A Broken Cosmological Compas
Through my algorithmic travels over the past few months, I have been following a thread of thought that emerges from time to time in between YouTube and TikTok. It started with two viral trends that came after each other. The first is Blake Lemoine, a software engineer fired from Google after going public with his concern about the sentience of Google’s AI chat, LaMDA. News items that proliferated throughout my online feeds from social platforms to news websites, were followed by a wave of criticism by scientists, journalists, and people involved in the artificial intelligence field. These critiques mainly centred around the argument that LaMDA, “a family of conversational neural language models” is designed to do exactly that: sound like a person. The second trend was mainly TikTok based, where users would film themselves being “main characters” in public spaces, in some some kind of performative assertion of agency by doing unexpected things, while making fun of confused passers-by for being NPCs (non-playable characters).
There isn’t perhaps much correlation between the two trends from a thematic perspective, other than perhaps both of them being some kind of speculation on the agency of an ‘other’, where the speculator does not have much data on the agency of what they are speculating on (whether that is AI, or an overwhelmed passer-by) other than their own experience of the behaviour of this “other”. What piques my interest is not the similarities between the so-called NPCs and the AI, as much as it is the similarities between the cognitive methodologies used to deal with the claim of AI sentience by “experts”, and the ones employed by TikTok users exporting their first person gaming strategies into the physical world.
From this perspective, methodologies of thinking about AI and the onto-epistemologies they are premised upon, start to reveal their saturation with racial and colonial logics embedded within considerations of agency based on the hierarchical ordering of the world according to differences and similarities to the self as a transparent observer. But for now, knowing that I tend to enjoy the occasional theoretical physics video after a long day of writing, my algorithmic friends decided to take me further.
I was recommended a short explanation on the intersection of theories of consciousness and quantum physics, and finding it interesting, I ended up reading Roger Penrose’s 1994 paper ‘Mechanisms, Microtubules, and the Mind’. This is a short paper based on a lecture he gave introducing the main ideas of his book Shadows of the Mind, where he develops the concept of ‘Objective Reduction’; a non-local process that sheds light on the decay rate (instead of instantaneous collapse) of the superposition of multiple possibilities into one. Accordingly, locating non-computable action (what physics cannot totally understand yet) in the movement from the Quantum as an enclosed computable level, to the Classical as another enclosed computable level. As such, Penrose puts forward the argument that the randomness of this collapse, as it is commonly understood, is due to an inability (until now) to isolate the experimental system where this collapse happens from the environment where it occurs. He goes on to argue that this isolation could potentially be happening in microtubules (a sub-cellular structure) in the brain, opening a new scientific facet towards the computation of consciousness.
As with AI experts and NPC hunters, the onto-epistemological stance is the same: hiding within an invisible self from where concrete judgements can be passed on consciousness and agency, while failing to acknowledge the logics within which the awareness of the thinker emerges. In the case of AI experts, the judgement is already out; we put the data in, so we decide on its agency. In the case of some TikTok users, in order to feel empowered, we imagine everything outside of ourselves to be simulated. In the case of a Nobel prize winner, judgement cannot be achieved at this point, we need more resources, and more capital, to chase down the generative complexities that complexity produces.
All in all, we seem to be falling deeper into individuated microtubular vortices, into self-oscillating complexity, increasingly imagining consciousness, agency, and sociality to be something separate from us, and separate from our environment.