Research

Empirical findings on AI workflow effectiveness.

Start here: StayFresh Thesis for the site's core doctrine and the fastest path through the archive.

2026-04 What is Prompting: Operational Constraints
Prompting is a constrained pipeline where tokenization, max tokens, and control tokens form the operational boundaries that determine agent behavior.
2026-04 Context Is a Budget
More tokens do not create more truth. Context should be spent where it changes the next decision.
2026-04 Protocol Before Personality
Shared artifacts and exit conditions beat charisma, cosplay, and fake seniority.
2026-04 Specs as Shared Reality
Agents do not share understanding. They share binding artifacts, versions, and tests.
2026-03 Reward Engineering for Coding Agents
Why coding agents optimize the rubric more than the prompt.
2026-03 Reward Hacking in Coding Agents
How poorly designed metrics produce plausible but unstable code.
2026-02 AGENTS.md Effectiveness
Evaluating repository-level context files for coding agents.
2026-02 Formal Verification with Agents
Property-based testing and specification generation.
2026-02 Enterprise Agent Design
BCG framework for production-grade agents.
2026-02 Persona Anchors
Character-based instruction patterns for consistent agent behavior.