LessWrong (Curated & Popular)
"The inaugural Redwood Research podcast" by Buck, ryan_greenblatt
27 Jan 2026
Chapter 1: What is the main topic discussed in this episode?
The Inaugural Redwood Research Podcast by Buck and Ryan Greenblatt. Published on January 4, 2026.
After five months of me, Buck, being slow at finishing up the editing on this, we're finally putting out our inaugural Redwood Research podcast. I think it came out pretty well. We discussed a bunch of interesting and under-discussed topics and I'm glad to have a public record of a bunch of stuff about our history. Tell your friends.
Whether we do another one depends on how useful people find this one. You can watch on YouTube here, or as a Substack podcast.
There's a YouTube video here in the text. Subheading.
Chapter 2: What challenges did Buck face during the editing process?
Notes on editing the podcast with Claude Code. Buck wrote this section. After the recording, we faced a problem. We had four hours of footage from our three cameras.
We wanted it to snazzily cut between shots depending on who was talking. But I don't truly in my heart believe that it's that important for the video editing to be that good, and I don't really like the idea of paying a video editor. But I also don't want to edit the 4 hours of video myself.
And it seemed to me that video editing software was generally not optimized for the kind of editing I wanted to do here, especially automatically cutting between different shots according to which speakers are talking. Surely, I decided, it wouldn't be that hard to just write some command-line video editing software from scratch, with the aid of my friend Claude. So that's what we did.
Chapter 3: How did Buck and Claude Code develop their own editing software?
We, which henceforth means me and Claude, first used DeepGram to make a transcript of the podcast that includes timestamps and a note of who's speaking. Then we generated IDs for all the different lines in the transcript, leading to a huge file that looks like this. There's a code block here in the text. We wrote code that lets us edit the podcast by copying and pasting those lines around.
We also wrote code that automatically generates cuts between the different shots using crappy heuristics that Rob Miles tells me are bad. The DSL compiles to a giant from PEG command, which is then executed to produce the final product.
This process produced the video and audio for almost all of that podcast, with the exception of the intro, which Rob Miles kindly sat down with me to manually edit in a real video editor. Things I learned from this process Claude Code had a lot of trouble getting from Peg to work. This makes sense, it seems pretty confusing and Claude wasn't easily able to check its work.
In the same spirit as the many American men who believe they can beat a bear in a fight, I often have the delusional belief that instead of learning existing software, I should write my own software to do the same thing.
Chapter 4: What insights did Buck gain from creating custom software?
AI agents are now good enough that this is not totally impractical. Video files, especially really long ones like these, are really large. This means that lots of things you want to do are constrained by disk space, network speed, and computational power.
I ended up doing all the work over SSH to my beefy desktop. Heading.
Chapter 5: What are the implications of AI in video editing discussed in this episode?
Podcast transcript.
The podcast transcript is omitted from this narration. This article was narrated by Type 3 Audio for less wrong. It was published on January 4, 2026.