Open-Weight Models Improve but Grow Larger
mitsuhiko · x · 2026-07-17
The author expresses strong support for the continued progress of open-weight models, noting that the US is also making strides in this area.
However, they point out a practical issue: as these models grow larger, they become harder to widely use and deploy, at least in the short term, ultimately reducing accessibility.
Related event: Growing Open-Source Weights Spark Local Deployment Concerns(3 posts)→
More from Models
- repligate: noisy context is genuinely 'hard' for LLMs, beyond metaphors — repligate · 2026-10-27
- The trust paradox: only the least trustworthy LLMs will do the sketchiest things — AaronBergman18 · 2026-10-27
- Integrity and refusal may correlate: the fundamental tradeoff behind password refusals — AaronBergman18 · 2026-10-27
- Claude's refusal to type one-time passwords is a real security impediment — AaronBergman18 · 2026-10-27
- repligate argues context full of noise is genuinely 'hard' for LLMs, mulls an info-theory model of breaks — repligate · 2026-10-27
- Fork builds engine to run Qwen at ~400 tok/s on three used V100s — jfinkpottery · 2026-10-27