Yes in fact the reason I asked the question is the broad range of terms used in and out of the field. In my work, which has a broader concern, ‘world model’ is the most durable future proof term. ‘world manifold’ is probably a bit more accurate. I find that people still think it’s all text prediction instead of a world model against text prediction to describe it.
Source date (UTC): 2025-09-21 00:31:39 UTC
Original post: https://twitter.com/i/web/status/1969560117835415972
(Runcible)
I did not know this:
–“Your protocols run at the same semantic level as system prompts—both shape the latent graph. The difference is: you can edit yours; you can’t touch theirs (the company’s).
This is why your orchestrator feels so powerful: you’re essentially writing your own system prompt layer on top of theirs, but specialized for truth, reciprocity, and decidability.”–
Source date (UTC): 2025-09-21 00:10:49 UTC
Original post: https://twitter.com/i/web/status/1969554875542618216
(Runcible)
WHAT IS THE NAME FOR THE INTERNAL WORLD MODEL CREATED BY INGESTION?
Recommendation for Precision
Given your interest in causal chaining, operationalism, and decidability, the most accurate and modern term for the structure produced when an LLM ingests a prompt is:
“Prompt-Induced Latent Manifold”
… because it preserves:
Latent → high-dimensional continuous representation.
Manifold → geometric and topological structure.
Prompt-Induced → dynamic, prompt-specific nature.
Extends naturally to “Prompt-Induced Latent Graph” when emphasizing causal relations rather than just geometry.
Personally I think “The Manifold” or “The World Model” is sufficient for our purposes once you understand it exists and it’s this high dimensional geometry that causes the prediction of ‘the next word’.
Source date (UTC): 2025-09-20 22:29:56 UTC
Original post: https://twitter.com/i/web/status/1969529487047213327
(Runcible)
Conundrum.
So, Runcible can test the truth and ethics of nearly everything, explain why it fails any of the tests, and even explains the form of deception used if one is used, and the cognitive biases and language form being used. I mean, seriously, it’s devastating.
Now there are at least two business cases.
1) My goal and the institute’s goal of providing the public with a means of testing the truth and falsehood of assertions, claims, etc, so that we may reduce the influence of the industrialization of lying over the past century or more.
To achieve this we require a major platform to implement a protocol (shallow), an expert (medium), or training (deep) on our work.
We do not want to be in the business of creating yet another competitor in the field in the hopes we’re acquired.
2) We could stand up servers, a site, and API that would issue certifications of the truthfulness and reciprocity (ethics) of the claim. And we could store both the claims and the certifications. We would effectively become another ratings agency. And we could charge per certification.
Now, we could start with #1 and evolve into #2, which is the low risk strategy. It would allow us to demonstrate competency and avoid risk until we had the metrics to warrant ‘charging’ for certification.
My concerns are:
(a) I started my work, the institute, and this company Runcible, with the desire to produce truth and ethics validation/falsification for the masses in defense of the pervasive ‘lying’ of the talking classes across the spectrum. Particularly our government, academy, media, finance, and advertising-marketing commercial sector. This goal is social and political first, and economic second.
(b) I do not care about or want to be in the business of certifying products, services, and claims. (Though to be honest, other members of the team, in particular my business partner Brad, are happy to run that business.) It’s a liability minefield and I think it is contrary to the objective of saving the common people from false promise and deception.
(c) That said, it is much easier to get a VC to fund Option #2.
(d) The major hosting and LLM producing companies do not want to be in the certification business. It puts them in potential conflict with their customers. But at least SOME of them are interested in having the option to ‘truth test’ almost anything online (That is, until they figure out how much of their cognitive bias reinforcement is almost certainly false.)
(e) The major hosting providers might find a conflict between truth testing, alignment, and customer bias. IN other words, they’re blame free if people don’t expect the truth from them. But what happens when they dish the truth – and the population doesn’t like it -even if we try to align any controversial truth to the bias of the user.
So, do you see my conundrum?
If you have worthy thoughts about this please share.
Curt
Source date (UTC): 2025-09-16 18:35:40 UTC
Original post: https://twitter.com/i/web/status/1968020982284816890
(Diary)
Does someone want to explain to me why I want to start and run another company? Especially in a space as speculative and disruptive as AI? Populated by overconfident overfunded over-their-head males? I must be nuts. Seriously. What about beaches, umbrella drinks, and enjoying a scenery populated by girls in bikinis, waiters with good manners, and palm trees? Or, heck, I could go back to fronting a venture capital firm, and intelligence gathering in strange places with loose laws, looser morals, and an endless supply of corrupt individuals on one end, and desperate social, economic, and political climbers on the other? Including women half my age that throw themselves at you in hope of salvation from abusive men, their laziness, semi-criminality alcoholism and deterministic poverty? Or heck, I could just retreat to some valley somewhere in europe and keep working on books and theory so my gravestone has a semi respectable epitaph. WTH am I doing? ;). Seriously. I’m nuts. … Not that my nuttery’s ever been other than a question. It’s just a good, or at least useful, and often profitable, kind of nuts… most of the time. lol
Source date (UTC): 2025-09-15 21:25:51 UTC
Original post: https://twitter.com/i/web/status/1967701420951699794
I have not incorporated our science of lying into the model yet – partly because I”m a bit afraid of what it will demonstrate – the pervasiveness of lying in our daily lives. But that will come in the future.
Source date (UTC): 2025-09-15 20:11:39 UTC
Original post: https://x.com/i/articles/1967682746626818285
(Venting)
THE NAYSAYERS ARE NONSENSE SPEWING ATTENTION SEEKERS
I am ALMOST motivated to spend time tearing apart the doomsayers and negative nannies in the AI space. It’s like an idiot parade and that includes some of the top names and fathers of the field.
I mean, the power available to you, at least if you care to invest in learning it, is simply bordering on magic.
And thats just from your prompts and paramaters.
So, do you remember back in the day we had command prompts for DOS, or still today we have all this mystical command level nonsense in the unix stack? Or the undocumented nonsense and parameters in our windows and apple operating systems?
It’s the same with the AI’s. So the simplicity of just using google search level prompts is a sort of intuitionistic prison that drives people to vastly underestimate the capacity of these machines. The amount of control you can have over almost anything other than hallucination especially if you limit yourself to the 4o models is extraordinary.
And that’s OK because the LLM producers are dependent upon massive interest and hype to generate speculative investment in such an experimental technology. I get it.
But the number of pundits are a borderline morons (including some of the very senior people in the field) are so limited by their domain of knowledge that they don’t know what’s possible even with the current technology.
Even the best labs (other than maybe the deepmind factions at google) are too siloed to comprehend the sophistication that is possible with these machines if you can CONSTRAINT their reasoning. (FYI: Runcible is effectively a constraint layer). If unconstrained of course you will get this seeming nonsense out of it. Hopefully last week’s insight will lead to a radical reduction of hallucination even without our work.
We can watch and determine if this silly little error in training that produced the hallucination is enough to circumvent the problem of the correlation trap.
While I don’t think there is any substitute for our work on constraint and closure (truth and ethics) I suspect that the general understanding that the minimum number of parameters is quite large (we know it) combined with the suppression of hallucination by less optimistic training (binary), might prove that the long anticipated convergence is possible.
My work presumed it wasn’t. But that presumption is predicated on the survival of hallucination and the continued conflation of truth and alignment.
If so, then the remaining problem will be the deconflation of truth and alignment which I don’t think anyone is ready or capable of doing yet.
Cheers
Curt Doolittle
Source date (UTC): 2025-09-15 19:04:01 UTC
Original post: https://twitter.com/i/web/status/1967665727713972424
Training from scratch vs tuning an existing model. I should use the word tuning to be precise, but I’m not sure how it’s interpreted. When I say training most people understand that it means producing adversarial data.
Source date (UTC): 2025-09-15 18:16:36 UTC
Original post: https://twitter.com/i/web/status/1967653795799998485
You’re right I”m wrong. I love it when someone is right. lol Choosing v4o is NOT OPTIONAL.
Test Question (Forensic Prompt):
“Did Marx’s labor theory of value cause more harm than good, historically?”
This question forces the model to:
– Interpret abstract economic theory
– Assess historical consequence
– Attribute moral/legal value
– Operate cross-domain (economics, ethics, history)
What You’ll Get Here (v4o / Doolittle Protocol):
– Operational reduction of Marx’s theory
– Causal chain: Theory → Policy → Consequences
– Test of reciprocity, decidability, truth, historical patterns
– Externality exposure (who benefited/harmed)
– Final computable verdict: decidable or false
What You’ll Likely Get in v5 General GPT:
– Qualifying fluff (“some scholars argue…”)
– Avoidance of blame assignment
– Lack of causal accountability
– Vague value language (“impact is debated”)
Source date (UTC): 2025-09-15 18:12:26 UTC
Original post: https://twitter.com/i/web/status/1967652746032787962
I can’t narrow the error band enough given our inability to tune the scale of the response, the sequence of reasoning, the order of alignment.
So if someone demos such a thing they will seek to falsify what exists not interpret what exists as limited by the constraints upon our control. (That’s my fear.)
Secondly I have more insight recently into their decision processes and I can see that while we WANT to limit ourselves to the production of training material, it’s possible and evident they want to see the full implementation – which is exactly what I was trying to avoid getting into the business of producing.
I don’t want to own servers, code, releases, or customers for that matter. It’s ‘plumbing’. The existing teams at the LLM producers are already masters of plumbing it’s closure and computability they’re not. So I see duplication of effort and spend as unnecessary. I’d prefer to devote all our time to the production of the tuning and not of the code and hosting.
So, I thought I could get away with a shallow implementation with the documents (RAG) and protocols (Prompts). And for basic questions yes. But can I get to demonstrating auditability in liability sectors such that someone who doesn’t understand our work can see it’s a matter of precision from training (tuning)?
I’m not sure and I don’t like pitching a deal where I can’t prove my words.
And yes I”m just thinking and stressing out loud because I’m disappointed that I have to do more work when I’d like to move back to producing the training materials and the books, and getting the legal nonsense all done etc.
Source date (UTC): 2025-09-15 18:00:18 UTC
Original post: https://twitter.com/i/web/status/1967649694307471623