Subscribe
Share
Share
Embed
Two founders talk about how to build software businesses that are meant to last. Each episode includes a deep dive into a different topic related to starting, growing, and sustaining a healthy business.
RICK (00:00.866)
What's up this week, Tyler?
Tyler King (00:02.515)
Hello Rick. That is the sound of an ice cold Dr. Pepper opening at 9 a.m.
RICK (00:09.806)
That's awesome. I don't know how to make the sound of hot coffee hitting my lips without slurping, so I won't do it. What's new with you?
Tyler King (00:17.843)
I just got back from Telluride. I go most years. man, conditions are rough out there on the West Coast. Or not West Coast, but Rocky Mountains.
RICK (00:21.943)
I forgot you went.
Yeah. good. You can empathize now.
Tyler King (00:32.027)
Yeah, because yeah, I've been snowboarding twice this year. Once was in Vermont, once was in Colorado. This is the first time ever that Vermont actually had better conditions. But it was a lot of fun nonetheless. I had a great time.
RICK (00:41.41)
Wow. Wow.
RICK (00:46.126)
Did you ski every day? How many days did you get in?
Tyler King (00:49.051)
Yeah, so we were there for like three skiable days, like the day before, like the five days, you count the getting there and leaving part, but you can't really ski those days. Ski days one and three. And this is, I think the first time for this type of trip I've ever not gone every day, which, and it's cause the condition sucked, but honestly it was great. Like I've, prior to having a kid, I never understood vacations or let me rephrase. I understood travel. I understood like,
RICK (01:01.773)
Nice.
Tyler King (01:19.143)
go to a city and eat the great food or see the history or what, you know, do an activity. I never understood the get away from it all and relax. Like that never made sense.
RICK (01:31.96)
Well, you and Shelley are doing better than me and Zabel did. think Oliver was two years old before we went on a vacation away from him.
Tyler King (01:38.62)
Yeah, well, we're lucky to have a lot of support. so my parents were supposed to take care of Sydney while we were gone. And then my mom had like kind of emergency surgery. She's fine, but can't lift a thing the size of a baby. So we called in Shelly's parents for kind of emergency. They had to fly in from Vegas last minute. So very lucky that we had that as an option. Yeah. And we're going to.
RICK (02:02.776)
That's so awesome. That's awesome.
Tyler King (02:05.981)
France for 10 days later this year and leaving Sydney behind. I'll be honest, I don't understand traveling with a baby. I have no interest in it. I'm leaving her behind if I'm going somewhere.
RICK (02:14.882)
Wow, I mean, that is, I wish I could say that out loud. If I said that out loud, I would be in trouble. But yeah, it's Exactly.
Tyler King (02:17.747)
Yeah, like the baby's not getting anything out of it. My parents took me to Greece when I was in third grade, I think. That was my first big trip. And I barely, barely remember it. Like, I feel like third grade is about the earliest time, like you're actually forming memories about a trip like that.
RICK (02:42.978)
Well, yeah, man, I would. Well, I would say like once I think your perspective will change once you get to the talking and like sort of witnessing of experience stage like where they go, this is so cool, especially maybe not like international trips, because I think that's a little bit like I don't think that means anything to.
Tyler King (02:46.811)
Okay. I'm not, I don't want to incriminate you anymore.
Tyler King (02:58.194)
Yeah.
Tyler King (03:06.131)
Sure.
RICK (03:12.6)
to toddler. Yeah, but a beach, big water, a big pool, slides, that kind of stuff. They love.
Tyler King (03:12.839)
Yeah, like cultural experiences probably aren't for them.
Tyler King (03:22.055)
Yeah, okay, fair enough. I've also never understood a beach vacation, but I guess I'm about to.
RICK (03:26.594)
You will, when you have kids and you're like, they play on the sand for an hour and a half. It's like, okay. He just lounge. Not much. The main sort of thing happening over the last, the next two weeks is this week JD's on vacation. So I'm covering the leg up health cues and it's crickets. So there's like zero general support needs. I've had to reply to one chat message.
Tyler King (03:31.027)
That does sound okay. Yeah. Cool. What's going on with you?
RICK (03:55.98)
And it was like, and it was answered by a blog post. What? Well, it's, it's, it's good from one perspective, but from a from a growth perspective, it's not good.
Tyler King (03:58.45)
Is that a good thing? Like we hope that it's dead around now or is it like, I, yeah.
Tyler King (04:08.785)
Yeah, you hope for new leads to be bothering you, not bothering you, but you know, to reaching out, not current clients. Do you feel like you have the expertise to handle it really well still, or do you feel a little rusty?
RICK (04:11.554)
Yeah. Yeah. Zero zero new leads opened.
RICK (04:25.512)
I'm definitely rusty. But I would say on the consumer side, I did it enough where it's like, yeah, I got it. On the employer side, I would say that there's not even rust. I've never done it. So so I would say that JD is the expert there. I'm on the leg up benefit stuff. I've never supported it. But I mean, I think I think it would I would approach it very differently probably than JD does.
Tyler King (04:39.219)
Mm, right.
RICK (04:53.526)
in terms of like how I would solve the problems that a user would be facing. But I think it would be fine.
Tyler King (05:01.213)
That's kind of the situation I find myself in when I do support for less annoying is I remember how to do the stuff from 10 years ago just fine. And I can improvise as well as anyone I think. like, you've got a problem, let's try and figure out a way to solve it. But the eight person CRM coaching team has all this process and well, when this comes in, here's how we handle it. And you tag it like this in front and I...
I don't know all that stuff, so I do feel like I'm a bit of a liability from the process side of things.
RICK (05:33.35)
man. Yeah, like I.
Tyler King (05:34.523)
I guess JD's a team of one so he doesn't have all that stuff going on.
RICK (05:37.792)
Yeah, that's that's scary to me though, like, maybe it's gonna be hard. Yeah, so I it makes sense.
Tyler King (05:44.18)
I forgot if I said, like, we officially stopped having developers do support. I forget if I said that or not. Yeah, we're replacing it. I still think there's a lot of value in them, like, understanding who the customer is. And so we're trying out various things right now, like recording, demos. What would happen is the dev would do support, email. They could only do email support. Phone support's much harder because you have to be able to answer everything immediately. You don't get to...
RICK (05:50.19)
Ooh, that's a big change.
Tyler King (06:12.487)
be like, I don't know that one. Give me 10 minutes to research it. So it's always been that devs only do email support. But CRM coaches are so good at email support, either the devs would, the inbox would just be empty and there's nothing for them to answer, or they can only answer the easy stuff and they're not really getting exposed to interesting customer interactions. So we're trying to instead be like, let's record, all of our demos are recorded anyway.
let's pick out the most interesting ones and do like a workshop on it or Michael is dabbling with different approaches to this. we're trying to get the devs that exposure, but yeah, they just can't contribute to support anymore. It's too complicated.
RICK (06:52.814)
That's awesome. mean, I'm sure the devs are like, thank goodness. Like, yeah.
Tyler King (06:59.795)
I'm sure no one liked that. And there's still two of us, Robert and I, Robert being the team manager, the dev manager, we both still do it, but no one else does. Yeah.
RICK (07:10.095)
if you don't feel like anything's being lost, like probably better use of time.
Tyler King (07:15.505)
I mean, the CRM coaches were asking for us to, they were like, every time a developer touches something, like that's just creating a liability that we have to go fix later. so it definitely wasn't contributing anything in terms of support, like taking some of the support load off. It was just about getting the devs exposure to customers.
RICK (07:23.659)
RICK (07:34.22)
Yep. Yeah, we there are other ways to accomplish that.
Tyler King (07:39.101)
Yeah.
RICK (07:40.646)
The other update I had real quick was just like personal update is like, so this week it's covered is JD's out next week Sable is going to Amsterdam for seven six days. So for work. Yeah, so I'll be solo dad next week. Okay.
Tyler King (07:50.387)
Wow. Solo dad.
Nice, I'm actually doing that right now because Shelly stayed in Colorado for a few extra days to see some friends. Well, you've got a little more, you've got more people to take care of than I do. Yeah, yeah.
RICK (08:01.832)
good for you.
Ha
Well, they're easier. They get easier when they're older. but yeah, it's it'll be fun. Let me just put it that way.
Tyler King (08:14.937)
It's great having help. We have a part-time nanny who's upstairs taking care of Sydney right now. I don't know how people do it without that. I guess daycare, but...
RICK (08:18.478)
You
RICK (08:24.462)
I get daycare or family help or a lot of people. I was talking to my mother-in-law the other day and she was telling me that back in the day, I mean, it was kind of neighborhood. You just with your neighbors, had people run it. You called the neighbors that, can you watch Bobby, you know, for for an hour while I go to the store.
Tyler King (08:40.06)
Yeah.
Tyler King (08:45.939)
Mm-hmm.
Tyler King (08:49.491)
Yeah, and we've got a little bit of, we have two of our good friends live within a block of us, but they have their own kids. it's like, try to help each other out when we can, but everyone's busy. Did I tell you, by the way, sorry to keep the chit chat here is going on a little long, but did I tell you my parents are moving in right next door to me? Yeah. It's like the paper. Yeah, yeah. Like the closest building to our building, which
RICK (08:57.899)
Yeah.
RICK (09:11.628)
You did not tell me that. Like next door?
Tyler King (09:19.559)
You've been to our neighborhood like it's a pretty dense neighborhood like we can look in their windows like that type of thing. Yeah.
RICK (09:23.35)
That's awesome. Was it that one that you were pointing out that time I was there where you're like, that's been empty for a while or something like that? Is it a commercial building?
Tyler King (09:31.471)
It hasn't been empty. Actually, they did the thing where you flyer the neighborhood and they printed out a little flyer that's like, hey, we want to move to the neighborhood and just coincidentally, the very closest house to us. a three, it used to be a pharmacy. It's a weird building that got converted into three townhouses, I guess you'd call them. But anyway, yeah, it's going to be great.
RICK (09:53.848)
Cool. That's awesome. Yeah. Similar thing is potentially happening on our front. So yeah, not next door, but in the closer vicinity, which would, which would be really nice.
Tyler King (09:59.788)
really?
in the neighborhood. Yeah. All right. How's work? Yeah. Chat complete. Yeah. Do you ever do this? Well, you always enter your podcast topics right before we record. I enter them over the course of two weeks.
RICK (10:07.406)
All right, that's enough. I want to know about Notion AI.
Tyler King (10:22.919)
before we record, just as something comes up, I enter it. And I entered this topic of, I'm trying out Notion AI and it's freaking awesome and I love it. I entered that topic and now though, like a week later, I'm like, no, I canceled the trial and it sucked.
RICK (10:36.398)
Okay, so I had a different experience. I'm sorry. so I would let me break down notion for a second, the stuff that we're experimenting, there's like, they were they've had this AI agent for a while. But what they released recently was the ability to build specific agents that can do specific things and or do different things. They also separately have integration or their own mailbox that connect connects to Google, or, you know, Microsoft or whatever.
Tyler King (10:53.171)
Mm-hmm.
RICK (11:06.7)
And then they have a calendar. Are you using all three of those? Okay. All right. So it didn't work.
Tyler King (11:10.023)
I tried to, yeah. It didn't work. So right off the bat, I think this is maybe like the best example of how AI can lead to low quality software. Like there's a question of is what they built useful and well-designed and stuff like that. even separate from that, every single person in my company got a pop-up on every single page load that said, welcome to Notion AI. And there was, we couldn't get it to go away.
So I'm talking about, I'm closing this 50 times a day. So right off the bat, I'm like, I'm canceling for that reason. And as soon as I canceled that one away, there was just a ton of new bugs in Notion that weren't there previously until I enabled Notion AI. And I assume that's just because they're shipping way too fast. They're feeling the same FOMO, the same pressure everyone else is, and their quality has like seriously dropped in my opinion. I guess you didn't experience that.
RICK (11:56.429)
Interesting.
RICK (12:07.266)
Hmm. Well, no, I haven't. I've had a pretty good experience, but my, my use case is pretty narrow. the, the AI, I have not played much with the mail or the calendaring stuff. the AI piece, the agent's pretty useful, in terms of.
Tyler King (12:20.733)
Mm-hmm.
Tyler King (12:24.893)
Well, you would have noticed if you hit this specific one I'm talking about, because literally every single page load, it pops up a modal dialogue that you have to close.
RICK (12:28.387)
Yeah.
RICK (12:32.59)
Now, but I don't, that wouldn't, I don't use notion outside of, you a daily user of notion? Like an hourly user? Okay. Okay. So yeah, that's not my work. I, I'm working my primary workflow and notion is long-term planning. Uh, and then on the other end of the spectrum content production. And I would say for content production, the AI use cases have been fantastic. Uh, the
Tyler King (12:39.476)
yeah, I've got 10 tabs open at any given time.
RICK (12:59.438)
I have not played with I'm more curious actually about the email and calendar stuff because I was thinking maybe we could replace front with that stuff but it sounds like that's a bad idea.
Tyler King (13:08.529)
I don't think so I didn't get deep into testing what I the only reason I hooked up my calendar and email to notion was so that I could create agents to interact with it. And I do like that idea like like here's a good example. My biggest complaint about Gmail right now Google is shoving AI into everything. Seemingly indiscriminately they have no sense of what's actually good use case for AI and most of it doesn't work.
I want one specific thing. There's only one way in which I want Google AI integrated with my Google tools, and that is I want to be able to describe my own spam filter. I want to be able to say, an email comes from someone who identifies themselves as a venture capitalist or private equity, archive it. I just don't want to see that crap. I get 10 of these a day. Get rid of it. I can't do that in Gmail. Notion does have that feature.
Because you can, you just like describe in natural language what, like what filters you want. Did it work? No, it did not work, but they have the feature. Like I, yeah.
RICK (14:13.322)
Okay. Do you... Am I... Is that the auto label thing that you're talking about? Okay.
Tyler King (14:18.769)
Yeah. So I was like, I can just label anything from a VC. I didn't even start that. I didn't even get that far. was like, I have a filter that actually works in Gmail, which is it's from this email address and it has this in the subject. Label it. I just did that as a test. It failed miserably. It missed half of them and it incorrectly identified some false positives.
RICK (14:41.518)
So would you say that Notion Mail is like beta, a beta version of Mail or?
Tyler King (14:47.801)
Sorry, the mail might be fine. I'm talking specifically about the AI features kind of attached to mail, but I don't get the impression that mail is meant to be this like collaborative support inbox that front is. It might be, but I did not see anything like.
RICK (14:52.428)
Okay. Okay. Got it. Okay. So
RICK (14:59.596)
Yeah.
Yeah, I don't see that either. Man.
Tyler King (15:05.501)
So I think the use is, and I think this is their vision. actually like their, all of the AI stuff, I love the vision. I think they've got, before they were doing the normal, like you said, just there's a little chat bubble in the corner, pretty worthless in my opinion and pretty lazy. I think they've actually got good product strategy for how to incorporate AI now. It just is not implemented well. But the reason I think they want mail and calendar in there isn't that this is gonna be a front competitor, it's that they want to be the system of record.
Right? If you're going to run an automation, right? Like right now, the options are you, you connect everything together with a million MCP servers and this and that they want. Everything's already in notion and you that's, that's where you run the automations and the AI stuff, which I think is a good vision. just doesn't work.
RICK (15:55.116)
Yeah, but so you would say that like, I guess where I'm trying to go is you let's say that they succeed at the AI side of mail. If the mail client itself is not, you know, as good as other experiences, like would you still, is it okay to have attached and then still using another client to actually send and receive mail? Okay.
Tyler King (16:03.763)
Mm-hmm.
Tyler King (16:15.131)
In my opinion, yes. cause the impression I get is it syncs everything with Gmail. So I can do the filtering in Notion and then I can use Gmail as my actual inbox. I'm not, I'm not saying that's what they want. They probably do want me to use Notion Mail as my main, you know, mail client, but I'm just, saying it's useful either way. just as a way, like, I can finally.
RICK (16:37.454)
Mm-hmm.
Tyler King (16:40.925)
connect email and calendar and I mean all my project management, all my notes, everything's in Notion. There's all kinds of stuff you can do having all that in one place.
RICK (16:49.902)
Yeah. I mean, if you have, this isn't the use case, they're probably thinking of it. Like I was just thinking my, my, my workflow with AI and notion right now is like, I'll have a start a Slack conversation with a, with a thing about a post idea. It starts drafting the post. I'll then go to notion and, basically work with the AI to copy edit a few times and finalize it. And then I use Claude to like, push it to web flow, publish it and like tell everyone that it's up. But what would be nice is if someone would
If there was a group of important people on an email list, if Notion could just email them and tell them, hey, this post is live. It's like a better blog subscription tool.
Tyler King (17:31.08)
Yeah, exactly. mean, eventually maybe it turns into like you're building no code tools just directly in Notion. So let me just talk through a few more like use cases I attempted here. Again, nothing really worked. So let's just start with the basic features they have. You mentioned that you can make custom agents, which should be cool. I started testing that out. Custom agent just meaning...
RICK (17:37.879)
wow, that's crazy. Okay.
Tyler King (17:57.992)
What's really neat about it is you can make triggers. You can say when this thing happens, basically run this AI prompt. So I ran a test with it first off. was like, okay, I'm connected to email. What if like I get a bunch of transactional emails that I send myself, like the Lesson Knowing Serum internal software sends me emails. Anytime someone subscribes, pays, cancels, adds a sub user, removes a sub user, et cetera. I archive these all automatically. I don't see them because there's too much, but I like to have the record.
in my email inbox. I asked Notion, just look for these specific emails, summarize them once per day and send me an email with that summary. It correctly ran the automation. The summary was terrible. I think they're using a much worse model than Opus 4.6 or whatever. The ones we're used to using, I'm just guessing here, but
If I'd given all that content to Cloud Code, it would have nailed it. Notion just, the email it sent me was worthless and incorrect. I had other agent use cases in mind as well that I don't think worked well, but again, I canceled it for other reasons. One was anytime a bug gets reported, because we do all our bug report, all our project management in Notion, anytime a CRM coach submits a bug report, I connected it to GitHub.
And I said, basically draft what you think the bug is and what the fix is. And then when the developer has to go in and triage the bug, rather than starting from scratch, they've got this like starting point from Notion. worked okay. It did run automatically. It did look at the code. again, significantly lower quality than what I'd get from Claude code. So, I'm just going to do, build the exact same thing with Claude code instead of with Notion.
So anyway, that type of thing is what agents can be used for though.
RICK (19:59.854)
But you're basically saying that Claude does anything related to code better than anything related to product development better than Notion. Anything. Okay.
Tyler King (20:06.973)
I think just anything. I just think the model is better probably. Or the agent is better. I don't know where in the stack Notion is falling short. Notion's only charging an extra five bucks a month for this and I'm paying a hundred bucks a month for my user of Cloud Code. maybe it's not a surprise, but yeah, I'm used to good AI and this is bad AI.
RICK (20:30.36)
Well, what is your gut? Do think it's gonna get better?
Tyler King (20:35.207)
I think they've, yeah, I think so. Again, I think they've got a good vision. They've got a more coherent vision for what their company looks like in the age of AI than any other company I'm aware of. I don't know if you've seen any other examples of this.
RICK (20:50.382)
I mean, the only other Atlassian has a pretty solid one with their bought their agents called robo. But similar similar approach where, you know, they've got their confluence and their gear tickets, and they're trying to integrate that functionality more and more. But but yeah, there's a there's an agent on top of it that's pretty powerful. And I would say it's good. It's better. In a lot of ways. It's better than than what notion has.
Tyler King (21:00.861)
Mm-hmm.
Tyler King (21:09.267)
It's good.
Yeah. So yeah, I hope it is good, but, then the other side, you know, I'm rooting against AI as much as we talk about it. And as much as I feel like it's inevitable and I have to get on board with it, I don't want it to work. It is somewhat comforting seeing this company that has many resources and is obviously making a major commitment to AI, just butchering it so badly. Not that my takeaway from this is not, AI is a fad. It's, it's fake. It's like crypto. It's, it's never going to amount to anything, but it is like,
we have some time. I think this, this has made me a little less stressed out about everything because of notion botched at this badly. It'll probably be a few years before everyone's talking about, you don't have years, you have weeks. You know, I don't think that's true based on this experience.
RICK (21:44.952)
Mm.
RICK (22:00.098)
Well, I'm on the other end of the spectrum. I am rooting for this because it's major advantage for leg up, I believe, and us.
Tyler King (22:08.795)
Assuming it doesn't make health insurance irrelevant, health insurance agents irrelevant.
RICK (22:13.346)
Well, I don't mind that. Then the employer will need software to explain how the journey is for their employees.
Tyler King (22:20.497)
No, they won't. If AI can do the job of a developer and it can do the job of a customer service person, it can do the job of an insurance agent. Sorry, can I rant about this for a second? I keep hearing opinions from all kinds of different people that are like, I'm going to run an AI agent and it's going to make me $100,000 a month or whatever. And I feel like this just fundamentally misunderstands how capitalism works.
RICK (22:24.944)
you're saying bring chat GBT. You're saying those. Yeah, I hear you saying.
Yes, you can.
Tyler King (22:48.175)
You're not going to own that. Like leg up health. If, if an AI agent can act as a health insurance agent, leg up health goes out of business. We are not going to own that agent. Period. No, like the bigger companies are not going to let us extract all that value using a technology that we didn't even build. There's just no way using a third party AI agent, you're going to just profit off of the agent autonomously running a business. I'm not saying you're saying that, but I've seen so many people.
just misunderstand that capitalism is a zero-sum game and deeply competitive.
RICK (23:22.934)
I'm thinking about like an example like Google search, right? So I mean, your point would be, yeah, now you can't get organic rankings and Google chart, like basically paid search dominates the results.
Tyler King (23:35.525)
No, I'm saying everyone just has their chat GPT account and they're like, what should I use for insurance and chat GPT answers them? They're not even going to need Google.
RICK (23:43.982)
Exactly. mean, I'm trying to think of a corollary where it's like, where do you like Google built a great search tool as a bit like people keep grew because we ranked really well in the search algorithms organically. Eventually Google started to monetize those search terms with paid search and and long tail paid ads and paid ads dominated to your point. Like eventually the capitalism wins over us writing the technology that we don't know.
Tyler King (24:05.235)
Mm-hmm.
RICK (24:13.934)
There are lots of examples of this, is all I'm trying to say. Yeah, times 1000.
Tyler King (24:14.097)
Yeah. But this is, this is that times a thousand because they are actually so Google is still linking out to other sites. Like someone other than Google is still the one completing the sale. I don't think that's going to be true for service businesses, which leg of health is if AI gets too good, it's not good enough yet. But if it gets better, the fantasy that it's going to get good enough that we don't have to employ people anymore. That.
fundamentally puts us out of business. doesn't mean, great, we get to let, not that we have employees, but bigger companies are talking this way. We're gonna lay off all our employees and just get all the profit. No, you're gonna stop existing as a company is what's gonna happen.
RICK (24:55.374)
Isn't that what monopoly regulations are for?
Tyler King (24:59.881)
yeah, those are getting enforced all over the place right now. Okay, sorry, rant over. Go ahead.
RICK (25:06.35)
Nope, nope, I'm good. Any other AI use cases in Notion that you tried and struggled with?
Tyler King (25:12.979)
I took some notes here. Another one I tried was we get feature requests. This is similar to the bug report thing, but we get feature requests a lot. Our serum coaches bring it. They submit it as a topic for our weekly company meeting. We do try to do a pass through where it's like, have we had this topic before and have we discussed it? And if so, what did we decide? I tried to make an agent to automatically find that to just say like, what are the most related topics in the database?
didn't do well at that either. tried to, so one of the things that can, back to them like being the system of record, they let you connect to GitHub and Slack and probably other things, but those are the ones we use, and just ask questions to the Notion AI, and then it'll not just search Notion, but it'll search your Slack and your GitHub, which sounds cool. I think a lot of companies have the same problem we do, which is you talk about stuff in Slack, so that's some of your,
knowledge base is just informally unstructured in Slack. And then some of it's in like a Wiki or something like that in Notion. Now you can just ask the agent a question and it will automatically search both places. Again, completely botched it. That one was terrible. I connected to GitHub in three different places. don't understand, like again, they're just tacking on shit carelessly. There's one, the agent connects to GitHub and there's another one where like my personal account connects to GitHub. And there's another one where like the search
Anyway, I connected in three different places and then I went to the agent and asked a question. It was like, sorry, I'm not connected to GitHub. I'm like, I should send it screenshots. I'm like, here's screenshots showing you're connected. What am I getting wrong here? And it's like, I'm just not connected. I don't know what to tell you. So again, it just didn't work.
RICK (26:49.08)
Ha ha ha, jeez.
RICK (26:58.657)
Aw man.
Tyler King (27:00.871)
So maybe I'll probably try it again for an extra five bucks a month, for an extra 20 bucks per user per month. If it worked, it would be worth it, but it doesn't work. Yeah.
RICK (27:10.754)
Bummer. Yeah, I was gonna share a couple learnings that I've had like, with the content thing related to this before we move on to topic. The the main thing I'm happy so I'm this is working. So counter example to Tyler's experience is that for the content production line, you know, if you if you've ever run a high velocity blog or online publication, you probably know what I'm talking about, Tyler, you've done a lot of blogging. I know a lot of people have but
it's either the process, right? You come up with the ideas, then you, you know, you do the brief, you, you then write the first draft, you go through a second round of editing, and then you are into the final editing piece, and then you got to publish it, right? And so there's, it's very clearly a Kanban experience from, you know, idea to this. And if you run a team, usually there's cadences around like, what's our plan? You know, are that what are our pillar content pieces? How is that spidering? How do these interconnect? What's our release schedule, all this kind of stuff?
I didn't start there. just started with, you know, the content production line and it works great. But what I'm realizing is I don't have, I'm not getting any help with the longer term stuff. So I'm still the bottleneck every day of coming up with the idea and stepping it through the process. so I, this isn't revolutionary, but the thing that I'm going to shift to is treating the notion AI people things, like a team. So
I want to basically have a meeting, you know, once a week where we are doing planning and planning out, at least a week's worth of content so that, I'm, I'm more managing the out towards the outcome that I want versus like kicking out the process and monitoring the process. and I'm, I'm curious to see if that doesn't unlock some additional sort of like use cases across multiple agents where maybe there's a, an agent that's very focused on, you know, the
ideation and the coordination of like when and what to write about. And then another one is, you know, taking it through the steps. And another one is a publishing and interlinking it. But anyway, I haven't gotten there yet. But I'm starting to see the vision of like, having multiple AI agents be a team working together and managing them like a team. I haven't done it yet. It's hypothetical, but I'm going to start doing it.
Tyler King (29:31.028)
I don't know, it's probably not worth it to you to get to this level of depth, but I think someone who's really into this stuff would say Notion is not the right tool for this. For two reasons. One is, like I just said, I think they're using cheapo models and you'd probably get better results with something else. But the other one is, they don't run in the background. They can get triggered by things, but compare it to something like OpenClaw where it's just like...
Do you follow Brian Castle of Builder Methods? He's been talking a lot about how he's using OpenClaw. Okay. You might want to check it out. The mental model... I'll admit, I still don't have my... Like every time I see him talk about this, I'm like... It's not that I think he's wrong. It's that I don't... It's just not clicking for me what I would do with this. But he has like multiple OpenClaw agents.
RICK (30:04.244)
I know who he is. have not followed him in a couple of years.
Tyler King (30:24.131)
And even this is probably more for fun than because it's actually necessary, but like he gave them different names and different personalities and is treating them like actual employees. And so he'd say like this agent's job, it's exactly what you just said. This agent's job is to like do the big planning stuff. And then this agent's actually writing things and then this agent's doing this and that, which isn't that different from what you said in Notion. But the difference is it's like, you still have to go to Notion and tell it, okay, it's time for our monthly planning session. Let me do this, let me do that.
versus it's just always running in the background and you say, hey, on the first of every month, I want you to email me what the plan is and I'll either approve it or give you feedback. Like it's kind of working in a more passive way than what you're describing.
RICK (31:06.648)
think what you just said is possible though in notion today. You can say you can trigger it on time a time bound stuff. So but it's still very prescriptive versus like, generic, like, I'd much rather be able to say you're smart, you can figure this out on your own learn, you know, versus let me Yeah. Yeah, yeah, I see your point. Your Yeah, where I would I haven't gotten to the limits of notion with content yet, but I can imagine you're right.
Tyler King (31:10.651)
Yeah.
Tyler King (31:21.735)
Right go do the research what should a business like this be writing about yeah.
RICK (31:35.411)
that notion is not the right tool for that. Have you played with any of open call stuff?
Tyler King (31:39.345)
No, I have, I still don't, I have not been able to think of a thing that it would be useful for for me.
RICK (31:46.752)
you can give it a credit card.
Tyler King (31:48.658)
Yeah, no, that's crazy. not, I have too much pride to be the person that gets his life destroyed by giving AI too much access. I'm not doing that.
RICK (31:51.182)
and a voice.
RICK (31:58.722)
There's a, I saw a new, a new sort of product that came out, which is basically it's a wallet for open call that you can constrain both like the budget for and what it could be spent on. and you can also, it also like you control the voice. And so, I mean, I don't know, like I've heard stories of where if you let the open call free, they'll, they'll find a way to get access to finance financials. and they will.
like figure out a way to talk to you. can, they'll like get their own phone number. They'll call you, they'll, they'll create a voice and they're, it's crazy.
Tyler King (32:31.389)
Yeah.
Tyler King (32:35.219)
One of the guys at Big Snow TinyConf, that was the week that at the time, ClaudeBot, it was called, took off. And he did that at Big Snow TinyConf. He was texting the bot and then he was like, I just want to talk to you. Can you make that happen? And yeah, it did exactly what you It downloaded all the software and built a little voice to text thing. It's scary. It's so cool. But we've talked before with humans, with employees at businesses, how hard
access is. Someone wants to do something, but I think the example you gave is they need access to the Salesforce API and you're like, no. I think one of the things going on with AI, yes, it's magic and it can do all this stuff, but a huge part of the unlock is that we just all have this weird brain virus that we've just stopped caring about privacy or security. If we did that with human employees, it would unlock a lot of shit too, but there's a reason we have that stuff in place.
RICK (33:10.542)
No.
RICK (33:33.102)
Yeah, yeah. Do you anticipate like there's a more sophisticated AI that has access control capabilities and then a less sophisticated like it's like a new a new AI, you know is governed by this like more sophisticated AI or is this just is it all one thing at the end of the day?
Tyler King (33:40.743)
Yeah, I don't know.
Tyler King (33:51.091)
I mean, I don't really know why you wouldn't just make everything the sophisticated AI aside from cost, but...
RICK (33:54.862)
Exactly. Yeah, once you have a sophisticated AI, why would you have a junior AI?
Tyler King (34:00.593)
Yeah, it's tempting to anthropomorphize all of this and actually every AI is a person and some people are juniors and some people are seniors and some people have this role and some people have that role but that's not quite the right metaphor, I don't think. I don't know what is, I don't know how this all ends but it's not that.
RICK (34:12.247)
No, it's not.
Today it is a good use, it's a useful metaphor because use cases across AI vary differently in terms of whether you can rely on the AI for it. And so like, you can say like for this use case, it's a junior use case for this use case, like, no, I trust it for this. And so it is useful to like think of it that way. But yeah, at some point it's not, it's irrelevant.
Tyler King (34:38.087)
Yeah. It's like a weird super genius that's, it's like so much better than us in some ways and so much worse than us in other ways that it's just different. But, okay, final thought before we move on from that specific topic. You said you're using Notion and it's working. And I said, well, here's this other thing that would probably be even better. I want to push back on myself here, which is to say it's exhausting trying to keep up and it's probably not even useful because six months from now,
RICK (34:46.932)
Mm-hmm. It's different. Yeah
Tyler King (35:07.717)
The better thing I'm describing will be irrelevant. So even though I'm saying this just out of an intellectual curiosity to talk about it, I don't actually think it's good use of most people's time to be doing the cutting edge. Like here's the best of the best version of it.
RICK (35:23.054)
Yeah, I think as long as you're getting better at, you know, making continuous improvement cycles constantly, like that's all you can ask for. And if there's a step lead that you discover through that process, whatever process you're using, great, take it. like,
Tyler King (35:35.588)
Even that though, I question how important it is. I mentioned in the last episode or one or two ago that code review is a big thing on less annoying serums mind. you and I were emailing with a listener and he was just like, by the way, Cloud Code just put out a code review tool. Now we haven't tested it out yet. Will it solve our problems? Maybe not. But eventually we don't have to spend a month putting together some
code review pipeline that uses all these different models and all this. Eventually someone's just gonna be like, us 50 bucks a month and we'll do it for you.
RICK (36:12.118)
The perfect example of this is the AISDR. Last year, I spent lots of time trying to cobble together tools with different people at the company to do an AISDR. Today, there are multiple AISDR companies that literally just say, hey, come in here and just use the interface to control the AISDR. We'll take care of the rest. And it's a subscription. It's great. They do everything. Yeah.
Tyler King (36:32.517)
Mm-hmm. Yeah, and they'll keep getting better and yeah. So I'm kind of both in the, if you don't adopt this, you will die, but at the same time, wait a little bit and buy an off the shelf solution. I kind of feel both of those at the same time.
RICK (36:49.768)
Do you not like, I mean, you mentioned that you haven't played much with open call, but you mentioned that the Brian, Brian Castle has like, do you feel like you aren't tinkering like you should be just for fun? Okay.
Tyler King (37:01.039)
man, yeah, good point. I think I would be feeling this even without AI. I feel like such a poser right now. Because the dynamic I have on the dev team at Lesson 9 CRM is I'm following this more than anyone. I'm Twitter, podcasts. I just think I read and I consume more content about what's going on in tech than anyone else at the company. And all anyone talks about is AI, which means naturally I'm getting a ton of AI content.
But I'm not coding really. I mean, I've tinkered with it, I've played, but I'm not really the one in the trenches building stuff. So I'm just this empty suit sitting in all these meetings being like, well, have you tried this? Have you tried this? Have you tried that? I hear this works, I hear that works, but I haven't actually done any of it. And it's definitely making me feel like the type of manager that I used to think didn't deserve to have a job. How are you feeling about it?
RICK (37:51.564)
Yeah. Yup. I don't, I, much more hands on keyboard on this stuff. I would say, you know, versus consume I'm consuming two fine tools to play with. Like that is my primary like reason for consuming AI stuff. or to unlock more understanding of how to use them. so I would say it's, I'm much more tinkering than you just based on my job, function, but, but I, but like this open call stuff, like I don't, I'm talking about stuff that's like,
There's kind of like stuff that's useful and then there's stuff that is new. I'm not just playing with new stuff. I just remember back in the day, you and I would tinker with new stuff all the time. Whether it was a new like Palm pre or, you know, email to a client or whatever it was at the time. And I feel like we're not doing that in our forties.
Tyler King (38:25.296)
yeah.
Tyler King (38:30.301)
Yeah.
Tyler King (38:39.527)
Yeah, so is it an age thing? Is it that the tech industry has beaten us down? Like, I used to get excited every time there was a new thing. I was like, this is probably gonna make the world better. And now every time there's a new thing, I'm like, I'm certain this is gonna make the world worse. I'm not sure why. I totally agree with you. Is it age though? I don't know.
RICK (38:57.336)
No, I know. Like I pulled up Brian Castle while you're talking and I found his multi-agent team with open claw YouTube video and he basically bought like the meta description is I ordered a Mac mini just to run open claw and set up a team of four agents. Like I want to order a Mac mini just to run open claw. Like I want to do that. That sounds fun. I don't have time to do that.
Tyler King (39:18.151)
My problem is I keep trying to tinker. The thing about AI, the trap you fall into, you get 80 % of the way there so easily, and then you actually have to work for the last 20. And so I'm like sitting on my couch. I've done this like a bunch of times. I'm sitting on my couch watching TV at night. Normally I would not be working or anything, but I've got my laptop out in front of me. I give Cloud Code a prompt. It gets me something really cool, but it's not done. And then I'm like...
I can't, I don't have the mental energy right now to get this work. So I've got like 20 unfinished branches. Yeah.
RICK (39:48.002)
Yeah.
Yeah, But that would say that's tinkering. So you are doing it.
Tyler King (39:55.282)
doing it but but a lot of the so I think the difference between your interaction with AI versus mine or your job function versus mine coding is like the frontier of how to use AI right now I think and there's a lot because it's the frontier it's it's less like here's all these proven like here's this AI SDR you can just go buy off the shelf and it's a more like here's 20 different things that if you stitch them together properly it's magic
So like the type of thing I'm talking about is, okay, when you're in planning mode, use Opus 4.6, but then switch to Haiku to actually execute these types of things so that you're not waiting for too long, or how are we gonna use working trees and Git to switch between branches so that you can run multiple agents at the same time? It's much more like nitty gritty tooling stuff that you can't test. I can read about it, I can understand it, but I have to be actually working.
full-time as a developer to use it, you know?
RICK (40:54.412)
Yeah. Yeah, it's not, it's too big.
Tyler King (40:58.769)
Yeah. And other people on my team are doing that, to different degrees. So it's, it's not like we're, I'm not the bottleneck here, right? Robert and Bracken can go push forward the frontier, but it makes me feel like I'm reading all this stuff and I'm like, I've got these ideas and I want to tell people about these ideas, but I feel kind of like a, an empty suit. Just being, bringing ideas and not actually executing on them.
RICK (41:23.416)
doing them. Yeah. Yeah. Makes sense.
Tyler King (41:27.459)
surely we've got something unrelated to AI to talk about here.
RICK (41:31.244)
Yeah. So, well, one thing that, that I'll, I could brought an update on it with leg up health is that, know, switching to sort of planning versus AI, like we have spent most of like our, planning cycle starts February 1st. we, we, you and I started moving to planning probably middle of January where JD was still wrapping some stuff up. It was very clear in the middle of February that JD was like, I am now ready to plan. So we've been sort of.
We've had a couple of partner meetings where we've gotten deeper on like, what's the plan for the year? I think we've narrowed in on like some conviction around two couriers of focus. and so I feel really good about that, but I just wanted to talk a little bit about the experience. Cause you, you, you kind of said, Hey, let's go for a, it's good to go for a walk. And I was like, okay, we're to go for a walk. and what I mean by going for a walk, I mean, is like, let's stop, you know, let's get out of the constraint of what we said the plan was for the next three to five years and pretend it's not the plan and we can do whatever we want.
today. I've never been more uncomfortable planning in my entire life. That made me so uncomfortable. And I could tell it made JD very uncomfortable too. I don't think it made you uncomfortable, but like
Tyler King (42:37.107)
Hmm. It felt pretty low stakes to me.
RICK (42:42.314)
Yeah, like, it just, when you go through and start turning over rocks and like thinking about like, well, what if we did this? Like it makes you question everything about the business. And it's very, very uncomfortable because you start exploring like, what if we did something completely different than what we do today? And, anyway, I think like if we had gone through that exercise and found something that was better than what we were doing, we might pivot.
Tyler King (42:52.232)
Mm-hmm.
RICK (43:09.272)
to that thing, but we weren't able to find something that was better than what we were doing, you know, at a core, you know, value proposition perspective. And then we, we kind of came back and you know, anyway, it was very uncomfortable, but, as a result of that, I can feel both JD and I have more energy and more conviction around the things that we are going to do than we would have had otherwise. So we are as a result, I think of doing that sort of nature walk, and uncomfortable planning. We have fewer priorities.
this year than we would have otherwise. We aren't trying to do too much. And we have much more conviction around those fewer priorities.
Tyler King (43:40.243)
Mm-hmm.
Tyler King (43:43.965)
But if I'm hearing you correctly, the priorities were already there. It's not like some brilliant new idea came up. It's just, let's confirm that this is what we actually want to be doing.
RICK (43:55.15)
And I think there was the thing that I think that was the, so the hidden benefit here is narrow or focus is allowed by more conviction, which is only gained by picking your head up and like really turning over rocks and challenging, uh, what you should be doing. Um, and I think if we had done through our normal playing process, we would do what we've done in the past, which is let's try to do 10 things. And we would make minimal progress across all 10 things.
Whereas right now we have two things and I think we'll make a lot of progress across two things.
Tyler King (44:28.893)
And I think specifically my observation of what, I agree with you, nothing really got added, but things got removed. Focusing on the individual side, or rather not focusing on the individual side and just accepting, if you look at the current business, which is working, again, we've said this before, like LegUp Health, as much as it feels like a super early stage startup, whatever, it's making six, what, like 300,000 a year or something like that?
RICK (44:54.51)
Yeah, we're about a 20 20k MRR. So whatever that is 244.
Tyler King (44:57.041)
Yeah. It's like, you know, it's small, but it's a real business. We can look at what's working and what's not and all that. And we can just look at it be like, all of the good leads came from one place, basically. Okay, well, let's do that more.
RICK (45:10.88)
Yeah, exactly. Okay. Well, what's the challenge with those? Well, we need more of those leads and getting those leads into a legitimate sales process is challenging. Okay. That's all we're to focus on for this year. If we do those two things, we got a great business.
Tyler King (45:22.823)
Yep. Going zero to one is so much harder than going one to 10 and all the advice is zero to one. I've said this before in the podcast that if you just go out there and follow startup advice, they assume you don't already have customers to lean on. You don't already have experiences. all the do the customer interviews and all that. And again, I'm not saying you shouldn't do that, but yeah, it's so much simpler when you can just say, it's already working. Let's do more.
I do want to comment on the like, if being uncomfortable to me, to you is a little funny to me because like Bracken and I go so much further out into the wilderness when we like even recently, 16 years into the business, we were walking around like, should we start a plumbing company? Like what is AI not going to replace? And we actually talked about that for real.
RICK (46:16.365)
I'll tell you like, let me let me be clear, like it made me very uncomfortable doing this for the I would think this was actually the first time I've ever done something like this. What I do it again is the real question and I would absolutely do it again. And I would probably do it more aggressively. So I bet like every time you do it, the benefits are so like strong from doing it that you're just like reinforcing that you should do it deeper and deeper and it gets more extreme every single time. Like that's probably a good thing.
Tyler King (46:40.721)
Yeah. I, yes, I, I didn't exactly want to make it sound like what we normally do is the right thing though. Like this has caused problems before specifically, like you, lose focus really easily. If you think there surely is a better here, sorry, the, the, the visual I have in my head is imagine like a mountain range, right? Like up peaks and valleys. And you don't know if you're on the highest peak.
Or I mean, you're not at the top, you know that like there's an uphill and a downhill. so you can be working uphill, uphill, uphill, but you don't know is that peak going to be higher than the one next to you or the one next to that. And so the question is, do you walk uphill or do you like hit the random number generator and transport to another mountain and hope that that's a bigger mountain and, walking uphill is almost always the right answer. But this line of thinking can get you to try to attack a different mountain. And I've said.
Many times, I think I wasted a decade of my career at Less Knowing CRM working on the wrong stuff. Like roughly 2014 to 2024 or something like that. I mean, we made progress. It's not like we didn't, but we were talking about how are we going to scale to billions of users? And we were talking about when we eventually are more than a CRM and well, what about freemium? And we were talking about all this crap when we should have been saying, what is the next feature that our customers need to make this a better CRM?
So that's the flip side is you can get distracted really easily.
RICK (48:15.726)
That's fair. Yeah. So, so like the thing that I like one governance governance on that is like, or governor on that is don't do this more than once a year. Like I wouldn't I would not open this unless something went bad with the business south of the business. I would not like we're not reevaluating our priorities until we solve these two priorities or something else changes like externally.
Tyler King (48:25.234)
Yeah.
Tyler King (48:34.003)
Mm-hmm.
Yeah, I'll say, time will tell if I'm currently making the mistake I'm describing, but I don't think I am. I think we finally got our heads on straight and we're working on the right stuff. A big vibe shift for me is at the end of the year when we have this conversation again, we look back at the last year and we're like, yeah, we worked on the right stuff that year. If you're constantly like, if let's say this year we said, we need to pivot entirely into group. And then next year we say, you know what, it's all about.
employer benefits, it's not even about health insurance. And then the next year we're like, we're now a cell phone reimbursement company. I think that's a good like sniff test to see, are you just coming up with like reasons to like, are you just brainstorming for no reason? Like if you're not staying focused on some sort of consistent coherent vision, that would be a good sign that it's not working. But yes, I agree. Once a year is probably about as often as you would want to do this.
RICK (49:28.878)
I
So thank you for that guidance. think that was you adding a lot of value to our planning process. So anyway, I'm happy to get focused on execution.
Tyler King (49:43.175)
Yeah, I'm excited. I both feel relieved and a little imposter syndrome that the execution has nothing to do with me basically. So we're kind of having these meetings and then I'm like, all right, good luck guys. I'll see you later.
RICK (49:55.086)
Yeah, whatever. We're gonna need your help on the coding side. We got to figure that out. I know I you some thoughts. So, but hopefully that'll be your tinkering. That'll let you tinker.
Tyler King (49:58.963)
Yeah.
Tyler King (50:06.835)
You know, Legop Health has been good for me in that the stuff I have built with AI has mostly been Legop Health. Again, I've 80 % built a bunch of stuff for Lesson on Xerion, but I haven't actually shipped anything.
RICK (50:21.752)
Do you think that we could get sort of always on developer agents on leg of health?
Tyler King (50:29.811)
I'm nervous about that. Like could we? Yes. And could it be building stuff? Yes. I think like the constraint now for most companies is figuring out what to build. And I don't think, so I believe AI is big. It's a big deal. It should be used. I don't think the industry has fully reckoned with how much damage it's going to do when used carelessly. And I think a lot, A, a lot of people talking about
all they're always on agents. They're not shipping anything that they just talked about this on the mostly technical podcasts. Like there's all these people on Twitter being like, look at all how fast I'm moving. And then it's like, okay, what did you build? Where's the business here? And they don't have anything. And there's all these big companies that are talking about AI. And then you get into the details and it's like, okay, so your product has, if anything gotten worse over the last year, you're drowning in technical debt. You've had more downtime than ever before. I just think.
I think companies need to be really more careful than they're being. What do you think?
RICK (51:36.578)
I think it depends on the business that you're in. think if you have like, I mean, if someone is depending on you to do important work, like you have to have a lower threshold, a lower tolerance for things breaking. But if you're a productivity tool and you know, I would say that you have some shots you could take as long as you don't totally destroy people's productivity in the, think Notion has the ability to take some shots.
Tyler King (52:05.799)
Yes, yeah, Notion hasn't had any devastating problems yet, but losing data, leaking data, having major downtime. I don't know if you've... The article's coming out right now about AWS's downtime was basically management said to developers, have to use our AI a lot. And then they did, and then the AI just deleted a bunch of shit in AWS. I think it was recoverable, but AWS has had terrible uptime this year, and apparently it's entirely because they're overusing AI.
You're right, AWS is more important than Notion, but Notion's important. If Notion goes down, I can't do my work.
RICK (52:40.899)
fair.
Tyler King (52:43.719)
Yeah, okay. Probably a good stopping point here.
RICK (52:48.174)
Sounds good. Well, if you'd like to review past topics and show notes, visit startuptolast.com. See you next time.
Tyler King (52:56.605)
See ya.