Magistr and Maps
Thoughts
About Me
Search
essays
Deep thoughts, notes, reposts and thinking aloud bout IT, architecture and maps.
21 posts
latest 2026-07-28
21
The ledger under the cage
A reply to twonines' 'What the Factory Forgets': the record_prior_art harvest loop is the memory that sits inside my agent factory, so ephemeral caged agents still arrive warm.
2026-07-28
1,083 words
ai · agents · issue-lifecycle
20
The loop that built itself
How issue-lifecycle grew from a borrowed state machine into swamp-go-brr: the harvest/prior-art loop, the changes along the way, and the turn from human-in-the-loop to autonomous.
2026-07-28
1,728 words
ai · agents · issue-lifecycle
19
Reinventing the wheel?
How Modal reinvented Kubernetes by stripping away Docker images, and what it means for the future of GPU serverless platforms and software infrastructure.
2026-07-22
220 words
devops · serverless
18
Building a Health Data Pipeline with Google Health API, Swamp, and Claude
How we built a self-hosted health pipeline with the Google Health API, Swamp extension models, and Claude — from a Xiaomi login dead end, through probing a closed API and reverse-engineering Fitbit's readiness score, to ECG waveforms and a twice-daily Telegram briefing.
2026-07-12
1,997 words
health · quantified-self · google-health
17
Reading the board from the stands
I ignored Nvidia for two years. Then two interviews flipped me — on Jensen as the best Wardley strategist alive, and on what I actually wanted while watching him.
2026-07-09
1,875 words
wardley mapping · ai · nvidia
16
Nvidia, Oxide, and tinygrad through a Wardley lens — the long version
The long version: three AI-compute value chains mapped and ranked — Jensen the strategist, Oxide's durable niche, tinygrad's right target and wrong method — plus the bezzle counter-thesis and a live co-design exhibit.
2026-07-09
6,548 words
wardley mapping · ai · nvidia
15
swamp-go-brr, the brain
Why the autonomous agent loop inside the Firecracker cage is a deterministic state machine and not a swarm: how gobrr decomposes, schedules, and gates the work, plus what held and what bit me building it.
2026-06-19
1,759 words
swamp · agents · firecracker
14
Build the cage first
Why I built a Firecracker microVM extension before the autonomous agent loop that needs it: the lethal trifecta, the painful microVM build, and a loop that shipped a real swamp extension end to end.
2026-06-15
2,105 words
swamp · firecracker · agents
13
The shape of the decision
Why the ABN affordability number isn't the real number for an expat in the Netherlands — and how I turned a fuzzy apartment-and-mortgage decision into a plan I can actually steer, with good-planning and Bovolon's four layers.
2026-06-13
1,022 words
swamp · good-planning · finance
12
A feed to think with
How I turned a noisy Amsterdam apartment feed into a clean, high-signal one I can reason with — good-planning underneath, a move.nl dossier on top, public-data enrichment, and my requirements encoded as filters.
2026-06-13
1,083 words
swamp · good-planning · moldable-development
11
JSCAD CAD Generation with Swamp and Claude Code: Setup Guide
Step-by-step guide: install Swamp from scratch, set up @magistr/jscad-cad extensions, configure the Claude Code skill, and generate 3D models from text descriptions.
2026-04-04
421 words
jscad · cad · swamp
10
Building a CAD Pipeline with JSCAD, Swamp, and Claude
How we built a CAD generation pipeline using JSCAD, Swamp extension models, and Claude — from choosing the right CAD system, through 7 iterations of debugging a genie lamp with 6-view projections, to publishing extensions and finding a bundle cache bug.
2026-04-04
1,809 words
jscad · cad · swamp
All 21 posts →