# As We May Search

arXiv: https://arxiv.org/abs/2606.29652 · HTML: https://mlsanity.com/paper/2606.29652/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Privacy and security of RAG, one of 15 directions in Retrieval-augmented generation.
- Builds on 22 of its 62 references that are in the map.
- First posted on arXiv 28 Jun 2026.

## Placed in

- Retrieval-augmented generation / [Privacy and security of RAG](https://mlsanity.com/rag/privacy-and-security-of-rag/)

## Builds on (most cited first, 22 of 22)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Billion-scale similarity search with GPUs](https://mlsanity.com/paper/1702.08734/) (2017)
- [MS MARCO: A Human Generated MAchine Reading COmprehension Dataset](https://mlsanity.com/paper/1611.09268/) (2016)
- [BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models](https://mlsanity.com/paper/2104.08663/) (2021)
- [Efficient and robust approximate nearest neighbor search using Hierarchical Navigable Small World graphs](https://mlsanity.com/paper/1603.09320/) (2016)
- [ZeroSearch: Incentivize the Search Capability of LLMs without Searching](https://mlsanity.com/paper/2505.04588/) (2025)
- [Text Embeddings Reveal (Almost) As Much As Text](https://mlsanity.com/paper/2310.06816/) (2023)
- [Power Hungry Processing: Watts Driving the Cost of AI Deployment?](https://mlsanity.com/paper/2311.16863/) (2023)
- [The Good and The Bad: Exploring Privacy Issues in Retrieval-Augmented Generation (RAG)](https://mlsanity.com/paper/2402.16893/) (2024)
- [The Text Anonymization Benchmark (TAB): A Dedicated Corpus and Evaluation Framework for Text Anonymization](https://mlsanity.com/paper/2202.00443/) (2022)
- [How Hungry is AI? Benchmarking Energy, Water, and Carbon Footprint of LLM Inference](https://mlsanity.com/paper/2505.09598/) (2025)
- [PersonaRAG: Enhancing Retrieval-Augmented Generation Systems with User-Centric Agents](https://mlsanity.com/paper/2407.09394/) (2024)
- [MeMemo: On-device Retrieval Augmentation for Private and Personalized Text Generation](https://mlsanity.com/paper/2407.01972/) (2024)
- [SoK: Privacy Risks and Mitigations in Retrieval-Augmented Generation Systems](https://mlsanity.com/paper/2601.03979/) (2026)
- [Anatomizing Deep Learning Inference in Web Browsers](https://mlsanity.com/paper/2402.05981/) (2024)
- [Differentially Private Synthetic Text Generation for Retrieval-Augmented Generation (RAG)](https://mlsanity.com/paper/2510.06719/) (2025)
- [Variations in Relevance Judgments and the Shelf Life of Test Collections](https://mlsanity.com/paper/2502.20937/) (2025)
- [FedRAG: A Framework for Fine-Tuning Retrieval-Augmented Generation Systems](https://mlsanity.com/paper/2506.09200/) (2025)
- [HyFedRAG: A Federated Retrieval-Augmented Generation Framework for Heterogeneous and Privacy-Sensitive Data](https://mlsanity.com/paper/2509.06444/) (2025)
- [Efficient Privacy-Preserving Retrieval Augmented Generation with Distance-Preserving Encryption](https://mlsanity.com/paper/2601.12331/) (2026)
- [Ten Blue Links on Mars](https://mlsanity.com/paper/1610.06468/) (2016)

## Abstract (arXiv)

The sensitive information in personal documents, legal files, and medical records is among the most valuable things to search, yet current retrieval-augmented generation systems still require sending content to remote servers. We propose local-first IR, a design philosophy where indexes, models, and inference reside on user devices, treating remote services as optional. This paper makes four contributions: (1) a framework organizing retrieval architectures along three dimensions: privacy and control, capability, and accessibility, (2) experiments on consumer hardware across five benchmarks, scaling from 1K to 1M documents with dense retrieval, BM25, and hybrid fusion. Dense retrieval keeps over 91% nDCG@10 up to 100K documents, with approximate HNSW indexes extending this to 1M with only 2% quality loss; a 7B local language model reaches within 4 points of a cloud baseline on answer quality, (3) competing perspectives for and against local-first IR, informed by experimental evidence, and (4) a research agenda identifying open problems. The real tradeoff is scope rather than quality: what matters is what you can search, not how well you can search it.