full-title:: [Edge 380: Inside SELF-Discover: Google DeepMind's LLM Reasoning Method for Solving Complex Tasks](https://omnivore.app/me/edge-380-inside-self-discover-google-deep-mind-s-llm-reasoning-m-18e60b73f4d)
site:: [thesequence.substack.com](https://thesequence.substack.com/p/edge-380-inside-self-discover-google?isFreemail=true&post_id=142822544&publication_id=54309&r=z2w67&token=eyJ1c2VyX2lkIjo1ODkyMTU2NywicG9zdF9pZCI6MTQyODIyNTQ0LCJpYXQiOjE3MTEwMTg4ODAsImV4cCI6MTcxMzYxMDg4MCwiaXNzIjoicHViLTU0MzA5Iiwic3ViIjoicG9zdC1yZWFjdGlvbiJ9.3tbCxwSX7OY06e94zth4T3shVeiXvMQleb3-gdXP5F0&triedRedirect=true)
author:: davide.diruscio@gmail.com
labels:: [[Newsletter]] [[LLMs]] 
date-saved:: [[21-03-2024]]
source:: [[Omnivore]]
state:: [[Reading]]

	- ### Highlights
	  collapsed:: true
		- > continues evolving as one of the most fascinating areas in generative AI with research papers pushing [⤴️](https://omnivore.app/me/edge-380-inside-self-discover-google-deep-mind-s-llm-reasoning-m-18e60b73f4d#e4724bcc-979c-4da4-9f05-380cbedcdf23) 
		  omnivore-note:: [[IDEAS]]  #omnivore-note-color
		- > method, decomposition-based prompting, draws from the human ability to break down complex problems into smaller, manageable parts and then address each [⤴️](https://omnivore.app/me/edge-380-inside-self-discover-google-deep-mind-s-llm-reasoning-m-18e60b73f4d#d7dc0606-700b-4026-b37c-21061d0d0e90)