Agentic AI: Redefining How We Interact with Technology

11 snips

Aug 6, 2024

Amir Behbehani, Founder and Chief AI Engineer at Memra and former AI research scientist at Google and Meta, dives into the transformative potential of agentic AI. He unpacks the challenges of current AI systems, highlighting memory management and data orchestration. Amir introduces an emergent five-layer AI stack for enhanced functionality and discusses the significance of vector and graph databases. He envisions a future where AI reshapes the labor market, encouraging entrepreneurial endeavors through improved productivity and automation.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

Agentic AI vs. LLMs

Agentic AI focuses on imperative commands, enabling automated systems to perform actions.
This contrasts with current LLMs, which primarily engage in interrogative conversations.

INSIGHT

Emergent AI Stack Overview

The Emergent AI Stack consists of five layers: CPU (LLMs), long-term memory, short-term memory/context, agentic layer, and application layer.
Each layer builds upon the ones below it, with the agentic layer leveraging the memory layers.

INSIGHT

Vector vs. Graph Databases

Vector databases are suitable for general inquiries, while graph databases excel with dense information.
Memra vectorizes knowledge graphs, combining the strengths of both for precise retrieval.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

Welcome back to the first episode of season 3 of the Talking AI podcast, formerly known as the Built Right podcast!

We’re kicking off with Amir Behbehani, Founder and Chief AI Engineer at Memra, a platform that helps build, deploy, and automate workflow through AI. He and host Matt Paige look at some of the hurdles AI needs to overcome to fully integrate into enterprise, such as lack of proprietary data access and memory. Amir delves into the science behind these challenges and potential solutions. He introduces the concept of ‘agentic AI’ and the emergent AI stack, made up of five layers, aimed at enhancing AI functionality through memory management, data orchestration, and agentic reasoning.

They also discuss the role of vector and graph databases, the importance of memory and context, and exploring future use cases in a marketplace with agentic AI. Amir also speculates on the transformative potential of AI in the future labor market - hinting at more entrepreneurial endeavors as individuals leverage AI for improved productivity and automation.

Enjoyed this episode? Don't forget to subscribe to our podcast for more insightful discussions on the latest in AI technology! Follow us on social media for updates and exclusive content. Your feedback matters—leave us a review and let us know what topics you'd like to hear about next!

Key moments:

Introduction to Amir and Memra
Challenges with current LLMs
Emergent AI stack overview
Layer-by-layer breakdown of the AI stack
Agentic AI and memory management
Future of work with agentic AI
The future of Memra

Key links:

Mentioned in this episode:

AI Opportunity Finder

Feeling overwhelmed by all the AI noise out there? The AI Opportunity Finder from HatchWorks cuts through the hype and gives you a clear starting point. In less than 5 minutes, you’ll get tailored, high-impact AI use cases specific to your business—scored by ROI so you know exactly where to start. Whether you're looking to cut costs, automate tasks, or grow faster, this free tool gives you a personalized roadmap built for action. 👉 Try it now at https://hatchworks.com/ai-opportunity-finder/