessarily want to do. So I do struggle with some of the philosophical discussions of this but
- 16:33honestly I think how I rationalize it is like somebody is going to do this and I'd rather it be me
- 16:41because I would approach it with sort of the care and attach it to detail that I think most people
- 16:46are not going to do. They're just going to try to build a product but they don't account for the
- 16:50sort of nuance of it all. Yeah and if I think about business, how would you use it in business?
- 16:57An example would be I mean so many different ways like we're already deploying some of this but like
- 17:04let's say you were a kind of person that's not really good at reading understanding behavior
- 17:08and you really struggle with social interactions. Our software can be overlaid in an interaction
- 17:13and give you real-time feedback on what you might not need to do and what not need to do.
- 17:16It can catch your blind spots. It can coach you individually on specific movements, language patterns,
- 17:23tonality aspects. It can be integrated in a wide range. I don't know exactly what direction
- 17:30I'm going to take this. I have three, I have like a couple of options. The first option is
- 17:37do what I'm doing now which is integrating versions of this inside of my own software and
- 17:42really focusing on social interactions. The second is building like a model company,
- 17:47a company that builds and license models and then the third option is building an infrastructure
- 17:52company. So building like the open AI or the anthropic of human behavior. But the thing that's
- 17:58hard is like I've coached a bunch of venture-based founder startups over the years and it's really
- 18:04thinking like do I want that life and do I want to play that big of a game and what I give up when
- 18:09that happens. Right now I'm in Austin, I travel, I do whatever I want. If I build an infrastructure
- 18:15company, I move into San Francisco. There's no other place that I can be. I'm working and I'm
- 18:22marching towards something and you know the value as I spent a very long time executive coaching,
- 18:29a wide range of clients, very big startups to private equity, to like neurosurgeons. So I kind of
- 18:38know what works for me. So you just got to really make that conscious choice about how you're going
- 18:43to do it. But I'm doing some version of this. I just don't know the scale that I want to do it at.
- 18:48So at the moment it's more the foundation or technology layer. And there's also like just for
- 18:54any of you who start a startup, like I don't really believe personally for me at least.
- 19:01I don't believe in the whole raise money and let us build something. I believe more in
- 19:07let's build something, let's get to solid ARR, MRR and then let's raise money to explode it.
- 19:13And you know, so all of my corporate clients, all the money I'm making just gets deployed to this.
- 19:21I can imagine how big is the operation? Are you raising out? So you're not raising out some money.
- 19:25It's organic at the moment. Yeah, it's organic. It's probably, I mean, it's definitely more than
- 19:32probably next year. I'll probably be at like 1. I'd say $6 million of, I guess it's like my money,
- 19:39like being diverted to this. I work with three, well, two AI focus engineers, people with like
- 19:47really good backgrounds in computer vision, model creation, all of that. I've got two engineers
- 19:53and then I got two other people on it. But the truth is, the reason why it's going so much quicker,
- 19:59I would have never been able to do any of the stuff that I'm doing right now without
- 20:05like Claude code. Like that has changed the game entirely. This wouldn't have been, I've had this
- 20:12vision for years, but I would have always had to raise ridiculous amounts of money to accomplish this.
- 20:19I'm pretty much like a full stack engineer, so I understand
- 20:23development environments and I don't trust myself with a lot of things, but Claude code,
- 20:28I'm pushing to get like last week, I push like 60,000 lines of functional code. It's just absurd.
- 20:37It's like, it's really absurd. You needed just a couple of years ago, you needed a whole team
- 20:41to do that. Oh, forget it. It's not only an expensive team, but like the feedback loops,
- 20:47like right now, like I have a very like detailed eye. So like let's say we just push some sort of
- 20:52development and I'm like, well, it's breaking on this, it's breaking on this. Now I click the
- 20:57function key on my keyboard and I'm like, listen, you're breaking this, this, this. I want you to
- 21:01switch the framework to using pixel spacing at four pixels, use this JavaScript framework, use this,
- 21:07try it. And I sit there for like two minutes. And then I, I don't know if any of you are really like
- 21:13if anybody listening to this is, you know, there's kind of, let's discuss this sort of state of AI,
- 21:17right? Like there's, there's AI, which in a lot of ways is really these foundational models are
- 21:24powering everything, right? So if you sign up for some sort of app and you're using it,
- 21:28and people will make jokes that it's so funny because people will be like, this is better than
- 21:32open AI. It's like, this is open AI. Like this is, or this is cloud or whatever. It's just
- 21:38modified for the user. But I think where AI can be really sort of helpful is in this building stage.
- 21:46I, it's not like you say a couple things and a full product gets built. But there is no need
- 21:51right now for my team to have a junior developer. So like I'd rather pay my more senior people more
- 21:57and max them out and have them use cloud and all the other tools that we have. But it's, it's a real
- 22:04nuance process. So like, if you've ever coded with cloud, you've, you have this file called like
- 22:12cloud MD, which is like this mark down file that like says everything, like optimizing that file
- 22:17is incredible. And then you can get really like, I'll just say one thing. Like any of you who are
- 22:23in GIT or whatever, like in developers, you could do something called work trees. So work trees
- 22:29basically take, how do I explain this in basic terms? They let, they take an asset and they split it
- 22:37to three separate environments, right? And the three separate environments don't interact with
- 22:43each other technically. So you can work on them simultaneously. And then you merge them. So what
- 22:48I'll do is I'll have like our API layer on the left, our user interface here and are like,
- 22:54oh, they're onboarding process here. And I'll work on them simultaneously.
- 22:59What would you say that 10 x is the speed? It's, it's not even 10, I don't even know the number.
- 23:04It's like, it's like 10 x in the 100 x, you know? And even some developers on my team,
- 23:13like I've had to be like, let me show you something. And there's like, holy shit. And any of you
- 23:18out there listening to this right now that are cross functional, meaning you understand design,
- 23:25you understand a little bit about marketing, you understand a little bit about product development,
- 23:29your technical, but not super technical. You are sitting on a gold rush and gold mine of
- 23:35opportunities because I believe the people with the diverse perspectives for the first time
- 23:41ever can build at speeds that are unprecedented. Yeah. I agree. We may see the one person
- 23:49billion dollar company very soon. Oh, you're going to see it. Yeah. Yeah. Small teams,
- 23:53what they're doing, I'd speak to a lot of founders every day. And it's, it's incredible how fast
- 23:57they can move with very small teams of really good people. And as you say, you know, the downside
- 24:02is that for junior people, it's very hard to get into the labor market where senior people with a
- 24:06lot of experience like us get empowered, can reduce cost, move faster and build cool stuff. And,
- 24:13you know, what we, what we need to keep in mind is that AI doesn't do the judgment for us.
- 24:18And, you know, in my TED talk, I contrast the sort of the raw intelligence that we're getting in
- 24:24our disposal, fast, cheap, abundant. But the wisdom and the human judgment, that's still up to us.
- 24:31And that is a relatively rare quality. So really appreciate you saying, I want to make sure that
- 24:36what I'm building is actually serving humanity rather than just being, you know, a cool business
- 24:41that makes me a lot of money. Yeah. It's like, I've never been one of these people that's like
- 24:47money. I like money or whatever. But if I made the amount of money that I made now for the rest
- 24:53of my life, I would have an, you know, pretty great life. I just like challenging
- 25:02problems and I like helping people. And like, it's been so great that that's been my career
- 25:07for the last like 20 years. But yeah, I am definitely, I wasn't concerned a year ago.
- 25:16A year ago, if you were to have this conversation with me, we did. It's like, yeah, like, I would
- 25:22be like, yeah, it's going to be like, yeah, it's not there yet because I don't think
- 25:29modern society is ready for what's coming. I gave a presentation to about a thousand business
- 25:37owners in Vegas a couple of weeks ago. I gave a talk about non-rule behavior 45 minutes.
- 25:43People started asking questions and somebody asked me a question about AI and I kind of went on
- 25:48this rant and the rant was everybody in this room is screwed if they don't start understanding what
- 25:53this is, right? So, you know, people, the audience was a little bit like agreeing and then on
- 25:59hesitant. And then I, this one guy kind of looked at me weird and I was like, listen, would you be
- 26:04willing to go through an experience right now? And he was like, yeah, I'm totally dead. So he
- 26:09came on stage, I took out my computer, I plugged it into the thing and I was like, tell me,
- 26:14so what do you do for work, right? And he owned a manufacturing company and I was like, what's
- 26:18your biggest problem right now? And he starts talking about like their CRM and how to get leads
- 26:23and all this stuff. And over the next 15 minutes, I asked him questions and I build a system. And
- 26:31he was like, yeah, I don't need to hire somebody now. And like, everybody started clapping and I was
- 26:36like, this is not a gimmick. You're going to have people do this across every industry. And
- 26:43those of you who are not adopting are going to get destroyed by somebody who is thinking how to
- 26:49use these tools. And I really believe, I strongly believe that a society needs to prepare for some
- 26:55sort of universal basic income. Society needs to find a way to like tax AI robots because I think
- 27:03there's the gap between wealth is going to start to get so profound. And we got to use this
- 27:09technology to take care of people that are not using it or else we all lose. So like, I'm a big
- 27:16believer and I don't think there's enough conversations about this. And I hope they're like
- 27:22in America, my country, they're not talking about this at all. But like, I hope it starts to become
- 27:27part of a global conversation. So we're that we're ready for when we're replaced.
- 27:33And yeah, I think so here in Europe, there's a lot of discussion about AI and the threats of
- 27:38it. But I think people focus on the wrong things like you. Are they thinking about, you know,
- 27:42a wild white god humanity or that stuff and I'm glad that people are thinking about this. But
- 27:47you and I can't do anything about that anyway. So I think for the interested public and business
- 27:51people and to pronounce a section, we should focus on exactly what you're saying. How do we give
- 27:55people purpose if their job's going to fall away? How do we actually pay for all of that? Because
- 28:01the current model of taxation in the Western world is we text work at a very high rate and then we
- 28:07give it to some people in need who are retired or school children, you know, for education or people
- 28:14who can't work for some reason. And that's going to blow up. And it's already blowing up because
- 28:20you can see how much tax take there is across Europe, often over 50% of what the country generates
- 28:26in terms of GDP, but still countries piling up debt. So you can see the model isn't working anymore,
- 28:32but no one has the guts to really say, it's not working. We need completely, we think how
- 28:37the tech space is generated, who we text, how or what robots. And what do we, how we give purpose
- 28:45to people who cannot maybe survive a thrive in this particular new brave new world?
- 28:52Yeah, I mean, it's hard, right? Like, it's like, I could imagine, like, I get a lot of value from
- 28:57talking to someone they're like, Blake, you know, you change my, imagine to the point of where I'm like,
- 29:02yeah, I'll talk to you, but my machine's going to be better. Just talk to it. It's like, it's a scary
- 29:07proposition, right? Or just that's what I'm afraid of more than anything is that I don't think
- 29:16people get how much work and the things that they do to live, they get meaning out of that. So
- 29:22some people listening to this right now would be like, listen, if you paid me 100K a month and I
- 29:27hadn't 100K a year and I don't have to do anything, I'd be the happiest person in the world. I get
- 29:30that. But some of some of us would be absolutely like, well, we don't have no purpose now. There's
- 29:35nothing to truly create, nothing truly to do. And it's just, I'm seeing it in so many ways.
- 29:44I'll give you a really weird example. So I have a board game night every Tuesday that I play with
- 29:49friends, right? And we play this game called moon raker. And it's like a strategy cooperation
- 29:55game. And I was sitting there and I was like, you know, it'd be fun. I'm like sitting on the
- 30:00computer and I'm like, I'm sitting on the game and I'm like, thinking through optimal strategy.
- 30:05I got to just, I won. So anybody on the game board game that listened to this, I won last round.
- 30:10But then I was like, you know what? It would be cool to build a model and sort of tested on like
- 30:16GTO and Monte Carlo simulations. See what the optimal strategy is, right? So, you know, that used
- 30:22to be something that like the kind of fun little projects that I would think about and maybe do.
- 30:26I went into claw code and did it yesterday morning. And I had it run 10 hours of simulations
- 30:32and helped me solve the understanding in in five minutes. Like I built it in five minutes,
- 30:38I executed it all day and I read the results on optimal strategy. And it was really made sense.
- 30:43And it was probably the most optimal strategy. So it's kind of like when these things are so good,
- 30:48all of your ideas can be executed by this thing that's more sort of profound. But yeah, I personally hope
- 30:56I don't know if I want to, I don't think people really understand the implications of what super
- 31:01intelligence really is. And I don't think they're doing good AI safety stuff for that. So I don't
- 31:07know if I necessarily want to live in a society with super intelligence. Like it's really hard to
- 31:12understand that wrap your head around that because you don't know what you don't know. But imagine
- 31:17something orders of magnitude smarter than the smartest human. And imagine someone who has access
- 31:24to not only that raw intellectual capacity or that raw processing power, but also to all of the
- 31:30information that us humans have created. It's like freaking Battlestar Galactica coming, you know.
- 31:35Yeah. Yeah. So what can people do to, you know, to compete, not compete with the machine,
- 31:43but to use the machine to compete as I put it. Yeah. So I would say first, you have to acknowledge,
- 31:49first you got to drop all your filters and sort of biases and all this stuff for these things.
- 31:54Yes. Right. So so many people are getting into conversations about, well, this isn't good,
- 31:59this isn't good. This is it. Or I don't think we should use this yet. You don't have to. But if
- 32:04if you have a process, like there's a lot of processes in my company that I still do not trust
- 32:09any LLM to handle. It's me. It's me answering. I still answer client. I don't like automate client
- 32:16bots with somebody sends me an email, anything like that. But what I can do is use AI and use these
- 32:22systems to replace some of the process. So for example, I can use a system that reads an email,
- 32:28order categorizes the priority. And if it's the highest priority to text message me,
- 32:33you couldn't do that before because without the LLM, you couldn't understand the context of the
- 32:37message. So if I were to get an email from like one of my like private clients, it could be like,
- 32:43hey, Blake, check this article out. And another email could be like, hey, Blake, I really need to
- 32:47talk to you this week that different priority levels that we couldn't understand. Now we can.
- 32:53So I think you have to just understand that you're at it. If you're not integrating this in some
- 32:57way in your life, you're just at a major disadvantage. And I don't mean to be like rude or disrespectful.
- 33:04But if you think cutting and pasting into chat GPT or clawed interface is using this in your
- 33:11life, you're not getting the full power of what's happening. It can go way more than that with
- 33:17workflows and all that stuff. And I really believe that the people that understand these tools
- 33:24that lean into these tools that understand how to use them, whether it's 10%, 20%, 80%
- 33:30are going to significantly destroy the people that don't.
- 33:35That's what I also believe. And the way I approach this is non-technical person is not to think
- 33:42about how do I automate emails. But I approach them with different perspective to say, what would
- 33:47it take to save me an hour a day? And that made me love that. So a few years ago, right? That framing
- 33:52made it way more interesting for me because it's like, oh, I would laugh an hour back a day.
- 33:56Cool. Yeah. How can I do this? So I've figured it out from that perspective. So if people looked at
- 34:01this, there's my suggestion. So how do I, you know, what's my big pain point here? Can I save time? Is
- 34:07that a process I really hate doing the, you know, menu or invoicing or whatever it may be? And then
- 34:13just find a way and you could just go to YouTube and there's so many great tutorials, webinars,
- 34:19et cetera, most of it for free, some, you know, small, you know, small money really. And so you get
- 34:24a wonderful education almost for free. And can you apply it to a specific problem you face and maybe
- 34:30join forces with another few people who are working on similar problems and just exchange what's
- 34:36working for you? That would be a relatively straightforward, forward way to approach this,
- 34:41rather than, oh, I heard I should be doing agentic AI, but I'm not really sure so I don't know
- 34:46what to start at least to that paralysis. Yeah. I love that approach. But I push back on one
- 34:52thing specifically for you. Don't let the perception of not being technical stop you. So based on,
- 34:59like I've had two conversations with you, I read your LinkedIn posts, I read your emails,
- 35:06I would be willing to bet a large sum of money that you're probably way more technical than
- 35:12you think you are and you don't need to be that technical because you can ask it to do every step.
- 35:19So this is one of the things that I do. I have, I have something called a learning file. So basically,
- 35:26I'm technical compared to most founders, but I'm not CTO level, right? So what I do is every time
- 35:33I reach a challenge when I'm when we're developing stuff or there's something I don't understand
- 35:39every session, I have it right another markdown file that's called TeachBlake. And it explains
- 35:46everything that it did and gives me resources to dive deeper. So like at the end of every day,
- 35:51I go into that sheet and I read and I drop that into Claude and I have conversations like,
- 35:57oh, I don't really understand why the API request was failing. Can you explain that to me?
- 36:02And then I'm learning in this massive quickly feedback loop. So it's kind of crazy. It's like
- 36:08right now my AI thing is I watch about 40 minutes of random videos on YouTube that I find
- 36:13interesting every night. And then I spend about an hour reading those log files and diving deeper
- 36:21like and just sort of understanding different things. And I'm telling you, my rate, my speed over
- 36:26the last three months has been insane. Like I'm just understanding way, way, way, way, way, way,
- 36:31way more. And then it's really just I surround myself with a lot more technical people and have
- 36:36conversations. But I believe people like yourself that have like this cross disciplinary understanding.
- 36:42That's a byproduct of coaching and seeing so many businesses. I believe you're the most deadly
- 36:49because you see the matrix more than the engineer like the engineer doesn't see maybe the nuance of
- 36:56product and marketing. They just see engineering. So they look at it this way and they're like,
- 37:00oh, well, it's useful for this. But I don't really like how it shows this function. And I don't like
- 37:04this doesn't make sense. And did it. But when you see the bigger picture, it's like,
- 37:10like you get to see how all these things sort of connect.
- 37:13Well, I appreciate that. Like, I appreciate it. And yeah, I'll take that to heart. Maybe I shouldn't
- 37:17describe myself as an entrepreneur. Yeah, no, I think you'd be able to pick it up so quick.
- 37:21AI augmented. And I agree, compared to most people, you know, because I pushed for this. And I like
- 37:26to push myself in a way that, you know, they're forcing functions like the TED talk again,
- 37:30months ago, right? So, you know, I don't know, in three months, I have a TEDx talk. I'm talking about
- 37:35AI leadership. So I better sharpen up, right? I better get that. That could be perfect.
- 37:40Yeah, there could be part like a three-month sprint for yourself. And what I would do,
- 37:44this is where you start. I would go into for this. I usually test all the models. But like,
- 37:49this would be, I think, a better clawed thing. Like, use Opus 4, but here's the deal.
- 37:54Give it all the context you have on you. Explain what you did. And you probably have us written
- 37:58down somewhere. And say, in three months, I'm doing a thing on this. I want to become more
- 38:03technical. Where do I start? I want you to design a daily plan for 45 minutes a day or 30 minutes
- 38:10a day on what you do. And you just follow the plan and interact. And that's the thing. And really,
- 38:16if you talk about education, it's hard as an educator. I make videos and do all this stuff.
- 38:21It's really how people learn. People learn through feedback loops and through a mechanism of
- 38:27like exchange and all that. And that's what these things provide us. And we didn't have that before.
- 38:32Totally. I think learning will completely change from this, you know, just in case to just in time,
- 38:37just in case. And there's a watch the whole video course. And then let's see how I apply it.
- 38:41And it's not really clear. I'll apply to my situation. So you and I have video courses on
- 38:45communication skills, but then you're like, but do you do the right thing with my hand or do
- 38:49I structure my messages correctly? Well, you have no way of finding out. Now with this approach,
- 38:55you cannot easily learn just in time. So you learn a little bit. Then you encounter a challenge,
- 39:01go back. Ah, okay. That's how I've become it. And as you said, it absolutely
- 39:06fast tracks supercharges the speed with which we learn. And also you and I as educators and coaches,
- 39:11we can employ that to our businesses as well. And I'm looking into that as well. We have custom
- 39:16GPs, et cetera. So you don't get the bad average of the internet. You can use my custom GPs,
- 39:21which are properly trained on particular texts, et cetera. But there's certainly more. And the
- 39:27technology allows us to do even more than that. For me, that was just the first step to integrate AI
- 39:32into what I do. Exactly. And like little steps, little, little, little, little steps. Like I
- 39:40tell my friend one of the big steps that I have. So on my phone, the, you know, Apple has that
- 39:45button. So when I click that button, I get chat GPT. So that I'm like using that as much.
- 39:56And just sort of like checking things out there. And also sometimes you don't use it to ask questions,
- 40:01uses an aggregate. Like I'm going from my 40th birthday to the Mediterranean. So I was like,
- 40:05oh, I'm doing something fun. I'll show you. For my 40th birthday, I'm having chat GPT and
- 40:10Claude buy me a gift. So I'm going to say based on everything you know about me, I want you to
- 40:16get me a gift under, you know, two K or whatever. I'll pick a number. And then my wife's going to buy
- 40:21it and wrap it up. And then I'm going to open up with the gift is. And next thing they will buy for
- 40:26you. I mean, I technically could do it that way, but I want it, I want it wrapped. It can't do that.
- 40:34I can't do that. Yeah. Yeah. Yeah. Um, with the Stoke and Jones job, I will run whole businesses.
- 40:39And we said, well, there's still a person who has to go to the notary to set up the business. So
- 40:43that's not going to happen. Yeah, there's, there's still, there's still bottlenecks, but they're going to
- 40:48they're going to be reduced and reduced and reduced and reduced until they aren't any exactly.
- 40:54Exactly. Wow. This is so good. Um, like, thank you. And yeah, of course, if you're interested in
- 41:01what other cool shit Blake is doing, go back to the first conversation we had last year, we talk
- 41:06about as background as a poker probe, psychology professor, um, working with humans and human
- 41:12behavior, nonverbal behavior, what's working, what maybe isn't working. And I took so much away
- 41:18from this conversation, Blake. So first of all, um, what you taught me about, uh, you know,
- 41:22baselines, which taught me about where, you know, what the state of tech really is. A lot of
- 41:26claims out there that may not be substantiated and super cool that you're working on on this frontier
- 41:32of AI and human behavior. And, you know, I love to hear that you want to make sure that this is
- 41:39deployed in the right way. They actually serves humanity because they're, you know, the the
- 41:43changes we will we already see coming are immense and we will all have to learn to deal with that.
- 41:49But I think everyone listening to this can be hopeful because it offers all these opportunities.
- 41:53And even if you're young, I think you don't want to go for some typical young person career where
- 41:57you go to business and learn the ropes for years, that's probably not going to work anymore.
- 42:01But if you learn AI, as you said, over the next three months, you absolute killer. And you can really,
- 42:06you're probably no more than 99.9%, and you will never have to worry, well, you will not have to
- 42:11worry for the coming years about finding jobs or building a business because you'll be
- 42:15had as short as above everyone else. It is that rare opportunity, I think last time it happened
- 42:20was in the 90s when the internet came up, when we, you know, I was young back then and we had the
- 42:27sort of first move advantage, if you like. So we were there, but I, you know, as a back then,
- 42:33non-technical person, non-technical person at the making, you know, it didn't maybe, you know,
- 42:39I'm still off the start of some scale ups and so on, but now it's very obvious what's happening
- 42:45to me. And therefore, I think I may be better prepared than I will expect that.
- 42:49Yeah, I mean, look, you're like your own little clawed summary right there. That was a pretty good
- 42:53summary. That was a great job. Thank you. So Fem's last words I always ask, what's your best
- 43:00piece of communication advice? I would say, I don't know what I said, I don't remember what I said
- 43:06last time, but I am more and more convinced that if you're communicating with, oh, I think I
- 43:15know what I said, so I'm going to switch it. Establish range. So the human body is this incredible,
- 43:22beautiful apparatus. And what happens over time is you tend to stay in your sort of like
- 43:29homeostasis communication zone. And I, with working with people for so many years, it is so
- 43:36interesting how when you push their comfort zone, how much range and how dynamic they can actually
- 43:40be as a communicator, if they just push past that fear and practice. And every leader needs a wide
- 43:47range to deal with the sort of goals and objectives of being a leader. Sometimes you need to motivate,
- 43:53sometimes you need to slow down and apologize in a right way. Sometimes you need to calm people
- 43:58down. Sometimes you need to hide people up. And that all comes from creating more range. So I would say
- 44:02the more you create range and push your comfort zone, the better, more effective you'll be.
- 44:06Yeah. I love that. And you actually, one of the people I quote the most, because you,
- 44:12yes, absolutely, you said the best communicators have range. And I often have these conversations,
- 44:17I like to stay in this box. So like says, the best communicator is a range. And it is so true.
- 44:22And no one has said it better. So thank you for that. No, and that's where, but then I'll give you
- 44:28one last thing on that. So that's where people, people will say, but then they'll give me an example
- 44:32of someone. And they'll say, but this person is x, y, and z. I'm like, you're not understanding
- 44:37the invisible world. Because sometimes somebody has so little range, but so much status, it doesn't
- 44:45matter. Yes, they could do whatever the hell they want. And we're superman like talking for four
- 44:49hours and we're not going to sign this voice. Yes. Andrew's like in this sort of communication
- 44:54range. And it's like almost cute when he steps out of it a little bit. It's like, oh, there's
- 44:57the sweet Andrew. But he's, you know, Andrew Iberman. So he doesn't need that robust range.
- 45:03Exactly. And, um, yeah, totally for that one medium. I'm sure he has more rights than we know,
- 45:08but yeah, so good. Like, thank you so much. Thank you. Always a pleasure.