OpenAI's Open Source Considerations

On Untitled Linux Show, David Ruggles talks with Ken McDonald and Jonathan Bennett about Sam Altman's thoughts on open source. Has OpenAI not done enough with the "open" part of its name? Check out the article at techcrunch.com...
Full episode at twit.tv/uls188
You can find more about TWiT and subscribe to our full shows at podcasts.twit.tv
Subscribe: twit.tv/subscribe
Products we recommend: www.amazon.com...
TWiT may earn commissions on certain products.
Join Club TWiT for Ad-Free Podcasts!
Support what you love and get ad-free shows, a members-only Discord, and behind-the-scenes access.
Join today: twit.tv/clubtwit
Join our TWiT Community on Discourse: www.twit.commu...
Follow us:
Bluesky: bsky.app/profi...
X: x.com/twit
Mastodon: mastodon.socia...
Facebook: / twitnetwork
Instagram: / twit.tv
TikTok: / twittok
LinkedIn: / twit-llc
#opensource #openai #samaltman

[@BriefNerdOriginal]

The ludicrous ad-hoc licenses that companies like Meta call "open", are something that lawyers should look into.

[@flowqi212]

The term "open-source" has been used in such a terrible and wrong way in regards to LLMs. The "open" thing about most models is that you can download and run them. If you are lucky, then there is a rough description of the training method and there almost never is any "source" available like training code and data 😑.
About medical data. It needs to be anonymous before the training process and thus shareable or there might be a chance to extract it from the resulting LLM!