Writing
field notes · research · what came out of the workNotes from work in progress and the longer papers behind them: AI agents, distributed systems, and the things that break.
Novlatest guideNovember 2026Nº 06
From Zero to Serving a Model on Kubernetes, on Your Own Machine
A tested, hands-on guide to running an LLM on Kubernetes on the laptop or desktop you already have: kind, llmfit, Ollama, GPUs and k8sgpt.
field notes · dispatch ledger6 entries
05Testing Celln's BoundaryFour boundary probes and one successful computation: what a small Celln demonstration shows, and what it leaves untested.Aug 2026 · note
04The Agent Environment Should Be a LeaseWhy I built Celln to lend agents a defined capability set, and what that execution model still needs to prove.Aug 2026 · note
03Post-Kubernetes Infrastructure for GenAI WorkloadsField notes on Modal's million-sandbox announcement, what it says about Kubernetes' assumptions, and the coming decoupling of coordination from execution.Jul 2026 · note
02The Sticky-Note Problem: Making Agent Handoffs ReliableWhy agent handoffs lose context, and how shared evidence and explicit ownership can make coordination more reliable.May 2026 · note
01Agents Need Somewhere to Share Their WorkWhat makes agents useful as a team, and how a shared, policy-controlled workspace might help.Apr 2026 · note
research papers · revision ledger1 current
v2.2 The Synthetic Membrane: A Coordination Layer for Multi-Agent AI SystemsRevised position paper defining a coordination contract for shared evidence, selective access, and action ownership, with explicit limits and an evaluation plan.September 2026 · paper