Battle of the NLE AI Assistants 1
Support ProVideo Coalition
Shop with Filmtools

What Do You Think? Let Us Know.

Subscribe
Notify of
guest
2 Comments
Oldest
Newest Most Voted
Erik

Great write-up, and the 2:31 vs 30 seconds stringout is the most honest number I’ve seen in one of these tests.

Disclosure: I build an AI assistant editor for Final Cut (Cutler), so I’m not in this particular fight, but your results match what I keep seeing. Mechanical bin work is where an experienced editor with shortcuts wins every time. Where the AI pays back is the part that takes hours: watching everything and choosing takes. And “use the better take” from a transcript alone can’t see a blink, a soft focus or a bad frame, so it has to look at the picture too.

Your line about not wanting an AI editing “based on a feeling” is the key one for me. My answer has been to make it write down a reason for every clip, so you can disagree and overrule in seconds instead of reverse-engineering what it did.

On cost, agreed, it should be visible. When I measured it on the API path, the visual pass on a 300+ clip project came to about $0.38 and the reasoning about $1 per cut after a few rounds of notes. On a subscription plan you only see a percentage, which is exactly your point.

Looking forward to part two.

You Might Also Like