AI PROMPT ·
Flyctor 机器人进化模拟
为 Flyctor 机器人实现一条简单规则：每当它在不碰撞任何物体的情况下到达目标时，就强化导致其成功的行为；每当它陷入困境、撞上家具或原地打转时，就降低该行为的权重。使用 GPT-6 Astra 将视频转化为简单的感知信号，并利用重构的苍蝇大脑模拟来选择下一步动作。
PROMPT
Implement a simple rule for the Flyctor robot: Every time it reaches a target without hitting anything, reinforce the behavior that got it there. Every time it gets stuck, crashes into furniture, or drives in circles, reduce the weight of that behavior. Use GPT-6 Astra to turn video into simple sensory signals and a reconstructed fly-brain simulation to choose the next movement.