Active Inference AI - The AI Paradigm Shift Dr David Bray, PhD Wants You to Know About

Speaker 1 – 00:00

Foreign.

 Speaker 2 – 00:13

Hi, everyone, and thank you for joining us today for another episode of the Spatial Web AI podcast. I am so pleased to have our guest here today. He is a very special guest. We have Dr. David Bray with us today. David, welcome.

Thank you so much for being on our show.

 Speaker 1 – 00:31

Thank you, Denise. It’s great to be here with you, and I look forward to the conversation.

 Speaker 2 – 00:35

Yeah, well, so I have to tell you, a friend of mine, Brian Hayashi, when I was talking to him last fall about you, he described you as a treasure. He also said that you’re like the real life young protagonist in the movie Real genius, only you have a. Your heart’s in the right place. And he also said that it’s rare for him to see people who have such your level of understanding of both tech and geopolitics. You’ve been named one of 24Americans changing the World by Business Insider. You’re the former CIO of the fcc, as well as a long list of accomplishments. So, yes, I would love to have you tell our audience a little bit about yourself and, you know, some of your experience and really what is your driving purpose in the world right now, Because I know.

 Speaker 2 – 01:40

I know it’s big and it’s impactful, so I’ll let you unpack all of that.

 Speaker 1 – 01:45

Well, that was very. So here. It was very humbling. And I would say, you know, I think for me, I was just very fortunate. I had two parents. One was a Methodist minister, but he. He was definitely not dogmatic, and he was one that recognized it’s all about personal faith and personal narrative. And while I’m not particularly a religious person, I find there’s value in understanding the personal narratives and journeys that people are on and celebrating them. I think that’s probably how we initially collected, is that you are advancing active inference and helping people understand about it. And for me, one that’s something I care passionately about as well, because I think that’s the better way forward. I think Generative AI will later be looked on as it had some value, but it also had a lot of dead ends.

 Speaker 1 – 02:30

And so anything I can do to help with that. But my mom was also a schoolteacher, but she was a schoolteacher that loved to have fun. And she was definitely not the traditional preacher’s wife. There’s one story, the bishop came to

visit them. It was the late 70s, so fake nails were in style. She’s lighting a candle for the dinner. Her fake nail catches on fire, and most of us would probably discreetly blow it out, but she turns with her thumb on fire and says to the bishop, want a light? You know, and my dad said he saw his short career pass before his eyes, but the bishop left, and, you know, so my journey has been one.

 Speaker 2 – 03:09 Your sense of humor.

 Speaker 1 – 03:10

Exactly. I mean, you gotta have a sense of humor. That’s also how you know you’re not burnt out. You know, if you can find a humor even in some of the most challenging moments, you know, it keeps you going. And in fact, in a past life, I used to do counterbioterism, which, of course, you know, I can remember, like, thinking, like, what type of world requires me to even be in a role of doing counterbioterism? But separate from that, we often used to, whenever we saw each other and we passed by in the hallways, we’d say, are you having fun yet? And that was fun in quotation marks oftentimes.

 Speaker 1 – 03:37

But it was really just a test to see if somebody would find some reason, even in the most challenging moments, the most stressful moments, to recognize that we’re human and we can’t get burnt out. And so, I don’t know. I mean, the short version is my driving purpose is to make sure free societies still exist 10 to 15 years from now. And there are free societies in which you are free to think differently. You’re free to disagree with each other. You’re free to have differences of perspectives and opinions because that’s. That’s. That’s goodness. And again, my mom was Catholic before she agreed to marry my Methodist minister father. And then he was jointly assigned to a church that was jointly Methodist, Presbyterian, Episcopalian at the same time. So, you know, I realized that a lot of people mean well.

 Speaker 1 – 04:21

There’s just differences and the words they use and the rituals they do. When I went off to college, I was friends with a lot of people that were both Jewish as well as Muslim, and so got invited to several seders. And so I celebrate that we humans, even if we think we have differences, we actually have much more in common than we do in differences. But I think we’re also in a world right now where we may have built a whole lot of technologies and other. Other systems that currently benefit from accentuating the differences in groups as opposed to similarities in groups. And so that’s why I say I’m both a nonpartisan that’s sworn O to the US Constitution, but also a globalist, which generally means I’m shot at from all sides.

 Speaker 2 – 05:08

Which brings me to a comment you. You made to me recently about Dorothy Parker’s short poem, Neither Bloodied nor bowed.

 Speaker 1 – 05:18

Yes again, that’s again, a sense. She had a sense of humor back then. You know, it’s like, you know, I’m doing it because you got to have a sense of fun, but you also got to be yourself. And I know no other way than to be

someone that tries to create a big enough tent that brings the differences of opinions together. And you know, just to give a little side, I’ve actually had a group and I continue to sort of help with a group that’s got, and now it’s got 300 different technologists and data people that have promised to be nonpartisan, non attribution. So we don’t really talk about who specifically said what and non absolutists. And it’s now about 300 people and it’s very lively as you can imagine, here in 2025. But so far they’re not killing each other, which is good.

 Speaker 2 – 06:01

That’s great. Yeah. And you know, it’s interesting because besides being able to recognize and appreciate differences in people, because to me it’s, you know, it’s interesting because when you think of active inference AI and this distributed network of AI that’s coming, it depends on the differences, it depends on all of the variations to be able to kind of push back on each other and come together so that knowledge and wisdom and all of that can grow. Otherwise it doesn’t, you know, but with people, with humans, I feel like we’re stuck in this kind of situation these days where people, you know, they recognize differences but they take an affront to it, you know, and, and that doesn’t allow growth at all. And it’s really disheartening because people also seem to be coming from their own agenda, tied to their Persona.

 Speaker 1 – 07:06 Oh yeah, everyone’s got.

 Speaker 2 – 07:08

And that complicates it even more because then they become passionate about the wrong thing, in my opinion, 100%.

 Speaker 1 – 07:15

Well. And I look back, you know, I started off as a person who built computer simulations. That’s what led the government offering me a job when I was 15, which was kind of crazy. Then later did stuff when I was 17 that was classified also in that area. But I was doing simulations, if anything, to try and understand what. Understand natural systems, but also understand what we didn’t know. One of my first simulations was this was 1992, 1993, ozone layer deterioration and then later predicting where forest fires might go back in 1995. I couldn’t fully declassify it, but it was more to beg the question of what are we missing? What is absent in our understanding of this phenomena than to say, I know everything. And I think what we’re facing right now as a world. And you’re absolutely spot on about active inference.

 Speaker 1 – 08:01

I mean, if you go back to the founding United States, I mean, the founders had vigorously vigorous disagreements about how to organize, and that was actually valued. I mean, the Federalist Papers. I’m a big fan of the debates that go back and forth. I think the sense of uncertainty, the sense of uncomfortability at either the speed or scale of change has caused people to dig in because they’re feeling anxious. And the brain can only feel fearful for so long before it has to either shut down or it starts to get angry. And so people have shifted from being anxious about the future to now being angry and, like you said, feeling grievances.

Speaker 1 – 08:36

I mean, the latest Edelman Trust Barometer from 2025 says not only is trust down across the board in ways very similar to what we saw in the 1890s with the original Gilded Age, when there was massive technological progress, but now people have a sense of grievance against different groups. And I think, again, that’s. If I was to be empathetic, I would say it’s because they’ve shifted from anxiety to feeling like I need to present that I’m right, you’re wrong, you have wronged me. And I just am trying to encourage people, like, look, free societies are at their best when we’re curious and we’re learning from each other, because as you said, if we become echo chambers of thought, that becomes echo chambers in reality, and that’s a source of stagnation as opposed to growth.

 Speaker 2 – 09:21

Yeah, yeah. And. And it’s interesting because I think that people take the. I think our society has kind of gotten into this pigeonhole of if someone has a different opinion than mine, it’s not that curiosity it, like you’re talking about. It’s, oh, that means there’s something lacking in me or something. You know, like they take it as a personal offense that they have to stand up to rather than a way of learning and sharing and growing and. And that it’s okay to have different opinions. It’s good to have different opinions. And I don’t know how we change that. I feel like social media has been a huge factor in conditioning that. And so how do we change it?

 Speaker 1 – 10:09

It’s interesting because I was having a conversation with someone from the Buffett Institute yesterday about a possible future event we could do, and were actually addressing that Question. And I said, what if the underlying problem is we have built a free Internet? I’m doing free in quotation marks because we know it’s not exactly free. Whose underlying model is ad revenue. Yeah, and ad revenue. You don’t usually buy something when you’re told everything’s all good, nothing to worry about, everything’s fine. We know that even if you don’t tell the algorithm explicitly, and I don’t think people are that.

 Speaker 1 – 10:49

As the algorithm searches for what is the most likely way to get the most viral engagement, we know that what happens is the way to have strongest viral engagement is to get one group angry and another group angry in that group’s anger and have it be this reverberating wave that goes back and forth. The second number way is to have fear, like fear. And anger goes back to our evolutionary root as a species. Again, evolution is a blind watchmaker. There’s no intentionality. But in our evolutionary history, 150,000 years ago, 100,000 years ago, those of us that paid attention to that moving grass and said there might be something that’s going to eat me or come after me versus those that saw the moving grass and said there might be a wonderful surprise, you know, which ones are more likely to survive.

 Speaker 1 – 11:33

And so, you know, evolution has resulted in us being creatures. That content that is angry and fearful unfortunately goes halfway around the world before a content that says, look, it’s more nuanced. It’s going to be okay. You don’t have to worry. And that’s where I know you and I are both friends of Carl Friston. I think his diagnosis that we may look back 10 to 15 years from now and say, through a combination of the Internet followed by the smartphone and now generative AI, we have built an infosystem or infosphere, where we are overloading our brains in ways completely foreign to what is best for them. I won’t say it’s the equivalent of mental carcinogenics, but might be close.

Speaker 2 – 12:20

Well, we’re continually exhausting our brain energy capacity. Like, I know I, that I have moments where, like, I pick up my phone and, you know, it’s because it’s all engineered to take us down these rabbit holes of attention capture. You know, like, I’m trying to check the weather, and before I know it, I’m 20 minutes into a bunch of notifications that came in and I’m like, I just wanted the weather. And, you know, there are times where I just Pick up my phone and I’m like, I just can’t. I, I can’t take it.

 Speaker 1 – 12:53

But I think that’s actually healthy. Like when I, on my phone, I turn off all the notifications with the exception of texts and calls. So if you text me or if you call me, I’ll get a notification. Everything else, you send me an email, the app says nothing. Like, I feel. No, you know, because otherwise I know if it showed me a number, I’d be like, I have to check. I have to check. So, but I think so maybe we’re going through as a society, we’re going through the same thing we saw with the radio. Like when the radio came out. Marconi, as one of the, as the inventor of the radio, said that the radio was going to cause world leaders to talk to each other on a daily basis. There’d be peace and understanding because nobody would like, ever misunderstand each other.

 Speaker 1 – 13:28

Yeah, I don’t think radio achieved that peace. But then later, come 1933, some pundits were saying the radio is going to be the end of Western civilization as we know it. It’s going to be the dictator’s toolkit. Yeah, I don’t think that happened either. And so if we look back at the history of the Internet, smartphones, social media, we had this wave of it’s going to lead to truth and understanding. We’re going to find everything online and we’ll all be enlightened and then social media will all be goodness. We’ll all connect and we’ll overcome geographical barriers. And maybe what we forgot was to factor in the fact that humans are humans and we do wonderful things. We do a lot of benign things, including cat photos.

 Speaker 1 – 14:05

And sometimes if things are out to sell us something, they’re going to unfortunately play upon some of our more sobering tribal tendencies as opposed to our uplifting whole of humanity tendencies.

 Speaker 2 – 14:20

Yeah, very true. So let’s shift a little bit over to AI, because I know we have a lot of opinions in this space and I saw you for the first time and the last time out in Half Moon Bay here in California back in October at the Constellation Connected Enterprise event there. And it was really interesting because a lot of the feedback that I was getting there and you know, I was there to speak about active inference AI, which is completely different methodology than deep learning, which is all the current AI that we’re seeing. And a lot of the feedback that I was getting there was that, you know, it was AI, you Know, AI fatigue. Right. You know that.

 Speaker 2 – 15:10

I guess the year before, which I wasn’t at the event the year before, everybody was excited because they knew they had to figure out how to integrate AI and it was all brand new. And then this year, after having to dig in for a year and realizing that it falls short of performing in the way that they need to actually get the results they want, people were just like, don’t talk about it. But yet they know they can’t ignore it. And, you know, that was a great opportunity for me because it really opened the door wide for people to be receptive to. There’s something else coming, and it solves all these problems. When you were there with me, somebody had asked you, what do you think of it? And you’re like, I’m

bullish on active inference.

 Speaker 1 – 15:59 Yes, let’s caveat. Generative AI.

 Speaker 2 – 16:01

I would love to have you explain to our audience what is attractive about active inference AI to you. What drew you to this?

 Speaker 1 – 16:09

Sure. Well, so first I would say you’re absolutely spot on. I remind people that over the last two years, if you’re experiencing AI fatigue, or specifically generative AI fatigue, that AI has been around in different flavors since the late 1950s. The only reason why we’re all talking about it right now is marketing, marketing. And there’s a whole lot of marketing dollars that wants to make you think that this is somehow the next revolution for Silicon Valley. And I think probably where I first got connected with Carl Friston was actually the summer and fall of 2023. And so again, partly when I was hired by government, early age, it was at the time it was involving expert systems and decision support systems. So I have seen AI ways of excitement and AI winters and you know, in machine learning.

 Speaker 1 – 17:02

Yeah, there’s a lot of interesting things. And then now if what’s possible with generative AI. And so. But I also did as an undergrad and also as a graduate student, I did evolutionary biology. And so I think what attracts me to active inference is it is not simply throwing data against the wall. Massive numbers, amounts of data that is questionable about how did you get that data or not. I often tell people we may be repeating the lessons of Napster circa the early 1990s, where it’s unclear how are you respecting intellectual property? How are you respecting equity of the people that may have been artists or musicians or people that created that data?

Speaker 2 – 17:40

Yeah.

 Speaker 1 – 17:41

Two, as one who has worked in the policy space, I saw in the 2021-2024 a whole lot of AI policy that was attempting to get around the fact that generative AI will probably never really be explainable. The whole nature of neural networks, and I often caution, just because they’re called neural networks does not mean it’s actually like the human brain. But regardless, neural networks, by their nature, are not explainable.

Speaker 2 – 18:10

You know, one of the things that stuck out to me at that event was there was one speaker, and she was great. She, you know, she had a lot of great input, but one of the things she said to the audience is, you just gonna have to get used to not knowing. I just thought, oh, my gosh.

 Speaker 1 – 18:29

Well, that’s a very Silicon Valley perspective. And I say that as one who has interacted with him, because I see them come to DC every once in a while. I’ll give you another. This was before the wave of generative AI, but we had people that came from Silicon Valley, and they were on a circa 2014, it was 2015 panel, and the Silicon Valley person was telling government, your customers want to be surprised and delighted by how much data you know about them, and they want targeting, targeting. And I’m like, no, don’t think so. And sometimes the Valley is doing really great things, and sometimes I’m like, yeah, that. No, you know, and so I agree. I think the.

 Speaker 1 – 19:13

The unknowability, while it’s probably true, you know, we don’t have explainability for a lot of human phenomenon, the answer is not to get comfortable with the. The unknowing and unexplainability. The solution is to both improve our machines and human systems to have more explainability, because otherwise we are going to continue on what. What is already tracked in the Edelman trust barometer for 2025. Trust is down across the board partly because people don’t feel like they have agency, but also partly because they don’t like how information is both delivered to them and how information about them is used. And so I actually see active inferences hopefully turning around that as more companies start to adopt it.

 Speaker 2 – 19:49

Yeah. You know, and I think that’s a great point because, you know, I mean, the whole way that active inference works with the free energy principle is to close that gap on uncertainty. But also these agents can report on their level of certainty or uncertainty. Right. So you can trust this system. It’s. It’s something that you can trust. And, you know, I think when it comes to people with fear, Right. You know, fear largely comes from uncertainty. It’s the unknowing. Right. I mean, that’s really where fear comes in for people. And so I think that this really does have the potential to, you know, kind of improve on both of those levels.

 Speaker 1 – 20:28

100 well, and I tell people, if you define trust and the definition I use, if you define trust as the willingness to be vulnerable to the actions of an actor you cannot directly control, then that actor could include an individual, it could include an organization. I mean, we all have elements of our life, but also includes governments, includes militaries, includes civil societies, and includes machines. And so I think that the biggest question is, you know, and I don’t fault Silicon Valley.

 Speaker 1 – 20:59

They’re focusing on the immediate and they’re trying to get MVPs out the door to get return on investment, but they haven’t stepped back and said, as I unleash this tool or this algorithm into the global marketplace, is it going to increase or decrease the receptivity of people to want to be vulnerable to the actions of this actor that they cannot directly control? And I think for the most part we have spent the last 30 years releasing technologies that remove people’s sense of agency and that’s not good. And again, we saw this the same thing in the 1890s, so it’s not like we haven’t been here before. But the solution was in the end of the day to turn it around and figure out how we do that.

Speaker 1 – 21:41

And so I’m a big believer you actually don’t do that through top down approaches because that itself removes agency. We’ve got to find better ways of using these tools so that they are human centric, that they’re uplifting individuals just as much as communities. And they do give you a sense of choice and agency that you can then impact what you want done either in your personal lives, at work or in the global world.

 Speaker 2 – 22:05

Yeah, and that’s honestly one of the things that’s most attractive to me about the Spatial web protocol. You know, this new protocol layer right on top of the same Internet we’re talking across right now, you know, but it, it bakes in this ability to have that.

 Speaker 1 – 22:23

Yeah, well, and what I would say is what I love about what you’ so, you know, after meeting Carl, I got pulled to. I don’t normally do Davos because there’s usually 100 people in a room meant for 50 people having a conversation that nobody can hear. But this was, it was the chance to do five dinners with him and others, including some folks from Versus on Active Inference. And this was January 2024. And what was interesting was were testing the market.

 Speaker 1 – 22:52

I had a hypothesis that the market wasn’t ready, but were trying to see and we found that for the most part y there are at the time, and I think even to this day, although we’ll see now, with Deep Seq out there’s a lot of people doubling down on this idea that if I build a moat based on either my data and, or the fact that I require high degree of gpu, that house that’s somehow going to be a way I will maintain a marketing benefit. And so active inference in some ways is seriously disruptive for that business model. But now that you see, and again, I have no doubt that a lot of the claims of deep seek have been seriously exaggerated for whatever public relations purposes.

Speaker 2 – 23:33

Yeah.

 Speaker 1 – 23:34

But it may indicate that there will be some investors and backers looking for the next as opposed to doubling down on the old. And I think what I love about what you’re doing is you’re making it accessible and educatable in a way that can be reached by people. Because I found even after a two hour dinner with about 20 people and Carl is very eloquent, people were still wanting to learn more. And so this is a space where you got to do a lot of education.

 Speaker 2 – 24:00

Yeah. You know, and that honestly that’s when, that’s what I’ve seen for the last two and a half years. Like I, you know, I’ve kind of been watching this tech unfold over the last seven, eight years and two and a half years ago I realized, okay, they’re actually doing it and you know, it’s a couple years out until it becomes tangible. And I knew that, you know, this is a huge mindset shift. And with that kind of a shift, it takes time to change minds. It takes time

to introduce new ideas and new ways of thinking, especially when they kind of fly in the face of the norm. And right now, I mean you have all these leading giant tech companies. They are, their entire business model revolves around the centralized approach, which is very.

 Speaker 1 – 24:53

Unlike the early days of the Internet. It is the opposite of how the Internet early got started, right?

 Speaker 2 – 24:57

Yes, exactly. But that’s how they’ve, that’s how they have grown to be these giant tech companies because they do take advantage of the fact that all of the transactions take place under their umbrella. So they own it all of the data, everything. So it’s no surprise that they’ve wrapped their hearts around deep learning, the giant database that has to process all the Queries.

 Speaker 1 – 25:17

They know how to monetize that business model. Yes. Or they think they do. Again, it’s worth knowing they haven’t made a profit. As far as I know. They may have made revenue and they may have made profit off of cloud. But I often when I brief people right now, I say it’s worth knowing these companies as far as we know, have not made a profit off of generative AI alone. And so every action and every message you see from them might be explained through that lens.

 Speaker 2 – 25:38

Yeah. And so I, I personally think that because, you know, you can’t put the genie back in the bottle. Right. We know from especially like you just this year seeing some of these benchmarks and tests and stuff that Versus has been putting out. I mean they are, they use 90 less data, they’re, you know, 100 accurate. They’re using, you know, a fraction of a model size for, you know, it’s just, it’s a no brainer because this deep learning approach, this giant database with these superpowered GPUs, I mean, you know, they’re talking about opening up nuclear plants and doing all this stuff to sustain it. It is not sustainable. Yet you have a deep seek, I feel like is the canary in the coal mine. It’s like kind of showing the world this is not sustainable. Even in the deep learning sphere.

 Speaker 2 – 26:37

We’re doing it better right over here you still have something that blows that away as far as energy efficiency. And then this decentralization which does what you were talking earlier, it puts the power back in the hands of the people, the agency, over your data, over your being able to decide those things and what’s done with them in your life. But all of that flies in the face of this model. That’s, that’s so I think we’re going to see a lot of Kodak and Blockbuster moments. I think that a lot of these big tech companies, they’re going to hold on as tight as they can to that centralized old way of thinking they’re untouchable. And then you’re going to have people cropping up out of seemingly nowhere.

Speaker 1 – 27:23

Like right when they’re there. When I think that’s the fascinating thing is the US is at its best when there is a healthy tension between the fact that, you know, again, in the United States Constitution we have freedoms, we have decentralized. I mean we are highly decentralized. I tell people we have the United States has more police jurisdictions than the Canadians have police officers. That is how decentralized we are. But, but that is actually a source of goodness. That’s actually how we do checks and balances. That prevents any one person from getting too much power. And we did throw a war, we overthrew a king. So we don’t like king like individuals. But I raise that because there’s that tension because right now I think there are those like you said, that have figured out how to monetize if they do centralize.

 Speaker 1 – 28:09

But we know that centralized path, you know, not only does it take away freedom and takes away choice over time, but it also is creating the very trust, barriers, the divisions and everything that we’re seeing. I find it absolutely fascinating that the People’s Republic of China, which is a centralized autocracy in terms of government and I realize their businesses may not be as, as disconnected as we separate our private sector. I think Deep Seq was a very well planned psychological operation to undercut in some respects the U.S. Companies.

 Speaker 1 – 28:45

Now they’re doing it because here’s what I would say is geopolitically, if generative AI requires massive amounts of data, time, compute and money to sustain itself, it may very well be that private companies by themselves will not get to any source of profitability with generative AI, whereas autocratic nations that prop up their companies and everything like that will. So I, that’s where I think we actually need to step back and say look, if we really care about freedish societies and we care about what that is, what we need to do is actually push for decentralized approaches that are privacy preserving, that still give people choices. And I’m not going to say the specific methods, although I think active inference is one of those. We may discover others.

 Speaker 1 – 29:31

Yeah, and what we want to do is of course because we don’t want this to be bankrolled by the government long term. You can’t have this be, you know, a continuous source because that’s not necessarily good news. I think the business model is instead of you making money off of people using your platform, it is much like Red Hat where Red Hat helps with Linux but they don’t control Linux, they make money off the consulting of how to use that. And again the challenge is for VCs. It’s hard to figure out how do you do a VC scalable market off of consulting. But that might be what we need for a future in which we do have freedom and we are allowed to disagree.

 Speaker 1 – 30:08

And ultimately, I know people talk about agentic AI and I always cringe because when they say the word agentic AI, they don’t really mean agents. They mean things tethered back to the central mothership for the most part, unlike, like. But this would actually allow us to have a future in which we each have a hundred, if not more agents doing different things on our behalf. Yeah, and it’s. And I. And I actually raised this back in 2015 when I was visiting Taiwan and comparing Taiwan’s 24 million people to China’s 1.2 billion people at the time. I said, do we need to have almost a digital golden rule, which is do unto others as they give their consent and permission to have done unto them?

 Speaker 1 – 30:54

And I think that only becomes real when we can have agents where we can have the confidence those agents are actually working for us versus some centralized mothership.

Speaker 2 – 31:04

Yeah, totally agree with that. That’s why to me, you know, it makes so much sense when you combine the. The power of the spatial web protocol, you know, the hyperspace transaction protocol, where now transactions are at every touchpoint in this digital twinning of digital representation of all the physical things so that the programmed contacts can be there, the programmed permissions can be there. They’re in the hands of the user rather than, you know, just a blanket consent to the giant company that you need.

 Speaker 1 – 31:38

You can bound it. You can say, like, when I step into my house, I don’t want the following things to happen when I’m in the geographical boundaries of my house.

Speaker 2 – 31:45

Yeah.

 Speaker 1 – 31:46

Oh, you can also have restaurant owners that say, the moment you walk into the restaurant, you know, you can’t have the following actions happen in my restaurant. Or when you walk into an airport, because an airport is a public space. You can also say, look, if you’re going to board that plane, you have to give some data to prove who you are. Otherwise you can’t board that plane. And so it allows us to really have choice architectures defined by both geography and time.

 Speaker 1 – 32:08

And for things like medicine or law, as you know, that don’t necessarily have geographical boundaries, you can still express policy formulations we could actually have, whether it’s business process modeling for medical care or even medical policies that say, here are the things that the individual has given their consent to that they do want to be revived if the following conditions are present. But after, you know, after more than 30 minutes of attempting to revive them, if they’re not showing things or signs or anything like that, then take this action instead. Yeah.

 Speaker 2 – 32:38

And being able to set even expirations on your consent and everything which will be totally possible in this new protocol existence powered by These individual agents, you know, and there will be kind of a zero knowledge proof environment too to where you don’t have to give a third party any sensitive information. It’ll be known within the checks and balances of the nodes themselves. I mean there’s so much more security. I, I always tell people who are like, oh my God, AI, you know, oh, you know, they’re just, it’s like we’re just trying to use all these emerging technologies in the most unsecured environment of the World Wide web.

 Speaker 1 – 33:18

But we’re what could go wrong?

Speaker 2 – 33:20

We’re progressing into a space that will be much more secure and then these agents will actually be able to have a deep understanding of the world in real time as it changes moment to moment, you know, so they’ll have a grounding layer that is not there with deep learning. There’s no understanding of the actual world now in the moment.

 Speaker 1 – 33:42

Oh no, exactly. It’s simply pattern matching. Really advanced pattern matching about the world. Yes. You know, any, anything that looks like causation is simply just really advanced correlation with generation.

 Speaker 2 – 33:53

Yeah, absolutely. And that’s actually, you know, that’s the beautiful thing about what we’re entering into here. And so it’s funny because, you know, getting back to like education about this space, I, that’s what I see as the most important thing right now is letting people understand what this difference is, what this, what is really coming.

Because let you know, the genie doesn’t go back in the bottle. This is coming. And I think this deep SEQ thing was interesting because it was a canary in the coal mine. It let people know Nvidia is capable of falling.

 Speaker 1 – 34:33

And Nvidia may have an interest to convince you that you have to have exclusive GPU chips to do these AI models as opposed to and with active inference.

 Speaker 2 – 34:42

It’s all the processing gets distributed across the Internet, across the network at the edge devices, the computer, the phone that, you know, processing these AI queries, the compute necessary just being distributed. So it’s going to be a huge disruption and it’s to me, I, I also see it as we are embarking on potentially the most innovative time period ever for mankind. Because it’s an open source dream. It puts it into the hands of anybody to be able to develop these AIs and the fact that HSML is a common language that’s going to bridge all emerging technologies across the network. It’s not just AI, it’s everything.

 Speaker 2 – 35:30

The biggest blocker has been lack of interoperability so to me that’s what really excites me about the future and what’s coming is I think that there’s going to be innovation in ways that like logically we can say, you know, this is going to happen. But we don’t know.

Speaker 1 – 35:47

Oh well, we don’t know. Well and I think the. So, so I clearly I, I support the same vision you have. I’m also one who has seen. Well, so you know, we mentioned Dorothy Parker, but let’s also go to E. Cummings. E. Cummings has a poem where he says, oh humanity, how I love you. You are constantly putting the secret of meaning of life in your back pocket. It sitting down on top of it and smushing it. And so I think, you know, as we’ve got this really wonderful, well thought out promise that actually does show a better way. I think that’s when I then have to go up to the balcony and say, okay, in the United States right now we have vested interests that see doubling down.

 Speaker 1 – 36:31

You know, it’s odd to think of generativity as the status quo, but it is the status quo. Doubling down on that is either going to help them get more money and, or more power. And I think as a realist, you know, I often tell people, you know, when anyone says the system’s broken, I’m like, no, the system is perfectly calibrated to produce the results you’re seeing. You just may not like that. And that the way you change the system is you got to figure out how to change the incentives or the motivations that are resulting in the emergent outcomes you’re seeing. And so again, I love that you’re doing the education.

 Speaker 1 – 37:05

What I am trying to figure out is what we really need to find is a one or more multiple US Companies that are willing to amplify what is already the proof points you mentioned versus there’s also how. So there’s some companies out there that are showing this and of course there’s always the active inference conferences that Carl does too. It’s got to be a company that is willing to say, I am willing to place a bet that Gen is not the future. This could be and I think there’s a lot of evidence for it. And then that’s when you’re going to start to get the catalyst to make it happen.

 Speaker 1 – 37:41

Because I think right now what’s holding us back is, and I, I recognize there are people that are forward leaning, but there’s also a lot of people in Silicon Valley that are actually want to be the second mover. They don’t want to be the first mover because you get shot at. So the question is we need to find a large enough first mover to give this airtime moment. Right. That maybe recognizes that either A they were behind the curve on generative AI, they don’t want to be the behind the curve on active inference or they’re realizing that they’re hitting the limits of what generative AI can get them. Whether it’s in terms of again like you know, Deep Seek was a shot across the bow.

 Speaker 1 – 38:16

I will be interested in long term one when we get the whole story how much it really cost and how much time it took. Also Deep Seek, I’m also interested whether you call it AI colony collapse or model destabilization because it’s an AI trained on the AI, there’s some risk there.

 Speaker 2 – 38:34 Right, right.

 Speaker 1 – 38:36

I would not trust Deep Seq for some really, you know, it’s something that impacts life, liberty or property.

Speaker 2 – 38:43

Yeah, you know it’s going to be interesting to see you know because I know that like as far as with Versus and you know their genius platform where you can model these agen and they’ve been in beta and they opened it to the wider developer community last summer but they’ve got some big players who are part of this beta. NASA, JPL is in on it. Volvo, you’ve got Cortical Labs and you’ve got a lot of people doing some major things Analog and they’re working with Analog to do the Abu Dhabi Smart City project. This morning I just saw something announced with a telecom company in Australia to, that they’re working with to go after fraud, telecommunication fraud using the technology.

So it’s going to be interesting to me.

 Speaker 2 – 39:38

I think once people really see the impact, the real world impact of what this is capable of, I just, I think it’s a no brainer, you know. But I, you know, I, I do think you’re right in the sense that you know, we need to see more and more support of that. But again, I think that comes down to education too. Right. You know, where even these large companies they barely understand what deep learning.

 Speaker 1 – 40:06

Is other than I can monetize it and everybody else seems to be doing it. So I will monetize it. Yes, exactly.

 Speaker 2 – 40:12

You know, and honestly that’s one of the reasons why like I just, I just did a soft launch of my education platform. I’m actually going to be, I this week. I’m inviting a lot More people in. And later this week I’m doing a big announcement to.

 Speaker 1 – 40:30

Nice. Excellent. We’ll have to amplify it. Excellent.

 Speaker 2 – 40:32

Yeah. But I’m offering a, an executive program. I’ve created seven courses and to me these courses are the prerequisite knowledge that everybody needs to know to understand what these changes are, what this is capable of doing and what it means to, you know, individuals or businesses or, you know, anybody who’s going to be making decisions based on this technology. You know, you have to understand what, you know, what this is and all the various aspects of it. So, you know, when it comes to.

Speaker 1 – 41:08

Well, you’re doing the right thing. I mean the number of times in my life, I mean, I think the thing that I always fall back on is it was. So I joined the biochar’s preparedness response program In November of 2000, the Agile Manifesto at the time. So Agile development, which we now all take as again, it’s not a panacea. But most folks would say if you’re not at least thinking about where agile fits and then there’s sometimes when it doesn’t, then you’re missing a boat. But initially the Agile Manifesto came out, I was an early adopter. I was the IT chief for the Biotarism program. But I was being told follow the five year enterprise architecture, follow the three year budgeting cycle.

 Speaker 1 – 41:44

And so I literally had an email that I sent in June of 2001 where I said we don’t have a deal with Mother Nature or bad actors not to strike until we have our IT systems online. But I was still being pushed or like you’re being disruptive.

What’s this Agile thing? Why are you adopting it? And it was actually scheduled weeks in advance for me to brief the CIA and the FBI as to what we would do technology wise should a buyer trans event happen. That just Happened to be 09-11-2001, 09:00 in the morning. We obviously responded, you know, the briefing was postponed because 834 the world changed, stood down from my own. October 1st I briefed the say on October 3rd. First case of anisrback showed up on October 4th and then everything else went from there.

 Speaker 1 – 42:21

Had we not done Agile development, we would have had to handle 3 million environmental samples and 300,000 clinical samples by fax versus digital. But that’s where I think it’s sort of similar here, where you know, one can, one can ask the question, had 911 not happened. And again, I didn’t want 911 to happen. But had it not happened, I would have probably been perceived as a heretic and after two or three years, gone on and done other things. I think with generative AI, the question is, you know, again, you’re doing the right thing, you’re getting the education out there.

 Speaker 1 – 42:54

And maybe it is going to be more of the deeper scrutiny as to can you actually defend your business model with this massive compute strategy or is that going to be undermined by geopolitical actors, in which case, dear United States, dear Canada, dear Europe, Australia, New Zealand and other free societies, maybe you need to invest in the next as opposed to doubling down on the current.

 Speaker 2 – 43:18

Right, and the next being the self organizing, adaptive, intelligent intelligence automation that actually.

 Speaker 1 – 43:25

Builds cumulatively uses much less compute and uses much less energy. Exactly.

 Speaker 2 – 43:29

Yeah, yeah. Oh gosh. Well, okay, so let me ask you this then. You know, what do you think that people should be focused on right now in this space? If you’re talking about a government or if you’re talking about an enterprise or

even just an average citizen, what, where do you think their focus should be as we transition in this space?

 Speaker 1 – 43:54

So I’ll do the easiest one first, which is an enterprise, small, mid, or maybe small to mid size, possibly slightly big enterprise, you know, talking to friends that are CIOs or CTOs. I mean they’re already underwater and they’re getting bombarded on all sides by hype, by fear, by fud, fear, uncertainty and doubt. So I would say one, take the time to get smart where you can. I know that’s hard, you know, and I often get people to say, like, where’s the one place you go to get all this information? Like there is no one place. But you know, obviously you can go to proactive inference, but I mean, there’s obviously other tools. I mean, you know, yes, they’re underwater with cyber security hits, hardware questions, all this sort of stuff, supply chain. So I think take the time, that would be worthwhile.

 Speaker 1 – 44:35

And I would say simultaneously, enterprises should be making the case to their boards and say we need to spend 5%, maybe up to 10% to start placing some strategic bets. And part of that should be not just placing bets on the existing generative AI approaches because again, AI has gone through waves of, you know, expert systems, decision support systems, machine learning and now deep learning. We would be, you know, we would be remiss to not think there’s going to be other approaches so include some active inference in it. But the other reason I tell people is if you don’t make that case to your board that you need to be spending 5 to 10% on some future leaning things while 90% doing the must do. Now the boards will create some other position that somebody else will do it for you.

 Speaker 1 – 45:18

You know, and so I think it’s important the individual that’s a little bit harder because a lot of us are really, I mean most people are just trying to get through the day. They’re just incredibly busy. They’re bombarded by how do I make sense of it? I would say if you have bandwidth, educate, and that’s where obviously you’re providing that. I would say even if you don’t take the time to recognize, we have been through waves of history before where massive changes happened. It has felt anxious, it has felt turbulent. And so if you are feeling anxious about the here and now, one, know you’re not alone. But then two, it’s going to be okay.

 Speaker 1 – 45:57

You know, it may not feel like that, but I think that’s where I think for everyday people, I’m almost like, you know, give yourself the permission to recognize that yes, you do feel anxious, but it’s going to be okay. And that there are forces that are going to tell you it’s not because it serves their interest, but it will. And then finally for governments, well, I think it’s harder because government is trying to reach for levers that it used in the past. But in the backdrop of the last 30 to 40 years, I tell people we have succeeded in giving people and organizations the capability of the CIA and the KGB circa the mid-1970s, but we haven’t upgraded society to deal with the fact that you have those capabilities. I mean, your smartphone can call anybody at a moment’s notice.

 Speaker 1 – 46:47

You can, you can geolocate yourself or if you have permission, geolocate others. You can download commercial apps that have satellite footage that’s at 0.25 meter resolution that’s possibly 15 minutes recent. You know, that old. And so I tell them like President Carter, President Reagan, Presidents Bush 41 and 43, President Clinton would have loved to have had your smartphone as part of the Situation Room in the White House. And so for governments it’s, you can’t use the same levers you did. And so you’re going to have to figure out ways to work not just with the private sector, because if you only work at the private sector, they may steer you to dead ends that serve their own financial interest.

Speaker 1 – 47:28

You got to work with the private sector, but you’ve also got to work with other voices raising future perspectives, whether that be Startups, whether it be people like yourself, other people that are saying, look, if we simply double down on the status quo, especially for free societies, we may find that we get ourselves in really bad geopolitical waters where we have gone down a blind alley and other nations that can be more force directed have out competed us. And that to me would be a loss for free societies if we find ourselves out competed. But I also realize in the original Cold War between Soviet Union and us, we just had to be more competitive than the Soviet Union and they had their own war challenges.

 Speaker 1 – 48:10

I think there are now state actors and non state actors that have figured out ways at the moment to be faster in their competitiveness than the US because they’re leveraging whatever strengths they have. And again, I’m a globalist at heart, but I do celebrate the United States and free societies.

 Speaker 2 – 48:28

Yeah, absolutely. And you know, I think that one of the ways to kind of preserve those qualities is through community. And you know, as a global society, you know, we can have global community that then we can take back to our regions, our environments and nurture them to have, you know, to maintain those, you know, positive qualities and everything else. And actually that’s another aspect of what I have launched here with this learning platform is there’s an entire community kind of social networking side to it. So people can go and create, you know, you can post in your newsfeed, you can talk to other people, you can follow them, you can meet people who are also interested in this and find out what you’re doing and how you can kind of mesh together.

 Speaker 2 – 49:19

But you can create focused groups and invite people into it and have very focused conversations about different aspects of technology, different regions, different industries, you know, all kinds of things. So I think that’s, I think that’s really necessary when you’re talking about such a big shift and kind of keeping everybody aligned towards, you know, being a force for good in this.

 Speaker 1 – 49:47

Well, and I’m going for, yeah, I’m all for, I mean, again, you and I both resonate. I believe that positive change agents illuminate the way forward. I’m also a real politic person, so I would say. So I won’t, in the mid-2010s, I won’t give the exact date range. I visited a generally friendly nation to the United States and I got to meet with the Deputy Minister of National Defense. We had a great conversation at the end. We gave gifts and you’re not supposed to look at them until you Leave. At the end, I looked at what he gave me and he gave me a battery powered USB device and I’m like, really, really? This is 2015. So I waited till I got back home.

 Speaker 1 – 50:22

Send it off to some people that came back about two weeks later and they said it has some of the worst malware on the face of the planet. It’s trying to call home. And I’m like, yeah, love you too. But I raised that because, you know, even as you work to try and align mutual interests, I think there’s a phrase that says there are no, you know, there are friendly nations, there are intelligence agencies are friendly nations, foreign nations, but there are no friendly foreign intelligence agencies. And I raise that because we would be remiss if we didn’t see the value of both. Yes, we should collaborate and work together where we can and align interests, but also fences make for good neighbors, they say.

And so that leads to the. You know how you talk about the geospatial protocol.

 Speaker 1 – 51:03

We may be in a world which we want to have certain things happen only when you’re in US Periphery because there are certain laws that protect people on US soil that are not present overseas. But similarly, if I. Fortunately, I knew enough not to plug in that battery operated device because I suspected that it was probably not a good thing. But you could also say in the future, if you have this device from overseas and all of a sudden it starts doing an exfil to a foreign IP address that does not appear to be in the geospatial boundaries in the United States, maybe you should stop that.

 Speaker 1 – 51:33

And so I actually think in the conversations about active inference and how it can help and the geospatial protocol, I think there’s actually a huge cybersecurity dimension both for software and hardware, where that can help too, which is just pragmatic that we will find ways to collaborate, but every nation’s also going to try and eke out an advantage too.

 Speaker 2 – 51:54

Yeah, yeah, that is so true. Oh, well, it’s another day of the adventure.

 Speaker 1 – 51:59

Exactly. And that’s all you have. You have to say, welcome. Let’s see what we can do. And what is it? It’s the equivalent of the serenity prayer, which is, you know, give me the courage to change the things I can. The certainly accept the things I can’t and the wisdom of the difference. I hope a lot more people lean into the moment just like you. And I want to thank you again, Denise, for everything you’re doing, because this is how we build the next. So thank you.

 Speaker 2 – 52:22

Thank you, David, and thank you for, you know, you’ve shown me an incredible amount of support and you know that means everything to me and I really appreciate it. And yeah, it’s together, it’s mutual.

 Speaker 1 – 52:36

So onwards and upwards together there.

 Speaker 2 – 52:37

So, yes, do you have any final thoughts you want to leave with everybody with or how can people get in touch with you?

Speaker 1 – 52:45

So the best way is LinkedIn, but please say you’re not a bot because I get way too many bots for some mysterious reason on LinkedIn. But I would also say, you know, let’s give a shout out to our mutual friend, Ray Wong. Yes, Be a Constellation connected enterprise this year. It’s always great to meet in person. So a shout out to Ray and again, please be bold, please be brave, and most importantly, please benevolent in building the future ahead. Thank you.

 Speaker 2 – 53:09

Love it. Thank you so much, David. Thank you for being here and thank you everyone for tuning in. We’ll see you next time.