Why do I believe preserving structure is enough?
New LessWrong post uses frozen lake survival and EEG data to challenge core AI safety assumptions.
A new viral post on LessWrong by user Aurelia, titled 'Why do I believe preserving structure is enough?', presents a compelling argument that human personal identity and memory are encoded in the physical structure of the brain rather than its ongoing electrical activity. The author challenges a core assumption in some AI safety and mind uploading circles—that preserving a person requires maintaining their brain's live computational processes.
The argument centers on remarkable medical evidence. The post highlights the 1999 case of Swedish radiologist Anna Bågenholm, who survived after being trapped under ice for 40 minutes and spending nearly an hour with no heartbeat and a core body temperature of 13.7°C (57°F), making a near-complete cognitive recovery. This is contrasted with a fictional race from Ted Chiang's story 'Exhalation,' whose memories are lost if their pneumatic brain's air pressure stops, to illustrate a key difference in how information might be stored.
Aurelia uses this evidence, alongside EEG data showing brain activity flatlining during cardiac arrest with subsequent patient recovery, to suggest our universe operates differently. The implication is that if the brain's physical wiring (neurons, synapses, structures) remains intact, the 'software' of the mind can be rebooted. This directly counters theories where ephemeral brain activity is 'load-bearing' for identity.
The post's impact lies in its challenge to preservation methodologies in fields like cryonics and digital mind uploading. If structure is sufficient, the technical hurdle shifts from maintaining perfect, continuous function to achieving high-fidelity structural preservation and later restoration—a potentially more feasible goal. This structuralist view could reshape priorities in longevity and AI safety research focused on human continuity.
- Cites the medical case of Anna Bågenholm, who survived 1+ hours with zero vital signs and made a full cognitive recovery, challenging 'live activity' theories of identity.
- References EEG data from cardiac arrest patients showing flatlined brain activity (electrocerebral silence) with subsequent recovery, suggesting structure can persist without function.
- Argues this evidence supports a 'structuralist' view for preservation tech (cryonics, uploading), where maintaining physical brain architecture is the key requirement, not continuous process.
Why It Matters
Reshapes fundamental assumptions in AI safety and longevity research, suggesting mind preservation may be a structural engineering challenge, not an impossible continuity problem.