Active Working Memory: The RAM of Agentic Systems

Part 2 of the Building the AI Memory Stack series

When I published the first article in this series, I thought I was writing about context windows.

The more I wrote, the more I found myself bouncing between documents. I had the glossary open in one browser tab. The Memory as Infrastructure article was open in another. A third tab contained notes about Context Hydration. GitHub was open with the SDK specifications, and a handful of Architecture Decision Records were sitting beside my editor.

None of those documents individually contained “the answer.” Together, they formed the temporary collection of information I needed before I could make progress.

Then something unexpected happened.

I was no longer writing about context windows.

I was reconstructing a memory hierarchy, and the context window was only its first layer.

That is why this became a series.

In Part 1, I argued that the context window is best understood as the CPU cache of an AI system. It is an execution surface, not a memory system. Like CPU cache, it is optimized for fast access and exists only for the duration of the work being performed.

But caches do not populate themselves.

Neither do context windows.

The Working Set Before the Work

Before I could write, I assembled a working set.

Browser tabs and open documents

- Sovereign Systems glossary
- Memory as Infrastructure
- Context Hydration notes
- SDK specifications
- Architecture Decision Records
- GitHub repository
- Article draft

            |
            v

    Active Working Memory

            |
            v

      Context Window

            |
            v

         Reasoning

That collection was not my long-term memory. It was a temporary working set assembled for one task. My brain still had to compare ideas, notice contradictions, and produce something new. The documents simply gave the reasoning process the information it needed.

Agentic systems work in much the same way.

Before a model begins reasoning, documents have been retrieved, tools have executed, state has been restored, policies evaluated, and responses normalized. The prompt is usually the final artifact produced by an orchestration layer, not the beginning of one.

By the time the model receives its first token, dozens of retrieval, filtering, ranking, and assembly decisions may already have been made.

That assembled execution state is what I call Active Working Memory.

Active Working Memory Diagram

This article is about the second layer in that stack.

Models Reason. Applications Assemble.

One sentence captures the distinction this entire article is trying to make.

Models reason. Applications assemble.

Those are different responsibilities.

A model does not retrieve documents.

A model does not decide which Git commit matters.

A model does not know whether a tool response is stale.

An application does.

Consider a concrete case. An agent is asked to update a Python SDK.

It retrieves the ADR describing the package boundary.

It loads the glossary definition of Active Working Memory.

It checks the current implementation on GitHub.

Only then does it build the prompt.

People often talk about “putting something into the context window.”

That wording quietly suggests the context window is responsible for finding, selecting, and organizing information.

It isn’t.

By the time inference begins, the application has already decided what the model will, and will not, be allowed to see.

Working Memory Sequence Diagram

The context window does not begin the process. It receives the result of the process.

Many failures blamed on the model are actually failures of context assembly. The wrong evidence was retrieved. A stale document won. A constraint never made it into the working set. Those are architecture problems before they are model problems.

RAM for Agentic Systems

If the context window is the CPU cache, Active Working Memory is the system RAM.

Layer Primary Responsibility Typical Owner
Durable Memory Preserve knowledge Storage / Application
Active Working Memory Assemble task state Orchestrator
Context Window Present selected information Runtime
Model Perform inference LLM

Each layer has a distinct responsibility. No layer substitutes for another. A larger context window does not repair poor selection, and a better model cannot reason over evidence it never receives.

Search Finds. Context Assembly Decides.

Search answers one question.

What information exists?

Context assembly answers another.

Given this task, what information belongs together?

Those sound similar. Architecturally, they are completely different.

Active Working Memory presents that second decision to the model.

Two systems can use the same model, the same context window, and the same knowledge store yet produce very different results, because one assembles a concise, relevant working set while the other floods the model with loosely related information.

The model is identical.

The memory architecture is not.

An Architectural Boundary

Once Active Working Memory is treated as a real layer, context assembly stops looking like prompt engineering and starts looking like systems architecture.

Information crosses this boundary only after decisions have been made about relevance, authority, recency, format, and priority.

Every unnecessary document increases Context Tax. Every verbose tool response competes for attention. Every missing source creates a blind spot the model cannot recognize from inside the window.

The context window can only reason over what Active Working Memory hands it.

Looking Ahead

The working set I assembled while writing this article disappeared as soon as the article was finished.

The article remained.

That distinction turns out to matter.

Agentic systems face the same decision. What belongs only in today’s working set? What deserves to become tomorrow’s memory?

That is where Part 3 begins.

Models don’t assemble context. They inherit it.

Working memory is not where knowledge lives. It is where knowledge collaborates.

Facebooktwitterredditlinkedinmail

The Resilient Farm

(The Final Gamble and the Zero-Waste Finale)

The primary harvest is over. The bins are mostly empty and the air in the cellar is thick with the smell of fermentation. In a traditional vineyard, this is where the data collection stops. But in The Agile Harvest, the final 5% of the season is where true resilience is built.

A Resilient Farm does not just survive the primary market. It captures the value that everyone else leaves on the ground.

The Final Gamble: Ice Wine and the Almanac

While the Pinot Noir is tucked away in tanks, the Riesling is still out there. The crows are getting thick in the rows and the leaves are beginning to drop. Holding out for a sustained freeze is a calculated risk, but it is the high-hanging fruit of the season.

The Scribe is no longer just watching sugar levels. It is watching the long-range forecast for a very specific tactical trigger.

The Late-November Exchange:

Farmer: “Scribe, the Almanac says we have a cold front coming. Is it going to be enough for an Ice Wine pick?”

Field Agent: “The Farmer’s Almanac and local weather stations are aligning. We have a 70% probability of a 17°F dip on Tuesday night. To qualify for a certified Ice Wine harvest, we need that sustained freeze.”

Farmer: “What is the backup if the freeze is too shallow?”

Field Agent: “If we miss the freeze, our acidity is still high enough for a Late Harvest dessert wine. I have already updated the Digital Twin for Block #4 to include both pathways. I have also alerted the picking crew to be on standby for a 3:00 AM start on Wednesday.”

Beyond the Juice: The Zero-Waste Pivot

Resilience also means looking at what remains after the press. Once the juice is gone, you are left with the pomace, e.g. the skins and seeds. In a standard operation, this is waste. In a Sovereign system, the Scribe tracks this material as a new category of premium inventory.

By linking the life story of your estate-grown, high-provenance grapes to the resulting pomace, you open up high-value secondary markets from culinary grapeseed oil to nutrient-dense grape skin flour to distillery wash for regional grappa producers.

Because the Supply Chain Guardian has already verified the production standards and the cold chain of the original fruit, your waste product suddenly has a premium pedigree. The data you collected in July is still generating revenue in December.

Closing the Loop: The Material Memory

As the vineyard goes dormant and the winter rains begin, the Scribe’s job shifts again. We have moved from 1880 census records to the 2026 harvest. We have pivoted markets, guarded our provenance, and eliminated waste.

But there is one more piece of the farm’s infrastructure that holds a story: the Barrels.

As we pull old barrels out of rotation to make room for the new vintage, we are not just looking at bulky trash. We are looking at Material Memory. In our next series, we will step into the Salvaged Barrel Studio. We will use computer vision and the same Sovereign architecture to grade, reclaim, and upcycle the wood that held this year’s success.

The Agile Harvest Series

The harvest is in, but the work is just beginning. How are you capturing the final 5% of value in your business? Reach out on LinkedIn and let’s talk about building a resilient, zero-waste operation.

Facebooktwitterredditlinkedinmail