Let's Build an ACOG - Part 1

Overview

Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!

Grab it

Explore the concept of building an Artificial Cognitive Entity (ACOG) in this 43-minute video by David Shapiro. Delve into topics such as cognitive control in large language models, the importance of natural language in thought representation, and the basic architecture of superintelligence. Learn about the thousand brains theory, designing artificial cognitive entities, and the potential for multiple minds in one brain. Discover the advantages of using timestamps for memory labeling, the power of superintelligence, and techniques like prompt chaining with GPT-3. Examine the process of updating memory indexes, stacking and searching memories, and the relevance of blockchain for AGI. Gain insights into building cognitive control loops and the overall framework of an artificial cognitive entity.

Syllabus

- Introducing the Artificial Cognitive Entity
- The Telltale Brain by VS Ramachandran
- The need for cognitive control in large language models
- The importance of natural language in representing thought
- The basic architecture of a superintelligence
- The possibility of multiple minds in one brain
- The thousand brains theory
- Designing an artificial cognitive entity
- Censorship in a cognitive loop
- Creating a supercomputer
- The advantages of using timestamps to label memories
- The power of a superintelligence
- Using GPT3 to generate prompts for further information
- Prompt chaining with GPT3
- Updating the index of memories
- Stacking memories and searching them for later
- The relevance of blockchain for AGI
- The search for relevant memories
- The cognitive control loop
- Building a cognitive control loop
- The framework of an artificial cognitive entity