0:00Sam Nadler says he built a tool that lets recruiters with no video editing experience cut up candidate videos themselves and generate links to share with clients, built with a product called Remotion.
2:01Sam Nadler says clients usually get a CV, rate and experience, but a minute of video shows how a candidate speaks and thinks. Editing used to require video editors and took too long to do routinely.
4:26Sam Nadler says the recruiter uploads the video and pastes the Ntrvsta transcript. AI finds nine clips across English communication, role fit and personality and drive, and the recruiter trims about three into a reel under one minute.
7:14Sam Nadler says rendering was the hardest part and still takes five to ten minutes, versus hours or days with a video editor, and outputs an S3 link to share. He built it mainly in Codex using Remotion.
9:00Jordan Metzner calls Remotion a JavaScript video library in React with a player and editor. Sam Nadler says he shaped his template in natural language with Codex and Claude Code, with zero editing skills.
12:02Jordan Metzner says he makes product feature trailers synced to music, turned a 65-page slide deck into a narrated video with Remotion and ElevenLabs, and has seen Remotion used for automated documentation.
13:09Jordan Metzner says servers left on led to a $35,000 bill and one developer using a new Cursor feature cost $2,000. He expects efficiency scores comparing token use with each worker's output.
16:56Sam Nadler raises the argument that heavy AI use could cost as much as labor. Jordan Metzner says running agents is not free, people are buying Mac Minis to run more of them, and it is too early to see the payoff.
Remotion Video Editing as Code: An AI Highlight Reel Builder: the key questions from this episode, answered with quotes and timestamps.
Sam Nadler [0:00]: So I created a little tool that gave our recruiters who, I'm assuming, most of them don't have any video editing experience, the ability to cut up these videos and be completely self sufficient, generate links to share with our clients.
[0:15] Theme song: Built this week, breaking it down. Built this week, we show you how. A fresh idea, a clever tweak, you lock in true. Built this week.
Sam Nadler [0:25]: Hey, everyone, and welcome to Built This Week, the podcast where we share what we're building, how we're building it, and what it means for the world of AI and startups. I'm Sam Nadler, cofounder here at Ryz Labs, and each and every week, I'm joined by my friend, cohost, business partner, Jordan Metzner. How are doing today, Jordan?
Jordan Metzner [0:42]: Hey, Sam. Happy to be back. Another fun, exciting episode. So much going on in the world of AI, and this week is nothing different. New models coming out. New exciting news from all the Foundation Labs, and some news from NVIDIA as well. So, yeah, it's been a crazy exciting week just like each week before has been in the AI race.
Sam Nadler [1:03]: Absolutely. Every week, there's something new to talk about, new releases. Before we get into the agenda, don't forget to like and subscribe. Hit that button. We have new episodes out every single Friday, so just stay with us and see what we're building, see what products and tools are out there. And this week, we're gonna cover a tool I built for our recruiting team that I think is fun to highlight, really kinda gives them some video editing power in that that they can, be self sufficient and no longer rely on our video editors to cut up little videos. I'll do a demo of that tool. Then we're going to jump into how I built that tool. There's a product out there called Remotion that we've started to use that will be fun to cover, and then finally, we'll cover the latest and greatest news. Does that sound like a plan?
Jordan Metzner [1:58]: Awesome. Yeah. Let's go. Let's jump into it.
Sam Nadler [2:00]: Alright. So just to jump in really quick, a little bit of context. As you know, at Ryz Labs, we help source and provide great candidates of all different types to our clients. Those candidates can be from an engineering background, a data background, project management, product management, whatever, you name it. We are interviewing and sourcing hundreds of thousands of candidates. One little kind of element that we liked to provide some of our clients when it made sense is a little sizzle reel of our of the candidate we're presenting. So, in general, when a client receives a list of potential candidates, they get their CV, the standard, their CV, etcetera, potentially their rate, their experience, their studies. All that is pretty typical. However, wouldn't it be nice to be able to see that candidate, to see them speak, to see their thoughts on certain problems, and just to give you get a little better interpretation or impression of that candidate if you could see a minute of them before you decide whether to interview them or not. So we've done this occasionally when we had, like, a pool of candidates that we wanted a client to review.
Sam Nadler [3:21]: However, the process took a really long time. We had to, leverage our video editors to edit the videos, and it just wasn't something that, we could do all the time. So I created a little tool that a lot gave our recruiters, who, I'm assuming most of them don't have any video editing experience, the ability to cut up these videos and be completely self sufficient, generate links to share with our clients. So, this is all theoretically powered by our Ntrvsta tool, which performs, AI interviews, both technical and nontechnical of candidates. So that's like, we have these videos of these candidates. This product just takes those videos and packages them up into a branded sizzle reel. So it's not fancy, but it puts the power of the video editing directly into our recruiters' hands. So as they prepare these candidate lists, then they can just ship off these videos when they feel it's appropriate. So it's just as simple as I hope you can see my screen. This highlight reel builder.
Sam Nadler [4:26]: I know I'm not that great at creating names, but you basically drop in a video. I have one prepared here. And as soon as it loads, this is a candidate that took an Ntrvsta interview within the last forty eight hours. You just hit next. You fill in the candidate's name, and then you fill in the title. And then you take the transcript, which from the Ntrvsta interview, and you paste it right in here. And this was all my design. Not sure if it's the best design. The beauty of putting in the transcript is with AI, it will analyze the transcript and identify their best responses, their best little sizzle clips. So I say find best clips, and it'll give me nine clips theoretically based in three different categories, and there's probably a few different ways I could have thought about this, but this is kinda how I approached it. So as you can see, it already generated those clips for me. So the first cohort of clips is on English communication. The second cohort of clips is on role fit, and the fourth I'm sorry. Third cohort of clips is on personality and drive. And as you can see, there's, like, this wave format.
Sam Nadler [5:40]: There's probably an even more automated way of doing this, but I wanted to provide some editing tools for our recruiters to really identify that clip of the candidate that made the most sense, that put them in the best light, that really highlighted either their communication skills, their technical skills, the project management skills, whatever. So you have these little clips, these little editing things you can work with here. But, basically, the recruiter, puts in the transcript, then can listen to these clips, adjust them a little bit so they don't end on an, or something like that. And then once they find kind of the three best clips and you could add more. You could make it six, but generally speaking, I'm trying to keep these two under one minute. As you can see here, it's forty three seconds, and I've selected three clips. So, generally speaking, three clips is around fifteen seconds each. Maybe you could get away with a fourth one. For the purposes of this demo, I'm not gonna go through each clip and listen to it, but you just hit next. You get a preview so you can see, the candidate name, the role.
Sam Nadler [6:47]: You could choose the aspect ratio. 90% of the time, this is for, a link within an email, so landscape feels appropriate. You can do a preview. Now the video won't be rendered in the preview, but you could theoretically see the graphics and the audio just to make sure it flows naturally. And then next, and then you can just get a render settings and then begin to hit the render of the video. Now this is, for me, one of in the development was one of the most challenging parts was to get the video to render in a reasonable amount of time. It still takes probably five to ten minutes, so it's not that fast. But it's, theoretically way faster than engaging with the video editor, giving them all the context, waiting for them to turn it around based on their workflow. So five to ten minutes versus several hours, if not days, feels like a reasonable compromise. This same exact candidate, I had already pre prepared so we don't have to wait, and it comes with a link that you could theoretically share, an s three link you could theoretically share directly with the candidate or client.
Sam Nadler [7:56]: I'm sorry. So imagine a client is just receiving an email with three candidates, all senior data analysts. They get their CVs and three links to their video, and it looks like this.
Candidate [8:12]: And as a data analyst, what I've been doing is basically taking care of the whole process from data extraction, so ETL, to modeling and also visualization.
Sam Nadler [8:28]: Perfect. So as you can see, it's a little clip of this candidate. And, we think, hopefully, it will help increase the step from, presenting to the client to getting to client interviews. This was mainly built in Codex using a tool that we're about to dive into called Remotion. So before we go into Remotion, Jordan, any feedback on how to make this better? Anything I missed? Or just general advice?
Jordan Metzner [9:00]: No. It's pretty awesome. I mean, I kinda feel like we're giving away a little too much secret sauce, to be honest. I don't know. Yeah. I mean, I think we all agree that, like, video is a great format for understanding candidates and, like, reviewing them and whatnot. And then, as you said, kind of historically, editing video was left for still being done exclusively, kind of offline via heavy applications like, Premiere or Final Cut or iMovie or something else like that. And, then went to mobile with like CapCut, etcetera. But, this is all video editing as code, which makes it like so much more interesting, especially kind of the possibilities of like kind of creating infinite amounts of video slates. So, yeah, let's jump into like what is Remotion and how you use this tool.
Sam Nadler [9:40]: Yeah. So I think, you told me about this tool and just like, plugging it into Codex. And I actually think I started with Claude Code. Was really, really simple. And then it was just finding just toying around with, my own feedback on how I wanted the video to look. There was absolutely zero editing skills, but just like, playing around with the colors and how I wanted the graphics to be presented just in natural language, he was able to take that and help me build this template. But when did you learn about Remotion, and how have you used it in the last couple months?
Jordan Metzner [10:15]: Yeah. So like, I mean, I think of Remotion as kind of like a JavaScript video library. And, you know, before AI, using a JavaScript based video library might be like pretty cumbersome to kind of like write out in JavaScript a bunch of things you want a video to be. And then, yeah, I found out about it through like maybe like in the vibe coding, early vibe code days of like Lovable and Bolt or something like that and someone was using it as a video editing tool. And I just kinda saw the power of it. And it's interesting that there's not really that much other competition, but yeah, I think of it as like a video editing or video creation library, but if you go up to the products page, can see that it has kind of a bunch of other features built into it, like a player. It has an editor, if you saw that up there on top. Yeah. So it has its own editor tool that you can kind of build on top of. And it's really got a bunch of really cool dynamic features all kind of, in React that you can integrate into your app.
Jordan Metzner [11:07]: So in the case of Ntrvsta, we have this awesome video asset. And so I think you've done a great job of kind of leveraging Remotion to make these kind of reels if you might wanna call them that. But, yeah, it seems like it's just the early days of Remotion, early days of kind of video creation via code, and, I think, like, if Remotion gets a little bit better, your prompts get a little bit better, I think your videos will get significantly better. And it kinda goes all way down, but, yeah, pretty cool. Great job.
Sam Nadler [12:02]: Cool. And how have you I know you've toyed around with it. Can you give me a couple use cases outside of the one we discussed?
Jordan Metzner [12:08]: Yeah. So when I build, like, a new product feature, I've been trying to make, like, Remotion video trailers, so that's been really cool and been pretty effective. I usually am able to find I'll just grab like some music from our royalty free library and then just use that as the backdrop of the video, and I'll use the waves file of the music so that it'll hit in the video and make these kind of like really cool trailers. And what else? Recently, I did this video where I took like a 65 page slide deck and turned it into a video using a Remotion and ElevenLabs for text to speech, and just kind of made it into like a narrated kind of video. Yeah. We've I've seen people use it for kind of like automated documentation solutions. So like, when you build something, you just like have the video go straight into your documentation settings. But yeah, it's really cool. It makes really nice animated videos.
Jordan Metzner [12:58]: It's really fun to play with, and I think we'll just, we're just gonna continue to keep using Remotion and build more and more features honestly because it's just it's really fun to use and it builds really cool stuff.
Sam Nadler [13:08]: Awesome. Well, that queues us up for the news. So let me this is kinda top of mind for us. So we're on internally here on a big push to pretty much ramp everyone in the organization on AI regardless of job title, regardless technical or nontechnical. We've got multiple people pursuing, I wouldn't even say side projects, but just for the sake of this conversation, like AI driven side projects, if their role isn't inherently technical, just on a really big push to ramp our entire organization on being very AI savvy. And this article stood out to us. You finally figured out AI at work. Now here comes the bill. AI output, which has turbocharged productivity and promises to change the nature of work, might seem like it materializes out of thin air, but it's really the work of data centers churning through prompts and interpreting them in elaborate and expensive, if unseen, process.
Jordan Metzner [14:04]: Yeah. So, basically, the real question, now we're giving everyone access to all of these tools. I think, we've given everyone access to Claude Code, Claude Cowork. We have people using Codex. We've had ChatGPT for a while now and other models. I think everyone has Gemini access, but, not all the models are the same and not the cost of use of all these models are the same. And, currently, like, we're just kind of telling everyone just like use whatever you want, whatever model you want, but, optimizing your model used for output will be important if you kind of see that your spend gets out of hand and like, we've seen kind of leaving servers on accidentally, how that left us with like a $35,000 bill. We saw recently how one developer used a new feature in Cursor and left us with like a $2,000 bill. So, this has hit us like in the bottom line for sure. And, I think both, these companies as well as ourselves are putting in like kind of more guardrails.
Jordan Metzner [15:01]: But, at the end of the day, I think like you're gonna look at token input and output and then output of the worker and, kind of get an efficiency score of kind of like is this worker like, using these additional software pieces in a way that provides, good value for the organization? Or, are they just wasting money kind of thing? And that's that's an interesting discussion and a new format and a new way to manage, I think.
Sam Nadler [15:25]: Yeah. Totally. I mean, I don't think we I think you mentioned that we've had two little surprise two well, a big surprise and a little surprise. But, there is going to be it's going to be managed both, I think, how an employee uses AI, both on, like, are they adopting? Are they maximizing their time on the job? Are they, like, driving things forward that, could be faster with AI? And also, how much they're spending doing it. Are they being reckless? And, I think, like, our internal organization, we haven't gotten to the, throttling or managing what tools certain teammates get based on, their role or what models they have access to yet, but I couldn't you know, we're one big bill away from it being an important part of how we provide access.
Jordan Metzner [16:14]: Yeah. I've heard of large organizations giving users access to using LLMs and coding, but only a certain amount of tokens. And so they only use it kind of, like, selectively as an example rather than just using it going all out. And so, yeah, I think, like, kind of cost consciousness, obviously, open source is exciting in this area where if you can host your own models and you can get those kind of on, old servers or servers you already have, then you maybe can save money in that way and kind of drive up efficiency and allow everybody to become more kind of token consuming. But, and then at the same time, obviously, these models are getting cheaper. Right? So, using what previous model gets you today for the same model, you get just way more output, right, and way higher quality as well.
Sam Nadler [16:56]: I feel like I've heard a recent argument that the cost of these tokens is actually one of the I don't know if I agree with this, but it's one of the best arguments that this, like, potential impending, major disruption to the job market won't actually happen? Because if we all switch to AI and we're using AI so much, it's it could be equal or, more costly in terms of token than, like, just, like, the labor costs. Have you heard similar arguments?
Jordan Metzner [17:27]: No. Not exactly. But I think, like, it's obvious that everyone's getting more and more, like, agentic and having, like, more assistance work for them. And, obviously, the cost of running these agents is not free. We talked about OpenClaw in previous episodes, but, OpenClaws having OpenClaws and whether, you know, I just heard of a friend buying a Mac Mini and then Mac Ultra and, people just trying to get, more servers basically to run more agents. And, I think that'll spike up and probably then come back down again. And, it only comes at the cost of whether or not like you're getting better output or, kind of like, more efficiency or whatnot. So, if everyone's spending money on all these tokens, like what are you getting for that? Is like your roadmap moving faster? Are you like, replacing the need for more engineers? Are you making more money? Are you saving money? So, I think it's still too early to see kind of, like, what the shakeout is of all this.
Sam Nadler [18:18]: Totally agree. Well, great little episode, Jordan. Thanks for watching my demo of the highlight sizzle reel that I built for our recruiters. Thanks for covering Remotion. Any final thoughts before we close out the episode?
Jordan Metzner [18:33]: No. Great episode. Another exciting week. I think, like, it seems that, like, more and more people are building at, all different levels, and these tools are getting more and more accessible, think, through Claude Cowork, through, the deal that we saw last week with Microsoft doing a deal with Anthropic to get, copilot, co work happening, and it's just exciting that, kind of the more people get access to these tools, the more democratizing it becomes just like, the Internet has been and smartphones have been as well. So but, yeah, it'll be exciting to see what comes next week, and looking forward to next week's episode as well, Sam.
Sam Nadler [19:04]: Great, Jordan. See you, everyone. Bye.
