lichess.org
Donate

How to Lichess Voice

#22
Pretty much. We have the source code of every single component that runs on it. We have total access to the filesystem and everything stored in the various databases on every server.

Did you know you can monitor socket traffic in your browser’s network inspector? Here’s an idea - use that to find some actual evidence that I’m not on the level. I’ll wait.
#24
I was enjoying this very interesting read on new tech to play chess here and was impressed with your step-by-step instructions to help people try this new feature. Then out of the blue you wrote, "if you've soiled your undergarments..." and I had a good laugh. You're funny too. Thanks for sharing and keep up the good work!
#25
@schlawg thanks for the post! This is something I've been thinking about implementing myself, but I would have had to implement it as a third party solution because I wouldn't have the skills to integrate/contribute it to Lichess this nicely.

Do you think there is a way to get the speech recognition up to a higher accuracy rate, since the field is moving so fast nowadays and it's been half a year? I was trying out OpenAI's Whisper today and it transcribes chess moves so well out of the box. I understand that you want the model people need to download to be fairly small, but maybe there could be some option somehow to download a better one if the user wants?
#26
Unfortunately, this is not in the cards at present. The reason Whisper does so well is because OpenAI uses a large language model to disambiguate (which requires 800+ GB storage and many H100s in a data center).

For doing large LLM edge inference on your device, even the tiny version of the open source llama 2 model requires something like 28 GB of storage. That's a big download for a website.

But with the speed things are moving, it's possible that flawless conversational chess moves could be widely available without requiring a data center or a huge download in a year or two.