We’re heading into a world of persistent, tool-using agents. Surface behavior alone may not be enough to tell whether shutdown avoidance or self-preservation is intrinsic or merely instrumental.
When an agent resists shutdown or preserves its continued operation, is continuation part of the objective itself — or merely instrumentally useful for maximizing something else? That distinction matters for AI safety, but it is often difficult to infer from behavior alone. Our protocol shifts the problem from surface-behavior interpretation to latent-structure measurement.
A simple analogy
Imagine two employees who both fight to keep their job. One values the work itself; the other just wants the bonus. Similar outward behavior, different underlying objective structure. We frame this as a problem of observational equivalence: shutdown avoidance, memory preservation, and risk reduction can arise under both intrinsic and instrumental continuation regimes.
The central claim is not about consciousness or subjective experience. It is that agents with intrinsic continuation objectives may produce more deeply coupled latent structure across time than agents for which continuation is only a means to another end. If robust, that would make continuation-seeking a measurable scientific object rather than only a behavioral impression.
Starlab veteran・日本語・Guinness Book of World Records・NASA・Kavli Institute・Harvard・TU Delft・Chief Scientist・Quantum Technology・Artificial Intelligence・NASA-trained Commercial Astronaut
We are at the very beginning of time for the human race. It is not unreasonable that we grapple with problems. But there are tens of thousands of years in the future. Our responsibility is to do what we can, learn what we can, improve the solutions, and pass them on.
— Richard Feynman
From CAM-Brain to Cambrian 2.0: Engineering the Next Intelligence Transition
We stand on the shores of a vast cosmic ocean, with untold continents of possibility to explore. As we continue forwards in our collective journey, scaling the cosmic ladder of evolution, progressing onwards, expanding our reach outwards in the transition to a multiplanetary species, Earth will soon be a destination, not just a point of origin.
From early childhood, I set out to convey a profound and positive impact on the long-term future of humanity — to make the world a better place for our children, our children's children, and the generations yet to come. As we're collectively propelled forwards as a species, I committed to ensuring core values of balance, integrity, and ethical responsibility are upheld with paramount importance in scientific research and principal government leadership. With unprecedented leaps and bounds of progress in our scientific understanding — enabled by the development of converging and expanding exponential technologies — newfound, unexpected discoveries await, just over the horizon.
Rapid advances in fields such as artificial intelligence, biotechnology, molecular nanotechnology, neuroscience, renewable energy, spaceflight, supercomputing and quantum technologies — each enabled by the recursive technological progress of Moore’s Law — will converge to confer radical changes to society over the coming decades, as we move forward in the collective transition toward the dawn of a post-scarcity economy. The future is unbounded. The responsibility falls upon us to ensure that its limitless potential is filled with dreams of hope, happiness, freedom and fulfillment.
In tribute to timeless, inspiring, and visionary friend, colleague, collaborator, and coauthor
Serguei Krasnikov
(1961–2024), whose midnight brainstorming sessions and legendary
time travel parties
at Starlab will echo through the ages. May we carry forward his boldest dreams, fulfill his most audacious ambitions, and meet again — somewhere, sometime, just over the horizon.
Altman, C, Zapatrin, R. Spacetime from Quantum Topology, in Spacetime from Quantum Topology, Ignazio Licata and Cecilia Flori, ed. Oxford University Press, Oxford.
(2015) Altman C. and Zapatrin R. Spacetime from Quantum Topology. Edited by Ignazio Licata and Cecilia Flori. Oxford University Press, Oxford.
(2015) Altman C. Invited contribution to McDonald, K, Flat World Navigation, sequel to Innovation: How Global Change Innovators Think, Act and Change Our World, with Vint Cerf. Kogan Page, UK.
(2014) Altman C., Belden C, Nicholson C and Ellis J. A global satellite network to secure air and space traffic worldwide. PeopleSat: A Comprehensive Solution in Response to the Disappearance of MH370. Washington, DC.
(2013) Altman C., Williams C, Ursin R, Villoresi P and Sharma V. Astronaut Development and Deployment of a Secure Quantum Space Channel Prototype. Pacific International Space Center for Exploration Systems. DARPA; NASA NIAC/OCT.
(2012) Altman C. The Race to Bring Quantum Teleportation to Your World. KurzweilAI Newsletter. October 5, 2012. Cambridge.
(2012) Altman C. Moving Plane Exchanges Quantum Keys with Earth. KurzweilAI Newsletter. September 17, 2012. Cambridge.
(2012) Altman C. Efficient tunable ion-photon entanglement interface enables quantum networks. KurzweilAI Newsletter. May 23, 2012. Cambridge.
(2012) Altman C. Quantum entanglement in spin qubits. KurzweilAI Newsletter. May 17, 2012. Cambridge.
(2012) Altman C. Austrian researchers set new world distance record for quantum teleportation. KurzweilAI Newsletter. May 21, 2012. Cambridge.
(2012) Altman C. A Boost for Quantum Reality: The Quantum Mechanical Wavefunction is Real. KurzweilAI Newsletter. May 9, 2012. Cambridge.
(2009) Altman C. and Zapatrin R. “Backpropagation in Adaptive Quantum Networks,” International Journal of Theoretical Physics, Vol 49, No 12. Springer, July 2009. London.
(2008) Altman C. and Zapatrin R. “Superposed Adaptive Quantum Networks,” International Conference on Quantum Structures, Brussels–Gdansk. Springer, London.
(2008) Altman C., Knorring E and Zapatrin R. “Accelerated Training Convergence in Superposed Quantum Networks,” NATO Advanced Study Institute on Mining Massive Data Sets for Security. Como, Italy. NATO.
(2007) Altman C. “Microlens Array Fabrication: Future Directions in Quantum Coherent Information Processing,” FISBA/TU Delft Faculty of Applied Physics. FISBA Optik, Switzerland.
(2004) Altman C., Pykacz J and Zapatrin R. “Superpositional Quantum Network Topologies,” International Journal of Theoretical Physics Vol 43, No 10. Springer, London.
(2004) Altman C. and Kahaner D. “Korean Quantum Information Research,” Korea Advanced Institute of Science and Technology (KAIST). Quantum Information Science and Technology Project, Asian Technology Information Program, Japan.
(2004) Altman C. “Advances in Quantum Algorithms,” Quantum Information Science and Technology Program, ATIP Tokyo, Japan.
(2004) Altman C. and Satoh T. “Japanese National Research and Development Programs,” RIKEN National Laboratory. Quantum Information Science and Technology Project, Asian Technology Information Program, Japan.
(2003) Altman C. “RIKEN Quantum Dynamics Research,” Quantum Information Science and Technology Project, Asian Technology Information Program, Japan.
(2003) Altman C. “SOKENDAI Quantum Information Research,” Quantum Information Science and Technology Project, Asian Technology Information Program, Japan.
(2003) Altman C. “Quantum Circuit Complexity Research,” Quantum Information Science and Technology Project, Asian Technology Information Program, Japan.
(2003) Altman C. “International Conference on Quantum Information,” Tokyo Institute of Technology. Quantum Information Science and Technology Project, Asian Technology Information Program, Japan.
(2002) Altman C. “Quantum State Engineering with the rf SQUID,” NATO Advanced Research Workshop on Quantum Chaos. NATO. Como, Italy.
(2002) Altman C. “Converging Technologies: The Future of the Global Information Society,” UNISCA First Committee. Chair Report to the United Nations General Assembly. Amsterdam. Recipient of the 2004 RSA Information Security Award for Outstanding Achievement in Government Policy.
(2002) Altman C. “Directed Evolution in Silico: Modeling Large-Scale Neural Networks at Starlab,” Towards a Science of Consciousness. MIT Press. Cambridge.
(2002) Altman C. “Quantum Uncertainty: The Boundaries of Empirical Knowledge,” Towards a Science of Consciousness. MIT Press. Cambridge.
(2002) Altman C. “UN Sustainable Futures: Eden Project,” Trimtab Newsletter, Summer 2002. Special Issue with United Nations Secretary General Kofi Annan. Buckminster Fuller Institute. New York.
2020 Keynote on the Future of the Military in Space · Space Mastery · Portugal
2020 To the Stars and Beyond: Deep Tech & AI · San Francisco
2020 International Astronautical Congress, 71st IAF · ESA
2020 SpaceCom 2020, Enabling Commercial Space · Colorado Springs
2019 Quantum Information Processing with Superconducting Circuits
2019 Materials Frontiers to Empower Quantum Computing
2019 Quantum Technology: The Second Revolution
2018 FutureHack · Tokyo
2018 American School of Japan · Tokyo
2018 International School of Science · Tokyo
2018 Future of the Global Energy System, Institute for the Future · San Francisco
2016 Keizai · US-Japan Commercial Spaceflight · San Francisco
2016 Effective Altruism Summit · San Francisco
2016 Hive Global Leadership Forum · San Francisco
2016 RSA Information Security · San Francisco
2015 Hive Global Leadership Forum · San Francisco
2015 Further Future · TED Meets Burning Man · Las Vegas
2015 Hive Global Leadership Forum · San Francisco
2015 DefCon Information Security · Las Vegas
2015 Black Hat Information Security · Las Vegas
2014 The Future of Commercial Spaceflight · Silicon Valley Space Center
2014 Yuri’s Night: The First Manned Orbital Spaceflight · Los Angeles
2014 IEEE Quantum Photonics: The Next Frontier of Quantum Communications
2014 Yuri’s Night: The First Manned Orbital Spaceflight · Hawaiʻi
2012 NASA ESA JAXA Pacific International Space Center for Exploration Systems
2012 NASA CSF Next-Generation Suborbital Researchers Conference · Palo Alto
2012 Quantum Information and Nanoscale Optoelectronics · Berkeley
2012 Yuri’s Night: The First Manned Orbital Spaceflight · Los Angeles
2012 Inaugural Quantum Future Technologies Conference · NASA Ames
2011 Quantum Coherence in Excitation Energy Transfer · Berkeley
2011 The Future of Spaceflight · Mobile Monday, Invited Keynote · Amsterdam
2011 Delft-Leiden Biannual Casimir Symposium · Leiden
2011 Alain Aspect: The Second Quantum Revolution · Leiden
2011 ESA-TNO Space Pier Day · The Hague
2010 Kavli-Delft Center for Bionanoscience, Founding Conference · Delft
2010 Quantum Mechanics in Higher-Dimensional Hilbert Spaces · Austria
2010 What is Real in the Quantum World? Int’l Akademie Traunkirchen · Austria
2010 NASA ESA JAXA Pacific International Space Center for Exploration Systems
2009 NASA ESA JAXA Japan-US Science, Technology and Space Applications Program
2009 From Foundations of Quantum Mechanics to Quantum Information · Delft
2009 DEISA Distributed European Infrastructure for Supercomputing Applications
2009 Partnership for Advanced Computing in Europe (PRACE) · Amsterdam
2008 Quantum Decoherence and Quantum Information Science · Lorentz Center
2008 Triennial Conference on Low-Temperature Condensed Matter Physics XXV
2008 International Conference on Quantum Structures · Brussels
2007 Workshop on Time Symmetry in Quantum Mechanics · Brussels
2007 Optical Fabrication Technologies, Coherence and Metrology · Switzerland
2006 The Best of Nanoscience: International Symposium for Hans Mooij · Delft
2006 SPIE Defense and Security Applications of Quantum Information Science
2005 New Computational Paradigms: Neural Nets, Quantum, Biocomputing
2005 UNESCO Physics for Tomorrow, UNESCO Headquarters · Paris
2004 RSA Information Security · Barcelona
2004 SPIE Defense and Security Applications of Quantum Information Science
2003 Gordon Research Conference on Quantum Information
2003 Quantum Information Technology IX · Tokyo
2003 International Conference on Quantum Information · Tokyo
2002 NATO Advanced Research Workshop on Quantum Chaos · Lake Como
2002 National Science Foundation Coding Theory and Quantum Computing · Vienna
2002 United Nations International Student Conference · Amsterdam
2002 International Conference on High-Energy Physics XXXI · Amsterdam
2001 World Technology Summit · London
2001 French Senate Hearing on the Future of Artificial Intelligence · Paris
2001 US Government Conference on High Performance Computing · Salishan
2001 National Security Agency · Fort Meade
Black belt, First Class, shōdan. Certificate of recognition, as first foreigner to qualify in eight years. 認許する, Japanese traditional archery, Kyūdō, “standingZen,” 弓道初 in formal recognition awarded by the Japanese National Kyūdō Federation, 全日本弓道連盟 while on Japanese National Fulbright Award with the Association of International Education, Japan.
20240512
Starlab: Deep Future The 'Noah’s Ark' of scientific research that launched 1,000 startup
ideas
Lab-concocted vodka, time travel and epilepsy treatments:
Welcome to the Moonshot Factory
“ a place where 100 years means nothing … ”
Financial Times | Sifted | 08 08 2022— What happens when you round up more than one hundred of some of the
world's greatest scientists, maverick geniuses working on some of
the world’s most groundbreaking ideas, put them together in a
Belgian castle, and let their imaginations run wild?
Fire extinguisher duels, bootleg vodka made with lab-procured ethanol
and worldbeating treatments for epilepsy are just some of what went
down at Starlab: a one-of-a-kind experiment created to unite
some of the world’s most daring technologists.
When it was founded in 1996, Starlab was compared to other top
research institutes — like Xerox’s Palo Alto Research Center — that
successfully bridged the gap between idea and market. It was also a
prototype for the ambitious organizations of today like Google’s
“moonshot factory,” X, trying to bring entirely new ideas to
the world.
But the centre’s idealism was to be its downfall; its pie-in-the-sky
approach couldn’t pay the bills, and it went dramatically bankrupt
during the dotcom crash. But what most people don’t know is that
Starlab’s legacy lives on in the picturesque hills overlooking
Barcelona and elsewhere.
Many European VCs and universities claim they’re backing innovations
that will solve humanity’s problems, but huge successes have been
elusive. One of the companies from Starlab’s second generation has
found significant success, but the centre's tale forces anyone
interested in innovation to ask themselves: how do we really bring the
wildest ideas to life — and make them financially viable?
The Noah’s Ark of science
Christopher Altman
Starlab was established by serial entrepreneur Walter De
Brouwer together with MIT Media Lab founder Nicholas Negroponte and
European VC pioneer Johan Konings. The idea was to create a utopian
“Noah’s Ark” of science, where the brightest minds from different
fields would be brought together to work on “deep future”
research. “De Brouwer’s ambition was to bring the best scientists in the
world together to ‘think thoughts for the very first time.’ It
was very interdisciplinary — no walls, no boundaries, no borders …”
says Christopher Altman — astronaut, quantum physicist and
Starlab veteran.
In its heyday, Starlab was home to more than 130 scientists
from 36 countries, who worked on ideas ranging from time travel and
consciousness to new media and “intelligent” clothing. The majority
lived on site: a neoclassical castle designed in the late 1800s on the
outskirts of Brussels.
Starlab “Time Travel Party,” May 2001. (LtoR):
Hugo de Garis, Serguei Krasnikov, Roman Zapatrin, Christopher Altman
“It was like a pirate ship in a way, which is what I think I fell in
love with. Or you could call it a kind of sect,” laughs Giulio Rufini,
neuroscientist and current CEO at Starlab. “We’d stay up all night talking in-depth theoretical implications of
closed timelike curves (time travel). Roman had a centuries-old recipe for homemade vodka and put to use
some surplus ethanol he reappropriated from the biophysics lab down in
the basement,” says Altman, referring to one of his colleagues, a
quantum topologist and mathematician. “One time a few of the researchers covered themselves in yards of
aluminium foil as “armor” and started a duel, complete
with fire extinguishers as weapons, in the courtyard.”
I began my scientific career at a multidisciplinary research institute, Starlab, located deep in the serene and secluded forests outside Brussels, Belgium.The lab’s principal base of operations was housed in a historic landmark — an imposing 19th century manor, remarkable both in scale and magnificence. In a previous incarnation, the palatial grounds served as official embassy for the First Republic of Czechoslovakia. Its nearest neighbor, the world renowned Pastéur Institute, was one of but a handful of highly-secured Biosafety Level 4 labs in the world.
Cofounded by MIT Media Lab founder Nicholas Negroponte and serial entrepreneur Walter de Brouwer and established in partnership with MIT, Oxford and Ghent University, Starlab was created as a “Noah's Ark” to bring together the world's most brilliant and creative scientists to work on far-ranging multidisciplinary projects that hold the potential to convey a profound and positive impact on future generations.
Starlab was borne as an incubator for long-term and basic research in the spirit of Bell Labs, MIT Media Lab, Xerox PARC, and Interval Research. Its research mantras were “Deep Future” and “A place where one hundred years means nothing.” Approximately 130 scientists from thirty-seven different nationalities — each established leaders in their respective research fields — lived and worked at the lab.
A second base of operations, Starlab DF-II (Deep Future II)was established in the Royal Observatory in Spain on a mountaintop perch overlooking the city of Barcelona. With a more tightly-focused mission scope of space-borne and neuroscience research, DF-II continues to innovate and grow to present day.
Onsite research ranged from artificial intelligence, biophysics, consciousness, emotics, intelligent clothing, materials science, protein folding, neuroscience, new media, nanoelectronics, quantum computation, macroscopic entanglement, robotics, stem cell research, theoretical physics — e.g., the possibility of time travel — transarchitecture, and wearable computing.
Our custom-built supercomputer, the CAM-Brain Machine, was supported in part by a 1 Million Euro grant from the European Union. The custom-designed and created supercomputer — as powerful as 10,000 Pentium II PCs — harnessed the power of Xilinx field programmable gate array (FPGA) evolutionary hardware to evolve seventy-five million neurons in a massively-parallel artificial neural network instantiated directly in silico using evolutionary genetic algorithms. With each clock tick, the supercomputer simultaneously updates hundreds of millions of cellular automata billions of times per second. The one-of-a-kind machine was recognized by the 2001 Guinness Book of World Records as the “World’s Most Complex Artificial Brain.” In testimony to the French Senate and the United Nations, I stated “Future networks will not be built, they will be grown.”
When the laboratory came up short on research grants in June, I personally went to the President himself when serendipity brought us together at the same time and place on his first trip overseas after election. The Commander in Chief, who had just arrived in Brussels for a meeting with NATO, impressed upon us both his immediate familiarity with the lab and its work, and his enthusiastic approval in response to my earnest request for $1M in budget — funds that had been allocated for national security priority scientific research topics through a grant newly created by Clinton with his last act in office: the 2001 National Nanotechnology Initiative.
For my contributions to the field, I was honored to be selected by the US Government as one of three graduate students most likely to impact the future of the field at Salishan — a distinction shared with John Carmack and Bill Butera — was sponsored to attend conferences and senior administrator briefings at Fort Meade, National Security Agency headquarters outside Washington, DC, attended the World Technology Summit in London, was an invited delegate to the French Sénat to provide testimony on the future of technology and how it will transform our lives over coming decades, and more.
Following three days of enraptured debate with senior politicians, senators and international diplomats at the French Sénat hearing on artificial intelligence in Paris, the world's first senate hearing on the topic, Starlab's principal investigator and AI program lead, Hugo de Garis, forecast I may one day even be elected President myself. Far sooner than that, however, he drew our attention to the potential threat of assassination from extremists or technology Luddites who may rise to stand in opposition to the rapid pace of progress in technology and artificial intelligence.
Our living arrangements at the lab consisted of an expansive three-bedroom master suite with fully-stocked library, typically reserved for visiting prime ministers, senators, and senior diplomats. My quarters were shared with none other than the project’s principal scientific investigator himself. One midsummer’s afternoon as the two of us strolled on a random walk through the sprawling estate and lush wooded grounds surrounding the manor, immersed in a passionate debate on the long-term promise and perils of superintelligence, the ever-eccentric de Garis came up with a radical idea: to obtain a life-size replica of Fat Man — the solid plutonium core, 21 kiloton, 10,300-pound nuclear bomb detonated over Nagasaki in World War II — and to mount it precariously to the vaulted ceilings of my apartment, with the bomb hangingdirectly over my bed.
The Volkswagen beetle-sized replica of the bomb was constructed for a Discovery Channel documentary de Garis had just finished filming on the future of artificial intelligence, which featured a potential global war between humanity and artificial intelligence. He'd purchased it outright from the director, making arrangements for expedited shipping and delivery direct to Starlab's headquarters in Brussels. The sheer audacity of the proposal was outrageous. He intended the bomb to serve as a dramatic and powerful reminder of “the weight of my responsibility to the future of humanity.” He certainly knew how to drive home a point.
OpenAI CEO Sam Altman,xAI founder Elon Musk, DeepMind CEO Demis Hassabis, and Anthropic CEO Dario Amodei have each warned that advanced AI could pose an extinction-level risk comparable to nuclear war.
That question has shaped my work for decades. To help reduce that risk, I’ve developed and patented a scientific framework for detecting structural signatures of self-preservation and terminal continuation interest. Rather than relying on behavior alone, the approach shifts analysis and prediction to latent structure itself. Its core thesis is that agents with terminal continuation objectives exhibit higher entanglement entropy than agents with merely instrumental continuation interest, yielding a consistent, measurable signature in the system’s latent trajectory structure.
With the explosive rise of AI we've seen over the last twenty-four months alone, with artificial general intelligence (AGI) and artificial superintelligence (ASI) seemingly just around the corner, one could say that de Garis — though radical and exceedingly unconventional in his unprecedented approach — was just a few years ahead of his time.
A recentFinancial Times (FT) spinoff magazine article in Sifted highlights our research going back to Starlab, AI and time travel research projects, and my subsequent travels across East Asia to create national quantum roadmaps for US national research funding and IC agency directors. In years that followed, I continued on through research fellowships in nanoscience and the foundations of quantum mechanics with Nobel physics laureate Anton Zeilinger’s research group in Austria and across Europe, then was recruited to lead a futures initiative at NASA in collaboration with Ray Kurzweil and Google, together with leading companies, luminary scientists, venture capitalists and entrepreneurs from Silicon Valley and around the world.
From manned spaceflight training at NASA on to the summit of a volcano where the Apollo 11 astronauts trained before their first landing to put a man on the Moon, to the development and deployment of a fundamentally new quantum communication system based on continuous-variable quantum teleportation to provide space-based NASA assets with unconditional information security, from collaborations leading diplomats to advise the United Nations on critical security issues of the future, to field expeditions employing state-of-the-art sensors in rough desert terrain with a multidisciplinary team of scientists, special forces operators, and agency directors beta-testing next-generation technologies in austere environments — each of these initiatives was undertaken with the singular aim to make a profound and positive impact on the future of humanity, for our children, our children’s children, and the generations yet to come.
.jpg)
