Amazing handheld attention sinks + low quality education during the pandemic seems like a deadly combo. I wonder if reading rates will bounce back at all given a few years.

This implementation is fine but the text to speech is meh, Google’s uses https://google-research.github.io/seanet/soundstorm/examples/ which is unfortunately not open source

This is hilarious. you should do standup.

Sorry for the ambiguity, I intend to use this for audio applications (specifically I want to shift 20 kHz to 100 kHz down to human hearing range, tunable by a pot.)

Given the price and low supply I think I’ll go ahead and try to wind my own transformers–thanks for the video, seems perfect!

For the majority of human history, we’ve eaten around wood (around a campfire, a hearth, etc), it makes sense it would become intertwined with our food palette

99 woodcutting!

Cool!

In my experience, mypy + pydantic is a recipe for success, especially for large python projects

What are your thoughts on Javier Milei?

https://lissajous.space

Cross modality is what is missing. We have models that can produce text, hear things, and see things really really well. Facilitating communication between these individual models, and probably creating some executive model to utilize them, is what’s missing. We’re probably still a few years from beginning to touch general intelligence

How can setup get any easier than apt install jellyfin and then going into a web UI to add a few folders?

I’ve been using jellyfin everyday for a few months on my (very tiny) debian server and have never experienced a memory spike like that. Handles music, HD video, even network streams without a hitch

Why does anyone use Plex when Jellyfin exists?

Just like the trillions of parameters that make up machine learning models that can speak or create images