A selective mechanism in encoding allows for a learned aspect in creating long-lived memories, particularly demonstrated on DNA in the original paper. This mechanism enables handling super long sequences, offering an alternative to the episodic nature of transformers where managing context windows is a challenge. The potential to build up, save, and fork states over time presents interesting opportunities. Contrary to earlier assumptions, the A100 utilized for training has more SRAM capacity, aiding in keeping states closer to computational processes.

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