Interesting! Lightdash introduced data apps yesterday as well. The results look pretty similar. Personally, I’ve always had problems with data apps, because they were kind of messy and difficult to organise/keep updated. With a product like Hex or a lightdash semantic layer behind it, it could work better.
I didn’t find a blog post, but here is a YouTube video and the docs page.
tl;dr: Hex now allows you to use the full power and flexibility of AI code generation, while using the trusted data, context, and governance of Hex's platform.
---
We've been a customer for a number of years now and I personally think the way Hex is approaching the intersection of self-service, vibe-coding/AI and governance/maintainability is directionally really smart.
As an admin, being able to see what our team is asking and get feedback on where context is poor (unclear dbt models, not enough context/guides to disambiguate inconsistent language etc) has been awesome. Then this latest stuff effectively bringing a replit/v0/lovable codegen combined with the jupyter-like Python/SQL cells is a real cherry ontop to handle the "why can't we just connect claude cowork directly to Snowflake" question that comes up
Interesting! Lightdash introduced data apps yesterday as well. The results look pretty similar. Personally, I’ve always had problems with data apps, because they were kind of messy and difficult to organise/keep updated. With a product like Hex or a lightdash semantic layer behind it, it could work better.
I didn’t find a blog post, but here is a YouTube video and the docs page.
https://docs.lightdash.com/guides/data-apps
https://m.youtube.com/watch?v=cLLjfDAvEz8
tl;dr: Hex now allows you to use the full power and flexibility of AI code generation, while using the trusted data, context, and governance of Hex's platform.
---
We've been a customer for a number of years now and I personally think the way Hex is approaching the intersection of self-service, vibe-coding/AI and governance/maintainability is directionally really smart.
As an admin, being able to see what our team is asking and get feedback on where context is poor (unclear dbt models, not enough context/guides to disambiguate inconsistent language etc) has been awesome. Then this latest stuff effectively bringing a replit/v0/lovable codegen combined with the jupyter-like Python/SQL cells is a real cherry ontop to handle the "why can't we just connect claude cowork directly to Snowflake" question that comes up