Anthropic built its own version control system in five months with seven people and migrated its monorepo off git. The interesting part is not Scion – it is the diagnosis underneath it: git's cost model is priced for an offline client that no longer exists in an agent-driven fleet.
Category: AIML
Spec Kit vs AWS AI-DLC
I ran the same expense-approval project through Spec Kit and AWS AI-DLC. One is a lean artifact pipeline; the other is a stateful process runtime. Here is where each approach helps, where it gets heavy and how the implementation differs.
AI Is Not the Plan
I keep hearing two versions of the AI story. In one, AI gives everyone a personal tutor, doctor, researcher, and software engineer. In the other, it takes our jobs, floods the internet with deception, and concentrates power in a handful of companies. Both versions miss the important part. AI is a capability. It is not … Continue reading AI Is Not the Plan
What Actually Runs When You Walk Away: KiroCrew’s Autonomy Modes, Compared
Once a KiroCrew session is running, there are four ways work can proceed without you driving every turn: Autopilot, the Task Runner, Monitor, and /goal. The docs present them as separate features, and I could not tell which one I was supposed to reach for — so I spent an afternoon working through them question … Continue reading What Actually Runs When You Walk Away: KiroCrew’s Autonomy Modes, Compared
Bots Club, Ten Years On: The Controls That Take It Into Production
In May 2016 I wrote a short post called Bots Club. One idea, one line: bots should and can work collaboratively as we humans do. Bot to bot, no human in the channel. I built three Hubot demos for my team — bots processing a ticket, bots handling an alert, bots watching cost — and … Continue reading Bots Club, Ten Years On: The Controls That Take It Into Production
