Gil Salu

Founder of Cyaxios and the person behind TN. I build the protocol, write the docs, and think out loud about why AI adoption stalls and what makes automated work trustworthy.

TN came out of a recurring problem in applied AI work: agents need real data access and a record of what they did, while privacy and governance teams need evidence they can actually review. TN is the answer I wanted to exist. Every log entry is signed at the source, sealed per reader, and chained so it can be verified later. You hold the keys.

Cyaxios is the firm behind it. The same thinking that shapes the protocol shows up in how I write about knowledge, organizations, and the gap between what teams say about AI and what they actually do.

Team

• Paige Grimes
  PhD student, Computer Science, University of Colorado at Colorado Springs
• Marios Petrov
  PhD student, Computer Science, University of Colorado at Colorado Springs

Writing

Essays on AI adoption, organizational knowledge, and the systems that make automation worth trusting. The full archive is on Medium.

• Where Institutional Knowledge Lives June 2026
• Conway's Law vs. the Agent April 2026
• Transformers v5 Will Kill the Last Excuse for Renting Your Intelligence Layer December 2025
• The Nobel Wake-Up: AI, Workslop, and the Myth of Productive Output November 2025
• Baking, Game Design, and Knowledge: In Praise of Soggy Bottoms and Other Necessary Failures November 2025
• Embrace the Human Spark in AI: What Japanese Masters Knew in 1995 About Knowledge Creation November 2025
• Why AI Initiatives Feel Fake: The Gap Between Process and Practice October 2025
• CAP Theorem Wasn't a Suggestion October 2025