# AI Agents Reshape Developer Platforms and Demand

**Podcast:** alphalist.CTO Podcast - For CTOs and Technical Leaders
**Published:** 2026-08-13

## Transcript

Hello friends, this is the Alphalist podcast.
I am your host, Tobi.
The goal of the Alphalist podcast is to empower CTOs with the info and insight they need to make the best decisions for their company.
We do this by hosting top thought leaders and picking their brains for insights into technical leadership and tech trends.
If you believe in the power of accumulated knowledge to accelerate growth, make sure to subscribe to this podcast.
Plus, if you're an experienced CTO, you will love the discussion happening in our Slack space where over 600 CTOs are sharing insights or visit one of our events.
Just go to alphalist.com to apply.
Welcome to the Alphalist podcast.
I'm your host, Tobi.
And my guest today started out as a soldier in the U.S.
Army in the late 90s, wrecking Unix boxes and writing the military's internal applications.
And went on.
to run engineering at GitHub before becoming the CDO of Netlify, the company that hosts a big chunk of the internet you use every day.
So that's it.
Welcome to the show, Dana.
Really happy to have you here.
Yeah, thanks for having me.
I'm excited about this conversation.
I mean, you know, it never gets old talking about AI, Toby.
I tell you what, okay, maybe it gets old.
But today we're going to make it new.
Let's make it new.
So maybe before we dig deeper on the trenches of AI, I'd like to start with my regular startup question.
What's your nerd path?
When did computers first feel like magic?
And why?
Why did you catch fire early days?
I was born with a keyboard coming out the womb.
No, I wasn't.
I mean, I...
I never, ever, ever dreamed of this.
Like, you know, more power to the people that did come out of the wind with keyboard.
I wasn't one of them.
I kind of stumbled on it.
I always really loved math, but my passion was art.
And I was like, you know, I'm going to go be a famous artist.
You know, that's a hard gig to get into.
You got to know some people.
I didn't know anybody.
Well, I didn't know how to party.
And so art school didn't work out for me because of my love of partying.
And, you know, on a whim, and people are like, seriously, on a whim, I joined the U.S.
Army.
You said it earlier.
And this was in the late 90s when the internet was really becoming a thing.
We didn't have this concept.
It was like right at AOL dial-up when people were getting CDs.
In fact, I came in a little bit before that, so I'm aging myself.
But honestly, what bit me was when you join the Army, and you don't really think about joining the Army, you have to be smart and say, well, what could I do in the Army?
And I was like, computers need air conditioning, Toby.
So my bite of the computer bug.
was based on my comfort, which I wish I had it in Europe right now.
It's hot over there.
But I've always loved math and I love tinkering.
And to me, like this translation of going from art, which is you can create anything with a paintbrush.
Code is no different, you know?
And like for somebody that's creative, like it feels like I'm talking out of both sides of my mouth.
There's never been a better time to be creative on the internet because you can do anything.
You can absolutely do anything.
And it's...
It's pretty amazing.
So that was my first bite.
I've been doing it ever since then.
Thousands and thousands of euros for AWS every month.
You know that feeling.
You're building fast.
New services, new instances, new databases.
Everything grows.
And so does your AWS cloud bill.
That is where Blox comes in.
Blox guarantees at least 20% less on AWS.
No long-term contract, no infrastructure changes.
How do they do that?
big group buying discounts and AI optimization.
I have known Blox co-founder Andreas for more than 20 years.
I trust these guys.
They will help you getting cloud costs under control.
So if you use AWS and want to know how much you can save, run the free cloud check at blocks.cloud slash alphalist.
That's B-L-O-C-K-S dot cloud slash alphalist.
I guess then you really have a great time these days, right?
Where with AI, you literally can create anything.
But we're chatting about that later on.
So what did you do in the Army then?
Like, what did you, like, if you're able to talk about it?
I mean, it's a bit dated, but...
Definitely talk about it.
Well, you know, you give somebody that doesn't have, has never done anything like this.
I mean, what did we have back then?
Microsoft front page.
I did some of that in high school.
So like...
I was not one to be trusted, but in the U.S.
military, when this field was emerging, my specific title was Information Systems Operator Analysis Developer.
It's very long.
It's a very long title because I'm like, I don't know, they're figuring out computers.
But what I actually did was I managed the Department of Defense, what we called back then installation support modules.
All it was is a fancy name for software that we built and deployed to a private network to manage.
the movement and the records and all the processing of troops globally.
And it was really awesome.
I like to say, you know, I've been in SAS from the beginning.
This was really the very junior leagues.
So my first job was changing backup tapes.
Like for anybody that remembers, there were physical devices that you would have to go back up to.
There wasn't automation.
I was a human with a cart that went there once a day and put tapes in.
So you were before those robots.
There were no robots.
There were those tape robots, right?
And then I started, I was so lazy.
You already heard me like join military for air conditioning.
Then they're like, girl, you can change passwords.
I was like, hell yeah.
And you know, this is the military password changing, not easy back in the nineties.
So I automated it.
And that's, that's really when it bit, I was like, wait a minute, I can use code to make me even more lazy.
So I've been on this DevOps path from the nineties because I'm like, Why waste your time doing stuff that a computer can do better?
And here we are.
And so, you know, it was really just a lot of curiosity and growing and driving through the foundation that the U.S.
military actually gave me.
We had to learn from like routing to Windows.
And I remember when Windows came out, Windows NT and Windows Server, we were like, oh my gosh, so cool.
We can move.
It's not cool.
Okay.
People don't even know how that stuff works anymore.
Except a couple of nerds out there.
But here I am today, man.
And then you managed engineering at GitHub, which also is quite a step, right?
Yeah, I don't know if GitHub's having regrets still from this, but I love you, GitHub.
No, it was awesome.
So at the time, oh man, I didn't even know about it, but GitHub was going through an acquisition with Microsoft.
And I had met Jason Warner, who is awesome.
And now Jason Warner, the Jason Warner term.
And like we were at a conference or something and we were on a panel and I was, you know, running my mouth as I do.
And he was like, you should come work at GitHub.
And I was like, yeah, maybe I should.
I don't know.
I've been doing this startup jam for a minute.
Maybe I should go see because y'all are going on to this next chapter of like big business.
And I was like, man, I don't know no big business as you can see, but maybe I do.
And it was just like, for me personally, an opportunity to go and learn something at scale.
Like I've done growth, I've done startups, I've done IPOs, M&As, but working for like a hyperscaler, I never did that.
And it was a wild journey and it was an incredible ride when I was there.
I was the VP of product engineering and then Sam Lambert, who's now the CEO of...
PlanetScale was the VP of platform.
And then we just kept cell dividing GitHub to where it is today.
And like, we were just a little part of that narrative, but it was a really good time.
And, you know, being the home for all developers, it was just a natural translation to go from that type of trying to build tools and workflows for any kind of tech.
So really focusing on Jamstack and front-end development, which is really interesting because like I am...
until agents arrived, I was a terrible front-end developer.
You got to have some patience for CSS and you got to have some taste.
I'm just like, no way, man.
I'm like, Claude, you can CSS all day.
I'm a platform engineer.
I'm a platform engineer.
But as a platform engineer, you want somebody that thinks about automation.
I've spent decades thinking about...
How do we take out all that noise just so we can get good ideas?
Or maybe not great ideas, but fun ideas out there.
Great.
And how do you think, looking back, looking at the world now, how will GitHub look like in a few years?
I mean, software development lifecycle is changing quite a lot, right?
Will we still use it?
mostly power the platform for our PRs for codecs clarified?
Or how will it look like in a few years?
Do you have an opinion on that?
I mean, when I was at GitHub, we still had to have sticky notes for Git commands.
I'm going to tell you what, we're hardcore developers.
And it's like, hold on, let me mine the think tank to remember how to do this.
Ain't nobody got time for that today.
And I'm going to say, why should you?
Why should you have to think about that kind of stuff?
Now, does that mean that the problem of version control and rollbacks is going anywhere?
Hell no, it's gone worse.
Like if anything, I'm just like, oh Lord.
But I think that it's become really a commodity.
And I think GitHub still serves a huge purpose for being the home for all open source.
We still need a playground to share.
and host and curate ideas beyond ourself.
I just feel like the entry point's going to be different, you know, because if you think about, there's really this emergence that's happening and there's not one side to rule them all.
And I think that there's going to be problems that need to reach in and still have some human creativity behind the architecture and the greenfield footprint.
But I'm going to tell you, That's going to become rarer and rarer.
And most of us, self-included, are just going to use an abstraction because most things that took that idea to reality have now become something that just exists and is expected.
And it's expected.
Like, man, people expected.
They're just like, why didn't this work?
And I'm like, well, DNS has to propagate.
Like, try and explain DNS to somebody that doesn't know DNS is still a challenge.
And so, like, I think that...
There's going to be a need for social collaboration to democratize the web, to keep it open in a place to have those conversations and share that.
But I don't know if it's not going to be the same people, it's not going to be the same users, and it's going to have to get a lot easier.
And that's what society is demanding right now.
Maybe same users, but all users equipped with agents who then, I don't know, like...
Yes.
Translate between them where no one really understands the real content, right?
Like what the agents post.
I know, I talk for myself.
Sometimes I'm contributing to open source where I feel like shit when I really use Claude to post a PR and then have absolutely, or let's say very little clue what...
actually was changed and how exactly um but this is why i still feel like until the gap is solved and the trust is that right large language models like we believe they get smarter right like i hope we do i mean they should be right the concept is you keep training it it's going to get smarter on that decision making We're not putting a cap on the training.
We're not saying like, oh no, reserve that for humans to think about.
We're not doing that.
And so unless you believe that there's a ceiling for artificial intelligence, you have to have a belief that someday it's going to get good at that and you're going to have confidence.
And there's this question of where do you need the expertise?
Like, okay, let me ask you this, Toby.
Do y'all have Teslas where you live?
Teslas.
Yeah, Teslas.
Self-driving cars.
Absolutely.
Yes.
Not self-driving, unfortunately.
Do you operate one?
I would say yes.
You absolutely have no way and understanding of how any of that shit works.
But yet you get in it every day and you drive it.
How is this any different?
You just trust that car not to run into a wall.
There is no difference.
There's absolutely no difference.
It's just when does that trust become defaulted?
And we're in this awkward, weird time.
Like, I don't trust any of those models at all right now either.
I kind of trust them.
I still am like, thank God I've been doing this so long because I know it's under the hood.
But my job and our job as technologists is to make it safe.
Isn't that what open source has always been about?
Sharing ideas in a safe way with expertise around the world to go and balance and check.
And yeah, maybe some of those experts are going to be agents.
But that's, again, unless you're using agents that you've put a ceiling on for learning, then yeah, I wouldn't trust them either.
But this comes back down to like, are we being short-sighted on where we see the future?
Because, you know, previous to all of these advancements, like we weren't doing any of this, but we're doing it now.
And I always point to these self-driving cars.
I'm like, half y'all ain't all car mechanic.
Developers come to me.
I'm like, are you a mechanic?
You get a car every day.
You operate a vehicle.
No.
All right, then.
What is the difference?
That's a good comparison.
That's a good comparison.
And yeah, I mean, there's a lot of debate also about like Javon's paradox, right?
And how this is actually changing or if demand is actually changing or exploding, if a price of a resource becomes very low.
And right now it looks like that, right?
While, I don't know, I sometimes feel a bit strange when someone tells me about a new app that he or she has built, right?
Like, and wants to install that on my computer.
Not because it doesn't feel safe, but because sometimes my mental capacity doesn't seem to be made for that, right?
Yeah.
I have my macOS and it's fine.
I use my browser.
I use my email.
Sometimes I use this or that tool, but maybe I don't just need more.
Maybe I just don't need what you have just built.
Yeah.
And I think that's like, you know, again, it's one of the challenges, right?
When you build software is people vote with their feet and determine what becomes commoditized.
And eventually you can be a holdout.
But if the rest of society have made that decision, you eventually have to get in.
Or you just quit participating.
And it's really difficult to do that in technology, right?
I think that you become a ludidite and it's like, cool.
I hope you have the infrastructure to support that.
I mean, some of the basic skills of survival, we've lost as humanity.
So I'm like, y'all are going to have to trust technology a little bit.
But this is the challenge, right?
Of having that expertise, being uncertain, unknown.
And like us as developers and engineers, like...
Our whole persona has been like, protect production, do no harm, be skeptical.
Maybe even a little cynical because you got to be like, really?
But that's how you have good guardrails and boundaries.
But I feel like, you know, you have to ask those questions.
Do you need it or don't you need it?
And we are going to hit that limit.
Like somebody was asking me, they're like.
now that agents can like do more is the expectation that our 40 hours go to 80 hours and that we should be just like harnessing and orchestrating these agents all day because like we can now.
And I'm like, isn't AI here to serve us?
I don't know, man, maybe call me old school, but I'm with you a little bit.
I'm like, what do you need?
And what do you don't need?
Like really, what are you trying to do besides live at this point?
So, you know, just because you can build it doesn't mean you should.
But that doesn't mean you should stop being able to allow anybody to build it because it's kind of like the needle in the haystack.
If everybody can build, man, there's going to be some gold up in there.
But when we continue to gatekeep, when we continue to gatekeep, then we may be missing out on something awesome.
So, you know, it's tough.
I'm also not, I don't think that we should gatekeep.
I just feel that, yeah, a day has 24 hours and there's so much capacity, only so much capacity to create really, right?
If you really want to, if you really think about a problem deeply, then often your mental capacity is exhausted after a while, even though you can hit fast forward.
That helps quite a lot.
But yeah, I don't think, or so far, I don't see like really great products popping out of the blue.
Yeah.
And I think this is like, from my perspective, what really changes, right?
The constraint has moved like really forward and is basically moving way forward.
Like even product, I think it's not really what slows us down now, but ultimately the demand for what we're creating, right?
And really working into that demand is challenging and will become more challenging with models changing, with Google changing, with clicks going away, with websites going away, whatever.
Let's see, right?
Let's see what the world has for us.
I agree.
And it's interesting, too, because I think that that's where people are safe.
I mean, a lot of it is just random acts of slop, which is cool.
I'm just like, okay, but we're killing the environment.
So maybe we should like...
I don't know, think about the GPUs we're doing for Random Acts.
But on the other hand, this is where I think like, you shouldn't be that scared.
There's going to be a lot of builders, but a lot of their stuff is going to be bullshit.
The stuff that people need in the market value, people that care deeply and have real solutions, they're going to come out on the other side.
My hope is just they can be empowered to rise above because their idea is so much better that we don't have to deal with crappy.
And it's the same thing.
It's like when you get two competitors, you're like, oh man, these two offerings are so similar.
are you going to go with one on price or is it because of your likability?
And it's like, oh man, it depends.
It's like, I like this one enough because the price is right.
But if you really have a great product-minded sensibility and the demands there, let's give you the tools to go not win on price, go win on being the best.
And I don't know, man.
I think the real challenge is trying to find the pieces that matter.
I mean, I've been doing this silly little game and I'm running out of ideas.
I'm not really, where I was like, I'm going to build 30 apps in 30 days.
Well, I would tell you, I'll call it all the apps that maybe like two are actually useful, but I was able to do it.
And I'm just like, oh my goodness gracious.
If I did have a good idea, I'd be unstoppable.
Luckily I'm a CTO and I leave that work for the creatives.
Luckily, yeah.
Recently, you released an article on Newstack that was, I think, received quite a few comments on Hacker News.
That was writing or title was, writing code is no longer the job, which is, I would say, arguably true.
Why did you get so much backlash?
And was there like, what did the angriest, commentary of it get right.
Hacker news, hackers don't care.
Like, oh my gosh, talk about an echo chamber of somewhat dev toxicity.
I'll say it out loud, put it on your podcast because I believe it.
Do I have to like, I don't know, play the game like I care because I actually don't.
But I do see it as a forum for people that have an opinion to come share it.
And here's my two cents, right?
We can all pretend that AI is creating jobs in the future.
Every technologist.
And every big hot tech CEO, CTO keeps saying, don't worry, don't worry.
Thousands of jobs.
I'll tell you, that is easy to say.
But apart from end traffic, right?
I would say that's easy to say because that's a theory because it's not in reality.
But reality is the IDC showed that 95,000 developers lost their job in the first five months of this year.
That actually happened.
Okay.
Those are facts.
That's a thesis.
People are getting upset because they don't want to believe that the thing that they worked really, really, really hard for is not going to be accessible and done by other people.
And a belief that they are still more equipped and able to do it better than an agent or a machine.
And like I said, I hope I'm wrong.
Let me just say this.
I hope...
And I am so freaking wrong.
I really do.
I do not want this to be true.
But the reality is just like any other technology.
We use the car as an example.
We can go and cast this with any other advancements where people say, oh, no, it's not going to happen.
Where are newspapers?
When was the last time you bought a newspaper?
When was the last time you saw somebody pay a lot of money for somebody to sit and edit something by hand?
All of these things where we build technology and advancements to take away workflows, there is no difference to this happening to the art and the craft of development.
It's becoming automated.
And it is.
And so people that want to argue with me, argue with me.
Like, prove me wrong.
Like, I don't know.
Like, I sure as shit don't know everything.
I'm just trying to base it in facts here by looking at history over time.
An example, another example for you, okay?
I don't know if y'all have a problem over there, but every damn kid in the U.S.
has an electric bike now.
Shouldn't we force them to ride non-electric bikes because it's better for them?
Hell no, there will never be a kid buying.
They will always ask now into the future for an electric bike.
Okay, like we can't ride manual bikes, but nobody's going to.
We can write code manually, but you know it's going to be faster letting an agent do it for you.
And what people want is the demand for the experience to come quicker.
And it's always been time to market and market demand.
It's not about what you build.
It's about how it's used once you build it.
And like, I love the craft isn't like anybody else.
Like I do.
But again, like unless you have this belief that we're going to cap agents and artificial intelligence from learning.
Yeah, jobs will be protected.
You can go be a developer.
But I'm going to tell you why create harder problems.
for things that don't exist.
This is when I was like going from GitHub.
I'm like, man, heck is complex.
Like you got Kafka, Kubernetes, you have message queue, you got all these inner systems.
The internet is mostly just websites and apps with data stores.
And yeah, there's hardcore needs for some laws of physics to be manipulated.
But like we're reaching that limit because processes can only go so fast.
Memory is only that big.
And we are actually going to hit a part where it's like, cool, where are the new vectors for experiences?
So I don't know.
I didn't see any of the spiciest comments.
I might have to go and look what those, what they were hating on me, but I don't like hate on me.
Like, again, like I'm just another talking head that's trying to balance out all the voices too, because it's like, nobody knows where we're going.
Nobody does.
It's like, nobody can predict it unless somebody has figured out time travel.
Let's see if you do, right?
We give it a try later on.
Love it.
But I'd also say that a lot of the thinking about craftsmanship and software is also a lie, right?
Like if I look at many folks working on...
or that used to work on translating, I don't know, from a single page web app to some GraphQL API to then store it with some data store and some, like, I don't know, thing that was invented like 10 times to a database, to Postgres.
Ah, now it's Postgres, no longer MySQL.
Like, I'd say, if this is really craftsmanship, then...
yeah it's it's it's okay that craftsmanship happens without me exactly it's like why do we keep reinventing the same problem just so we can say we and i love being an inventor right like that's that's how you grow great ideas is allowing people to participate and take them to the next level but it's like what are you trying to do we're just being nerds now i love being a nerd for nerds sake but like what are you trying to do and so like yeah it's devices because this is our personal identity this is who we are You know, just like a doctor walking around saying they're a doctor, like a developer walking around saying they're a developer, you've carried that badge of honor, you've paid the dues of production.
It's hard to give up that control.
But this is where we're shifting what we do.
We need to be experts because we're going to have a whole workforce that we're orchestrating.
And it's going to get better and better.
And if we don't know the traps, we can't train it to get better.
But I think at the end, the end user, again, mostly is going to be the every person.
Like, think about that, right?
Like, we have Wix and Squarespace, been around forever.
And now you can go and tell any chat or any agent, build me an experience, and it works.
And it's like, wow, who are we to say somebody can't get to do that?
Who are we?
Who are we the ones that get to say that we can't do that?
I want to be on the other end to say, how can we make it safe?
Like, I want to be like, you know.
The AI engineering that's doing the self-driving cars to make sure it doesn't go into a wall.
Like, I'm not going to overly worry about anything else except the do no harm part, because that's now where my expertise lies, because the operators aren't going to understand it at all.
And most of us play that part as well.
And that's okay.
It's that balancing act.
It's an interesting time.
It's a very interesting time.
And I think what we can do as practitioners is participate.
Don't let one model to rule them all.
Start thinking about fiscal responsibility, environmental concerns, the parts that we actually can control as humans and where we need to make systems safe for everyone.
Oh, don't say thank you to Claude.
No, you tell it to get it right.
You're like, come on, this is terrible.
Do it again.
But it's interesting because I look and see how a lot of people are doing agentic coding.
I don't know about you, I still go in my ID.
I hop back and forth because sometimes I'm frustrated.
I'm like, man, I'm not going to ask you for the fourth time to do this.
I'm just going to do it myself.
So we're still not there.
So people need to simmer down.
They all right.
I recently saw a guy who built some presentation tool, which was amazing in the background.
It was able to really edit in a company's style guide.
keeping up with the company's story, I was able to really edit slides using a prompt.
And he then said, yeah, and then I moved using a prompt, I moved this icon from this side to that side.
And he was super excited about that.
But I mean, ultimately, that's the downside, right?
That you really get lost in that loop or in that prompting hell.
Yeah.
to tell you, like a lot of us that are engineers at businesses, like again, I come back to, you know, in 2012 when Google released the SRE handbook, it had this premise of self-healing machines.
How does this any difference?
Like you're not going to put these agents in software factories and everything because we ain't there yet.
But like DevOps has always been the practice of utilizing development workflows for operational tasks.
And now the operational task layer is a lot wider.
It's just a lot wider.
It's like where you used to have to know something.
Well, guess what?
You don't even know that anymore.
And, you know, as a, I'll claim myself a full stack developer, which I just admitted I'm a terrible front end, but like, come on now.
It's like, you know, as well as me that you are doing, if you do know one area better than the other, that you are unstoppable now.
And I just can't deny the fact.
that I actually feel way more productive than I have in a lot of years doing full experiences versus handing the baton back and forth.
And a lot of companies, the reason Jamstack even came in is because marketing want to change the icon from this to that.
I'm like, let them do that.
You want them to come talk to you to change the icon?
Absolutely not.
Well, many engineers actually wanted that, right?
Because they wanted to stay in control.
And I think that's what also...
A very good learning is you can't stay in control, unfortunately.
Absolutely.
If you look at what's happening now, you can't, right?
There's no control anymore or little control.
So isn't your vision to build this system for staying in control?
I think you have to approach it in both ways, right?
the bitter trap is not building for the future, right?
And like looking and building for right now and it becoming obsolete.
So I'm constantly going like, where's the rough edges?
Where's the future going?
Where's the rough edges?
Where's the future going?
I think that there's two parts here.
I think one part is there should be a defaulted set of trust, you know, reliability, governance, et cetera, privacy that's just baked in.
Like, like we know people.
may not know the nth degree.
It's like, this is why we have SECOP professionals.
It's like, ask an engineer to get into that side of business.
They're like, I don't know, YOLO.
So I think that what I'm trying to do is really serve both parties.
One where you're like fast iteration and you're just like, I know that I'm getting these ideas out, but I'm still fleshing them out.
So you have this baseline of like, I'm not worried about that stuff.
I know it comes with brakes and seatbelts.
I'm golden.
But...
let's say you are an organization where you're all on this team sport, right?
When you're playing a team sport and everybody is now a solopreneur, you're going to need a little bit of control.
But the control that you need is the guardrails that are important to your business.
And I'm not going to tell you where you should have a human in a loop.
You get to determine that.
And so when I think when you go from like, I am working more solo.
I can do more things because I have control of the box.
But as soon as it's team sport, then you're like, okay.
And if it's team sports towards a particular outcome and it's not a small set of people, like you're going to want even greater guardrails.
But again, it's going to be interesting to see how even that starts to play out and where we see larger organizations have that lever.
A year ago, I asked a whole bunch of enterprise companies.
I said, Who's adopted AI or who's opened up their development practices outside of their technical teams?
And like, like a lot of teams are like, heck no, like that's, that's bonkers.
We spent years thinking about our CICD, our workflows, the, the protected IPP, our certifications.
And this year I asked the same question and the same room, everybody answered their hands.
So, you know, we can again say it's not here and it's not coming.
But what I'm trying to do is saying, it's here, it's coming.
They're probably going to do something not great.
Let's go protect them.
Because I don't want to be in the news for some noob like destroying databases or leaking out information.
But I also don't have a lot of control on that because that's your problem and you should be educated.
But I'm like, is it not?
Because again, that lever of responsibility is now back where we were in the 90s because it used to not be just buy, it was build.
And coming full circle from all these years, you're like, man, we're back to VMs and cloud computing and mainframes when we went to microservices.
But it's probably going to come back around to small language models and stuff on devices.
Yeah, let's wait a year.
Let's see where we go, man.
But that's my argument for it.
So what does that concretely mean then for your platform?
I mean, let's if I...
evolved as a static hoster, I think, throughout the years.
I still know Netlify CMS, which was amazing or is amazing.
I guess it still exists and it's an open source CMS that you guys built.
And there was like JAMstack.
Many of my listeners may also know as a term.
So basically JavaScript and or static websites that were like fully generated often from YAML files or from...
really steady content, rich experiences.
And how did that evolve from then?
How does that look like in the agentic age now?
Well, I mean, again, it's like stuff can still be static websites.
Internet, listen to me.
You don't need all the bells and whistles.
You think you do because they're available.
Again, hopefully agents solve this.
But it was a natural evolution as open source maintainers build more dynamicy and we start seeing things like...
router and we start seeing serverless functions and compute and these experience become non-static but alive and dynamic, we've naturally gone where the development practices have gone because we've been able to still maintain this separation of concern, which is really still, I think, more important than ever than in an agent world is like, okay, protect that darn back end.
It's actually got agents running around.
Don't give them access.
So really thinking about the architecture, but the evolution where we've gotten to is like Netlify has done a great job of saying like, let's get you from your code base into the internet.
We're going to take care of the deployment, the packaging, the deployment and the serving.
Well, now it's like, well, why not the front door connected up?
And now it's like ID8 all the way to internet.
So imagination to reality, we're on the same mission we always have been, which is.
The internet will continue to be composable even more now.
There's even more SaaS tools.
There's even more endpoints.
Like, shoot, there was what?
Just APIs and it was SOAP and JSON.
Now there's MCPs.
Now there's skills.
Now there's recipes.
Now there's going to be some new endpoint.
There's going to be something.
I don't know.
I still love WebSockets.
It's coming back.
But my point being is like, it's still the same problem we've always served, which is people that build.
should not have to worry about anything from getting that idea out to reality in a safe, secure way with the ability to manipulate it and collaborate with others in a recorded, revised way.
It's the same thing.
It's just now you can do it even faster.
And it's pretty freaking fine.
And from the command line or from using an agent or whatever.
CLI!
CLI!
I mean, it doesn't matter.
Without touching CSS.
Oh, never.
No way, man.
I'm like, come at me with your padding.
I was like, close enough.
No way.
It's amazing.
I think that, again, like we've just been able to like supercharge our mission now, which has always been like, the internet's freaking hard.
DNS is hard.
Guess what?
Builds hard.
You shouldn't have to worry about, you know, you need to worry about your fabulous idea just working.
And that's all I want.
And making the right choices for that fabulous idea.
So you're not spending a ton of money and you're not spending a ton of effort trying to learn how it works.
You're just knowing it works.
Now you can't go spend effort on how it works, but like do that when you have capacity.
It's hard to do that when you're in the firefight.
And that's where we're trying to really ensure that like you could deploy with no error because in a world where anybody can deploy, that doesn't mean they can debug the errors.
So it's my job to make sure.
that you don't get the errors and you can just go, go, go.
Oh, basically, your job changed or your company changed from providing the best possible developer experience, like developers deploying to production to agent experience?
I mean, agent experience is the human experience, okay?
We have redesigned our platform for agents because at the end of the day, it's not just developers that are operating them, it's everybody.
And what we found out in...
good agent experience allows for a great human experience because if agents can do the work, the humans benefit from the outcome.
When the agents can't do the work, then you're wasting everybody's time.
And if you don't have the wherewithal to go make experts in everything, then like you're not going to win because think about this.
If you can enable somebody that is a skilled, skilled marketer, product manager to move fast, they're going to be better at their part of the job than writing the code.
If we can enable that code to generate so that they can do their job, they're the experts.
Developers, unless you're really product-minded, you're going to be the bottleneck.
This is all about, again, like finding the right battles to pick and making sure you stay focused on the end experience.
And, you know, maybe you're an internal developer and your end experience is the release, you're a release engineer.
Great.
How are you going to make that frictionless?
How's release engineering for a developer any different than release engineering for a human that's not a developer now in the age of agents?
Because, like, why do we have SREs and DevOps functions?
If developers can do it themselves, we still do.
But so it's now our job is to create systems to define architecture, stuff we've already been doing.
but not necessarily right in the bits and bobs.
Because like you said, it's like, I always tell somebody, I'm like, oh man, I don't trust this code review agent.
It just runs it through.
And I was like, oh, and you trust Jimmy that just approved a 500 line PR that came in two minutes ago and he did it from GitHub mobile.
I'm probably going to trust the agent a little bit more than Jimmy on the 500 line code.
Like, come on, y'all.
Like, come on.
Like, let's be real.
Yeah, I think the reality is also that complexity kind of hit us or increased so much in the last years that an abstraction layer is actually needed to kind of handle that complexity, right?
Like, I don't know.
And even though like the problems are all, or many of the problems are all like history repeating.
There has been layers and layers and layers have been added on the way.
And you kind of need, I mean, the DevOps function by itself is in a way absurd, right?
So this, I think, really was just about time to have another abstraction layer in front of us.
I mean, it was going to happen, right?
It's going to happen.
Until again, you know, we have everything as heads up displays and everything's holograph.
I don't know.
They're working on it.
Like this is what we're working with right now.
Okay.
And so I don't know.
I'm here for it.
I'm here for all of it.
I'm here for the silly apps.
I'm here for the serious apps.
I'm here for the experts that already know what a product should do to be able to give you 90% of the way.
Like I'm here for that.
Where then my job comes and says, how do we scale this effectively?
How do we support the user experience with reliability?
And where do we find all those rough edges?
Because computers will always break and they will always fail.
And so, again, I still feel like there's a huge jobs to be done, but why not let the experts in their field, like, be the experts of their app?
I don't know.
Isn't that what we always wanted?
Yeah, absolutely.
Absolutely.
How did the demand actually change on your platform?
I mean, there are plenty of others now, right?
There's Vercel.
I don't know how much you compete with whom, right?
There's Render.
There are a few others.
AWS also introduced certain services for sure.
There's GitHub pages, etc.
There's a lot of stuff happening.
And how did...
I still feel that the demand exploded quite a lot, also with Argentic now being around.
How did that hit you?
And how did that change your business?
It's, I mean, you know, it's like old times.
I tell you what, like it's exploded.
And I rested on the fact that we have been really successful over the last decade of removing friction.
We're known for it.
And we're known for commercial professional grade software building and deployment and production readiness.
And now getting to leverage that at a wider breadth, like it's been great.
We're going to continue on again to serve that.
And there's more competitors than ever.
And they're actually different.
Like, you know, in the past, you would have maybe looked at Netlify, you know, even said it as a SASSing infrastructure hosting company.
Because, you know, content delivery network system and caching, like not everybody can do that.
And not everybody is going to do that.
They're going to try to do it though, because why wouldn't you?
And I think for us is keeping our eye on humanity and remembering that.
Building software is now the world's best team sport.
And some of those teams are teams of one and some of them are teams of 20.
But it's not just going to be skilled developers that have a background in computer science.
It's going to be curious people that are hungry to share, that are going to do the work to prop through all the roadblocks.
And the platforms that win are going to enable those people to really reach reality with their thoughts.
And I think that there's going to be a huge amount that doesn't matter.
You know, a lot of it's going to be a toy.
I mean, you were here for MySpace.
Remember when everybody learned HTML?
It's happening all over again.
I think it's going to be interesting.
We're intersects with like social, social media apps, the internet websites, commercial business.
But we have to wait and see.
But at least our mission will be.
to be able to be prepared for that with the right tool sets to be able to, you know, allow that participation for whatever that may be.
So yeah, that's kind of how we're looking at it.
Let's see, right?
You told me that you're thinking a lot about the future.
What are your thoughts on, I mean, code is going away, we have a new layer in front of us, and how does this look like in two years from now?
I think like...
If you would have asked me two years ago, I would really mess the question up or the answer would be so wrong.
Because everything is basically almost the same from my perspective, just slightly different.
Will it be the same in two years?
Will it be the same but slightly different?
How dramatic will it be?
I absolutely have no flippant idea.
I don't know.
No way.
No way.
I mean, honestly, if we think about the trajectory of technology, like I've mentioned in other fields, it's just going to get easier.
It's going to get easier.
Remember when like you first started using Excel and then like the real power users are like, check out this macro.
Nobody has time to learn macros.
And now Excel, you're like, I don't know, do this calculation, make this macro.
Software will get easier.
It will just get easier.
But where that line comes and what?
What new tools and utilities will make it better?
That's the real question.
There's some cool stuff out there, man.
And I still think there's a lot of work to be done, like Tanstack AI.
Man, that framework's nuts.
Like Jack Harrington, I get the privilege to work with that guy.
They're working on crazy stuff.
And like, it's for the agent though.
But again, it's the agent that's going to do it for the human.
And the human's never going to be on the wiser.
Because like, you care what framework your website is.
Do you like really care or do you want the website just to do what you want it to do?
Give a shit.
Like, and the only one that does is like hardcore developers that invested a lot of time and space.
And I will give it to the open source maintainers.
You better think your shit's precious.
You're maintaining it.
So like, you could have a high opinion of it, but I'm going to tell you, most of the world don't care.
They don't care what database flavor you use in.
They don't care what framework, what streaming system, what, what, what, it doesn't matter.
You just want it to work.
And that's reality.
And I think in two years, maybe stuff will be simpler again.
I think also so many people that care just shouldn't care, right?
Like, I think if you try to control something too much, sometimes it ends up controlling you or slowing you down.
So like the worst version of yourself.
And that is what I saw with SaaS, especially with so many SaaS apps I've seen.
If you overthink it, then it's not going to help you, right?
It's true.
And I found in my 30 days of building 30 apps, the apps that people love the most are the real simple ones that do something very, very specific.
And the ones where I've actually put too many brain cells in, nobody cares.
They're like, whatever.
I care about it deeply.
I'm really close to it.
I've thought about all the things, but like, I didn't do the market research and see if the demand was there.
And so I think, I think again, like to your point, like, you know, like spend a little time making something useful.
Don't make it just for you.
Um, and, and keep it simple because the most simple solutions are the ones that are maintainable, the ones that bring delight.
And I think that's just like a premise we've moved far from.
What's that old XCC comic where it's like software developer and the tire swing.
You remember that one?
It's like, ask an engineer.
Oh, man, I'll send it to you.
But it was like an old comic, an old web comic.
And it was like, ask an engineer to build a tire swing.
And it was like ropes and pulleys.
And it's just like, you need a rope and a tire.
Like, you know, and what are you going to get on?
And so like, I think keeping that concept.
But that's interesting because now when you have agents doing that, you should have clearer telemetry.
You should have better observability.
This is why I do think agent experience is like key to it.
Because again, like.
We can go build a bunch of abstractions, a bunch of stuff around it because we're solving problems from our own little purview.
An agent doesn't have that opinion unless you've told it to.
And I do hope stuff gets simpler and patterns become more easy to diagnose because computers will fail.
And so I think having a good investment on agent experience, which ties together telemetry, observability, ease of use, friction, and pain so that you can know.
and clear those error logs and again, focus on that experience.
And so in two years, let's hope it's all simple and apps are useful.
And I don't know, everybody's having fun and not as scared as we are right now because like, it's hard.
It's hard.
It's hard to choose your livelihood on this field and not know where it's going.
Yeah, absolutely.
But what do you think will happen to like, I don't know, the big, big open source stuff?
Like, I don't know, Postgres.
Will there be a day?
where like AI is not going to use Postgres anymore because it can just develop its own database for whatever purpose or is that I guess not right like somebody's trying that right now but I'm like why the hell would you go reinvent the wheel man like what are you going to get out what are you going to get out of building your own like flavor okay you to control your own destiny which means controlling your own database on a cloud provider because you still have to have servers somewhere out there in the freaking universe.
I'm like, I don't know.
Unless, again, this is where you can get real deep into like, is it fun to be nerdy or is it actually useful?
And like, here's my thing.
Like, if you want to go work on that tech, go find the company that says, I am changing the way that agents write data.
They're going to come up with, it's not going to be about creating my own database.
It's going to be something novel and new.
It's going to be totally novel and new.
And there is an open question, man.
It's like you brought up Netlify CMS.
I'm like, does CMS doesn't even matter anymore when you can just go change the code?
Why do you got to store it in a weird ass content management system?
Why do you need a CMS when I can go right now onto Netlify, do an agent run, change any content I want.
If I don't like it, I'll change it back.
I don't ever know.
There could be a CMS in there.
I don't freaking know.
But nine times out of ten, it's not going to choose one.
It's just a little blob in a blob store.
Yeah, yeah, yeah, yeah.
Yeah, well, if it becomes a team sport, then maybe more than, then there's more to that, right?
To more reasons for a CMS, but otherwise you're true.
You're right.
I think what on the other end will get lost if you don't reinvent the wheel will, like, I think stuff will look the same.
A lot of stuff will look exactly the same as other things that were invented already and that it's going to be a bit boring.
I don't know.
It is.
Shopping on Amazon.
I'm going to tell you, it's so boring.
But this is where we don't know.
How are humans going to do their shopping?
Are they going to go to a website?
Are they just going to be like, my claw bot, you heard me say Jimmy.
My claw bot's name is Jimmy.
Is it going to be like, yo, Jimmy, buy me some new shoes.
And then Jimmy's going to send me a list of shoes.
And I say, do it.
Like, personal assistants are really all.
It's going to happen.
But I do think, like, you know, I don't know.
I don't want it to be boring.
I mean, hopefully as a video podcast, I wore a really crazy shirt today.
Because, like, guess what?
AI ain't going to come up with this.
This is the best part of humanity.
The best part of humanity is, like, I, right now, And AI can hallucinate too, but we don't call them hallucinations.
We call them intentions as humans.
I can do whatever the hell I want.
Nobody has guardrails or skills or recipes on me.
So we still have a job to be done.
So I don't think we'll ever get that boring.
But I'll tell you, when it spits out one of my apps and I'm just like, oh, come on.
I'm like, what is this?
Like everybody's using Claude Design.
I get it, but I would rather not.
So I'm hoping that we can go.
Keep it weird out there with the real creatives and have them come in and really still play their part in making it better.
But that's the one thing we need to protect against is like not having boring experiences.
So now I really want to know which shoes you're wearing and which app was the most popular one in the 30 days challenge you did for yourself.
Okay, man, I'm a true hippie.
I ain't got no shoes on them in the house.
Come on now.
And my most popular app.
All right, let me look.
Let me look at my projects.
This is a live reading of all the money that I am spending.
My most popular app wasn't a part of the 30 days, but it's still just eating up all my money.
It's called Who Was Cancelled?
It will tell you who's been canceled on the internet.
Unfortunately, that is a very popular app.
I was like, dang, man, I got to go get me some BC money for that one.
All right.
My second one is, where is it called?
It is called Animalify.
You can turn yourself into an animal and it tells you what type of animal you are.
So it's real dumb.
It's real dumb.
Like everybody loves a picture of themselves anamorphotized.
And it's like, yeah, you can go and chat GB2, but you can also just hit my little app and do it yourself because it's really cute.
People want to know who's canceled and if they're an otter.
And so like, you know, it's still very human centric.
Okay.
Very cool.
I have to look up your apps afterwards.
Maybe you can send me a list and I'll put it in the show notes.
Go get my LinkedIn.
I'm releasing every day.
I got seven more days of dumb apps to build.
Great.
Looking forward to that.
Yeah.
I think the world is missing a reverse talk app, which you can use to record yourself and then...
Learn reverse talking, right?
That would be cool.
There you go.
I needed an idea.
People have been throwing random ideas.
I was like, I'll build it.
I'll build it.
You watch me.
Look at that.
Look at that.
A year ago, a year ago, I wouldn't be able to do that.
I wouldn't be able to sit here in literally one hour and have like a fully formed, you know, DNS named cert bot end-to-end application that works.
No way.
No way.
You would have been like, you may be in a day, but that's the reality of what's happening right now.
Yeah, yeah.
I know.
It's so hard when you think.
But coming to my outro question, as I'm already super keen to hear your answer, let's imagine in the Netlify MCP server, there's a little Easter egg that someone in your...
team whispered me about.
And it allows you to, if you call that like certain command called time machine, it allows you to travel back in time, not forward, just back, unfortunately.
Otherwise, we would all look like two years ahead, right?
And imagine we could now land in the late 90s, I think it was, or early zeros when you had your first day in the army.
Yeah.
And you're sitting in front of a rack of Unix boxes and you get to whisper one thing into your younger self's ears.
What would it be?
Invest in Amazon and Google.
Oh, no.
Come on.
I'd be like, girl, NVIDIA, who knew?
You love GPUs too.
No.
Honestly, I would tell myself to go harder, man.
I don't even know.
If I could tell, I'd be like, dude, you were going to do it, but if you even just buckle down a little bit more, you could do even more.
So I don't know.
I'm a very go, go, go.
I would probably just cheerlead myself into it.
I'd be like, girl, you didn't even know.
You should probably go learn some machine learning.
dust off that python book you've looked at three times in your life it's time to go so i don't know man i you know i i love where i'm at i love what i'm doing i think i would just probably go harder i would just go harder i would i i would be like get those backup tapes out of my way robots can come and change these tapes i would have been right i would because like i don't know i've always you know, I guess naturally gravitated towards like, how can we give people back time to be people?
And, you know, that was the beginning of my, my journey of like, I don't want to sit and like, I don't know, change passwords by hand.
I don't want to deploy by hand.
I don't want to roll anything by hand.
I don't know.
I want to play my guitar, come up with bad app ideas.
And hang with cool people that together we can do something better.
And that's where I'd love open source.
So, you know, I don't know.
It would be interesting.
I would tell myself, go harder, contribute to more open source.
And like, I don't know, like, shoot.
Like I said, learn some more Python.
But now I'm like, maybe not learn Python.
But I don't know.
I never fully embraced it.
I started in C.
I like to go the hard route first.
Thanks a lot, Dana.
And have a great day.
Very nice talking to you.
And I hope we meet you to all soon.
In the real world, maybe.
Thanks for having me.
I'll check out your shoes then.
You'll see what I'm wearing.
That'll be cool.
Don't worry.
It was great to be on there.
Have a great day.
Bye.
Thank you for listening to the Alphalist podcast.
If you liked this episode, share it with friends.
I'm sure they love it too.
Make sure to subscribe so you can hear deep insights into technical leadership and technology trends as they become available.
Also, please tell us if there is a topic you would like to hear more about or a technical leader whose brain you would like us to pick.
Alphalist is all about helping CTOs getting access to the insights they need to make the best decisions for their company.
Please send us suggestions to cto at alphalist.com.
Send me a message on LinkedIn or Twitter.
After all, the more knowledge we bring to CTOs, the more growth we see in tech.
Or, as we say on Alphalist, accumulated knowledge to accelerate growth.
See you in the next episode.