How did we come to the place that the most transparent and open models are now coming out of China—freely sharing their research and source code—while all the American ones are fully locked down
The US is more law and finance led and the first thing seems to be to get an IP advantage and raise money. China I guess less so - they are famously lax on IP and everyone copies everything.
Short cheeky answer is that capitalists need to capture value and communists don’t. Less cheeky answer is that this is a good opportunity for China to make sure the world isn’t dominated by US-sourced AI models.
However in another way the US probably offers more free inference than China. What good is an open 600 billion parameter model to a poor person? A free account with ChatGPT might be more useful to them, though also more exploitative.
> How did we come to the place that the most transparent and open models are now coming out of China—freely sharing their research and source code—while all the American ones are fully locked down
I built a website that lets you browse Pokemon ENS (Ethereum Name Service) names, view their registration statuses and recent sales. It's a small but engaged niche
https://matcha.video is an in-browser transcription-based video editor. Remove silence and keep only the takes that you want with just a couple clicks. Export your edits to an XML format that you can bring into FCP or Davinci Resolve to finish. Free to use, data stays on your machine.
Note: this is a WIP, so if you have any experience working with .fcpxml, I'd love to hear from you!
I love this. In a quick test, I noticed the transcription missed on quite a few of my "filler" words that I might want to edit out, although it did get some of them.
It took me a little while to figure out how to "edit out" words.
EDIT: Turns out should I have read the instructions before uploading my video it would have been straight forward :D
I appreciate you testing it! The whisper model that I'm using indeed doesn't seem to pick up filler words, I'll try to add those ASAP. I'd also like to add background noise removal as a preprocessing step in case the input audio isn't the cleanest.
Were you able to import the result into your editor?
Awesome! I think that makes you matcha's first official user. Please let me know if you have any other feature requests or use cases you'd like me to look into
I've been a YNAB (You Need A Budget) user for years now, and they have a community run extension which includes Sankey Diagrams and different visualizations and reports. Highly recommend!
reply