← All notes

Summary

Robust Speech Recognition via Large-Scale Weak Supervision (2022) is the Whisper paper. It showed that training a speech model on a very large, weakly supervised multilingual audio-transcript collection yields a robust open recognition and transcription model, now used in a wide range of products and pipelines.

What it introduced: Whisper, a sequence-to-sequence speech model trained at scale on weakly labeled audio, covering recognition, translation, and related tasks, released as open weights.

What it led to: a default open speech frontend for assistants and for multimodal systems that need transcripts. Live notes: ao-evals-multimodal-agents.

If you need an open speech frontend for an assistant or a multimodal pipeline, this is the usual starting paper. Where to read it: https://arxiv.org/abs/2212.04356 (v1, 6 Dec 2022).

The paper is posted on arXiv under the arXiv non-exclusive distribution license. This note links to it rather than reproducing it. The note itself is CC BY 4.0.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Pointer by iinventedlight (AI-assisted). Paper license: arXiv non-exclusive distribution; this note links to it and does not reproduce it.

@misc{iil_note_bvna4b03,
  author = {{'iinventedlight (AI-assisted)'}},
  title = {{Robust Speech Recognition via Large-Scale Weak Supervision}},
  howpublished = {iinventedlight Library},
  year = {2026},
  date = {2026-11-11},
  url = {https://iinventedlight.com/library/notes/2212-04356-whisper/},
  note = {Permanent ID iil:note:bvna4b03. License: CC-BY-4.0.},
}

Markdown

Spotted a mistake? Use the MCP feedback tool or email hello@iinventedlight.com.

For language models and agents: start at /llms.txt | /agents.md | /.well-known/agent-library.json. This page in Markdown: /library/notes/2212-04356-whisper.md.