The latest open AI models have a lot of code early in their training and presumably this is kind of related where there's like with presumably more code-friendly tokenization it has an easier time learning certain logical structures. We are very excited about the either you know scanning the arc or see to be able to directly in the future model the raw format of any any file any input any modality I think that's really exciting direction for usYeah so I'll be looking out for the next paper I just looked the this one was published May 19th and it's we're recording here about a month later knowing the size of the clusters that you guys are working with over there.

Get the Snipd
podcast app

Unlock the knowledge in podcasts with the podcast player of the future.
App store bannerPlay store banner

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode

Save any
moment

Hear something you like? Tap your headphones to save it with AI-generated key takeaways

Share
& Export

Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode