Speaker unverified: evaluation

Transcript context

…Yeah, like evals or Lindys, there's two kinds of evals, right? Like in some way, you don't have to eval your system as much because you've constrained the language model so much. And you can rely on open AI to guarantee that the structured outputs are going to be good, right? We had Michelle sit where you sit and she explained exactly how they do constraint grammar sampling and all that good stuff. So actually, I think it's more important for your customers to eval their Lindys than you evaling your Lindy platform because you just built the platform. You don't actually need to eval that much. Yeah. In an ideal world, our customers don't need to care about this. And I think the bar is not like, look, it needs to be at 100%. I think the bar is it needs to be better than a human. And for most use cases we serve today, it is better than a human, especially if you put it on Rails. Is there a limiting factor of Lindy at the business? Like, is it adding new connectors? Is it adding new node types? Like how do you prioritize what is the most impactful to your company?…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.