Graph Engineering

@agenticamit

12.1K followers · usually 3.9K views a post

Views

25.2K

Likes

1.0K

Comments

81

Shares

81

SoundTikTok sound · original sound - agenticamit

Why it worked

The video introduces a novel and potentially important concept in AI, "Graph Engineering," which is likely to pique the interest of those in the AI space. By framing it as the "next evolution" and linking it to established concepts, it creates a sense of forward-thinking and essential knowledge.

Summary

The video introduces the concept of "Graph Engineering" as a new term in the AI space. It explains that Graph Engineering is an extension of loop and harness engineering, and is considered the next evolution in the field, drawing on graph theory.

Structure

1Introduction of "Graph Engineering"
2Explanation of its relation to loop and harness engineering
3Positioning Graph Engineering as the next evolution in AI
4Mention of graph theory's resurgence

On-screen text

Graph Engineering The Karpathy Loop, Improved 1000x by Itself The Anthropic Playbook Agentic Software Engineering Practice 2023 The Big Picture Three layers. Three jobs. One system. Harness Build the environment around the model Environment -> Feedback -> Flow Loop Design the repeated work-and-feedback cycle Graph Make the workflow topology explicit and controllable

Transcript

Just a quick PSA to everyone in the AI space right now. There is now a new engineering term that you have to learn, graph engineering. And it's an extension of loop engineering, which is an extension of harness engineering, which is an extension of memory engineering. And you get the point, graph engineering is the next evolution. And just as history predicts, graph theory comes right back into the fray to make things just that much better.