问HN：‘自然语言分析’是未解决的问题还是只是不够吸引人？
像Looker、Hex和ThoughtSpot这样的工具多年来一直在努力让非技术用户能够使用分析功能。大型语言模型（LLMs）重新引发了这个话题的讨论。但大多数新一波产品仍然要求用户首先拥有数据仓库——这意味着它们的目标客户是数据团队，而不是运营人员。
我关注的是一个不同的切入点：既做数据仓库，又做查询层，目标是那些从未拥有过这两者的用户。直接与Stripe、Shopify、Meta广告、HubSpot等进行同步，并通过自然语言展示所有内容。
我很好奇HN（Hacker News）对此的看法：真正的障碍是数据仓库的搭建、查询的复杂性，还是对输出结果的信任？你认为是否有某个用户群体是真正受益于此的？
查看原文
Tools like Looker, Hex, and ThoughtSpot have been trying to make analytics accessible to non-technical users for years. LLMs have reopened that conversation. But most of the new wave of products still require you to have a warehouse first — which means they're selling to data teams, not operators.<p>I'm looking at a different wedge: be the warehouse AND the query layer, and target the segment that has never had either. Sync Stripe, Shopify, Meta Ads, HubSpot, etc. directly, and expose everything through natural language.<p>Curious what HN thinks: is the real barrier warehouse setup, query complexity, or trust in the output? And is there a segment of user you think this genuinely serves?