Caring in the age of AI

All right. Far too kind. Thank you, Chris. Um, so I wanna tell you a little bit of a, a history story. It's a little bit, it's a little bit grim. It is real. You can look it up later. In the sixteen, seventeen hundreds, um, in England and Wales, they had this belief, they had this tradition, um, whereby when somebody died, their sins could be taken away by somebody else.

And the way that they believed that this worked is they would bring someone in, usually from the outskirts of town, they would bring someone in, and they would eat a meal over the body of the deceased. Pretty grim. They don't do it anymore. What they believed was by doing this, those sins of the deceased would go to the food, and then that person would eat those sins and ensure that person, the deceased, a happy afterlife.

This is called a sin eater. This is a real thing you can look up. This is a sin eater. This is what we are supposed to do, okay? We are supposed to be, as product people, as developers, as entrepreneurs, we are supposed to do that. Not the sins, but the pain.

We are supposed to eat the pain of our customers. We are supposed to absorb the pain from our customers. Every product has pain somewhere. Every product has some kind of pain, and you can't just, you can't just get rid of it, you have to shift it. You have to shift it to yourself or to your customer.

That is the only, that is the only question, is who is going to experience the pain? And my challenge to you today is you need to be the person, you need to be the company, you need to be the product that eats the most pain on behalf of your customers. Okay? This is the thesis of my talk, which is Be the Sin Eater.

It's very memorable. It's very grim. It paints a picture in your head. Your job is to eat the pain for your customers. My name is Aaron Francis. I'm the VP of Marketing and Community, and I want you... at Laravel, and I want you to raise your hand if you've ever had an email that you were waiting on not show up.

You were waiting on a shipping receipt, you were waiting on a password reset, you were waiting on something, and it didn't show up. Exactly. Now, I want you to raise your hand if you know, without looking it up, what DKIM stands for. Yeah. Fewer, fewer people. There's pain, right? SPF, DKIM, DMARC, return path, IP reputation.

Turns out, n-not five minutes ago, they got on stage and said, "Hey, we're gonna eat more pain." As Resend, they've eaten this pain already, right? They've, they've looked at this and they've said, "This is brutal. We're not gonna make our customers experience it. We will experience it. We will absorb all of that pain and suffering so that our customers don't have to." Do you know that there are still email engines that render with the Microsoft

Word rendering engine? Painful, right? You want your button to look beautiful? Did you use little bitty tables that can render in Microsoft Word? Painful. Resend has eaten that pain. They've looked at this and said, "I am not willing to put this onto my customer." Where there is pain, there is opportunity, okay?

This is what Resend has done, and I was standing backstage earlier, and they were talking about Forge, and I was like, "This goes in my talk. This goes in my talk." We now own... Not we, I don't work there. Resend now owns their own infrastructure, right? Absorbing more and more pain from their customers so that we never have to feel it.

I'm a customer of Resend. I don't feel that pain. I am certain that they do. I do not wanna run email infrastructure. I do not wanna make a mistake and get my domain blackholed forever and have to call Spamhaus. I don't even know who they are or how to get ahold of them. Resend has said, "We'll eat it." The pain didn't go away, it just shifted to somebody else.

Uh, in my, in my career, uh, I've been a marketer for three months and a developer for twenty-five years, so, uh, most of what I'm gonna talk about is, is development. In my career, I've done a ton of database work, and you know what's terrifying? Huge database tables that need to be migrated. And if you've never experienced this pain, praise be, okay?

It's brutal, right? You've got a hundred million, five hundred million rows in the table, and you think, "I wanna alter this table to add a new column," right? This is something we're all scared of, and this is something you should be scared of. When you run an alter table on certain engines, it locks the whole table so that it can add the column, the index, drop the column, whatever, right?

Some of the engines are getting better at this, but this is still terrifying. This is still terrifying because when you accidentally lock your table, your five hundreds start blowing up. Your 503s start blowing up. Everything is piled up behind this exclusive lock. PlanetScale looked at that and said, "We'll eat it. It sucks, and we will absorb that pain on behalf of our customers." And they said, "Not only will we absorb that pain, we're gonna go a step further, and we're gonna make that migration undoable." Why

not? You make the migration, it's safe, no downtime. You realize, "That was actually not the right migration. Let me roll that back." No problem. The pain exists. The question is, who's gonna bear it? Recent PlanetScale have decided we will bear that pain. My contention is where there is pain, there is commercial opportunity, right?

If there is pain there, someone will pay you to eat the pain because they don't want to. They want to be absolved from that pain. They want to enter their life happy and free while you stay behind full of pain. Pretty optimistic pitch, honestly, okay? So I wanna talk about, I wanna talk about another one that is slightly different, um, and we'll, we'll come back to this one.

Nobody needs, uh, nobody needs domain search to be fast, okay? If you're a healthy, well-adjusted individual, you buy domains once or twice a year. I'm not gonna, I'm not gonna ask for a show of hands, okay? So if you buy domains a couple, a dozen, two dozen times a year, eight hundred milliseconds, one point two seconds, not that bad, right?

Somebody at Vercel looked at it and said, "I'm gonna eat it. I want it to be better. I feel a small amount of pain here. I'm doing a lot of domain searches while I'm building this product. Eight hundred milliseconds is a long time." It's not, but they felt the pain, and they said, "Instead of, uh, imposing this pain on our many tens of thousands of customers, I'm going to eat it.

I'm going to absorb it myself, and I'm going to bring delight and a pain-free life to my customers." This is what I was gonna talk about for Laravel. I was gonna talk about, um, kinda Laravel the open source framework. So Laravel, uh, we're the makers of a PHP framework. We're also the makers of Laravel Cloud, which is commercial hosting.

This is what I was gonna talk about. This is still true, right? This is still true. For fifteen years, we have absorbed pain on behalf of developers. If you've ever glued together your own stack and decided, "Well, what auth do I want? What ORM do I want? What queues do I want? What mailer do I want? What, uh, file storage do I want? Now, how do I put those together?

Shoot. I don't know. That sucks." Laravel has been eating that pain for fifteen years. Okay, so I did talk about it, but what I want to talk about now is Laravel Cloud. I was up last night until twelve thirty. I got up this morning at four thirty, not because I was nervous about this talk. This talk has been in the bag.

Why I was up late and up early, the worst, the worst PHP vulnerability in all of history dropped this morning. Today, we've got our VP of, uh, product is at Cloudflare speaking. I am here speaking. We've got a board meeting tomorrow. We've got salespeople in town.

We've got DevRels in town. It happened this morning, and we had no insight into what was about to happen. OpenAI and the PHP Foundation knew. The rest of us were screwed. So what happened was last night... I really do love our team. Um, last night, our whole team, infrastructure, marketing, app developers, everybody

stayed up all night absorbing pain, eating the sins of PHP, of which there are a few. They were up all night. They're still going right now. I just checked Slack in the back. They're still going. We decided this is our responsibility. We could say to our customers, "You gotta watch these releases on php.net.

Good luck figuring it out." Or we could say, "We're gonna set up our playbooks. We're gonna run smoke tests in our dev regions. We're gonna get our messaging ready. We're gonna get our emails ready. We're gonna do as much back channeling to the foundation as we can to get even a scrap of information about what's going to happen." And we decided, you know what? We could tell customers, "Hey, the images are patched.

Go redeploy." And we decided, forget that. We're gonna patch it without them ever knowing. We're gonna do it... We're gonna do it live. We're just gonna freaking do it live. We're gonna patch it on behalf of everyone. We decided, we stared into the abyss and decided this is our problem, not our customer's problem. We chewed glass and walked backwards into hell so that our customers could wake up and have a good day.

This is eating pain. Every great product is someone choosing to eat pain. Not every product. Not every product. Every great product is someone choosing consciously, "I'm going to absorb this." The people should not have to feel this pain. I wanna talk about, um, I wanna talk about craft just a little bit.

Um, you can go into IKEA right now, and you can get a bookcase for seventy-nine dollars. This is not a denigration of, of IKEA's furniture. This is a modern marvel of manufacturing. This is a modern marvel of capitalism, that we can walk into an IKEA right now and get pretty good furniture.

For your twenties, early thirties, it's all I lived on was IKEA furniture, and it was pretty good. Three hundred years ago, not possible, right? Mass production gave everybody furniture. That's great. This still exists, right? This still exists. Somebody named Keeta made this one. Low, walnut, tile inlay.

Nobody needs that. Right? This part, nobody needs. Doesn't hold the books better, doesn't take away any pain. But you look at that and you know there was a human being with a heart that made that, right? Somebody cared enough, somebody paid the cost to cut that out and put that inlay in.

This is the same thing as eating, eating the sins, eating the pain. You are deciding, "I will bear the burden. I will pay the cost of this so that my customers have a more delightful experience and a more delightful life." And so you may be thinking, "Brother, we're talking about B2B SaaS.

Put the SaaS in the bag. This is a Wendy's, okay?" I get it. Like, I, I, I get the feeling. I get the feeling that like, what are we, what are we doing? This guy's crying on stage about software. Like, what the hell is wrong with this guy? I wanna tell you another story about...

This is, this is a tough one for me to get through because I love him so dearly. There was a monk 350 years ago. His name was Brother Lawrence, and he lived his whole life in the monastery. And Brother Lawrence's job in the monastery was to do the dishes. That's it. That was Brother Lawrence's entire life.

"I will serve my brothers by washing the dishes," okay? So depending on what you think of B2B SaaS, maybe washing dishes, right? Like maybe not, "I'm gonna go off and defend my homeland," or, "I'm gonna, I'm gonna do something brave and noble." Brother Lawrence washed the dishes, and you know what he said about washing the dishes at the end of his life?

"We ought not to be weary of doing little things because it's not the greatness of the work, but the love with which it is performed." Your job may be in B2B SaaS. Do you want to finish your career and look back and say, "It wasn't the greatness of the work.

We hosted applications, we sent emails, we did databases." Sure, maybe that is great. "It wasn't the greatness of the work, but it was the care and the craft, and it was the love with which I did my noble duty." Right? This is your life. Do you want to end your career and say, "Ah, I did as little as I could, and I got by, and it was all right"? Or do you wanna end your life and look back and say, "The love with which I performed my job"?

That's what we're talking about here, okay? Mass production made quality evident. It is not bad to have low-quality software. It is not bad to have average-quality furniture. What is true is that if you walked into an Ikea and you saw Kida's bookcase, you would say, "That's different." You would see it from across the roo-- You'd see it from the parking lot.

This bookcase is different. This was built with love and care and craft. So the fact that software is having its Ikea moment right now may be depressing. You may look and say, "Everybody's ripping out crap all the time." That's fine. Rip out as much personal software, rip out as much whimsical software as, as you want.

There is still room to put your heart and soul, your care and craft into your product, and it becomes way more obvious. It becomes way more obvious when you're standing out. When you say, "I'm gonna own email-sending infrastructure," it makes the vibe coders that are like, "I'm gonna rip off Beehiiv." It just makes it so obvious who cares about what they're doing.

I don't wanna own email, and I don't want a vibe coder to own email. I want somebody who freaking cares about it, right? So my question is, what is, what is the thing for which you will pay? What is the tile that you will inlay? What is the thing that will delight your customers that is maybe not entirely necessary? Maybe it's 50-millisecond domain searches instead of 1,200 milliseconds.

What is the thing that you can pour yourself into? Because when everybody says, "I c- I cloned Linear in a weekend," what's left for us, right? If you can clone it in a weekend, so can everybody else, right? And that's okay. Like, if you wanna rip personal software, have fun. It's the golden era of making little things.

I'm happy with that. If you wanna build a product and a company that you're proud of, you have to do something else. You can't just say, "I'm gonna rip off somebody, and I'm not even gonna... I'm gonna tell the agent, 'Go. I'll be back after lunch,' and then come back and see what I did." So what do we do? Like, practically, what do we do? The hardest thing.

Do the literal hardest thing possible. Own your email infrastructure. This worked out great, this announcement, this timing. I just g- keep using the example. Do the hardest thing possible. Don't do things in the hardest way, but find the gnarliest, ugliest, most complicated problem that keeps your customers up at night or your potential customers, your target market, find the thing that they're scared of and say, "I will eat it.

I will chew that glass." And there, in that chewing of the glass, is economic value. You set yourself apart. You do the hardest thing because the floor has moved up, right? It used to be being able to write code, and I know this because I used to use my hands to type code. I know this.

It used to be that that was a superpower, and it made me feel awesome. It's like, "I'm a developer. I can will things into existence." So can everybody now. The floor has gone way up. So if you lived right above the floor in the good old days- You're toast. There's still room up here. Nobody cares. Nobody cares.

If you care, you can stand out pretty well and potentially make a lot of money. Both things are good, right? So, what does this mean in terms of, um, what does this mean in terms of building your product? I think about it like this. I want you to increase your vertical ambition. Everybody's like, "You need to be-- You need to increase your ambition." Okay, which direction?

Vertical or horizontal? You see a lot of, um, people on X.com, the everything app. You see a lot of people out there saying, "I'm-- I can-- I'm gonna be more ambitious. I'm gonna make 15 products." Oh, so close. So wrong. The exact wrong takeaway, right? 'Cause if you build 15 shallow products, you don't care about them, nobody else cares about them.

You fool yourself into thinking you've been ambitious. What I want you to do is increase your vertical ambition. I want you to continue to mine the depths of your customer's pain and go lower and lower until you say, "This is literally the most painful thing I could have ever built, and now it's done." So, don't build 15 products. In fact, at Laravel, we're starting to turn off signups for old products that Taylor built a long time ago.

We are increasing our vertical ambition on Laravel Cloud. This is our vertical ambition play. We are doing as much as possible to make sure that the Laravel Cloud experience takes away all of the pain of developing and deploying applications. I do have a s- a minor caveat.

I'm not talking about breadth, right? Breadth is fine, shallow is the problem. And an interesting thing is when you, when you increase your vertical ambition, some of that breadth comes for free. So, I'll tell you one, uh, one big co example, which is Laravel Cloud, and then one indie hacker example, which is me.

Laravel Cloud, uh, which this may surprise you, historically, we've supported Laravel applications. It's right there. Thank you. Thank you for that. That was a joke. Um, and we've gone as hard as possible on making that experience delightful, down to connecting the database, up to front-end caches, everything in the middle, right?

We've gone as hard as we possibly can. When you flip on the Resend integration in Laravel Cloud and you don't have the Resend package installed in your application, we warn you. We don't wait for you to deploy, try to send an email, throw a PHP error at you. In the UI, you turn on, "I want email," and it says, "Oops. Hey, we just scanned your dependencies.

You don't have, you don't have the email dependency. Do you want it?" We've gone as deep as possible on the experience. Now, we're realizing, wow, we've built this huge cathedral. What if we hosted Ruby on Rails as well with that same depth, that same care? I'm not talking about shooting out 15 different projects.

I'm saying, what, what breadth does this depth allow us to do? Rails, Python, Rust, 'cause everyone's writing Rust for some reason, Node. And so we are increasing our breadth, but we're not going shallow. Every, every single framework that we support, we have hired someone to say, "What do you, a community expert, expect to see here?" Not three Docker files and a trench

coat, right? Not like, "Hey, it can work." No, it is built specifically for you. This is the thing where we'll scan your dependencies, and honestly, most people will never know. Ninety-eight percent of people, when they're like, "Oh, I wanna add email to my app," they add the package, and then they turn it on. The two percent are gonna think, "Ah, a human was here.

A human, a human went through this flow, and they made this better for me." That's big co. What about your own projects? Everybody's got a project. Come on. What about your own projects, right? If you wanna rip out 30 things that you never maintain because it makes you happy, do it.

That's fantastic. If you wanna build whimsical software because it's silly and it g- it makes you smile, do it. That aside, if you're trying to, if you're trying to build something that is commercial and makes you money, you have to go deep. I write a desktop app, um, that is a terminal multiplexer. It's like your Agent Hub terminal multiplexer.

I've been working on it for about a year. Early on, I decided to do the hardest thing, which was I'm going to write my own terminal, like an idiot, okay? Not a great idea, but it was... I was experiencing pain with xterm.js, which is what everyone pulls off the shelf, and they're like, "Aha, I have a terminal." Yeah, true, but you quickly run into the edge cases there.

So what I did was I wrote my own terminal. Everything lives in Rust on the back end, and the front end is a thin client. This is what I'm talking about when I say depth gives you a little bit of breadth for free because now mobile apps, thin client, peer-to-peer desktop apps, thin client, all of the hard terminal state is stored in one place, and I'm getting this breadth for free.

It's just like, "Oh, we'll just implement this protocol, and now I have a terminal on my phone." It's great. So the depth compounds into breadth. So I encourage you, find the depth, go deep, find that customer's pain in the pit of hell and eat it. Now, back to Vercel, back to domain search.

Mitchell Hashimoto says, "The pursuit of excellence does not need justification." People may be confused by you pursuing excellence. You owe them no explanation. You can try to win them over, but you owe them no explanation. If you wanna make something fast, if you wanna make something beautiful, if you wanna make something performant, do it.

You do not have to explain yourself. This is pure, noble, and right to make beautiful things and put it into the world, to make performant things. If you're an infrastructure engineer and you don't care about product, you don't care about marketing, you don't care about UI, it is still pure, noble, and right to pursue excellence in infrastructure, and you don't owe anybody an explanation.

Last one. This building is the first freestanding building built specifically for jazz. In the US, the first one built specifically for jazz. None of us knew that. Somebody knew that. Somebody said, "I am going to put my heart and my soul into architecture that is built for jazz because I am on the pursuit of excellence." And

in fact, this is a great work, but it doesn't matter. The love that went into this work is what matters. The person who built this gets to finish their career, perhaps finish their life, and say, "I put myself into my work. Well done, me." And that is what we want. We wanna end our careers by saying, "I put myself...

I gave my best work, and I am proud of the work that I've done. I absorbed the customer's pain. I chewed glass and walked backwards into hell because I wanted to make other people's lives better." That is what we're after. That is what care and craft in the age of AI is.

This is my, this is my challenge to you. Do not let the doomers get you down. Software is not dead. Software without a heart and soul is a lot cheaper, but software that has your human fingerprints on it is as alive as it's ever been. Thank you.