15 years of extremely well paid engineers and yet we still cannot delete a track from Recently Listened. ಠ _ ಠ

Recently listened is an automatic list like a log, not user curated. I would never expect to be able to delete things from that list.

I don't need an audit log for my audio. What I do need however is to remove something so it doesn't influence the algorithm. I may not remember every time I play music for a kid to turn on the right mode.

That would also then require them to invalidate caches and rerun scripts to rebuild those lists and recommendations, which costs extra compute.

Didn't think it was possible to ruin Tetris, but... I stand corrected.

I'm not the developer but I'd venture that it's a WebRTC "metadata track" over the same MediaStream the audio is going through.

Metadata tracks aren't formally part of the spec but there's ways to either inject it into the video frame data payload or encode them to audio.

There's libraries like LiveKit that abstract this and make meta tracks easy.

I don't think any web browser is able to make use of WebRTC metadata tracks.

If you buy the adorable monochrome console called the "Playdate" https://play.date it comes with a bunch of games included, including a cute elevator game where you help penguins to different floors while controlling the elevators yourself using the crank: https://play.date/games/flipper-lifter/

Exactly. The system should run in a forcibly limited scope of the current repo only or add permissions for scope to include other org/user repos.

It can't leak a private repo if it can't open it to begin with.

The main library here in Montreal has a sick makerspace with 3D printers (plastic and resin), wood CNC machines, a digital embroidery machine, button maker, shirt press, hole driller, laser cutter, vacuform and vinyl cutter: https://square.banq.qc.ca/fablab

It's a pretty dope library. They also let you borrow movies, videogames for all consoles and even board games, vinyl records and a few music instruments.

And here I thought this was gonna be a Playdate GPT client app.

Ever tried SSH'ing via "Mosh"? https://mosh.org

A hello-world example of what it looks like, and what running it would validate, would be nice.

It's all very abstractly described in the README and that intro page you linked.

Fair enough! I'm probably doing a demo video that'd help with this. Maybe can condense that in text form into the readme as well.

Or why not stay fully local with WebLLM... https://webllm.mlc.ai

That looks great! I also thought about calling the Gemini nano model embedded into Chrome (only extensions can do that). But after some testing on smaller models I found that anything smaller than 9b can’t really handle the complex tool call schema I use.

Qwen3.5 4b is quite good but still gives messy json quite often. But it’s very promising!

Maybe after one more model iteration or some fine-toning we can go fully embedded?