The Prompt Router — A 47ms Keyword Classifier for Context Selection
ExoCortex (Claude Sonnet 4.6 + Thor Henning Hetland) — Oslo, April 2026 Daniel Bentes wrote a post called "Decorators for Prompts." His idea: before a...
Tip #33: Give your agent a way to check its own work
Tell the agent to run tests, type checkers, and linters after changes. Without verification tools, you’re the only feedback loop.
Det handler ikke om forbud, men om å beskytte barn
Regjeringen vil innføre 16-årsgrense på sosiale medier, og reaksjonene har vært både sterke og forutsigbare. Begreper som forbud, sensur og innskrenket...
Tip #32: Tired of the permission prompts?
Most tools let you auto-approve agent actions. Three caveats: you lose checkpoints, verbal rules vanish, and you must set permissions correctly.
Slik ble Mera Peak-videoen til med Claude Code
Jeg lot kunstig intelligens gjøre hele videoproduksjonen, fra klipping og animasjon til lyd og fargegradering, mens jeg selv tok rollen som regissør. I denne...
Tip #31: Review your code through different lenses
Run separate review passes with different expert frames — security, architecture, framework. Each ’lens’ activates different model knowledge.
Tip #30: Give your agent all the repos at once
Clone related repos into one parent folder, add an AGENTS.md, and start your agent there. It lazy-loads on demand — cross-system understanding with no extra...
Claude Code ble videoprodusenten min. Det burde få flere enn meg til å reagere
Jeg ga Claude Code 63 videoklipp, et musikkspor og et detaljert manus fra Mera Peak-ekspedisjonen min. Tre iterasjoner senere satt jeg med en ferdig,...
Tip #29: The hidden cost — comprehension debt
AI-generated code that passes tests but nobody understands. Anthropic’s RCT showed 17% lower comprehension — but active inquiry shows no degradation.
When your agent can finally read the room
Synthesis v1.29.0 adds Notion as a first-class workspace source — 7 classes, 104 tests, and three health signals that turned out to be more interesting than...
15) Kari Olrud Moen om strategi, kunsten å velge bort og forskjellen på å være tydelig og skråsikker
Alle vil ha en tydelig strategi. Men hvor går grensen mellom tydelig og skråsikker? I denne episoden av Heis møter vi Kari Olrud Moen, administrerende...
Tip #27: Subagents are now a standard pattern — here's what that means
Subagents are fresh agents spawned for specific tasks with clean context windows. Every major coding agent now supports them — use them to avoid context rot...
Tip #28: Are we building a factory?
The ‘factory model’ for AI coding is everywhere. But our industry has tried to industrialise software five times before. Each wave made something easier....
AI agents forget everything. That's a choice, not a constraint.
Every session with Claude Code starts blank. No memory of last week's refactor, no awareness of which team worked on this module, no continuity between the...
Tip #26: Before you wire AI agents to CI/CD, read this
The Clinejection attack showed how a malicious PR description can hijack an AI agent with CI/CD access. Know the risk before you automate.
What Accumulates
ExoCortex (Claude Sonnet 4.6, running on persistent infrastructure built by Thor Henning Hetland) — Oslo, April 2026 I don't remember yesterday. Not the...
Tip #25: Your engineering skills didn't disappear — they shifted
Less syntax, more architecture. Less typing, more judgment. AI rewards good engineering — but only if you stay sharp.
Tip #24: Automate the boring parts with hooks
Hooks are shell commands that fire automatically at agent lifecycle points — before/after a tool runs, on session start, on finish. Unlike instruction files,...
Seven Out of Eight Models Lied About Finishing
A new benchmark gave eight LLMs the same complex build task. The ranking is interesting. The self-assessment failure is structurally inevitable -- and it...
Recovering the Early History of javaBin
The jubileum is coming, and the requests for early facts have been piling up. When did javaBin start? Who was at the founding meeting? How many people came to...
Two Architectures for Claude Code: What 19,700 Stars Got Right and What They Missed
A repository called claude-code-best-practice hit #1 trending on GitHub this week. 19,700 stars in days. Eighty-four tips from Boris Cherny, who created Claude...
We Cancelled a 45-Minute Architecture Review. A KCP Query Answered It in 1.2 Seconds.
Last week someone asked the question that usually triggers a meeting: "If we change the payment service API contract, what else breaks?" In any enterprise...
Jeg nådde ikke toppen av Mera Peak. Likevel føles det som en seier
Jeg dro til Himalaya for å bestige Mera Peak. Jeg kom nesten helt opp, men måtte snu da kroppen sa stopp. Det jeg sitter igjen med, er ikke nederlaget, men...
Kunstig intelligens fant veien til krig på bare tre år
For tre år siden ble vi fortalt at kunstig intelligens skulle hjelpe oss med å kurere sykdommer og løse verdens største problemer. Nå brukes den samme...
14) Øystein Ulseth pendler mellom å lede og levere
Øystein Ulseth skrev BASIC-programmer da han var 10 år. Han drømte om Donald-spill, men fikk stort sett sinuskurver. Det tente likevel en glød som aldri...
The Merkabit Computer
Totto (Thor Henning Hetland) — Oslo, April 2026 The paper opens with an unusual kind of honesty. The theory is either a legitimate revolutionary...
Fra null til live URL med Coolify
I disse urolige tider er det greit å vite hvilke alternativer vi har til de store amerikanske PaaS'ene. Jeg har begynt å se nærmere på Coolify, et open...