Memory and reflection, in brief

About this track: Clawdemy uses the open-source TradingAgents framework (by TauricResearch; Yijia Xiao, Edward Sun, Di Luo, and Wei Wang; arXiv:2412.20138; Apache-2.0) as a real-world example of how an agentic AI system is structured. Our goal is to teach the architecture: how specialist agents, a research debate, a trader role, and a risk layer are coordinated by an orchestrator. We are not teaching you how to invest, trade, or pick stocks, and we make no claim that this or any AI system is profitable or predicts markets. This content is for education only and is not investment, financial, or trading advice. Rules vary by country, and this is not advice anywhere. The Trading Agents Lab capstone is a simulation you explore yourself with your own AI provider key; it does not place real trades. The framework’s own authors state it plainly: “It is not intended as financial, investment, or trading advice.”

What you’ll learn

A decision is not the end of a good system; the best teams learn from how their calls turn out. This lesson reads the last loop of the framework: memory and reflection. The key insight is that the work is split across time. When a run finishes, the system records the decision in a plain logbook and marks it not-yet-judged, with no model involved, because you cannot grade a call before you know how it went. Later, on the next run for the same company, the system looks back, measures what actually happened over the following days against a broad market baseline, and writes a short, honest reflection on the cheaper model: was the call right, what held or failed, one lesson for next time. That history is then fed back into the next run, and only the final gate (the portfolio manager) reads it, so memory lands exactly where the system commits. The throughline is transferable: record now, judge later against a real outcome, keep the lesson short, and put it where the next decision is made.

Where this fits

This is the seventh lesson of the track and the last of the architecture walk. Lessons 1 through 6 built the team and the wiring that connects it; this lesson closes the loop by showing how the system learns from its own track record. Next, the track finishes with the hands-on capstone, where you run the whole system yourself in a safe simulation.

Before you start

Prerequisites: lesson 6 of this track (orchestration and shared state), since the memory context is read into that shared state for the final gate. You do not need to install or run anything; the project can be read in your browser, and every excerpt is quoted in the lesson.

By the end, you’ll be able to

Explain why a decision can only be judged once its outcome is known
Describe the two-phase memory loop (record immediately, reflect on a later run)
Identify how reflection is grounded in a real outcome against a market baseline
Recognize that only the final gate receives the accumulated memory
Apply the record-then-reflect pattern to your own AI workflows

Time and difficulty

Read time: about 10 minutes
Practice time: about 15 minutes (a self-check, a design exercise, and flashcards)
Difficulty: deep (this track is the advanced end of the catalog; the foundational agent tracks are the on-ramp)