Made some improvements to the X-like interface for LLM chat, and slapped a name on it: Lapin. Give it a whirl at https://t.co/JZNJms3gQD :) You can dive into rabbit holes by expanding on "tweets"

TIL the importance of pimped out hardware

the impacted services list just keeps growing and growing...

Moving back to Korea soon - anyone have recs for good private office solutions around Jamsil or Samsung? Just need reliable ethernet + electric Maybe @harpriiya or @Davidjpark96 know?

be grug-brain

Well that's that

Snipped from

Bruce lee and the midwit meme

Writing a memory allocator - v2. Just a simple optimization on top of the existing memory allocator + a short exploration of ptmalloc2. https://t.co/GpyqD62Ou6

Writing a memory allocator - v1. A continuation of our exploration from v0 without the assumption of fixed-size blocks. https://t.co/3NKPjeYsUX

Writing a memory allocator over the next couple weeks. v0 makes a simplifying assumption that all memory demanded is fixed-size. I'll relax this assumption in v1 of the allocator. https://t.co/BRGSov8UQr

Continuing last week's exploration into cache misses:

Experimenting with cache misses:

https://t.co/Fzq0ww9CX6

Some snips of spec layouts

Maybe smaller font opens it up small. Getting rid of the date helps too because I don't care about the dates in this context

Looks too cramped in general

https://t.co/S3foT7R2KA

This is a sweet detail; keep the tool-tip visible once it shows up.

Lots more I want to look into. Honestly not sure how I would make the onboarding process better. For a tech-savvy audience this onboarding process makes sense (it's pretty clear that Linear is for software workflows; you have the option to connect to Github right off the bat)

Sort of like Raycast: