ELSEIF
Your brief EB
183 stories from 71 feeds 32 clusters Refreshed 8 minutes ago next pull 13:20

DEV TOOLS Signal 185

Experiences with local models for coding

WHY IT MATTERS

For engineers evaluating local models, RAM is the binding constraint and quality is still hit-or-miss compared to cloud-hosted models. The gap between autocomplete and agentic coding viability is substantial—smaller models may work for completion but struggle with the tool calling that agentic use requires.

Written by elseif from the cluster below · every claim links back to a source

The three things worth knowing

01

RAM is the core constraint; models in the 15–25 GB range run comfortably on 48–64 GB machines, but pushing toward 30 GB or beyond risks crashes or severe slowdowns.

02

Response speed has improved dramatically over the past year on Apple Silicon, though it degrades as conversation length grows.

03

Tool calling for agentic coding remains unreliable, with frequent failures, though models can sometimes self-recover; autocomplete is far more viable than agentic use on local hardware today.

THE CLUSTER

Same story, 1 feed.

ORDERED BY FIRST SEEN
Martin Fowler Viability of local models for coding Open ↗
Martin Fowler Experiences with local models for coding Open ↗