Rick (00:01.656)
What's up this week, Tyler? I had one of those nights where I woke up at 4 a.m. and could not go back to sleep, so I just read a book for three hours. In bed. Yeah.

Tyler King (00:03.44)
Not a whole lot Rick.

Tyler King (00:12.066)
Ugh. Like in bed? Wow. Is that, I mean, obviously you'd rather be sleeping, but is that like, sometimes I find when I'm like, like when I come into the office and no one else is here and it's just me, it's like a kind of alternate reality and I kind of like it. Do you enjoy that at all?

Rick (00:30.178)
when I'm not sleep deprived.

Tyler King (00:31.514)
Yeah. So, no.

Rick (00:35.321)
I mean, it was nice to read the book, but now I'm like, I have a lot to do today and I've got to figure out how I'm going to buy some energy. I'm going to go work out after this after this podcast. I'm going to go do a bike ride.

Tyler King (00:43.194)
Nice, nice. That would sap energy from me, but I'm happy for you. Speaking of reading, I just started reading, there was this series of fantasy books, I read it back in middle school, which is the last time I ever read seriously. I basically stopped between middle school and age 40. And I'm like repicking up that series I was reading back then, and that's been fun. It's The Wheel of Time, yeah, The Wheel of Time, which, well known, I know that's not like an obscure pick.

Rick (01:05.262)
What was it?

Oh yeah. Amazon Prime series. That's, mean, that's a, that's like 12, what is it like 10 books?

Tyler King (01:12.804)
I hated the show. The show was terrible. Yeah, I think when I was reading it, there were six or seven and I didn't make it all the way to the end. I think I got maybe four or five in and they're, you know, 800 page books. So I'm going to be reading this for the next decade, but I'm into it.

Rick (01:28.588)
Yeah. So the guy who finished the books is a named Robert or Sam Sanderson, Brandon Sanderson. Yeah. So he he is a BYU professor of fantasy writing, which I think would be such a cool course to take in college.

Tyler King (01:34.374)
Brandon Sanderson, I just read some of his stuff, yeah.

Tyler King (01:45.936)
You've gotta be a real badass for someone to hire you to be a professor of fantasy writing.

Rick (01:49.974)
Yeah, he's a badass. But anyway, he I'm reading that's the book that I'm reading one of his series right now. And it's you know, they're the Stormlight Archive is the series. So it's it's the book the second book of the series, which is called I guess called Words of Radiant. Radiance. Miss Porn's great. Yeah.

Tyler King (01:55.365)
Which one?

Tyler King (02:05.82)
Nice. I just read Mistborn. Have you read that? Yeah. And then I was going, I was deciding, am I going to read more Brandon Sanderson or am I going to go read the Wheel of Time? I, hey, I decided to wheel of time because Brandon Sanderson is the one who finished that. Like that's a funny coincidence. But then, yeah, his website is like the Stormlight Archives is like my good one, right? That's kind of how he presents it is like, if you're serious, that's what you should read.

Rick (02:19.446)
Yep. Yeah.

Rick (02:26.157)
Yep.

It's intense. I highly recommend it. you know, in 10 years when you're done with Will of Time, can look at that. I feel like Will of Time is one of those that you're going to probably start and then be like, I need a break and then come back to it.

Tyler King (02:36.251)
Nice.

Tyler King (02:41.508)
Yeah, my memory is the first couple, this is from 30 years ago, but the first couple I was like glued to and then it gets the middle and they're still good, but there's so many characters doing so many things that like an 800 page book, each plot line only moves a little bit and it is kind of tedious at that point.

Rick (02:52.194)
Mm.

Rick (03:00.238)
When you read Mistborn, did you read the trilogy or did you read just the first book in the trilogy? Oh nice, that's great. Yeah, I really, it's a fast-paced book.

Tyler King (03:03.388)
Mm-hmm, the trilogy, yeah. It was really good. I think one of the reasons I stopped reading for fun is that when I got to high school and it's like, okay, now you're reading a bunch of books and you're analyzing them and it's great. A, it kind of takes some of fun out of it, but B, I think I became like a snob where I was like, like what I was reading was just trash fantasy and like there's good literature. And then I tried to read that for fun, but it's not fun. Like good literature is not fun to read. So what?

Rick (03:32.224)
It's a great way to fall asleep though.

Tyler King (03:33.74)
Yeah, fair. Like, I remember I had to read Anna Karenina twice in one year and I was like, I'm done reading for the rest of my life.

Rick (03:42.094)
Well, we'll have to like talk about, well, not on the podcast, but we'll have to talk about some of the reading that you're doing. We're reading the same stuff. This is fun.

Tyler King (03:50.362)
Yeah, yeah, that's cool. What's going on business-wise?

Rick (03:56.162)
Well, we had a record month in November, which we hit 25K in monthly recurring revenue, which is really exciting, well beyond our goal for the year. It's coming with a very challenging open enrollment though. we're, you know, I feel we had a partner meeting on Monday and you were there, so I don't need to belabor it, but like, you could tell Dady's like, man, he's slammed. and he's...

Tyler King (04:17.498)
Yeah. And it's a combination of things, right? Like he's every open enrollment, he's busy, but it's like busy plus is it really working kind of double whammy, right?

Rick (04:28.3)
Yeah, yeah. Yep. And it's unique, this open enrollment because of the government shutdown was over these subsidies that are expiring at the end of the year that have a massive impact on our customers. And you never know what the actual impact is going to be until it hits. it's causing, mean, if normally it would be like a quick 15 minute, like, hey, you're your choices, would you like to pick or self enrollment? It's now like,

This isn't working for me. I'm going to cancel. Like, what are my options? And so now JD's having to go, here are like 15 different things you can do. I have to asterisk every single one of them because they're not real health insurance. And so he's figuring out how to do that, that spiel and it's multiple calls per person. And so the effort required, I'm sure they're not, or maybe they're more. I mean, it's likely that they're more because the margins are higher. But it's just,

Tyler King (05:05.756)
Mm-hmm.

Tyler King (05:16.336)
And I assume the commissions aren't as good either. Yeah.

Rick (05:27.886)
massively more time consuming and, you know, emotionally atoning because these people are lives are being wrecked. mean, if health insurance goes from $100 a month to $20,000 a year, I mean, that's more than your mortgage sometimes. And it's like, OK, like, what am I doing? I can't afford this. And if you have a health condition, like, I mean, it becomes an even larger issue.

Tyler King (05:52.878)
Is the law still in place that, like I know one of the big things with Obamacare when it passed was that if you don't get insurance, have to pay a penalty. Is that still true?

Rick (06:02.19)
Nope. They took the penalty the way they got, didn't know the penalty still exists, but it's zero dollars. So technically, yes, there is a penalty, but it's zero dollars. Yeah, I don't know why. I never know and understand why they do it that way. Like why is there something different about like changing the amount in a statute versus like, yeah, it's like, okay.

Tyler King (06:12.23)
write a check for zero dollars to the government.

Tyler King (06:22.054)
I assume that's it, It's how like a lot of laws can pass that are technically tax code changes. They're not law, you know, things where it's like, well, it's easier to get it into the bill that

Rick (06:31.47)
Mm-hmm.

Yeah, so that's what they did.

Tyler King (06:36.624)
Cool. Not, well yeah, sure, whatever. I don't know if I think that's cool or not, but.

Rick (06:41.414)
But anyway, it's a bittersweet sort of, I don't know the right word is, like it's hard right now. And I don't feel like I know how to help. So I'm trying to be like emotional support, but also it's kind of like we kind of have to let JD power through this, be supportive when he asks for help about specifically, offer help, offer support, and then wait until February.

Tyler King (06:50.043)
Yeah.

Tyler King (07:08.026)
Yeah, I did appreciate it. He's pretty open about how he's feeling and that if someone were like acting like everything's great, this is fine. This is, and he, but he also wasn't like, this is terrible. He said he's a five out of 10. He's like, this is, this is hard and it's fine. Like it's not going to ruin me. if he was like, I'm nine out of 10, I would be like, you're lying. You're about to burn out.

Rick (07:21.869)
Yeah.

Rick (07:30.76)
He was like irrational about a thing. I don't know. Yeah, well, I think we just got to get through it. It's a one time shift in the market. I don't think it gets worse from here. If any changes that happen will provide opportunity and it doesn't seem like they're going to. I mean, what would really suck? Here's the worst case scenario. They basically come, you know, at the end of open enrollment, say.

Tyler King (07:33.979)
Yeah.

Tyler King (07:53.227)
They undo it all.

Rick (07:56.994)
The subsidies are back. And so JD has to go back to all these people and go, good news. We did all that work and it was a wasted effort.

Tyler King (08:00.252)
my God.

Tyler King (08:06.202)
Yeah, so weirdly you're rooting for it to stay fucked up.

Rick (08:09.802)
I don't know. mean, yeah.

Tyler King (08:11.548)
I get what you're saying, that would be good for America, certainly. That would be good for the people. It would be hard for JD.

Rick (08:19.692)
Yeah, this open enrollment period would be Harvard J. It'd be better for our business long term. It's better for the market. It's the right thing to do. I think should do it. So hopefully they will.

Tyler King (08:30.123)
yeah.

Rick (08:31.17)
Yeah, that's what's going on in my world. I'm like trying to be helpful, but I'm also like, OK, I probably don't need to be throwing things at JD right now. I just need to be like helping where I can.

Tyler King (08:38.864)
Right? Yeah. That's so hard when you like have an idea and you're like, I just want to send this over to somebody, but you can't like the boss sending an idea to someone is no matter how you phrase it, going to be interpreted as a command to go do it right.

Rick (08:45.582)
Nope.

Rick (08:52.502)
Yep, yep. What's on your plate right now?

Tyler King (08:56.246)
yeah, just kind of a variety of little things. still just trying to keep my head above water with this, parental leave three and a half day per week schedule. So I'm definitely looking forward to January when I'm back to five days. but yeah, so like project man with three and a half days, I'm, you know, I'm cutting out the stuff that doesn't have to happen. One of the things that does have to happen, both, which both means I have to do it and I can't, someone has to do it and I can't delegate it. Those two things, right.

Project management is a big one there. And there's a lot of good stuff going on at the company. Like our dev team is doing great right now, but man, it's almost a full-time job just being a project manager. Not even a product manager, just like a, well, I don't know, the two blend together, of course. But I did run into an interesting challenge that I don't think we've really dealt with before, which is the devs are working on roughly three different big projects right now, like about

two devs per project, but they're all kind of intertwined. And we've been having a lot more like, okay, there's two different things are happening, but both of them require redesigning the header on this one page. So both teams redesigned the header separately to the same thing, but they just both did the work, you know, or like, when do we ship something like there there's all these different feature flags being like,

If you turn this on, it turns 100 % of this feature and 20 % of this other feature on because the code's so intermingled. Does that make sense to you, like as a non-coder, does that make sense, like what I'm saying?

Rick (10:37.45)
Yeah, I think what you're saying is that people have different like sort of domain, like there's just more complexity and there's different sort of knowledge of the code base. Is that what you're saying?

Tyler King (10:49.628)
No, like it's the projects themselves are different in nature than they have been. So, and so this kind of brings me to maybe like, if I'm trying to pull an insight out of this, I've said many times on this podcast that in the past, I think we were working on projects that were kind of too small and low impact. Like, you know, the type of thing where 5 % of our customers will be 1 % better off.

Rick (10:53.72)
Welcome.

Tyler King (11:15.472)
And we should be working on things where 50 % of our customers will be 5 % better off. know, like, like forms was a big project that actually unlocked new functionality in the CRM as opposed to just like, when you're, when you're doing this thing, it's one less click than it used to be.

Rick (11:30.572)
Got it, so it's not a nature of the complexity of the code base, that it's more like the thing that you're trying to do is bigger and touches more core things, which requires more people.

Tyler King (11:40.218)
It's not that it requires more people. It's that if you're working on a big, the first part of what you just said, a hundred percent, they're bigger and they touch more things. But if you imagine you're working on three different projects at the same time, if they're all small isolated things, the chances of them overlapping are very low. So in the past, I could just give different projects to different people. They would go do them. They're all in their own little bubble and it was fine. Now they're all working on projects where it's like this project touches every page on the site, but they all touch every page on the site.

Rick (11:55.47)
Hmm.

Rick (12:08.718)
So they're colliding? Got it. Okay, did not follow. Now I get it. That was probably me, actually, sleep deprivation. Yeah.

Tyler King (12:09.956)
Yes.

Sorry. No, well, probably a little of both, yeah, no, I'm still like, it's just a new experience for me, but in a way it's really hard. And I think the devs are getting a little frustrated cause they're like, they're used to me always anticipating everything. Cause with a simple project, I can just in my head, if you have experience as a developer, we're not doing rocket science here. It's like simple B2B crud stuff. I can think, I can see a project, turn it into notion cards, plan the whole project.

And I don't get everything right every time, but it's normally pretty close. But this time a devil do something. I'll be like, fuck, I forgot that someone else is working on this thing that actually can you undo that? Because that just hasn't really been a problem before the upside to it though. Yeah, sorry. Yeah.

Rick (12:53.966)
Yeah.

This is the nature of product management though, isn't it? Breaking these things down into small things and identifying the commonalities and then making that its own project that isn't actually a feature. It's building towards the feature. It's like setting up the feature, you know?

Tyler King (13:16.122)
Yeah. Yeah. I don't, and I've, I've gotten a whole new level of respect for like, how could you possibly run a big company with thousands of developers? part of me sitting here, like, even if we could, even if we were growing revenue, I'm not sure how much more we even want to grow the team. Cause I get why software quality seems so bad everywhere. And like we have the advantage that I can hold everything in my head and I know every developer and I know what every developer is working on. even still.

We have all these weird, it's not quite bureaucracy, but like that type of inefficiency happening. How does it work when you have two teams or 10 teams doing this? But the good thing, like the silver lining to all of this is I've, while it is harder work and the devs are maybe a little more frustrated because of it, it's making me realize this might just be a fundamental characteristic of the right type of project.

Like if this isn't happening, that might be a bad smell. Not that like 100 % of projects will work this way, but if I am looking at it, I'm like, none of them are stepping on each other's toes. Is that a sign that we're thinking too small and we're not working on projects that will really improve the product in a meaningful way?

Rick (14:32.418)
That's a good point. Also, I guess now that you're noticing this, is there not opportunity to get more efficient with each of the projects by focusing on the commonality, really thinking about the things where they overlap and getting those right as one project so that you're not doing it twice?

Tyler King (14:46.341)
Yeah.

Okay, I had not thought of that before, but my immediate reaction to that is that sort of works against a different thing that we've gotten a lot better at, which is back in the day, I would be like, here's a big project, make all the mockups. I didn't make Notion cards or anything. I just handed it to the dev and I said, build it. Or even then I would make Notion cards, maybe like build the whole thing and we'll ship it all at once. One of the huge productivity boosts we've had as a company was

I started putting a lot more thought into how do we break these into the tiniest little, not just tiny pieces, but separable shippable pieces such that at any given time, almost all the code that has been written has been merged into our main branch and shipped. That saves a lot of hassle where if you imagine two people work on big project, well, sorry, there's two failure modes I want to talk about. One, two people are working on big projects at the same time. Both of their code bases don't get merged in for a long time. And so they're, they're working on two like wildly different

versions of the code. And then when they try to merge them both in, there's just a million conflicts. I realize that's not what you're saying, but that's one thing that we avoided. But the other one is, there's this natural instinct to be like, well, while we're doing this, let's also do this. And let's also do this. while we're in this code base, I know we're going to need to do this other thing for this other project six months from now. So let's just do it now. And it just slows everything down because you're letting perfect BVNMT a good.

That's my concern with what you just said is that we, blocks us from moving fast because we're trying to get everything in place before we ship, you know? Am I interpreting you right though? Like.

Rick (16:25.688)
Yeah, yeah, now it makes sense. Yeah, yeah, I think what I'm saying is like, make the perfect plan. And you're saying, yeah, but there's a cost to that. And I think you're right.

I know. think, I think like there's probably we're talking about this is two extremes, right? There's probably a middle path where you go through and say, are there opportunities for us where in certain areas to invest ahead of time to save us time, not by like over planning, but because it's just like so obvious the high leverage that we should do this first or we should think through this dependency first.

Tyler King (16:45.948)
Yeah.

Tyler King (17:04.347)
So let me run through, I think we actually did that like too late on this. I, we first had the collision and then I was like, here's how we could have done it. And we did it, but let me just give the example. I'm not sure if this will make sense without being able to see it, but one project we're working on is saved views, which is basically any report we're putting tabs across the top where you can like save a set of filters. So like, I want to look at my leads.

from last month will be one tab. And then you click on the tab next to it it's like, I wanna look at all the other users leads from the last quarter, whatever. This is a pretty common pattern, Notion does this. A separate project is we're building the Kanban board. Both of these require removing the sidebar. We have all these controls on the sidebar of these pages and we're moving them to the top. like instead of, we're just moving everything, but we're moving them for both projects. What I should have done was been like,

Project number one is move all this stuff, ship that, and then there's a Kanban project and there's a Save View project from that. We ended up doing that, but kind of clumsily. But that's the type of thing you're talking about, right? You figure out where these things overlap, pull that out as a separate project, ship it, and then the other projects get built on top.

Rick (18:12.12)
Yeah.

Rick (18:20.296)
while you were talking, I wrote down a note so didn't forget it. I was thinking like, well maybe the solution here is just like, know that this can happen. Try to like be prepared for this, like as part of the planning, but don't over prepare so that when it happens, you can quickly correct like what you did. And maybe that's the right thing. like, it's kind of, a lot of the things that I've learned on this podcast with you and in the early days of Leg Up Health is don't over optimize for hypotheticals.

Tyler King (18:47.548)
It's so hard. Yeah.

Rick (18:50.286)
I think like where it's unclear to me because I just don't know as much about coding is like when is something a hypothetical versus like, no, this is actually like inevitably going to happen. We should actually prepare for it. Like, is that clear or, know, I guess, I guess a minute that becomes clear. You want to do something about it. But but if it's not clear, then it's hypothetical.

Tyler King (19:06.161)
Yeah.

Tyler King (19:11.664)
This has been one of the main ways in which I have failed as like a product visionary or what that sounds really highfalutin, but you know, like the person who's coming up with the product strategy in the past is I have all the, used to spend a lot of time thinking about like the long-term future of the, like where do we want to be five, 10 years from now? And then anytime an idea would come up in the present, someone would be like, someone requested this small change to how tasks work. And I would be like, that sounds good with how our product currently is, but that's kind of.

either incompatible or redundant with how it's gonna work 10 years from now, so we're just not gonna do it. But the thing is then we never, like 10 years ago we were saying that, we still haven't improved tasks. So like, I am now starting to realize if there's like a one week thing that makes a feature much better, even if it's inconsistent with the long-term vision, most cases you should just do it.

Rick (20:02.018)
Yep. Yeah. My version of that is it's the opposite. It's over engineering things for future use. So like in our plan dictionary or database for our clients, like adding all the fields, you know, that we could possibly need so that we have them versus like, what's the minimally viable thing that we need? Because every time you do that, you add complexity to your brain that you have to manage to the product that you have to manage. And anyway,

Tyler King (20:10.651)
Yeah.

Tyler King (20:30.716)
And how often do you actually get back to it? Like, not often. Yeah. Yeah. Can I also just to, can I segue to a different thing that this conversation made me think of? Talking about the like, you want to, there's this temptation to keep pulling the thread. Like you're doing, you're doing one thing and so, and you're like, well, that kind of touches on this other thing and that touches on this other thing. Let's do them all at once. And I think it's generally good to not.

Rick (20:34.109)
1 % of the time, 5 % of the time, it's very infrequent.

Tyler King (21:00.86)
follow that instinct to say, you know what, no, we're just gonna do the one thing. But the situation where I find myself feeling the opposite is we have a lot of tutorial videos to explain how our features work and videos are really hard to make and basically impossible to update. Like if you have a help article or a blog post, you can just go in and be like, this sentence needs to change because something changed. You pretty much have to remake the video from scratch every time.

And you know, have like, call it 10 really core instruction guide videos and probably another 15 to 20, like you go to this random settings page and there's a video that explains how it works. So we make all these videos and these are super valuable. I often think maybe we just shouldn't have videos because they're so hard to update, but no, they're really good. Our customers praise them all the time. But then over time, like we have all these videos and for a brief moment in time, they're correct.

And then we start changing things and it's like, well, that's not a big difference. Customers will figure out that button used to be at the top right now it's at the bottom, right? They'll figure that out. But over time, the videos are just completely wrong. With the changes we're working on right now, especially Kanban and saved views, the two things and moving, getting rid of the sidebar that I was just talking about, we kind of have to redo videos for that. And almost every video, almost all 25 of those videos show one of these pages one way or another.

So now I am thinking like, should we go do all the other things that are gonna fuck up the UI before re-recording all these videos?

Rick (22:38.264)
How much does it cost to record the videos? Hours.

Tyler King (22:40.73)
We have many hours. It's very time consuming. Yes, I'd say it's 10 hours per video. These are not like quick loom videos. These are like well-produced. Like one of our CRM coaches has their 20 % time does it. Now we can do quick ver- like when we ship a new feature, what we normally do is I record a loom video that we put there. And then the CRM coach like when she has time goes in and makes a better version of it.

Rick (22:44.59)
more than 40?

Rick (22:51.886)
times 25.

Tyler King (23:09.732)
You can make a quick and dirty version, that's not what most of them are.

Rick (23:14.254)
Yeah, well two things. I have two thoughts. The first is I think it seems strange to have the tail wag the dog in terms of actual value to the customer. It seems like whatever is best for the customer in terms of the product development life cycle and the business should be the way.

Tyler King (23:15.846)
So yeah, it's all good.

Tyler King (23:20.251)
haha

Tyler King (23:33.071)
Agreed.

Tyler King (23:39.824)
These are things that we would have been doing three months from now, not five years from now, but yeah.

Rick (23:42.894)
Okay, but but it seems weird to like say we're gonna do this because we don't want to remake videos It seems like you should figure out how to make videos more effectively and more efficiently Yeah, well you could cut the you could cut You could make less less produce videos or I mean, I this is the second thought I had which is I Read an AI newsletter religiously just every day just to force myself like I got to look to stay up on the tools that kind of thing

Tyler King (23:53.744)
I don't think it's possible. Sorry, it's possible to make worse videos more efficiently. Yeah.

Tyler King (24:07.772)
Mm-hmm.

Rick (24:11.904)
One of the tools that keeps coming up is, that's really interesting to me, is a bot that you basically have, you prompt to build these videos for you. these help, it's for a SaaS app basically, and you basically prompt it to go learn to do things in the app and then build content, help articles, videos, et cetera, for you for that use case. Now, I'm sure there is like an overarching like,

human version that's better, like for individual like feature, like I think of like some of the best video content I have seen is the Webflow content. They have like this, like this guy who has like the perfect radio voice just sitting in it. Yeah, exactly. So like, I don't know how close it is to that level, but like, it seems like this is something that AI could assist with getting to good enough if for a SaaS app. I'll have to go through my, I have like a list of 30 tools.

Tyler King (24:45.244)
Mm-hmm.

Tyler King (24:53.052)
Yeah, it's like funny. There's a lot of humor in it, right?

Tyler King (25:06.46)
Do you know what that's called?

Rick (25:10.37)
But yeah, I can look it up.

Tyler King (25:11.938)
I, my initial reaction is like heavy, heavy eye roll to that. But I mean, I should check it out. I'm, I'm very dubious, but I'll check it out if you, if you can find that.

Rick (25:22.456)
What would be cool though is if you could have AI at least do the polishing of the production and then have, you know, come in after it and edit it with the human, like sort of rawness, which is maybe less produced. That would be an interesting sort of flipping of the script versus like what you're doing right now is like you're doing the raw and then, which is like the quick and dirty loom and then you're passing it to a human to produce. I just wonder if the AI can't somehow accelerate the production.

Tyler King (25:35.26)
Mm-hmm.

Tyler King (25:48.283)
Yeah.

Yeah, it maybe could. I do think like when you were talking there, I was getting greedy. I was like, awesome, awesome. All our competitors are gonna start making these dog shit videos and we're gonna have even more of an advantage. Like saving time is sometimes the goal and improving quality is sometimes the goal. And I do think like not every company, but for some companies doing this type of thing with a real human element is probably worth it. But yeah, if you don't mind finding that at some point, we should mention it.

Rick (26:00.408)
Yeah

Rick (26:16.141)
Mm-hmm.

Tyler King (26:20.826)
on the podcast when we do, because I'm guessing some listeners are interested as well. Cool.

Rick (26:24.556)
Yep, I'll find it. I thought it was a good idea for a, you know, a AI tool. Like, I mean, that's kind of cool, right? Like if you have like a really.

Tyler King (26:32.954)
Yeah, if it works, it's cool. I'm doubtful, but yeah, if it works.

Rick (26:37.826)
Listen, I don't know how much you want to talk about AI today, but like I could talk about it. I am very impressed. the, the, my adoption of AI to do things and thus my productivity is like skyrocketing exponentially right now. and it continues to get better every day. So I'm not, I'm pretty high on this stuff and like try, like, even if it's just like, you don't use it to do the whole thing, you use it to accelerate a piece of it, just like start with an outline. It's, it's like so high leverage.

Tyler King (27:02.339)
Yes.

Well, so let's talk about this. So I agree. there's no question. AI is magic. There are some things it's so incredible at. I don't want to come across as an AI skeptic at all. I think it's possible that the models are not going to keep improving. Like it's like chat GPT at least has really like not gotten a lot better as far as I is like my usage can tell. Maybe, maybe that continues. Maybe like, but let's imagine for a second at plateaus. It feels to me like.

2008 and the movement to SaaS, which is to say there is this technological shift that was required of like everyone having broadband internet, server, AWS, easy deployment of web apps, all of this underlying technology, the Ruby on Rails, all these frameworks, all of that had to exist first. But then once that existed, there was a separate

thing, which is taking all that technology and using it to reproduce, like recreate every actual product. Right. So there's an on-prem S like SAP on-prem, like, ERP system, make a web version of that, make a web version of, well, Salesforce already was, but like make, you know, take thousands of different like legacy B2B apps and bring them, use this new technology to bring them into the cloud.

I kind of feel like that's where we are with AI, which is to say the underlying technology, phenomenal, but no one's actually built any of these. No one's actually moved it into the cloud, so to speak yet. And almost everything is like one shotting prompts. It's like, give the prompt, it gives you the video and you can reprompt and it'll regenerate the video and you're done. But like, and maybe the tool you're talking about does this, but I doubt it. What you'd want is it doesn't generate a video. It, there is a...

Tyler King (28:57.552)
video editing software that the AI controls and it makes the video with like layers and timelines and then you can go in and edit it. I have yet to see any AI tools that work that way. Have you?

Rick (29:12.526)
Let me just like recap what you're saying to make sure I understand using the video as an example. Let's take chat.gbt. Let's say make me a video chat.gbt. The video comes back and you can only reprompt a reproduce video. You can't actually like take over the video production and finish it for lack of a better word. What you're imagining is take with listia. If they were going to reinvent themselves, what they would do is they would take a prompts to edit the video, but then let you edit it on top of that. Like it would be like

You're editing it. The AI is editing it. We're working together. Is that what you're saying?

Tyler King (29:44.858)
Yeah. Yes. And like, let me add onto that. An analogy for this might be, sorry, my fan in my office just got a lot louder. Can you hear that?

Rick (29:57.198)
It came on and then I think it's being buffered out right now.

Tyler King (30:00.09)
Okay, well hopefully the post-production removes that. Bitmap images versus vector images. So like you can have a photograph that's like got all these pixels and it's, let's say it's a picture of a flower and like the flower is pink but the background is blurry green with all these grays and whatever. And you can go into Photoshop and like apply filters to it and stuff. like fundamentally Photoshop doesn't have any information about the flower. It just has the pixels.

Or you can do a vector version of it where you're drawing it like an SVG or something like that. And you're like, okay, the flower is actually made up of all these little shapes. And if we want to scale it up or down, or if we want to change the color from pink to red, it's not like trying to interpret all these pixels and turning the pinks to red. It's like, I know how to draw this flower from scratch. I'm going to turn it into a red flower. That's like the metaphor for what I'm talking about with video where the AI would be generating the vector image, not

bitmap image so that then you can edit the vector later.

Rick (31:02.562)
I got it. Yeah. No, think I think it's I would be very surprised if there are not tools doing that. What? How well they're doing it. I don't know.

Tyler King (31:10.266)
Yeah, I'll at least say I haven't seen it. This is my main complaint about most of the AI tools I've seen is they're very lazy about the tool. There is no tool itself. It's just sending a prompt to a model, the model sending a thing back. But this is where I'm seeing a lot of improvement like cursor, the code editor getting way better than it used to be. Because back in the day, I feel like you'd give it a prompt and it would just one shot, here's some code changes. And now you give it a prompt and it's like,

First, you can tell it's doing something behind the scenes or it's like, first thing is we need to find the right file. And that's going to use some AI to figure out what grep command to send. But the grep command is a normal terminal command. It's not AI. It's just like the AI is calling a real thing. The real thing is getting results, passing it back to the AI and then the AI is like having this exchange between real code, normal tools and the AI doing its magic. Cursor has gotten way, way better because of that change from my perspective.

I have, it is happening, but I feel like a lot of the ways in which AI promises are falling short is products that aren't doing that really.

Rick (32:16.59)
Yeah. I'm thinking about like a windfall example of AI. windfall at its core was, is an AI company. you know, we do model data attributes for us households. So like that's a model, like that's predictive AI for, for lack of like, it's a machine learning, and LLM. Yeah. Well, not, not, no, no, not LLM, but we do leverage LLMs for our generative AI. So once you have a really strong data set,

Tyler King (32:34.789)
Is it LLM?

Rick (32:44.278)
And you have a customer using that data set for a use case. can then insert generative AI concepts to make use an action of the data for the customer's use case. So an example of this would be, let's say you are a nonprofit fundraiser that has purchased WinFall. We now have an application that you can go into and we know you've connected your CRM. So we know about, you your past donors. know about your prospective donors. We know about like the history there. And we also have our data.

match to that. We can do two things. One, we can sort of go, you know, here's some ideas of segments that we could create of your database that have opportunity for increased fundraising based on historical fundraising, based on our data, based on, you know, all the things. We can also have like a chat prompt that actually lets you explore the data and then automatically create these segments versus.

If you go into HubSpot or any segmentation tool, building a segment from scratch, like clicking all the attributes, like there could be 30 attributes you want to like go through heuristics to make the segment. It's very time consuming. And also you're limited by what you can think of. so in this case, like what, what Windfall will do is we will actually create the segment for you, show you the results of the segment, but the segment is then editable or clonable. and so it's an example, like it's, it's, it's, not necessarily as advanced as a video editing or

or a, what do you call it, image editing, but it's the concept of like, we're gonna help you get to a first version and then allow you to edit that. That isn't text, right? It's actually using the tool to make it better.

Tyler King (34:19.612)
Mm-hmm.

Tyler King (34:26.042)
Yeah. Yeah. So great example. That's awesome. It sounds like that's a, that is a very modest. And I mean, this is a compliment. That is a modest use of AI, which is to say, I don't for a second think bullshit. That doesn't work. Right. Like, yeah, you can explain to that. You can give a prompt that says here are all our filters. Here's a little information about the data. Tell us what filters to turn on. And I think that's a great example as well, that you have to have the UI, the app with the filters.

so that the AI can control, you can almost think of it like a recording studio with all those knobs and stuff. And the AI is telling you which knobs to turn. It's not just giving you music, right? So yeah, it's certainly possible. I think even if AI models stop improving right now, which isn't, I think, unrealistic, right? With how expensive all the, like, there could be a situation where the bubble bursts and there's just not enough, like the data centers stop getting built.

there's not enough GPUs and like stuff just kind of stays where it is technologically for a long time. I still think there's a decade of opportunity similar to the SAS wave. I think we're just scratching the surface of how this stuff's actually being applied. just, people are getting so far ahead of themselves about what it can do, but we haven't actually done most of that work yet.

Rick (35:47.15)
Yep. No, it's, it's, it's like, I have a hypothetical. I just, have you were, have you, um, what came out last month was, um, there's a guy named Benedict Evans, um, does an AI eats the world presentation. Like, you know how that, that old software eats the world, uh, blog he upstates it. think once a one, every other season. Um, have you seen this? Yeah, I'll, I'll send it to you so you can put it in the slide notes. think everyone should look at it. Like if you're listening and you're interested in this, cause it's like, it's very insightful in terms of the trends. Um, one of the, the

Tyler King (35:59.355)
Yeah.

Tyler King (36:05.488)
I know his old stuff, but I haven't seen anything recently.

Rick (36:16.974)
There are two things that are very interesting to me from the presentation. one is that, the initial wave of application for AI is that it's mostly, automation use cases or smarter automation use cases or higher leverage automation use cases, in software. it's not necessarily like replacing the human. just making the human like 10 X faster or some multiple multiple faster.

So, that's where the hanging fruit is right now and where the application is. And so if you're trying to go, I think that's kind of sobering to what you're saying. It's like, it's not the next thing, which is like. Generation, like it's like human replacement or like exponential generation of new ideas. It's actually taking what we're already doing and making us do either it faster or more or better. and, I buy that. And I think there's a lot of applications within all of our days for that. and.

I find new, I'm still finding new ones every day, but it isn't a standalone necessarily like thing that's just like so obviously like transformational. You have to actually go and think through like, what are the things that I'm doing today? What are the things that AI could replace or automate? But a human can still supervise. Where do you actually need a human to supervise? It requires this like thoughtfulness, but it's more of like an operational question more so than like a, what I would call like a product development question. Does that make sense?

Tyler King (37:33.724)
Mm-hmm.

Rick (37:44.582)
So that was interesting to me from the thing and I said a lot of opportunity for me in the revenue operation space that I play in the second big takeaway is in commercial real estate We've had this like interesting thing with kovat, know of like one the doubt like kind of the reduction or flattening of Commercial real estate development. How do they fill these big buildings that they built right? second Retail is stalled out too because people are buying online

Tyler King (37:45.179)
Yeah.

Rick (38:13.996)
Right? And the shipping crap, like the boxes are coming to your house versus like, so what is starting to outpace? What is the leading area of commercial real estate development? Data centers. Yeah. And it's like, yes. So I'm having matrix visions. I'm having matrix visions.

Tyler King (38:17.756)
Mm-hmm.

Tyler King (38:24.09)
I data centers, assume is what you're gonna say. that's terrible. They're all empty. There are no humans in them. Yeah. Yeah, there's a separate discussion from is it useful to like, is it good for us? And it's like, to me, definitely no. absolutely. If I could hit a button to stop all this progress, I would. I'm not gonna bury my head in the sand. Like, there is no such button, but it doesn't lead to good places, I don't think.

Rick (38:34.342)
Terminators. Yeah.

Rick (38:54.286)
I don't know, you remember Paul's version of unlimited wealth and the pie gets bigger and we figure out how to address this stuff. I have an optimistic view.

Tyler King (39:01.414)
But the pie has gotten bigger and bigger and bigger in our lifetimes and it has not helped the bottom 50 % at all. Like I don't doubt that AI could lead to abundance, but I don't think that abundance would actually get distributed to people.

Rick (39:14.808)
There is, I'm not gonna, I don't have a strong opinion here, but I would just say like devil's advocate is that the, some portion of the bottom 50 % have the, like today can live like a king did a thousand years ago.

Tyler King (39:28.112)
Yeah, that was true when we were born as well. I'm not saying like capitalism has never done anything good for humanity. I'm saying like late stage capitalism that we have right now is completely corrupted. Like it's, for example, it's supposed to be the case that competition drives prices down. There are so many different examples of industries where it's like, how come that's just not happening there? You know, like I believe in the vision of it and that vision has been true in the past.

Rick (39:30.232)
Yeah.

Tyler King (39:56.656)
But right now I'm very, very skeptical that if the three biggest tech companies on earth develop AGI, that that's gonna turn into general abundance for all people in the world.

Rick (40:06.808)
general abundance is different than what I'm saying for sure. But I would just say like, yeah, I just generally think that like technological advancement generally allows like comes with the ability to deal with the ramifications of tech, the said technological advancement. That's an optimistic view.

Tyler King (40:09.306)
Yeah.

also... sorry again.

Tyler King (40:24.764)
I mean, the argument is it's always worked out and it has, but does that mean it always will? like, so Paul's example from his book was like, humanity thought we were going to run out of, but for people listening, Paul was our former boss who wrote books. He was like, rubber was the example. We started using rubber to make car tires or whatever and other stuff. And then it was coming from the rubber tree and there's no way to supply the world's supply of rubber with rubber trees. then.

Rick (40:39.299)
Mm-hmm.

Tyler King (40:53.666)
Everyone was like at the time the world's about to end like we're about to run out of rubber and then we just figured out how to manufacture rubber out of other stuff. It's not technically rubber. Yeah, it's happened a million times over and over where there was this panic of some resource getting used up or this is going to destroy jobs and all that. Fully acknowledge every time in the past it is either been a net positive or it at least hasn't like no single negative has overwhelmed the positives.

Rick (41:03.298)
Same thing happened with call, know, call, yeah.

Tyler King (41:23.59)
But like that doesn't mean it's always going to happen, right? AGI in particular, I don't think we're in danger of AGI anytime soon. AGI would be a completely different thing from any other technological advance ever, in my opinion. I'll also say, I think work is noble in and of itself. A lot of people are like, I want AI to automate away work so I can make art all day. Fuck that, I wanna work all day.

I'm bummed about the fact that I might not get to work anymore.

Rick (41:52.846)
Yeah, yeah. Well that kind of brings me to like a more philosophical question. I'm interested in your take on this. Do you feel like, I don't know, let's just say that this thing's inevitable. Are you working in the right domain today to be able to continue to work? Where should you be working?

Tyler King (42:06.841)
no, absolutely not.

should be working at Google. Or you should be buying stock in Google. That's the thing is there will be no... Sorry, you're saying if AGI happens, right?

Rick (42:12.039)
Yeah, Ed, one of the big three. Yeah.

Rick (42:19.15)
I'm just saying let's assume that AGI is inevitable for a second and like that plays out like the maximum possible thing plays out in terms of what AI can do.

Tyler King (42:22.298)
Yeah.

Tyler King (42:31.224)
AGI blows up. So right now the narrative is, AI is not going to replace people. It's just going to like supercharge you, but you still need a human to supervise it. AGI by definition means that's not true anymore. There is nothing a human can do that AI can't. That's by my understanding, that's kind of a simplistic definition of what AGI is. So there are no more jobs, like period. And like not only that, but not only do I not need to exist because less knowing serum doesn't need me, but we no longer have any customers because no small business exists anymore.

Just there is no more economy. And again, that could be a version of abundance where we just have robot like servants waiting on us all day and we get to do whatever we want. I doubt it turns out that way. But no matter what, there's no justification for any human labor at all at that point, right?

Rick (43:16.206)
I guess the question becomes like what do we do? What is our purpose? a big part of our purpose. mean there's Maslow's basic needs, right? Like those are taking account. You still have these other needs, right? Like what do you do? Where should you be like? Like what's the next advancement?

Tyler King (43:31.58)
Yeah, I think humans evolved to need challenge in their... This is why I think it's good for people to work. I know a lot of people don't like working and like, sorry, it sucks to be you, I do, good for me. But like, humans did not evolve as much as it seems like it would be an enlightened existence to not need to work. We evolved to get bored very easily and I would not look forward to a world where...

Rick (43:44.43)
You

Tyler King (44:01.612)
all labor and not just I'm not just talking about your day job from nine to five I'm talking about you don't have to do chores anymore you don't have to call up the bank and whatever like you just every single thing is automatically taking care of you for you I think that for many people myself included there's just no like that would just be a bummer of a life

Rick (44:21.11)
And what would you do? I guess that's I'm like, is it that you would do? You'd make up work. Yeah.

Tyler King (44:23.226)
I'd make up work. This is what games are, right? Like, why do we play games? It's fake work. It's like, have to think and I might lose. Yeah, I guess. Assuming the robots let us, right?

Rick (44:29.782)
Yeah, so we basically become an awesome, we play Call of Duty all day. Yeah, like pretend war. Basically it's sport, it's video games, it's competitive gameplay. Basically it's fake stuff. Fake war, fake war.

Tyler King (44:42.524)
Yeah, I think so. How many video games are there that, like, I play Stardew Valley, it's simulating farming. Like, there's this, the truck simula- there's all these games that are just simulating mundane day-to-day life. Yeah.

Rick (44:58.222)
Yeah, that's exactly what, yeah, that's what we do. We'll play games all day, which is basically what, Ready Player One?

Tyler King (45:04.437)
I haven't read or seen that, is that what it...

Rick (45:06.849)
Hmm. I don't know. I haven't seen it in a while, but basically it's a virtual. We live in a virtual reality that dominates the world.

Tyler King (45:14.748)
Yeah, I mean, I guess you were saying the Matrix earlier, like in that utopia that the robots create in the Matrix, they're still working boring desk jobs, right?

Rick (45:17.208)
Mm.

Rick (45:24.779)
Yep. And I think it's real.

Tyler King (45:27.63)
Well, this really went off the rails here.

Rick (45:29.163)
Hahaha

Rick (45:33.998)
Tyler King (45:34.98)
You got anything else quickly or should we call it?

Rick (45:37.486)
Well, I think like I just kind of there's a I think we're I think we're very far away from the world that you view like I don't know if in our lifetimes Yeah, like we're gonna experience that so in the short term like I still think like it's a big opportunity to adopt AI to automate things in the short term and to be vigilant about where there is professional opportunity both within your current startup or new startups or other startups

Tyler King (45:46.5)
No, I...

Rick (46:07.022)
And I think automation is the obvious one right now, like speed, human efficiency. like, I'm curious like when, I don't think that's going to change for a while, I what I'm trying to say. And so that's where I would focus.

Tyler King (46:19.74)
Yeah, I agree. I have no direct expertise, but it seems like all of the experts agree. LLM, whatever that means, is not going to lead to AGI. Maybe there's some other AI thing that will come up, but like, they're not just gonna keep improving these models and eventually have AGI. It seems like pretty much everyone's in agreement about that. But yeah. And I...

Rick (46:41.036)
Yep. Thanks for going. I've been wanting to talk about this for a few weeks, three months. And like, I feel like this was the right time to talk about it. But

Tyler King (46:46.496)
Yeah. Yep. Let me also say, I probably sound like a doomer about all this stuff. For a long time, since 2010, I've been bummed for the future of entrepreneurship that they're just, the wave of opportunity you and I experienced when we were entering the workforce has not existed since then. Everyone thought it was gonna be crypto. Not everyone, I never did, but.

A lot of people were talking, they got so excited about crypto because they were like desperate for something. They were like, there's no opportunity out there. Everything's just mature and boring. Let's do crypto and crypto is all bullshit, obviously. I think they're, but Bitcoin's bullshit too. It's worth $0. People are just gambling. Okay, moving on.

Rick (47:24.696)
I mean, Bitcoin's not bullshit. Okay.

I mean, if you were smart enough to invest in Bitcoin.

Tyler King (47:34.812)
If you invested in Ponzi scheme and got out at the right time, that worked out too. You can make money on all kinds of bullshit, Yeah, well, I mean, if you can give me an argument for why it has intrinsic value, I'll be happy to hear it. But my point, a lot of people got excited about a bunch of, forget Bitcoin, like the NFTs and all that stuff. AI actually has created, again, I think a lot of people have this perception that

Rick (47:38.956)
It's been a decade! It's been a decade!

Rick (47:48.046)
I can't do that.

Tyler King (48:03.524)
AI is already over, that all the land has been grabbed already. But I want to just reiterate apps that turn the knobs rather than generating the music. That's the opportunity. There are a million different ways, like different businesses like Lesson Knowing CRM that could be built on top of that. And that as far as I can tell, it's still wide open.

Rick (48:27.118)
Yep, agreed.

Tyler King (48:29.798)
Okay, I'll get off my soapbox.

Rick (48:32.534)
Anything else? Do you have any other quick ones?

Tyler King (48:35.098)
No, I'm all worked up here, Rick.

Rick (48:38.626)
I'll send you the link for the latest AI eats the world presentation so you can put that in the show notes. I'll look for the tool as well. And if you'd like to review past topics and show notes, visit startflask.com. We'll see you next time.

Tyler King (48:42.618)
Okay, thanks.

Tyler King (48:50.854)
Bye.