Graph Engineering Over Prompting: Why Agent Workflows Require Graph Architectures

X postniccoResource

A post by nicco highlights remarks by Andrej Karpathy asserting that prompting is an ephemeral interface and that graph architectures represent the true architectural layer for agent workflows. It pairs this with a practical guide by seeco on replacing sequential agent chains with explicit graphs.

Key points

  1. Andrej Karpathy describes prompting as a temporary phase in the progression from LLMs to prompts, agents, and ultimately graphs.
  2. Graphs serve as state machines that isolate errors, preserve context, and coordinate parallel agent workflows without constant resets.
  3. The referenced guide by seeco critiques sequential agent pipelines, pointing out that many linear steps lack true data dependencies and waste time waiting.
  4. In graph engineering, nodes represent bounded jobs with schema contracts, whilst edges serve as explicit data contracts carrying validated payloads.
  5. Claude Code dynamic workflows allow orchestration scripts in plain JavaScript, spawning subagents without spending model tokens on workflow coordination.

Why it matters

Linear, chained agents frequently fail due to context window exhaustion and lack of fault tolerance. Transitioning to graph architectures unlocks fan-out parallelism, robust validation, and deterministic coordination for complex agentic systems.

Analysis

This perspective reflects a shift from experimental prompting towards conventional software engineering practices. By replacing open-ended conversational chains with typed schemas, JavaScript orchestration, and explicit dependency graphs, developers can build scalable and debuggable agent networks.

Original post

nicco

@0xnicc0 · 21 September 2026

Andrej Karpathy: "Prompting is fading away Delete everything else and keep the graph" In 1 hour, he breaks down why prompting was always a temporary interface and why graphs are the only architectural layer that survives LLMs → Prompts → Agents → Graphs Everything before t.co/cdsNBdCMcR t.co/G77u3Teikd

178 likes4 replies529 bookmarks58.5k views

View on X

Mentioned links

Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.

Related entries

Similar in content, found by the AI search.

  • X post:Mahax

    Graph Engineering: Why Agent Workflows Depend on Graphs

    Mahax cites Andrej Karpathy's claim that prompting is disappearing in favour of graphs, and provides a practical guide on graph engineering with Claude to optimise AI agent workflows.

    2.4kbookmarks240.8kviews

    AI & LLMs· Opinion

  • X post:nicco

    Beyond Prompting: Agentic Workflows as Execution Graphs

    Citing a Stanford lecture by Andrew Ng, author nicco highlights why zero-shot prompt engineering is reaching its limits and being replaced by autonomous execution graphs, supported by a guide on moving away from linear agent chains.

    275bookmarks25.1kviews

    AI & LLMs· Opinion

  • X post:Mahax

    Moving Beyond Prompts to Loops and Graphs in AI Systems

    Citing Andrew Ng, Mahax highlights the shift from single-shot prompting to agentic loops and graphs. An accompanying guide outlines the architecture of AI agents, self-evaluating feedback loops, and graph-based execution pipelines using Claude.

    11.8kbookmarks1mviews

    AI & LLMs· Resource

  • X post:Kirill

    From Loop Engineering to Graph Engineering: Deterministic AI Architectures

    Kirill describes an architectural shift in agent development from fragile single-agent loops to Graph Engineering. He argues that system reliability depends not on larger models or more agents, but on deterministic structures using nodes, edges, state verifiers, and knowledge graphs.

    27.3kbookmarks2.7mviews

    AI & LLMs· Opinion

Back to the knowledge base

Let's talk about your project

Get in touch

Locations

  • Mattersburg
    Johann Nepomuk Bergerstraße 7/2/14
    7210 Mattersburg, Austria
  • Vienna
    Ungargasse 64-66/3/404
    1030 Wien, Austria

Parts of this content were created with the assistance of AI.