AI Empowerment: Understanding the alien intelligence of GPT

This week

GPT and other language models haven’t yet reached human-level intelligence—but they are intelligent, though in ways that may be fundamentally alien to us. Grasping this alien intelligence is essential for understanding their trajectory, ensuring AI safety, and using them effectively.

Join A B Carter for a discussion of how large language models are trained—and what that reveals about their unique cognitive capacities.

We're discuss:

• Simulators — LessWrong
• Janus' Simulators - by Scott Alexander - Astral Codex Ten

Supplemental:

• What Is ChatGPT Doing … and Why Does It Work?—Stephen Wolfram Writings: excellent overview of LLMs but not for the faint at heart.

Agenda
6:00-6:15 Socializing and networking
6:15-7:30 Main presentation & group discussion

General info about this meet-up
A group for people to meet and discuss the development of AI, how it might impact society, and what can be done to facilitate humans flourishing in the coming age of AI.

Discussion and topics are aimed at cultivating high-level, philosophically-informed perspectives on the topic of AI impacting human existence.

Come join us!

Brought to you by AIGS Canada and Trajectory Labs.