Posted here many times https://news.ycombinator.com/from?site=github.com/localsend

Yes, this is being done, and has been done. A variety of linguistic atlas projects record audio and make written phonetic transcriptions: https://www.linguisticatlasproject.org/ There are also many dialect and accent arechives, such as https://www.dialectsarchive.com/ and https://accent.gmu.edu/.

Outstanding. May his fax endure forever.

Your malarkey sounds the same in this language, too.

people out here applying their genius with things to with no profit incentive

How many times will this particular problem be solved? 500? 2000? 10,000?

Please share the other places this has been "solved".

I'm a daily user of Wispr Flow, in part because all previous attempts to try and reimplement it myself have been terrible. Tortoise and all of the other suggestions people offer up have terrible performance in comparison.

Here are just a few I found in just a few minutes of looking:

https://github.com/Ljove02/AIT-AiTranscribe-MacOS

https://www.soink.ai/

https://hitoku.me/draft/

https://altic.dev/fluid

https://www.typewhisper.com/

https://github.com/kitlangton/Hex

https://github.com/cjpais/handy

https://spokenly.app/

https://www.trykoe.com/

https://saga.deepgram.com/

You can find more here under the "Dictation" tab:

https://docs.google.com/spreadsheets/d/1HtJN4oQ6oBDFmFaF4Qeq...

* AIT-AiTranscribe-MacOS: Parakeet + Whisper based. Parakeet isn't anywhere near as accurate as Wispr Flow (or as fast), Whisper is way too slow.

* Soink: Not OSS

* Hitoku: Not OSS

* Fluid: Parakeet + Whisper again. Also Apple Speech which is even worse than all of the other options

I'm not going to dissect every example of why this absolutely is not an already solved problem.

Still sounds bitter even if you were right. I would reconsider this approach.

Fair. Though I'd argue there's a difference between bitter and just telling the truth with receipts. I took the severance, built something new, and haven't looked back. If it reads as bitter that's probably on me as a writer.

There goes two decades of muscle memory.

Aren't there a lot of these? Like a lot a lot?

Yes, this is common.