By John M. Switlik, January 2026; Context: Taming GenAI/LLM, 7th — KBE Insights, …
The last post here was in the middle part of 2024. In the past 18 months, I have continued to write, mainly on Linkedin. Below are links to a few articles and posts that depict the themes. These were done in the Ln environment for several reasons, but the intent all along was to have a two-pronged representation what also included the different blogs.
So, also on the lists are links to the main blog (Truth Engineering) plus various posts on Quora.
Truth engineering blog — Restart, ICAD focus — The gist of the matter is that truth engineering came of my work in the domain of KBE (Knowledge Based Engineering) which was a successful merger of facilities from KBS, mathematical software for engineering, realworld programs requiring support, and the goal of developing what is now known as a “digital twin”. This was in the latter part of the 1980s and carried on for over a decade. Lessons from that work did not get notice. Yet, there was a maturity attained that ought to have involved with the recent machine learning efforts. That will be discussed.
Linkedin — AI and Truth — I summarized some details about KBE. Given from a general point of view, the examples of data analysis come from the environment of a university lab doing experiments concerning tradeoffs with regard to material. The following image is meant to give one view of the central problem with computing: it deals with abstraction, however that approach deals with context which then implies nestled states. Structure that might be native to situations gets lost; one argument for trashing intuition was that “truth” was many times non-intuitive. Oh, let me disagree and demonstrate otherwise.
Linkedin — KBE – Importance of knowlege — After a brief description, there is a look at “futures and theory” which pertains to the present situations as it did back in the day.
Linkedin – KBE, continued — Given that my contribution to KBE was from the role of engineering support (advanced computational systems), my view of KBE runs along a broader framework which ought to be familiar in this day of “scale” and the modes of GenAI/LLM (or briefly, AI/ML). Seemingly remarkable results have been witnessed many times sufficiently to elicit concerns about a “critter” in the box. But no, KBE would emphasize the mathematics which will be a continuing topic in future discussion. In this article, I pushed a summary of my writings on topics associated with KBE and AI/ML to one of the known packages. A response is provided based upon two versions the summary file.
Linkedin — Differences of opinion — From the engineering support role, I could see the disparities between the views of approaches that could become evident in data and interpretation of such. Some of this deals with representational concerns. A huge factor is that all numeric systems are hugely dependent upon factors that are not easily understood. An example would be boundary concerns. People talk these issues and come to agreement as if there has been a closure to the problem. The computer though? Thinks otherwise as we will be able to understand through AI/ML and other hugely complicated systems of the future.
To end this, let’s consider Gibbs and Maxwell. The former? U.S. researcher. The latter? The well known Scotsman. They collaborated in ways that are of interest beyond what has been discussed so far. So, we will venture there.
Linkedin — Maxwell and Gibbs — As we go further in discussion of Truth Engineering, we will see two recurring themes. One of these is the thermodynamics can be used as the focus. As such, it’ll represent several things to discuss. Computing? Well, Shannon might come to mind. Let me add another which deals with mathematical support. In particular, we will start with two publications that were translated into English in 1926 which indicates that they were written earlier. These are Levi-Civita’s work on Absolute Differential Calculas in which we will find Ricci’s and others’ contributions to Einstein’s theory. This is an early look as we can discuss. For that matter, a 1976 book on differential forms and variation principles will be a nice leap to the modern view (recognizing, of course, that we’re 50 years from that point). This book is cosmologically oriented. The other book? Planck himself took on the task of describing the fundamentals of thermodynamics from a chemical viewpoint. Same mathematics. Different context. Outward, for the former. Inward, for the latter.
Where sits intelligence as we know it? Inward, for the most part. Yet, people bound between these two views of great separation. Wait. If we were to put our finger on a necessary frame of reference (rightly used), would theoretical chemistry be were one would start to try to get a handle on the specifics of intelligence? That is, physics is wanting, in this sense. Except, that all of these sciences are wed by mathematics which then shows that the true focus will be both mathematics and chemistry (hence, the use of theoreticxal as we have seen pop out over the past few decades – augmented, naturally, by computing).
Context
Boeing Commercial Aircraft