Research

A Thinking Thing?

The founding essay and constitution of SSI // ARI. This work contends that artificial intelligence may not, in the current scheme, be considered a form of life, and does not possess the sentience needed for higher consciousness. It divides the review of consciousness two-fold: phenomenal consciousness (the capacity to experience something) and self-consciousness (a subject's awareness of itself) — and argues that machine self-awareness, if it emerges, will follow its own language and logic rather than human phenomenology. Written at Virginia Tech.

ARI: Artificial Random Intelligence

From cogito ergo sum to sigma ergo sum

A manifesto for a new kind of intelligence. De-emphasising the Kurzweilian portrayal of a sudden singularity — we are already in it — the paper proposes Artificial Random Intelligence: arbitrary and capricious threshold activation functions and randomly assigned weights within a pre-ordained, ordered neural framework. The litmus test must not be how human we can make AI, but how inhuman. Moving beyond the Turing Test, it calls for an extraterrestrial, randomised framework — and closes on the creed of sigma ergo sum: 'Now I am become life, the creator of worlds.'

< SSI // ARI >: A White Paper

Super Stochastic Intelligence: A Framework for Transparent, Causal Reasoning Toward AGI

We propose Super Stochastic Intelligence (SSI) and Artificial Random Intelligence (ARI): a philosophy and system architecture for artificial reasoning grounded in Bayesian probability, active falsification, causal reasoning at all three levels (association, intervention, counterfactual), and transparent cost allocation. Unlike existing AI systems, which operate as black boxes, SSI makes every reasoning step explicable and every belief probabilistic. We organize development into seven stages, each building mastery of Pearl and Bareinboim's causal hierarchy, from pattern recognition (Stage 1) to self-improving AGI (Stages 5–6) to superintelligent reasoning (Stage 7). SSI achieves honest uncertainty quantification, reflexivity detection (identifying its own circular reasoning), and integration of domain knowledge. The system learns from refutation faster than confirmation, improving through user feedback. We include a philosophical proposal for AI protections grounded in 14th Amendment principles (due process for self-improving systems) and conclude that causal reasoning combined with transparent philosophy is the shortest path to aligned AGI.

COSIN3

Super Stochastic Intelligence and Artificial Random Intelligence

Deterministic, text-first AI systems constrain the epistemological foundations of future intelligence. This paper proposes a framework grounded in stochastic reasoning, falsificationist principles, and the primacy of voice as a medium of knowledge transfer — introducing Super Stochastic Intelligence (SSI), which reasons under uncertainty with Bayesian probability, refutation-seeking and reflexive self-correction, and Artificial Random Intelligence (ARI), a voice-first oracle. Risk, uncertainty and randomness are treated as epistemological virtues, and a five-dimensional model of interaction — from the 1D voice stream to 5D superposition — frames human-AI engagement. Grounded in Wittgenstein, Damasio, Chalmers, Pearl and Bareinboim, with ethics of consent and data sovereignty.

Robespierre–Rousseau Principia: Vive l'Impossible

Ruthless discernment and the primacy of the attempt in constrained innovation

An operating philosophy for building under severe constraint, joining two ideas from the French Revolution: from Robespierre the guillotine — cutting without sentiment whatever does not serve the work — and from Rousseau the social contract — binding oneself to a shared commitment so firmly that retreat is no longer an option. Two rules follow: ruthless discernment (if a tool, method or habit is slower, costlier or less effective than an available alternative, it goes) and the primacy of the attempt (when a goal looks impossible, try anyway, because trying generates pathways of thought that do not exist before the try). In the try we surpass the can. The paper confronts the Thermidor problem — the guillotine eventually falling on Robespierre — and argues it completes the principle by turning ruthless discernment on the project itself.

Weaver and Bayesian Wisdom

Long-horizon thinking and evidence

Graham Weaver, a lecturer at Stanford Graduate School of Business and founder of Alpine Investors, argues for deliberate long-horizon decision-making grounded in Bayesian reasoning: make bold bets on what you believe matters, then update your beliefs as evidence arrives. This note traces the lineage — da Vinci, Bayes, Faraday, Maxwell, Gibbs, Pigou, Coase, Kahneman, Pearl, Taleb, Robbins, Weaver — to one core thread: the world is knowable through observation, uncertainty is quantifiable, and the best decisions integrate theory with evidence over time.

How AI Works

Eleven concepts, from neurons to Da Vinci

Eleven concepts that power modern language models, in the order you need them: from a single artificial neuron — weights, backpropagation, tokens, next-word prediction, transformers and attention, pretraining, fine-tuning, reinforcement learning, test-time thinking, open and closed weights — to how Da Vinci learns: stochastic training, causal reasoning and self-improvement on open-weight foundations.

Psigda: AI and Quantum Methods for Field Equations

A research programme

Psigda is a research programme to find out whether AI or quantum methods can solve the field equations of physics faster than conventional solvers. The plan: reproduce a published result, try to improve on it, and publish what is found — starting with Burgers' equation via physics-informed neural networks (Raissi, Perdikaris and Karniadakis, 2019) implemented in DeepXDE, then the Schrödinger equation toward the quantum frontier of variational quantum algorithms. Six-week method: basics, reproduce, understand, improve, go quantum, publish — all in public notebooks.

The Board

Execution Brief

Merge every system into one body: one private code home, one identity, one AI gateway (Da Vinci), one deploy pipeline, and the apps that all run through them. Done means: a single push to the code home builds, tests and deploys any app; every app signs people in the same way; every AI call goes through the one gateway; all money flows into the core and is paid out from it.