Translate between plain English and deliberately over-engineered “Claudish” with two tiny ProgramAsWeights functions.
pip install programasweights --extra-index-url https://pypi.programasweights.com/simple/import programasweights as paw
to_claudish = paw.function("ca9d5165b6c8e6615529")
to_english = paw.function("e469f61ccab2699fbd51")
print(to_claudish("Only owners can merge."))
print(to_english("The honest shape is asymmetric: the data is correct; the format is hard to read. Correctness landed; legibility did not."))The programs download once and then run locally. Pass no max_tokens; PAW
stops naturally at EOS.
For a small command-line wrapper:
python translate.py to-claudish "The release can go out after Alice approves the final report."
python translate.py to-english "Here’s where I’d hold the line: do not launch until the tests pass. Green is the gate, not a suggestion."Copy either spec into your own PAW program or adapt it for another model or workflow.
Inspired by gvzdv/claudish-to-english. This is an unofficial parody project and is not affiliated with Anthropic.
The translation functions are described in Compile by Training (§6.3). If you use this demo in research, please cite that paper and Program-as-Weights for the underlying fuzzy-function programming model.
Compile by Training: Turning Natural-Language Specifications into Local Neural Functions
EMNLP 2026 System Demonstrations · Preprint
@inproceedings{deng-etal-2026-compile,
title = {{Compile by Training: Turning Natural-Language Specifications into Local Neural Functions}},
author = {Deng, Yuntian and Nie, Pengyu and Shieber, Stuart},
editor = {Agrawal, Sweta and Abad, Alberto},
booktitle = {Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations},
month = oct,
year = {2026},
address = {Budapest, Hungary},
publisher = {Association for Computational Linguistics},
url = {https://aclanthology.org/2026.emnlp-demos.78/},
doi = {10.18653/2026.emnlp-demos.78},
pages = {830--839},
ISBN = {979-8-89176-461-3}
}Program-as-Weights: A Programming Paradigm for Fuzzy Functions
NeurIPS 2026 · Preprint
@inproceedings{zhang2026programasweights,
title = {{Program-as-Weights: A Programming Paradigm for Fuzzy Functions}},
author = {Zhang, Wentao and Hotsko, Liliana and Kim, Woojeong and Nie, Pengyu and Shieber, Stuart and Deng, Yuntian},
booktitle = {The Fortieth Annual Conference on Neural Information Processing Systems},
year = {2026},
url = {https://arxiv.org/abs/2607.02512}
}