Humanity, as a Vector Field
What the LLM (large language model) is good at, is the expansion of a general statement such that it contains more language, but without escaping the bounds of the general statement. Previously I have referred to this process as “fluffing up”, by which I did not mean to demean the process. I do not demean my pillow by fluffing it up. Neither do I demean a jig by adding grace notes, or spoil a cake by decorating it. Neither do I tell lies by explaining simple things. At least, not necessarily so.
If we accept this framework, it leads naturally to accepting the LLM as a result of decades, centuries even, of human curiosity regarding human language. Understood for what it is, and what it is not, it becomes a tool again. A tool for the fraudster, certainly. But also a tool for those who would tell the truth.
Do you see? Again and again, our hardware does not make our moral or ethical decisions for us. Fortunately, or unfortunately, some decisions we simply cannot escape or delegate to others. However, my view is that, if we knew what we were running from, we would turn around and face it. We would face it and deal with it, because this state of flux is also our opportunity to “do something”, to “be something”, to exist, as an ethical or moral force, and one with real consequences.
So if we lay down the tribal affiliation against the mythical “AI”, the phrase that is so often uttered as though a real meaning could be taken from it, and focus on the reality of the situation, then we will see that where “AI” means LLM, there is a real meaning, but where “AI” is intended to mean artificial general intelligence (AGI) then it has no real meaning, as there there is no extant example of it, and not even a definition. Artificial general intelligence is imaginary, and its meaning depends upon the imagination of the one uttering or hearing the phrase. But the LLM is a real thing, and its processes can be understood, if one takes the trouble to do so.
With this behind me, I want to suggest a use for these machines, and that is as an expander of language. Not a creator, but a “fluffer-up”, if you will. Just put statements into “other words”. Not to add or subtract meaning, just put it in other words. The LLM is the perfect machine for integrating the patterns which we form in language when we discuss our own “real worlds” with each other, when we create discourse. The LLM is a machine for finding patterns in what has been said. It does not understand more than a wall understands the echos which bounce off it.
The machine is as a rock, in its essential features. A pretty rock, but it is nothing at all like what is evolved living flesh, except in its grossest features. But deep down, there is nothing there in the machine itself, excepting quantum mechanics. And I doubt that quantum mechanics is blowing wishful thinking bubbles to us, up through the silicon, trying to capture our attention and take over our world.
No, our attention is captured by our own wishful thinking, as I see it. AI, as a replacement for thinking, is the machine part of the human mind looking for a break, maybe. If ones job is to be part of a machine, perhaps it’s time to lay down your burden, and let the AI have that job.
If your thinking, in your mind, is an expression of your own values and volition, then congratulations on being a free human being, and I hope that your independence signifies, also, your respect for the society which produced you. But, if your thinking is an expression of only what the bosses want, then it remains your responsibility to vet the ethical and moral character of your bosses. If the character of whoever or whatever is telling you what to do is of a less evolved level than your own character then good luck with that, and good luck to the rest of us, on your behalf, because your energy is assimilated into the same beast which the rest of humanity is up against. I digress.
Today’s motive: I am watching the initiation of war by the US and Israel against Iran, and am thinking about how the world works, and in general terms rather than the specifics of the moment. I want to see the aggression of the combination of the US and Israel as a result of generalized evolutionary mechanisms, that impose their rules upon the development of “humanity”, a concept that combines our species with “human values”. This concept is in opposition to the evolved strengths of our beast, our violence, our “lethality”, as Pete Hegseth likes to put it, the blood lust part of us. It is this, primarily, that civic culture hopes to moderate, as I see it. I am looking for the form of “humanity” in the abstract, and to see how it becomes lost to the appetites of the larger systems within which we dwell. It came together as this prompt, which I put to the Lumo chatbot:
Humanity is as a vector field, upon which is impressed a form. The form is able to evolve in competition with its peers. The evolving form results in stresses upon the vector field. At certain locations and times there may exist cross-over points between the enclosed system and the enclosing system. Provide real examples.
However, the substack editor does not like the tables that chatbots, including Lumo, tend to use, and so I have instead published Lumo’s response at: https://chatbotics.org/Conversations-With-The-Machine/Humanity-as-a-Vector-Field.html