Egg Syntax (legal name Jesse Davis) is an independent AI safety and alignment researcher based in Asheville, North Carolina. After 15 years as a professional software developer specializing in Clojure, they transitioned in 2023 to full-time technical research aimed at decreasing existential risk from advanced AI. Their work spans mechanistic interpretability, LLM reasoning capabilities and limitations, and a novel research agenda investigating whether frontier LLMs develop something functionally analogous to a persistent self or character. They are funded by Coefficient Giving and the Long-Term Future Fund (LTFF) and are active contributors to LessWrong and the AI Alignment Forum.
Funding Details
- Annual Budget
- -
- Monthly Burn Rate
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
- Fiscal Sponsor
- -
Theory of Change
Egg Syntax believes that AI companies may successfully build ASI but that no one currently knows how to do so without an unacceptable level of existential risk. Their research focuses on understanding the internal structure of frontier LLMs—particularly whether they develop persistent functional selves, and the limits and nature of their reasoning capabilities—with the aim of generating technical knowledge that could inform how to build or constrain AI systems more safely. Better mechanistic understanding of what LLMs are and how they work internally is treated as a prerequisite for either solving alignment or making a credible case for pausing dangerous capability development.
Grants Received
from Long-Term Future Fund
Projects
No linked projects.
People
No linked people.
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.
Details
- Last Updated
- Mar 21, 2026, 7:48 PM UTC
- Created
- Mar 19, 2026, 10:42 PM UTC