phil

"prefer duplication over wrong abstraction."

truth.
timestamp: https://www.youtube.com/watch?v=8kMaTybvDUw&t=889s

🐵<🐦‍⬛> ← 🙈🙉🙊 → /phil →  github.com/philcockfield →  
ƒ(🌳) →  🧫 /sys  ←  @cell <T>, 🧠 human.

🎯 

i find that coding agents make this even more true, for multiple reasons:
- it writes the code for you (so no need to worry about saving time)
- context collocation (no need for llm to jump between the extracted code, often in different file or files)

fortune cookie casts, memes, parenting, building mini apps, /framedl, /framedl-pro & /calmer

I do think emphasis should be put on “wrong” abstraction. 

Super interesting angle @ds8

I think the capacity to clean up duplication once a “good” abstraction is discovered, with LLMs, is a hugely promising direction too.

Duplication can be cleanly replaced….abstractions generally require deeper untangling.

Oh boy, TRUTH!

And amplified by the fact that making up (a shitty, or any) abstraction has an implicit joy and satisfaction to it - so we do it more…and then the messy untangling trap is set.

oh, hundred percent... but it's severely underappreciated how easy it is to come up with an abstraction that's wrong and not notice until too late