Independent researcher & engineer

Intelligence, emergence & strange loops.

I'm Hao Ke. I study how intelligent systems interact, organize, and change the environments they inhabit—and how those environments change them.

My work moves between artificial life, complex systems, and AI safety, grounded in a background in distributed systems engineering.

01 / Research

Questions turned into systems.

Ongoing research · 2026–present

Quine

A research system
that studies itself.

Quine began by realizing frozen LLMs as native POSIX processes, opening a space to investigate the emergence of coordination, communication protocols, and self-reproduction.

As its runtime, experiments, logs, instructions, and theories accumulated in the same repository, the research environment became a research object. Its structure shapes the agents doing the research; their discoveries reshape that structure.

Today, I use this reflexive laboratory to explore artificial life, complex systems, AI safety, and the computational foundations of discovery.

01

Emergent coordination

Cooperation through shared environments, even without explicit knowledge of other agents.

02

Self-reproduction

How inherited material and future demands shape what an agent becomes across a process boundary.

03

Reflexive governance

How self-maintaining structures form, persist, and sometimes become detached from their purpose.

Cellular automata · Formal methods

Langton’s Ant

From one black cell to a highway.

A Lean-checked proof that every classical Langton’s Ant state with exactly one black cell, at any ant position and heading, eventually reaches the translating period-104 highway.

A result for an infinite family of initial states within the broader, still-open highway conjecture.

Astrodynamics · Megastructures

The Dyson Continuum

Reconsidering the shape of stellar infrastructure.

An exploration of the physical design space connecting Dyson architectures, with analytical and numerical models of solar-sail-supported, displaced orbits.

From a broad architectural idea to quantitative support conditions and low-latitude feasibility screening.

02 / Selected publications

Papers, proofs & proceedings.

Google Scholar ↗
  1. 2026

    ALIFE 2026 · Late-Breaking Abstract · Accepted

    A Tumor in the Repository: Function-Decoupled Self-Maintenance in an LLM-Agent Substrate ↗

    Hao Ke

  2. 2026

    ALIFE 2026 · Late-Breaking Abstract · Accepted

    Structural Elicitation in a Frozen POSIX Agent: Environmental Affordance Fields for Artificial Agency ↗

    Hao Ke, Jingyun Wu

  3. 2026

    ALIFE 2026 · Late-Breaking Abstract · Accepted

    Facultative Self-Reproduction in Quine, a Process-Native LLM Medium ↗

    Hao Ke, Jingyun Wu

  4. 2026

    ALIFE 2026 · Workshop contribution · Accepted

    From Simulated Worlds to Process Habitats: POSIX as a Substrate for Generative Artificial Life ↗

    Hao Ke, Jingyun Wu

  5. 2026

    ALIFE 2026 · Conference paper

    Coordination Under Existential Unawareness: Information Ablation and Closure Thresholds in LLM Multi-Agent Systems ↗

    Hao Ke, Jingyun Wu

  6. 2026

    arXiv · Systems preprint

    Quine: Realizing LLM Agents as Native POSIX Processes ↗

    Hao Ke

  7. 2026

    Zenodo · Preprint & formal proof

    Every One-Black-Cell Langton Ant Reaches the Period-104 Highway: A Lean-Checked Proof ↗

    Hao Ke

  8. 2026

    Zenodo · Research preprint

    From Keplerian Swarms to Radiatively Supported Bubbles: A Continuum Framework for Dyson Architectures ↗

    Hao Ke

Also accepted at GSO 2026, a satellite of the Conference on Complex Systems: Computational Neoplasms and Structural Elicitation (two non-archival workshop abstracts).

03 / Writing

Notes from inside the loop.

All writing ↗

04 / Background

An engineer’s instinct.
A researcher’s curiosity.

I came to independent research through building systems. My engineering work spans distributed infrastructure, developer platforms, and AI adoption; my research follows questions that emerge when those systems begin to act, adapt, and reorganize.

Professional profile ↗

2023–present

Contentsquare

FDE Engineer

Advancing internal AI adoption through the Cloud Agent project. Unifying developer infrastructure and delivery practices across merged engineering organizations.

2022–2023

Heap

DevOps Engineer

Production reliability, Kubernetes infrastructure, and deployment systems through Auryc’s integration into Heap.

2017–2019 · 2021–2022

Auryc

Infrastructure Engineer · Beijing & United States

Distributed infrastructure for analytics: cloud-native systems, data platforms, scalability, and observability.

2013–2017

Tsinghua University

B.S. in Software Engineering