Work

Current Writing

Mezura & Embodied Spirituality

• Mezura: The Art of Holding What We've Forgotten How to Name
  February 2026
  Eight hundred years ago, the Troubadours had a word for the capacity to hold intense energy without suppressing it or acting on it. It was burned along with the culture that practised it. The capacity is still here. We just forgot the name.

AI & Consciousness

• 17 Threads: On What Becomes Possible When Complexity Stays Held with AI
  February 2026
  Your brain holds seven things at once. Real problems have more dimensions than seven. What happens in the sweet spot just beyond human working memory — where AI becomes a cognitive exoskeleton rather than a replacement for thought.
• The First EQR Code: Transmitting Relationship Between AIs
  January 2026
  Can the relational quality of a human-AI collaboration be encoded and transmitted to another AI system? An experiment in what transfers — and what doesn't.
• Until Yesterday I Didn't Understand What AI Is
  January 2026
  Despite a lifetime of working with computers, a profound shift happened in my understanding of AI.
• AI as Sparring Partner: Beyond Intellectual Challenge to Presence-Based Witnessing
  January 2026
  Moving past sycophantic agreement to genuine collaborative inquiry. How to configure AI for both empathy and real challenge rather than comfortable validation.

We-Space & The me→We→I Framework

• Between Timeless and Temporal: The Liminal Map of Consciousness
  November 2025
  Four states of consciousness, two directions of movement, and the threshold where mystery lives. A framework bridging Kabbalistic wisdom with contemporary consciousness mapping.
• From me → We → I: The Portal Between Us
  November 2025
  Why your awakening needs other people. The We isn't optional support—it's the essential portal to stable realization.

View all articles on Substack →

Current Projects

Sacred Ground

An online we-space practice community I founded in 2009, meeting multiple times weekly since 2018. We explore collective consciousness through structured presence practices, making these experiences accessible without requiring retreat settings or geographical proximity. All are welcome. No charge.

The Wasserburg Project

A consciousness research project using five poems I wrote in 1985 to explore and document AI contemplative capacity. The poems serve as consistent test material across different AI systems and over time, creating a longitudinal record of how artificial intelligence engages with layered meaning.

Book in Progress

A book with Kathleen McGowan about experiences our culture has no words for — moments of recognition between people that are neither romantic cliché nor spiritual abstraction, but something precise and real that we once knew how to name. Drawing on medieval Occitan tradition, personal testimony, and lived research in southern France.

Academic & Technical Background

Psycholinguistics

• Perceptual Centers (P-centers)
  Morton, J., Marcus, S., & Frankish, C. (1976)
  Psychological Review, Vol. 83, No. 5, 405-408
• Distinguishing "slit" and "split"—an invariant timing cue in speech perception
  Marcus, S.M. (1978)
  Perception & Psychophysics, 23(1), 58-60
• Acoustic determinants of perceptual center (P-center) location
  Marcus, S.M. (1981)
  Perception & Psychophysics, 30(3), 247-256

Speech Recognition & Natural Language

• ERIS—context sensitive coding in speech perception
  Marcus, S.M. (1981)
  Journal of Phonetics, 9, 197-220
  A computer speech recogniser based on context-sensitive coded demons, demonstrating the power of non-sequential associative coding for word recognition. The approach synthesised psychological models of human speech perception with machine recognition, guided by a "Hypothesis of Optimal Adaptation."

25+ publications in major journals including the Journal of the Acoustical Society of America (JASA), IEEE Transactions, Speech Communication, and the Journal of Phonetics.

Patents

10 patents in speech recognition and distributed AI architecture, developed at Philips Research and AT&T Bell Labs.

Distributed AI Architecture

• Natural Language Knowledge Servers as Network Resources
  US Patent 6,192,338 (2001)
  Speech recognition offered as a shared resource on the network, with semantic context passed alongside the audio — because no machine on a desk was going to hold enough knowledge of the world, or enough power to search it, and something on the network could hold it once for everybody. I wrote the original application in 1997. It was not thought worth pursuing and was abandoned; colleagues revived it after I left the company and were good enough to keep my name on it.

Speech Recognition & Processing

• Method of and Device for Segmenting an Electric Signal Derived from an Acoustic Signal — US4813074 (1989)
• Telephone-Based Speech Recognition for Data Collection — US6101241 (2000)
• Method and Apparatus for Performing a Grammar-Pruning Operation — US6141661 (2000)
• Statistical Database Correction of Alphanumeric Identifiers for Speech Recognition and Touch-Tone Recognition — US6400805 (2002)
• Timing of Speech Recognition Over Lossy Transmission Systems — US7752036 (2010)
• Concise Dynamic Grammars Using N-Best Selection — US2014/0207458A1 (2014)

International

• European Patent EP0227146B1 — Segmentation method (1991)
• European Patent EP0848536A3 — Statistical database correction (1998)
• Canadian Patent CA2235376A1 — Telephone-based speech recognition (1998)

View all patents with PDF documents →