News for hackers. Stuff that matters.

agenticup

i guess inference engineering, like dpsark or dflash specific speculative decoding technqiues
can the orniths self scaffolding could learn to scaffold the rlm loop?
Re:Qwen 3.6 27B is the sweet spot for local development on Tuesday June 30, 2026 @02:13AM (#48729078)
qwen 3.6 27b and qen35b a3b work like magic, if we get dpark speculative decoding versions of these models it will further improve the throughput