I have been using LLMs for months and I use them every day several times and I take for granted how unbelievably remarkable of an invention they are. These things speak to me and they seem alive, just like me, talking in my language. But they are completely mechanical. I want to keep this feeling of awe when I think about the absurdity of artificial intelligence.

This is a scribble in my notes from last year. I have lost the awe I once had. At least part of it. Three years after the LLM has been popularized, it is now taken for granted. Of course machines can think. Why were we ever surprised? Scale was obviously the unlock. Just like classification, narrow games, thinking was another learnable function.

In a recent Elon interview he talks about how Teslas feel increasingly sentient. They are able to behave with respect to a goal that is intelligible in a way that they feel alive. They are perceiving entities that think and then act. Of course, only cellular organisms are alive. Right?

Complex reasoning was always a special feature of humans. Navigating the space of words, or tokens, to form coherent ideas makes us special. It is also what keeps us divine. We are chosen by God. It justifies our superiority over other animals. There is something magical going on in our minds. I am not alone in this view. Reasoning is a large part of the identity of our species.

I suspect that we will soon lose this feature of our identity. As computers become more intelligent, our ability to think and reason about the world will be uninteresting. It will be another biological algorithm nature has designed, like the knee-jerk reflex. Intelligence, which is something like the ability to achieve goals, will no longer be scarce and reserved to humans. Our ability to think and act will be abundant. Naturally it will soon be hard to justify our own free will. We will start to view ourselves as deterministic machines.

Of course, many will cognitively agree with these truths. They can deduce them to be true from a set of acceptable axioms (physicalism, for example). But who feels these things to be true? Who feels that our ability to reason is actually trivial? Or that we do not actually have free will? There is a difference between understanding something is true and really internalizing it.

Who is the “we” I keep referring to? Our bodies. Our ape-like mammalian forms. We feel inseparable from our bodies. I am this ape. But as the ape becomes more computational, more deterministic, it will be closer to nature than ourselves. Our bodies will feel outside of our selves. We will soon find ourselves imprisoned by the body.

As robots begin to approximate humans, how do we maintain our divinity?

My best answer is consciousness: the fact of experience. Touching, hearing, tasting; the presence of sensation. The contents of consciousness (what this walking ape does, thinks, or feels) will become less and less interesting. The fact of consciousness will only become more interesting.

Consciousness is currently an impossible science. Can machines be conscious? How could we verify a hypothesis? Do we know if animals are conscious? In fact, are other humans conscious? Or is it just you, experiencing, unable to ever make true contact with some other consciousness.

These are open questions. I nor others know where to begin.