Nobody Knows

Kent back here. Host of the podcast Still Burning, where we have conversations with geeks who still care and are still doing something about it. I want to contrast that with the geeks who have kind of given up. They learn some skills. They practice the skills. They're not changing anything. Maybe they're even working in the same codebase, in the same style, on the same team.

The same rhythm, the same tools. The same language, year after year after year. I don't have anything against people like that. It's just not my people. And the problem right now with that approach to being a geek is everything's changing. You're not guaranteed that you're going to be able to keep doing exactly the same kind of thing, even if that's what's sustained you for ten or 20 or 30 years.

The rules of the game are changing. The technology landscape is shifting. We don't know what the shape of the new landscape is. We don't know what the shape of the landscape is going to be a year from now.

So adaptability brings value in ways that it hasn't before, especially compared to just executing on the same kind of things you've executed in the past. We've all been parachuted into a wilderness. Nobody knows the answer. Five years ago, if you'd said, oh, we have too many defects in production, I would have been able to press play and given you the same advice that I gave ten years before and 15 years before and 20 years before.

Here's how you align authority and responsibility. Here are the tools that you use. Here are the ways to think about it. The thing is, nobody knows the answers. If you have too many defects in production and you're doing augmented development, how do you address that? Nobody knows. We know some things in principle, but we don't know things in practice in the same kind of way that we have in the past.

So it behooves all of us to approach this changing situation with a curious combination of confidence and humility. We don't know the answer. Nobody knows the answer. We'll try some things out. Now, trying things out is a skill on its own. Just as executing a known strategy is a skill on its own. So the kinds of things that we'll emphasize in the conversations here are things like curiosity, cheap experimentation, trying out bad ideas.

If an idea is cheap enough to try out, you don't have to prejudge it. You just try it and see what happens. And 99 times out of 100, it's going to turn out to be a bad idea. But that one time that it turns out to be a good idea — oh, that's so sweet. Because one, it's a genuine discovery, which just feels great. Two, you have no competition because nobody else is dumb enough to try that idea. Three, you're going to be contributing to your community.

Everybody's trying to figure this stuff out. The people who aren't busy saying "don't bother trying to figure this out, it's all going to go away" and the people who say "don't bother trying to figure it out, it's impossible anyway" — we won't talk to those people. But we will talk to the kind of geeks who still care and are still doing something about it. There are young people doing amazing things right now, and they still care, and they're still doing something about it.

I use GeePaw Hill's definition of geek, which is: a geek is someone who is highly technical, highly creative, and highly desirous of being both technical and creative. That means you can apply geekdom to all kinds of activities. You could be a baking geek. You can be an art geek. You can be a music geek. We're talking mostly to programmers, but we'll see where this conversation goes. I think once we get the conversations kicked off, we'll discover that the geeky mindset applies to a much wider range of human activities than we suspected.

I think one of the things that's challenging for a lot of geeks is that they haven't chosen the disruption that they're living through. Five years ago, you could have highly refined skills. Expect to spend the rest of your career using them only to find that — for example, I've written two books on how to write code that humans can read. Well, go on. Just not a skill that has any leverage anymore.

I'm sad about that because I loved crafting code that could be quickly scanned, easily dove into, changed in various kinds of ways. I felt good when I made progress on those dimensions, and now I just don't exercise that anymore. Kind of like using a clutch or remembering phone numbers. It's just not a skill that matters anymore.

But I think a lot of the dislocation that people are feeling right now is because they didn't choose this. They got good at something. They were able to support their families, save for their futures based on those skills, and now those skills have been devalued by changes introduced from the outside. And yes, that genuinely happened and it happened to all of us.

The question is not can we hang on to our horse-shoeing niche service until the last horse dies? It's how are we going to adapt. A lot of people are applying a conservative approach: what's the least I can change and still keep making my house payments? I don't think that approach serves the individuals involved well, and it certainly doesn't serve our community well.

We're all trying to figure this out. And the people that I'm going to have conversations with here are the kind of people who still care and they're still doing something about it. They're the kind of people who say — what I got good at ten, 20, 30 years ago is not good enough anymore. So what is good enough? What is going to make a difference now? How can I learn from my community? How can I contribute to my community? How can I teach the juniors? How can I learn from the juniors?

One of the most exciting things about augmented development for me is how often I get fantastic advice from people who are half my age. I'll say, well, here's my approach to this. And one of the younger folks will say, perhaps that's not going to work. And I just think that's fantastic. I'm excited to learn from them. But I also need to figure out which of the principles that I built my career on really are evergreen, and which of them were context-dependent in a way I wasn't aware of.

The thing about the genie is everybody's world has just expanded. If I want to write a database from scratch, I can write a database from scratch. What would have been an impossibly big task may or may not be an impossibly big task anymore, depending on what exactly I'm trying to accomplish. The skill of imagining things that would be valuable to create suddenly has much more leverage. The skill of picking which ones to explore suddenly has much more leverage. Knowing when to abandon a project suddenly has much more leverage because you're starting a project a week instead of a project a year.

I've been surprised by the vehemence of some of the reactions to augmented development. Is that caution, or is that ego? I was not an early adopter of augmented coding. People were doing it, raving about it and raving at it for probably a year before I saw Gene Kim and Steve Guy give a demonstration of what they called vibe coding. And that was the spark that got me started.

Now I try and take a Buddhist mindset to everything I do in my life. And that helps me understand some of the most emotional reactions to augmented development. If you spent ten years getting really good at TDD — all of a sudden the workflow matters. But picking the next test, for example — that's a subtle, nuanced skill. In an augmented world, it's not clear where the leverage comes from. Even fundamental questions like how many tests you need to make pass at one time — nobody knows.

When I feel resistance in myself to trying a technique with augmented development, or to stop using a technique, I think that's a question of ego. I'm attached to it. I worked really hard to understand. One of my drives is to really understand topics deeply — which led me to spend 20 years of my life writing a book about software design. And once I really understand something, I'm reluctant to say: it doesn't really matter.

Some of the things that I learned through that diligent digging really don't matter anymore. Some of them matter more than ever. Sorting out which is which requires me to take a clear-eyed look: here are the things I'm good at, but let me sort that into things that for sure matter, things that don't matter — bless and release, as my music professor Bob Trotter used to say — and the things in the middle where nobody knows.

If there was an alternative title for these conversations, it would be nobody knows, because that's the facts of it. We have these new tools with new trade-offs, and the tools are changing every day, and the tools themselves are causing the tools to change even faster.

A mystery to me is not passing 100% of tests. I find it so valuable to just know — to have absolute certainty. But increasingly we're working in a world where software mostly works. Is that okay? Is that a transitional state? I don't know. That's the kind of thing I expect to explore in these conversations. Software defects used to be a bad thing. They're still a bad thing. Under what conditions? What can we do about it? Nobody knows. So we're all going to find out about this stuff together.

One of the exciting things about augmented development is the possibility that designers and product people can make their decisions and implement them directly without the intervention of a programmer. Even more exciting is the ability to pair across disciplines. The augmented development tools give us a chance to break down silos. They also give us tools to enforce the silos more directly. I hope we don't do that. I hope we use them to make connections between people.

We're not going to have easy conversations here. We're going to have conversations between people who disagree about things. We're going to have conversations where values and principles contrast, maybe even conflict. I expect to have skeptics on. If you want a gentle pat on the head, this is not going to be a conversation for you. This is encouragement, a pat on the back, sometimes a kick in the butt: hey, go out and try it.

Nobody knows. Nobody knows. So if you say TDD can work with augmented development, or TDD is the only way to do augmented development — try it. Everybody needs to be in that try-it mindset. Ten years from now, 15 years from now, who even knows what part of development is going to be a human activity? Nobody knows. But for right now, we're in this terrifying, fun, exciting, changing, highly creative state where the rules of the game have changed and nobody knows what the new rules are.

I've been asked: how do you keep up with everything that's changing so much? My snap answer the first time I was asked that: I don't. I'm too busy staying ahead. That's who these conversations are for. In the words of the late Reverend Jesse Jackson: I'm a tree shaker, not a jelly maker. I'm here to try stuff out, figure out what works for me, talk about it, listen to what's working for other people. When their conclusions contradict mine — fantastic. It's an opportunity for both of us to understand more deeply the context in which the techniques, principles, and skills operate.

I'm a programmer. I hadn't written lots of programs for the past ten or fifteen years. I have rediscovered my love for programming doing augmented development. I use AI for everything I can think of — learning about eigenvalues, finding the best seafood in Porto. The thing I love more than anything else in my intellectual world — finding an interesting topic and digging into it — is suddenly much easier now.

What scares me about doing this podcast? No real care. That's my biggest fear. I think we're at a time of great leverage, great opportunity, and great risks. I'll put my two cents out there and it just won't matter. Nobody will be listening. Nobody will do anything based on what I say. That's my greatest fear — that the negative consequences of this technology on an activity I love so much will barrel ahead and there's not a goddamn thing I can do about it.

Well, we're going to have conversations. Human connection, real conversation between people who don't have the exact same perspective always creates value. The more of that I can stand — because it's not easy for me — the more value I can create. So that's why there's a chair here. That's why there's a blanket, because it's kind of cold in spite of the fire. And this seat will be occupied by people who are ready to have conversations about what it's like to be a geek, to still care and still be doing something about it.