Underlord Adds Context-Aware Skills for Better Captions (Full Transcript)

[00:00:00] Speaker 1: What else is coming that you're excited about? It's related to context, actually. We want Underlord to basically understand when it needs to use a particular set of steps and preferences and workflows to accomplish a goal, depending on what the context of the goal is. To give you an example, there might be a caption style that works well with long-form podcasts, and it's probably subtle. It's an accessibility thing. But there's a completely different kind of caption style for a clip. You might want to be one word at a time, very punchy. And so if you're creating clips, you might want to pull a skill for clip creation. That's actually very different than a skill for long-form podcast creation. And this idea of splitting that context across use cases to get it actually in each use case without confusing it for any other use case is a thing that we're investing in. It's amazing.