Loyal to the Result, Not the Tool

0:00 Lee Rodrigues

It does not matter how much time and money you've sunk into this. What are we doing now? I'd rather keep learning than defend a habit, or defend an old workflow that I used to use. And

0:09 Sam Rogers

It's a great way to put

it.

Welcome to Signals and Subtractions. I'm your host, Sam Rogers. Loyal to the result, not to the tool. That's where we're headed today. Every one of us has some kind of tool that we keep because it's the one that we know, right? The one that we go to. And most of us also have some cool AI tool that we picked up because it was new and cool, and then we just kept using it because we'd already moved things into it.

Both of those are loyalty. Neither one is about the result. So I've brought back someone who's been at this table since the very first episode, and we love to talk tools. Lee Rodrigues, welcome back.

1:03 Lee Rodrigues

So glad to be back, Sam. I I feel like we talk about the tools people actually use here and I end up leaving every one of these doing like four or five Google searches and downloading something and joining something else and being better for it afterwards. So excited to be here again.

1:18 Sam Rogers

Yes, and that's that's why we do not just the signal of this new cool thing, but the subtraction. What's the thing that we can stop doing now that doesn't make sense? because you can't just add your way to success, right?

One signal, one subtraction from each of us. Lee, you are up first. What is the signal that you're watching right now?

1:38 Lee Rodrigues

something we've talked about for decades together is you don't build a house with a tool. You build a house with a blueprint, everybody who works for you to build that house has a workflow. And that workflow involves a set of tools that are reliable. They do what they're supposed to do.

But when it comes to like what I'm doing and and we're launching a performance support tool for those job seekers, job pivoters,

how do you adapt and figure out what you're going to do with this changing world? we've given ourselves a rigorous timeline. I'm doing a lot of podcast interviews in Riverside, much like the tool we're using right now. Riverside, fantastic for capture. chopping that podcast into smaller clips and exporting it out. Fine-tune editing, wrong tool.

2:18 Sam Rogers

Yeah.

2:19 Lee Rodrigues

Not good for that. Sam, have you tried fine-tune editing in Riverside? How'd that work out for you?

2:22 Sam Rogers

You know, I've I banged

I I bang my head against that particular wall every week making this show. Yes. Yes, I have.

2:28 Lee Rodrigues

Yeah.

2:29 Sam Rogers

Yes.

2:29 Lee Rodrigues

It it's like, you know, a a table saw is not a sander. They're not the same thing. They're for different purposes. So my workflow in the past was capture in Riverside, download to Final Cut Pro and finish in Final Cut Pro. capture in Zoom, download to Final Cut Pro, a bunch of different capture methods, capture on my DSLR into Final Cut Pro.

And as I've been doing more and more of this content, doing more interviews, the text-based editing is really powerful. I can search for the word that I'm looking for to hone in on the clip that I'm looking for, because in the performance support we're building, it's on a task-by-task basis throughout the workflow. So I don't need 10 minute videos. We're talking 35 seconds to get the concept, the story, and then all the performance support of how to do this and all those job aids and everything are underneath that. So you really want just a nugget.

just a little bit, and the text-based editing is great.

Well my workflow with Final Cut, download, cut it.

Upload it, get a review, get approval, and then render it out and get it published. Well, I tried Descript and once you're done editing, getting it for review happens on the server side. I'm not watching a render, I'm not watching an upload. I get a little message, all that's done while I'm doing other things. That's like a 20 to 30 percent time savings. The tool itself has a comment feature within it.

Or I can send it to my partner who can then give me comments, improvements, what they want. I can make those changes, render on the cloud when I'm ready. I can go straight to YouTube from Descript. So all of a sudden, all the thing I've been using Final Cut Pro. I was a Final Cut Pro instructor in the Apple store from 2010 to like 2015. I love the tool. It's great. I know seven, I know ten, I have no premiere, right? These are all tools I've used for a long time. They are so slow compared to what I'm doing.

Right now, just the downloading, the getting it all ready. And as much as I love Final Cut, as much as they have promised to create a text-based editing experience, I give them an F minus. It is horrible trying to do that.

But when the landscape changes, the workflow changes. And when

4:36 Sam Rogers

Yes.

4:37 Lee Rodrigues

the workflow changes, the tools that we used before, despite your comfort with that tool, might not be the most efficient way.

So, what I learned is rather than trying to get a tool in practice with it, I hit myself with a deadline, had five or six client projects, and decided to teach myself to Descript the stressful way. And it turns out

4:57 Sam Rogers

yes, yes, on the deadline.

4:59 Lee Rodrigues

you learn very fast when there are deliverables waiting to be sent out that have invoices attached to them.

My old loop, feedback, open the project, import more footage, render and repeat. If I have all the footage already in Descript, importing footage is three seconds. It's already on the cloud. Clip that move it over here. With Final Cut, that media has to be ingested. Then it has to process it, it has to work with it, which is all great in the classical way of doing editing.

But with these podcasts, this is not cinematic 4K footage going on a big screen. It's going on a screen as big as you're watching it on right now on YouTube.

And you don't

5:36 Sam Rogers

Yeah.

5:36 Lee Rodrigues

need any bigger than 1080p to do that. And or I don't. I know some people like to edit in 4K, but sadism. You know, why? Why? Why why do the files have to be so big? So

5:46 Sam Rogers

Well well

let's let's pause on that for just a moment because you and I have had this particular discussion for like I don't know at least a decade

what is the result that you're going for and how do you optimize for that?

For all the viewers out there, I want to make sure that it's good quality, but I don't necessarily need the best quality, best for what? Right? So best,

6:07 Lee Rodrigues

Yes for what?

6:07 Sam Rogers

best for fill in the blank versus like the best that is possible, I don't know of anybody who's like shooting IMAX podcasts, for example.

so two different points here.

one is knowing when to optimize and and what you're optimizing for. And the other is like the threshold by which you're saying, okay, now it's time. Is that just like I hurt enough? Or is this something that you do like once a month or once a week? Or, you know, how how do how do you figure that out?

6:39 Lee Rodrigues

Well, you've tapped into one of my favorite internal conversations, which is perfectionism. And the way we monitor perfectionism in the professional world is you need to be clear on the output. I used to train my team when I took on a team of six designers at Sunnova before you go too far into this development.

Take the final authoring tool, create a sample of the export, export it, upload it to final delivery. I don't care if it's a three-second video. You and I had a video we made at YouTube that said this video works. BABAM!

7:12 Sam Rogers

Right.

7:13 Lee Rodrigues

And it was just to make sure that if we upload a 1080p export, how will that look as the final delivery?

So for my output, I'm putting videos into Vimeo and videos onto YouTube that are going to go onto our knowledge base. And that knowledge base video beneath it is going to be the text that supports it. So I have an end product of a story that wraps the context around this content. So the content beneath it makes more sense, which is going to be a 30-second video. So the same thing we did at YouTube, I did it here. I recorded interview in 4K.

Edited it in 4K, added all my effects and everything, exported it out. The final cut. Finished one in 1080p, did it in Descript, exported it out. Guess what, ladies and gentlemen? It doesn't matter.

But more importantly, my workflow had a ton.

of beautiful titles and transitions and everything on it that I had spent hours building a template, put all this together by hand, Claude giving me all these great ideas for how the design will look, get it all together, look fantastic. And my wife looks at it, she's my partner, she's building all this, she's the expert on the camera, looks at it and says, wow, that looks very corporate, very polished, a little AI-ish.

Is that a story, or is that showing off how many graphics you can fit in one thing? And I'm like, not exactly the feedback I was after. I thought you would say, wow, this is zippy. And she goes, No, it feels like, you know, it would come out of the Googleplex, you know, someone who had millions of dollars to build this. Isn't this supposed to be just me telling a story? how many hours did it take to add all that flash and fizzle to these videos versus a simpler version I made

on Descript

Which is just some good call-out text, a couple slides that give you, we're talking about this. This is the key learning of the word. This is the concept. This is the quote. Nothing else. Maybe an image or two, but nothing else. No builds of text coming on screen. And it turns out that stuff that corporations like, the text building on screen, doesn't always impress people. They they look at it and go, it looks like AI slop. And I was like, great, because.

I spent 30 hours on a video, and my Descript video took about four hours, and they both achieved the same objective. So the end result was an end run around, a bunch of editing and work that made it look over flashy. Now, granted, this was the first one and prototyping and trying all kinds of stuff. The second one's would take a lot less time. But as you said, your podcast could be, you know, two, three, four hours to edit. If you're spending

Seven hours on a three minute video. Wow, what happened?

9:37 Sam Rogers

Yeah, yeah. there's varying levels of quality that can be measured, not just in terms of like file size and pixels and qualitative kind of feedback like like Christine had in in your case, but like measurable with the end user on the actual delivery platform. So you and I are

Often working, within corporate networks to deliver learning content, training stuff. the thresholds there, especially the architecture there, it's much better now than it used to be in in many ways.

10:10 Lee Rodrigues

Right.

10:10 Sam Rogers

But there's some there's some things that are still true, which is that if your end result audience, if the target you're trying to reach,

Can't tell the difference between A and B, then arguably there is no difference between A and B, right? put 30 hours into this one and four hours into that one. But at that point, that's 26 hours of waste because the target can't tell the difference between 30 hours and four hours. Or like in your case,

the preliminary target, says, you know, I actually like the four-hour one better than the one that takes all the flashy sparkles so back in episode nine you said that the most powerful tool that I have is real data, not assumptions. Is this just maybe the

Same kind of extension of that a few months later.

10:59 Lee Rodrigues

Yeah, and you know it brings up

term I use to measure this is the MED, the minimum effective dose. Water boils at a set temperature. When you increase the temperature beyond that, do you get better boiled water? What do you get out of it? You just boil

11:14 Sam Rogers

Ha ha ha

11:16 Lee Rodrigues

more water into steam, right? It doesn't get you anywhere. why, right? data is I'm trying to get this good enough.

That works for what we're doing and battle my own internal perfectionism. And my own internal perfectionism, when I worked in the corporate environment, I would send over a video that I had just finished rendering in Final Cut. And I would say to myself, I have a list of seven things I want to correct on this video to make it perfect. And my director replies back and says, Love it, publish it.

Guess who's not fixing those seven things? They're good enough, right? And it takes

11:44 Sam Rogers

Yeah. Yeah. Yeah.

11:46 Lee Rodrigues

the audience to look at it and go, damn, this is better than anything we have. I'm like, yeah, but this audio isn't mixed right. And this title comes off on the wrong thing. And it's like, dude, it is fantastic. Ship it. if we're left to our own devices, we will refine away.

That's what it comes down to.

12:02 Sam Rogers

that's some great, very specific examples of the kinds of things that you're wrestling with and and kind of the overall signal that you're tuned into. so here's mine, which is kind of the other half of Lee's. every platform changes under us. the ones that we own don't. So

so Lee, you were actually there for this one. our very first live stream failed when the encoder told us that everything was fine. it happened just before we were going live here. we are live on YouTube, but not Substack, because Substack's live stream stopped connecting. and this happens frequently. it stops taking video uploads in the middle of something.

And you know, I can test it and it works, but then I go live and it naturally it doesn't work. I've also had other issues like in ramping up this podcast where streaming to YouTube as a new account. You don't get to do that for like 30 days. So that had to wait, right? I had was in an Apple sign-in loop for two days to get this

Podcast podcasting on Apple Podcasts. LinkedIn Live took three tries to start that one episode that you and I did with

13:11 Lee Rodrigues

Right.

13:11 Sam Rogers

Christine. now none of those were disasters, but every one of them was a dependency on a tool that I didn't own, that was changing under me and not living up to promise, or at least my expectations.

The site for this show is the one thing that held because it's mine. well, it's ours, the joint repo that we have on GitHub for all the co-hosts. so the rule became: if the site publishes first for everything, then the platforms get linked to that downstream.

as long as the result lives somewhere that I can own, then swapping the tools out really doesn't cost anything. I can stream as I have to Facebook, to X, to whatever platforms will accept the RTMP. And as long as we're streaming somewhere, then we're streaming somewhere every Wednesday, like usual.

and so my signal for this week is that it's only the thing that I built.

That actually has been reliable for the last 16 episodes. Every other thing that I've used, including Riverside, including Descript including, you know, Final Cut, which I've used Camtasia for things. Yes, I've used every tool that you've listed.

the more we can build our own, which is actually possible now, the fewer places we have to worry when it breaks, because we have alternatives that we can toggle to quickly. I'd love to get your thoughts on that, Lee.

14:47 Lee Rodrigues

Well, as you said, there are variables in our workflow. I like that term, the the platform changing underfeet. It gives me a vision of like Kid Icarus or Mario Super Mario Brothers, where you duck on a platform and it turns out you fall through it. That's

14:59 Sam Rogers

Yeah, yeah.

15:00 Lee Rodrigues

kind of what happens when someone runs an update and takes away the feature you were using. And it's like, well, you were a cool tool. So I think that's where it comes down to.

Be focused on your workflow, your output, and recognize any of these tools can be substituted.

Are we ready for subtraction?

15:17 Sam Rogers

I'm ready for the takeaway I don't think you've been on an episode where we had takeaways. This is something that I introduced on episode twelve. for every episode of this show, there is an asset that you can make actionable. So something that you can actually do

To be able to put to practice what we're talking about in this episode. here's a little generated video that I actually did right as you were speaking your signal, Lee. this is the takeaway for this week's episode, which is two decisions and an owner. what are we using as

evidence to say the thing that we're deciding, the thing that we're optimizing towards is appropriate is something that we want, and forcing that decision to actually have an owner that's making that call.

it's really useful to say what counts and what doesn't count as evidence for them and being able to have a clear verdict. So what I was just showing is available at sigsub.show slash takeaways I'll put the link in the show notes. it's a great way to be able to take some of what Lee and I are talking about.

and make it into something that you can actually practice, use, and and I'd love to hear back from anyone who's using it as well to say how it's working

for you.

16:29 Lee Rodrigues

this is the lesson that happened to me a couple of years ago that continues to return. And I'll put it to you this way: when you finish a product and you're sending it out to the org to review, remove the word review from your vocabulary. I

16:41 Sam Rogers

Hmm.

16:42 Lee Rodrigues

don't want you to review it, I want you to test it and approve it.

We use the word review, and someone looks at it, and they will give you the worst two-word that you can get as a designer. Looks good.

16:55 Sam Rogers

Right.

16:56 Lee Rodrigues

That means probably didn't even look at it, and we're gonna find out something's wrong with this after it's launched when it's a stressful replacement, right?

17:01 Sam Rogers

Right. Right.

17:03 Lee Rodrigues

That's typically what that means is gonna happen. When you tell someone, I'd like you to test this and approve it, here is the judging criteria.

something we learned at Google. If you're sending out something for review and making a second request in that same email, you have failed. You

17:20 Sam Rogers

Yes.

17:21 Lee Rodrigues

have failed. You have set yourself up for failure. It's gonna

17:24 Sam Rogers

Yes.

17:24 Lee Rodrigues

be a mess. Begin licking your wounds now. If you want approval on something, One objective per piece of content. And if you have more than one,

If you're gonna ask for approval on a piece of content, then ask a follow-up question, just shoot yourself in your own foot because they're gonna answer one of the two. So approve, not review. And here's what we're looking for.

17:45 Sam Rogers

So given everything that we've been talking about thus far, Lee, what stopped making sense? You threw your extra subtraction in, but I know you had something else prepared to talk about today.

17:55 Lee Rodrigues

I dropped Final Cut Pro and my loyalty to a workflow that used to work.

The way I was doing it before made sense. Capture with the DSLR, cut it down from a storyboard, right? But when you're capturing an interview with a conversation, now I'm looking to capture specific pieces.

All of a sudden a non-text-based, non-linear editor is an excessively large amount of skimming the timeline doing the stop play, stop, play, stop, play, stop play. Now I can throw in the words I'm actually looking for that I remember being said when I recorded it.

Back in the day, we were doing interviews and we would ask a question, record that question. That output was going to be the video because it'll go into an e-learning course. Well, I'm doing.

Performance support now. It's much smaller pieces. I'm only talking about the task that you're doing right now. And how it used to be was, well, here, watch this sixty minute video. And these days it's like Are you are you serious right now? You want me to watch a six minute video is probably

18:51 Sam Rogers

Yeah, a a six minute video is a big ask.

18:54 Lee Rodrigues

a lot, right? welcome to short attention span theater. it's just where we are, right?

I had invested a lot of time in it. And my wife has an MBA and she will remind me of sunk cost all the time. It does not matter how much time and money you've sunk into this. What are we doing now? I'd rather keep learning than defend a habit, or defend an old workflow that I used to use. And

19:15 Sam Rogers

It's a great way to put it.

19:17 Lee Rodrigues

and things change. They shift and they change. That's fine. But currently.

I'm trying to put a bunch of small clips together, keep them consistent. I just was not getting the return on investment for extra polishing.

And frankly, you told me this when we first met at YouTube. Videos have two speeds, good and off.

19:35 Sam Rogers

Yeah.

19:36 Lee Rodrigues

That's that's all there is.

19:38 Sam Rogers

It's true. It's true. Yes. So here's my subtraction. it's getting back to that loyalty idea of loyalty to the surface. so tools and platforms are what I'm talking about when I say surfaces. They're where the the copies land. So the result is the artifact.

And the surfaces multiply this. Artifacts don't. So, like making the video for this show I'll upload it directly to Substack. I'm going to upload it to YouTube. It's going to be like the same artifact across those different surfaces. Lee's dropping the loyalty to the old tool, you know, to Final Cut Pro. I'm dropping the loyalty to

Any tool, old or new. Loyalty belongs to that result. It does not belong to any tool. Which means that every tool has to earn its seat as evidence. does it do this well? Does it do it here? Does it do it now? Is it actually better in a way that I can measure? Again, measuring with like the target audience in mind? Or is it just faster?

Is it just shinier, So a critical thing with AI here, to bring it back, because I see people make this mistake often, is the tool's opinion of itself is not evidence. And you'll see this in the takeaway that I made for this week, too. remember that asking Claude.

About some of these things that we're talking about is different than conducting the test yourself. We talked about this in the previous episode that Lee was on, episode nine. Like there's some parts that need to stay human that we need to do, the AI's reporting of something may or may not be trustworthy. in the same way.

That you could ask a human that is maybe new to the team, or maybe junior, or maybe just doesn't know as much, not necessarily bad intent, but they might report to you something that's like, yeah, I checked this part here, but they didn't necessarily know to check the underlying data. So keeping track of your own data, remembering

the AI assistant maybe doesn't quite count. Like you want to check on it yourself. No different than working with humans. And what it costs is the comfort of that tool that we know. and Lee, I'd love to ask for Final Cut Pro: what's the delta between I think I probably shouldn't do this anymore versus I'm not.

doing this anymore. Like there's some emotional distance in there, I'm guessing, that took up some calendar time.

22:10 Lee Rodrigues

Well, it's it's the minimum effective dose, right? You look at the output that I got from one tool and I look at the output that I got from the other tool. And if you have not effectively logged how much time, reluctantly or happily, put into these two products, I like to create a little spreadsheet that tells me here is the amount of time I put through each of these processes. And what changed it all.

was what you just said about the AI, what I ask my AI is I like to use black hat thinking. Tell me things I'm doing wrong. Tell me some mistakes that are made.

if you ask it to tear this apart, and that's what I do with mine, and Claude was like, This is great, but where's the render time on this, the active render time on the computer? And I was like, I wasn't counting ingestion into Final Cut Pro. So when I counted ingestion into Final Cut Pro on the top side of taking the hour webinar.

and ingesting it in, which by the way is like thirty five minutes. If you try to edit it while it's ingesting, it stops, you get gaps, you get glitches. It's not

23:04 Sam Rogers

Yeah, yeah, yeah.

23:05 Lee Rodrigues

done chewing up the footage yet.

it was about a six to seven hour process on one editor and about three on the other.

23:14 Sam Rogers

just so you know, you should probably go back, in fact, everyone listening should go back and listen to the most recent episode with limited edition Jonathan, where he talks about his video generation process. All that stuff that's programmatic, like easing the text in and out. You can actually do that via AI now and commands to an editing system such as Descript or like I'm

I'm doing even locally with Da Vinci Resolve. it's actually less work, far, far less work than it used to be. But this is the very argument for this show, which is that the new tools are the same as the old tools. They don't deserve our allegiance. What we need to stay focused on is the end result, not necessarily how we get there. It may involve

Some rearranging of how it is that we think of the workflow And we're talking in this case about specific examples to do with video editing, but this applies to really any kind of work where it's not done in one sitting in two minutes or less. Like if there's handoffs that go to other teams, if there are multiple tools that you're working with to create whatever assets you're working with.

It's the same thing of staying focused on what makes the impact at the end result layer. Not that it got approved, but like, did it actually move people did it actually make the target move in a detectable way? Preferably in the right detectable way, and not just.

you know, run screaming from the AI-ness of it all. well, thanks, Lee. It's always great having you here. I really appreciate you co-hosting this week. if people want to reach out to you, where would you send them?

24:54 Lee Rodrigues

System, the number two, focus dot com. Sign up for our newsletter. Cool stuff is coming.

25:00 Sam Rogers

All right, and I will put that on screen and in the the show notes as well. thanks as always. thank you as well, dear viewers and listeners. one thing to do this week pick one tool that you're keeping, old or new. Write down the result that it's for. Then write down the evidence that makes that result better. And cross out anything that came from the name of that tool.

If there's nothing left on the page, that's not a tool decision yet. You're just exposing your own loyalty. The takeaway is available at sigsub.show/ takeaways. The two decisions one is the one that we're talking about. But for today, that's our show. If you're listening on audio right now, please tap follow. If you're watching live, open your podcast app, search for signals, ampersand, and subtractions.

follow us there so that the next one just magically shows up next Friday. On Sunday is issue 73 of our newsletter. One signal, one subtraction, one analogy, about five minutes to read. Everything on the show lives at the one thing that's never broken sigsub.show. One signal worth watching, one subtraction worth making. Now go find yours.