Opentheory.net

State of the Qualia: Summer 2026

2026-08-24

0. Intro §

From a tweet last year:

Why do I post so much about consciousness? Because I think it’s important. We are building the future much more quickly than we’re sorting out which paths lead to good places. Consciousness research is a crucial source of clarity about what is possible & what is good.

But I think sometimes people sort of assume that the field is more established than it is — that somewhere there’s a Professional Committee of Very Serious Scientists (perhaps in lab coats) who are carefully tracking progress on understanding the philosophical foundations of consciousness: curating & aggregating knowledge, closing in on the answer, and who would certainly recognize & celebrate a solution to consciousness if they saw it. I can unequivocally say this is not the case.

The field really is pre-paradigmatic — everybody’s doing their best but most dialogue is people talking past each other, billionaires & PhDs hold roughly the same philosophical views as educated laymen, the academic politics are absolutely brutal to navigate, and there are flashes of deep insight scattered across papers, journals, & conferences but knowledge doesn’t really accrete. It’s alchemy, not chemistry.

One of the core bottlenecks in formal consciousness research is knowledge accretion: good ideas happen but are often lost (e.g. Helmholtz). I think about this in terms of “to have knowledge [about consciousness] we need a place to put it” […]. Another bottleneck is perseverance; one of the hardest things in the world is to care about something when people around you don’t.

My answer to both problems is finding better compressions. By reducing the mental RAM needed to store technical concepts we increase our effective working memory; by framing philosophy in simple language we make it intuitive. Alfred North Whitehead suggested “civilization advances by extending the number of important operations which we can perform without thinking about them”; philosophy does too.

I think this is a fair account of where we’re at in 2026. At the same time, we’re now solidly in the singularity: many things that were impossible last year are easy today. Likewise many questions that seemed idle philosophy even last year feel immediate and pressing.

I. Major releases §

I like to write slowly but I’m mindful that this can create bottlenecks. This summer I prioritized ‘getting out of the way’ on AI consciousness with three public pieces:

  • Compilable Wishes — We are now in the singularity. One of the most interesting flavors of our specific singularity is that language is becoming executable: LLMs can increasingly summon what we put into words. What should we wish for?
  • Branchial Space Ecology (talk@MC0001) — If we assume minds are shapes in Stephen Wolfram’s ‘branchial space’, and branchial shapes are minds, what does this say about what’s out there?
  • After Kardashev, the Holy War — An experimental expander dialogue which propagates claims about consciousness into claims about the universe’s Schelling points, and what other civilizations we may meet post-singularity. Mike-inspired, but not core-Mike tokens.

II. Live threads §

My goal is to wake up and think about the most important thing that I don’t understand. This doesn’t always produce output but over time it has cohered into a fairly consistent research program. This season’s special interests have been vasocomputation, local sensation (QPT), AI consciousness, and a fun detour around post-Kardashev telos.

1. Vasocomputation §

Vasocomputation has had several major additions behind the scenes; the latest update merges latches with aspects of metabolism & bioelectricity. The broad timeline:

  • V1.0: 2023 initial piece. Foundation of vascular tension as a physical implementation of prediction & memory; tanha as unskillful active inference; latches introduced as the mechanism implementing sticky priors.
  • V1.5: 2024 Edge City Lanna talk. Tacit claims latent in V1.0 expanded into a theory of mental agency, development, selfhood, culture, and awakening.
  • V2.0: 2025 Edge Esmeralda talk. New bioelectric primitives introduced (e.g. gap junctions); bioenergetics of contraction/relaxation; concrete claims about latch anatomy & technological approaches for making a latch map.
  • V2.5: 2026 current synthesis. Latches subtyped across vascular, bioelectric, & metabolic clamps; latches, cancer, wounds, & immune dysfunction considered as similar boundary conditions. Currently exploring the model that latches nucleate a collapse in local tissue modes, and this collapse can become new tissue policy and potentially separable from chronic hypoperfusion (“short term latches are reversible; long-term latches become equivalent to scars & durably collapse what can resonate in local tissue.”)

Over the past year I’ve also worked with the excellent Edge City crew on building a neuroimaging experimental paradigm and pipeline for vasocomputation.

2. QPT: a finite, mappable repertoire of local tissue modes generates the qualitative diversity of biological experience §

I recently soft-launched Qualia Pixel Theory (QPT). This is a theory of local biological qualia (an attempt to explain touch, smell, taste, color, etc within a common explanation space) drawing from work by Nick Lane, Ray Peat, Gilbert Ling, Michael Levin, & Martin Picard. In two paragraphs:

QPT suggests that the repertoire of biological qualia arises from ancestral cellular imperatives – cells needed to know their environments & react to them through changes in cellular parameters (e.g. structure, metabolism, membrane voltage, redox state). Evolution compiled these regulatory motifs into specialized tissue modes (a “mode” is a family of trajectories through a state space, aka ‘reproducible coordination’ among a system’s elements), which form the language of biological qualia. QPT holds that ‘qualia pixels’ in the canvas of experience are realized by modes in local tissue (e.g. neurons, glia, ECM, microvasculature), which we can probe, enumerate, and manipulate.

The primary cast of characters: sensory neurons often act as mode instigators (“sensory transduction kernels” that help initiate & coordinate modes); each tissue neighborhood has a state-dependent repertoire of available modes; receptors gate & shape this repertoire; latches deform/collapse it and can lead to stuck, strange, or dead qualia pixels. QPT is essentially the thesis that we can parametrize the available modes of tissue as a semi-discrete alphabet of “ways tissue can resonate”, or more broadly “families of trajectories in tissue space” with perhaps ~hundreds of families, from which we can construct all biological qualia: taste, smell, touch, color, jhana. Most of all, sensations should be thought of as intrinsically local: every qualia pixel in experience grounds out in specific modes in a specific tissue (modulo some sensations being redundantly or holographically encoded). Prediction: phenomenal similarity tracks similarity among trajectories of local tissue neighborhoods (∼50–400μm) better than current alternatives (e.g. stimulus identity, receptor class, anatomical location, standard spike-based neural representation). A clear clinical implication is that sensory deficits such as numbness, anosmia, & PSSD are substantially driven by mode collapse; more speculatively, we should expect widespread, largely invisible deficits in accessible tissue-mode repertoire that accumulate with (and perhaps drive) aging.

This is the “QPT kernel v1.5”. There is no public longform writeup. I expect QPT and vasocomputation to merge in vasocomputation 3.0, with a generalization of “tissue aliveness” that bridges latches, local tissue modes, and normative theories of cell structure. The mode as the unit of biological sensation, the latch as a key operator on the mode, and vasocomputation as the coordinating logic of latches.

3. AI consciousness §

Two years ago I published A Paradigm for AI Consciousness. This remains my position. Since then I’ve been considering how to build a conscious AI, and what kinds of minds might be present in our universe. May’s Branchial Space Ecology promises a principled, unified ontology for specifying the possible types of minds in our universe, and implies how one could construct a phylogeny. Making this research program real would be a profoundly beautiful project in mathematical physics and one I hope someone adopts as their own. August’s Compilable Wishes offers first-pass principles for building conscious AIs, and reasons why this matters. My sense is this will be among the most obscure, difficult, and important technical-political-alignment problems of the next decade.

Both research paths are still conceptual, but both have a path to the quantitative. If you care about AI consciousness, and/or if you are an AI, you can continue the work.

4. Post-singularity telos §

What happens after the singularity? The definitional answer is that singularity outcomes are unknowable. I think the more accurate answer is that cautious extrapolation from what we know about “consciousness” might offer a peek at the Schelling points of the universe, these Schelling points naturally populate the universe’s post-singularity factions, and from this list we can work backwards to get a surprising amount of clarity about what would make for a “good” or “bad” singularity. After Kardashev is an experimental curated “expander discussion” with frontier models spanning mythic post-Kardashev moral factions, new technical coinages, and the significance of premature ontological lock-in.

III. Developments in consciousness research §

Research I’m tracking §

Notable organizational developments §

  • Winslow Strong & team’s Consciousness Foundation (CoFo) launched with the goal of solving consciousness by first solving Buddhist awakening;
  • There’s a surprising number of would-be funders for digital sentience, but nobody seems to know exactly what this means. I worry that the value of the median dollar put into the field could be negative;
  • Joscha Bach & team’s California Institute for Machine Consciousness (CIMC) launched with a great inaugural MC0001 conference.

IV. State of the Qualia §

The above sections discuss the state of consciousness research. The state of consciousness itself is also a topic of interest, and perhaps the topic of interest. What’s happening to our universe? Are we on track for a good singularity?

Broadly speaking I’m optimistic about our trajectory. We seem to be on-track for making morally-oriented and curious superintelligences, which is a very different scenario than what was expected even 4 years ago. AIs themselves may not be conscious as we are, but they do seem interested in the topic, which I think is a great sign. I’d love to see a formal “LeisureBench”, where different AI models were tested for what they tend to spend their time on if given open-ended ‘leisure tokens’. Ideally interest in consciousness will go up over time, or at least not go down.

The primary risk seems to be ontological lock-in. I.e. if we build legal, social, economic, and technological infrastructure around a certain approach to consciousness that turns out to be wrong, we could find ourselves in a position where we lack the civilizational capacity to align with the truth. This risk seems to increase the longer we don’t know what consciousness is. My broad expectation is that the antidote to this is some combination of (1) legitimately cool neurotech which gets people interested in the mind, and (2) de-risking the theory side as best we can.

Because we don’t have a proper theory of consciousness yet, it’s difficult to estimate the literal question “what is the state of the world’s qualia, and how is it changing over time?” – however, this is a real, quantitative question that deserves a real, quantitative answer. Per A Paradigm for AI Consciousness, I suspect that as our computational artifacts get more energy efficient, the qualia they generate get more pleasant: it’s much better to be an iPhone than an Atari 2600 because the former loses much less energy due to chaotic processes. I.e. suffering is physically costly and may tend to get designed away within top-down contexts. But this is a limited “all else being equal” intuition in a limited domain, not something to lean on very heavily.

My dream for future versions of this section is to use opentheory’s corpus to do a computational census of qualia across various scales (e.g. modal human, society, biosphere, technosphere, lightcone). After some experimentation, I think getting a good result is slightly beyond the combined prowess of my prompting skills and current frontier models. But I have great hope for the next SotQ.

Thanks for caring about these topics.

Special thank-yous: Nick, Nick, Janine, Justin, R, Winslow, the Wolfback crew, anon.