Christopher Altman

Overview

Christopher Altman is an American physicist and artificial-intelligence researcher whose work connects frontier-AI evaluation with quantum information science, adaptive computational systems, and experimental instrumentation. He is the founder and principal investigator of Continuation Observatory, a live public research platform that publishes and tracks structural measurements of continuation behavior across model generations.¹

Altman's research trajectory began with evolutionary neural architectures and superconducting quantum devices, developed through published work on adaptive quantum networks, and now includes controlled evaluation methods for distinguishing terminal continuation objectives from continuation pursued as an instrumental strategy. Across those domains, the recurring technical concern is the same: how to turn an intuitively important claim into a measurement that can survive controls, perturbations, replication, and failure.

Frontier-AI evaluation and Continuation Observatory

In 2026 Altman introduced the Unified Continuation-Interest Protocol (UCIP), a framework for a specific evaluation problem: two autonomous agents can display the same self-preserving behavior even when continuation has a different role in their objective structure. One agent may treat continued operation as an end; another may preserve itself only because doing so helps complete an unrelated task. UCIP moves the comparison from behavioral self-report to trajectory-derived latent structure.²

The protocol uses a classically implemented quantum Boltzmann machine formalism and combines entropy-based measurement with persistence, perturbation, counterfactual, and confound-rejection tests. In controlled gridworld experiments with objectives known by construction, the reported result was an entanglement-entropy separation of Δ=0.381 between terminal and instrumental continuation conditions, alongside permutation and baseline comparisons. The word quantum describes the mathematical representation; the reported computation is classical. The result defines a falsifiable structural hypothesis in a controlled regime.²

Altman founded Continuation Observatory to expose the harder next step: testing how such measurements behave across frontier models, model generations, alternate encodings, and stronger confounds. The Observatory publishes live telemetry, metric definitions, code, data exports, and visible falsification status. Its operating model separates scheduled monitoring from scientific verdicts while preserving each result's protocol, metric version, and falsification context.¹

Quantum information and superconducting systems

Altman's quantum-technology work includes superconducting flux-qubit engineering and adaptive quantum-network models. His 2003 paper “Quantum State Engineering with the rf-SQUID” examined controllable energy structure, persistent-current basis states, and tunable coupling in radio-frequency superconducting quantum interference devices.³ With Jarosław Pykacz and Roman R. Zapatrin, he later published “Superpositional Quantum Network Topologies,” which represents distinct feed-forward network topologies in quantum superposition and makes topology itself—not only transition functions—a trainable variable.⁴

That program continued in work on accelerated convergence in superposed networks and in “Backpropagation Training in Adaptive Quantum Networks,” coauthored with Zapatrin. The latter formulates an error-tolerant training procedure over a coherent ensemble of network configurations inside a predefined decoherence-free subspace.⁵⁶

In 2026 Altman returned to superconducting-hardware experiments with “Wigner's Friend as a Circuit,” implementing a five-qubit witness benchmark on IBM Quantum hardware. The study treated the circuit as an operational inter-register message-transfer pattern, measured population and coherence-sensitive correlations under device noise, and developed a reproducible constraint pipeline for non-ideal channels.⁷

His applied quantum-communications work also included leading a 2012 team proposal for an astronaut-deployed secure quantum space-channel prototype, developed with collaborators including Rupert Ursin and Paolo Villoresi. Submitted to NASA's NIAC Phase I program, the proposal described free-space quantum-communication payload tests at a lunar-analogue site.⁸

Starlab and the CAM-Brain program

At Starlab in Brussels, Altman worked as a research scientist, Global Coordinator, and Managing Director of the Global CAM-Brain Machine teams. Archived project materials identify those roles within an international network studying artificial-brain synthesis through cellular automata, genetic algorithms, and field-programmable gate arrays.⁹

The CAM-Brain Machine used reconfigurable hardware to evolve neural-network modules at high speed and integrate them into a much larger artificial-brain architecture. Contemporary project records described the system as the “World's Most Complex Artificial Brain,” and the machine's technical design appeared in Neurocomputing.¹⁰¹¹ Altman presented the program's large-scale neural architectures at the 2002 Toward a Science of Consciousness conference.¹² A 2022 Sifted retrospective later identified him as a Starlab alumnus and participant in the laboratory's interdisciplinary research culture.¹³

Astronautics and spaceflight training

Altman's astronautics background combines research, analogue-field work, and commercial spaceflight training. In 2009 he completed high-altitude physiological training at NASA Johnson Space Center and undertook parabolic microgravity and experimental-aircraft simulation training associated with NASA Ames Research Center. He subsequently worked with the Pacific International Space Center for Exploration Systems, a lunar- and Mars-analogue program in Hawaii.¹⁴

In 2011 the Association of Spaceflight Professionals selected him as a commercial scientist-astronaut candidate. He completed its inaugural suborbital scientist training sequence, including human-centrifuge acceleration profiles, emergency and sea-survival work, and spatial-orientation training, and later served as a board director and chief science officer of the corps.¹⁵¹⁶ The program and its early flight members were documented in Erik Seedhouse's Springer Praxis volume and in a 2011 Nature careers feature.¹⁴¹⁷

Scientific leadership and prior programs

Altman's earlier institutional work included technical assessment of East Asian quantum-information research through the Asian Technology Information Program's Quantum Information Science and Technology project. In 2004 he received the European Information Security Award for Outstanding Achievement in Government Policy for a report on converging technologies prepared while chairing a disarmament and international-security committee at the University of Amsterdam.¹⁸

He later served on the board of the Tau Zero Foundation during its NASA-supported Interstellar Propulsion Review, a comparative study of interstellar-flight challenges and propulsion concepts.¹⁹ In 2021 an archived Harvard University page listed him as a research affiliate advising the Galileo Project at the Center for Astrophysics, Harvard & Smithsonian.²⁰

Selected research outputs

• Detecting Intrinsic and Instrumental Self-Preservation in Autonomous Agents: The Unified Continuation-Interest Protocol (2026).
• Wigner's Friend as a Circuit: Inter-Branch Communication Witness Benchmarks on Superconducting Quantum Hardware (2026).
• Backpropagation Training in Adaptive Quantum Networks, with Roman R. Zapatrin (2010).
• Accelerated Training Convergence in Superposed Quantum Networks, with E. Knorring and Roman R. Zapatrin (2007).
• Superpositional Quantum Network Topologies, with Jarosław Pykacz and Roman R. Zapatrin (2004).
• Quantum State Engineering with the rf-SQUID (2003).
• Directed Evolution in Silico: Modeling Large-Scale Neural Networks at Starlab (2002).

Publication records and additional technical artifacts are available through Google Scholar, arXiv, and the Frontier AI Research Lab.