Part 3/16:
Cognitive architectures are particularly valuable in controlling autonomous or semi-autonomous agents—ranging from robots to virtual entities like NPCs in video games. Shapiro points out Mars rovers as classic examples. These robots operate with a degree of autonomy, performing planning, reasoning, and problem-solving based solely on sensor input, memory, and self-directed decision-making.
Similarly, he discusses rockets, which utilize telemetry to make real-time hypotheses and optimizations during spaceflight. While some may classify these as feedback systems rather than true cognitive architectures, their autonomous decision-making processes align with the foundational principles.