We Paid Professional Editors to Tell Us Our Product Sucked. Here's What Happened.
Cutback launched Selects by handing it to editors from Beast Games and George Janko, asking them to be brutal. This is what they said.
TLDR: When Cutback launched Selects, we didn't ask whether professional editors would like it; we paid them to tell us if it sucked.
When Cutback launched Selects, we didn't ask whether professional editors would like it; we paid them to tell us if it sucked.
Most AI video tools are launched into the consumer market, gather five-star reviews from people who make one video a week, and declare themselves revolutionary. We made a different call. We handed Selects to editors who cut Beast Games, and George Janko, and told them, directly, to tear it apart.
This was not a soft launch. It was the biggest marketing spend in Cutback's history, and it was designed around a question we couldn't fake the answer to: Does this actually work for people who have been editing professionally for decades?
Check out the full video below:
The Problem With AI Video Tools
There is a very specific way professional editors react when you show them a new AI editing tool. Their eyes glaze over. They've seen a hundred of these. Most tools that claim to automate editing can generate a short social clip, add auto-captions, or remove silence at a basic level, useful for someone running a podcast from their bedroom, less useful for a team cutting a major streaming production or a high-volume documentary series.
What they actually need is a tool that understands footage the way they do. One that recognizes when a speaker is repeating themselves across takes, builds a storyline that reflects how the interview actually flowed, and handles the mechanical preparation work, syncing cameras, building stringouts, pulling selects, without requiring the editor to babysit it. That gap between what AI tools promise and what they actually do for working professionals has made the industry's best editors deeply, reasonably skeptical.
Tom Kim, Cutback's co-founder and CEO, knew this going in. Professional editors, he noted, have been working with the same tools for decades. They'll abandon a new product over a single extra click, a hotkey in the wrong place, a sync that drops a frame. The bar is not "work most of the time." The bar is "earns a place in the edit suite."
What We Built, and How We Tested It
Selects is a pre-editing tool for long-form production, not a clip generator. It ingests raw footage, transcribes speech at word-level accuracy within 9 milliseconds, identifies speakers through diarization at 97-98% accuracy, detects topics and subtopics across the entire project, analyzes visual content to detect scenes and camera angles, and builds a structured stringout with the best takes already placed before an editor opens an NLE. The output hands off directly to Premiere Pro, Final Cut Pro, or DaVinci Resolve.
The theory was that Selects could compress the most time-intensive and least creative part of the edit, everything that happens before the first real cut, into something that runs in the background while the human editor focuses on storytelling.
The test was whether professional editors, given the product with no guidance and no cheerleading, would reach the same conclusion.
We ran a campaign we internally called the roast. We gave Selects to editors whose credits include some of the biggest productions in the industry. We asked them to be honest. We told them to say if it sucked. We filmed it.
What The Pros Had To Say
Rez Karim, a professional editor, put it in terms that land precisely:
"Most AI tools require users to constantly "tweak a setting" and "fight the software." Selects "handles the boring prep in the background," lining up A-roll while the human editor focuses on the creative work. That is the distinction. Not that the AI replaces the editor, but it removes the hours of mechanical labor before the first real cut so the editor can start at the interesting part.
Nas, an AI tech specialist who has seen every tool that has come across the category, said:
"First AI editing tool I've come across that feels like it was actually built by people who edit."
That line is the one we care most about. Not because it is flattery, but because it describes the design decision behind everything in Selects. The syncing, the speaker recognition from a single audio file, the topic detection that groups related content across a two-hour recording, those features exist because editors told us, repeatedly, that those are the problems that consume their days. We built toward the friction, not away from it.
One editor described how Selects would "integrate so well into our workflow." Another noted that an assistant editor could complete "6 episodes in a day" using the tool. Sam Wasserman described the potential for AI to handle "early assemblies and strong outs," freeing human editors for the work they actually want to do.
The throughput numbers reflect this. Cutback's data shows approximately 60% less editing time per project on average. One user noted that a 37-minute podcast project that used to feel like their "biggest video" now represented work they could triple in volume without adding headcount.
What This Means for the Industry
There is a version of AI in post-production that replaces the editor. That is not what Selects is. Selects shows you what AI in post-production actually replaces and is designed around the premise that the most valuable thing a human editor does, narrative judgment, pacing instinct, the decision to stay on a speaker's face for two seconds longer because something is about to land, is not automatable and should not be. What is automatable is everything that happens before that.
The result is a shift in where editors spend their time. The prep work that previously defined the first half of any long-form project now happens while the editor is doing something else. The creative window opens earlier. The output volume increases without the burnout that typically comes with it.
We think that is what a useful AI tool looks like in a professional context. Not a replacement, but a partner in the prep so that the real editing can begin effortlessly.
Frequently Asked Questions (FAQs)
Q: What is Selects by Cutback?
A: Selects is a pre-editing AI tool built for professional long-form video production. It ingests raw multicam footage, transcribes at word-level accuracy, identifies speakers, detects topics, pulls the best takes, and builds a structured rough cut before the editor opens Premiere Pro, Final Cut Pro, or DaVinci Resolve. It is designed for editors who work on podcasts, interview series, documentary content, and high-volume YouTube production.
Q: Is Selects used by professional editors?
A: Yes. During its launch, Selects was tested by editors whose credits include Beast Games and George Janko, to name a couple. The tool was specifically designed for professional post-production workflows rather than consumer or casual content creation use cases.
Q: How much time does Selects save in post-production?
A: Cutback's data shows approximately 60% less editing time per project on average. The reduction comes primarily from automating the pre-editing stage, syncing cameras, transcribing footage, building stringouts, and pulling best takes, before the editor begins working in the NLE.
Q: Does Selects replace video editors?
A: No. Selects automates the mechanical preparation stage of editing, syncing, transcription, topic organization, and rough cut assembly, so human editors can focus on narrative judgment, pacing, and creative decisions. The tool is designed to expand what editors can produce in the same amount of time, not to eliminate the editorial role.
Cutback Team
Share post
