一位开发者在 Hacker News [1] 上发布了一款名为 Show HN 的免费应用程序，旨在帮助纽约居民通过自动堆叠优惠券、现金返还和返利来降低杂货支出。该应用目前仅覆盖纽约市约 690 家商店，且无需用户登录即可使用 [1]。此外，该软件集成了基于训练过的 Llama 模型的 AI 购物助手功能 [1]。开发者同时向社区提出一个问题：在处理杂乱的多源零售定价数据时，应优先保证数据的时效性还是覆盖率？[1]
A developer has released a free application designed to help residents of New York City save money on groceries by automatically stacking coupons and cash-back offers [1]. The app, titled "Show HN," is currently available without requiring user login credentials [1]. It covers approximately 690 stores within the city limits [1] and features an AI shopping assistant powered by a trained Llama model to assist users during their shopping process [1]. In addition to these functionalities, the developer has posed a question to the community regarding data management strategies for messy, multi-source retail pricing information: whether freshness or coverage should be prioritized [1].