Will ISRM supercharge AI? | John Paul Schell | #190
Will ISRM supercharge AI? John Paul “JP” Schell is a former biotech scientist, systems theorist, AI entrepreneur and author. We talk about the nature of consciousness, ISRM Labs, the efficiency of systems, his Interactionist Self-Regulation Model, the complexity and danger of our
Publicerat 10 sep. 2026. Fler avsnitt finns på poddsidan.
10 sep. 2026
Will ISRM supercharge AI? John Paul “JP” Schell is a former biotech scientist, systems theorist, AI entrepreneur and author. We talk about the nature of consciousness, ISRM Labs, the efficiency of systems, his Interactionist Self-Regulation Model, the complexity and danger of our
Transkript
- SPEAKER_00 Gotta make sure everything is recording. Don't want to mess that up. Yeah, yeah, yeah. All right. It's like start acting awkward. JP Schell, welcome back. Cool, man. Thank you. Tha
- SPEAKER_01 So it was like a lot of learning and a lot of stuff that I'm just like, I became very kind of critical of like, what is true? Like, why should I need to be learning this in order f
- SPEAKER_01 It's just that we made all these constants and we made all these mathematical observations that we're sticking to, but that was during a time when we didn't have these kind of mill
- SPEAKER_01 where your your neurons are actually developing way. You have like way more like if there was a baby in the room, like there would probably be more neurons in that baby's head than
- SPEAKER_01 The connection between your eye and your occipital lobe on the opposite area. When you have injuries of traumatic brain injury, like we've seen with football players and whatnot, a
- SPEAKER_01 So if you do have this like traumatic injury that then takes out this huge portion, like naturally, like the system is just trying to integrate what should have been here. Now you
- SPEAKER_01 this stems from just like I don't know, just kind of philosophical curiosity of looking at us compared to LLMs and agentic systems. And it was surprising to me, but there's a lot o
- SPEAKER_01 But you're broadening, I think, the level of models that you're having access to. Serotonin, when it is a selective pressure, you essentially refine your models down to what is abs
- SPEAKER_01 And so like I mean, and them being, like, if it has a face, it's too closely related for me to think that, like, that, like, the proteins that are in that system are not going to d
- SPEAKER_01 When you consume something especially as closely related as a cow or a pig, and we can go to the extreme. Why do you not eat humans? It's like we know why. It's because you get kur
- SPEAKER_01 But this is coming from an idea Yeah, we needed meat. We absolutely needed. There was bottlenecks in our survival as human beings. Think about the Mongolians coming across the Poli
- SPEAKER_01 But here's the reason why I would say the dangers aspect is pretty realistic and pretty well known now of why are they recommending people not to eat red meat is because there has
- SPEAKER_01 oh fuck, like this is just, this is not supposed to, like it's Legos that fit, but those Legos like melt when they get in there. I don't know, that's a bad example, but it's just,
- SPEAKER_00 I think so, in some degree. In some degrees, you have these self-replicate, I think that there's stability there. There is optimization of constraining energy. Like that seems to b
- SPEAKER_01 a topography where you have kind of basins of stability. And like this is where we're talking about like this idea is not new. It's coming from biology in the sense of like cell st
- SPEAKER_01 then they go together and then they just complexify into boom. And it's like, no, he's kind of showed that you have certain kind of behaviors of individuals and of the system's app
- SPEAKER_01 That was the trade-off that we made. We go like, oh, we're having complexity. But in that complexity, we're losing some of our regenerative capacity. And despite there being good e
- SPEAKER_01 And of course, there was an MIT study that said doing DTI, the thing I did, that white brain matter is decreasing in people who are using GPT regularly. But I would argue that if y
- SPEAKER_01 It's just now it has a virtual... You're saying the LLMs are a snapshot of reality. I think in some way- They provide you a snapshot of reality. No, I mean, I'll tell you what they
- SPEAKER_01 And that's the problem is we don't know the original weights of what the stuff that was trained into these models really is. I mean, that's the thing is how many hours of Katy Perr
- SPEAKER_01 to how you should live your life in a better way for you. It's how to get... It's clickbait. It's how to get more attention and how to get more users, how to get more funding. So i
- SPEAKER_01 Sadly, that's being done on a U.S. first or this nation first type of mentality. And that's not going to fly in the future because it's all connected. And all the people, whoever y
- SPEAKER_01 I'm sorry. I know that this is a big part, but we can't use our resources for making a diversity of cars. Maybe things will be plentiful enough. Maybe we start mining an asteroid a
- SPEAKER_00 And now we're like, all right, well, let's not... And now AI is the most expensive thing. It's like the most worth, which is insane. I mean, Kimmy 3 just came out. It's going to pu
- SPEAKER_00 With this advanced AI, you can literally make digital clones of people, perfect voice, perfect video, everything, and in real time respond to or talk to other people. Yeah, I tell
- SPEAKER_01 the attacks, the cybersecurity attacks, I mean, not to mention some of this crypto stuff. Can you imagine rug pulls that are just being done by completely unethical agents that are
- SPEAKER_01 but I could totally crash the car into this construction site and I would probably not be considered at fault. It went through the mind, or at least through the thinking, dot, dot,
- SPEAKER_01 And who knows, you know, I think that like when you, like, I think when you, when you essentially type in, you know, whatever your prompt is, you know, to this giant LLM, I think y
- SPEAKER_01 to Have the least use the least amount of energy in order to achieve the most in whatever dynamic way you're trying to do it. We do it with meat. Now we're doing it with silicon. T
- SPEAKER_01 No, your brain is running off of around 20 to 40 watts. And it's like, what is it? It's like petawatts now that are going into answering one question that gets the acronym wrong. A
- SPEAKER_01 this is how many tokens I spent. Did I... Is my predictive error decreasing? Did I align more to the model or am I more misaligned to the model? And that's where you can then make
- SPEAKER_01 The way that like modern engineering systems already behave in the sense of let's try to minimize uncertainty to an extreme. Like the idea is always like let's get chaos out of the
- SPEAKER_01 Things are kind of extrapolated from nothing. This is having too many tokens for a problem that probably could have been solved with a lot less. You shouldn't be using the same mod
- SPEAKER_01 a ledger is transparency. Using that transparency is what makes it like optimizable. Having capacity to self-observe how much you drifted in that particular call, probability call,
- SPEAKER_00 at these factors like uncertainty and coherence and energy. And energy. Yeah, yeah. So two of the free energy theory are in it. It's not exotic math or anything, but it's basically
- SPEAKER_01 She's done some really awesome stuff that I gotta get back to her about. Different domains where you have adaptive change. That adaptive change can be this recursive thing. Now we'
- SPEAKER_01 because it could be less symbols that pack in more information per subunit. So apparently, yeah, it's just language is the thing. And that's the thing is like, we're getting away f
- SPEAKER_01 So, oh, you use this model rather than you use these models for like, and this can be translated into like different Industries as well, cryptocurrency. You have really good models
- SPEAKER_01 So it's context dependent. You know, I will refrain. You know, not waste the energy. So this transparency that ISRM offers, would that Also be a safeguard against like an iRobot or
- SPEAKER_01 And now you're going into this kind of meristem state. It turns out that all this very early stages of agriculture, like plant botany, also parallel what pluripotency and the whole
- SPEAKER_01 If I hit a rock, now all of a sudden there is a high predictive error because I'm putting that energy there and now there's resistance. So your coherence to the model of being able
- SPEAKER_01 a young quantum physicist who's in Savant territory for sure. And a couple just other like quantum physicists who have actually been really intrigued by what ISRM could bring to th
- SPEAKER_01 Correlation. And now we have to be real scientists about it because AI can confabulate a lot and it will assume a lot of non-reality now. And that's the big dangerous part of how w
- SPEAKER_01 There's a lot of awesome positivity in the world. And overwhelmingly, it seems like the world is going in a positive direction. I'm a big Hans Rosling fan, so I want to believe it.
- SPEAKER_01 of these phase changes right now. And yeah, the singularity probably happening. That's cool. But also like, you know, it's just, I've been, probably actually the last time we spoke
- SPEAKER_01 Not that it has any directive or cares about us or anything like that. But when we become a parent, would we become a parent to it? And do we do enough that it would actually be so
- SPEAKER_01 And I don't know, I mean, like, I don't know. I'm kind of against that just because I feel like I'm not... It does dig into my... We're losing a lot by giving up so much to the age
- SPEAKER_00 Yeah. Yeah. Yeah. Yeah. We're just a creature like any other creature, I believe. We're not special in that sense, but we are very, very special in the sense that we are trying to
- SPEAKER_00 We've actually sent, what is it, Voyager 2? It has left the solar system. It is insane. We have landed stuff on an asteroid hundreds of thousands of miles away. We do such cool shi
- SPEAKER_01 like any physical reality and are actually only existing in a perpetuated virtual realm of their own. And so it's like, And then I look around and I see people walking everywhere.