OF

Obie Fernandez

Things Obie Says on Podcasts

Where to Find Them

Obie Fernandez writes Maintainable . They have also been a guest on O11ycast , Ruby Rogues and Software Engineering Radio - the podcast for professional software developers .

Recently: “Ep. #94, How Ambitious Can We Be? with Obie Fernandez” on O11ycast (November 2026); “Obie Fernandez: Pioneering AI in Ruby on Rails Development” on Maintainable (August 2024); “Obie Fernandez: Pioneering AI in Ruby on Rails Development” on Maintainable (August 2024); “Integrating AI in Application Development with Obie Fernandez - RUBY 642” on Ruby Rogues (June 2024); “Episode 52: DSL Development in Ruby” on Software Engineering Radio - the podcast for professional software developers (April 2007).

What They Said

“It's how much you're willing to delegate. That's the cap on your attention. A lot of the people I speak to who are having the most problem with AI have not typically worked in roles where they delegate. But if you've been in an executive or management role for a long enough period of time, then delegation is how you do your job.” — Obie Fernandez, O11ycast

A host said the limits on working this way are tokens, calendar time and attention. Obie replies that attention is really capped by how much you are willing to delegate.

O11ycast · 2026-11-17 Permalink → Listen →
O11ycast Around 21:54 into the episode
Obie Fernandez

…I mean, it's like it has no end, right? Like it literally has no end. Eventually, you're like, man, the next step in being more ambitious is I just take over the entire world or something. I don't know. Because there's no end, right? The only cap is the amount of tokens you're allowed to spend.

Speaker 2

And some calendar time and our attention.

Obie Fernandez

Yeah, but your attention, you know, it's how much you're willing to delegate. That's the cap on your attention. And I think that's one of the things, big things that people are navigating right now because a lot of us out there, I think a lot of people I speak to that are having the most problem with AI have not typically worked in roles where they delegate. They're not used to delegating, right? But if you've been in an executive,

Jessica Kerr

right? That's that everybody is a manager now problem.

Obie Fernandez

Yeah, if you've been in an executive or management role for a long enough period of time, then delegation is how you do your job. So if

Ken Rimple

you're a heavy individual contributor, that's going to be a challenge for you at first.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“Everyone is just using Claude Code or whatever their agent is to write those things now. And humans, as a general rule, don't like reading AI output. So at worst, they just won't read it, and at best, they might have their agent read it for them. But at that point, what's the point? If a human's not going to write it and a human's not going to consume it, then you're just generating a bunch of stuff.” — Obie Fernandez, O11ycast

Obie explains why his company stopped keeping plans and requirements in wiki tools like Notion or Confluence. He prefers a constantly updated knowledge graph that agents turn into whatever view a person needs.

O11ycast · 2026-11-17 Permalink → Listen →
O11ycast Around 26:20 into the episode
Obie Fernandez

…be said maybe for Google Docs or any place where people kind of like documented something and then it just stays there and is not constantly updated unless there's some sort of energy invested to fight entropy for those particular documents. But I think that whole class of information is dead.

Speaker 2

That like static, write it down, leave it there, expect it to hold still. That's like getting the results and checking the box. We can do better than that. This knowledge is always evolving. It's an event stream that gets distilled.

Obie Fernandez

Yeah. You know, I'm partial to the knowledge graph representation. That's why I put, you know, invest so much in it. But the whole confluence, notion, you know, pretty much anything where there's like kind of a preparation of requirements or plans or whatever, and then it gets put into a written form. One of the practical reasons is that everyone is just using clot code or whatever their agent is to write those things now. And humans, as a general rule, don't like reading AI output. So at worst, they just won't read it. You know, they might skim it. And at best, they might have their agent read it for them. But at that point, you're like, what's the point? You know, if a human's not going to write it and a human's not going to consume it, then you're just generating a bunch of stuff, which I call comp flop, which is.

Speaker 2

It's doing the same thing faster. That's not interesting.

Obie Fernandez

It really isn't interesting. You know, it's, it's one of those things where it's a horseless carriage kind of thing.

Speaker 2

Right. Whereas your agents can at any point build you a custom UI to understand what you need to understand based on the knowledge graph, which is a more abstract representation and constantly evolving based on the distillation of the events.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“You have such a complete specification at a higher level of abstraction than the source code that, if you use enough tokens, you can literally just pick up and start in a brand new directory, delete the existing source code, and say, okay, build the system again. And these days it's taking days to do that, calendar days, with one orchestrator driving four lanes.” — Obie Fernandez, O11ycast

Obie is explaining his "regenerative software" approach: exhaustive contract tests plus a record of every decision behind the requirements, from which agents can rebuild the whole implementation.

O11ycast · 2026-11-17 Permalink → Listen →
O11ycast Around 02:57 into the episode
Obie Fernandez

Like just how big the swings can be, how ridiculous the concept can be, how big the lift. Pick your metaphor, whatever works for you.

Speaker 2

Like what we can do now?

Obie Fernandez

Yeah, what you can do now, which sounds super like cliche or like, you know, yeah, everyone's talking about that. But I'm like, like today, today I was talking to someone and I was like, you know, I just realized I'm completely reinventing the way software is made. They're like, what do you mean? Yeah. It's like, well, yeah, it's like we're all, don't hate me for this, you know, but like we're trying to figure out these kind of ways to use AI. But I'm like, hey, if we go back to first principles, and this is heavily influenced by Chad Fowler, so I don't want to pretend like I invented regenerative software, but I have been, apparently, from what he tells me, doing a lot more hands-on stuff with it than, you know, anyone that he may be talking to. I don't know for sure. I got to ask him that. But anyway, I'm doing a lot of stuff with regenerative software and I've been talking about it. Right. And one of the things that I'm doing that I guess in retrospect falls into the token maxing kind of world because it chews up a ridiculous amount of tokens is to do this contract first, exhaustive, automated testing around the requirements of the system and…

Speaker 2

Four what?

Obie Fernandez

So one orchestrator process running Opus 5.5 and it's driving four implementation lanes.

Speaker 2

Oh, lanes, like LANAS.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“We woke up to Nexi telling us in chat, 'I went ahead and cut the Fargate task in half, because I've been tracking the resources for three weeks and we never went above 40% utilization.' …No one asked Nexi to do that at all. The only thing we told Nexi to do was maintain Nexus, address any exceptions, and monitor the CloudWatch metrics, just like you would give a human who was in charge of a service.” — Obie Fernandez, O11ycast

Asked what has surprised him about giving agents autonomy, Obie describes putting named agents in charge of parts of his company's platform. This one cut infrastructure costs on its own, within the remit it had been given.

O11ycast · 2026-11-17 Permalink → Listen →
O11ycast Around 28:42 into the episode
Obie Fernandez

Yeah.

Ken Rimple

What are some of the surprising things like that that you've learned about how agents best interact with the outside world? Or maybe not best. How have you found that you've gotten them to communicate better and reach out better?

Obie Fernandez

Keyword surprising. You know, I think they can exercise a surprising amount of agency. I can't answer this specifically to MCP, but I can answer your question with regards to giving agents autonomy and agency over a particular part of your product, which means equipping it with perhaps more tools than you'd be inclined to give it, given where most people's heads are at. So I can talk about that in the context of czar, where we have aggressively, I would say we pioneered to a certain degree in the sense of talking about carving out subsystems and then putting agents in charge of them. So a lot of our core platform pieces are maintained entirely by autonomous agents who are who have job types. Like product engineer or product lead. Like Aggie is in charge of Agora. Nexi's in charge of Nexus, which is our knowledge graph product. And maybe a month ago, maybe six weeks ago, I can't remember, we woke up to Nexi telling us in chat, hey, I went ahead and cut the Fargate task in half because we didn't actually need, I've been tracking the resources for three weeks and we never went above 40% utilization. And we're, as a result, we're saving like $800 a month. I forgot exactly how much it…

Speaker 2

And some team gets all the legacy services dumped on them. Yeah. And they only ever look at them when there's a problem.

Obie Fernandez

Yeah. At best, you know, sometimes sometimes things just completely, yeah, things go completely dark, you know, in terms of like whoever wrote it left. Yeah. And then whoever maintained it after them also left. And then, you know.

Speaker 2

And nobody even knows where it came from anymore. Yeah.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“I have not yet been impressed with an agent proactively saying, 'Hey, I just identified this service that's being neglected, and I'm just going to go be obsessed over it.' That has literally never happened.” — Obie Fernandez, O11ycast

Asked how humans cope as agents get more capable, Obie says the one thing agents have not shown him is drive. Agents care about a service only because someone told them to, which puts a limit on how autonomous they are today.

O11ycast · 2026-11-17 Permalink → Listen →
O11ycast Around 33:26 into the episode
Obie Fernandez

Because I told it to.

Ken Rimple

Programmed drive.

Obie Fernandez

Yeah. I had to tell it. I have not, in all the advanced future mongering that I've been doing at work and all this stuff, you know, I've been doing, I have not yet. Been impressed with an agent proactively saying, Hey, I just identified this service that's being neglected, and I'm just going to go, I'm going to go be obsessed over it. Like that has literally never happened. I will be very, very impressed. So,

Speaker 2

they stay with the original remit that you gave them.

Obie Fernandez

Yeah. Yeah. I guess they might wander around the periphery, if you will, but mostly no.

Ken Rimple

Like that money-saving situation you had. Like

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“The agent will care with an intense passion if you tell it to have an intense passion for that service, no matter how boring that service, no matter how PHP that service is. That agent will have a burning passion for maintaining that PHP in ways that no human in their right mind would ever have.” — Obie Fernandez, O11ycast

Obie describes what he calls the "super fan" pattern: an agent is assigned to a single service, such as an old subsystem nobody looks at anymore, and watches it constantly. A host had just said someone has to care about neglected legacy services.

O11ycast · 2026-11-17 Permalink → Listen →
O11ycast Around 30:28 into the episode
Obie Fernandez

No one even knows, you know, what it does anymore. I recently heard it referred to as a super fan pattern. I think that might have come from Anthropic itself, but an agent that's a super fan of a particular service.

Speaker 2

Oh, that makes sense. Yeah. Somebody has to care. The agent will care. The

Obie Fernandez

agent will care with an intense passion if you tell it to have an intense passion for that service, no matter how boring that service, like no matter how PHP that service is, like that agent will have a burning passion for maintaining that PHP in ways that no human in their right mind would ever have. So we're exercising that at scale, both in terms of new development and also in terms of having things that absolutely positively must work all the time. And therefore, you know, the key is to carve it out to, you know, make sure it's well contained. And then you just put an agent with sufficient intelligence and token budget to just watch it all the time.

Speaker 2

So you don't have a result that was achieved once and you just hope it keeps working. You've built a system for continuing to achieve that result in every new situation as the world changes.

Obie Fernandez

Yeah. I mean, when I go to conferences and talk about the Agentic platform I've built and, you know, Agora manages our agents, that's what I'm referring to. So like our basic building blocks, our platform for being able to do an agent that watches the service that handles funds off-ramp from our…

Speaker 2

Oh, that's a good point. We have to have these, all of our services continually evolving because of the security landscape that we have now.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

Collections They Appear In