Skip to content

19: Your Old App Is Full of Holes

• 46 mins 51 secs

About this Episode

A year ago, the Speaker Deck offline app stalled out. This time, Kris went from idea to App Store submission in about three weeks. Meanwhile, John moved hundreds of terabytes of Fireside's storage to a new provider and built Nunabot, a bot that handles the tedious upkeep across 17 apps. Security fixes, dependency updates, performance tweaks. He went from averaging four contributions a day to 30, and merged 110 changes in a weekend from a campground.

Then we get into the other side of all that speed. A wave of security vulnerabilities hit last week, and AI is making both the good guys and the bad guys faster. We talk about what every app owner should be checking right now, whether 2FA is worth the friction, and how Kris taught Momentum's founder fellows to run AI reviews on their own sites.

Plus, the investor meeting that was way scarier in Kris's head than in the room, and why John's first question is always "how could this break?"

Show Notes:
00:00 Intro: a new tagline
01:11 Why the Cybercab episode took off
02:25 Embers: Speaker Deck offline app in three weeks
07:55 Moving hundreds of terabytes off B2
11:05 Nunabot: a bot for the boring work
12:42 From 4 contributions a day to 30
17:13 A week of security fires
21:32 Should every app have 2FA?
24:30 The student becomes the teacher
30:56 110 changes from a camp chair
32:39 Into the Fire: the investor meeting
36:23 Optimism vs. "how could this break?"
44:33 Recommendations and a challenge

Standing in the Fire is produced in South Bend, Indiana.

Transcript

John 00:05

Hello. You are listening to standing in the fire where myself, John, and Kris, the owners of fireside dot f m, talk about what it's like to be in the heat of entrepreneurship. I've changed it again. And I... The last one I said I was not gonna change it and we were not gonna discuss the intro Yep. Like we do every time, but I've decided at least one more time we're gonna discuss it because I was like, the heat of SaaS growth just... It doesn't Yeah. It doesn't work for me. Because I'm like, first off, you know, fireside is not growing. So that's weird. Other SaaS that we have are growing. But... So, like, that feels that feels wrong. But then additionally, like, you know, the heat of SaaS growth is just like a weird phrasing of things. That was probably a claudism or something.

Read Full Transcript Hide Full Transcript

Kris 00:47

So Probably.

John 00:48

I was like, you know what? Let's just call it entrepreneurship. We talk about momentum a lot.

Kris 00:51

Yeah.

John 00:52

We're both involved in that. You know? I'm not nearly as involved as you, obviously. But I thought maybe first we would just go into, like, some of that, like, Amber's recent progress on Momentum, on Speaker Deck Yeah. On our apps and stuff like that instead of going into, hot takes like last time or other things.

Kris 01:11

The hot takes, I don't know if you saw, are doing well. When you... Are they? Your hot takes on auto auto robotic travel.

John 01:21

Oh. Yeah. Yeah.

Kris 01:22

Okay. All the the cabs.

John 01:25

Well, what... Did... Is it doing well because of anything I said or because it had an awesome cover? Because I feel like the

Kris 01:31

cover

John 01:31

was pretty good, and I was not involved with the cover. Like, the whole, like... What was it? I I rode in cyber cap and I cried or something. Like, yeah. I feel like that was really good. I immediately screenshotted that and posted it. And so that... Maybe that Yeah. Helps doing a little extra or something. I don't know. Yeah. Maybe just

Kris 01:46

I think when it has emotion and it also then has an opinion Yeah. That that creates... Could create divisiveness is Oh. Two ingredients. Yeah. Because some people dislike automations and AI and Elon Musk.

John 02:00

So They wanna

Kris 02:01

watch it. Yeah. They can... It can create battles. I haven't comments. Looked at

John 02:05

Well, we we probably should look at the comments. I'm sure that they are, you know

Kris 02:09

On the shorts is where they leave them typically.

John 02:12

Okay.

Kris 02:13

Yeah.

John 02:13

YouTube shorts. Yeah. People don't wanna actually get the whole story. They just want Yeah. The quick the quick snack, the junk food. Nice. So do you wanna go into your first unburst? You mentioned before Speaker Deck.

Kris 02:25

Yeah so I think it was the only last episode we had talked about how Eric and I were working. Tried to work on an offline app for Speaker Deck

John 02:33

Yeah.

Kris 02:33

Maybe a year ago and it was just too cumbersome to try to get to work because we needed to work without Internet. And I was like, I was like, it's conference season again for that audience like October, November. Like, what... Can we try it again based on the stuff we've done the last couple months?

John 02:49

Yeah.

Kris 02:50

And within, I don't know, maybe it was like two weeks from when I suggested that we he, you know, we get back on it. We had a app in the beta testing and test flight had probably 50 people signed up to do it maybe 30 actual like downloaded the app and we're getting feedback and I think it's like we already now submitted it for a full approval.

John 03:12

Oh cool.

Kris 03:12

So like within now let's say three weeks went from we knew the we already knew what the core idea was what we needed to do. It was more just downloading off files for offline and viewing them without internet was the part where we kept struggling with. Yeah. And yeah it's it's like it solves a very specific niche need but yeah, it works. And I don't like, don't know. That that was like, it used to be months and months of development time to get something like that out there.

John 03:41

Yeah.

Kris 03:41

So, yeah. I know sometimes we talk about stuff on here that we're gonna do and then I don't do it. This is one where it it got executed.

John 03:49

Yeah. You know, was just listening to our last episode on the way in because I I listened to him. I tried to like improve, you know, etcetera. I feel like the best way is to listen.

Kris 03:58

Yep.

John 03:58

And there was there was... And now I'm forgetting what it was. But there was a thing you were like, I'm gonna do this. And I was like, oh, I gotta call him out. And then I didn't take a note. And so I don't remember. So I'm glad that you did do a thing that we talked about last week too.

Kris 04:12

So it's just like I think it opens the just ideas of like what else can be done in that way that there's been an idea ruminating to do something.

John 04:21

Yeah.

Kris 04:22

And even if it's one of the things I don't understand because I think you guys have done... And maybe it's... You guys are using the same process with Box Out of... So the game you showed me, like, two episodes ago. Did you use TestFlight to do all that, or is that a completely different way that you got that to be on your phone? So I'm just thinking like micro apps Yeah. That may be good

John 04:41

for

Kris 04:41

a specific use case. It doesn't have a larger audience, but it solves a problem. How are... How did you deploy that?

John 04:48

Yeah. So I... You... With... If you have an account with Xcode on your machine, you can just... If you're on the same Wi Fi network or wire plugged in Yeah. You can just put a phone... An app on your phone. So that's one way to do it.

Kris 05:02

Yeah.

John 05:03

The other way to do it is with TestFlight. So I first just did it with a wire. Yep. And then I was on the same network, and so I started doing it that way. And then eventually I was like, which again, I'm saying this like it was over the course of weeks. It was like the first minute was plug it into, you know, and the second second minute was wireless. And then eventually, I was like, we'll just create test flight Yeah. Accounts and do all that stuff and set it up that way. So So

Kris 05:32

if you're doing, like, personal stuff, are you putting it all all in the same developer account?

John 05:36

So my personal stuff is going on the developer account. Yeah. Yes. We also have a box out developer account.

Kris 05:42

Yeah. Which is separate.

John 05:43

Yeah. Garrett is actually setting up a very good developer account for us, I think, right now because he's been working on an app, he's like, might as well put it in very good. Who cares? Yeah. See what happens with it. So I think, you know, that's the right path. It's very awkward and weird and you need, like... Sometimes some of the stuff you need, like, weird numbers and stuff. I don't know. Like, I don't know where you get them.

Kris 06:04

But But once you set up... It's not straightforward Yeah. For sure to set up an account. But once you get it, like, doing it again is gonna...

John 06:14

Is easy.

Kris 06:15

Yeah.

John 06:15

Yeah. Yeah. So... Yeah. I mean, so we use TestFlight and, know, I added once you... I think you... If you submit it through TestFlight to let, like, random people in, then it takes a little bit longer to get approved. They do some kind of AI automatic check and they say, cool. Yeah. Yeah. Like, you're not, you know, harvesting password or doing something, you know, very obviously bad. And so then they let you in a lot quicker for test flight, which is nice. If you do the full submission, you know, obviously, it takes a lot longer. And so we had we had, like, our screens app for Box Out. You know, it took a little while to get through there. It got rejected, like, the first time Yeah. Yeah. Of... I think it was because it... The sign up page, this is just something for you to think about for your approval. It may end up getting rocked because of this too. Our sign up page, we took it to our regular sign up page, which has, like, you know, email password, like, sign up, sign in, like, all these kinds of things. And Apple took that as like us trying to like skirt their payment system. Interesting. And so they... Like, we literally had to create a new sign up just for

Kris 07:25

Probably getting rejected then.

John 07:26

The Apple TV app Yeah. That was different and that followed the rules a little closer. Yeah. And they didn't say we're rejecting you because you're not giving us a cut. They they had some weird generic thing, then Claude was like, this is what they're doing. Got it. Yeah. But... Yeah. So that for sure have to do that. We'll see. Maybe not. It's entirely possible. But, like, that's what hit us on the Apple TV screens submission was that the first time. Think that's what Steve said. So Nice. Yeah. How about you? Do you do you

Kris 07:52

have any... I'll I'll think of some others too. But do you have any embers?

John 07:55

Yeah. I've got I've got quite a few, like, little things. So I feel like the biggest little thing that I did... Well, do... Which one do wanna talk about first? Okay. First, we'll just say we've been working on moving... We currently, for fireside, have all our data stored on a a product called b two. It's just works kinda like Amazon s three, which is just like storing big files.

Kris 08:14

Mhmm.

John 08:14

And we have an enterprise account with Cloudflare, and they have a deal between the two of those to, like, transfer between them is free, which is nice. We... Turns out we don't really need the enterprise account, and the enterprise account is very expensive. We actually just need, like, the business account, which is a couple $100 a month. So... But in order to do that, I had to move all the data off of b two onto r two, which at first, I was like, this is gonna be horrible. It turns out, again, because of the world we live in, it was not horrible. It was time consuming, but, like, it was not horrible at all. Like, we have terabytes and terabytes... Tens and... Or hundreds of terabytes of data, so we're talking, like, a lot of data. In the past, I don't even know how I would migrate all that. Like, it would have taken forever and been really hard. And instead, it was like, oh, Cloudflare is just, like, click a button and they install this thing they call super slurper, which just pulls it all over. And so, anyway, we we did all that, and then we set up, you know, automatic copying basically from b two to r two. And then I switched reads. So anytime someone went to our...

Any of our podcasts, not just ours, but, like, customers' podcasts, they would all the reads would come from there just to make sure everything was fine. And then this morning, I switched rights. So as of now, we're, like, all the way onto r two. We're also... We have v two there as, like, backup. And if we need to, like, move back, I can transfer the files that have moved in between and stuff like that. But, like, I don't think there'll be any issue. I mean, at this point, r two and Cloudflare... Cloudflare's half the Internet, and r two is one of their products, so I have a pretty high trust with them. So that was kind of a... Just a, like, a neat thing that I did this morning. I was just like I was like, you know what? Normally, I would do this really slow, but I was like, we don't get a ton of uploads in a day. Like, we get a decent amount. But I was like, it's not like reads where, like, we get... I mean, we move a ton Right. Every day, like, terabytes of data. And so I'm like, with writes, it's like, okay. X number of people are gonna upload.

And if I set this to, like, five percent or 25%, I'm like, I'm gonna have to wait so long to be like, are people having a problem with this or not? You know? I was like, I don't wanna wait. So I was like, I just go to a 100%? And, you know, Claude was like, okay. So we went to a 100%. And so, like, every twenty five minutes now, Claude just is checking errors. It's checking support. Yeah. It's checking performance, all this kind of stuff, and saying, like, cool. You're still good. Cool. You're still good. And then if it finds a problem, it'll just automatically fix the problem and make a PR, and I have to, like, approve it to go out and stuff. So so that was, a a tiny thing that went out. The big thing that went out is I started working on this thing called Nunabot, which is basically just me, not incarnate, so the opposite of incarnate. So out of carnet. But it's basically just like... And it's something we've talked about a lot here, but, like, things that AI is really good at. And, also, like, all the things that are on our plates. You know? Like, if I only had to work on, like, one app Mhmm. Probably... Maybe I wouldn't have built Nunabot yet.

Kris 11:05

But

John 11:06

I'm like, I have... I mean, there's, 17 repos that I end up in, which means there's basically, like, 17 apps, and it's probably three, four, five, six somewhere big ones, you know. And I was like, oh my gosh. There's just like so many of these and they all have dependency updates and performance things they need to tweak and errors that happen and like all these kind of like just the standard things you do to run a SaaS. And I was like listening to another podcast and they were talking about automating some of that stuff and then I was listening to another thing. All those mixed together and I was like, alright, I'm gonna start working on a thing that will do all these tedious things so then myself and everyone else on the team for all of these products can just focus on the higher level things. Yep. Yep. Because, like, that's what we wanna do. Right? Like, we want Yeah. We wanna say, okay. Let's spend as little as possible time on lower value things, but they need to be done still Mhmm. And as much possible time on higher value things. And so my idea was basically find some way to automate that.

So I spent a bunch of time kinda thinking about it, brain dumping, using whisper flow, voice dictation, threw that into Claude, made a prompt, pulled all the stuff in, and I started running it just manually. I'd say, okay. Run Nunabot on like one... On Box Out. Run Nunabot on speaker deck. Run Nunabot on flipper cloud. Like, did that on repeat for a little while, built up some confidence, and then I was like, okay. Run it up to five times on every repo that I'm in. And I was like, cool. And then I... One night, I just was, looking through them, and I'm like, merge. Yep. Merge. Yep. Merge. Merge. Merge. Merge. And I was like, wow. This is this is doing good stuff. And it's stuff that, like, normally would just never get done or I would have to do it.

Kris 12:41

Right.

John 12:42

Right. And so I just started, like, I was running again. Merge. Merge. Merge. Run it again. Merge. Merge. Merge. And so, like, you know, we've talked about... I've mentioned the productivity gains I feel like I have gotten from AI. You know? So, like, my seventeen year, sixteen year baseline without AI was, like, four contributions a day, eight on a given workday. Mhmm. So proves that I didn't work on the weekends and stuff a lot. And then, you know, AI kinda hits and you get, a 30% bump with, like, auto complete being better than it used to be, but I'm still in an editor. Mhmm. Then I'm, like, pencils down in December. I don't don't write code anymore. AI writes my code. And I moved to, like, four times. So I'm up to, like, you know... Let's say I was at, like, 16 or 17 contributions a day. You know? And so then it's like, okay. Cool. Like, you can see AI made a a bump. You know? And then in October, there was another huge bump for me. I just figured some stuff out, and I hit, like, 30 a day. And I was like, wow. I am cooking. I was averaging four a day for sixteen years.

Kris 13:45

And Yeah.

John 13:47

Now, you know, in October, I'm averaging 30 a day. I'm like, that's that's seven... What? Don't do... Don't do public math. You know? Almost eight times. Almost eight x. And I'm like, that's almost an order of magnitude. Was like, wow. That's a lot. Nunabot, it's insane. I have... So I would average about 1,500 contributions a year. Mhmm. I did almost a thousand last week. So, like, you know, it's just... It's insane once you start to automate. Now none of these are, like, they're not massive Right. Needle moving things. They're not... It's not taking on things that, like, I actually need to use my brain for that AI is not necessarily as good at and needs to be kinda shaped based on a lot of context that I have not put out. But I'm like, this idea that like something can run-in like fixed performance, security, dependencies, marketing Mhmm. Things that can do new landing pages, it can do SEO, it can do geo which is like AI SEO and stuff. The fact that it can just kind of do... Take all that stuff off my plate. And I got to where I... Once I had merged like the fiftieth one, I almost just like looked at the title and looked at the content and I started just merging them.

And then I was like I was like five in, and I was like, wait a second. I haven't looked at the code. Maybe I should be looking at the code. Like... Yeah. So that that was, like, I think my biggest ember was just, like, creating this little bot that can, like, crank out the tedious stuff for me, and it's really good at that. And it's not fancy. It's just a prompt and a bunch of things glued together.

Kris 15:17

It

John 15:18

it literally looks like the first, you know, SpaceX rocket that's got crap all over and looks ugly. You know? But over time, now I see, okay. Here's the path to making it so that it can do do this kind of, like, more on its own and less with me. Yeah. And probably, honestly, get to the point where I don't look at the code. What I do is maybe I I... You know, it's collecting the work to do. Mhmm. And then it's prioritizing the work, and then I say yes or no to the work. Then Before it. Yeah. And if I've said yes or no, you know, then it just... It goes all the way through. Takes it to prod, deploys it. I don't look at anything. And I I honestly, I think before the end of the year, maybe before the end of November, that's the way I'll start working for for this level of work. Right. Right. You know? Like, very Yeah. Yeah. In size small. Yeah. So that was my biggest one. Nice. Yeah. What I... And I've

Kris 16:07

noticed what... What's gotten different. So historically, when I'm working with software developers over the last whatever ten years, most of the time is like waiting for them to get the feature built. And now I feel like it's more waiting for CI test to run. Yes. So like you're just sitting there for twenty minutes, you know, and like at night Yeah. I'm like, okay, like I wanna try to get through stuff, but I did then have to sit there and wait until then I watch TV or whatever. And, like, that's that's the bottleneck right now is waiting for those tests to run.

John 16:40

Did you see what I did this morning?

Kris 16:41

I did not yet. Okay.

John 16:43

So I I fixed that. So for speed back. Yeah. So this this morning, I just, was, like, speed this up.

Kris 16:50

Yeah.

John 16:50

So it... It's it's not merged yet.

Kris 16:53

Okay.

John 16:53

But I've switched it from Heroku CI to GitHub actions and parallelized it. So it was taking sixteen to seventeen minutes on average on Heroku.

Kris 17:03

Yeah.

John 17:03

It will now take under four minutes on box outs.

Kris 17:05

And so we need to get... Well, anyway, away from Heroku at some point. Right? Because it can definitely... Yeah. And not to railway.

John 17:13

I don't know if we talked about that, but... Oh my gosh. Yeah. So they... This is the downside of AI. So everyone's more productive, including research people for good and security vulnerabilities and also security vulnerabilities for bad. So railways environment variable, which is their secret... There's basically their secret storage. So everyone's tokens that have access to do all the things in place of a human signing in. Some amount of them got hacked. They didn't know which ones, and so they're like, everybody has to rotate all your secrets. It was a mess, and it took, like, two days out of my vacation. Just literally wrote... I just started my highest revenue app, rotating all the secrets until I got through all my personal apps and down to the tiniest thing. It was, yeah, like a

Kris 17:58

historic Yeah. Something The last week though for everything. Yeah.

John 18:04

Yep.

Kris 18:04

Yeah. I I mostly had PHP issues. Yep. Like that was my probably most exposure. Luckily the hosting services I use automatically update everything but I did email some clients and had to do some stuff just because I wanted to make sure that they felt like they were in the know of what was happening. Yeah. But, yeah, it's... That's...

John 18:28

It was It was bad. I mean, it was it was PHP, which is... I mean, they... They've stated right now it's 43% of the Internet Yeah. Websites. It was... Ruby had a small one. Node had a had a big one so several versions of Node got affected. It was just like on repeat.

Kris 18:48

When I I was almost thinking over the weekend because I think yeah it was last weekend that I was like should we put something in our like member portal Because because when, like, you're on top of it. I... I'm on top of certain things. Like, I I I see it because of the things I follow. But, like, the members who have apps, like, do they follow it? Like, because depending on where do they how far to remove they are from the technical side, they may not. Yep. And they don't know if the developers are. Yeah. To just send like, hey, guys. Like, your apps and check your, you know, websites to make sure that it's... Things are getting updated because there's huge flaws.

John 19:24

Yeah. Oh, massive flaws. And those are the good people. Those were the good people who were, like, reporting them responsibly, working with the, like, top down, you know, the biggest PHP host site and, you know, WordPress sites, etcetera, in the world and, like, responsibly. And then there's, like, you know, railway where they got hacked by someone not being responsible who stole, you know, AI tokens, crypto, like, you name it, a whole bunch of stuff. And that that stuff's just gonna keep happening on a repeat. I mean, I I would... I told the Box Out guys on Tuesday at our meeting. I was like, honestly, it's terrifying. Because, like, I'm just seeing all kinds of stuff get... Even just... Who was it? It was one of the podcast hosts announced Libsyn or not... I don't know. I don't think it was Libsyn. It was another one. They just announced that, like, someone... Basically, they they weren't correctly handling the podcast lock tag or, like, locking mechanism. And so someone, I think, got access to someone's email, used that, or I don't know. Somehow they managed to transfer the podcast. They had a security hole. That's what it was.

So anyone who was on the platform already could somehow magically inject something and transfer a podcast from one account to another. It was something like that. So they got access to it. They put it all into AI to recreate the host's voice. This is a very popular podcast that, like, makes money. Yeah. So then they started making new episodes for that host, which the host was not involved in at all, and submitted for Dynamic Ads approval because they're getting all these downloads. Yeah. You know? So I don't know if it was, like, a dead podcast that the host wasn't... You know, but it was still popular or whatever. But, like Wow. Literally... And so, like, you know, my AI news, like, showed that, and it was like, FYI, you might wanna make sure, know, Fireside is respecting this. I'm like, yeah. We are. It's fine. But I was like, oh my goodness. So... Yeah. So, I mean, Nunabot's been opening up, like, a ton of security vulnerability

Kris 21:27

things,

John 21:28

like, all over all the apps that are just, like, in...

Kris 21:30

Intense

John 21:31

Yeah. To find.

Kris 21:32

Do do you thoughts on... I I I thought of this over the last week, I think, with some of these things. Like, what should we add to maybe the login? I know some things are annoying and some things aren't really as secured just maybe as perception. But having that like speaker actually add like a 2FA option Yeah. Of something whether it's a text or an email. I know it's super annoying for users but is is having it better than not having it? I think actually because when a corporation sent sent me like a checklist of security questions and one of them was do you do you offer that? I was like, oh, like, should we? You know, just to make another layer of security. Do Yeah. Do you guys use that on any of the apps that you have currently?

John 22:13

We need to. Yeah. We need to add like OTP one time

Kris 22:17

password

John 22:17

Yeah. Types. Oh, so boxed out technically does... Anyone new who signs up doesn't sign up with a password. They just sign up with a one time password code. And so Okay. Yeah. So we, at some point in the last two, three months, switched to that. So, like... And if you... If you're... If you have an older account, it will still ask you for your password and stuff. But we basically have given the ability to be like, no. I don't want a password anymore. Just send an email to me, give me a code, paste it in, then you're in, and you're in. You have access. So that's what we currently do. I mean, there's a there's a bunch of stuff on top of that. You know, it's like, well, if codes have been sent, you know, it's like require, like, a code in order to change your password. Require your current password in order to change, you know, up to a new password. Whenever the password changes, invalidate all emails, links, sessions. Like, there's a a ton of things like that. Rate limit, how often people can submit password, use Yeah. You know, bcrypt. Like, there's a bunch of security things like that. So... Yeah.

I mean, I think it's gonna be going forward for the next several months, you know, no one being sure, a, how to regulate AI without falling behind dangerously Mhmm. And b, how to stay secure Mhmm. Over the next little bit, it's gonna be dicey. Yeah. Yeah. So it's just gonna be a little bit crazy. So what I'm doing is I'm just like, okay. Let's get a lot of that stuff going, and then I'm going through all my apps and being like, okay. What is every single thing that I should be encrypting? What is every single... Like, least privilege. That's the biggest thing that you typically wanna do for security. So you would... I wanna create a new AWS key that has full access to your AWS account if you're just putting files into s three. You wanna isolate it to just that. So even for speaker deck and stuff like that, going through and making new keys for all of that stuff, even though there's not any issues currently. But, like, the app is 15 years old. Like Yeah. You know, it's entirely possible. So that that... That's the kind of stuff that that I I would recommend for

Kris 24:20

people right now. But, yeah, it's pretty bad. Nice. I don't if we're staying on the same topic right now, but I had another one that I thought Go for it. So the the student became the teacher last

John 24:30

week

Kris 24:30

Okay. Of... So I I... We we have, you know, the the five founder fellows and so I I taught them how to set up their

John 24:43

Oh, yeah. Okay. I had no idea where this is going. Yeah. Yep.

Kris 24:47

And use g stack reviews Nice. For for their websites and for their copy. And it was interesting because, like, I sometimes think I'm behind on some of that. Yeah. But then knowing that the fellows, like, because of all the different things they're responsible for Yep. They're not staying on top of it either. And so going through one of the guys who... He has an app that he has, obviously, a public facing marketing site to get people to download. And we we focused on two things. One was going through G Stack review for technical and SEO things and then go also going copy but copy based on him having like a target certain target audience. And I sent then I sent him the results because we went through a class like it was a little it took a little long. It does. But I but I wanted to show them step by step of what I was doing and they were like blown away. Yeah. So That's awesome. I I like... I'm not an expert, but I I knew enough and it was like cool now that hopefully they can move faster and get get more things done now that like they have like that extra tool versus just like... And and and I caution them on...

And I know I just switched sentences in the middle, but of not trying to just, like, use it to continue, like, redesign or or rechange the copy, like, every week.

John 26:06

And Yeah.

Kris 26:06

But just, like, we're... Keep refining what exists because I think you can get caught of because it's so easy to redesign their website like every week and you'll just be doing that and then you'll never actually be making a difference.

John 26:20

Yep.

Kris 26:21

So we'll see like tomorrow or early next week of has anyone gone through and and done anything but I thought that was pretty fun.

John 26:28

That is cool. And I feel like that's that's like the best thing about AI or the worst thing I guess. Yeah. The worst thing is you can feel insanely productive and do tons of things and those things don't move the needle. Correct. Yep. Like, you still... Like, I feel like the people who are really successful with it are the people who stay laser focused on, this is what will move the needle, and I will use AI to help me do that faster. So yeah, think it's a good good thing a, to just show them how it works and how to do that, but then b, to like follow-up and see like who did it and who didn't. And I think G Stack is really great for like people who aren't technical because it helps them do technical things.

Kris 27:08

Yeah. And like focus on... Yep. Yeah.

John 27:11

It's very good at like being like this is what's important when you're building apps and websites.

Kris 27:15

And I think working within a framework of something. You you can put some guardrails of like saying stay in this lane.

John 27:21

Yep.

Kris 27:22

Because otherwise it's just too endless of what Agreed. Yeah. Yeah. That's one of the things that

John 27:28

Nunavut uses a couple of g stack skills. So one is it uses simplify from Claude which just kind of shaves, you know, a bunch of the cruft off and stuff like that and just improves the architecture, and then it uses G Stack QA, which is a great one. Mhmm. So it'll click through the whole app and make sure that everything still works, and you can tell it to record screenshots. Mhmm. I have it put the screenshots in so I don't have to like boot the app and look at it. Like I can just look at those and be like, no, that's wrong or or right. And then the last one it uses the the G Stack Codex one which is kinda weird but it basically allows you to do an adversarial review in Claude using Codex. I find building in Claude and reviewing in Codex is when I'm most productive. Interesting. Yeah. The last little while. Cool. Yeah. So the last two, three weeks. So a couple of those are G Stack ones, and then I've added a bunch of... You know, there's, like, CSO, which is, like, chief security officer. There's... Cloudflare just came out with some security ones, and so I've been recommending those to people as well just to get, like, a baseline of, like, is there anything massive?

Kris 28:30

You know?

John 28:31

And I'll be honest. Like, our apps... A bunch of my apps are ten, fifteen years old, and they 100% had security vulnerabilities that were severe. Yeah. But no one's found them.

Kris 28:44

Right.

John 28:44

You know? But it's just like what was standard at the time Yep. Is just different. Because think we even had just some, like, version upgrades that, again, it was made whenever and just need to up... It wasn't even significant, but it was like, oh, yeah. You're using 1.2 and it really should be like two point o now.

Kris 29:02

Yep.

John 29:03

Yeah. And that kind of stuff is just so easy now. Like, I... Ruby came out with the vulnerability. I don't even know what it was for. I just saw that it happened. Yeah. So I just literally started a new cloud session. I said, go through every one of my personal apps Yeah. And upgrade them all to the latest Ruby and the latest Rails and automatically deploy them. I don't care. I don't even wanna be involved. And if there's errors, fix the errors and

Kris 29:23

keep

John 29:23

deploying and do it until they're back up and running. Because I was like, on those, I don't care. Can Right. Right. Rather have it done and patched Yeah. And have ten seconds of downtime.

Kris 29:32

Yep.

John 29:33

You know, on Fireside and other ones, would be a lot more

Kris 29:36

Right. Careful.

John 29:36

So on those, I was like, make PRs, and then, you know, Gary and I got them all merged out for all the stuff. But it did it, and, like, thirty minutes later, all my apps are updated to the latest Rails, latest Ruby, all this kind of stuff and I was like, again, like that's just a great use case for like a bot. It should just be running all the time and say keep me up on the newest software and then make sure that it doesn't error when it goes into prod.

Kris 29:59

Mhmm.

John 29:59

And that it's fine. So I just think being able... I've actually been thinking lately that I wanna have, like, deploy bots that basically automatically run. Right now, I use my own agents, like, locally, but it's kinda hard to keep track when you're doing Nuna bot because I don't have, like, a task for each Okay.

Kris 30:15

ER. Yeah.

John 30:15

Yeah. Right. Right. That knows that it merged. Yeah. And so I I want automatic running of, like, okay. New deploy went out. This is what was in the deploy. I'm now going to exercise all those paths in production. Oh, error. Okay. Roll it back. You know? Or no error. Cool. Like, let's go forward. To the point that I think I'm actually gonna give it, like, I'm gonna create, like, a setup Fireside account, BoxSight account, speaker. Whatever, where that... It has full login capabilities for that one account

Kris 30:46

Okay.

John 30:46

So that it can thoroughly test the whole way through. Yeah. So it has the right permissions. And it might require multiple of those accounts. Like Box Out has tiers. You know? Speaker Deck might have, like, free and pro.

Kris 30:56

Yep.

John 30:57

So you can create one of each of those users so that they can kinda exercise everything Yeah. Correctly. So... Yeah. And then you can stuff them full of a lot of data, see if things got slow, etcetera. Stuff that I'm doing locally in development, but I'm like, it should happen in production. Yeah. Like, that's the best... That's where users are. So make ad in there. Nice. So... Yeah. That's... I think those are all my my logs. Most of them were just like... I mean, Nunavat did a ton, which was cool. Even just like... I think it was like Friday and Saturday or Saturday and Sunday, I merged like a 110 PRs from like a camp chair. Jesus. Like it was literally I had the most contributions in a single day in my entire life. And it was like I was walking the campground like while my kids were fishing with, you know, grandpa. Yeah. They love fishing with grandpa. They're doing that. I was like, I'll go for a walk. Okay. Nice. So I walked with them to the fishing place and then I just walked around looking at all the rigs. I like to see the different campers and how they do stuff. And I was just like, merge. Oh, look at that one.

Kris 31:54

If that wasn't a dad comment.

John 31:57

Yeah.

Kris 31:58

Yeah. Yep. Is that...

John 32:00

You know? I was in my New Balance. Yeah. My New Balance shoes, my white socks that go halfway up my calf, my cargo shorts, my tucked in T shirt with a belt. Yeah. No.

Kris 32:09

No. You you could almost be

John 32:11

on those...

Kris 32:13

Is it progressive commercials? Yeah.

John 32:18

Those are the best. They are verifiably. They're so good. I didn't know that Flo or, like, the mayhem guy from the other one would be, like, beatable, but those are so... No. We don't need to tell them about their hair.

Kris 32:29

Yeah.

John 32:29

Their hair is okay. It's not your business. Yeah. That's very true. So... Nice. Okay. So how about we go into what's been

Kris 32:39

hard

John 32:40

Yeah. Or easy? You've got you've got it hard.

Kris 32:42

Yeah. We we had a, you know, important... Well, they're all important, but like an important, you know, investor meeting earlier this week. Yeah. And there was just a lot of, like, prep and angst, you know, going into it because you don't know how people are going to react to updates. You like you don't really have a concept because you like talk to folks along the way but unsure if the progress is what people expected if it's not what people expected and there's a I mean the whole day like up until literally like 05:45.

John 33:18

Tommy's tum tum was hurting? That was my... That was what I'm thinking. I was always picturing that.

Kris 33:23

Very true. I mean I didn't really even eat that day.

John 33:26

Yeah.

Kris 33:26

And then you know I'm getting calls at like 05:00 and then 05:30 from like two advisers. Yeah. And so I took the calls because I was like, I wanted their input because I was actually changing the slides a little. I actually ended up changing them based on some of those calls that I had. Yeah. And it was the... But it's also like, I don't know, keep not finishing the sentence, it ended up being great. Think I don't know your perspective on it, it ended going way better than I expected to. Yeah. You know, did not that I had low expectations, I didn't know what to expect. And I think the receptiveness of the... And it might have been because of the preparation and understanding as much as I could the information beforehand so I wasn't like saying I don't know to a bunch of questions.

John 34:10

Yeah.

Kris 34:12

But things like that just like, yeah, cause lots of like stomachache along the way.

John 34:17

Oh, yeah.

Kris 34:17

And it's... When you're around like... I wouldn't say they're all like like friends and colleagues and people you trust like you shouldn't maybe be as worried about that. If they're believing what you're that is you're trying to achieve. But now like it's not that it's all finished and like oh like move on but at least at this point you know I don't know if I slept well that night

John 34:39

but way.

Kris 34:40

Yeah. I I ate way too late because I didn't even eat at dinner. But you know at least you know a lot of the anticipation was on and I say unneeded because maybe it helps get it to the right spot.

John 34:52

Yeah.

Kris 34:52

But it was way worse the thought than it was the actual thing that happened. I think that's probably true in a lot of cases of just like what I mean our community manager is gone this week.

John 35:05

Yeah. Oh yeah.

Kris 35:06

Out of town. Well and

John 35:07

I hadn't seen her but I hadn't thought about. Yeah.

Kris 35:10

And I was on Tuesday which again was the day after this meeting. Yeah. I was just like super frustrated because I was like I don't think things are ready. Like I I didn't think things were prepped and ready for having a week without our community manager. Hey. It's day two, and we're still here.

John 35:27

So Surviving.

Kris 35:28

Yeah. Coffee is made, and the door opens.

John 35:31

I did hear someone pumping it, and it sounded like it was empty right before I came down. Oh, no. And I was like, do I... Should I go tell Steven quick? Yeah. But I was, yeah. He may not have

Kris 35:40

the ear for that yet of like, you know, yeah, but things are, you know, with when we have, you know, in this case, like 40 of 44, 42, you know, of the 45 offices rented, that's obviously good seeing like that that when we went to other spaces before we opened like just for looking for examples like I think we're killing it way more like and a lot of these things and other people work them way longer to get there But because I think of all this, the the things that I see we still wanna do, it doesn't it doesn't always feel like we've made it as far as we want to. But I think comparatively, it's like way better than what should have maybe been even expected.

John 36:23

Yeah.

Kris 36:23

But I mean, I think I think from the beginning, I think, you know, you and the

John 36:27

original whatever you wanna call partners. Yeah. Guys were... I am extremely conservative. Mhmm. You guys were very aggressive on your timeline. And I remember that was my first reaction. I don't know if you remember that. The very first time you showed it to me. Yeah. I was just like

Kris 36:42

First or first, second, third, fourth.

John 36:44

Yeah. Exactly. I was like, I I don't see how this is possible. Yeah. You know? Like, it just it just because it was so aggressive on how fast you're gonna feel

Kris 36:52

like it's taking optimistic. Time. I think aggressive and probably too optimistic. I'm like

John 36:57

Yeah. But that... That's... I would classify optimism and aggressiveness as very similar Yeah. Things. You know? Like... Because, I mean, in general, I'm... I would consider myself an optimist. Some people might call me not. But I I would say in general, I believe things will be better.

Kris 37:11

Yeah. But but I would also say... I'm not saying you're not an optimist, but I but I think you

John 37:16

often

Kris 37:17

challenge things. Yes. So that's your first go to.

John 37:21

My go to is how is this gonna break?

Kris 37:23

Yeah.

John 37:23

And how

Kris 37:23

do we

John 37:23

avoid it breaking so that we can have the optimism that we want. You know? So I don't know what would

Kris 37:28

like like It's not pessimistic. It is that.

John 37:32

It's definitely not that.

Kris 37:33

It's a 100 questioning.

John 37:34

Yes. Yeah. My thing is, okay. I hear your optimism, but here are some reality things. Yeah. Here are the challenges I would see. Yep. I don't know if I'm right or not, but that's what I would guess. And I I definitely go to that first. I think that's just because I... Being a developer for twenty years, like, that's what you have to do. Like, you have to be like, okay. I'm gonna build this thing and I know I can build this thing and it's gonna be great but like how could it not be great so that it And do

Kris 38:00

you you think you do that for yourself? Because I would say it's for me to be like that for other people's stuff. Right?

John 38:07

Gotcha.

Kris 38:07

I think it's less easy for for me to be like that with my own in some circumstances. So like... And then maybe I'm gonna contradict myself in the next sentence. But but I I was thinking like when, you know, we had a couple events yesterday. Yeah. The same day. And I was concerned with the things that could go wrong and how to preemptively prevent those things from happening. Yeah. And I think we did a really good job of that. Like everyone was happy with both things that went on. There were no challenge or issues. And I kept not kept thinking but I was trying to with Steven like let's do this so that in case something happens like we don't have to worry about it. It's right there for us to like have access to or get or adjust if we need to. But I think... But I I address that differently than when it's like what we're just talking about now. Yeah. Maybe where it's my idea and I I can't I can't use that same thought on my own idea as easily as I can on like, other people's stuff.

John 39:08

Do you think it's like you don't you don't want to? Or Yeah. I don't know. Yeah. I don't know.

Kris 39:11

That's what I was curious about. If I don't wanna be critical of the thing

John 39:14

Like, you do... Because you want the thing to happen. Yeah. So... And if you get too critical, maybe you wouldn't do the thing. Or Yeah. I don't know. I I definitely feel like I rip on myself a lot. Like, I would say, you know, like, the inner critic on my ideas and my execution and things like that. What I've, you know, what what are they... They're just like all the self help people talk about, like, you know, like, oh, give yourself some grace. Yeah. You know? Like, stuff... So I've been... I've I've tried to do that. Yeah. I think a little bit more to just be like, look. What do I like to do? I love to build stuff.

Kris 39:45

Mhmm.

John 39:45

I love to help people. I love to tell people about the stuff that I built that helps them. Mhmm. So, like, those are the three things I really like. All the other things, like, are things that I could be good at if I spent time on. I believe that. Yeah. Like, optimistically, I'm like, I could be good at sales. Like, I can schmooze with people and, like... But but, realistically, like, when I'm on a sales call with somebody for, like, flipper or something like that and they, like, ask me about a thing, I'm like, no. We don't have that thing. Like, whereas, you know, like, I feel like Robbie, he'll be like, oh, well, you can do it this way. And it's like, when I would look at it as a developer, I'd be like, but that's super awkward. Yeah. But he still has solved the problem for the customer and they're happy. And Yeah. Because they don't... They haven't thought, could there even be a better way than that? You know? So I've actually tried to take that a little bit on on some calls that I've had in the past, you know, like, month month or two and be like... You know, someone's like, oh, I'd love to have, like, multiple instances of Flipper in the same app and do these different things.

And I was like, well, we can't do that today. And then I was like, can we do this today, Claude? And Claude was like, well, you could do it this way. And I was like, that's what I was thinking.

Kris 40:44

Yeah.

John 40:45

And then I was like, well, you could do it this way. And they're like, oh, well, they... That'll work. And, like, I was like, but I've got some other ideas, and then I can kinda kick back and forth. And so at least they're they're unblocked, and they can start using us, they can pay us. And then we can make the experience better. And that's almost a... That's a better thing than being like, no.

Kris 41:01

Right. Right.

John 41:01

You So, yeah, definitely, I mean, I think what I've tried to do lately is be like, look. These are the things I'm good at. There's lots of other things that I wish I would spend time on, you know like email a single person, you know about like you know would you like to use my

Kris 41:18

product,

John 41:18

know, or like any kind of sales?

Kris 41:20

I wanna try that over the next like couple weeks of of doing that because I think as we talk about certain aspects of growing, whatever the thing is. It could be anything of, know, Alex Shermosey says if you spent an hour whatever ninety minutes just going versus like one after one at a time. Like if if we... If I emailed all the people that I think I know who might be a good candidate for renting space or Yeah. Being co working member, whatever it and just brute force it and didn't let anyone interrupt me. Mhmm. It has to... Like it has to either work or learn from it of... Like one of those things is gonna happen.

John 42:04

Agreed. Which means that at some point it will work. Right. Like if you keep doing it. Right. And I think that's the difference between people who grow like massive businesses or who like like get to... Not that I... I don't even want to do that. I don't want a massive business. Yeah. But like that seems more like a headache. But I think those people who hit those targets and stuff along the way or grow really quickly that you hear all the stories about, I think that's their superpower is, like, they either like that or they just... They are able to, like, get in that mode and just do it. And I'm like, I don't know why I can't. I guess because I'm comfortable and that makes me uncomfortable and then I I don't really wanna be uncomfortable but I'll have to, I you

Kris 42:38

think I avoid sometimes because I'm I'm worried about them managing the the the response back because I don't have a good system to Oh, yeah. You know, track that. Yeah. Which I know there's a thousand CRM systems out there that are built to do that.

John 42:53

That goes back to like the responding

Kris 42:58

immediately

John 42:59

which is not always possible when you're in meetings and things like that but that definitely makes a big difference too is, like, the faster you respond and stuff like that. But, yeah, I I feel like, overall, though, like, you know, aggressive, optimistic, you know,

Kris 43:13

pessimistic,

John 43:13

not aggressive, whatever, however you wanna look at it. Like, optimism is the right choice. Mhmm. Because, like, in general, if you believe something bad's gonna happen, it's probably gonna happen. And if you believe something good's gonna happen, maybe it will happen, maybe it won't. But it's way better to, like, believe in the future being better or something working out than, you know, something not.

Kris 43:33

Yeah.

John 43:33

Like Yeah. Yeah. Because... Yeah. It's just... I don't I don't... You're not gonna get anywhere if you don't believe that in optimism. But I also feel like the best thing you can do to increase your chances of the optimistic thing happening are actually figuring out what can go... Like, reducing the risk of failure. Like, reduce the things. That's what we do as entrepreneurs or anything. Like, I'm just accepting that word now. I don't wanna say it, but I hate it every time. I feel gross. I like in in my mouth a little, but but like as an entrepreneur, all you're trying to do is just reduce risk. And if you can reduce the risk, the chances of something failing, you can reduce the downside, and even if you can't increase the upside, you have a higher chance of success. And so that's why I would go to that first, the the challenges, is because I'm like, okay. What are the things that can make... Like, I hear what you're saying. That sounds awesome. What are the things that could maybe make it not work? Mhmm. And so I go to those first because I'm like, once we get past those, then it's just all upside, you know?

Kris 44:31

That's that's

John 44:32

what I

Kris 44:32

would

John 44:32

probably tell myself.

Kris 44:33

That's why. But I don't know if it's accurate. Do have a recommendation or Mhmm. Any recommendation for... I would say when we talked about it and we talked about it a number of episodes, but my thought and recommendation is if you haven't used like some of the G Stack tools to ask it how to install it for you. Yeah. And then just use those because don't don't don't play in hard mode. Yeah. Don't play like you use the tools that are there to, like, help move things faster.

John 45:04

Yeah. I'm glad... I'm also really glad to hear that they are helpful because like

Kris 45:07

Yeah.

John 45:07

I I found them helpful, but there... There's... For me, there's almost so many bumpers on it that once in a while, I'm just like, I can't do this anymore. Like, I just gotta go back to,

Kris 45:15

like Yeah. It asks a lot of questions.

John 45:16

It does.

Kris 45:17

Which sometimes can be Yep. Like, just move on on the thing.

John 45:20

But Yeah.

Kris 45:21

But I

John 45:22

think for people, I... What I've noticed is for people who, like, who haven't built for twenty years, having the bumpers is really helpful.

Kris 45:31

Mhmm.

John 45:31

And so I think it actually makes it possible. It it puts you in the mindset Mhmm. Of the CEO, the engineer, the designer, like, of all those things, and I think you actually end up with a really good result. Mhmm. It's just a lot slower. Mhmm. And sometimes the slowness drives me insane. But, like, generally, though, I think it's a great tool. I I really think it's fantastic. I... I'm impressed with it. I actually submitted my first PR to it. Nice. Like, in the last week or two. I... They... The Codex review doesn't have like... You can't pick the effort level. Okay. And I'm real picky about effort levels because if you go too high, it gets insane. So I submitted a PR basically to Yeah. To do that. That was also a kind of fascinating process as well. But Nice. But, no. I I would say my only recommendation is honestly, like, and maybe I'm gonna do this for next time. I'm just gonna, like, go whatever hormozy or whatever I'm gonna do. I don't know if I'll do an hour because that's a long time but just do one thing.

Kris 46:29

Yeah yeah okay for next time. We're gonna both try to do a thing that we can report on and see. Won't We probably have the results by then. Okay.

John 46:36

But

Kris 46:37

that's gonna be it. Okay.

John 46:38

I love it.

Kris 46:39

If you

John 46:40

haven't,

Kris 46:41

leave a comment, like, subscribe, helps the algorithms. And we'll catch you next time.

Download Transcript

Your Hosts

John Nunemaker

John Nunemaker

When not playing basketball or tinkering with watches, John works at Box Out Sports, Flipper Cloud and Fireside.fm as a programmer/owner.

Share This Episode