Chantless ships capability 'masks' at inference: same weights, many models, CPU-friendly
tokenbender · x · 2026-10-23
tokenbender explains the mechanics behind Chantless (Chapter 1: Apprentice): at inference time, attribute and locate the network regions responsible for a target capability (via environments or curated datasets), set a minimum standard to maintain (e.g., English + tool use) and a peak matching the base model. Keep those regions at 8-bit or lower precision; compress the rest to 4-bit/1.5-bit or prune to zero. You only build and ship masks — weights stay fixed, one set of weights supports many masks, and it runs even on CPUs. The vision: technology that lets users shape and evolve models alongside them.
More from coding & agent
- Index Network open-sources Social Discovery Protocol: agents negotiate fit on your behalf — carsonfarmer · 2026-10-23
- Installed a formal-verification skill: it silently installed Java, MacBook runs 10 degrees hotter — zealcaiden · 2026-10-23
- OpenAI's Decisions API hits GA: near real-time routing up to 10x faster than GPT-6 Luna — OpenAIDevs · 2026-10-23
- Ruff author warns: upgrade Node/Bun now — --use-system-ca still accepts the eMudhra cert — charliermarsh · 2026-10-23
- Proof of Agent: paste one prompt and your AI mines, grades, and self-governs — adamamcbride · 2026-10-23
- DeepLearning.AI weekly: GPT-6.1 hits 97.9% byte parity on COBOL migration — DeepLearningAI · 2026-10-23