The Most Interesting Thing in AI
Neuro-Symbolic AI - with Nicholas Thompson and Ohad Elhelo
Sponsors in this episode
No sponsor reads found in this episode's notes.
Host-read ads that the notes don't mention can't be found this way.
Show notes
In this episode, Atlantic CEO Nicholas Thompson speaks with Ohad Elhelo, co-founder and CEO of AUI, about neuro-symbolic systems. Apollo-1, AUI's neuro-symbolic foundational model, promises both the fluency of generative models and the reliability of symbolic reasoning, an AI method that uses symbols and explicit rules to represent knowledge and perform logical inference. Could this hybrid architecture finally unlock task-oriented AI at scale?
(00:00) Introduction
(02:57) A rift in the AI community
(04:59) Why neuro-symbolic AI could succeed where chatbots have failed
(12:04) How AUI built its model
(19:11) How to solve the hallucination problem
(25:46) Business use cases for neuro-symbolic AI
(30:48) What happens once Apollo One launches
(35:03) The most interesting way Ohad has used AI
Learn more about your ad choices. Visit podcastchoices.com/adchoices
From the show's RSS feed.