Who Does the Marketing?

Morning.

Hey good morning Ed. How are you?

um doing great.

Alright, we we have a bunch of stuff to dig in and we started ten minutes late. Or no, half an hour late. Um my bad. So so what's Yeah yeah, what's what's going

Yeah. Yeah. Great day to start late.

on with scalpels?

Uh yeah. I mean nothing today, unfortunately. Um they're still on track to launch on Saturday. I mean, I don't know. I just picked the date out and didn't look at where it fell on the calendar. Um Yeah. Well, it's

You gotta pick a date, but but not a Saturday though, and not a Friday either.

it just felt right. Like I'm gonna launch on Halloween. It'll be a cool story. Didn't realize that Halloween was a Saturday.

Ha ha ha ha

You gotta pick

a Tuesday.

Yeah, well I'll probably still be launching on Tuesday. I mean this this is how it goes. It's not like push the button and say, it's done. That's when the you

Mm-hmm.

know that's the day that I get serious, I suppose. Um but yeah, I mean with all this stuff

Yeah, yeah, yeah. But s uh by the way, before

we d dig in, is that a is that a new mic? Did you have that mic before?

yeah, no, this is um I'm in my office. So usually on Thursdays I'll work from home. And so

Right.

when we do this, usually I'm at my dining room table. Um but I have an office up on the nineteenth floor of a tower in Basel overlooking the city. Um right

Very very fancy.

next to the train station. It's really a uh I my little nook here doesn't have a nice view.

Um, but I have there's space upstairs where I can sit and just look out over the city. It's a it's a great place to be creative. Uh it's a nice place to have meetings when I do occasionally have meetings with local clients. You know, um it's a it's a great environment for that. And um

Yeah, yeah, nice, nice.

and it's just it's so much better for me to I try to be here about three days a week because

Mm.

I find it so much easier to work in a in an office than at home because

you know, when I'm at home there's laundry, there's, you know, dinner to prep,

Kids.

you know, yeah, cleanup from the kids, like all that stuff. I I get into the office, all that's just out of my way. And uh so yeah, it's kinda like chaos free place. And I came here

Yeah. Yeah. Yeah. I should really put

put some dishes in the in the dishwasher, right?

Yeah. Yeah. So I'm here today because it's right next to the train station. And as soon as we get done recording, I need to jump on a train. I have a meeting this afternoon in Zurich with a prospective client. Then I have to find somewhere quiet to have a big project kickoff call with a client in Chicago. And then from there I'm gonna scramble to Zoog, which is a little town just outside of Zurich.

Uh-huh.

And

uh I'll be speaking at the Laravel Switzerland meetup this evening. Yeah.

Exciting. What are you gonna talk about?

Um it just experiments. I'm just gonna basically walk through a lot of the experimental stuff that led to scalpels that led to the products. So I'll go through the you know, accidentally discovering the Go could be used as a universal runtime on Laravel Cloud. Um, I'm gonna dig deep with them on my philosophy uh that we should always build everything twice. Um

Mm-hmm. Yeah,

That you just, you know, yeah. And just kind of the

yeah, at least. Yeah, yeah, yeah.

joy of experimentation. And I have a kind of a fun project that I'm gonna I'll I'll give you a little preview of it because nobody will hear this before then anyway. Um, as part of my not really marketing for scalpels, but just sort of uh you know, in in order for us to really grow, uh, you know, as an industry, the one of the big opportunities I think is that a lot of kind of people who have never considered

software as a solution to their problems can now

Mm.

or should now start thinking about software as a solution to their problems. Um so I I have this sort of uh an attempt to do some sort of AI for normies uh type thing. And really what it's going to be is just uh I'm gonna show off just fun things that have no profit motive. There's no commercial value and it's just fun things that you can do with AI and um kind of create like a directory of

you know, people who these are a lot of people out there that are doing kind of fun things with AI that have nothing to do with shareholder value. And

Yeah.

um and just kind of uh do some resources for people who are, you know, like my mother. Um, you know, she's she's working on this um this big family tree project. And um yeah, so there are there are some things that I could

there you go. Just pull in pull in Claud.

teach her to do using AI that could make sort of this

You know, sh you know, she grew up when she grew up, so it she did the really nice, you know, family tree. Um, like you would expect.

Mm-hmm, mm-hmm.

I'm like, well, you know, we can turn this into something a little bit more interactive and turn it into a digital family tree. And, you know, let me show you how we can do that. And of course, this is not

Yeah.

something that anybody would want to ever productize, but um, you know, if you're looking for, you know, interesting ways to engage with technology in a fun way that doesn't really have any uh

you know, profit value, I think that's that's probably a good thing. Because the fact is what whatever we think about AI, and it's not it's not a perfect technology, right? It does have it has uh serious economic implications, social implications, environmental implications. Um but it it yeah, it is it is perfect. But I mean it's

It's a perfect technology ad. It's perfect. Cloud5 Opus five point five is perfect. What are you talking about? It is perfect.

it is kind of the reality. And to the extent that it could be added, it can be used to enhance joy

and give people who have maybe some you know, this one of my frustrations with the the whole AI art war, I don't like that it stole art from, you know, starving artists. Like there are ethical problems with it. But but my problem with the whole, you know, you're not an artist if you're using AI. Um, I think people

It's silly.

I think a lot of people have creative ideas, but they don't have the manual dexterity or the training or the knowledge to actually make those things come to life.

And to the extent that people are using it to avoid paying artists, that's a problem. But when you have, you know, a mom who wants to be able to create, you know, coloring book pages for her kid, um, you know, I don't fundamentally think there's anything wrong with that. And I think that that is really, you know, that sort of application, it really does sort of enhance life. And so if we're going

Mm-hmm.

to be stuck in a world where we're going to completely disrupt the economy and threaten all kinds of jobs and all this other stuff, we

Mm-hmm.

might as well get some

just sort of useful, fun benefit out of it. Um, so yeah, that's another one of the experimental things I'm doing.

Yeah, yeah, y y you know, you know what

I did a few weeks ago? So my mom turned eighty-five. So I was like, great, um, you know, Opus five point five. Here's a folder of old pictures, make a kinda like pop-up book video from it, because my mom's turning eighty five and

Mm-hmm.

I just went for it. It was really good and I sent it to my mom and she loved it. And that was it. Like it

Yeah, yeah. And that's the sort of thing

it it did not take me a long time, you know? But it was nice.

That's the sort of thing that we

know how to do and we can teach people how to do and it's like, you know, we're we're all using this to make a bunch of money and to move really fast and um, you know, your your life doesn't really

Well make a bunch of money,

I don't know, you know, like

I'm making a lot of money for my clients. So, you know, that

Yeah yeah yeah yeah yeah.

that works. Um, but yeah, I I think that uh, you know, it this is gonna be kind of part of everybody's life going forward. And um, you know, while the major labs are focusing on, you know, shareholder value, um, you know, people are kind of generally

Of course.

frustrated and scared and, you know, disgusted by AI for s sometimes for very good reasons. and I just kinda wanna show the other side. Like it's yeah, it it in a lot of ways it sucks, but it's actually a fun tool. We can do some

some cool, interesting things with it. And as long as we have to live with it, we might as well get some enjoyment out of it.

Yeah, yeah, yeah, yeah. So Segway, perfect. Um, make a lot of money. Scalpels.

Yeah, so um I uh since I'm actually launching I have to

Is launching launching this week apparently.

yes, since I'm doing that, I have to actually take the whole monetization plan from WIP to something that's real, leaving it pretty much intact as it is. Um I have added a couple of additional paid benefits. Um so I've been working with how to create um

You know, 'cause the the lab is just sort of like I have AI just throw out my ideas, then we do experiments,

Yeah.

and then I have the AI write at the experiment. And uh sometimes it becomes becomes a product, sometimes it doesn't, and I get on and I stream and I talk about it and you know, that sort of thing. But it's all very like unplanned, impromptu. And so a lot of the things that I don't turn into products, I do actually end up harvesting the the results of those experiments and turning it to things that I do for my clients.

Um so I

Right.

did a lot of experimentation around test segmentation. Um, you know, figuring out in a giant slow bloated test suite what tests should run every time CI runs. The the tests that give you real confidence that you didn't break anything. And then there's this whole other

Yeah, yeah.

big bucket of tests that's tests that are good to have, but I don't need to spend 20 minutes waiting for CI to pass. So maybe we just run these every night. uh Yeah, and and then a bunch of tests we can delete and then

Or even even a bunch of tests we can probably delete. Like like is are you making that a product?

And then tombstoning tests. Well, I mean, I don't know. I it's kind of hard to it's kind of hard to productize it. I've been I've been tinkering

Or o or a thing? Something? Yeah.

with trying to make some deterministic scripting to to kind of make it easier. Uh, but a lot of times it's sort of a judgment call and it comes down to, you know, your own personal risk tolerance, like how many tests do you feel comfortable deleting, you know, because all point to a test suite, right? Just to give you confidence that your application continues to work after you've made changes.

And so how far can you pare things down before you start losing that confidence? And I think that answer is going to be different for everybody. So I don't really

Yeah.

want to make that decision by bundling it into a product. But what I can do is I can create like real planned, scripted, thought-through educational content around that. And so

Yes.

um there will be that sort of content will be in Scalpels. And for something that there's going to be like

The CI optimization course that I'm it's gonna be like a little short, like six tutorial course. The first three will cover the basics,

Uh-huh.

they'll be free and available to everybody. And then if you're paying it in some way, you'll get the the next three. And the next three will come with um, you know, actual scripts that will help you to do like tombstoning of code and tests. And uh we'll kind of dig a little deeper into how I actually do this segmentation for my clients. Cause now I've done this for several

several client projects and gone from you know 40 minutes in CI to one minute in CI and I did it without

That's your headline right there.

yeah and I I did it without really I mean there was a small reduction in confidence right there's there are definitely things that are not being tested in CI that could break.

Mm-hmm.

Um but then when you combine that with the fact that I have all of these sites being monitored with hone and wake and so I'm seeing their logs, I'm seeing their exceptions.

I feel like I can l loosen my grip a little bit in CI, knowing that I can react as quickly as I can if something goes wrong on the other side. And the things and the places

Yeah, yeah, yeah.

where I have kind of loosened that test coverage are areas of the app

Mm-hmm.

where if something breaks, it'll throw an exception. It might annoy a user, but it's not going to touch money, it's not going to touch privacy, and it's not going to touch data integrity. It's going to it's going to do things

Yeah, yeah.

like, you know, I got a 500 where I try to click this button. Fine, whatever. Like come back in 10 minutes, that'll be fixed.

Yeah. That

So you know, just

makes a lot of sense. Can I just say scalpels.app? I should we should say that scalpels.app. Um I like the new H one.

Yeah, scalpels.app. It's very good.

AI, it's very good. AI made building software easy and owning software hard. That's good. I like it. That's

Yeah.

true.

Yeah. Well I figured that w I finally figured out, you 'cause we've been talking a lot about my struggles with the front page because I do kind of serve

Yeah.

multiple masters. I have like the the end user that I'm giving all the stuff to for free. And then I have the the freelancers and the agencies that I want to be able to connect with those end users in order to help

Yeah.

um connect, you know, companies that are trying to adopt AI with experts who really know how to use it. Um,

Mm.

you know, and that's really fundamentally my my uh

monetization plan, I have to drive a

Yeah.

lot of those end users in to make it valuable for the freelancers and agencies. I have a pretty good marketing plan for that, I think. Um, you know, I'm just kind of broadening the the portfolio of tools. Um and it's a lot of it comes down to I've identified where people are saying, I have this problem, I'm frustrated with this, you know, my there's not a SaaS that really solves this well for me, or I have to pay way too much money. And I'm just gonna advertise

these things there. And it's like here here's a free alternative to thing you hate paying for. And people will click, they will install it, you know, 'cause this there's no reason not to, it's free. So um you know, so I I don't have any

Well, there's a whole

bunch of reasons, you know, but I mean,

Yeah.

you know, like being free is not why people do things, right? But but yeah, yeah, I I know what you mean. Uhhuh.

No, it's well, it's free, but unlike so the

the big difference the big difference between like how I'm doing free and how most people do free is I'm trying to reduce all the friction from free. You know, I kind of get the um

Yeah.

you know, a lot of a lot of SAS that has like you know, we have an open source host it yourself project and you know, we'll we'll put it into a Docker container for you and good luck, you know, Godspeed.

you know, and I'm taking it and like, okay, we're gonna put it on really good managed hosting. You install it with one click, everything's properly provisioned, and I get to lean, I get all kinds of great stuff for free, right? Because Laravel did such an amazing job with that cloud dashboard that once it's

Yeah.

there, I just tell users, hey, you need to tweak something, you need a little bit more compute, you need to, you know, increase the number of of uh Q workers that are available, just jump into cloud and do it there. I don't have to build any of that UI because there's no way I could do it better than they did anyway.

So

Yeah, yeah. No, I I get it. Um

So Scalpels is launching.

Yes.

What's the plan? On Saturday, is that what it is? Uh-huh. Alright,

Uh yeah, so the plan is yeah, the plan is Saturday. Um

it's Thursday.

Yeah, I mean I think the the big plan is that well, I think I forgot to feature flags the things that should have been feature flagged. So the there are some things that should be feature flagged and not shown in the app. That th those feature flags will go away on Saturday. Um but I'm actually

Okay.

not sure that they're feature flagged and they're probably appearing if you were to log in now, which isn't a big deal. Um so everything will be fully functional.

on Saturday and um

I'm I'm creat creating an account right now.

Yeah. You're creating an account.

Yeah, you're gonna be you're gonna hit the still sort of friction UI. first thing you're gonna have to do before you can do anything else is put in your Laravel Cloud token and connect your GitHub. Because those are two things I need before I can install any apps for you. But now that Scalpus is more than just the apps, I need to think through that whole onboarding flow and that's what I've I've been working on. Um you know, fun challenge. I built this whole thing with deploy to Laravel Cloud in mind.

And so I never bothered to

Yeah, yeah, yeah.

create any good, easy way for me to, you know, provision hone locally from scalpels to be able to like exercise that whole thing in a local environment. So now when I find any any quirks in the UI, I I find it in production. I to go back to my agent and say, I'm seeing, you know, like screenshot production, say, fix this, push it back up so I can see it again. So I

Yeah.

fix that. Now I have uh two other

install paths that I've created. One is a local install path so that when I'm running locally I can just provision any of these apps to my local environment, which gives me the ability to iterate very quickly on UI, which I need to do because it it is still pretty rough. Um and Sarah, my friend in person, has um she's just finishing up some some other work, so she'll get on it. There so that I mean that'll still be getting a paint job,

Uh-huh, uhhuh.

you know, after launch. Uh and then I'm introducing Laravel Forge as a uh

as a place that you can install to.

But but leave that after Saturday, right? Don't do it don't do it all at once.

Yeah, well, I mean that's those pieces

are done. You th you think that like adding you know, the ability to launch to Forge is like an outfit worthy and I should save that for

Yeah, yeah. Save it. Do it like like like little you know, then you have something to talk about in a week.

Yeah, well so the reason that I'm doing it is because um I we need for us, primarily Len on his side, needs uh social scheduling because he does the this week in PHP internals and he has this B2C SAS that has to pump out

Yeah, yeah.

content constantly. So he needs social scheduling. And there's this open source product called Postive, I think it is. Um yeah.

You mentioned it last time. Yeah, yeah. Okay. And that's

that's an app you can now deploy on Laravel Cloud?

You can

Yeah, so I don't have it in the Scalpels catalog yet, but the it's built, so I can add it to the catalog whenever. The problem is that the floor for that is like sixty dollars a month. Because of the resources that it requires, you would not be able to run that for less than sixty dollars a month on cloud. You could run it f you could run it for like Yeah. Well, I mean in this case in this case it's sixty

That seems like an issue. 'Cause I might as well pay somebody uh you know to some degree, right? Yeah, yeah.

a month, it would still be worth it for us because

what we would need from a paid service. you know, we can get things from that sixty a month that are, you know, more than what we would get for a paid service of a hundred a month. So

pff f fair enough, but you would

still also have to manually go into like Facebook, create an app into and do do a bunch of work yourself.

Yeah. Right. Yeah. I

mean it's there's there's definitely work. this is one of those like

Yeah. Yeah.

um it's it's not for everybody. I think a lot of the stuff that I'm doing here in Scalpels is not for everybody. And in fact, part of the the mission at Scalpels is to Well me scalpels is for everybody, but individual things are not necessarily for everybody. You know, it's because I it's

Scalpels is for everybody. It's for i it's it's very good. Scalpels.app. Right. Fair enough.

because I do want to be really honest about the trade-offs between you know, choosing a scalpels product versus

choosing the paid product. And and every one of the product pages does actually it has a section of why you should keep paying for the product that you're paying for. Because they do come with advantages. So the biggest one obviously is that somebody else is managing it for you and maintaining it. You don't have to think about any of that stuff. I try to make that as easy as possible by putting you onto Larvel Cloud where they take care

Mm-hmm.

of all of the infrastructure stuff. So don't have to think about that at all. Um and I have the opt-in automatic update. So I'm thinking about the code. So you don't have to think about that at all.

But there's still a piece

Yeah, yeah.

of you own it. So like if it breaks, especially if you're, you know, not paying us for any sort of a support tier, if it breaks, you're on your own to figure it out. I mean, you can obviously, you know, open an issue and we'll do our best. But um, you know, it's it's a different thing when you're dealing with open source versus when you're dealing with uh, you know, somebody you're paying a hundred bucks a month to. You're gonna get better support f you know, from them or more timely support. Um and this is one of those this post-it is one of those things where

Uh part of it I wanted to see because of the it's sort of an institute tech stack, it's it has three different databases. It uses temporal as a long running process, which, you know, by default isn't available on cloud. so it's one of those like, can I actually take this thing and shape it into an installable bundle that can run on cloud? And I did it and it worked. I'm like, we'll just use this for us. But if I'm gonna add it to the catalog,

Yeah, yeah.

you know, obviously first I have to tell people like the floor for this is sixty dollars a month. It's gonna cost you a lot of money if you do it this way.

But by adding uh Forge as an install target now, you can probably do this on like a six dollar a month uh VPS with Forge. I mean I haven't actually I haven't actually installed on Forge on Forge to see what that floor is there, but you know, um, you know, I have this the the basically part of my process is that I I install with the absolute bare minimum requirements, see what blows up, and then kind of increase till I get to where things will run.

reliably with real production workloads so I can load test everything at these various various levels. That's how I come up with the floor calculation for all these things. So

Mm-hmm.

I'm gonna do that on Forge and see where it lands, but it's probably you know six to ten bucks or something a month. And then it becomes a much more reasonable thing, right? To you know then you can host it yourself.

Then it's much more reasonable. Yeah, yeah.

For sure. Okay, so so launching on Saturday. So you g we we got the videos coming out, right?

I do, yeah. I have videos. Um I'll there are a lot of videos that are just the um kind of like the product videos that you've seen. Uh it's that same voice and the same animation and stuff like that. Um

Uh-huh, yeah, yeah. Yeah.

and yeah, so I'm doing I'm doing a bunch of videos around not necessarily products, but um to go with the various landing pages that I'm launching. And we're gonna kick off some advertising.

So you're gonna be like like online

on on your laptop on Saturday, like doing stuff? Not even? Yeah,

No, it's Halloween. I have kids. No.

I I that's what I was thinking. Like, you know, if it's launch day, surely you have to be there, no? No, you just could get it all scheduled, alright? Alright, alright, alright.

No, I'm just I'm just hitting I'm hitting start on some ad campaigns. I'm not, you know.

Yeah, I mean

Um

I I don't know. I mean I'll probably, you know, Monday, Tuesday, Wednesday, I'll probably do like a some sort of live something, you know, maybe do some office hours for

Oh yeah.

you know, I might open up Riverside and have it so the people could just join and say, Hey, you know, come in, install this thing, play with it, see what you think, ask me questions live.

you know, that's sort of put myself

Yeah. Yeah.

in the line of fire type content. Um but

Yeah. Can I can I give you

like uh concise feedback 'cause meanwhile I signed up for an account.

Sure.

So first thing I tried to install was Wake, but then it told me now I have to pay you.

Oh yeah, that's because Wake is

It's

oh Wake shouldn't be experimental. Um so yeah, but the yeah, so the experimental

okay. So th so that will be fixed, right? Um

ones I make people pay me for and th I've I think I've explained the reason in the past, because when when I build these things they go up on day one. Like literally they it finishes building, it goes up.

Fair enough, but it doesn't say so.

Like it should say so in the thing and all like it 'cause

Yeah.

it's like and then and then I go let me see if I can do hone. So hone it lets me do. Okay, cool. Um

Yeah, you can do hone. You can do hone. You would

you would because you already have it running in cloud, you'd probably run into an error because it already exists. Um because you installed it the other way. But offline I can

Yeah, but it's just a new

box, right? Like what what does you know, what does home care if it runs

No, it'll it'll try to it's because

it's it's because of the way it um it names the application. I don't give you and I probably should, but I don't give the ability to name the application. Uh I choose the name of the application.

Y

you should because me as a as a Larvel Cloud user I would like that.

I mean that would be I I would at least wanna see a name pre filled and then I can change it.

Yeah, so my first iteration of this I did. I had like you could you could set the name, you could choose like it had dropdowns for all the different sizes of all the different resources and things like that. And it it yeah.

I don't need the sizes so much but but just the name

the name matters, you know, 'cause it's like how I organize my things.

Hmm. Yeah, I don't. I never that's the thing because I've gotten so good with the API and the CLI. I don't ever go into the the cloud dashboard anymore. So for me it's like, you know, not that big a deal. So okay, I'll add

Right. Right. All right.

I'll add the ability to to add a name and I will put like locks or something over the um experimental ones for free users so that they can see that those are not available to them.

Wake, I

Yeah.

feel good. I have so I like to um use these things myself and have at least a couple of clients using them before I say, Yeah, it's good enough that I can just make

Mm.

it free. 'Cause I don't um you know, if somebody's and again, just to reiterate, I think I've said this before here, but if if somebody's paying me something, I don't mind like jumping a call with them and, you know, helping walk them through something or whatever. So these experimental ones that nobody's really used for a a real workload,

you know, I don't want to give those away for free because then I have just everybody installing them and I get flooded with people who have ini you know, those initial issues. But um, you

Yeah.

know, anybody who's paying me, yeah. Schedule a call and we'll get together and and work through the the quirks.

Yeah, yeah. It it has to show it because

right now like half of them seem to be experimental. I have to click on each to find out. Um

Yeah. I think

yeah, so it it says once the uh once the dialogue opens where where you would otherwise be able to install it. So I just need to put some sort of a lock icon or something.

Yeah, yeah, yeah, yeah.

Um

Um, all right, cool. So Saturday Saturday launch. Um

Let I'm very curious to see to see kind of the traffic and the results and how next week goes. Um

Yeah.

shall we talk sh I I have another topic. Shall we dig into that?

Sure, yeah.

Okay, so I'm gonna brainstorm uh like brain dump what the topic is and then let's talk about it. So the topic is this. Um so we are all like you know, us and many people are trying to build businesses in this new AI world. Um

So we're building products. Uh I think a lot of us are now uh solo founders, basically. And so, you know, some people are very good at marketing, some people are very good at tech, some people are very good at whatever, architecture, you know, product or or w whatever it is. Um, it's it's hard to

Have so many functions, even though we have AI, it's hard to have so many functions and priorities into in one person's head. And so when you and I talk about this, the last, you know, we used to talk, I think, more about tech, and now we talk more about marketing because you know it's kind of like that phase,

Mm-hmm.

right? But we cannot all be fantastic marketers, we cannot all be fantastic.

product builders, I don't know, like like that it that is the topic.

Yeah.

Even though we have AI and sorry, I'll add one more thing. And from my perspective, I'm trying to automate as many things as I can in loops and I think you're probably doing similar things.

Yeah, I mean I'm I'm not as far as marketing goes, except to the extent that I'm getting, um, you know, I think I've told you before, we do have somebody who does some marketing for us. And um

Yeah.

when she has a little bit of time, which should start this coming, you know, in in the next week, a little more time to pay attention to what I'm doing. Um she so basically she like will create lists that here are the things that you should be doing to take this thing to market, and then I'll do those things, but then I don't have any real creative like um

I don't have a good instinct for it. So if I don't have somebody telling me, you know, you need to create this kind of content, be posting this thing here, you know, be planning ads to run here and here, like I would just be lost. So um yeah, I'm not I'm not really doing much automation because of like I think I've said before, like I don't feel good about the idea of trying to automate something that I'm not good at doing manually.

Yeah.

Um so my my hope is that we can get to um and I've various

paths potentially to get us there, uh to where we can ramp up the um our our commitment to paying actual marketers to help us do this stuff because it's not my wheelhouse. And I I've long thought, you know, kind of to your point, there are just so many different functions that go around launching something or go around running a small business or a SaaS or anything. Um

Having a business, yeah.

yeah, that uh

you know, y from the outside you might look at and say, well now we have AI, so AI can just be that generalist to allow me to take the one thing I'm really, really good at and it can fill in all those other gaps for me. And I'm just not a hundred percent sure that it's quite there yet. And so it'll be interesting to see to the you know, to what extent you find that you could be successful automating things that aren't necessarily in your wheelhouse. Because maybe I'm wrong. Uh maybe it is further along than I think.

I think

I I think there's okay, so there's 'cause there's a bunch of interesting things that things there. So one thing is, for example, um if you can simplify things, it's much easier to make them automatic. And so the very simple example is can I simplify my my bookkeeping, the way I get paid, the way I write, etc.,

Mm-hmm.

so that that thing just runs and I literally have to send

files to my accountant once a year and they file my taxes and I'm done, right? Or

Right.

or can I pay a service of 50 bucks a month like to keep track of that and then you know I have Stripe, etc. So that's an easy example, and I think a lot of people like get that set up in a very efficient way. Like it's simple and efficient.

Yeah, but that that makes sense because

but that's that's sort of operational and doesn't require any creativity. That doesn't you don't need an instinct for that. You just need to know that I need to get I don't

Well, y uh don't don't say that to an accountant. Everything everything requires

you know, engineering requires creativity. You are in one of the most creative probably maybe one of your most creative periods in in maybe of your life. I don't know, I I haven't known you that long, but you are in a very creative period, right? Marketing is creative, everything is creative.

Yeah, no, I mean fair enough, but I it is

it is creative, but I think, you know, uh to some extent though operations is pretty dry. Like upgrading servers,

Yeah.

like that's an operational thing. You can write a script that does those those uh those things automatically. Bookkeeping.

Yes, yes, yes. Exactly. So Laravel Cloud made

the whole server stuff automated. If

Right.

we can automate bookkeeping, great. Like or or minimize so we so basically you have to think about it thirty minutes every year, that's it, right?

Yeah, but I think

there's a okay, so I'll I'll say it this way. I think there's already an established playbook for that. Like you you

Yeah.

need to be able to collect all of your incoming and outgoing money records, categorize

Yeah.

them correctly, and put them into a format that you can hand off to your CPA.

Um, like we we kind of know what that has to look like. So we can kind of play around with the tooling and the automation to make that happen because we have a full understanding, even if I in my head don't have a full understanding of that process, that process is really well documented somewhere. So there's a clear path to success. Like if you do this

Yeah.

with your receipts and your invoices, and you then

then at the end of the year you will have everything that you need for your accountant. With marketing, it's little bit different because it's not really necessarily a clear path to success because it depends on your, you know, who your ideal customer is, it depends on what

Yeah. Th there were like

fifty playbooks and everybody every company tries to figure out their playbook and it takes four and it never settles. Yeah.

R right. So so with accounting, because there's kind of one playbook, you can

automate that. Um with marketing, you could automate one. I'd have I'm sure I could go to like Justin Jackson's Marketing for Developers course. I could take that and I could find every piece of what he advises to do in there and I can write automations to make those things as simple as possible.

Um, but is Justin Jackson's approach in marketing for developers gonna be the thing that actually will get scalpels off the ground and making real money? Maybe, maybe not. I

We don't know.

don't know. I don't know enough about marketing to know whether or not that particular playbook is the playbook that I should be automating around. Um so yeah.

Yeah. Yeah. Yeah. And and even a marketer

would not know until they try it, right? Like

Well,

that's kind of that's sort of my instinct on it is that uh,

Yeah.

you know, a good marketer is going to kinda do what I do when I'm coding, right? It's a lot of like I throw some stuff at the wall and I see what sticks and you know, then I I extract that knowledge and I I refine it and I turn it into something better. And I know how to do that with code. I don't know how to do that with marketing, advertising,

So

that sort of thing.

I agree. So let me ask you question. So as people so we are building businesses, I think, you know, you and I are probably both better at technology and product than we are at marketing. Um

But how do we find you know what I'm saying? Like it's easy to say I'll just hire a marketer to do this for us once we get a bit of revenue, but but but that doesn't mean that is the right marketer that's gonna you know what I mean? Like that's not that seems very, very hard to me. To find somebody who can actually push stuff the

Yeah.

right way.

Well, I mean that that's true, but it's it's kind of the same thing as hiring a developer, right? You just kinda have to look at their track record and you know, have they been successful in the past and or have they, you know, just stumbled and stumbled and stumbled. And it's kind of that's a little bit hard to gauge, I think. Um, in in both cases. And so to some extent it is just sort of luck, but I do feel like throwing some money at the right person is going to give me better results than me trying to do it myself.

Yeah, yeah, I do agree with that. Um I just I just do not know how to identify the right person because it's the right person for this very particular business and business model that each product will be different. Or for or even you could say for this ICP.

Yeah, well, but I don't think that there's necessarily a person to ICP match because I mean, especially now

Maybe not.

especially now that we're in the age of AI, so let's let's look at it as let's let's go back to that sort of uh the marketer to developer analogy, right? Two

Yeah.

years ago, if you wanted to do you know, a project in Rust, I would not be the person that you would want to talk to.

Yeah, me neither.

Today I might be the person you'd want to talk to because

Yeah.

I have launched, you know, for myself. I've launched some Rust projects at work, and I've launched some successful stuff for clients. Like I know how to do this stuff now. Um,

Yeah, yeah.

you know, because now I have this this tooling that helps me to do it a lot better. But it's my the fundamental knowledge that I had from years of PHP and Perl and JavaScript and C that allowed me to um to once I had the tools to take kind of take that like.

that learning curve and wipe it away, you know, allows me to kind of become now totally stack agnostic. And it might take me a few extra iterations to get something right in Rust versus what it would take me in PHP, but I'll get there and it'll be a great product. And I think maybe it's the same thing with marketing. If you have somebody who has that those years of experience and the, you know, all the successes and failures that come with actually trying to do marketing and succeeding sometimes or failing sometimes. Um,

And then you give them these tools so that they can iterate faster. Like they're they're gonna look at it and say, My instinct says this is your ICP, this is who we're gonna go after, these are the channels we're gonna use. And then, you know, very quickly respond to what's working, what didn't work, pivot, shift, whatever. Um,

Yeah, yeah.

but with with all of those years of experience, they can do that in much the same way as I could go into, you need something done at.NET, fine, let let's let's knock this out.

Yeah, yeah. Maybe, hopefully. It it's just interesting to me that for a s for a successful small business, right? You need a bunch of things, you need legal accounting. What all that stuff is extremely automatable. You need product and tech which we can build whatever we want right now. And then you need sales slash marketing, which just seems

Still hard.

Well it is it is hard and I think it's it probably even harder now than it was before because you know you're you're battling for attention. Um you know, like every time I roll out a tool in Scalpels within a week, I've seen five or six other people who are like, Hey, I created this free open source thing that does this. And it's exactly the same thing that I already built. And the the big

But is it coincidence

or is it because because you built it and they're like, I can build that too?

Uh in some cases I think they saw what I did and just did it because the you know, they follow me on some social media and you know the shape

Yeah. Yeah. Yeah.

of what they did is pretty similar to what I did. But the advantage that I have is that, you know, I built it around this whole one click is all framework. Um and also I'm not trying to

Yeah, exactly, exactly. And you'll maintain it.

yeah, I mean it uh nobody's paying me for the software, so I don't really care like

You know, competition doesn't bother me at all. And in fact, what I think it actually kind of gives me some confidence because when I see three or four other people who go after the same thing that I've already solved that I've solved well, um, it

Yeah.

tells me that there are enough people thinking about this problem that it's worth it for all of these attempts to clone whatever, uh whatever the incumbent is.

Yeah. Yeah it's it's it's an interesting world. I do I don't have answers. I do have like it it's funny. Um I'm gonna start using scalpels. I I I I do I have been building up this set of primitives for for products so

Analytics, right? This and that, all the little bits that you need. And and you know the one thing I found is like the most useful thing is um they all use MCP, so I just like connect the thing, and then I can and then now now my cloud can say, and they're all set up the same way. So my cloud can say, Well, let me see how that campaign worked. let me look at analytics. here it is, and it all just works. And so once you have a set of primitives around building products.

This is my current thinking. Um you can build products very fast and very easily. And I'm starting to I I have a theory which may be wrong. I'm not sure if that's right. Just an early hypothesis, which is maybe maybe there's opportunity to say, maybe we build very niche products for very niche use cases that are underserved. Um because you know.

Sure, AI can build anything, but but that very specific niche product needs some very specific niche data and a niche use case, and those people haven't been helped yet and might not be helped for years. So that's

Yeah.

where I see opportunities. I'm just looking for more and more niche use cases, you know, like niche use cases with money.

Yeah. Uh that's definitely a good way to go. I'm I'm taking a

And and once I can

build yeah, sorry, once I can build a product very efficiently and I can und so then then my next challenge is I need to understand the niche use case. But yeah, go ahead, sorry.

Right. Yeah,

so I'm I'm taking a slightly different approach. I want to put this toolkit in place. I am building these primitives, um, not so

Yeah.

much so that I can then knock out a bunch of um just products and to throw out in the into the world. Um I do think that's a valid thing to do, especially if you can find a way to actually, you know, reach those people and market them effectively, it you know, it's great. Um I you know, m my initial reason for building scalpels, and I think I've told you this before, was that

Mm-hmm.

I keep solving the same problem over and over and over for my clients. And so

Yeah.

after I build the same thing three or four times for three or four different clients, it's like, why do I keep building the same thing? Why don't these things just exist in a way they can just be spun up, turned on, connected, and it just works?

Yeah.

Um, so really it was initially, in fact, I bought the scalpels.appdomain like three years ago, and that's what I use it for. It was for a very long time just an internal toolkit where

there were just these little miniature tools that I would be able to spin up, you know, w not one click on cloud, but like a, you know, composer create project, whatever. Like it was

Yeah.

everything was built. I could just pull it in and run some scripts that would customize it for whatever the the client's needs were. But instead of doing, you know, four weeks worth of work to build something for a client, it was like, I'm gonna take what I've already built and adapt it for them in a day. and then AI made it.

you know, kind of change the calculus on that and and um Laravel Cloud with their API, um, you know, again was a was another big change where I was like, you know what? Everybody should have this toolkit. Everybody should have these these primitives available to them that they can just pull into whatever they want to do. Um and so, you know, for for us as we continue to take on clients, um, you know, I want I I love that I have these things. It just makes uh

makes it so easy is especially these like I'm having a sales call or sales meeting today and I kind of know the shape of the problem that they're trying to solve and there are three things already sitting in scalpels where it's like, okay, the first three things on your list of things you have to solve, I have a solution for that I can install on one click. Now let's talk about the hard things we haven't solved yet.

Um and so I think that gives me a

Mm-hmm. Yeah, yeah.

real advantage when I'm having these these sales conversations because you know they're kind of looking at this whole big scope of work and I managed to lop out a a decent chunk of that scope of work in a way that still leaves all the data on their own servers. You know, it's not like,

Yeah, yeah.

you know, we can we can shortcut that and especially, you know, in European companies, as you know, because you're also in Europe, there are a lot of concerns around data residency and privacy and things like that. So

Yeah.

Um, you know, being able to

So

say, this is just gonna live on your layer of cloud in Europe, you know, you don't even know what to think about that stuff, you know, just tick these boxes, let's move on to the hard part. Um and I think for us it'll be great. I think for other agencies and freelancers who, you know, and I know a lot of agencies and freelancers do have their own personal toolkits. Um, so you know, maybe they're this isn't really solving a problem for them.

But for those who don't, for the ones who are just kind of rebuilding everything from scratch every time they start with a new client, this is something you just kind of have in your pocket that you can bring into any one of your meetings.

Yeah, scalpels.app. It helps you win new customers.

Yeah, well it's it's helped us it well, it's it's helped us with customers. It's helping us retain customers. So

Basically, right? Um let's let's let

but let's talk about that. Let's wrap up maybe with that, like 'cause 'cause about a a month or two ago the the agency world wasn't feeling that optimistic. You know, some people were shutting down, right? Um

Mm-hmm. Yeah.

like how do you feel things are going?

Um, yeah, I mean, it's it's different now. and I don't know if it's different on a like a macro level, or if it's just that I've I've managed to rethink in my you know, I've I've managed to reimagine what an offering from an agency could look like and should look like in

Mm-hmm.

in this age. I think the challenge was, you know, we we were all sort of unprepared for how quickly um

code and hours with our fingers on the keyboard uh was no longer gonna be something we could sell. You know, we we kind of thought,

Yeah, yeah.

you know, we we might have a couple years left before we have to start rethinking things. Then all of a sudden like November, December of last year, it all just kind of went away. Like now nobody has to pay anybody to apply fingers keyboard. Now it's all down to thought leadership. It's down to architecture. It's down to, you know,

strategy, you know, kind of making the right decisions up front so that you can deploy something and be confident that's gonna stay secure and it's gonna stay performant, things like that. Because the actual like I need this button to do this thing, that's a solved problem. Like anybody can do that now. We just need to know, you know, our our role is making sure that button is put somewhere that is going to continue to work and not become a big maintenance burden uh down the road. And I think that the the recent PHP apocalypse has has further solidified it.

'cause I think this

Yeah, yeah.

because it was such big news and it was unlike in mainstream media and you know, companies who maybe weren't aware that they were running PHP and now are figuring that out, and companies that have been have kind of taken this attitude that like, yeah, I know it's out of date and it's, you know, whatever, but it works and it's making me money and I don't really want to mess with it, you know, that's not what I want to be spending my money on.

You know, now they're kind of waking up to, yeah, we actually should be spending our money on that. And the great thing is now that we can go to them and say, Yes, you have this thing that's running on PHP five point six and uh you know, we managed to patch it together for you this time for this particular vulnerability, but it still has all of these other vulnerabilities that aren't patched because it's not like everything is gonna patch back to five point six, just this one big one has. Um so

Yeah, yeah,

and and plus you could we can make it much better for much cheaper than than

Mm-hmm. yeah.

than we could ever before. So if this was like a hundred two hundred K projects the last few years that

Yeah, it's now significantly

you you didn't want to take on, now it's now it's a fifty K project and plus it'll be much better and redesigned and all the good stuff and you can just do it.

Right. And also it's the sort of

thing where we don't have to put our fingers the keyboards to do it. It's like we just um

Yeah.

you know, they can they can pay us just to kind of

So

plan and supervise that process and so it can become a lot less expensive. Um

So

do you feel like like how how has how have your client conversations changed? Like do you charge like just a lot more per hour and do less hours? Like what's what's the difference?

so I've gotten away from hours entirely because it's so hard. What I'm really

Good. So so what are you choosing for? It's not scalpels, we know that.

right now. No, it's not scalpels. I've gone to the subscription model where basically you just pay me a monthly retainer, and then I just become the guy who it and and now is so initially when I switched to this, it was you know, I'd have a a client who's on like a $5,000 a month retainer.

And that kind

Yeah.

of buys them one week per month of some attention from me. So like

Yeah. Yeah. Yeah.

or one week one one day a week of some attention from me. So like every Thursday I'm on this client because they're paying me, you know, five thousand a month. Now

Yeah.

I can actually be attentive to them every day in smaller chunks because I have things kind of worked out so that I wake up in the morning.

And I have like here are the the things that are the front burner with every client and I can spend just a little bit of time and deliver a ton of value for that client in a small amount of time and then move on to the next one.

Is that because

because you have all your agents have fingers in their systems and and you have telemetry and stuff, is that

Yeah, so the telemetry is there. Um, and then I have um you know, it used to be that I would just have like a weekly meeting with them and that was kind of it. But I have in a couple of cases now I have a Slack app that just kind of lives in their Slack and they can just

Yeah, yeah.

dump stuff into the Slack app and um that that feeds into my

I love that model.

that feeds into my system so that why I wake up in the morning I know, okay, this client is talking about this um and this is somewhere that

You know, I I can and should add value. And so then I'll ping them and say, okay, I've I've caught up on what you're thinking about with regard to this new CRM or whatever it is that you're thinking about doing. Uh here's the conversation

Yeah, yeah.

I think we should have, and I help them get things on track. And um so a lot of it is

Much more responsive. Can

you turn that into an app on scalpels?

It is well it it is sort of so that's um

It it it's feature flagged, right?

no, it's it's one of the experiment experimentals capstan that has um that's one of the one of the components of it. It's um a Slack app I can't really turn into a scalpel's thing. I could, maybe. Does

Well, you you y you can

if if you just let me connect and and whatever so that I don't have to set up a sta a stack app, essentially.

Yeah, so I've I've

kind of avoided doing that. And maybe I shouldn't avoid doing that. Maybe I should actually have like Slack apps that are just integrated that you can you could use. I have to think through how that would work. Um but I know that like GitHub has manifests for apps. And so like to the extent that I might need GitHub apps for things, I could create a manifest that you would be able to create an app from. So basically I create the recipe for the app. And I think Slack probably has the same thing, right? So I could probably create a

Slack app manifest that defines all the permissions and the you know everything that would need to be

I haven't seen it. They

don't seem that open to Slack people. Um I last time I I looked at it, it's still like a lot of buttons you have to click and a lot of like, you know, whatever.

Yeah, and I I don't know. I mean I

Yeah, who knows? It's fine.

it's something that I could certainly explore product I'm saying. It's not gonna be super high on my list, but um but yeah, it's stuff that I've I've built for us and I think it's you know, super useful that um now I can just kind of be in the conversation without having to be in their slack, right? I'm not

You know, the problem is if you're not charging a ton of money, if like if you're just paying me five thousand dollars a month, I don't want to be in your Slack having your people pinging me fifteen times a day.

No.

Like I just want to um I want to be able to wake up every morning and see is there an opportunity today for me to deliver real value to you? And if there is,

Yeah.

then I'll jump in, I'll get my hands dirty and I'll do it. and then the the rest of the time that I spend with them is a lot of just kind of fine-tuning how their own

you know, because they're using hone or whatever. They're they're using some sort of telemetry, uh, exception tracking, automated GitHub issue handling

Yeah, yeah.

through their own agent pipeline. And I go in and I just kind of change the oil on that and I I tweak some things and I look through, you know, oh, I can see why this is happening. You have this in your Claude MD. This is how we were doing things back in August of 2025. We don't do this anymore. Take this out, and everything will work better. You know, just sort of that sort of thing.

And what's funny is I think that ultimately that ends up in me spending a lot fewer hours with any given client, but it feels to them like they're getting more value than they were getting when they were paying me for just writing code.

You see,

that is perfect. I I feel that is the way forward. Like less hours you're still getting paid well. It's not hourly, and they are getting much more value. Like like

Yeah. Well I f

that to me is great. That that's great, right? Who doesn't want that?

I think the reason that the prospective clients

are starting to soften on this and be they're more willing to have this conversation is because um AI has kind of broken that whole hours model for them internally too. Um that they they understand that like I need to spend a certain amount of time highly focused on this problem and put a plan in place and put some automation in place, and then I'm gonna send off my agent, my agent's gonna crunch away into all of this work. Um so

Yeah.

Like the the whole idea of, you know, I'm gonna pay you per hour, it's probably gonna be like how how would I even track that, right? 'Cause I'm gonna spend like twenty or thirty minutes highly concentrated on whatever problem I'm trying to solve for the client, and I'm gonna hit a button and now I have to wait until the agent comes back with results so I can evaluate that and decide whether or not it's actually on the right track, whether I need to pivot or you know, whatever. So it's like do I have multiple clocks that I'm just starting and stopping as I shift from one client to the next? It's like the whole model just doesn't work anymore.

Yeah, yeah.

And I think that

the clients just recognize that it's not really working for them internally either because their their employees are not no longer working in like hourly chunks. They're working in you're getting to a more of a a uh value pricing um type the the they're thinking of things more in in in terms of value.

It's interesting. Have you have you seen

that? Yeah, have you seen that with with other agencies that you talk to as well, like engineering kind of agencies?

Um, well so like Joel and Yeah,

'Cause it feels like like a white model, yeah. Go ahead.

Joel Claremont and Aaron Soray from uh from Milwaukee, the no compromises guys, they've been out they've

Yeah, yeah.

had this model for I don't know, years now. And have so they they went from

They went from hourly to more of like a a value based pricing to just a a plain, you know, subscription. Like you just and actually on their site they literally have like a uh stripe button. You could just start a subscription for like ten grand and give them work without having a single phone call.

It's literally like a stripe and then I stripe my credit card and it's ten grand a month. My card could not handle ten grand a month.

Yeah. Yeah, you you swipe you swipe your credit card and assign them a GitHub issue.

It'll get rejected.

Yeah, well, I mean if you want airline miles, man, that's the way you to pay these things. Um so yeah, they but they it is that simple. You just start a subscription and open a GitHub issue and add it to them and they'll just knock they'll start knocking out issues. And um you know, I don't know. I think it it's I'm sure for them, just like for me, it's sort of always evolving. Like what do I actually feel good about delivering for that amount of money? And a lot of times it just comes

Yeah, yeah.

down to um just having expectations

set with clients. And you know, I I think I'm I'm sort of landing on, you know, five thousand is sort of the sweet spot for the I'm just the guy that keeps an eye on everything and tells you when you have a lug nut loose that's gonna cause a wheel to fall off.

Yeah, and and and the typical client there is we have like, you know, a bunch of engineers and we're deploying like crazy with AI and we wanna make sure stuff stuff it doesn't fall apart.

Yeah. Yeah.

Yeah. That's that's right. I mean, I'm not I haven't started trying to talk to any uh any companies that don't have any in-house development. we've never done

Yeah, yeah, yeah, yeah.

staff augmentations, so we've never had those conversations where it's like you're a totally non-tech company, you want to have this tech project built uh and run for you, and you have nobody in-house that's gonna gonna do that stuff. That's just not it's not a world that we've ever been in.

Um, we've always been sort of the um you have a hard problem you wanna solve, we'll come in and we'll solve that hard problem for you and then hand it back to your in-house development team. And now the hard problem that everybody wants to solve is we wanna continue being able to use these you know, the latest tools and the latest models and you know

a lot of what we do is just cutting through the hype, you know, like a couple of months ago it was everybody's calling me saying, What's a Ralph Loop? And it's like, Yeah, a Ralph Loop is something that will evolve into something useful. Like don't get you know,

Yeah.

there are people out there that are experimenting and let them experiment and I will

Isn't it a crunch

job? Yeah, who knows? Who cares?

Well,

no, not not exactly. But that's ultimately what it kind of turned into. Like the the whole Raoul loop thing went away, but the things that made up the Raoul Floop, you know, some of it went into agent harnesses and some of it went into tooling and things like that. So it's no longer called the Raoul loop,

Yeah, yeah, yeah, yeah, yeah, yeah.

but the things that work from it kind of remain and became sort of invisible components of the tools that we're using. And so a lot of the time that's the conversations I have to have. It's like, yeah, this is the the hot new thing that everybody's talking about, but

It's probably not going to be in this shape in a year. So let's just keep refining and, you know, slowly iterating on the things that are working for you because um

You know what what'll happen a lot of times with with teams that are kind of unguided in this, there'll be somebody who'll see like this new hot thing that everybody needs to do and you know they'll just grind to a halt and try this new thing and maybe it'll work.

And they'll go crazy. Yeah, yeah, yeah, yeah, yeah.

Maybe they'll see great results from it for a while. Um, but maybe it'll be catastrophic and they'll realize, you know, we've actually done nothing for the last two weeks of try retool around this whole thing that it turns out doesn't actually work.

And so, you know,

Yeah, yeah.

a lot of times it's just being that that sober voice of region of of a sober sober voice of reason that um that's worth it for you to pay somebody for.

Yeah, yeah. Higher ed higher

eds, he's a sober vo voice of reason and get scalpels ca it'll help you get clients. You see that's the pitch right there.

It's help it's helping me get

clients. If it helps you get clients, that's great.

That's good enough. Sh shall we wrap up? We're about an hour. Yeah.

Yeah, I gotta do a catch train. All right.

Um alright, catch the train. Uh be careful because you know, we had a cre train crash in Spain, not to freak you out. It was it it it was terrible.

Yeah, I know. It's it's awful. You've had a couple of them this year. Um

Yeah, I don't know. I'm starting to get freaked out. But um alright, well you have a good day and we'll chat soon.

Thanks. Bye bye.

Alright, see ya.