Well, give me, give me the… I mean, we talked a little bit about it last time, um, but I, I wanna kinda understand more about Runcible and, and the applicability of the tech. Maybe, I, I understand it in law, but I’m, I’m kinda wondering. I, I got to thinking it, you know, there’s gotta be applicability in other areas and disciplines.
Oh, it, it doesn’t matter. Think of it as things that can be reduced to cardinal measures, numbers, statistics is pretty good. But in things that have to be reduced to, to semantic measures or natural measures, AI, LLMs are not that good.
And so what we do is we create, uh, what is, is very complicated, not complicated. It requires a great deal of knowledge to understand what we’ve done. Let’s put, that’s put that way, right? So, um, uh, what it does is it, it looks, it treats, um, semantic systems, in other words verbal shit, um, as natural measurements instead of cardinal measurements. So just like you would say there’s before, there’s now before, there’s now just a bit ago, before, a, a while before, and a long time ago, those are, those are not cardinal measures. Those are semantic measures, right? They’re, they’re measures of, it’s a measurement scheme. The one we use most common to illustrate it, because it’s the one everybody understands, is, uh, evil, immoral, unethical, amoral, ethical, moral, virtuous, right? Right?
Now, the, the thing is, can you actually measure adherence to those terms? Turns out you can, and one of the great things about LLMs is the large ones, large parameter ones, have enough information to actually do that if you give them the criteria for those measures.
[Speaker 2 – 11:41]
So essentially, um, in, in s- in the dumbest possible terms, um, it’s we’ve turned, we’ve taken computing from its, uh, very, well, we would think of it as simple terms, which would be, uh, mathematics and programming, into linguistics, right? So, and we do this by tur- turn, converting, uh, ordinary language into operational prose, which is, um, just means a sequence of demonstrable actions that are, uh, equally testable. And so this machine does is it turns, takes your query, breaks it into that term, those terms, uh, and then reduces it into a series of statements, essentially a program. That’s what it does. It turns it into a program. Turns ordinary language essentially into what would, would… the equivalent of a program using semantic measures, and then it runs a series of tests. Now, uh, if you… I can actually do this if you can follow it. You have to tell me if you’re not following, okay? Um, [clears throat] in the simple terms of all spe- everything we say, if we’re not s- we’re saying it to other people. Why are you laughing, Luke? Um, wh- why have, w- everything we say falls under ethics, right? Now, if we say it to other people, it falls under ethics. But if it, we’re saying it to ourselves, maybe it doesn’t. It still falls under ethics. You’re still making an agreement with yourself. So once we get into ethics, ethics consists of a bunch of properties. I won’t go into them right now, but they’re basically non-imposition of costs on the interest of others. If we have– We get into testifiability, uh, the possibility. Excuse me. Possibility is we say, can this actually be bought, brought into being? In other words, if we reduce these claims, can these claims be brought into being? In other words, is it possible to do that?
[Speaker 1 – 13:44]
O- other actions you can take that actually bring this about in the real world.
[Speaker 2 – 13:47]
That’s correct.
[Speaker 1 – 13:48]
Yeah.
[Speaker 2 – 13:48]
Correct. Um, the next one is, can you testify about it? Which is, uh, i- and we, we… For each of these things, we have what are called gates. So you would say that they’re tests pro- possibly, you call them tests. And so for each of these statements, uh, reducibility, uh, possibility, uh, ethics, and testifiability, for each of those sets, there are a set of, uh, 10 to 15 tests that we sub-test that we run. Some of them are very complicated, like the long chain of externalities that occur from your actions, which tests the ethics, and some of them are very simple. Does this, is satisfy the, the necessity for naturalism, realism and naturalism, which are sound philosophical, but they’re, they’re not. Um, and, uh, so what we’d run is that suite of tests, plus what are we now liable for if under these conditions, and what are we, um-
[Speaker 1 – 14:56]
What can we warrant?
[Speaker 2 – 14:57]
What can we warrant within that liability? Then that’s the, that’s the decomposition and testing side. Then we do the explanation side, which is what ignorance, error, bias, and deceit are exposed in these questions. And so we actually explain things like this is a female position, this is a, this is a male position, this is a, this is a, uh, a Sinic civilizational position, this is a European position, et cetera. Um, and then we, we go into the… if a lie is being used or we can find ins- we simply find incentive. If we can find an incentive to lie, and then we go and explain those incentives and that force. It’s in… If I look– So the do- that’s just the, the basic statement about what’s been stated, right? That is if you see the length of output that we produce, it’s overwhelming. Humans can’t actually understand it. I had this json output from runcible. I gave it to Claude last night, and it made a couple of errors, and it’s then it finally understood. It was interesting to see it read our stuff.
That’s just level one. That just says, is this, is this truthful, ethical, possible, and warrantable? The next level is each domain, we’ll call it an industry, has a series of protocols or tests that they run. In other words, it’s a grammar or frame of reference. We do the exact same thing with a set of protocols for that industry. The next thing we do is we have clients, customers, organizations who have a series of protocols, rules, tests, whatever you want to call, processes, right? Then we run through those. Then the last thing is what is usually called safety. Because we do ethics right up front, safety isn’t a matter of ethics we don’t need any judgment on ethics. It already knows the answer, and it’s told you even why there’s a difference in, in ethical interpretation here because of sex, class, civilizational differences. Um, then but the next piece is, um, the explanation of, uh, of, uh, what we’ve just all said in terms that are eth- that are ethical, moral, and biased to you. In other words, t-t-to you. We can-
[Speaker 1 – 17:29]
To, to bridge back into your frame of reference and into the way that you-
[Speaker 2 – 17:32]
We have to bridge back
[Speaker 1 – 17:33]
… came to the-
[Speaker 2 – 17:34]
Right
[Speaker 1 – 17:34]
… system.
[Speaker 2 – 17:35]
So we’re not gonna tell you that, uh, you know, if you’re a flaming feminist liberal that you’re, that you’re wrong or something like that. We’re gonna say, “There’s a… Your position is this. The opposing position is this. The overlap is this, and the con- ongoing conflict is this.” And, uh, you, the only– And then we’ll resuggest, uh, means by which you can come reconcile those differences. If you follow that, that’s a really long chain of work, computational work.
[Speaker 1 – 18:07]
And that, and that sounded kind of in the political or ethical or kind of philosophical arena, but that’s just the example case we’re using here. This applies to-
[Speaker 2 – 18:15]
Right
[Speaker 1 – 18:15]
… scientific statements or any number of-
[Speaker 2 – 18:18]
Mm-hmm
[Speaker 1 – 18:18]
… concerns you might have.
[Speaker 3 – 18:19]
Let, let me ask, let me ask, can I ask a question or are we, can we-
[Speaker 2 – 18:23]
Yeah, no. As long as you get that’s what it does.
[Speaker 3 – 18:26]
Yeah.
[Speaker 2 – 18:27]
So, so in simple terms, LLMs hypothesize and we adjudicate. So Runcible is a wrapper around LLMs. It works with the LLM directly. It’s a wrapper around LLM that uses the fact that they are good at hypothesizing that we then falsify, right? So it’s the scientific method in computational form using natural measures.
….
[Speaker 3 – 21:20]
So we have three companies, right?
We have NLI, which is a volunteer organization, right? We don’t get paid. I think we’re all a volunteer organization, and we’ve put lots of hours in this organization.
In, Oversing, I’ve put millions into that organization, right? To build that, that type of a platform. It’s a million lines of code. It’s very complicated. It’s, broad functionality – as big as any of the other ERPs that are out there in the mainstream. However, it’s in beta. It’s in beta condition because we actually need an AI to complete it.
So Runcible is the third company. It’s the AI pr- company that we’re building to complete it. One of the reasons we’re building Runcible as we have is that we don’t have to own hardware. Because my view is that the hardware is a sunk capital cost that we’re never gonna, that these companies are never gonna recover.
So what we tried to do instead is come up with something that we could license or sell or whatever to some number of foundation model companies so that we can stay out of the hardware business and stay into the, uh, actual, um, user-facing business.
Um, so we have these three businesses. Uh, of these three, of course, uh, oh, uh, L- NLI is active and very active. I have shut down, the Oversing/Reality By Chanting was an, a UK company with development offices in Ukraine. I shut down the Ukraine and the UK businesses and transferred the assets back to me and then transferred those assets to Runcible.
So Runcible now contains both asset sets. The intellectual property Licensed to it from NLI, um, but assigned- There’s the NLI stuff that’s licensed to it, and then there’s the Oversing/Reality By Chanting that’s essentially granted to it. That has been, uh, s- that has been contributed to. It’s essentially my, uh, it’s treated as ownership, uh, as my, my ownership cont- contribution capital, right? So, um, so we actually operate right now as a volunteer organization running these things. So the burn we have is thousands, and that’s financed by Brad and I, right?
[Speaker 4 – 20:50]
Yes. No, for sure. Um, the… So I mean, I think maybe you guys could talk about the current state of your business operation, ’cause we can talk about the tech, um, a lot, and there’s definitely a lot to say. But I think that understanding where you, where you are with the business, your current burn capital needs, et cetera, and plans, that w- I think be helpful to contextualize.
[Speaker 3 – 21:20]
Yeah. Sure. So, so, so let me just go. So we s- we have three companies, right? Um, uh, we have NLI, which is w- a volunteer organization, right? We don’t get paid. Well, one of us gets paid. Uh, I don’t know if he gets paid anym- I don’t think Rod, he gets paid anymore, if I remember correctly. Um, but I don’t think so. I don’t think, I think we’re all a volunteer organization, and we’ve put lots of hours in this organization. In, uh, Run- in, uh, Overseing, uh, I’ve put millions into that organization, right? To build that, uh, s- that, that type of a platform. It’s a million lines of code. It’s very complicated. It’s, uh, viewed broad, broad, uh, broad functionality as big as any of the other ERPs that are out there in the mainstream. However, it’s, uh, it’s in beta. It’s in beta condition because we actually need an AI to complete it. So Runcible is the third company. It’s the AI pr- company that we’re building to complete it. One of the reasons we’re building Runcible as we have is that we don’t have to own hardware. Because my view is that the hardware s- is a sunk capital cost that we’re never gonna, that these companies are never gonna recover. So what we tried to do instead is come up with something that we could license or sell or whatever to some number of foundation model companies so that we can stay out of the hardware business and stay into the, uh, actual, um, user-facing business. Um, so we have these three businesses. Uh, of these three, of course, uh, oh, uh, L- NLI is active and very active. Um, I have shut down, uh, R- Runcib- uh, excuse me. O- o- the Overseing/Reality By Chanting was an, a UK company with development offices in Ukraine. I shut down the Ukraine and the, uh, UK op- uh, opera- b- uh, uh, businesses and transferred the assets back to me and then transferred those assets to Runcible. So Runcible now contains the, uh, both sets of asset sets. The intellectual property trans- uh, assigned it from, uh, from, uh, Re- NLI. Licensed to it from NLI, um, but assigned- There’s the NLI stuff that’s licensed to it, and then there’s the Overseing/Reality By Chanting that’s essentially granted to it. That has been, uh, s- that has been contributed to. It’s essentially my, uh, it’s treated as ownership, uh, as my, my ownership cont- contribution capital, right? So, um, so we actually operate right now as a volunteer organization running these things. Um, so we have no re- I mean, the burn we have is thousands, and that’s financed by Brad and I, right? Uh, so that doesn’t matter. Um, however, I have-
[Speaker 2 – 24:11]
You– This is my job is to build companies, right? So this is my 14th or 15th company, depending on how you look at it. Um, so we have Eric Adams on who, who’s a very seasoned COO, who’s done, you know, worked with me for 10 years and, uh, we’ve known each other for over 20, well, it’ll be 30 now. Um, uh, he has taken, done multiple, um… He understands our business. He has, uh, taken multi-co- multiple companies through sale. Um, he is the guy I use to insulate me from the, uh, daily hubbub of o-operations. Uh, I have Luke to manage the technology staff. I have Moritz to handle outreach. Um, and I have a bunch of other people in line who actually know how to run a busine- run these businesses. Starting up this business is no different from me starting another consulting company that is, uh, that is basically doing system integration work on n dif- n different companies’ platforms. Same fucking problem. The only difference is I own the goddamn platform this time, right? So it’s essentially a business I understand. Uh, just as, uh, you would tend to form a consulting company and try to get into as many markets as you have, we’ve laid out about 30 different markets we can penetrate, um, with Runcible. Um, and, uh, this basically requires we build cells. So you might, in a services business, you said, “I wanna build an offering.” An offering would be led by some person. They would, uh, be in charge of sales. You would put a s- a s- uh, sale support staff and some salespeople under it, and that would take some real resources ’cause it takes a, it takes money to employ all those people. In our case, it’s really a four to five-person team per vertical, and there’s only so much work to be done in each of those 30 verticals. So realistically, our job is to take our existing staff, form a number of cells, pursue the initial verticals, and go explo- exploit them. Setting up and doing this is sort of like doing the dishes for me and for Eric and for Ariella and for Luke, right? This is, it’s– There’s nothing strange here. This is our job is to go to Fortune 400 companies and help them do it. I mean, obviously, um, you don’t know the story of a- all the stuff we’ve done with Microsoft strategically to help save some of their products, um, and make them successful in the market. Um, so, uh, that’s what we actually do. That’s our current state. What we’re looking for is to go out for financing. The question about the financing has to do with, uh, our unders- our concern that what we are selling might be difficult to understand how revolutionary it is. Why are Mor- Moritz, why are you laughing?
[Speaker 3 – 27:12]
The inclusion of the word might.
[Speaker 2 – 27:15]
Yeah. Okay. Um, so that’s our basic concern. It isn’t concern for our history. We have a r- I certainly have a record of building multiple companies and growing them and selling them, right? I mean, the difference with this company is that it’s, uh, y- I usually sell a company at $100 million because in the services business, that’s about the level you can get to without going public because you actually can’t float the money. The, you can’t float the salaries, uh, without having access to public markets. That’s why consulting companies usually don’t get that big, is that they have to be able to go public. I don’t wanna be in a, own a public company. I’d rather sell it to a public company, so that’s what I do. All right? So this, the difference is this is a billion-dollar company easily, right? It’s probably a, it’s probably a lot bigger than that. And so, um, uh, so in that sense it’s a much, it’s more work the, it’s more work, uh, but with more capital. In other words, with more value appreciation than the companies I usually do. In other words, a services company, you only get so many multiples on the revenue ’cause it’s only so dependable. With a product company like this, I mean, the revenue up potential out of this freaking thing, starting with licensing but going through to the o- overseeing implementation, I mean, th- th- this is how big companies are built. So, uh, I see this as, um, uh, we, we look at it as six different revenue channels, um, able to accumulate into a god awful lot of money. We just don’t talk about it because it’s so fucking unbelievable. We s- I feel like an idiot when I even bring up the numbers. So, uh-
[Speaker 1 – 29:00]
Yeah, but that, that, that, that’s, that’s probably becoming less of a problem, Curt, because the numbers have just gotten stupid all over the place.
[Speaker 2 – 29:13]
So, so the where we are is we are, I have done this repeatedly in my career. I go build, I go set up a team. I set up a, I find an exploit in the market, I set up a team. I go and, uh, and, uh, raise some money and I go do it. In the past, I’ve been rich enough to do it myself or with a couple of friends. This is beyond my ability to finance. I was able to finance everything else, but I can’t finance this scale, this kind of thing myself.
[Speaker 3 – 30:12]
That’s what Curt used to call his Twitter, it’s a safe space for smart people
[Speaker 1 – 33:37]
There’s only two problems in ra- in raising, or two challenges in raising capital. One is, you know, does your… You know, can you talk the talk and explain what you’re, what you’re doing to a layperson or to an investor, right? Which you guys clearly can, I think. Um, and two, um, is the structure of the offering palatable? Like, the, you know, I often, there’s often a disconnect, right? So, like, you’ve got a great idea, but the terms are just not, are not right, and they’re not workable for an investor. You know what I mean? Like, for example-
Leave a Reply