请问HN:给人工智能代理提供数据库访问权限是新的商业智能工具问题吗?

2作者: vira283 个月前原帖
我正在咨询的一些早期到中期的初创公司都在问同一个问题:他们的人工智能/机器学习团队想要生产环境中的Postgres数据,但没有人确切知道该如何提供这些数据。 我之前为商业智能团队处理过类似的情况——使用读取副本,并设置较高的`max_standby_streaming_delay`和开启`hot_standby_feedback`,偶尔在主库上接受一些膨胀。这种方法效果不错。但对于人工智能/机器学习的需求,我感觉有些不同,虽然我还无法完全表述清楚,这也是我提出这个问题的原因之一。 我正在尝试调整几个方面: 1. 代理实际连接到哪里?是主库(带有行级安全性),读取副本,数据仓库(如Snowflake、BigQuery、Redshift),湖仓(如Iceberg、Delta在S3上),还是其他地方? 2. 如果你们<em>不</em>这样做——是出于合规性、成本担忧、糟糕的经历(如查询失控、提示中的个人身份信息)还是其他原因? 3. 我最感兴趣的是:这是否真的与给商业智能工具数据库访问权限的感觉不同,还是同样的问题换了个说法? 我并不寻求产品推荐,而是想从那些真正面临过这个挑战的人那里获得真实的见解。
查看原文
A couple of early-to-mid-stage startups I&#x27;m consulting with are asking the same question: their AI&#x2F;ML team wants production Postgres data, and nobody&#x27;s quite sure how to give it to them.<p>I&#x27;ve handled this before for BI teams — read replica with a generous `max_standby_streaming_delay` and `hot_standby_feedback` on, accepting the occasional bloat on the primary. Worked fine. But the AI&#x2F;ML ask feels different in ways I can&#x27;t fully articulate yet, which is part of why I&#x27;m asking.<p>A few things I&#x27;m trying to calibrate:<p>Where does the agent actually connect? Primary with RLS, read replica, warehouse (Snowflake&#x2F;BigQuery&#x2F;Redshift), lakehouse (Iceberg&#x2F;Delta on S3), or something else?<p>If you&#x27;re <i>not</i> doing this — is it compliance, cost fear, bad experiences (runaway queries, PII in prompts), or something else?<p>And the one I&#x27;m most curious about: does this actually feel different from giving BI tools DB access, or is it the same problem wearing new clothes?<p>Not looking for product recommendations. Trying to get a real sense from people who&#x27;ve actually faced this challenge.