Now
Updated 13 April 2026. This page is a dated log of the work, constraints, and decisions that currently have my attention.
April 2026
Working close to APIs, integrations, and operational reliability so the product surface gets easier to trust as it grows.
The important question is not only whether an endpoint works. It is whether the system stays legible for partners and for the team maintaining it.
See the Tillo case studyApril 2026
Stress-testing where agents genuinely reduce friction across planning, implementation, and review, and where they simply move ambiguity downstream.
The useful gains are showing up in specification quality and first drafts. Verification, architecture, and taste still decide the outcome.
Browse the experiments logApril 2026
Turning better briefs, acceptance criteria, and review steps into repeatable habits rather than one-off good sessions.
I want clearer intent and tighter feedback loops to be part of the workflow before any model starts generating code.
Read The Velocity ParadoxI like having a place for directional updates that stays concrete: what I am building, what I am testing, and what seems worth carrying into the next round of work.
If you check back later and the focus has changed, that is the point. This page is meant to show movement rather than polish.