The Changelog: Software Development, Open Source - Natural born SaaS killers (Friends)

Episode Date: January 30, 2026

We discuss the buzz around Clawdbot / MoltBot / OpenClaw, how app subscriptions are turning into weekend hacking projects, why SaaS stocks are crashing on Wall Street, and what it all means.

Welcome to ChangeLog and Friends, a weekly talk show about thinking outside the Dropbox. Thanks as always to our partners at Fly to I.O. The platform for devs who just want to ship. Build fast, run any code fearlessly at fly.io. Okay, let's talk. This is the year we almost break the database. Let me explain. Where do agents actually store their stuff?

They've got vectors, relational data, conversational data. conversational history, embeddings, and they're hammering the database at speeds that humans just never have done before. And most teams are duct-taping together a Postgres instance, a vector database, maybe Elasticsearch for Search, it's a mess. Our friends at Tagger Data looked at this and said, what if the database just understood agents? That's Agentic Postgres. It's Postgres built specifically for AI agents, and it combines three things. that usually require three separate systems. Native Model Context Protocol servers, MCP,

hybrid search, and zero copy forks. The MCP integration is the clever bit your agents can actually talk directly to the database. They can query data, introspect schemas, execute SQL, without you writing fragile glue code. The database essentially becomes a tool your agent can wield safely. Then there's hybrid search.

Tagger data merges vector suburb. similarity search with good old keyword search into a SQL query. No separate vector database, no elastic search cluster, semantic and keyword search in one transaction. One engine. Okay, my favorite feature, the for forks. Agents can spawn sub-second zero-copy database clones for isolated testing. This is not a database they can destroy. It's a fork. It's a copy off of your main production database if you so choose. We're talking a one terabyte database, Fort in under one second. Your agent can run destructive experiments in a sandbox without touching production,

and you only pay for the data that actually changes. That's how copy on right works. All your agent data, vectors, relational tables, time series metrics, conversational history, lives in one queryable engine. It's the elegant simplification that makes you wonder why we've been doing it the hallway for so long. So if you're building with AI agents and you're tired of. managing a zoo of data systems, check out our friends at tigardata at tigardata.com. They've got a free trial and a CLI with an MCP server.

You can download to start experimenting right now. Again, tigardata.com. What up, nerds? So one of the things that we say off and around these parts is the sophomore world moves fast. And this week has been a great example of that. On Monday, when I shipped change log news, I covered a tool called Clodbot, that's C-L-A-W-E-W. WD bot. On Tuesday, when we recorded the conversation you're about to hear,

it had been renamed to Maltbot. Today, as I master and ship this episode to the world, it's been renamed again to Open Claw. Turns out MaltBot just didn't roll off the tongue. So, you'll hear us struggle to say MoltBot. A whole bunch in this episode. Just Run Your Brain's Global Find and Replace Subroutine, and hopefully it'll still all make sense.

Where should we begin? Let's start at the start and take it away. Okay. My name is Simpson, Bartholomew J. Okay. That's just the start of an old Simpsons ramp. Oh, I love that. Oh, yeah.

If I knew the second of mine, I want to keep going. Better than that? Give me the real thing. Is that your best version? I'll start from the start and take it away. My name is Simpson, Bartholomew J. Okay. I don't remember any of the rest.

Do you have a Bart invitation voice? Like, are you like Arianna Grande where she can like imitate? Anybody, basically? I'm a talentless hack. Oh, man. I got nothing. Can you? You know what?

I don't even attempt, man. I mean, either. That's why I just called myself a hack. I can, I can impress some things. You're very impressive in general. Yeah, I mean, when you're impressive, you don't have to be an impersonator. Just be yourself. That's right.

And you impress. Well, thank you. Well, thank you. Well, I was impressed by all that's been going on lately. Have you been impressed? You know what? Goodness gracious. It really is an interesting time to be a software developer.

I do it. to say. It is uniquely interesting in so many ways. You know, just such a wild world we find ourselves in. One of the things that have been going on is this Mac minis sale. Oh, dang. We've been selling Mac minis for Apple. It's not us. It's just developers. Been buying them because of Claudebot, not to be confused with Claude Code, which I guess was confusing enough that Claudebot was renamed for, Prosterity Claudebot was spelled C-L-A-W-D. And I think it had to do with him having claws. It was like a whole lobster theme behind this thing. And I think it was like he had to claws.

And so he was a clawed bot. And he got very popular. At which point, Peter Steinberger, who might be coming on the pod at some point. Trying to make it happen. The creator of Claudebot decided to rename everything. trademark stuff. I can only assume Anthropics said,

hey, please rename this. At this point, Claude is pretty well established verbally, audibly, with Anthropic, right? Right. Yeah.

Different spelling doesn't really matter when you just say the word Claude, everybody thinks about anthropic. So now it's MaltBot. So, Claudebot mouled.

And why do we care? Well, because it seems like a pretty cool project. And one that has gotten And a lot of people excited. I covered it on news this week because it'll do a bunch of stuff for you now. Full confession, I have not set this up. Me either.

Not set this up. Gosh, darn it. You know, the list is long, Jared. You got a list of stuff you want to try, don't you? AI things to play with, man. I know. I just can't keep up with the things to play with. And I want to, though.

I do want to. So that's my excuse at least. The AI that actually does things, apparently. That's right. It emails for you, it calendars, it home automates, all from your favorite chat app. So cool lobster theme. And people are doing all kinds of crazy stuff with this. Well, not anymore.

No more lobster theme. It's gone. It says here, clears your inbox, sends emails, manages your calendar, checks in for flights. Now, this is something you've been, you know, upset a little bit about. And you can elite a little bit further. Don't check in my flights, man. All from WhatsApp. Telegram or any chat app you use.

Now, what a good, first of all, amazing copy. Okay, like when you make something describing it, because in your brain, you've got it, right? You've got all the details. You've got all the things that you're dreaming about for this thing. But then getting it out into a concise word, phrase, sentence, whatever you want to call this, right? Yes.

That's the hard part. That's a beautiful two sentence. This is what it does scenario. So good job, Peter. On that front. In terms of Claude. It's over with, man. Claude.

That's mold. Well, now it's malt. Call it multi. Well, you know, we have to, we have to go to the past to go to the future, right? You got to be like, where'd you come from? I just called it Claudebot in news. And the next day, literally the next day, it got renamed. Malt bot.

Molt. Because lobster's molt, I think. Is that when they lose their outer shell? Yeah, I think so. I only know this because of Futurama and Zoigberg. Zoidberg is always molting. Yes. It's the only reason I have any idea about lobster anatomy, otherwise I'd be clueless.

And I would imagine like any crustacean, when you don't have your crustacean, you are vulnerable to the elements. Right. You're not crusty. Yeah. You need that protective shell from the anthropics of the world. Exactly. And now he's protected. Now he's protected.

Well, nakedly protected. He's molded. Right. For what it's worth. I mean, it's just an open source project. So, I mean, how protected can you possibly be? Is it, though? It is.

Will it be? What do you mean? Protected or open source? Well, it's open source, but will it only be open source? It's just an open source project for now. For now. I mean, when things get this popular, so we should probably say, like, the reason we're talking about it, even though we haven't played with it, is because everyone's talking about it

because other people are playing with it. Here's a few choice quotes from developers whom you may respect. This is from the Kitezai. I can't remember what his name is, but he's on the internet as the Kiteze. And he's an opinioner. He says, create your own molt bot. Go in debt if you have to. I love the idea of this nerdy crab chilling in my attic in my Mac, on my Mac studio.

Just giving me, just giving me shit. I didn't pre-read that. I don't at least understand what that means at the end, but he's excited. He got you. Hey, Malti, give me something. And it just gives you stuff, you know. Just giving you the stuff. You know, whatever I want, just give it to me.

Here's LLM junkie. So, you know, he's showed his cards. That's right. But he says this is the legit, the only agent slash model slash whatever you call this that I have seen that's actually funny. So apparently this thing's got a good sense of humor. Demillion says a megacorp like Anthropic or opening. I could not build this literally impossible with how this thing works. Maybe that's speaking to the open source side of it.

I don't know. LTZ Reno says, ah, I was trying to resist and now I can't stop talking and adding things to MaltBot. Yeah, it gets addictive. And I mean, I'm just grabbing quotes because there's so many of them. There's probably, honestly, 300 on this page.

That's a lot. Well, I mean, if you go to docs, docs, not dox. Dox.molt.bot slash I guess automation might be the pull. Maybe just the docks.

Just go to the docks. Just go to the docks. No, I'm trying to. I'm reading the oral thing like, where am I out here? I just landed here. But like this left sidebar, first of all, it is a

well done documentation site thanks to Mintlify. That's how you say it. Mintlify. I love Mintlify. Service and open source. I believe you can self-hufficiency. Most Mintlify.

What is Mintlify? Like a docks site? Mittlify is a SaaS soon to be dead. If we talk about SaaS is dying, we'll see. That's here for now. I believe you can self-host Mintlify, but it is a documentation platform. You can have the service,

and I'm pretty sure you can self-host it. Pretty sure. I know that our friends over a recent user because that's how I learned about it because our friend Zeno is pretty well-versed in cool tech out there. and cool, I guess, soon-to-be-dead sasses. Yeah.

And there'll be, we'll get to that if we get to that. But I'm not trying to be negative here. That's the sentiment in the room. My gosh. Here in the left-hand sidebar of this mitlify at docks.mult. Dot, dot bot is just a plethora, a plethora of things you could do. Everything from like where you can deploy that, of course, the MacMany, like you mentioned, Jared. Yeah.

Docker, Nix, deploying railway. Render, North Flank, tons of CLI commands in here. So onboarding, configuring, doctoring, dashboarding, and installing. Why would you do that? Skills, of course. And then one of my favorites, Cron. Like, who doesn't love a good Cron job, right? Man.

All my Cron jobs get five stars. That's right. You can also have heartbeats, not just Cron's, but heartbeats. There's a difference. Very well done documentation. A lot. And this is an exemplary example. Isn't that what that means?

Exemplary. It might be redundant, but that's okay. Sorry about that. It makes sense to me. Okay, follow me here. Exemplary example. I'm leading. Are you going?

You doubled down on it. I'm doubling. Really well done, documentation. Like this, I'm very envious of this level of documentation, which is no surprise given the fanfare reviews. When you have docs this good. it's got to be good. So there you go.

No first-hand accounts here. Speculation. But lots of interesting skills you can install. I don't know. People are doing crazy stuff with this. Go to the website. Find out for yourself. You can make it do all kinds of things.

I'm still just happy with the coding bots, like just coding for me. I haven't gotten to like to automate the rest of my life bots quite yet. Right. I fear anything writing my emails. but I also embrace things helping me write emails faster. And so I'm tentatively excited. I did have a friend who read ChangeLog News already texted me and said,

yes, they bought a Mac Mini already because of that. And have I set up Claudebot yet now called MoldBot? I was like, no, and I just letting everybody know about it. I haven't tried it yet. But I am intrigued. So it's on my list. It's on my list. I'm going to buy a Mac Mini too, Jared.

but it's not because a Claudebot or Maltbot. It's because, well, you just want one. Well, you know, honestly, it's, you know, Tim, Technotim and I, we just covered this on Friends most recently. Home Lab, 26 edition. It's fun. Go deep with us. The thing he said was availability.

Now, I thought he was going to talk about availability in terms of services. He was talking about hardware. Like, he said, yeah, I mean, like, you just. That's what he means. Good luck finding. at a price you actually wouldn't afford. Good luck finding GPUs at a price you want to afford. And thankfully,

the Apple tax hasn't caught up to the bump of the hardware. And I assume there's still Mac minis to go around. And so you've got this. Now, I really wish the Mac Mini was an M5 because I hear the M5 is so much faster. And maybe that's coming soon. Maybe it's coming summertime and a year, who knows. But I really do think that the Mac Mini,

given the availability of other hardware, where RAM, CPUs, motherboards, you name it. It's all hard to come by. And all you're really trying to do anyways, potentially, is get some level of GPU or massive RAM amount to use for inference of some sort. And so if Claudebot or Maltz, sorry, I'm going to keep calling that. If Maltbot, gosh, it's like seeing Zulip, get me upset. I just want some inference.

I want it on my machine. Like, I want to have a good math. MacBook Pro, of course, but I don't want to overload my main dev machine to do stuff and my creator machine to do stuff. I want a dedicated machine for that. And GPS are hard to come by. And a Mac Mini has pretty much all you need in one little package. And it's, you know, thermal density is good. You know, all the, you know, dynamics are good.

Power consumption is good. It's SIP's power compared to other solutions. So, I mean, it makes total sense. And I want to run it for an 8N or, in this case, Maltz, but also just to run some transcripts, man. I can't burn through some transcripts. Burn them up. So it's also cheap to get into.

I mean, $5.99 starter price. A lot of people are saying who are buying these are saying, actually, you can most time when you want a Mac, you're like, yeah, skip the starter version. You want to beef that thing up. But a lot of people are saying that actually the baseline Mac Mini is good enough. for a certain use. Now, obviously, depends on your use.

But even for running Moldbot, they're saying that's probably just fine. I did check the Mac Rumors Buyers Guide. And it has been 455 days since the last Mac Mini. That's year and a half-ish. But the average days between releases is 732. So they're calling that neutral mid-product cycle. It is the M-4.

And the M-5 will probably come out, but not maybe next year, maybe some like late this year, perhaps. There's really no saying, but at 600 bucks to get started, they are available as far as I can tell. I mean, my friend just bought one today. And they're super small now. I had the previous version, the previous form factor. I mean, they literally look like an Apple TV now in terms of form factor. They're super tiny.

And, yeah, good computers. So I understand why people are doing that. Well, they also have the ability to have 10 gigabit Ethernet. Right. One of the fastest SSCs on the market. You could do a lot with even 32 gigs of RAM. I wouldn't buy anything less than that personally. Like, it's the $1,500 model.

I only know this because I spected out recently, but not because of this conversation or even Malt. But, gosh. It's not that bad, man. What a shame, man, to have to rename something like that. I mean, Claudebot was cool. that sounds cool but I can get it it's a it's a trademark and you know not an infringement but

what do they call it actually let me pull this up real quick on Forbes.com they called it trademark confusion yeah trademark confusion which I can see why because when I first saw it

I thought is this from Anthropic and then I saw how it was spelled and I was like no it's not anthropic but when you just hear Claude obviously it's identical sounding so It does make sense. So if you want my recommendation, this is what I would buy personally. And if I was going to buy a Mac Mini, it would be between two models.

Fully specced out because why not? Or what I would consider the base model would be the M4 chip with 10 core CPU, 10 core GPU, and 16 core neural network, 32 gigs of RAM, one terabyte of SSD, and always go 10 gigabit. Because that's just how you had to do it. That one there is 1499. Now, I'm not telling you to go buy this, but if you do, that's the one I'd buy. That's the one I would buy it personally. I'm not buying it, though.

I'm not buying it, though. Speck out the biggest one and then not buy it. Can you go more expensive than that? Maybe bigger, bigger. I have a 3090, an RTX 3090 from a while back that is still of good use. Now, the power draw is dramatically different. The sound is dramatically different. the architecture and the hardware requirement

is dramatically different. That's kind of fun to play with. But it's done. It's set up. I don't have to play with it anymore. You know, it's sitting there as part of the Proctumoc set up. So I can just spin up a VM and, you know,

pass through the GPU and call it done. So I've got what this could probably do. Although I think the throughput between the speed, I think, of the throughput through the machine is better than what I currently have. So plus, I mean, you can put this thing in a bread box. You can put it in your cupboard. You can put it on your hell of glove box. Yeah.

Put in your truck. Back pocket. Back pocket. It's a big iPhone. Yeah. Oh, a big iPhone with a worse processor. Okay. So I want to know is that is Peter getting a kickback from Apple?

I think you probably should. I mean, at least affiliate link through that thing. I don't think they really do. affiliates, but if they did, he should be linking over to Amazon with that affiliate link. Oh, gosh. Here might be the next best thing, though. When they released the latest next Mac Mini, and they're still a run on the Mac Minis, put him in the keynote.

Yeah, they demo. Here it is with Maltbot. Yeah. That's the next best thing. I mean, honestly. That'd be cool. Yeah, that would be cool. What else's been going on?

Well, I read this. post from Roberto Selbach. Okay. Your app subscription is now my weekend project. Oh, this plays into

the death of SaaS. It does. Your app subscription is my weekend project. Now, this is an old trope or meme, I guess, in developer world is like, the I could build that in a weekend

kind of a thing. Except for is it, his point is probably that that's now less false than it used to be. Well, here's some key points. The author with Zero Swift or Mac OS Experience built three functional apps, dictation, screen recording, and a marked on editor to replace $14.15 a month of subscriptions they had. Now, that to me is a really, really interesting new world, Jared.

Isn't it? Let me say that again. Zero Swift, MacOS experience, three apps replaced. 15 bucks a month. I believe it. I'm building a video editor right now. A MacOS native. I am.

I'm just like, you know what? Why not try it? Why not try it? Because I can spec out a video editor to work just the way I want it. And none of the extra stuff, of course, we use Adobe Premiere and all their tools.

And it's got so much in there. It's like, do I need all this? What if I had one that just worked exactly in my brain works? And I just don't need that anymore. Now, can I get there? I don't know. I'm not sure if it's going to get there.

So far, it opens up videos, it plays them, it has all the keyboard shortcuts I want. It does markers, ins and outs, stuff like that. But it doesn't actually do like the rest. Just like an editor for existing videos, like to create clips and stuff. And I never would have even considered such a thing. Because like, I don't, I don't know Swift. And I did write a Mac app probably 15 years ago, which was really simple. Yeah.

And it's called Detours. And its entire point was to manage your Etsy hosts for you without going the command line. So you could like detours. It's called detours. Yeah. Yeah, it was cool. It was very cool.

And I had to learn a whole bunch of stuff to do that, which was fun. Good side project. Put it out for free. People downloaded it. People liked it. Eventually, it actually didn't use Etsy hosts. It used a different Mac native system that does the same thing via an API that Apple built, which was really cool.

And that maybe not have to actually like edit that file and get pseudo-privileged and stuff. And then they deprecated it. And so eventually it quit working. I just didn't care enough to to make it work a different way. My point is that I had to learn so much. And it was so simple. Like literally you put, it's a table with like two columns, right? The address and then like the redirect spot, just like your Etsy hosts, which like space delimited stuff. And I learned so much objectivity to do that and packaging and X code and all this. I would never consider. to build a video editor. But, you know, your Adobe's Creative Cloud subscription is my weekend project, I guess.

I mean, there's so much there that Adobe does that this won't do. But man, just the, the hutspa, I don't know, what does it give you? It gives you, like, the freedom to just try. Well, that goes back to Tom Preston Warner's initial phrase around GitHub, which was a little bit more crass than that, permission to F up. Right. which I think is really interesting that you say that and I respond with that because look what happened with GitHub, you know, with open source even.

You know, when you had permission to just try things and permission to just create things and permission to just join communities and permission to fork things even, you know, a lot of liberty and a lot of agency comes into play. Yeah. And I would say a lot of courage. I think it's, you have to have courage to like bite off more than you can chew when it comes to even a weekend project. You know, like, why would I do that aside from this assistance because the depth is so deep and I may not come back to the same person, okay? Right.

I'm like, I might come out a whole new developer and I might not like myself, okay? Right. Yeah, I came back like in square brackets, you know? I was like, yeah, actually. I have spaces. No, come on now. I didn't even mind the square brackets after that. I'm like, you know what?

I'm like, you know what, I'm talking about, Jared? This stuff is atrocious. You just got used to it, you know? That's right. Yeah. So I am with, is it Frederico? What's his name?

Roberto. Roberto. My bad Roberto. That's okay. I think it's cool that we now have leverage so much so that we're willing to like dip our toe into other pools to see if the water is warm. And save ourselves. some money, save ourselves some time, save ourselves some pain.

I think you actually may have replaced more money than this. Now I'm looking at the details here. How many descriptions that he replace? Well, I'm looking at the details here and I wanted to say this one part because this is the part that I really enjoyed about his post. He says, in the past, I used to use Loom, which I like Loom. And if you go to Loom.com, I think you might like them too. And they're cool.

They really change things. And put a little video recorder in your MacMacon. you and it's like screen sharing for for sharing like captured to share right to document to share to i've used it in sales scenarios where i want to introduce myself to somebody or i want to share an idea and rather than the meeting i just tell them the five minute version of what i'm trying to convey skip the meeting still get the person you know and then you can respond in email or with your own video we can go back and forth if you want to which we've never done that but i've used loom before and had and and received value from it now i did question

The cost, because I didn't use it enough to make it worth it, but the value I got when I did use it was great. So it's one of those things where it's like high value, but not frequent enough to make it be like, oh, yeah, I should pay for this. Right. So he says, I used to use Lume, which costs 15 bucks a month. And so after creating Jabber, which is what he created,

I got excited and vibe coded real, which was the Lume version of it, simple demo recorder. It's, of course, open source on GitHub. Getup.com slash R-Selback or sellback. How are you want to say that? R-S-E-L-B-A-C-H-slash real, R-E-E-L. Cool.

And you open-sourced it. I love that. Well, if you do, if you take that exact same process and you multiply it out over time and space and you say now, I should say most developers and then even some non-developers. So like all these new people are going to be trying to do similar things that Roberto has done and that other people have done in small ways. And they're going to apply that inside their personal lives. They're also going to apply that at their work.

And look at all those subscriptions they have for their business and say, can we run this business a little bit leaner. Do we need all this? Can we build something ourselves? The cost of building your own is coming down. Could that lead to the death of SaaS, the death of software as a subscription? Well, this might be the nail in the coffin right here. Let me read this to you.

One more verbatim from Roberto. All right. Can't skip these words. These are good words. I quote, all of these $10 per month apps are suddenly a week. weekend project for me. I am an engineer, but I have never written a single Mac OS application. I have never even read Swift code in my life. And yet, I can now get an app up and running in a couple of hours. This is crazy. I concur, my friend. I concur. This is crazy. So what do you want to

say about the death of SAS? Have you experienced this personally? Have you replaced any SaaS? Have you killed a SaaS in your life, Jared? Not personally. No. I mean, even if it's just killed to you, it doesn't have to be killed for everybody else. Like, has it dead to you? Right. I haven't actually done that.

I'm trying to think I'm pretty light SaaS user. I'm trying to think of what my subscriptions are that are software subscriptions on a personal level. Now, we have a lot of change log subscriptions. We can probably talk through those. But honestly, I don't have very many subscriptions that aren't content. Like most of my subscriptions are content.

I guess we do use our groceries app, which is an in-app purchase. It's an iPhone app that we use in order to manage a shared grocery list with like, you know, what do we need and recurring and stuff like that. And it's like six bucks a year or something. Like it's not money that isn't well spent or like there's no pain for us. And so I think I haven't replaced that. one I haven't even tried. What else do I subscribe to? I can't even think of anything. What do you

subscribe to, Adam? What are your SaaS, your personal SaaS? Oh, not much. I mean, I would say the same. We're here in Texas and we have HEB in because we have HEB. They have an amazing app, which has amazing lists. And so all my grocery lists are done in there. That's a good idea by that. They have buy it again. This is like, bam, from me. Right. Right. Sweet. Okay.

Yeah. I think that's actually a really interesting place in, in technologies. You have these non-typical tech companies that are now tech companies. I mean, they have a really well-done iOS, you know, app for H-HB, a really well-done website even. And it's not just like, hey, we're a grocery store. It's like, hey, come and shop here. You know, they will deliver to your door. You can go and pick it up.

you know, all that good stuff and order in advance. My wife is a big fan of ordering advance and picking it up. I like to go in the store because I'm a, I have to touch it, feel it. You know, like I like to choose my meat and choose my fruits and vegetables. Sure. But I'm not against her ways. I like her ways.

She's amazing. Not many subscriptions, though. Honestly, I think, oh, you know what? Let me back up, my friend. Oh, he's got some. I got one. I forgot about this one. Well, friends, I don't know.

know about you, but something bothers me about GitHub Actions. I love the fact that it's there. I love the fact that it's so ubiquitous. I love the fact that agents that do my coding for me believe that my CI CD workflow begins with drafting Toml Files for GitHub Actions. That's great.

It's all great. Until your builds start moving like molasses. Get Up Actions is slow. It's just the way it is. That's how it works. I'm sorry. But I'm not sorry because our friends at name They fix that. Yes, we use namespace.so to do all of our builds so much faster. NameSpace is like GitHub actions, but faster. I mean, like way faster. It caches everything smartly.

It cashes your dependencies, your Docker layers, your build artifacts, so your CI can run super fast. You get shorter feedback loops, happy developers because we love our time, and you get fewer. I'll be back after this coffee and my build finishes. So, That's not cool. The best part is it's drop-in. It works right alongside your existing GitHub actions with almost zero config. It's a one-line change. So you can speed up your builds, you can delight your team,

and you can finally stop pretending that build time is focus time. It's not. Learn more. Go to namespace.com. That's namespace. Dot SEO, just like it sounds like it said. Go there, check them out. We use them.

We love them. And you should too. So I was going to call it Mac tuner. Remember this, Jared? Oh, yeah, Mac tuner. Okay, so clean my Mac. Yeah. So I got frustrated and it was not because of the cost.

It was because the, I guess it was kind of cost, RAM costs. Right. So here it is eating up my RAM and I'm like, you know, I'm hitting swap and stuff like that. I'm like, gosh, do I really need, you know, this thing running all the time? And so I was like, I wonder if I can vibe code a better version of this.

Because I mean, it's just scripts. All it is is a pretty UI on scripts. I'm like, really, what I want to do is maintain my Mac. And so I was going to call it Mac tuner, which was still cool. I've since changed the name because I'm more nerdy than that. I want it to sound like a system service. So it's called tuner D. T-U-N-E-R-D.

T-R-D. And let me tell you, one of my favorite features is in here is this, is that you do T-U-N-R-D. is a CLI, Tuna D, memory dash dash AI. What do you think that does, Jared? It's going to tell you how much memory your AIs are using. Man, it doesn't even tell me in like a table or a dashboard. It's a one-liner call out to Claude because I have it installed and cloud code, as a matter of fact, and haiku. So it's super fast.

It's almost instant. That's how fast it is. And instead of waiting for the whole entire thing to like be done and then print it to the screen, it streams it. It looks cool. So rather than me reading this table about how my memory is, now I'm reading words that are crafted

just for me in real time based on what I'm doing here. And it's like, hey, your system's doing great, a little overloaded here and there. Think about Dropbox is always one of the culprits. Oh, man.

That's what we got, a vibe code, a new Dropbox. Amen. That should be our next project. I kind of feel bad celebrating this. Okay. I really do.

I kind of feel bad at celebrating this. I did a fist pump and everything. Okay, for the audio listeners. He did. He fist pumped. Okay, you can't hear those. I fist pumped. Well, gosh.

There's a business SaaS. I guess it also is personal because before business, we had our personal Dropbox. And so there you go. And I pay annually, I think. But, you know, that's a form of monthly. That's a big software as a service for both individuals and teams, making tons of money.

money sitting there, not really doing much that I can see outside of the box, you know? That's right. Like, it's there now. I will tell you. Let me tell you what it's doing here. What's the last Dropbox feature that they added where you're like, oh, that's cool? I've been waiting for that. I can't think of one.

So the coolest feature they added probably in the last 10 years was selective sync. That was a big deal. Tuna D memory dash dash a out. Oh, gosh, this is so beautiful. I love this. I love vibe code. I'm just kidding. I'm not going to say it.

I can't go there. I can't say it. It's just too not nice. Can you share it so we can see it? Yeah, I'll screen share this. Oh my goodness. Did you know the original Dropbox Selective Sync 1.0

came out in 2010. That's crazy. That's 16 years ago. Okay. Do you see this, Jared? Yes. Adam Sikoviac at AS dash MBP

M1 That's right As soon as your initials and your Mac Mac with Crow and it's an M1 That's right Okay All right

Your naming conventions Prepare for this Tune or D That's it Okay Let's see it Analyzing Hiku

Oh You just some panel A little Just fine That's cool Oh That it's It's mad about

Granola They're running Too If you keep Chrome It just yelled about Dropbox. The format is a little weird here in this smaller screen, though.

What's going on here? Yeah. Well, you've got it bigger than that yet. Yeah. So maybe some touches and some feelies on that one there. Now, it's not upset about, let's run it again and see if we get a new result here. Michael, was it complaining about Dropbox?

It just complained about it. And I cleared it to do this demo. There you go. There you go. Now it's upset about Dropbox again. See, it's like, if you want the single biggest quick win, quick granola, I don't even why it's running. And Dropbox, 1.1 G.

That's Ram. What is granola and what's it doing? Or granola is What is granola? Granola like records It helps you with your meetings. It takes your meeting notes for you. Gotcha.

But like if I click those too, it's going to be like, man, Jared was really on fire today. Maybe so. Great takes. Well, hilarious even.

When you're not actively sinking, look at this. 2.4. gigs consumed. Zero downside. Or zero downside in that case, yeah. But then you could also do what could you do? Can you do Tuna D memory release or whatever?

Like free it. Now see now you get the whole dashboard. Would you see that down there? Now it's printing out down there too. But you get the full on dashboard. Look at this. Heavy. Google Chrome always.

Like look at these guys. These culprits here. AdobeCC, not even using it right now. Come on. You have a dash dash kill and you can just kill the stuff. You know? Well, I got it. Don't do that because it's going to knock you out of our session.

Don't do it. I haven't done this yet. Let's see what the dash help says. You sure you want to do this? Let's see here. We got some cleaning we could do. Execute. That's the word you chose.

Dash, dash execute. I'm just reading your docs. It says scroll down slightly. Apps uninstall Slack dash dash execute. Yeah. Oh, the reason why is I didn't want this thing to be, especially with uninstillations. I didn't want it to be like accidentally uninstalling slack or something like that. I wanted to be safe to run initially.

And rather than running like dash in for a or a dry run dash dash dry run, I wanted to actually make it so it was explicit. So somebody's installations are like system, you know, issue stuff where you had, you know, you had to do dash dash execute. So if you want to install something, you have to be explicit and put the slice your neck off situation here. Yeah, it seems like more like a guillotine scenario. I think I would advise like

I just seriously I wrote that copy Mac OS system tuning toolkit Get dialed maintain your Mac AI augmented And then dev is the It should be the version number And I haven't played with this enough

You have to give it a version yet So there's no version Dude Put it on the internet's Come on Well, let's have that time man Tudder D No you just say

You launch Claude and say I would like this to be on GitHub soon Please take care of that for me See the all commands run in dry mode Try remote by default. Safe by default. I like that. No accidental uninstallations or executions.

Cool, cool, cool. So there you have a little bit of your own, you know, killing of a SaaS because that was a subscription model. Now, the downside of this is like, hey, these are our people. This is our industry. These are our people. This is how a lot of us make our living is with this stuff. I was thinking about like uptime robot and all these things, you know,

uptime checkers. Like, you could go. code one of those up in a half an hour probably. That's going to work. Well, you could malt bought it. You could mulch on it. You got, you got heartbeat checks in there, which is that? So I have two minds.

There are two hearts of this where I get excited and I'm thinking like power to the people to build their own stuff and save money. And then on the other side, it's like, yeah, but I also have all these other people who've been building these software services and built lives and careers and hired a bunch of people and helped them make a career of themselves. themselves. And they're now threatened. Obviously, you can't stop the system once it's moving. But this is why people are talking about the death of SaaS. And I think mostly this conversation has been going on around recent stock market moves because the SaaS bundle of stocks have all

been trending downward over the last six months. We talk about Adobe. We talk about at last And we talk about Oracle, Salesforce, Service Now, DocuSign, et cetera, et cetera, et cetera, go ahead and find your SaaS ETF and you'll find all the tickers who seem to be trending down and not just with your typical market movements. I mean, that's what people are saying is like this seems like a sector that is losing value. And that can be concerning. and also a sector that needs to maybe find itself on the other side of the AI wave and some will find it through and others won't.

Your thoughts? I think there's some that I'm actually on Dr. Wealth.com right now. I just Googled quickly. SaaS bundle stock. Thank you for that new phrase of my lexicon, Jared. I didn't think of that phrase. I'm down with it. Keyworded it for you.

Yeah. Zoom. I'm not sure who would vibe code of Zoom. Yeah, I'm not sure if you can vibe code a Zoom Like maybe you could but there's a lot there There's a lot going on there Could you vibe code a docuSign? I think signatures

In the case of everyday folks Is so free Like I know Dropbox has it DocuSine might even have a free version But once you get into that market Where you for sure need to have like daily signatures You're not going to vibe code a docicine There's too much legality in there

I don't think you would at least not in 26, maybe 27, 28. Shopify, would you vibe code away as Shopify? I mean, that's maybe for your individual self, Rails. You got rails, right? That's what Shopify is. So on an individual level, you could do that, but would you replace the entire platform? I don't think that you could do that with a vibe code.

Could you vibe code Cloudflare away? Maybe in parts. I don't look at Cloudflare right so much as a SaaS as it is like internet infrastructure, but maybe it's in the bundle there. parts of it, yeah, for sure, Cloudflare. Twilio, would you vibe code Twilio way? I'd vibe code on it, not a way. Yeah, what if you find another, you know, base layer that's just cheap and doesn't just

has an API or just an MCP server, you know? All the extra stuff they built around it is where the money is. Yeah. I mean, I think probably you got CTOs and their CFO looking at this sort of strange new world and saying, what are our biggest? subscriptions. And they're like, well, we spend $300,000 a year on DocuSign. Like, I'm like, I just made the number up. No, I know. I'm, yeah, you're the guy.

As if I'm responding. Exactly. I'm the responder. I'm like, what? Yeah. Like, could we, do we need all of those features? No, we don't. What part of DocuSign do we use? That's right. And then they list out like the four things that we do. Maybe we do them a lot. And then like, you know, can you put two engineers on that for two weeks and see what they can come up with? Two days, Jared. Two days. I'm being generous. A weekend. Weekend project for two devs.

Or one dev and two clods. Or a dev, a clod and a vault, you know, and see what they can do. I feel like lots of people are doing that and are going to continue to do that more and more as these tools become more capable. And I do feel like we've crossed the threshold where capability is now viability. I mean, they're very useful now. It was not, it's not just the promise. It's the delivery. Are there problems?

Yes. Is it 100% amazing? No. Do you still need to know stuff? Absolutely. Like all those caveats aside, there's real value here, and it's starting to be realized,

and it's going to change things. I wonder if it's back to the days of service providers versus software as a service. What if we're in the day of bespoke software for me? Because how much SaaS, like you just said with DocuSign, when we start to examine, and audit, what do we actually use? Well, we only use four of the features we get from Enterprise.

But we have to have enterprise to have the API access. Right. So, well, we only really pay for the API access because that's what we need that. Because that's how we're doing X, Y, or Z. Maybe that's how we're doing syncing and we have our own backups for SOC2 compliance or whatever, for whatever reason. Disaster recovery. It's always, D.R. is always a good culprit there.

We paid for for disaster recovery, business continuity. And there's two reasons why you always. they spend too much money. Right. You know, and then you do that, but what if you're like, okay, great. Now I've got to go find to AI native,

well-informed engineers who wouldn't normally build it on their own, but now they can because they're augmented. So what if the world is full of bespoke software built just for us, almost in real time, potentially by service providers and or

the product market who has never touched a thing, vibe codes for the most part. And maybe you have a couple engineers sort of like making sure these things are legitimate and secure and the things that really matter to the business. Like that's one thing that the Damien Tanner and I talked about. He was saying how he was evaluating a CRM tool. And they actually used it for a bit. I forget which one he said they used.

I think he may have said it earlier. I'm not sure if he did. And he questioned. He's like, gosh, I don't really get a lot of value. they make me go to this web front end that's not really for me in this dashboard that's not really helpful but I do need to keep track of these things and so he said he just rift for like four minutes into like text a speech and yoloed whatever that was and out the other end came to this CRM they're now using and we speculated on that conversation like what if we're in the era of just in time just for me

interfaces. And it's not really the death of SaaS necessarily. It's like maybe it's always ill placed or misplaced or we still need the API. Maybe we still need some things from it. But the UI layer is for the, I think, I think Damien said this. I'm going to say it too. The dumb, slow humans. Like, you don't need that for a fast AI.

You don't need that web UI. You need an API, maybe a CLA, maybe an MCP server if you're, if that's not a bad word around here. You know, gosh, MCP. check yourself. You know, that's what his speculation was, was this world, like, just in time interfaces, is just from me bespoke. I can see that being the case.

If you think about it, what we're doing today, which is like, and which we've been doing for a couple of years now, which is like, I don't write sequel anymore. And Matt Ryer doesn't have to build his Grafond dashboards anymore. And the honeycomb thing writes its own stuff, you describe what you want. Like, that's kind of a first. of a bespoke interface based on demand. Like a SQL query is that. It's just a hardcore nerdy one, right?

Like I want to see all the sales where this amount was less than this amount. And then like make sure you don't include the ones from Australia, but then do that. And like you tell it that and it builds you that query and it shows you that data. And that is a bespoke one-off ad hoc interface to the information based on a human's desires. And as the tools get more tools to build with, it was very simple to say, I actually did this recently. Here's a good personal example.

I went into my OPPD account, which is Omaha Public Power District. That's our local energy partner. And they provide, I think it's like 12, that's more than 12. I think it's like 18 months of your history in this like wonky UI. That's a single page app that was built by some low-end dev,

you know, who just got hired or whatever, an intern project. And I'm looking at like the URL. And I want more information than that. And I want a better chart than that. But I'm because I want to know like, how am I doing now versus, you know, previous years. And so I can tell from their API calls because I'm a nerd. So I open up the thing. And I'm like, yeah, they're just like they're paginating this collection.

And it's just the UI that's limited. And the UI stinks. And I'm like, all I got to do is give that to Claude and say, look, this, here's my cookies. I hear the API calls. Just see how many months you can actually go back and then build me an HTML page that has a chart that provides like monthly, yearly comparisons, like all the things that I want to see. And like seven minutes later, I was looking at five years of data.

What? Yeah. I was looking at my whole history on my own website, running from inside my browser that's just sitting on my desktop, an HTML file. just in time bespoke for me. And that was cool. And then I was like, you know what?

I love that. I'm done with it. You know, like I don't need it anymore. I can open it up whenever I want. It's historical data, so it's not going to change. I can look at that. And then next year, when they have more data, I can go ahead and just rerun the things.

They grab the latest and throw it in there. Yeah. And that is, I think, more and more and more of that's going to happen. So I think he's on to something. And I think that is going to be a future. So the only pushback I have, which is not much. Because I fully agree. I just wonder who's going to do that?

What kind of person is going to do that? Do what? That's what you did. Like, no one else is going to have your intuition like you did because they don't have your software engineering history. Sure. But because you do, you can. And because you can, you do.

Someone else that doesn't understand your row structure or sees between those two lines like you did can paint that picture. So I think the availability is certainly there, but the kind of person who can do it for now is still largely an engineer or what we may call soon a former engineer. Gee, thanks. No, yeah. You know, I'm trying to say, though, like, like. Yes, I do know what you're saying. My wife's friend, no offense to them, they're not technical. They're never going to do that.

I get it. They're not the patience at all ever for it. I think there's like a gap right now. So the one insight that I had that most people wouldn't have is like there's more data in there. They're just not showing it to me. Right. And I could actually, I didn't necessarily have to give Claude any of that.

I could have said, here's a website that I'm logged into. Can you get more data and it might have been, it would have just figured that out on its own. Oh, that's true, Jared. Okay. Good concur. I like that. So I think my expertise probably gave it a leg up. and got me there faster.

But I think there's a gap and then we have a next generation. You know, we're going to have AI natives here soon. In fact, our kids are going to be darn near close to a. I have native, right? My kids are already saying things like, well, did you ask chat, you know? And that's always, that's not the chat in there. That's chat, it's not some,

the Twitch streamers chat, which is the other version of Ask chat, you know. And what they mean by that is like, any question this feels like it's impossible to answer, like we would never know or like, eventually you'd be like, can I get more data for my opioid, for my OPD, like, well, ask chat. You know, chat's going to tell you. And so they're going to be asking all kinds of questions that we wouldn't think of.

And so they might go there. Hopefully they go there and they use it for great good. Well, friends, this episode is brought to you by Squarespace. I love Squarespace. I'm a user of Squarespace. And Squarespace is an awesome all-in-one platform where you can stand up a professional site, offer paid services, get paid the whole thing without writing a single line of code or debugging CSS.

They've even got Blueprint AI now, which takes basic info about what you're building and generates a fully custom site with actual design recommendations. Not a template you have to fight with, but a starting point that already looks like you thought about how it should look. And for the data nerds out there, I know you. That's me too. Built in analytics are cool. See your traffic, see your revenue, see your bookings, see your sales.

Figure out where to focus all from one single dashboard. No third party apps. GDPR headaches, just there for you. And whether you're launching a side project selling a course or finally replacing that under construction webpage you've been kicking around, Squarespace handles the website part so you can focus on the thing you actually want to build and the content you want to create. So head to Squarespace.com slash changeload for a free trial. And when you're ready to launch, use our offer code, change log to save 10% off your first

purchase of a website or a domain. Once again, Squarespace.com slash changelog. Let me share one really uncanny story that I think was just really wild. We took this trip in October to Pennsylvania because I'm from Pennsylvania originally. And we had this, you know, we ended up getting stomach flu, which made the trip kind of suck really bad. I'll spray those details because they get kind of gory. But it was fun. It was a cabin in the woods in the mountains of Pennsylvania.

Beautiful. Like this area is beautiful. It's the fall time, autumn. time and it's when the leaves are turning. And so they were just beginning to turn. And so like this is beautiful mountain air, you know, just amazing, magical. And I think I had done some planning with my Claude. So my wife uses chat. I use Claude. She now uses chat and Claude. I still just use Claude because I'm biased. Anyways, that's a different, different concern. But because it knew I plan this trip and what we were doing, maybe it didn't even know all this. But I asked it because I

tell my son's stories at bedtime. And I said, well, you know, rather me make up this story about our trip because we want to reflect back on where we had been. We love to like reflect back on our fun times, the good old days, which was basically yesterday. And, uh, and I had Claude make up this story. And I, you know, similar to Damien, I just voiced in there dictation. Push the microphone button. Hey, Claude, you know, do this, that. This story went.

We went to Pennsylvania. here's where we're at. This is zip code. Kind of gave us some context and details that it can use to come up with some, you know, some kind of realistic version of it. This thing told us a story.

I swear it was there with us, dude. Like, that's how good the story was. So it retold this hike story we did at Ohio Powell. That's about a four mile hike. What we saw, the landmarks we saw, the bridge we crossed, like where we stopped at, where we got the ice cream from.

Like, dude, like it knew. these things. It retold us to I swear it was there with us. That's how good this word was. Like that is wild. That's not a nap. That's just entertain me stuff.

Right? Yeah. So wild. What did you have the voice mode on the entire time or something? How did it know how everything where you went? Oh, that's because I told it in the,

my prompt was just detailed. Got you. I think it had some chat history too. It may have pulled from. But I was like, are we going to go to this cabin or that cabin? But it wasn't like playing this whole trip for me. I think it may have had some data in our chat history potentially.

But my prompt to it was like we went to Ohio Powell, give me an Ohio Powell story where it took this hike. And I told it the loop we took. And so it knew the loop. It knew the place where we'd go get the ice cream and knew the different stops we'd take at like horseshoe falls and stuff like that. And so because it knew all that or it could go and find it and research it quickly, it just did. And it gave me a 20-minute story to read. He did my son. And it was the first time I'd read to him, like a story where he didn't fall asleep during the story because he was gripped by the realism of it being our story.

He's like, dad, did it just tell us it like what we did. I'm like, it basically did, son. Like it basically did. It was so close to being our story. So very, very interesting times. I mean, I don't know there's a SaaS out there for that. But maybe there's a sat. You know, like just in time storytelling.

I mean, you're taking an audible time right there. That's true. That's true. I would say the other one that's pretty interesting, if you haven't played with this yet, is Notebook LM. Oh, is pretty wild.

Our job, man, the podcasting. Now, you know what? I had this different take on it because what's the easiest way to describe it. Like, I was learning some stuff. And there's not a lot of content on some of the things I'm trying to learn.

And so rather than that, I will kind of like have this session with Claude, and I'll tell it to summarize it for me, and I'll shove that into Notebook LM. It'll create a podcast. Now, it's not that great. It's not that great of a podcast, but it... Yeah, I listen to one.

It will still do the audible version of the information, but... Right. It's kind of... I think that the future of whatever it is will be better. Current iteration, good enough. I see the value. So I'm mostly kidding when I say they're coming for our job.

I think it's silly. They call it up. podcast when literally it just means two people talking. Right. But I mean, I guess that's what a lot of podcasts are, you know? Two people talking. Yeah. So here we are.

But, and I was kind of offended because the way it opened up was like, they're talking about what they had for lunch and stuff. And I'm like, I know neither one of you had lunch. Like, I understand it would be valuable. And that's obviously just one example. It was the example that Google was using at the time. So I feel like it was fair judgment. But it makes sense if you have like a.

hairy topic or a boring topic, hearing two people talk about it back and forth is easier to consume and more enjoyable to consume than like sitting there and reading or even having an LM summarized for you. So I do see the value there. I think it's silly to call that a podcast, but I think it's probably the closest approximation to what people expect from a podcast, so I get it. And I agree that they are going to get better. My initial tastes were pretty sour. The way I think of it in my use case is not the replacement. It's the alternative. So when there's a lack of real-time content out there that's good on the subject.

Or I want to go into this particular paper because we wrote it. And I can't go back and read it. So I'll listen to it while I do the dishes. I'll just feed it the paper. Yeah. And these two AIs. have a discussion about it. And I listen to a lot.

So I guess in that case, it kind of is a replacement. But for me, it's like an alternative. I can't read it. But I kind of did. I would go and leave through it because I can't listen to it. I like to listen a lot. Like I'm a big audible user. So this is like right at my alley.

I'll do the dishes and listen to this versus that. And they don't give me the option to read it on the web. I can't just push a button and say read me this article. So it's the next best version of it. And it's been more like an alternative pattern. to the information, not a replacement of the original thing, to take its value away, or to, like, take somebody's lunch away.

You know, their web view. Yeah. Right. It's packaging. Yeah. I just think that the AI podcasters just need to be straightforward with what they are and who they are, you know, and don't try to be humans.

Just be like, yeah, we're two AIs to talk about this. And don't act as if you had lunch that day or that you have emotions about things, you know, because, like, that's just off-putting. but if you just be like, hey, I read this great article, tell me about it. And I haven't heard that version. I'm sure it's out there where it doesn't have any sort of personal story anecdotes. Just ditch the anecdotes and stuff and just like be who you are. I got no problem with AI generated content.

As long as I'm just straightforward that it's AI generated content and it's straightforward that it's, you know what I'm saying? There's just value there. It's the uncanny value of trying to act as if you are not that just confounds and offends. But yours didn't have any storytelling or anything about their lives and stuff. Like that's what I was listening to. I was like, no. Yeah. That's the part where I think it's even too sticky.

They, it's on repeat. So they will do, and these two, it's a guy and a girl in this case. Yes. They call the deep dive even. Welcome back to the deep dive where we're going deep. Oh my gosh. Our reader today has supplied us with a plethora of new information. It's like, I mean, excited.

They're very excited. They're very excited. I don't know how to take this stuff, man. I'm not sure where to where to categorize this and like, oh, no, this is terrible. This is super core. What the heck is this I'm listening to? I just not sure where this lands for me. I'm enjoying it, though.

I really am enjoying it. There you go. I guess that's what matters. Yeah. It's a new way. Some people will enjoy it. Other people won't and see if it makes its place in the world. Yeah.

I do think reading papers is often. often, often boring. And so if you can make that more approachable and accessible to more people, then you're going to help get that information out to more people. So there's value there. All right. What else? We've talked a lot about a lot of things.

Is there anything else that's exciting or scary out there or interesting in the world? I did have one article in this week's news that I thought tied into our conversation around when you brought up service providers. And I wasn't sure exactly what you meant by service providers, but the post was the future of software engineering is SRE. And his entire point, I think that was Swizzik Teller, his point was that as the creation of software becomes easier and easier and features and ideas and implementation get less important.

What gets more important is the ability to maintain, secure, and operate that way. which has been created over a long period of time. And I think that kind of dovetails into your point about service providers. And I wondered if I was picking up what you're putting down or if I was misreading what you were saying. Yeah, I think this is similar because I was saying, you know, you can have the product manager. And not saying that product managers can't or even product markers can't be developers. They totally can be.

But by and large, they're usually not. They're usually adjacent. And they're usually going to be vibe coding because they're not steeped in Swift or steeped in something or other. And they're certainly not going to be steeped in security. They may have security theory, but not be really that deep. Whereas the engineer is super steeped in it usually, especially in a shift left scenario. So I think the one you're referring to is the future of self-engineering is SRE. Yeah.

You know, reading this, I do agree with that because if you run your, if you, build some of these things for yourself. If you vibe code your own sasses, let's say, and you kill some stuff that is in your life, likely in today's world, 2026, you're going to be engineering or engineering adjacent, not like two or three steps Kevin Bacon away from software developer. Likely.

Yeah. And then you're going to be responsible for keeping that thing up or turning it into a reproducible product that you can conjure or call upon when it makes sense. Maybe you make an iOS app, maybe you do something else, maybe you make a web page.

Either way, you're going to be responsible for some version of uptime. So yeah, in that case, that does apply. What I meant by a service provider is this. There are businesses out there

that the large part of their secret sauce is somebody else's SaaS and their data trapped in somebody else's SaaS. And if, if they could just find a service provider, an engineer, a software developer,

to go and build their special thing for them. Well, now they take back control. Maybe they free up $15, $30,000 a year in would-be SaaS expenditures. And it's not because the SaaS isn't good enough. It's because it's not just in time and bespoke enough. I think there's a lot of use cases out there where there's people using SaaS because that's the only option.

It's not delivered in a box anymore. It's only delivered as a service via the wire, via HTTP. And not many of them are leveraging APIs. In fact, most don't even know what an API is. But I do, and you do, Jared. And what if there was a bunch of Adam and Jared's out there that's like, you know what? I'm going to go help my neighbor who runs this cool business that spends way too much on SaaS,

build their own thing and now have their own secret sauce and have that thing locked down. Then you get into the SRE land because you got to keep that thing up and you know this, right? Right. The Heroku days, right? You got to keep it up. You got fly. You got render. You got Roku.

You got railway. You got all these hosting providers. Or even if you got crazy, went out to Hester. Got your own box. Dialed it in. You know, put it in the cloud. Boom Shaka. You got your own box.

You got your own GP. You got your own CPU. You got your own RAM. You got your own disc. Whatever. And you know what you want. But the service provider, I think this coming future, there's a lot of value to capture in the place where you got people out there spending way too much on SaaS because they have

no other choice. And it's a dang shame because they're giving all that data to one, two or three service providers, SaaS providers that really don't care about them because they're not in the community. They're not bad Silicon Valley company or not bad, you know, a company in a different country even. Not saying that's bad, but they're not neighbors, like five miles, 10 miles away, impact my local community kind of neighbor. And I believe that if you're a software developer out there that is looking for place to place value, I would look there. What are the nearby communities and folks that are not visibly struggling, but they are spending a lot of money and maybe too much money for their own bottom.

line and in that way maybe struggling that you could be a good neighbor and a good service provider and build something cool for them. I think that could be the next big future. Yeah. Because I mean, how many more of us are way more enabled to build a video editor we would have never built or build tune or D, man, you know what I'm saying? And just do a little shell out, you know, a little shell out to, to Claude and Haiku and give me a one line or stream back and make it look cool. Right. You know, I mean, I'm using that. So, you know, in that case, that's cool for me. I think there's a lot of possibility in that front there.

And that's the area where I think maybe we'll shift. And I would just say that maybe SaaS, as we know, it doesn't die. It just changes. It's largely a web UI. They largely gate it. They say, well, if you want the API or you want a non-rate-limited API beyond a certain measure or access to your own data, call us because that's an enterprise ticket. And I'm cool with that.

That's how SaaS gets their value. I'm not anti that. I just think that the world is pushing back on that notion. Natural born SaaS killers. It's a good title, man. It's a good title. That's dirty to say, man. I want to tell you, audience, for those who can't see my face, I'm not, Jared's laugh a little bit.

I'm laughing. I'm smiling. I'm smirking. It feels a little dirty to talk about this. It really does. It's like a dirty word in a way. I feel like this is like taboo. Let's not do this.

We are. And we're not AI. We're real people. That's right. For now. And I'm still a plug puller at heart. So, yeah,

there's only so far I'm going to go. And then I'm going to pull that plug. Pull that plug, cut that cord. Right when they pick up weapons, you know, I'm just, I'm out.

Yeah. Hopefully it's not too late. Yeah. And if you are a sass out there, I think my advice to stop being such a gate on your API. If you want to be infrastructure, be infrastructure. Provide an amazing API with amazing uptime, great error codes, all that good stuff. Don't rate limit me when I hammered a little bit.

I'm developing on something. You just like suddenly make me stop because you can't handle it. Figure that out. Be my plumbing and be the layer and let me build some things on your things and have some data freedom. That's like figure that out. I think there's a world for that. I wonder if that's the I wonder if that's the way

to scale SaaS to the next level is freedom in the data, freedom in the API. We will see. All right. Thanks for hanging out with us and helping us work through all these things and listening along. Let us know.

If you agree or disagree or if you laughed or cried when we said various things, we hang out in Zulip, change log.com.com. Totally free. Totally cool. Totally cool.

And come converse with us. You can also email, of course. You can do whatever you like. And we listen. So that's all I got. Did you have any other topics? I'm kind of wrapping here without asking you. No, I think that's it.

That's it. I think we've said enough. If I stay anymore, I'll just get myself in trouble. Yeah. We did not want to get in trouble. Let's not do that. FCC. No, no, no.

Anthropic. We didn't say, we didn't say anything, okay? We didn't talk about Claudebot. That's not what happened here. There's some other bot. You chill. You chill over there. Don't cancel my account.

Okay, I keep my cloud code. All right, bye friends. Bye, friends. Okay, that's your change log for this week. Thanks for hanging with us. What do you think? Are the rumors of SAS's demise greatly exaggerated? Or is this a trend that's heading up and to the right?

Let us know in the comments. We love hearing from you. Thanks again to our partners at fly.io to breakmaster cylinder. to Tiger Data, namespace, and Squarespace for sponsoring ChangeLog this week, and thanks to you. We appreciate you. Next week on the pod, news on Monday, Tushar Jan from Docker, talking hard on images on Wednesday, and Mel Hussein here on ChangeLogging Friends on Friday. Have a great weekend.

Call someone you haven't talked to in a while. It's always worth it. And let's talk again real soon.