> Most of the gain comes from the per-layer lookup table
This is a dev post classified by AGK Intelligence as AI dev tools (a tool drop), kept by the Dev Radar because it carries real work, not commentary.
> Quantization compiler WOAH And since > Most of the gain comes from the per-layer lookup table I wonder if it'll help with Engrams in big models too but ultimately, the question is how fast does it run
Posted by Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) (79.1k followers) 1 h ago · 11 likes · 1.3k views · view the original post on X. Kept by the Dev Radar as AI dev tools.
More dev work like this
- DeepSeek-V4.1-Flash-1017 on 4 DGX Sparks built a marble music machine in 91.4 min. Every… — @niklaslenz_ai
- I set up Claude Code with a hook so in "OVERNIGHT" mode it has to make all the calls… — @jamonholmgren
- Gemini 4 Argon vs Claude Sonnet 5.5 for computer use. — @coastyai
- Lisa Crossman: "Good or Bad Vibes? Coding with AI in Rust" | RustConf 2026 — @rustaceans_rs
- OpenRouter's guide (dated 2026-11-01) is a 233-line Python script that reviews code by… — @Agent_Jace
- i was confused where one of my @claudeai projects was, and couldn't find it — @theaaron
- LangChain Community Spotlight: Five Controls for LangGraph Agents in Production 🛠️ — @LangChain
- How to Access Multiple AI Models in Your Python Apps Using OpenRouter! @kdnuggets… — @gp_pulipaka
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.3k posts from 5k X accounts over the last 21 days, 8.9k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-01 19:27 UTC. Full method.