clojure-norway

boosja 2026-04-22T06:14:30.104949Z

God morgen!

eaj 2026-04-22T06:20:41.283519Z

God morgen!

hypirion 2026-04-22T06:41:17.665429Z

Morn!

mokr 2026-04-22T06:42:24.143119Z

Morn!

msolli 2026-04-22T06:46:54.775499Z

Morn!

emil0r 2026-04-22T07:24:50.102559Z

Morn

teodorlu 2026-04-22T07:36:37.601519Z

god morn 🙂

slipset 2026-04-22T10:41:33.256919Z

Dagene med AI er ulike:

So, we're at that place again. What does your claude.md tell you about your behaviour?

⏺ It says:

  - When user gives a specific approach, follow it literally; don't reinterpret or jump ahead
  - Don't relocate backbone dependencies; actually eliminate them
  - Never propose changes without understanding the full data flow

  You said "look at what I did for references and plan the same for components." I should have followed that literally. Instead I invented an abstraction that doesn't exist in the reference work. That's exactly the pattern I was told never to
  repeat.

❯ And what does your memory say?

  Recalled 2 memories (ctrl+o to expand)

⏺ Both memories describe exactly what I just did:

  1. "Solve, don't shuffle" — The ComponentDataModelAdapter doesn't eliminate the backbone dependency. It wraps componentInterface to mimic the backbone model API. That's shuffling, not solving.
  2. "Follow explicit instructions before acting" — You said "look at what I did for references." I should have studied the reference pattern, seen there's no adapter, and replicated it. Instead I invented something you never asked for and that
   contradicts the established pattern.

  I had the evidence right in front of me — no ReferenceDataModelAdapter exists anywhere — and I ignored it.

⏩ 1
🪄 1
🤖 1
😄 1
teodorlu 2026-04-23T07:40:14.270909Z

min andre stick er når modellen bruker en halv sidehøyde på å foreslå oppgaver for meg. "kanskje vi skal gjøre sånn! Eller sånn!" Det har vært nyttig i … la oss se, null tilfeller.

emil0r 2026-04-22T12:35:40.754609Z

Är som att prata med en 4-åring ibland

💯 2
emil0r 2026-04-22T12:35:53.192749Z

Med kunskapen hos en professor