# Program-as-Weights: A Programming Paradigm for Fuzzy Functions

Source: [arXiv](https://arxiv.org/abs/2607.02512v1)  
Feed7 permalink: https://feed7.dev/p/2607-02512v1-1dr5458  
Published: Unknown  
Trust: Needs Review (needs_review)

## Why Included

Program-as-Weights compiles natural-language fuzzy functions (JSON repair, log filtering) into adapters for a frozen 0.6B interpreter — matching Qwen3-32B prompting at ~1/50th the memory, 30 tok/s on an M3.

## Source Summary

**The gist** A **4B compiler model** trained on **FuzzyBench (10M examples)** turns a natural-language spec into a parameter-efficient adapter that runs on a frozen **0.6B Qwen3** interpreter. The compiled function matches direct prompting of **Qwen3-32B** at about **1/50th the inference memory**, hitting **30 tokens/s on a MacBook M3**.

## Practical Implication

**Why it matters** The fuzzy glue you currently route to an LLM API — repairing malformed JSON, flagging important log lines, ranking by intent — could become a compile-once, run-locally artifact: reproducible, cheap, offline. It reframes big models as **tool builders** invoked once per function definition rather than once per call.

## Agent-Ready Context

**The gist** A **4B compiler model** trained on **FuzzyBench (10M examples)** turns a natural-language spec into a parameter-efficient adapter that runs on a frozen **0.6B Qwen3** interpreter. The compiled function matches direct prompting of **Qwen3-32B** at about **1/50th the inference memory**, hitting **30 tokens/s on a MacBook M3**.

**Why it matters** The fuzzy glue you currently route to an LLM API — repairing malformed JSON, flagging important log lines, ranking by intent — could become a compile-once, run-locally artifact: reproducible, cheap, offline. It reframes big models as **tool builders** invoked once per function definition rather than once per call.

**Watch out** Results come from tasks resembling the **FuzzyBench** training distribution; how compilation holds for messier or novel specs, and how you validate a compiled function's behavior before trusting it, is untested here.

## Context Map

- Layer: model
- Domains: coding, research
- Topics: open-models, model-selection

## Uncertainty

- Results come from tasks resembling the **FuzzyBench** training distribution; how compilation holds for messier or novel specs, and how you validate a compiled function's behavior before trusting it, is untested here.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.