Why keep making LLMs bigger? A case for DeepSeek and Qwen's small-model extremism
gnukeith · x · 2026-10-12
The author questions the "bigger = better" orthodoxy, arguing the ultimate goal for LLMs should be smaller and smarter—perhaps giant models exist mainly to distill into better small ones.
He praises the openness of Chinese AI labs: DeepSeek V4.1 Flash is a 196B-parameter (FP8) model, but its 183–204GB engram tables can be offloaded to SSD; Qwen 3.8 Flash Next lets you offload a 51B Prompt Lookup Engine (PLE), a specialized n-gram embedding lookup table. He also cites Qwen 3.8 27B as proof Chinese labs are pushing small models to their limits.
Related event: Debate rages over whether ever-larger LLMs are progress or a compute moat(2 posts)→
More from Infra
- Grid becomes AI's new bottleneck: Entergy warned Amazon of peak-time cutoffs, AWS built Project Falcon — kimmonismus · 2026-10-12
- 'AI always maxes out your GPU'? Devs push back on the myth — mgostIH · 2026-10-12
- Yandex reports data center damage causing service outages, 'Internyet' meme ensues — kingslyj · 2026-10-12
- Linus Torvalds is using Claude Code on a hefty Linux kernel performance improvement — PointmanW · 2026-10-12
- 4070 Ti Super beats RTX 5070 2x on MiniMax H3 int8 in local inference benchmark — smellyelon · 2026-10-12
- Strata Runs a Big Open Model Locally on a Normal Gaming PC — alex_verem · 2026-10-12