Basic Large Language Model understanding
This post was added as an extra which is deemed quite essential especially in terms of learning how Large Language Models (LLM) operate. The recent video shown here is obviously AI made! The video is from a very new Youtube account that’s quite likely trying to get in on the AI craze. Barely any views have been accrued for it so far. However what I liked is its very simplicity in presenting the ways and means LLMs work and that is important for any who may well have little or no understanding of AI – especially the LLM format. it was deemed these form an additional post (so that’s four instead of three) before the big one which is a quite deep yet insightful look at the issue of LLMs and how these reflect the human nature of communication. In other words, the notion that the way humans speak and communicate is no more than how LLMs work – and this is essential because both do not need silicon or brains to operate. Meaning in terms of humans these ‘natural’ forms of LLMs operate basically as a system that is independent of human thought.
Another post in this series was a question of generative speech – which looked at the current situation surrounding LLMs and the belief these reflected an automated entity inside humans – a form of parasitic existence since it was operating largely without its human owners agreement or knowledge. In this second post reference was made to two videos which were those by Chase Hughes – Language Has a VERY DARK Secret (23rd November 2025) – basically a non academic overview of this thing that existed in humans, and also Curt Jaimungal – Language Without Meaning: How LLMs Exposed Our Biggest Illusion (11th June 2025) which involved an in-depth interview with Elan Barrenholtz – one of the first academics to study the issue of whether speech (and communication) was largely a system that operated independently of the brain.
The post to follow those two features an in depth analysis of one where both Elan Barrenholtz and William Hahn put forward their views why speech is a thing that apparently exists independently of its user. Its quite deep so it was decided to make this the fourth post and this new one the third.
Anyway the following video gives a complete layman’s overview of how LLMs achieve a human like communication system. It also advises of the best ways of using these LLM systems – plus it describes the pitfalls of believing AI too much and how LLMs can ‘hallucinate’. For example one might ask, ‘who won the Nobel Prize for inventing blue bananas?’ There evidently is no such thing, but because such a question has been asked, LLMs make up answers that appear as if they are the truth! The video no doubt is a sort of groundwork for those who have little idea of AI/LLMs – and that is because the fourth post needs a fairly good knowledge of LLMs to understand the latest theories of communication in humans, especially the idea of parasitic speech. The one small thing about this video is like so many hearing people, it replicates the word dumb. Simply none of these, AI or hearing, can understand in a million years the sheer negativity behind that. And yes, its also used against Blacks too! However ‘dumb’ as used by hearing people is no doubt a very large LLM token. Its almost invariably the one that’s going to be used in many cases as an insidious reference to being lacking in intellect. In a sense it can be seen how AI could push negative words in ways not seen before and make them look ultra-normal – as if no-one had ever objected at all to these.
Next post is a theory that language and speech is not one’s own (featuring the video The Terrifying Theory That Your Thoughts Were Never Your Own). This is a comprehensive and detailed look at the new academic arguments on how LLMs reveal the inner workings of human communication.
