2.1 percent of the Bluesky posts about dictation that we coded (19 of 926) praise ai rewriting.

Measured by us on October 19, 2026, from recordings

Our own measurement

This number comes from our benchmark, not from another publication. The full record (the setup, every recording's checksum, every transcript and timing) is published, so you can recompute it.

Dictation App Problems: What Users Complain About Download the full record (JSON)

What it measures: Documents: Bluesky posts that discuss dictation or speech to text, gathered from public pages and coded with a fixed codebook. Theme: The app rewriting, cleaning up or changing what was said with AI: wanted or unwanted changes, tone, filler words.
How it is measured: Each document was coded by a researcher against the published codebook (themes and the problem/praise/mention polarity) before any count was made; the counts are calculated by code from the codes.
What it does not show: App Store pages show the reviews Apple selects (often the most helpful), not a random sample; Hacker News comments come from developers and founders. A theme counts only when the document states it.

In the data

A row of Dictation App Problems: What Users Complain About, as AI rewriting: 2.1 percent of documents.

Theme: AI rewriting
Reported as: Praise
Documents: Bluesky posts
Source: Bluesky
Documents coded: 926
Documents with the theme: 19

Used to calculate

The what people report about dictation tools: coded app reviews and developer comments dataset holds 248 data points, each a published figure traced to its source.