Rick (00:00.332)
What's up Tyler?

Tyler King (00:01.602)
Yo, yo, yo. If you hear screaming in the background, it's because we've got a squirrel running around our office right now. Yeah, it's actually been here for about a week, but one of the private offices is like has a closet to a utility closet and, there's no door on it. And so the squirrel got up through there and, it's been stuck in that private office. It's not like we're running around terrorizing people, but yeah, there's a, there's a squirrel running around.

Rick (00:08.27)
Are you serious?

Rick (00:27.374)
Did you call animal controller or are you gonna try to get it to go out a window?

Tyler King (00:30.252)
Well, we called our landlord and they're not doing anything. So I don't know what to do now. I mean, I guess, does animal control just do private commercial buildings? Will do anything? Yeah. This seems like a landlord problem to me, but new ownership in the building and they are doing nothing. Guy sold it. Yeah. So, yep. It's a bummer. He offered it to me, owning a 10 story office building does not.

Rick (00:39.926)
I don't actually know.

Rick (00:48.73)
your guy sold it.

You didn't want to be a buyer?

Tyler King (00:58.796)
sound like fun to me nor do I have any money with which to do that but I guess neat thing to daydream about.

Rick (01:07.224)
feel like most people who buy those buildings don't buy it with their own money.

Tyler King (01:10.762)
Yeah, sure. But then you're just in a totally different business. Actually, like unrelated to anything we normally talk about. have I told you the so the original owner who is kind of this old quasi retired guy, they tell you why he got into real estate in the first place. His main business was a rigging company, which is like, you know, when you're on the highway and there's like a massive like line of trucks carrying like a windmill blade or something like that.

Rick (01:14.381)
Yep.

Rick (01:26.678)
You never did,

Rick (01:37.603)
Yeah.

Tyler King (01:38.846)
That's what his main business did. So he had all these laborers for putting stuff on these big trucks, but they basically don't do any rigging jobs in the winter. And so he's like, how do I keep these guys employed? I'm going to buy a building and use these guys to fix it up. And so he owned like half of downtown St. Louis at one point, just to keep these guys employed so his rigging business could do well.

Rick (02:02.071)
seasonal seasonal business solve into what do you call it when you're like metropolitan like mega Lord mogul yeah but the business mogul that's hilarious

Tyler King (02:04.362)
You

Tyler King (02:09.39)
He's a mogul. He's a mogul. Yeah. Yeah, I don't have that kind of like multi-entrepreneurship in me. I got to stay focused. Anyway, what's going on with you?

Rick (02:23.105)
Yeah.

Not much. Just it's our end of quarter at windfall. definitely burning the midnight oil right now and trying to do a little too much.

Tyler King (02:36.76)
Seems like it's always the end of a quarter.

Rick (02:38.761)
Yeah, it is every three months.

Tyler King (02:44.302)
But like how much does the, like, is it a few days? So if it takes a month to prepare for the end of the quarter, then you're spending, you know, a third of all your time on it. Like, how, last two weeks?

Rick (02:54.253)
Yeah, it's usually the last two weeks that really ramp up primarily due to deal deal velocity and end of quarter goal pushes. So it's nothing crazy. It just means I need to stop trying to do other things while I get through that. So it's okay. But we've made some progress at like up health and I I forgot about my coding journey update. I did not do anything on that. I feel guilty.

Tyler King (03:00.792)
Hmm. Hmm.

Tyler King (03:10.168)
Yeah, sure. Cool.

Tyler King (03:22.062)
You didn't do anything. Shame.

Rick (03:24.245)
shame. Yeah, well, here's what I did do. I did use Claude a lot. We have so in order to get artisan built out artisan is our AI SDR tool that we set up for leg up health. It requires a lot of contextual prompting in order to work. And if you don't have that like really dialed in, it can only use the word hallucinate, like it can get the it can kind of go the wrong direction in terms of

what companies are targeting who who at those companies is reaching out to and then also what what it's saying to those people. And so

Tyler King (03:56.974)
And it's so it's finding people and generating a custom cold outreach email for each one.

Rick (04:05.227)
Yes. And I have exciting news about about that tool, which I'll get I want to talk about today. But that's the gist is basically it's there's a one point one point zero and then Ava 2.0, which is which is artisans new platform that's going out in May. But the 1.0 is basically email and LinkedIn outreach. And so the the the AI is doing

two things, it's identifying people to reach out to. And then it's deciding what to say to those people, you can guide it in terms of there's all sorts of like context you can give around like, who to target. And then you know what to say to that person, but it's very campaign specific. So if you want to target different people and say different things, that requires different campaigns with different contexts. And then for the outreach itself, you can set like, you can say, Hey, Ava, do whatever you think is best.

Tyler King (04:48.366)
Mm-hmm.

Rick (05:04.897)
which I don't think is the right way. Or you can like kind of in a flow chart diagram design the outreach cadence and then within each step, provide guidance on how additional guidance that's that's more narrow than the broad guidance you provide for the campaign about that step. And we'll probably have like a, you know, we'll probably take that approach is where we're, not giving her the exact email to write. But we're giving her the prompting per per email or outreach.

Tyler King (05:07.235)
Yeah.

Tyler King (05:34.018)
Yeah. So you'll say like, want a seven email cadence that the second one should be more conversational and the third one should have more of a pitch or whatever it happens to be.

Rick (05:44.457)
Exactly. And then she'll decide how to personalize it at the beginning. She'll, she'll decide the subject line based on what she knows about the person. She'll decide what the right call to action is based on, you know, that kind of stuff.

Tyler King (05:55.214)
What's the mechanism for, I assume if someone replies to one of these emails, you want to stop the campaign. How do you do that?

Rick (06:00.053)
Yeah, so they handle all that anytime there's a reply, the person's removed from any further outreach basically does a hard block for the account for us to ever reach out to them again and from artisan.

Tyler King (06:09.784)
Do know how it does that? Like, is it integrating with email, integrating with the CRM?

Rick (06:14.069)
Yeah, you're so they set up their way that the descending works is one of the biggest challenges right now with scaling up outbound is sender reputation. So if you send from your own domain, lots of spam, it can make your email go into purgatory. No, they set up 10 domains for us like go leg up.com go leg up health.com, etc. And so they are managing all that. So they get all the replies. And they have an inbox system that you go into and reply.

Tyler King (06:29.902)
So you have a separate domain for this. Okay.

Rick (06:43.83)
reply to stuff in.

Tyler King (06:44.952)
So they just send you a separate email saying, you got a reply, go in and answer it.

Rick (06:49.613)
Correct. Yeah, there's we'll probably use their slack integration to get notified of that. But but the exciting thing is on Ava 2.0. They're adding all sorts of cool features that are relevant to like health. You know what the most important feature that they're adding is?

Tyler King (07:07.692)
I'm hoping it's the ability to link to custom quotes.

Rick (07:11.019)
Well, they can maybe do that. I don't know. But what they what I was thinking was, one of the things that we want JD to be able to do is basically in while Ava is reaching out to all these people via email and LinkedIn is to make dials. But it's super hard to know who's where, where someone is in Ava's outreach in 1.0. And so they're actually launching a dialer where you can actually have a human step in the flow flow to make a phone call.

Tyler King (07:14.103)
Okay.

Tyler King (07:36.504)
Mm-hmm.

Sorry, I'm, I kind of high level get it, but so JD's calling or an AI rep is calling. Okay. And the idea, the idea is that you want the call to come right after the email. Like that's just a known sales technique.

Rick (07:43.809)
JD, human assisted calls.

Rick (07:53.934)
It could be anywhere, but you want it to be coherent with the outreach.

Tyler King (07:59.056)
so you're saying like, okay, it's two emails, then a call, then an email, then a call or whatever. And so it won't do the fourth email or the third email until the call. Okay. Gotcha.

Rick (08:05.185)
until JD does the call. Yeah, and it'll but more importantly, it'll basically provide JD with a call list every day. That he just works through and it does all the dialing for him. He just has to like, like log in. Yeah.

Tyler King (08:11.982)
Mmm. Wow.

Tyler King (08:17.836)
Yeah, that sounds pretty great.

Yeah, our customers ask us for like pre-AI, there were automated dialers, I think is the main term for this, where, you know, let's say you have an integration with the CRM and you put 10,000 leads in your CRM that you bought from some terrible source. And it, my understanding of how these tools worked is they would dial five, they'd figure out the metrics of how many people pick up, right? And so they dial, let's say 20 % pickups, they dial five at once. If two pick up, they just hang up on one of them.

They're just like, nope, sorry, someone else picked and, then it connects it to a human. This feels like a much, as much as like AI is ruining everything in terms of like spam, this feels like a much more human approach to that type of workflow.

Rick (09:05.549)
Correct. Yep. So it should be good for JD because should help him maximize this time on the right people. And then there's a bunch of other updates like so so I guess going back to like what I was originally talking about was for artists and to work, we have to have a lot of our messaging and industry like what is a good company to reach out to who is a good person at that company to reach out it's gonna be dialed in so that an AI can be successful. And JD and I made a lot of progress on that between the last podcast on this podcast.

Tyler King (09:29.582)
Yeah.

Rick (09:34.798)
had Claude write the first version, I was like, Oh my god, you're trying to do too much. So my kind of AI update this week is when you try to when you start with a blank slate with Claude, it tries to do way too much. It like tries to solve the all the problems. And so I'm like, stop, can we please? Over editing? Oh my gosh. And you can't and then you like want to wait for it. But at the same time, it's like you're consuming all these credits without even asking me why. And so

Tyler King (09:46.37)
Yeah, I think the term for this is over editing. Yeah.

Tyler King (09:57.229)
Hahaha

Rick (09:59.726)
I had a couple times I had to stop it and be like, just give me like a framework. I'll go do the like write the first draft of the first section, then I'll give it to you. can improve it. And so I kind of got once I got into that now Claude is smart too.

Tyler King (10:11.214)
Yeah, definitely, mean, it doesn't work 100 % of the time, but definitely telling it what the constraints are in advance, I feel like is a important part of using it. Especially with coding, this goes crazy because if you have a big code base, there's already bugs in it for sure, right? Or just like code that doesn't meet the standards and it'll just trace this to all these different files and it'll just make a million changes that have nothing to do with the project you have to do. So yeah, you have to say.

Rick (10:37.583)
Exactly.

Tyler King (10:40.29)
For starters, don't write any code. I will not allow you to write code that we're just planning. And then once the plan's done, each task needs to be like, don't do anything except this task.

Rick (10:49.165)
Exactly. And I just I oftentimes just want to get the conversation started. And so I jumped the gun. The other thing I don't like that it does is it produces word documents, and it takes a long time to produce them. And they're and I'm like, I just want it in markdown, man, like, please just give it to me a markdown. I wish you'd just put it in a Google Doc, but it can't seem to do that.

Tyler King (10:54.222)
Mm-hmm.

Tyler King (11:06.83)
So this is why one of many reasons to use Claude Cowork instead of the chat Claude. Well, first of it can write to markdown files, no problem. It's meant to work with files on your computer. Whereas the web one, as far as I understand it, I'll admit I have a shaky understanding of this, but it's got its own little sandbox with its own set of tools. So I just had Claude generate a PowerPoint presentation for me and it was the same thing. It was really slow.

Rick (11:12.022)
Okay, explain.

Tyler King (11:35.98)
And I asked it why it was slow and it's what it's doing is it's generating a JavaScript like it's writing a code base in JavaScript that then uses an open source library to generate the document. And so every time you want to edit the document in any way, it's rebuilding this code from scratch. Like, or that's what it told me anyway.

Rick (11:54.647)
Yeah, that makes a ton of sense. exactly what it seems like it's doing.

Tyler King (11:57.9)
Yeah, but if you just use co-work, can, or I use cloud code more than co-work, but I think they're more or less the same thing in this regard. You can just be like, here's a file on my computer. This is the one we're working on. Edit it. And if it's marked down, it's marked down.

Rick (12:10.559)
That's cool. Okay.

Tyler King (12:13.088)
I would strongly, if you just want a little baby step, just every time you go to Claude, use Claude co-work instead. It does everything that the chat version does, but then if you find yourself wanting a file, you can point it towards like a, so I would make a project folder in your computer and then each thing you're working on with Claude, will make a new directory, like make a new directory in there and say, here's where we're working. And it'll put all the files that it makes there.

Rick (12:40.983)
I'm putting a note down to just say like use cowork.

Tyler King (12:44.694)
Yeah, and I'm not, again, I'm not an expert on this. I'm sure some of our listeners are listening and either I'm getting stuff wrong or they're like, there's an even better way to do it. But I think that's a step.

Rick (12:53.943)
So one thing that's coming out of the messaging document is one, our, it's, it's getting into like who are top personas, like the, like the idea of a champion who's like the office manager versus the founder CEO versus like the VP of finance, if it's a larger company and it's all validated by JD's actual customer experience now versus hypothetical, which is the first time I think we've done a messaging exercise in the last five years. It's like not hypothetical. So,

We also have like six industry verticals that we're prioritizing. Top one being kind of startups and series A, series B companies. Sorry, and bootstrap companies over 3 million. thing that I want to take this to like the coding conversation again is I want to basically have Claude rewrite our website, restruct like, like review our website and provide improvement ideas.

But I'm wondering like is Webflow the right place? I'm coming back to I want to do this project for my personal website to learn to code. What is the right way to do this? Should I really build a CMS from scratch with Claude code or should I? Not do that and leverage a platform that has a little more structure, you know and call the you know, call the API. Do you know I'm saying?

Tyler King (14:16.13)
Yeah, I do. What's the goal here? It feels like outbound is the main thing anyway. Why does UpHealth need... Last time I said I was working on my website, you were like, this even worth it? Is it worth it for you?

Rick (14:20.279)
Tooth pr-

Rick (14:28.193)
Yeah. Well, I mean, people are coming to the website, we actually get a decent amount of traffic to by the way, did I tell you we got a lead from all that blog post I was right? You remember I was doing like 30 days of straight blog posts. We got an AI search to one of those blog posts, and it converted into an opportunity for JD.

Tyler King (14:46.862)
Cool. So someone asked the AI a question. That question was answered in this blog post. They clicked through. The thing that I think Google did very well in AI, you generally think this doesn't happen. People don't click through the sources, but I guess it did happen.

Rick (14:59.787)
Yeah, he was like, I was like, tell me about Utah. And it was like, ICRA. It was like an HRA related post. So that's how I know it was. We didn't have any HRA content on our site before that. And he requested a quote on the website. Yeah, so

Tyler King (15:15.192)
Cool. So your goal is to spin up a inbound SEO ish marketing channel alongside JD.

Rick (15:23.309)
Two, two, two items. One, I want the messaging that we're going outbound with to be consistent with the messaging that if they go, what's like up health, I'm gonna Google it before I reply to JD, that it's consistent. And I don't think it's inconsistent right now, but I just, I have a feeling like it could be more consistent, that kind of thing. I guess I'm actually less concerned about this for leg up, like ups website, I was more tying it back to the coding thing.

Tyler King (15:43.736)
Yeah.

Tyler King (15:49.737)
Okay.

Rick (15:51.502)
more like the approach because like I it's very easy to go edit what the like in 30 minutes the copy on like a pelt.com but it's just like yeah

Tyler King (15:58.85)
Yeah, that's what I was gonna say is you don't need to even touch the CMS for what you just said for leg up health. You need to change the homepage and nothing else. But, okay, why?

Rick (16:06.635)
Right. Yep.

Rick (16:11.501)
But with my coding project, my coding project was going to be build a new website for Ricklinquist.com. I'm still like hesitant to, I guess like I just need to go do it. And if it doesn't work, I can always, yeah, okay.

Tyler King (16:21.89)
Just go do it. It'll take a week. mean, AI is crazy and throw it out. I actually, have a topic very, very related to that, but yeah, I think in the past there was a, there was the sense of you have to plan and then you have to execute. And I think in some cases AI flips that upside down where it's faster to execute first, see what it looks like, and then use that to inform your plan. And then you re execute again afterwards.

Rick (16:48.013)
that's cool. All right, well, I'm gonna give a try. I hope I will have an update on my coding journey next week. So, and the only other update I had on like a health was we did bring on an intern. Did tell you this?

Tyler King (16:56.174)
Nice.

Tyler King (17:03.446)
Yeah, I saw that in Slack. know nothing about this. So tell me more.

Rick (17:06.733)
Declan is my brother. He is a great. He is going to play. He's like a senior in high school. He's going to play basketball in college. He's like super good, like straight A student, just like very good at just doing work and like having good attitude. And so he called me and said, I'd like to help you this summer. Could you have any work for me? And I talked to JD and JD was like, sure, this would be great. And so he's

Tyler King (17:10.708)
Nepotism, okay.

Tyler King (17:30.158)
Nice.

Tyler King (17:35.288)
So wait, wait, wait, most importantly, where is he playing basketball?

Rick (17:40.314)
Haverford. Yeah, which is in Pennsylvania. And then let's see. Yeah, the most important thing. But anyway, he started this week and he is we are we kind of JD and I are looking at two ways one, like, I think we'll get value in terms of helping JD build target lists that we can either feed into artisan or JD can work himself. And then the second thing is, we'll learn how to bring on this type of talent.

and have the resources to do it again if it works.

Tyler King (18:12.428)
Okay. So sorry if you said this and I was distracted by basketball. what specifically, like he's helping JD, but is there kind of constraints around what type of work he'll be doing?

Rick (18:21.739)
very specific. It is basically three buckets of things. The first is identifying. So in our doc, our document for artisan, we have basically three tiers of companies tier one is 15 to one to 50 employee companies in Utah that are in our six industry verticals. And so basically, like if he did nothing, other than this, identifying all of those companies in Utah, putting them in a spreadsheet, and profiling them across, you know,

What do we know about their number of employees? What do we know about their zip code? What do we know about their website? Like all this kind of stuff.

Tyler King (18:58.668)
Is this not what Artisan is supposed to be doing?

Rick (19:01.401)
Yes, our since doing this, but they're doing it at a black box, we don't get the data. And so they will do what they do, but we will not know unless someone responds, we will not know who they're reaching out to. yeah. I don't think so. Yeah, the and then so but but the the other factor is that we can trigger even if our son doesn't reach out to them on their own, we can actually direct our son to reach out to a list.

Tyler King (19:14.059)
and they never, there's not a way to get that information? That seems limiting. Okay.

Rick (19:30.773)
And so, we can be more my experience working with artists in other places has been, I call it like a retargeting campaign, but if you can provide it a list to target, the results are better.

Tyler King (19:44.175)
Okay, so Artisan is actually running a data business disguised as an SDR business. They're trying to build a really proprietary set of everybody. They won't share any of it, so you can't exfiltrate any of that info. And then that way, the only way to get access to their data is to buy their SDR service. That's pretty clever.

Rick (20:02.399)
I don't know that that's the they haven't stated that I'm sure I'm sure that like they're they're behind the scenes of vessel presentation is look at our co-op model. But but I I imagine that like, I don't know, I should I don't I just want to state for the record, like I don't actually I have not if it is possible to export people that are being reached out to that have not responded. I have not figured that out. But but I do know that like once someone responds, like they tell you who it is. it's like,

Tyler King (20:04.341)
come on. You know, that's what they're saying in their board meetings.

Rick (20:32.107)
Here they are.

Tyler King (20:36.334)
Okay, so he's, your brother is going to be putting together like data on kind of ICP type targets.

Rick (20:45.367)
So first is like company level and like we could go like that saves JD a ton of time. But like if I was gonna like, so top down, it's like generate lists of companies, find the people at those companies that meet our key personas, profile those people, and then identify the timing elements of when the right time to reach out to those people are. Those are like the three things. Top down.

But like I was going to say, like where we want to ramp him to, like, it's really hard to just start on day one and provide value. So we're kind of having him work down that list. Like by doing those three things, he's going to get good at knowing like what a good lead is. Ideally, what we're doing is every day JD wakes up and he has a list to call. Like that's the point of the internship is providing JD with the best possible list to call, uh, both in terms of contact information. So the JD spending his time calling versus researching.

Tyler King (21:37.624)
Yeah, that's great. mean, between Artisan and that, like the last couple years of growth at Leg of Health have been somewhere between acceptable and good. And yeah, nothing like either of these projects have ever happened before. It just seems like this has got to, I mean, what degree of working will it do? I don't know, but like this has got to ramp things up from how they've been.

Rick (22:00.809)
Yeah, that's that's my thought. Like I think we're they're all in the name of the same thing, right? Which is get JD conversations, right? And if

Tyler King (22:10.028)
Yeah. But like in a proven way, like he has done this in a less like assembly line way, like a less efficient way before, and it worked. so it's, you got to imagine there's a high chance that he'll be able to close. And we said basically needs to close one customer per month for the rest of the year to hit the goal. Right. Cool.

Rick (22:29.847)
Correct. Yeah. It's weird talking in these numbers. It seems so doable. The North Star we're using is if we can get JD one new opportunity a day, you ramp up to that.

Tyler King (22:41.262)
Opportunity means like they actually talk to him.

Rick (22:45.493)
like a meeting scheduled.

Tyler King (22:48.834)
And then, okay, yeah, and you can imagine all kinds of like, then the next step is you split Declan in two and one person's researching and one person's doing the initial outreach. And then if they get the meeting scheduled, then JD takes that meeting, which I like a lot. Cause like a different version of this. probably if I didn't know any of the sales stuff you've told me in the past, like what would make sense to me is like JD finds his leads, calls them, books a meeting, has the meeting all the way through. And then when you want to scale.

Rick (23:01.163)
You see where it goes.

Tyler King (23:17.346)
You have a bunch of JDs, which is bad for a lot of reasons. One, they're stepping on each other's toes and competing with each other. But the other one is like JD is a one of a kind person that if you said like the business growing relies on finding 10 more JDs, I'd be like, well, I don't think the business is going to grow necessarily. if it's, yeah, if it's finding one JD being fed by four or five talented, but not as like

Rick (23:35.181)
You gotta break it apart.

Tyler King (23:44.436)
all around generalists, knowledgeable as he is, that seems a lot more dual.

Rick (23:48.494)
Yep. Yep. One thing JD said when I built out a lot of so with the other project that it was in order to onboard the intern, you've got to build out a lot of very detailed, you know, here instruction. And so I built all that out. And JD's like, I should have been documenting this. I'm like, why? You didn't need to document any of this. And I'm like, it's like, oh, nevermind.

Tyler King (24:05.214)
the

Yeah, that's the wrong part of the business journey to start documenting stuff. Yeah.

Rick (24:13.325)
Exactly so. But you know that's those are my leg up updates so I didn't make progress on the coding front but a lot of progress on our North Star goals.

Tyler King (24:22.616)
Cool. Yeah, that's exciting. I'll give my leg up update, which is my task right now is to build a custom quoting portal so that JD can send when he's reaching out to these people or when Artisan is reaching out to people, each customer can have a, or each lead can have a link custom to them. That's like, we estimated this is your size. You know, here's some information and gives them an immediate quote on health insurance and leg up health and all that, or leg up benefits.

So yeah, I have not had a lot of time. think I've got the designs finalized there. I sent you a draft of it, which you more or less thumbed up. After that, Claude, or Anthropic rather, put out a new Claude design tool. I don't know if you saw this. So this was really useful. Me as a designer, I think I am good at UX. I think I'm good at having empathy for the customer, laying things out in a way that will make sense to them and so on.

And I think I can do visual design like 90%. Someone who's doesn't have a lot of design taste will look at my designs and think they're, totally professional and very good. But someone who does have design tastes will look at them and be like, I can tell they're just not quite there. I just gave my 90 % version to Claude and it gave me a 100 % version back. No conversation at all.

Rick (25:42.89)
You have, yeah, that is a terrific use case. So you can basically say, hey, I care mostly about the UX part of the design. Just make this feel like the super polished version, but don't take away any of my UX design.

Tyler King (25:47.819)
Yeah.

Tyler King (25:59.598)
Yep. And it changed nothing. All the copy is the same. The layout's the same. It just looks better. Now, what a real 100 % designer would say to this is they'd say, what it did is it made it look exactly like every other website. Like there's nothing interesting or creative or novel about it, which is true. I don't think a health insurance quoting tool needs to be like that. And neither does a CRM. Like nothing I work on needs artistry in the design. It just has to like convey trust, which is what that the polish does.

Rick (26:17.965)
Correct. Yeah.

Rick (26:27.019)
Yeah, that's awesome. Well, that's super cool.

Tyler King (26:29.801)
Yeah, so I haven't had much time this week, or I haven't had any time this week, so now it's time to turn it into code. I don't expect this to be a big project, because again, AI can do a lot of this for me, I just have to sit down and do it. But I hope to have that in the next couple of weeks, hopefully.

Rick (26:46.189)
That'd super. Thank you for doing that.

Tyler King (26:48.48)
Yeah. All right. Less annoying stuff. don't have a ton. We're actually shipping a bunch of stuff pretty soon, but it's all stuff we've been working on for a while. one thing, the devs are moving faster with AI. It is putting a strain on design and product management. I know, I think I've mentioned this before, but like everybody I thought was like two or three months away from finishing.

their project that like the big, at any given time, some people are working on little things here and there, but like there are a few big projects going on. I thought the big ones were two or three months away and they're actually like one month away. And I'm like, I don't know if I can design all the next projects in a month, you know, but me and the designer. So we're, I don't know. We, don't have any like insights to share here or anything, but like I'm feeling the pressure.

Rick (27:41.099)
Wow, your bottleneck has shifted.

Tyler King (27:42.988)
Yeah, it has. And I've, I'm also starting to kind of tease to the team. Like, I don't know what to do. We are not in a position to hire another designer or whatever, like just financially. so I've been like, what do you do about this? One option is well, developers need to do more different things. Like you're not just before it was like, I am. I'm overseeing the designer, the two of us together are designing it. I'm turning that into notion cards, specking the whole thing out. I'm leaving little notes like.

Make sure to check this part of the code because that's a mistake you might make all the way through and then assigning them to the design, the devs, and in many cases, testing them and making videos, announcing them. One possible solution to this is, okay, some of those things that I was doing now, the devs have to do, and it's less coding and more product management, product marketing, whatever. I don't think any of them want to do that. I think some of them would be good at it, but probably not everybody.

Rick (28:35.799)
Mm-hmm.

Rick (28:42.527)
It seems like an agent, a genetic use case like it seems like you could. It seems like there's some version of an agent that could do parts of this at least. You don't think so?

Tyler King (28:54.926)
People keep telling me that and I keep, every time I dive into specifics, I'm like, what exactly do you imagine it doing? And that's where the conversation stops. Do you feel like you can dive into this with me?

Rick (29:09.109)
Yes, I don't think it'll get you any further than anyone else has gotten you because I think ultimately it's a trial and error process. And I don't know where like, there's probably a lot of little things that you're doing that could get things further along and then even get feedback from a developer, another human. Like it's not like saying developer come in here and do this, but it's having, you know, the agent ask the right questions that other people can answer.

Tyler King (29:12.6)
You

Tyler King (29:37.742)
See the specific part of what I, cause I listed a lot of different things that I do. You're talking about the part of like going from, have a working, like a design we're happy with to we have notion cards. That part of it, is that what you're talking about?

Rick (29:48.728)
That could be one option. do feel like we're close to being able to have, you you have someone do the video for you that is good enough.

Tyler King (29:59.854)
I don't think that's true. I don't know. haven't seen anything like that. And sorry, let me say, we have an employee who does our more polished videos, but these are the CEO sharing on the user forum. A, there's a human element to this that's important, but even if that weren't true, I do not plan on it.

Rick (30:02.156)
Really?

Rick (30:17.419)
I wouldn't stop doing that. But you could generate the script and like spend a whole lot less time prepping that. Yeah. really.

Tyler King (30:23.618)
I'm not prepping a script at all. I'm one-shotting every single one of these. Yeah. Like these are very in...

Rick (30:28.909)
That'll get harder. That'll get harder as you remove yourself from this process.

Tyler King (30:33.141)
Yeah, I don't know that. Okay. I'm repeating things I've said before, but like if AI can do this stuff, we're out of business. There's no reason for like, I have to be preparing for a world where humans are necessary for something. There has to be something that AI can't do. If that's not true. A our customers are all going to go out of business. Cause who needs, if it can do my job, it can do an insurance agent and a travel agent's job too. But, and B why would anyone buy a product from a company?

If just your generic AI bot can do all of it itself. I just don't get that that is the future that we should be betting on.

Rick (31:09.259)
Yeah.

Rick (31:14.413)
That's fair. I just wonder if there, I don't think a generic AI bot can do this. I think you're going to have to be, you're going to have to design like maybe multiple agents at different steps that are very, very narrow that, that like create, like kind of break this large bottleneck of design into like different things. And then I like why you probably developers don't want to come in and do the whole design thing. There are probably elements where they could probably

Tyler King (31:26.509)
Yeah.

Tyler King (31:34.467)
Yeah.

Tyler King (31:42.1)
yeah, yeah, I agree with that.

Rick (31:42.35)
you know, dude, they could probably like an agent could say, this is ready for you. And then and like, there's all going to be human intervention across this entire design process. But it does seem like the you are right now playing the role of every, know, kind of like, like kind of kind of like supervisor at every stage, like it seems like an agent could kind of usher that along.

Tyler King (32:01.518)
So let me give you an example of this happening, but I will say, think your experience is more like you need a custom agent for every single one of these. I'm telling you use cloud code. It's going to, you're not, you're not going to need any of these agents anymore. like a version of this, think. So I just assigned out a card to a dev that doesn't have any designs. And I was just like, there are places in our code that we do things like this. Figure it out and run it by me. again, execution before planning, I'm not pulling myself out of the loop, but I'm saying rather than me taking my time designing it. Ask.

Rick (32:11.479)
Okay.

Tyler King (32:31.322)
Use Claude code and what the dev has to know, and this is one of the skills that I'm trying to get them to like get better at is like, they have to say, Hey Claude, here's the thing that I need done. Here's the notion card. We have other examples of this in the code. Like implement it the way you think we would implement it. Right. So that's just one small version of this, I think. I hear you on some of that. I also think, so one of the, I said I'm a

Right after this podcast, I'm giving the six month presentation and we talked about this in the last episode. And one of the things I'm saying is, okay, sorry, can I totally tangent here and talk about this for a second? Yeah. How do you get, like every company is giving, if they're giving an all hands talk, it's about AI right now. What do you say? And I hope every company is picking different approaches. What I'm saying is, a, what I just said to you, there's no point in preparing for the AI maximalist future. Uh, it is, I'm not saying it's wrong.

Rick (33:08.193)
Yes, tangent, tangent.

Tyler King (33:29.976)
I'm not saying that that future won't come to pass. saying that if it comes to pass, we're out of business, so nothing we do matters. There's just no sense. We should be building guillotines if that's the future.

Rick (33:38.565)
It's the same answer as like, you don't plan for your business failing. Like you just, you mentally prepare for it. Like what you're going to do if that happens, but like, then you don't think about it anymore. It's not worth it.

Tyler King (33:49.164)
Yeah, exactly. So we have to assume the scenario where there is a business and that means humans are still good at some things that AI is not good at forever, basically. And then within that kind of saying, okay, well, we're still selling to people. And so far the customer behavior has not changed at all. Like occasionally someone's like, hey, do you have an MCP server? But mostly they want the same features they wanted before. So the product strategy and really the company strategy is not changing. What's changing is execution. This is not a strategic change. It's an execution change.

And before this and the presentation, I'm doing a whole thing on like, capitalism is competitive. It's about to get more competitive. need to like, I'm kind of doing a rah rah, like we can't bear our heads in the sand. We got to, we got to get ready to fight here. But the fight is who can execute the best, not who can have the best AI strategy. That's my claim.

Rick (34:40.525)
I think that's been true for a long time. I don't think that's gonna change.

Tyler King (34:43.766)
Yeah, yeah. Yeah. Or if it does change, we're out of business. One of those two outcomes. Sorry, I just lost my train of thought. What were we talking about a second ago that I was trying to connect this to?

Rick (34:46.849)
Yeah.

Rick (34:53.678)
You're doing six month presentation rah rah and then you're going away from like the design bottleneck.

Tyler King (34:59.998)
Yeah. Sorry, I'm completely brain farting here. I had a segue to connect these two together. Okay. So anyway, I'm talking about the execution side of things and then, I'm blowing it. I'm sorry. My brain is fried.

Rick (35:15.981)
The engineer, it was about the engineer trying to get them to do more design stuff and maybe why that connects to the AI.

Tyler King (35:23.948)
Yeah. Thank you. You're doing your job here. anyway, I'm drawing a blank. I'm going to stop wasting our time here.

Rick (35:30.925)
That's it. This is a first everyone. Tyler never blinks. It's usually me. It's usually me that is like talks myself in a circle and can't remember what I was talking about. I knew it was gonna come.

Tyler King (35:34.774)
I

Tyler King (35:41.612)
sorry, sorry. I remembered. just got it. Thank you. You kept talking. One of the points I'm making is so I'm trying to say what are other companies doing? We almost this has always been my approach from day one is you have to look at everyone else. You have to study what they're doing and you have to do something else. Even if what they're doing is like quote unquote right from some standpoint, they're going to beat us at their game. We have to play a different game than them. I think what so the way I'm pitching this is what every other company is doing is five times faster. Half is good.

The speed is ramping up, they're shipping more stuff, they're shoving AI and everything. But does anyone feel like software is getting better? No, nobody does. yeah.

Rick (36:18.637)
And the buying software is getting exponentially worse. Like, I don't know if we talked about this last episode, but just quick segue here. Buying enterprise software is getting so bad. The people selling it don't understand the pricing, the packaging. It's like consumption, it's like no one can answer questions. Everyone's removing, like everyone's trying to like get rid of customer success as a function. Like on $50,000 per year contracts, it's like, okay, who is responsible for me at your company? I don't know if I could trust you anymore.

Tyler King (36:32.462)
They don't know what features it has.

Tyler King (36:43.768)
Mm-hmm.

Rick (36:48.161)
Like it's very, very painful.

Tyler King (36:48.812)
Yeah, I fully agree. so the bet is basically that, and sorry, those same people are saying publicly on Twitter and places, they're saying we are deliberately doing like, yes, quality is going down. We're shipping slop, but here's why. Because in six months AI is going to be so much better. It'll be able to come back and fix all this. That SDR that doesn't know the answer to questions. They don't need to know the answer. They're going to get laid off in six months and an AI will know all the answers. All the technical that you're accruing.

it'll get fixed by AI. they're doing, this is a deliberate strategy is to forget quality because the future will bail us out.

Rick (37:23.373)
There is a future solution to that that we can count

Tyler King (37:26.978)
Yes. So basically one of the big kind of bets that I'm placing here is that that's not true. That yeah, AI will, I'm sure we'll continue getting better. not saying that, but like that it won't get to the point where it just magically fixes all your technical debt with the push of a button. And so all of these companies are making a short-term deal with the devil that will not pan out. I don't know if that's true, but I think like playing the odds, positioning ourselves against what everyone else is doing is the probabilistically highest

percentage move for us.

Rick (38:00.717)
I think I could buy that. think that's interesting. So, and then, sorry, are you pitching like two times faster, two times quality? that like kinda same quality, huh? Cool.

Tyler King (38:07.234)
Just two times faster, same quality. So I actually have a slide on this that's saying like, you might be inclined to think that because I'm saying this, we should focus more on quality. don't think there's a business. I think we already probably over, sorry, let me pause. Product quality, yes. Like quality that customers experience, yes. But our customers are not canceling because we have too many bugs. They're not canceling because we have too much tech debt. We were already doing that well. So I'm trying to draw lines here where all of the gains from AI need to go to speed.

But whereas our competitors are adding speed and they're losing quality, we want to maintain quality and have less but still some speed increase. That's the basic formula. So sorry. Thank you for giving me time to get there eventually. To tie this back to what we were talking about, I think you're right that in the flow there are parts of it that AI can do. But I think given the strategy I just said, I think

Rick (38:47.159)
Yeah, that's great. Yeah, I like it.

Tyler King (39:02.83)
Part of what I need to be doing is figuring out what, where, where in this chain is a human necessary. Like so many people I talked to are just like every, what you do is you take every single thing and you make an agent for this and an agent for that and an agent for that. And then you pull humans out entirely. That can't be the strategy. It has to be figure out where the humans belong and keep them there. So that's the challenge for me, I think.

Rick (39:25.217)
Love it. Yeah, I like it. I agree.

Tyler King (39:29.562)
okay, cool. Yeah. Thank you. I'm like I said last time, I'm a little, it's, think it's both optimistic and inspiring as much as I can be inspiring, but like it's, it's 20 % that and it's probably 80 % doom and gloom.

Rick (39:30.189)
Good luck with your presentation.

Rick (39:43.639)
Yeah. Well, yeah, gotta be real.

Tyler King (39:45.262)
What are you gonna do?

Yeah. okay. We got, yeah. 10, 15 minutes here. I want to go back to something we talked about earlier, the, execute first plan later. A lot of our listeners have probably been doing what I'm about to say for months, but I just did my first time where I was, I was designing something. Sometimes when you design something, it's a pretty static, it's just a thing you look at on a screen and you can design it in Figma and get a, you can be pretty confident that it's good. Sometimes there's more of an interaction to it.

Like for example, if you were designing a video game, you can't just make mockups in Figma and tell if the game is going to be fun. I assume I've never made a video game. Even in the SaaS world, some things that you build are more interactive and you just can't tell by looking at the mockups how it's going to feel. So for the first time ever, just had Claude code Vibe code me the whole feature from scratch.

That way I could play around with it. could see what works. could tell where the, needed to be designed in Figma. So anyway, it was a good experience. I hope to do more of that.

Rick (40:53.421)
That's awesome. Are you gonna use the Claude design stuff or are you gonna just go straight to Figma?

Tyler King (40:59.608)
For this probably straight to Figma, I will, I'm sorry, I'm definitely gonna keep playing with it. I don't get the impression it's even meant to be really good for like UX kind of, we need a different word other than UI and UX, but like the workflow side of a SaaS product. I think it's very good at visual design and it's good at like marketing sites and stuff like that. None of the product like demo video for this cloud tool made it seem like it's meant to be.

helping you figure out user interactions and stuff like that. I could be wrong though, I'll try it.

Rick (41:33.737)
Okay. It's got to get there eventually, right? Like it's the same like concepts as a website navigation. It's just, right?

Tyler King (41:41.568)
Yeah, probably. even then it takes, so I use Claude pretty heavily for this as is like not the design tool, but just it's very good at looking at, mock-ups. So I'll be in Figma. I'll make a mock-up. I'll give it to Claude and I'll be like, here's three things I'm not sure about. Do you have anything to point out? it is a really good brainstorm partner, but it never comes up with the solution ever. That's my experience. It's really good at helping me find the solution, but it's, it's not coming up with it from scratch.

Rick (42:10.145)
And you'd rather design in Figma from an experience perspective.

Tyler King (42:14.306)
I think so. And Figma is what I want to give to the developers. Like Figma basically gives them the front end code that they can just copy into the real code base. Then again, they're not writing the code anymore. So maybe that has lost some of its value. I love Figma. People are talking about Figma being, yeah, it's getting hammered, which probably makes sense. like, you know, people like to pile on. They're not just like, it's being threatened by AI. They're like, Figma sucks. Figma does not suck. Figma is awesome.

Rick (42:17.165)
Hmm.

Rick (42:30.435)
It got hit pretty hard, I think, in last couple weeks.

Tyler King (42:44.206)
It's just also maybe in a bad business spot.

Rick (42:47.969)
Yeah, mean HubSpot between HubSpot and Figma like value loss in the last year and a half. It's crazy.

Tyler King (42:56.428)
Yeah, I have a slide because you pointed this out to me. have a slide in this presentation because I think some people at the company probably think I'm like a little too overreacting to AI. And I'm like, maybe I am, but look at all of our publicly traded competitors. And it's just this huge red, you know, everyone going down, down, down, like lots of people think this is happening.

Rick (43:17.773)
Can we talk about this for a second? you think, I mean, do you think Figma, just because of what AI is progressing, is that much less valuable?

Tyler King (43:29.582)
I think some of it's real. like, at the very least, I think a lot of companies, think like companies will do with one person what previously took two. And so even if nothing else, even if every customer of Figma's is still using Figma, they may have fewer designers. There just may be fewer seats. I think at the very least, that type of thing is a threat to them. Will AI completely replace them as the design process? I think the answer is yes for some people and no for others.

I saw a hacker news thread where people said something that sounds right to me, which is like a lot of people, the core designer probably still has a Figma license in the future, but the Rick in this, like there's a slightly different version of you at Windfall that sometimes works with designers and has a Figma seat just so you can do that. And now you would be like, no, I'm just going to go into the cloud design tool and whip it up. Like I don't, I don't need the Figma level designs. Yeah, it's got to hurt them a little. Does it destroy the business? I don't know.

Rick (44:25.805)
I mean, it's like when a stock loses 80 90 % of its value over the course of a year, like, at what point do you like that's pretty bad.

Tyler King (44:36.972)
Yeah, it's probably, probably, yeah. I again would never buy or sell individual stocks. I know better than to think I'm smarter than the market, but if I did, I think my strategy would basically be when other people are selling, I'm buying and vice versa. Like if there's 10 FIGMAs out there and you buy all of them right now, probably one or two of them actually suffers that fate, but probably eight of them.

Rick (44:53.997)
Yeah.

Tyler King (45:06.028)
bounce back and overall it's a winner. What do you think?

Rick (45:08.439)
Yeah. I don't know. I'm trying to form an opinion. just I don't spend time like like you I don't spend time thinking about individual stock investing nor do I generally ignore the macro, you know, on a day to day basis. So it's it's it's kind of once a month or once a quarter that I actually dig in deep on macro to inform me but like

Tyler King (45:29.797)
Yeah. I follow this stuff because I think it's a good way to follow the news. this is the economy drives so much of how the world works that I think like I listen to a few finance podcasts and they're like like stock market type podcasts, but not because I plan on buying and selling just because I think it's a good way to understand what's happening in the world. Um, one, one other, this reminds me of one other slide I have in the presentation, which is, uh, I'm trying to draw, like most of it's doom and gloom, but then trying to draw the like,

Rick (45:34.189)
Mm.

Rick (45:38.838)
Mm-hmm.

Tyler King (45:58.764)
What is our advantage here? Like what's the opportunity for us? And one of them is like SaaS is getting hammered, as you just pointed out. Figma more than most because it's so obvious how AI would replace it, but CRM certainly are. If we had a valuation, it would probably be half today what it was a year ago. We don't care. It doesn't matter to us. All of these companies I'm showing going in the red, it's the stock price that's going down. Their financials are fine. They're making, they're growing.

They're more revenue, more profit than ever. It's just people's expectations of them.

Rick (46:33.601)
Yeah, and that's that's kind of an interesting thing is that on top of all this, there is massive, massive bubble concepts on you know, over inflation of value on based on future expectations that is probably somewhat of the like what what is of this is just a reasonable correction on crazy forecast versus versus like doom and gloom.

Tyler King (46:40.417)
Mm-hmm.

Tyler King (46:43.971)
Yeah.

Tyler King (46:53.196)
Yeah, like how all the layoffs, like, you laying people off because the AI is replacing them? Are you laying them off because you hired too many people in the first place? But my point here is they have to care. It matters to HubSpot that their stock price goes down. Even if all of their baseline metrics keep going up, they pay their employees with stock options, right? All their shareholders are going to be upset and cause pain at the company, even if all their metrics look good, except their stock price.

We don't have to care about that. So like if every SaaS company is in the same situation, bootstrapped ones have this huge advantage of being able to ignore the panic in the market. It's worth something.

Rick (47:32.545)
Yep. Yes, for something.

What else you going on?

Tyler King (47:40.494)
I don't know. Let's see here. Yeah, okay. My next topic is probably bigger than we have time for here. So let me just... I just have a random idle thought that I'm not the only one thinking this. Have you seen all the security stuff with like Claude Mythos and stuff like that? Okay, so this is their newest model. I don't know how much of this is... Some of it is probably marketing hype, but there's certainly some truth behind it. They say it is so good at...

analyzing a code base and finding like security problems that they, aren't releasing it. They're like, we think if we would release this, would just, it would destroy the internet because hackers would immediately own every website out there basically. And then people are like, that's a hell of a claim to make. And you guys are known to lie all the time, but then they, excuse me. Then they like submitted pull requests to a bunch of open source projects.

that or they they exposed a bunch of vulnerabilities like privately to these various projects to say like, look, we found really big security holes in your stuff. This is real. And that happened like, at least to some extent, it's real. So like a that's scary.

Rick (48:58.485)
It's good in a way too.

Tyler King (49:01.334)
Yeah, we'll say more about that.

Rick (49:04.013)
I mean, what you're saying is people have unknown security issues that the human hasn't been able to detect. Like those are still vulnerabilities that like there's two sides of this, I guess I'm trying to say.

Tyler King (49:14.496)
Yeah, but it's kind of like, guess it's good. If they're going to have nukes, it's good for us to have nukes, but maybe it's better for nobody to have nukes, you know?

Rick (49:23.191)
Yeah, but that's not how the world works.

Tyler King (49:25.048)
I know, but like, it's, I'm not saying we can put the genie back in the bottle, but it's scary. This is a big, they're saying this is a huge leap up from what current AI models can do.

Rick (49:34.133)
What I'm wondering though is like, what does that mean? I understand like the downside here, but like what's the upside?

Tyler King (49:40.268)
Yeah. I don't know. I don't know that there's much there. What they're saying right now, they've released this new model to like 10 companies or like Google and Microsoft and stuff. and then they said their plan is to slowly. Like they just released Claude four point or Opus 4.7 and they're like, we took some of the security stuff from mythos and put it in there, but it's just like baby steps. So I think their hope is companies can harden their defenses iteratively over time without having like, all of a sudden.

So yeah, I mean, maybe the outcome of this is we actually end up with more or less like perfect, perfectly secure software. That feels like a pipe dream to me. don't know. Um, but okay. So let's go, let me ramble off a few things since we've only got a few minutes here. A few things. First of all, cal.com, which is an open source, uh, calendar, like it's a calendly alternative. Um, it's like a for-profit company, but they have an open source. They open source their code. They just closed source it and they said,

The reason is when your code's open source now, people can just have AI scan it and find like in the past, the upsides of open sourcing and the downside were symmetric, or even asymmetric in favor of the person who controlled it. Good people will find bugs and submit them to you and you can fix them faster than bad people can find them. But now like bad people will find every bug immediately because they just scan it. It's a push of a button. People are debating whether that's real or is this

Rick (51:06.615)
Yeah.

Tyler King (51:08.984)
They just didn't want to have an open source repository anymore. don't know.

Rick (51:12.919)
They probably have been waiting for like the right reason to do this. They finally got it.

Tyler King (51:15.436)
Yeah, very, very possible for sure. But that it does make me think as a, with code, I think there's an argument not to use open source libraries anymore, or not as a hard and fast rule, but like if you need something, the old move was just, there's an NPM library for this, just download that and use it to, we need drag and drop. There's a library for that. We need this. There's a library for that. AI can vibe code this stuff pretty quickly now and

While the vibe coded code is probably not any better than the open source code, it makes you a target of one, right? If everybody out there is using react, it makes sense for a hacker to find all the websites using react and target them all. If everybody out there has their own little custom version of everything, it makes, uh, it changes the security landscape a bit.

Just a thought I had. I don't know.

Rick (52:06.221)
Interesting. Yeah, interesting. just does I mean, if I'm going to do coding myself, what like how would you advise someone like me who's playing in coding that doesn't understand the security stuff? Would you say make sure you have Claude make thing it's secure before you release it? Or would you say like, just don't put anything in that you are worried about being insecure on the internet anymore?

Tyler King (52:31.584)
Yeah, but I think more or less the second, like you aren't tempted to put any, it's your blog, which everything's public, everything you're working on, whether it's leg up health or ricklinquist.com, it's all public anyway. Like what would the security vulnerability be? I guess the security vulnerability is someone could put malicious code, could make your server serve malicious code to your users. That would be the real, I think the main threat. Anyway, I'd mostly say the second.

Rick (52:55.661)
Yeah, someone could make me do bad things.

Tyler King (52:59.244)
Yeah. Yeah, potentially. I don't think it's a bad idea though. Like whether it's for security or just generally to learn anytime you're about to ship code, have club, look at it say, Hey, do do you see anything that could be improved here? It's not going to catch everything, but, it's worth doing that nonetheless.

Rick (53:18.519)
Cool. That's super interesting. Gosh. Every time I feel like I'm about to catch up, I realize how far behind I am still.

Tyler King (53:25.846)
It's crazy. mean, me too. I'm, I don't think it's worth being on the cutting edge, but being six months behind is worth it. I think. And that's, I feel like where I am.

Rick (53:37.867)
and I'm behind you, so I don't know where I am.

Tyler King (53:40.684)
I think you're ahead of me on like using AI tools. And I think you're behind me on like using the core models and like, like building your own stuff with it. Yeah. Like when you say, get an agent to do this thing, I'm like, I kind of know what that means.

Rick (53:51.819)
the depth.

Rick (53:57.367)
Yeah, like what I'm wider, you're deeper.

Tyler King (53:59.469)
Yeah, I think that's right.

Rick (54:01.783)
That's how we operate, generally. I've like twice your size in terms of width and.

Tyler King (54:03.411)
Yeah, probably.

You are literally wider. You are also deeper than me though. You are taller than me.

Rick (54:15.762)
That's true. Well, appreciate the time. Alright, see ya. Good luck. Kill it.

Tyler King (54:20.962)
Yeah, good talking to you. I gotta go give a presentation. See ya.