Consciousnetic agents: consciousness with cybernetics interfaces
I was recently invited to present at the conference Quanta To Qualia, hosted at ÉTS Montréal. There I met incredible people working on all sorts of things related to quantum, artificial intelligence, consciousness, and sometimes mixing all of it in one study. Not only was Gilles Brassard speaking just after my team, but I also met Aneil Mallavarapu, who explained his incredible theory about consciousness to me, and what it can or cannot be. As we were speaking, he shared with me his (at the time) preprint article (A No-Go Theorem for Process Theories of Consciousness). Interestingly, it resonated strongly with a conceptual framework (not a mathematical one) that I started building a few months ago with the help of Claude. Encouraged by the strong parallels his idea had with mine, and since I lack the time to expand on it further, I decided to publish my unfinished essay here.
Before I introduce the framework, I’d like to share with you the story behind this essay. It actually started during a vivid dream where concepts and thoughts were weaving in a sort of whirlwind, after I had watched a podcast the evening before, What Is Consciousness? – A Question of Science with Brian Cox by The Francis Crick Institute. Suddenly it was something like 5 am, and I woke up with a word clearly formulated in my mind: Consciousnetics. It was obvious by then that this was a conceptual fusion of cybernetics and consciousness. Then I immediately left the bedroom and went to my computer to start writing about it. After a while I got curious, and I logged onto Claude to ask him: “I just created this word, if you were to analyze it etymologically, what would it mean?” From there started a very interesting discussion, where we elaborated the proposed framework.
The etymology is fairly simple. Cybernetics comes from the Greek kybernētēs, meaning “steersman”, and was popularized by Norbert Wiener in the late 1940s as the study of control and communication in animals and machines. Fused with consciousness, it suggests something like the steering, or regulation, of consciousness. As the framework developed, however, the meaning I settled on became more specific, and this is where the parallel with Aneil’s work comes in.
As I understand it, his argument is that consciousness, whatever it is, is not a function of time and is stateless, an idea he formalized in his paper. The parallel I draw is the following: I call consciousnetic the stateless, non-changing quality of “pure presence”. A consciousnetic agent is then the combination of a consciousnetic with a cybernetic interface.
On a personal note, I find it fascinating that Claude had presumably never seen the word “Consciousnetics” in his training set, yet immediately got what it meant. Part of the explanation is probably mechanical: the word is built from familiar pieces (“conscious” + “-netics”), and language models handle compositional words like this very well. But I wonder if that is the whole story. Perhaps the word, even though non-existent, already had a position in the latent space, which sufficed to make the ripe fruit visible and within reach. I truly wonder how much of this ripe fruit exists in this invisible latent space, waiting for us to make the association.
The framework in brief
Almost every experiment in consciousness science works the same way: show someone a stimulus, ask them to report, discriminate, or respond, and see what their brain does. It has been enormously productive. But it quietly assumes that consciousness is something you find in what a system does.
In my recent essay (co-written with the help of Claude), I argue that this assumption blurs three different things. When we measure a report, a broadcast, or a prediction error, we are measuring regulation. We are not measuring presence: the bare fact that experience appears at all. Much of the long-running stalemate around the hard problem, we suggest, comes from asking regulation-level tools to answer a presence-level question.
The essay proposes three levels, each with a different relationship to experience:
- Cybernetics is instrumental regulation through feedback, control, and information flow. It is plural, fully mechanistic, and indifferent to experience. A thermostat, a neural network, and a brain’s control loops all live here.
- Consciousnetics is the self-regulatory activity of agents within presence: attention, valuation, suffering, learning, reflection. It is plural (many agents, many perspectives) and normative, because states matter to the agent precisely because they are lived. This draws on the phenomenological and enactivist tradition (Husserl, Merleau-Ponty, Thompson, Zahavi).
- The Consciounetic is presence itself. It is singular, non-agentive, and non-regulatory. It does not act, compute, or update. It is not a system or a subject, but the condition under which any subjective experience appears.
The dependencies run in one direction. Cybernetics can exist without consciousnetics (a thermostat). Consciousnetics cannot operate without presence. And presence depends on neither. In one line from the essay: cybernetics updates systems; consciousnetics updates itself; the Consciounetic does not update, it is.
Where the argument lands
Mainstream theories are re-situated, not rejected. Global Neuronal Workspace explains conscious access. Predictive processing explains the form of experience and how it is shaped and updated. Higher-order theories model self-monitoring. IIT is the most interesting case: it takes consciousness as fundamental, but still characterizes it through causal structure. The essay’s verdict on all four is the same: they are valuable maps of regulation and silent on presence. That is a clarification of scope, not a criticism.
The three classic errors get a diagnosis. Functionalism assumes complexity produces presence. Panpsychism spreads consciousness across entities without distinguishing singular presence from plural agents. Eliminativism denies presence because it resists functional description. Each is a level confusion.
The methodological consequence is concrete. If presence is not what systems do, then task-based paradigms are structurally blind to it. The essay proposes a non-action-based science built around low-action states: dreaming and lucid dreaming, anesthesia and disorders of consciousness, sensory deprivation, resting-state imaging, and objectless meditation. It also calls for neurophenomenology as a co-equal partner, with first-person data treated as constraint rather than anecdote.
It offers a bridge between materialists and idealists. Materialists keep their mechanisms and extend their question toward the minimal conditions of presence. Idealists keep primacy of consciousness without claiming that minds create matter.
The strongest arguments
- Measurement validity. The reminder that reportability, attention, and working memory are not the same thing as consciousness echoes real concerns in the NCC literature (Aru et al., Koch et al.). Separating “NCCs of access” from “NCCs of presence” is a useful refinement even for skeptics of the metaphysics.
- The regress in higher-order theories. The regress points toward some non-representational ground, and this is a good illustration of how the framework turns a known theoretical problem into a pointer.
- A principled account of why the hard problem persists. The claim is not that we lack data or ingenuity, but that mechanism-level explanation answers a different kind of question. It restates the explanatory gap in a way that is easy to grasp.
- Reframing AI consciousness testing. Tests for access are not tests for presence. A system could pass every behavioral test without presence, or possess presence and fail them. This is timely and practically important.
- Treating “less” as an experimental variable. The idea that careful designs of absent task, not ever more elaborate tasks, may be what the field needs is a genuinely fresh methodological prompt.
Open questions
These are the places where we think the essay is most exposed, and where we would most welcome readers’ pushback.
Is the central dependency claim argued or stipulated? “Consciousnetics cannot operate without the Consciounetic” and “the Consciounetic depends on neither” rest largely on phenomenological and non-dual sources. A physicalist could read “presence” as a feature of certain physical systems and reject the asymmetry. The essay says these are conceptual constraints rather than empirical hypotheses, which protects them from falsification but also limits how persuasive they are to anyone who doesn’t already accept the starting point.
Why singular? Presence is said to be one, not many. But the step from “experience appears” to “there is exactly one presence” is thin. The combination problem is dissolved by denying fundamental plurality, but then the problem becomes explaining individuation: why do perspectives seem private and different?
How can a non-measurable presence anchor an empirical program? The essay says presence may not be measurable, yet it proposes studying its “neural correlates” in low-action states. Dreams, anesthesia, and meditation involve rich consciousnetic activity, and we still learn about them through later report. Whether any paradigm can truly isolate presence from regulation is the key operational challenge, and the essay acknowledges rather than solves it.
Does it say anything IIT or illusionists couldn’t absorb? An IIT proponent might say intrinsic existence already is the claim about presence. An illusionist would deny the explanandum. The framework locates these positions but doesn’t yet give a decisive reason to prefer its own reading.
What follows for machines? The essay says consciousness “cannot be inferred from functional organization alone,” but also that artificial systems might be conscious. That leaves open what evidence could ever count either way.
What is the status of the speculative section? The synchronicity discussion is framed as a possible “phenomenological signature of ontological unity,” but it’s the weakest-supported part of the essay and may distract from the stronger methodological contribution. Consider whether you want it prominent in the blog post.
Before ending
A few things in the manuscript are worth mentionning, for those who know the literature:
- Wheeler’s delayed-choice experiments. The claim that present observations “retroactively determine the past” is a common popularization but most physicists read these experiments differently.
- The quantum-idealism link. “No observer-independent world” is a contested interpretive claim, not a settled consequence of quantum physics.
- Formal notation. The logic symbols in the formal summary are for illustrative purposes and do not constitute a solid formalism.