HackerNews中文版

我们创建了SocialCraft AI，旨在利用指数衰减模型来模拟LinkedIn连接的衰减，其中关系强度每90天减少50%。我们识别出的问题是，许多专业人士在LinkedIn上难以维持有意义的连接，这导致参与度下降和机会的错失。例如，如果您在180天内没有与某个连接互动，其关系强度将降至原始值的25%。这种衰减会显著影响网络效能和内容可见性。在技术上，我们实施了互惠账本和互动速度评分，以量化和跟踪随时间变化的互动情况。关键的决策是使用指数衰减模型，这使我们能够根据最后一次互动日期预测连接强度。这种方法相比于简单的指标（如连接数量或参与率）提供了更细致的关系动态理解。我们与其他社交媒体分析工具的不同之处在于，我们专注于关系的数学建模，而不仅仅是表层的参与指标。虽然像LinkedIn Analytics这样的工具提供了帖子表现的洞察，但它们往往忽视了驱动参与的潜在关系动态。我们的模型帮助用户理解何时重新与连接互动以维持其强度。目前，我们处于测试阶段，并根据用户反馈不断完善我们的算法。一些局限性包括需要更广泛的数据来提高准确性，以及整合LinkedIn以外的其他社交平台。我们还在努力增强我们的内容生成能力，以更好地与我们推荐的参与策略对齐。您可以在 https://social-craft-ai.vercel.app 免费试用SocialCraft AI。我们欢迎任何技术问题或对我们方法的反馈！

查看原文

We built SocialCraft AI to model the decay of LinkedIn connections using an exponential decay model, where relationship strength decreases by 50% every 90 days.The problem we identified is that many professionals struggle to maintain meaningful connections on LinkedIn, leading to a decline in engagement and missed opportunities. For instance, if you haven't interacted with a connection in 180 days, their relationship strength drops to just 25% of its original value. This decay can significantly impact networking effectiveness and content visibility.Technically, we implemented a reciprocity ledger and engagement velocity scoring to quantify and track interactions over time. The key decision was to use an exponential decay model, which allows us to predict connection strength based on the last interaction date. This approach provides a more nuanced understanding of relationship dynamics compared to simple metrics like connection count or engagement rate.What sets us apart from other social media analytics tools is our focus on the mathematical modeling of relationships rather than just surface-level engagement metrics. While tools like LinkedIn Analytics provide insights into post performance, they often overlook the underlying relationship dynamics that drive engagement. Our model helps users understand when to re-engage connections to maintain their strength.Currently, we are in beta and are continuously refining our algorithms based on user feedback. Some limitations include the need for more extensive data to improve accuracy and the integration of additional social platforms beyond LinkedIn. We are also working on enhancing our content generation capabilities to better align with the engagement strategies we recommend.You can try SocialCraft AI for free at https://social-craft-ai.vercel.app. We invite any technical questions or feedback on our approach!

社交工艺人工智能——你的LinkedIn网络连接得怎么样？