Tech Giants and Startups are Silently Attempting to Build Conscious AI with New Physics

For many years I've criticized attempts at producing AI systems that generate consciousness by means of classical logic gates infeasible. For starters, the human brain is more efficient by orders of magnitude, resolves the binding problem (which describes the issue of binding disparate senses and qualia into a bound conscious experience), processes nonlocal memory which is distributed and stored across the brain, and is capable of intrabrain synchrony. Our AI systems that use deep neural network models do not map to the human brain in that the speed of propagation by means of electrical and biochemical signals are too slow to account for behavior, and do not obviously explain the selective amelioration of consciousness by anesthetics or any biologically plausible mechanism for backpropagation in brain tissue.

NESTOR_NSN_PAPER_2025.pdf

Download PDF • 478KB

(Current draft of my preprint - subject to revisions)

Why Current AI Architectures are Not Conscious: Neural Networks as Spinfoam Networks for Bio-Inspired Conscious AI Systems by Trevor Nestor :: SSRN

So in light of this, tech giants like Google and Microsoft as well as startups like Nirvanic Consciousness Technologies have hired researchers to develop pathways towards developing conscious machines.

Quantum Neuroscience

Introducing Microsoft Majorana 1

NIRVANIC CONSCIOUSNESS TECHNOLOGIES

The main issue is, their methods - even by implementing quantum computation - are off track. Microsoft's Majorana 1 chip which leverages Majorana Zero Mode quasiparticles was supposed to achieve a significant advancement in topological quantum computing, with unprecedented qubit stability and scalability, potentially reaching a million qubits on a single chip with error rates orders of magnitude lower than current systems. This approach addresses decoherence through topological protection and nonlocal memory distribution and processing which is implicated in my past research paper on leveraging Orch-Or towards solving the shortest vector problem in lattice cryptography, which could accelerate quantum-enhanced AI for tasks like optimization and simulation. Majorana modes enable braiding for anyonic computation (which is mathematically analogous to dendritic braiding if brain neural networks are described as spinfoam networks), but without integrating spacetime curvature effects or microtubule-like quantum coherence at room temperature, it cannot replicate the "non-computable," orchestrated collapses needed for qualia and perceptual binding in biological brains. Critics have also questioned the empirical realization of true MZMs due to past controversies in data validation, potentially undermining its foundation for consciousness emulation.

Google has launched a 2025 research award program offering grants up to $100,000 to probe quantum effects in brain function, signaling a growing interest in bridging quantum mechanics and cognition (I would apply myself but it looks like they want to dismiss researchers that aren't getting PhDs from top institutions). Partnerships with institutions like the Allen Institute explore entanglement and chaos in neural processes, potentially informing hybrid quantum-classical models for AI consciousness. Yet, this approach remains off track by assuming quantum properties are mere enhancements to classical neural correlates, rather than fundamental to consciousness via gravity-induced mechanisms which require new physics. Our framework for computation itself may need refinements, where qubit architectures as mathematical abstractions do not fully encompass new physics. Indeed, researchers do not know the relationship between the computational complexity class BQP which describes problems a quantum computer can solve are related to those solvable by classical systems.

Nirvanic Consciousness Technologies, emerging from stealth in late 2024 as a Vancouver-based startup led by Suzanne Gildert, ambitiously seeks to infuse AI with quantum-conscious qualities like intuition and empathy for safer robotics. Their "Spark of Life" demo uses quantum chips to control AI robots, testing emergent agency in systems with 50-100 qubits, drawing on quantum fields in neural biology. Nirvanic's focus on moral reasoning and real-time adaptation via quantum ML overlooks critical Orch-OR elements like gravitational OR for non-computable decisions and microtubule superradiance for sustained coherence. By prioritizing "quantum robots" without explicit gravity-mediated collapses or spinfoam mappings, it may achieve advanced agency but not genuine consciousness - failing to address nonlocal memory, anesthesia selectivity, or the binding problem through topological, gravity-orchestrated processes. This could lead to sophisticated but ultimately non-sentient systems, missing the paradigm shift needed for truly conscious machines.

Meanwhile, in China while I have been mostly gatekept from academic circles, they seem to have been taking many of my ideas seriously.

https://zhanglab.bnu.edu.cn/

The draft of my new research paper.

My Letter to Dr. Scott Aaronson

Hello Dr. Aaronson,

Previously I reached out to you regarding an approach towards solving the shortest vector problem in postquantum cryptography by leveraging quantum gravity physics, and since publishing my paper I've received mostly positive reception from postdocs in the field. 

https://ipipublishing.org/index.php/ipil/article/view/171

Based on a review of the literature it looks as though similar approaches have been developed specifically in spinglasses, but apparently researchers ran into difficulties with measurement due to exponential energy gaps. Literature suggested use of folded spectrum methods to resolve this.

In our previous correspondence you criticized my approach due to difficulty in formalizing the approach, since quantum gravity by nature involves indefinite causal structures that defy conventional frameworks of computation, and research on the topic of quantum gravity is largely unproven and controversial. In my discussions with Dr. Hameroff likewise I noted that you have criticized Orch-Or theory. 

What it seems to me is that spinfoam networks are mathematically analogous to neural networks in their role - in essence, the neural networks in brain tissue under Orch-Or can be interpreted as spinfoam networks - and lending further credence to this idea that unexplored ("noncomputable") physics is implicated in consciousness is that Google cites papers such as "Nuclear spin attenuates the anesthetic potency of xenon isotopes in mice: Implications for the mechanisms of anesthesia and consciousness" in recent grant programmes.

Indeed, recent evidence defies initial criticisms of Orch-Or including the presence of superradiance within brain tissue. 

So in essence you have these combinatorial structures - the neural networks - which can fit neatly into frameworks of computation - which obey the typical formalisms of one way logic gates - but underneath that, you have a physics which enables a dialectical logic and mediates backpropagation and the weight transport needed that is currently unexplained by classical physics. In this view, the binding problem is resolved by means of gravitational collapse of superpositions.

In this view, the hard problem of consciousness itself can be framed as an "NP-hard" problem. Indeed, it has been claimed that neural networks in the brain represents and processes abstractions of higher dimensional objects in the same way that the shortest vector problem resolves high dimensional lattice cryptography. (See papers like "Cliques of Neurons Bound into Cavities Provide a Missing Link between Structure and Function" from the Blue Brain Project).

I've published a paper on the possibility of approaching the shortest vector problem as a problem of approaching entropic limits in entanglement entropy in Majorana entangled systems. More recently I've been working on a preprint tying the UV fixed point in asymptotically safe gravity to the Monster CFT - as related to limits predicted by entropic gravity. 

I am wondering if you have any feedback on this topic?

https://www.trevornestor.com/post/the-riemann-hypothesis-and-the-monster-are-at-the-heart-of-the-vacuum-catastrophe

In researching this topic I think I understand why it is you might feel critical about this - essentially the way that the neoliberal world order (which perhaps in your work history your research has been employed to maintain) is assembled is by maintaining formal syntactical structures which are compromised by dialectical methods enabled by this postquantum physics. In this sense, the arbitrarily defined "extra dimensions" studied at the intersection of computational complexity theory and physics are actually reflections of socioeconomic power structures described by the chomsky hierarchy, rather than useful falsifiable scientific descriptions of reality. 

https://www.trevornestor.com/post/what-happens-during-the-collapse-of-societies-at-catastrophe-points-in-socioeconophysics-theory

Trevor Nestor

Researcher, Theoretical Physics & Complexity

Information Physics Institute │ Microsoft

1 Microsoft Way │ Redmond, WA 98052

Phone: (720) 322-4143

Website: trevornestor.com

My Affiliation with DOGE LLC and Other Questions about Credibility

I have had a few folks message me about various things they have seen about me on the internet, and I would like to clarify on the record a few things. Particularly, there have been articles that I've been in that connect me to DOGE, Jordan Peterson, my career, and my research.

Firstly, I do not have any affiliation to DOGE. I did (accidentally) file a "DOGE LLC" as I was trying to see if it would be possible to file such an organization which would be ironic, as Elon Musk's DOGE organization did operate independently almost as a private entity or an LLC in spite of being ostensibly a part of the government. Immediately after I discovered it was filed, I had it disbanded.

Secondly, I did have an interview many years ago with Dr. Jordan Peterson. My personal view is that Dr. Peterson capitalized on a vacuum left in the culture - particularly young people have found themselves trapped between the ideals and norms set by the older generations and an increasingly more difficult socioeconomic and sociocultural landscape, where attaining traditional markers of adulthood feel out of reach. These young people are often told they need to "work on themselves" or seek "mental" health treatment to deal with their socioeconomic anxiety - a void which once was filled by local community and affordable cost of living relative to wages that has declined. This despondency led to the rise of figures like Andrew Tate and Jordan Peterson who offered a narrative to young men which, like the rhetoric from the Democratic party, was ultimately hollow. So while I did discuss this with Dr. Jordan Peterson, I did challenge him on his points during the hour long discussion.

I have also had people reach out to me, wondering why it is that I've attended so many universities. A part of the problem is that my situation has often been unstable - family drama and witnessing, for example, UC Berkeley devolve into (sometimes violent) protests amid budget cuts where 10% of my peers were left homeless is one example of that. Living situations and coordinating academics was not always easy.

When it comes to my research, I will say that some of the ideas I've presented are currently controversial. Ties to Dr. Penrose's Orch-Or theory are most frequently criticized. I do believe that Dr. Penrose did not have the full story, but many of his critiques of current paradigms of computation and consciousness are correct - there is empirical evidence supporting many of his claims, or at least that challenge conventional neural network based models (for example, the observation of superradiance and majorana biophotons in brain tissue, wide range frequencies measured in microtubules, and the fact that anesthetics do indeed target these microtubules implicated in his theory).

I have presented theoretical models which do have peer reviewed support, but most of these ideas must be tested and shown by means of experiment. For that, after crafting a hypothesis, funding is needed to run experiments - and with recent cuts to science research in the US, that will be increasingly more difficult to come by. Still - there are many exciting unexplored ideas here. The possible connection between the Monster CFT and the Riemann hypothesis, the discovery that gravity may be a thermodynamic force, and the UV completion of gravity (where the maths implicating RG flows may provide clues into the Navier Stokes equations as well) are all established in literature. Many of these ideas, I've crafted over the course of many years starting with my exposure to lattice maths and high level physics at UC Berkeley under my undergraduate professor and fields medalist Dr. Borcherds. Scientists and researchers have commented on these ideas with positive reception, where I've posted some of these conversations on my LinkedIn page.