The presenter sets up Claude Opus 4.8 to test two trading workflows with timed monitoring, using the same prompts as in earlier experiments. The one-hour window is a snapshot rather than a controlled comparisonEvaluation is the systematic assessment of a system's behaviour or results against explicit criteria..
The presenter reports different outcomes in the two accounts and says market conditions could explain the difference. He also describes having to intervene repeatedly when the agent stopped before the intended end of the test.
The presenter compares the experience with an earlier workflow and says further testing is needed. The video is a demonstration of agent behavior, not evidence that a trading strategy will be profitable.
Watch on YouTube




