Jev Edited This Video

A new AI model just launched. It's 400 times cheaper than Frontier AI models. It's called Jev. And what this means is that previous tokenheavy automations can now be run for much cheaper and much faster. I wanted to try this out on my existing video editing pipeline. I wanted to see how faster it could get if

I implement Jev into the mix. Hi, my name is Paul. I built aic systems just like these for my clients. And today I'm quite thrilled to show you how Jev improved my video editing pipeline and not only made it faster but also made it super cheap. Okay, so let's jump into the explanation. So what's the

difference between a normal LLM and Jeff? With normal LLMs, we'd give an input of type text and our LLM would spit out text back. Jeff is different. We give a textbased input but the output that comes out of Jeff is actually a JSON object. Now what's a JSON object? A JSON object is basically in simple terms

a list of labels. Now those labels help us classify that input and that's what we need. Imagine we have a huge timeline of a video and there's multiple takes. For example, this one specifically, I already had like maybe seven takes. And I want to classify these really easily by deciding which ones will make the cut

because I maybe stopped cuz I made a mistake or I didn't finish the whole sentence or I added a bunch of filler words. I will use Jev to classify all of these or make shortcuts in between to make it more efficient. So what does Jev change? Jev is a different type of AI

model. So, we can't compare it to anything else yet, but it's a gamecher because it's much cheaper because of that output I just told you about. And it's also super fast because it doesn't have those processes where it thinks a lot or requires to understand the context of the input. We don't need to