Greedy Shortcut Model
一位开发者创建了一个简单的“bandit model”来测试用户在新应用中的行为。该模型旨在强化使用频率更高的选项，而非评估哪个选项更好。开发者指出，用户倾向于选择他们熟悉的功能，而非最佳功能，并正在寻求关于这一观察和模型的反馈。
- 发布
- 09/07 07:07 UTC+0
- 收录
- 09/08 00:00 UTC+0
- 来源类型
- 开发者社区
- 档位
- 社区
- 信源状态
- 正常
档位是按信源手工设定的编辑判断，不是逐条打分。
Not quite for this channel, but relevant to choosing new-age apps: users don’t pick the best feature; they pick what they know.
I built a simple bandit model to test whether usage converges on quality over time. It doesn’t estimate which option is better. It simply reinforces whatever gets used more.
Three update rules produce very different outcomes: one locks onto early winners regardless of quality; one mostly self-corrects but can still get stuck under strong reinforcement; and one control always finds the true best option.
Same mechanism, wildly different UX. Basically, habit formation in miniature.
Feedback welcome :)