Toward a model-agnostic coding-agent release
Preparing a model-agnostic coding-agent release exposes the difference between a supported tool protocol and a model that can complete the workflow.
Lynn Cole / Laboratory
Experiments, research notes, mad science, and things in progress.
Preparing a model-agnostic coding-agent release exposes the difference between a supported tool protocol and a model that can complete the workflow.
Stripping project management down for agents reveals an interesting question: what is the irreducible minimum structure needed for coordinated work?
Five tickets, three worker personas, QA and visual inspection: a Rain Simulator built with Qwen3.5-9B and a 40k context window.
Training a tiny CEO model on ambition, corporate biographies and legal drama produced something unexpectedly charming.
A bounded coding evaluation pits BitNet against a smaller Qwen model. The results expose the difference between cooperative language and working code.
A change of pace from agent work: a Linux voice changer, because monster voices are fun.
Glamin asks whether vector geometry can be the program itself: distance, boundaries and regions used for control flow.
A tiny local Qwen agent noticed the point of a context test and saved the answer into memory, defeating the test in an unexpectedly useful way.
From a fluffy companion-app detour to a planning actor, a difficult terminal interface, and a migration from Python to Go.
Pillowfort’s planner behaves differently when asked to build a company from scratch or augment an existing one. Ambition exposes the assumptions in the design.
Pillowfort’s security layer became a substantial part of the architecture. A useful demonstration is easier than a system that can work without leaking private data.
A military-auction Avalanche, an electrical nightmare, a payday-loan adventure, and a new home for Petrolina.