What is the AI Agentic Systems Ladder on nirvacana.com?

Updated

The AI Agentic Systems Ladder is a framework introduced in a January 6, 2026 article on nirvacana.com, a blog whose tagline is "Thinking in Systems." The article uses the board game Snakes & Ladders—and its ancient Indian ancestor Moksha Patam, where ladders represented virtues and snakes represented vices—as a metaphor for how AI agentic systems progress or regress. The ladder is meant to help readers locate their own AI systems on a spectrum from simple to more autonomous, and to think about what moves a system up or down that spectrum.

The core metaphor

In Moksha Patam, the game board is not random in meaning: ladders lift you toward enlightenment, snakes drag you back. The nirvacana.com article borrows this structure and asks what the equivalent would be for AI agents. The excerpt poses the question directly: "what if there's a 'Snakes &...'"—setting up the idea that agentic AI systems also have rungs that represent progress and hazards that represent regression.

This framing fits the blog's broader "Thinking in Systems" theme: rather than treating AI agents as a single product category, it treats them as systems that occupy different levels of capability and autonomy, with identifiable forces pushing them up or down.

What the ladder is for

Based on the article's framing, the ladder serves two practical purposes:

  • Self-assessment — helping a reader judge which stage their own AI system or use case currently occupies.
  • Design thinking — prompting questions about what would move a system to a higher rung, and what "snakes" (failure modes, loss of control, brittleness) might pull it back down.

The article does not present the ladder as a rigid certification or a vendor benchmark. It is a conceptual model, consistent with the blog's essay style—the same style seen in earlier posts like "Demystifying Artificial Intelligence" (2017) and "Becoming a Data Scientist – Curriculum via Metromap" (2013), both of which use visual or structural metaphors to explain technical journeys.

How to use it

Because the source is a blog essay rather than a specification, the most reliable way to apply the ladder is:

  1. Read the original article at nirvacana.com under "Recent Thoughts" to see the specific rungs the author defines.
  2. Map your current AI system against those rungs—ask how much autonomy it has, how it handles failure, and whether a human stays in the loop.
  3. Identify your "snakes": the conditions under which your system would drop to a lower level of reliability or autonomy.
  4. Identify your "ladders": the concrete improvements (better tooling, evaluation, guardrails) that would move it up.

What to keep in mind

The ladder is a thinking tool, not a standard. The nirvacana.com article presents it through analogy, so the exact number and naming of levels come from the author's own framing rather than an industry consensus. Treat it as a lens for reasoning about agentic systems—useful for discussion and self-assessment, but not a substitute for your own evaluation criteria.