The idea of separating the agents surprised me. It seems like a sort of different take on a somewhat common theme lately which is like models talking to models in high dimensional space. Except you've created a hierarchy that can be trained end to end. So now we have kind of multiple transformers in a single architecture all operating under one loss function and one optimization function. And so it has this kind of general fork shape.

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