Subscribe
Share
Share
Embed
Kent Beck has spent 40 years watching and guiding engineers through their careers. He's made note of what it takes to help engineers stay sharp, stay curious, and keep building for decades. What he's learned is less about any one skill and more about adaptability itself; what it takes to be the kind of person who keeps up. Kent sitting down with geeks in front of a fire to talk about their careers, and what they're still doing to stay adaptable in today's landscape. Real stories, real engineers, real honest.
[01:00:00:00 - 01:00:15:07]
Kent Beck
Charity, welcome to Still Burning. It is great to meet you in person. I've read a bunch of your stuff. Likewise. We've interacted, but never in person. So, it's great to have you here.
[01:00:15:07 - 01:00:21:13]
Speaker 3
Honestly, it's kind of hard to believe that, but yeah, no, it's great to be here. It's a lovely place.
[01:00:21:13 - 01:00:23:02]
Speaker 1
Thank you. Thank you.
[01:00:23:02 - 01:00:25:16]
Speaker 1
when did you first know you were a geek?
[01:00:25:16 - 01:00:30:09]
Speaker 3
I was homeschooled when I was young. Okay.
[01:00:30:09 - 01:00:32:18]
Speaker 3
touched a computer until I went to college.
[01:00:32:18 - 01:00:35:21]
Speaker 3
I actually had a performance piano scholarship.
[01:00:35:21 - 01:00:52:22]
Speaker 3
I started to realize that the people who played piano for a living, you know, they graduated in music, and then they were still in their 20s, 30s, 40s. So, hanging around the music department, accompanying, giving piano lessons, then going to work at Subway. And I was just like, no.
[01:00:52:22 - 01:00:53:22]
Speaker 2
Okay.
[01:00:53:22 - 01:00:57:20]
Speaker 3
No. And that's when I started hanging out in the basement of the computer labs.
[01:00:59:05 - 01:01:02:17]
Speaker 3
And that's where I learned that I was a geek.
[01:01:03:21 - 01:01:10:20]
Speaker 1
the, did you have something in common with the people there? Did you stand out compared to the people who were there?
[01:01:10:20 - 01:01:41:01]
Speaker 3
I was on Doom. I was playing Doom or Quake, or I was on a Unix terminal with everyone there. And everyone there was a greasy, pimply, 17, 18 year old boy, maybe 19. Right? Okay. Because that was, and at that point in my life, I was just like, yes, I want to be wherever the women aren't. Because where I grew up, all the women were having babies and doing dishes. And I was just like, nope.
[01:01:42:04 - 01:01:49:13]
Speaker 3
So, you know, I learned some great things in the basement. And I learned some not so great things in the basement.
[01:01:49:13 - 01:01:50:16]
Speaker 1
Okay.
[01:01:50:16 - 01:01:56:13]
Speaker 3
But that's when I learned I was a geek. That's when I learned that geeks existed. And I definitely was one of them.
[01:01:56:13 - 01:02:06:08]
Speaker 1
Okay. So I use GPAH Hills definition of geek, which is someone who's highly technical, highly creative and highly desirous of being both.
[01:02:06:08 - 01:02:07:21]
Speaker 3
Oh, I love that.
[01:02:09:19 - 01:02:13:21]
Speaker 1
And like you, I also spent time in the music school.
[01:02:15:17 - 01:02:16:07]
Speaker 3
What instrument?
[01:02:17:11 - 01:02:18:23]
Speaker 1
Guitar. Oh, nice.
[01:02:20:16 - 01:02:27:15]
Speaker 1
And for me, I ended up, I did music
[01:02:27:15 - 01:02:29:13]
Speaker 1
and computer science every other year.
[01:02:29:13 - 01:02:30:19]
Speaker 3
Oh, interesting.
[01:02:30:19 - 01:02:32:07]
Speaker 1
And I just ended on the wrong year.
[01:02:35:15 - 01:02:49:09]
Speaker 1
Which helped me escape the, yeah, I mean, I know that person that accompanist, who's just been there for 20 years and loves making music, but can't make a living.
[01:02:49:09 - 01:02:50:13]
Speaker 3
Can't make a living at it.
[01:02:50:13 - 01:02:51:16]
Speaker 1
Yeah.
[01:02:53:11 - 01:03:00:11]
Speaker 1
We were talking before we got rolling. We live in, what's the Pogo?
[01:03:01:15 - 01:03:04:20]
Speaker 1
We live, we're confronted with insurmountable opportunities.
[01:03:05:20 - 01:03:09:05]
Speaker 3
Yes. Yes, we are.
[01:03:09:05 - 01:03:20:20]
Speaker 1
And, and so you talked about how a year ago, you couldn't necessarily see exactly where the genies were going.
[01:03:20:20 - 01:04:17:15]
Speaker 3
You know, I, I just wrote a little sub-stack post about this because, and it's called my hypothetical as a recon 2016 out because 2025, 11 months ago, Fred Haber and I were giving the closing keynote at a three grand 25. And our big pitch was, okay, SREs, it's time to learn the vibe code, pause for cynical laughter. Lots of people are like, Oh, vibe code, you know, time to do it because our big pitch was people aren't going to take you seriously when you try to slow things down or when you complain, unless your complaints are coming from inside the building. That was our big pitch. You can't really slow things down effectively unless you're inside the building to slow things. And I look at that now and I'm like, that was less than a year ago. I can't even connect with it. But I do think that like,
[01:04:19:01 - 01:04:35:15]
Speaker 3
I think it was reasonable up until sometime in 25. It was reasonable to be very skeptical about whether or how long it would take AI to generate code that was roughly as good as a median developer.
[01:04:35:15 - 01:04:36:06]
Speaker 4
Right.
[01:04:36:06 - 01:04:45:05]
Speaker 3
And at some point, I don't even know how many technologists I've heard, who will confess at least privately that, yeah, there was a point in 2025,
[01:04:47:10 - 01:05:02:05]
Speaker 3
when it turned a corner and it's no longer reasonable, reasonable. A lot of people seem to date that back to the cloud Opus 4.5 release, you know, definitely not reasonable. That's just in the past few months, dizzying.
[01:05:03:05 - 01:05:30:21]
Speaker 1
Yes. And you can try something and it doesn't work and you can feel all superior. And then a week later, you try the same thing in the same circumstance. This time, smooth as silk. Yeah. It's like, Oh, I mean, good. I didn't have to do anything to get better tools. I didn't have to learn new skills. But also, oh, holy crap. What what is what is this open up for me next? Yeah.
[01:05:30:21 - 01:05:47:12]
Speaker 3
Yeah. And so the advice that I gave in my hype, I can't go to S3 con 26 this year, but I was like, but if I was going, the advice that I would give is this is coming. It is changing the way we build software.
[01:05:48:15 - 01:06:00:14]
Speaker 3
And you should spend your time at work doing this if you can. If you can't do it at home. But this you want to run out to the waves. Don't sit there and wait for the waves to come to you.
[01:06:01:14 - 01:06:09:21]
Speaker 3
Run out to meet it. Because it's actually it's hard to master, but it's really easy to get started.
[01:06:09:21 - 01:06:11:01]
Speaker 1
Like the ukulele.
[01:06:11:01 - 01:06:28:18]
Speaker 3
Yeah. I love that. I love that. And it's never going to be any easier. It's correct. Always the easiest. Remember when you know, document databases MongoDB, you know, it's not it's not going to be easier. It's only going to be harder. And right now, we're all novices.
[01:06:29:20 - 01:06:33:22]
Speaker 3
And there's a real fellowship in that can be there can be
[01:06:33:22 - 01:07:11:15]
Speaker 1
can be. So I that's my hope for our community is that we're going to be learning and helping together. I think it requires some skills that people haven't necessarily exercised. Because we've been through this period where the like what you should do to do a good job was pretty known. Yeah. Maybe you chose not to do it. Maybe your employer didn't give you the incentives to do it. But it's not like anybody didn't know about testing. They didn't know about CI CD.
[01:07:12:21 - 01:07:18:18]
Speaker 1
Yes, but not here. Okay, fine. But you're making these complaints. There's a solution.
[01:07:20:06 - 01:07:23:22]
Speaker 1
Now you have complaints. But
[01:07:25:16 - 01:07:33:15]
Speaker 1
what is the solution? Yeah. So I like telling especially juniors nobody knows. Nobody knows.
[01:07:33:15 - 01:07:44:02]
Speaker 1
everybody's ignorance has been reset to 100. Yep. All at once. And it doesn't matter how experienced you are, or how, how smart you are.
[01:07:44:02 - 01:07:48:11]
Speaker 1
broad your experiences, how well you write, none, none of that stuff matters.
[01:07:48:11 - 01:07:49:13]
Speaker 3
Forgot race.
[01:07:49:13 - 01:07:53:05]
Speaker 1
Yeah, everybody's back to, to ignorance 100.
[01:07:54:13 - 01:07:57:07]
Speaker 1
And the only way out of that is doing some stuff.
[01:07:57:07 - 01:07:58:03]
Speaker 3
Exploration.
[01:07:59:14 - 01:08:05:02]
Speaker 3
I've gotten some spicy questions from folks like especially from SREs, you know, who are like, Hey, my value add for
[01:08:05:02 - 01:08:31:20]
Speaker 3
my entire career has come from this part of the software development lifecycle that we're saying is gone. Yeah. And I was feeling pretty anxious for SREs. But SREs seem to be doing great. SREs have always been very outcome oriented. Right. The people who I see struggling right now are the people, the software engineers who have taken the most pride in their craft. Yeah.
[01:08:31:20 - 01:08:32:23]
Speaker 4
Yeah.
[01:08:32:23 - 01:08:58:21]
Speaker 3
You know, and to an extent, I get it because of your entire career, the best proxy that we have had for maintainability has been, these are called beautiful. Is it beautiful? Does it make sense? Is it well architected? Is it, you know, idiomatic? We've rewarded people for this for so long, but I think it's really hard to give that up.
[01:08:58:21 - 01:09:35:15]
Speaker 1
Yeah. And an ego kind of definitely way. I mean, I wrote two books about writing code that people can read, implementation patterns and small talk, best practice patterns. And, and I took and still take great satisfaction in taking something, looked at it and I don't really understand it. And then just tweak and tweak and massage and tweak and then the fundamental structure of this logic is now completely revealed to anybody who wants to read it. Nobody's reading it. It just doesn't, that doesn't matter.
[01:09:36:23 - 01:09:39:04]
Speaker 3
Yeah. Yeah, it's interesting.
[01:09:39:04 - 01:10:00:00]
Speaker 1
And we have to get used to that. Now it's, it's not that understandability doesn't matter, but that, that kind of understandability to a human doesn't have leverage. If I could write code that the genie could read and humans couldn't read, yeah, exactly.
[01:10:00:00 - 01:10:02:17]
Speaker 3
But yeah, yeah,
[01:10:02:17 - 01:10:04:06]
Speaker 1
that is the right way.
[01:10:04:06 - 01:11:30:13]
Speaker 3
Yeah. Or so I have, so my ex co-worker David Paul, he now, he runs everything related to pull reviews at GitHub, but a decade and a half ago we were both at Parse mobile back into the service and, and I'm working over here and a series, right? And David's working over here on APIs and my, my somewhat tongue in cheek line, but not really software engineers has been like, dude, you don't know what is happening just by reading the code. Like, you know, you might think, you know, what's happening, but you don't know what's happening. And like, we learned this intensely at Parse because we had a million mobile apps running. We never got to say hi to any of those people. And I think right now it's like, I had forgotten that this is such a big deal because of that experience. But the number one debugging technique that almost every team has is, oh, something broke by the person who wrote it. Whoever, whoever wrote it, even if they wrote it a year ago or a minute ago, they are the one who's going to be able to figure it out the fastest and fix it. And what happens if you don't have that person, right? And I've been like, never, you know what you think the code is going to do, but unless you have instrumented it, also the code is not the system, right? The code, you might be like, I understand this, but you put it in one system versus a different system.
[01:11:32:04 - 01:11:51:19]
Speaker 3
And so I was ranting about that on Twitter. And David dived in, he's like, well, yes, but he's like, polar costs were never supposed to be about pointing bucks. He's like, they're about, do I want this in my product or not? And I was like, ah, fuck, you're right.
[01:11:54:02 - 01:12:06:08]
Speaker 3
There's so much there. Like, I do feel like what is it Boris, what's his face or I could do this like coding is basically soft now. And while it's a bit hyperbolic, I know what he means. Code review.
[01:12:07:14 - 01:12:11:12]
Speaker 3
What's your sense of what, where we are when it comes to code review now?
[01:12:11:12 - 01:12:12:12]
Speaker 1
My sense?
[01:12:12:12 - 01:12:14:01]
Speaker 3
Yeah, what does it mean these days?
[01:12:14:01 - 01:12:22:05]
Speaker 1
Oh, I've been anti asynchronous blocking code review for a very long time.
[01:12:23:12 - 01:12:25:07]
Speaker 1
I think it's applied
[01:12:27:20 - 01:12:35:12]
Speaker 1
to the, as geeks, we have a tendency to want to avoid human contact as a population,
[01:12:36:12 - 01:12:42:03]
Speaker 1
individual exceptions, sure, but as a population. And I've seen
[01:12:42:03 - 01:12:57:19]
Speaker 1
The fetishization of code review, blocking asynchronous code review. Because then I don't have to talk to people because I submitted my PR and now I get to wait until I get the feedback electronically.
[01:12:59:01 - 01:13:41:08]
Speaker 1
And extreme programming took the contrarian position that even if we're not biased towards human team and communication, there are ways to make that safe. And if we do, there's huge advantages to having social norms that grow and evolve naturally to having knowledge shared around and about. And so I'm not a big fan of code review. Here's the, I have a someone, a side project, everybody's got so many side projects now. So one of my side projects is called Wait, what?
[01:13:43:06 - 01:13:54:13]
Speaker 1
And the idea is the you make some prompts, the genie gives you the finger guns, you know, that's your boss. What a great idea. Awesome. I wouldn't have thought of that.
[01:13:56:04 - 01:14:00:11]
Speaker 1
And now I'm done. And wait, what?
[01:14:01:16 - 01:14:23:10]
Speaker 1
That's the question that I always have at that moment. So I want to build, it is in a sense a review tool, but it's it's a way for me to regain my situational awareness. I know what the pieces are. I know what just changed. I'll have some sense about what might change next.
[01:14:25:06 - 01:14:57:04]
Speaker 1
You know, did do we have new tests that are passing now that didn't used to? Did we have a test that was failing? And now it's passing? Did we need to make some structural changes? I want to be able to at a glance see the answers to all those kinds of questions. So that is review in a sense. But it's not this ritualistic. I put a piece of paper on the table and I slide it over to you. And then I wait for you to make your notes on it and slide it back over to me as comforting as that might be. Yeah.
[01:14:57:04 - 01:15:06:13]
Speaker 3
code review has just been so overloaded for so long. We use it as a we use it for pedagogical
[01:15:06:13 - 01:15:07:14]
Speaker 2
pedagogical.
[01:15:08:14 - 01:15:11:00]
Speaker 3
Thank you. So, my goodness.
[01:15:12:19 - 01:15:21:05]
Speaker 3
We use it for teaching people, you know, we use it for shared context. We use it for trying to hold a standard of, you know,
[01:15:22:19 - 01:15:23:00]
Speaker 3
consistency.
[01:15:24:00 - 01:15:45:23]
Speaker 3
But what really struck me about, you know, after I really I'm not a product person, I think I really tell that. But I did start thinking, yeah, there were those two reporters who vibe coded a clone of the slack out a given day. And it kicked off such a spiral of, oh, my God, it's software engineering. I was like, oh, come on.
[01:15:47:11 - 01:15:50:19]
Speaker 3
Please. Like my immediate thought is
[01:15:52:12 - 01:16:00:06]
Speaker 3
kind of billion people use that crappy little thing. Just come on. But after David posted his piece, I was like,
[01:16:01:14 - 01:16:46:10]
Speaker 3
if I am using that, do I want the app to just subtly change in small little ways all the time when I'm using it like this menu thing? No, that's actually just as bad. It's just terrible. Like there's a mental model. And the work that slack has put I mean, they famously so detail oriented, you know, your mental model of how I mean, when it's API's talking to API's, which I have always been a bit back end supremacist. I'm not embarrassed to admit that I and this is kind of embarrassing, I think, you know, because front end famously in the industry has been kind of dismissed, you know, so real, real engineering, you know,
[01:16:46:10 - 01:16:49:11]
Speaker 1
I've never heard it called that. Yeah, okay. Okay.
[01:16:49:11 - 01:16:53:01]
Speaker 3
You know what I mean? I know it's mushy. And recently,
[01:16:54:05 - 01:16:55:07]
Speaker 3
I started
[01:16:56:15 - 01:17:03:02]
Speaker 3
talking with my with my mouth, talking to other engineers, and blew my little mind.
[01:17:04:04 - 01:17:06:17]
Speaker 3
And I'm realizing, oh,
[01:17:07:21 - 01:18:03:00]
Speaker 3
actually, I think front end is much more complex than back end is these days, because they're dealing with chaotic human beings, times chaotic devices and rendering, and lengths of time that are like, oh, shit, every single one of those things is very open ended and complex. Whereas back end, we're like, it's mostly API's talking to other API's. And in the reason, I know, I'm blowing your mind. I'm trying to lecture you. But like, persistence problems have this nasty little habit of being permanent, or, yeah, or very hard to back away. And so we've Yeah, irreversible. Yeah. And so we put all this discipline around it. But it is absolutely it easier in some ways for the human if they can keep that discipline. And the front end side, yeah, they don't have to deal with permanent persistence in the same way. But testing it is virtually impossible.
[01:18:04:00 - 01:18:10:13]
Speaker 1
Yeah. So my observation about front end versus back end is every time I completed back end tasks,
[01:18:12:03 - 01:18:16:23]
Speaker 1
I get like point five new tasks that I need to complete on average.
[01:18:18:04 - 01:18:28:12]
Speaker 1
You know, I'm in the middle of doing something and go, I got to do this other thing. But it happens like half the time. Yeah. On the front end, every task I complete is five new times.
[01:18:28:12 - 01:18:29:08]
Speaker 3
Yeah.
[01:18:29:08 - 01:18:38:08]
Speaker 1
So the tree is just so much bushier. Yeah. And, and then you have a much harder. How good is good enough problem to solve?
[01:18:41:05 - 01:19:09:12]
Speaker 3
You know, it's so I wrote a blog post, this is March, I think it was last year, July, where I, it was about disposable code versus durable code. And at the time, you know, nine months ago, I think it was still reasonable to be like, the best way we know of to change durable code is by taking what is known to be good and making small iterative changes to it.
[01:19:09:12 - 01:19:11:21]
Speaker 1
Keeping it good at every step.
[01:19:13:04 - 01:19:13:09]
Speaker 3
Exactly.
[01:19:14:10 - 01:19:51:14]
Speaker 3
And it is, I think very clear to me now that this place is not disposable versus durable. There are also new workflows for durable code that are emerging that are based on a different model that are about you know, it's less clear exactly. But you know, we're worth, we're going it's pretty quick, but it's not clear to me that the new model of auto generating based on specs or whatever will replace all durable code models. Is it still being overly pessimistic or do you?
[01:19:51:14 - 01:20:11:15]
Speaker 1
Well, I think there's a big decision point that's not clear. This is one of these nobody knows, right? Nobody knows where do I have a thing? I have some specification of thing then I have the thing. And when I want a new thing, I go back to the specification and regenerate. That's the kind of disposable model.
[01:20:13:01 - 01:20:18:23]
Speaker 1
Does that work? I mean, anytime I've heard somebody describe that I just think you've never had a billion users.
[01:20:20:17 - 01:20:35:12]
Speaker 1
If you think you can do that. Yeah, exactly. It's just not gonna. But it's so legit. I mean, I could just be an old fuddy duddy and not get it and that's fine.
[01:20:36:22 - 01:20:46:18]
Speaker 1
Or do you have a thing that works and then you make changes to it and the new thing works but in a slightly different way. And so you have this path.
[01:20:46:18 - 01:20:52:15]
Speaker 3
And maybe you turn it around and claw it and you're like, well, don't accept unless it can perform within you know, this or something.
[01:20:52:15 - 01:21:03:06]
Speaker 1
Yeah. So I don't know which of those is going to end up predominating or what the landscape is like, like where you would choose one versus the other.
[01:21:03:06 - 01:21:28:06]
Speaker 3
My heuristic that I have used for a long time and we're talking 15 years of creating it is that the closer you get to laying bits on a disk, the more parent would be. For, okay, you know, and some people tell me that it's about dependencies and also I think persistence is different. But like operating systems, databases,
[01:21:29:09 - 01:21:37:21]
Speaker 3
the most rigorous testing I've ever done in my life has all been for stuff that because you don't want variance.
[01:21:37:21 - 01:21:43:21]
Speaker 1
You want, you know, like, and you have this, this reversibility problem.
[01:21:43:21 - 01:21:44:22]
Speaker 3
It's a reversibility problem.
[01:21:44:22 - 01:21:58:17]
Speaker 1
And then if you can't just press reload the browser page, and nothing bad has ever has happened ever. That's, that's one, that's one level of reversibility. It's fine.
[01:21:59:21 - 01:22:00:05]
Speaker 1
But yes,
[01:22:01:13 - 01:22:14:01]
Speaker 1
the more you the you put bits on disk or you speak with the outside world, you know, if you report the wrong numbers to the IRS, there's a also very true. That's a really bad thing to do.
[01:22:14:01 - 01:22:16:07]
Speaker 3
Whether crashing problem.
[01:22:16:07 - 01:22:17:16]
Speaker 1
Yeah.
[01:22:18:21 - 01:22:34:10]
Speaker 1
what are you doing? I'm going to make an assumption. I'm going to assume that you agree with me. But tell me if you don't, we're gonna have to fart around a bunch. Don't find out stuff. What are you doing to play?
[01:22:36:01 - 01:22:44:07]
Speaker 3
So I got into I have been a book hoarder for my entire life.
[01:22:45:18 - 01:22:48:17]
Speaker 3
Last year, I finally made a delivery. Okay.
[01:22:50:08 - 01:22:55:04]
Speaker 1
What kind of things?
[01:22:55:04 - 01:23:00:17]
Speaker 3
Just workflows for like, you know, syncing my library to good reads,
[01:23:01:19 - 01:23:30:16]
Speaker 3
trying to get a workflow where when I plug my device in it, it populates it correctly, and takes all the annotations off and uploads that and there's so many. It's one of those, oh, God, someone who loves units wrote this piece of software. You know exactly what I mean. But there's so many ways to do everything. And so I'm just like, yeah, everything's a giant mess. And so I'm happy to have fun with that.
[01:23:30:16 - 01:24:04:23]
Speaker 3
Also, I have been maintaining a list of phrases that I never want to read again and feeding them into grease monkey things like it's not X it's Y or things like we're an inflection point or just like all these things you're just like you were I know, I know who wrote that. They're all over. They're so exhausting, dude. They're so bad. And I really feel like you remember how like, well, you're working on Facebook like 10 years ago, remember when the X will shock you?
[01:24:04:23 - 01:24:05:23]
Speaker 1
Yeah, that's right.
[01:24:05:23 - 01:24:08:23]
Speaker 3
All this stuff is going to age very poorly, I think.
[01:24:08:23 - 01:24:10:12]
Speaker 1
you're still writing.
[01:24:11:20 - 01:24:13:23]
Speaker 1
You have a blog.
[01:24:15:02 - 01:24:17:14]
Speaker 1
Yeah, you're working on a new edition of a book.
[01:24:18:15 - 01:24:19:23]
Speaker 3
Almost finished. Yeah.
[01:24:22:09 - 01:24:29:06]
Speaker 3
Almost finished for the second time. Oh, sorry. Sorry. Stop it. That's not very nice.
[01:24:29:06 - 01:24:34:04]
Speaker 1
Because I was almost finished with one too. And I had to throw it all away.
[01:24:34:04 - 01:24:39:10]
Speaker 4
But
[01:24:39:10 - 01:24:40:05]
Speaker 1
here's a thing I've been thinking about.
[01:24:42:01 - 01:25:05:01]
Speaker 1
Books, like the cycle of a book, just looks it just seems like it's too long. I love writing a book, because it forces me to confront my own ignorance to understand, okay, what is the dependency graph of these concepts? Yeah, I can have the whole graph kind of in my head.
[01:25:05:01 - 01:25:09:10]
Speaker 3
That is, yes, there's this famous New Yorker editor,
[01:25:11:04 - 01:25:26:17]
Speaker 3
William McKeene. I love his book. But he wrote one of the great classics on writing and editing. And then after that, he wrote another book called Right to Learn. He's like, writing is just William Zincher. Okay.
[01:25:28:15 - 01:25:51:08]
Speaker 3
Right to learn. Writing is just thinking on paper. I don't know if that's true. I think I might have a lot of thoughts and jumble and I have great interesting conversations with people. I don't learn something until I've written it down. And then I'm staring at like, well, that is flimsy. Yeah, it makes sense. We have to figure this out, you know?
[01:25:51:08 - 01:25:58:09]
Speaker 1
Yeah. Or I say, well, of course, we assume we understand X. And then I realized, I haven't talked about X yet. Yeah.
[01:25:59:12 - 01:26:05:02]
Speaker 1
So I got to go change the table of contents and that causes me to change something else.
[01:26:05:02 - 01:26:18:09]
Speaker 3
You know, I started writing the second. And like, at first, I kind of didn't writing about observability for 10 years. Is there anything? Well, yes, there was a lot. There was a lot left to learn, you know?
[01:26:19:17 - 01:26:26:20]
Speaker 3
I hate writing. I can't actually say I love that you love writing. I can't say I love it, but I don't know how to think without it.
[01:26:26:20 - 01:26:43:10]
Speaker 1
Yeah. So the role of the book is not as a body of knowledge for someone to absorb as much as a series. It's kind of like
[01:26:45:06 - 01:27:48:23]
Speaker 1
the puzzle solving video games where, you know, you get to a certain place and you got to solve a puzzle to open the door. So for the author, you get that kind of experience. And then how do you explain something? I had to explain. So I've been working on this series of books about software design since 2005 and tried to explain coupling and collision and tried to explain and tried to explain. Finally came up with a good example. You mentioned you have a box of gears but they're all disconnected from each other and you turn one and it turns really easily. And then you add another gear and it hooks up with something. And now when you turn one, two of them turn and so it gets a little bit harder. But then the more gears you add and the more things that mesh with each other, you still think you're just turning one little gear, but you're actually turning the whole box and it gets harder and harder. Yeah. And I was like, oh, well, that just took me 20 years.
[01:27:50:00 - 01:28:00:02]
Speaker 1
Thank you very much. Oh boy, I know that feeling. So I published chapters of my books on my substat. Yeah. Are you doing that as well?
[01:28:01:04 - 01:28:11:09]
Speaker 3
So doing it with O'Reilly. So they have some rules, but we are going to be drip, drip, dripping out chapters over between now and June.
[01:28:11:09 - 01:28:14:14]
Speaker 1
Yeah. Yeah. I think that hybrid model where,
[01:28:15:16 - 01:28:32:00]
Speaker 1
where as the author, you, yeah, as the author, you, you get the experience of writing to learn as a reader, you get to, you get earlier and more frequent exposure and you get to provide feedback.
[01:28:32:00 - 01:28:44:14]
Speaker 3
Absolutely. Yeah. Yeah. I grew up on O'Reilly books. You know, I'm from Idaho and that's literally how I learned software. So I don't think I would be,
[01:28:45:16 - 01:28:53:23]
Speaker 3
I don't think I, O'Reilly has their way of doing things. And so I comply with O'Reilly's way because it's O'Reilly. But if it was just me writing, I would absolutely do that.
[01:28:55:11 - 01:28:55:16]
Speaker 1
Okay.
[01:28:55:16 - 01:28:58:13]
Speaker 3
They don't let you do that until, because they,
[01:28:59:15 - 01:29:02:09]
Speaker 3
we don't need to get O'Reilly. Yeah.
[01:29:04:13 - 01:29:05:16]
Speaker 3
They've been very good to me.
[01:29:05:16 - 01:29:08:02]
Speaker 1
Good. They've been good to me too.
[01:29:09:03 - 01:29:10:22]
Speaker 1
And, and that is how I'm doing it.
[01:29:10:22 - 01:29:16:02]
Speaker 3
Oh, really? Damn it. I should have asked for the 10 seconds. I know. It's nothing.
[01:29:16:02 - 01:29:19:03]
Speaker 1
I'm going to be in so much trouble now.
[01:29:20:19 - 01:29:21:09]
Speaker 1
Sorry, editors.
[01:29:23:16 - 01:29:24:01]
Speaker 3
Oh well.
[01:29:27:02 - 01:29:28:21]
Speaker 1
But it's, I think,
[01:29:30:20 - 01:29:59:07]
Speaker 1
episodic writing blogs, that kind of writing is, it's such a good match for the pace of change right now. Yeah. Cause you get, you get time, you get a day, five days, whatever, to step back and think about what's going on. Put some context around it. Make sure if you have some assumptions, you can validate them and then get it out there and have it immediately available. Yeah.
[01:29:59:07 - 01:30:00:15]
Speaker 3
Yeah. Totally.
[01:30:01:17 - 01:30:19:09]
Speaker 3
Yeah. I saw it just earlier today. Someone put something on X about how there's been more change in the developer tooling ecosystem in the past year than on all previous 10 years. And I'm like, well, yeah, it's actually seems like an understatement.
[01:30:19:09 - 01:30:20:12]
Speaker 1
More than 10 years.
[01:30:20:12 - 01:30:21:12]
Speaker 3
Yeah.
[01:30:21:12 - 01:30:22:17]
Speaker 1
I was around for it.
[01:30:22:17 - 01:30:45:14]
Speaker 3
I, I would say, yeah, it's, it does feel like the last three years have been like going from FreeBSD on bare metal to AWS in the cloud. At the same time as we've gone from like machine language to Ruby on rails. In three years. Both of them.
[01:30:45:14 - 01:30:50:00]
Speaker 1
So, so what does that do? I'm worried about the junior.
[01:30:50:00 - 01:30:52:18]
Speaker 3
I am so worried about the junior.
[01:30:52:18 - 01:31:02:09]
Speaker 1
I'm not worried about them in the, so I get these questions from a sophomore CS student. Oh, God. I don't know what to tell them. Oh, I do.
[01:31:02:09 - 01:31:03:13]
Speaker 3
What do you tell them?
[01:31:03:13 - 01:31:06:07]
Speaker 1
What an awesome time to become a programmer.
[01:31:07:08 - 01:31:44:16]
Speaker 3
I think that could be true, but that depends on somebody, someone being willing to hire them. I was watching, I don't usually watch podcasts, but for whatever reason, recently, a couple of couple of nights ago, I was watching one of the anthropic co-founders was on as replied. And he just straight up was like, Oh yeah, we don't even hire. There's no point in hiring anyone except the very super senior. We, no use for anyone in the middle. Let alone the genius. We want people who can just like set up 20 agents, then go for a run, have lunch, come back, check on their agent. He's like, there's no point. Anyway, be easy. You literally just like, no one else has any value to it. And I'm just like,
[01:31:45:23 - 01:31:54:19]
Speaker 3
wow. Did you hear yourself? Yeah. Well, I mean, and the thing is, it doesn't have to be true for all the smart people in the industry to do it.
[01:31:54:19 - 01:32:02:23]
Speaker 1
Correct. To follow that. And for the kids to get worried and really smart people leave the field.
[01:32:02:23 - 01:32:25:05]
Speaker 3
I'm worried about it with, with the layoffs too. People like, you know, block and gesture like, well, we're so gosh down productive now. We don't eat half our people. And I'm just like, really? Like these narratives have a way of taking up. Oh yeah. And it has to be true. Yeah. It's really, it's, that's what worries. I know kids are fucking smart. This is a great time to do technology.
[01:32:25:05 - 01:32:33:10]
Speaker 1
And they are going to figure out these tools in ways that we can't even imagine. Exactly. And they are going to smoke us when they.
[01:32:33:10 - 01:32:38:20]
Speaker 3
I believe you are right. But someone has to be willing to give them a shot. Right.
[01:32:38:20 - 01:32:44:23]
Speaker 1
Well, I think the juniors are more, more profitable now than they've ever.
[01:32:44:23 - 01:32:47:13]
Speaker 3
I have seen some of your recent writing and I was just like,
[01:32:48:14 - 01:32:48:19]
Speaker 1
yes,
[01:32:50:00 - 01:32:59:21]
Speaker 1
you have the opportunity. If, if they're, if they're encouraged, if they're incentivized to optimize for learning instead of production,
[01:33:01:18 - 01:33:23:02]
Speaker 1
they're going to learn so much faster and they're going to be native users of these tools who will use them in much more effective ways than us old one finger at a time people who, you know, we've converted and we feel really good because we can wear the AI hat now, but it's not the same as that being your native language.
[01:33:23:02 - 01:33:24:10]
Speaker 3
Yeah. I agree.
[01:33:24:10 - 01:33:48:06]
Speaker 3
I agree. And it's also, I think, and I think another lens that I've looked at it through, like it's just so clear that the way that we were brought up and taught to be technologists, it's just not going to happen again. Doesn't mean that we're not going to have technologists. They're not going to, they're not going to, and this happened before. I think an interesting sort of parallel is with infrastructure.
[01:33:49:14 - 01:33:55:18]
Speaker 3
You know, I mean, it used to be, you and I are probably same generation ish. I grew up
[01:33:56:23 - 01:34:04:02]
Speaker 3
getting, getting called in the middle of the night, having to take a taxi downtown with the power switch on the my SQL server, you know,
[01:34:04:02 - 01:34:05:22]
Speaker 1
and you get boxes from sun,
[01:34:05:22 - 01:34:27:15]
Speaker 3
you have to rack them and stack them, you know, and, and, and I can, and our people, our engineers who haven't learned, you know, Linux and all that stuff, are they as good at debugging systems as people like us? No, probably not. But has that stopped them from making a very profitable career? No.
[01:34:28:23 - 01:34:34:09]
Speaker 3
You know, do they learn what they need to learn when they need to learn it? If someone like us isn't around?
[01:34:34:09 - 01:34:41:04]
Speaker 1
So what can we, what can we do to encourage the juniors?
[01:34:41:04 - 01:35:03:02]
Speaker 3
So the main thing that I don't, I don't actually think they need encouragement. I think this is something I wrote recently was that every place that I know of, this is a small sample, everything's sample bias, but like the places that where I have heard companies investing in junior engineers in the last year or two,
[01:35:04:19 - 01:35:18:13]
Speaker 3
none of those places had this driven by managers, or execs, or accountants. It all came, no staff engineers or principal engineers. It all came from senior engineers,
[01:35:19:13 - 01:35:52:15]
Speaker 3
okay, people who were in those shoes in the last five to 10 years, who were just like, just like brought it up, and didn't let it go. And they were just like, I'm willing to do the interviewing, the recruiting, the mentoring that, and I really do think, I think that's smart, actually, you want to be mentored by someone who was in your shoes not long ago. Yeah. Right. But like those are the people who I see really agitating for change, engineers often act like they don't have power or authority over things like, but if you make yourself annoying enough,
[01:35:53:21 - 01:35:58:16]
Speaker 3
often, even if you volunteer to do some of the work, I've seen it happen in many places.
[01:36:00:02 - 01:36:07:06]
Speaker 1
So the story I heard that the corroborates this is, it was a large tech employer.
[01:36:08:07 - 01:36:16:02]
Speaker 1
And what they noticed is of the senior engineers that got promoted to principal, all of them had interns.
[01:36:18:07 - 01:36:24:17]
Speaker 1
makes perfect sense. Perfect. Because you get you get the interchange. I've spent a lot of my career coaching engineers.
[01:36:24:17 - 01:36:25:02]
Speaker 2
Yeah.
[01:36:25:02 - 01:36:29:05]
Speaker 1
And I always say, as a coach, you get more out of it than the student does.
[01:36:30:05 - 01:36:34:13]
Speaker 1
Which means the student is just screwed until they become a coach. But that's okay.
[01:36:36:18 - 01:36:55:23]
Speaker 1
And so, as I'm imagining being the senior in that interaction, you're going to see the juniors do some stuff that's wrong and you're going to correct them. But you're also going to see them do some stuff that's really creative and out of the box and you're going to copy them. And then they're going to go back to school and you can claim credit.
[01:36:58:02 - 01:36:58:07]
Speaker 1
Yeah.
[01:36:59:08 - 01:37:26:17]
Speaker 1
So I think that there was a missing middle to technology leadership right now where for the past 30 years, we haven't known what the answers were. And if you had one of the canonical problems, reliability, performance, so you just like you press play on this known CI CD test driven development,
[01:37:27:20 - 01:37:29:16]
Speaker 1
pairing, frequent integration,
[01:37:31:00 - 01:37:38:15]
Speaker 1
refactoring, you just press play. And all of a sudden, so that middle layer of leadership,
[01:37:40:11 - 01:38:12:02]
Speaker 1
they learned the playbook that somebody else wrote. But that's a very different skill than writing a playbook from scratch. And having been through the Rackham and Stackham era, we didn't know, we didn't have the playbook. We didn't know what the answers were. And the so there's a set of skills that people in our generation have about how to flail around productively, how to play, how to experiment, how to stop experimenting.
[01:38:12:02 - 01:38:14:17]
Speaker 3
How to be comfortable with things being super fucked up.
[01:38:14:17 - 01:38:31:04]
Speaker 1
Yeah, absolutely. The juniors don't know any better, but that middle layer of people who you would excel because you learned more playbooks more thoroughly, they don't they don't know about how to deal with that same level of ambiguity and uncertainty.
[01:38:31:04 - 01:38:55:23]
Speaker 3
There's something real to that. And I think what I worry about is we also like, in the early days of the internet, the business folks were also just kind of like, is there a there? There was a lot of flailing. And I worry that we're going to short circuit the exploratory phase to try to figure it out profit, profit, profit.
[01:38:55:23 - 01:38:58:21]
Speaker 1
Yeah, what is the ROI of these tokens?
[01:38:58:21 - 01:38:59:16]
Speaker 2
Yeah.
[01:39:01:06 - 01:39:03:05]
Speaker 1
Great question. Bad time.
[01:39:03:05 - 01:39:04:00]
Speaker 3
Yeah.
[01:39:05:09 - 01:39:05:18]
Speaker 3
Yeah.
[01:39:07:06 - 01:39:11:00]
Speaker 3
I love that the question you asked me was how are you playing with these things?
[01:39:12:05 - 01:39:17:02]
Speaker 3
You know, it's been so interesting just recently seen.
[01:39:21:04 - 01:39:30:08]
Speaker 3
We talked earlier about how the split between people who are thriving, people who are not is outcome oriented versus the love of the craft.
[01:39:32:06 - 01:39:46:22]
Speaker 3
But I think I've encountered a lot of these are not people who are talking about their experience in social media, but I've encountered a lot of people in real life who are not having fun. They're stressed. They're just stressed.
[01:39:47:22 - 01:39:54:23]
Speaker 3
And I absolutely understand why like the world is on fire.
[01:39:56:01 - 01:40:08:19]
Speaker 3
And things are changing so fast. But I think that I don't see this advice that is like very easy to give. But
[01:40:10:14 - 01:40:30:10]
Speaker 3
you have to run to the way you have. I think the best way to get a handle on your anxiety is to realize how easy it is to get started. And there's like, what makes me happy might not make you happy. Which is why you can't wait for someone to be like, here's your assignment. Go on. Right.
[01:40:33:22 - 01:40:35:02]
Speaker 3
You have to be.
[01:40:36:07 - 01:40:44:10]
Speaker 3
Yeah. I always feel I don't like when people tell me what to do. So I'm always like kind of resistant. But you got to try it, dude. You've got to see if you can find the fun.
[01:40:44:10 - 01:40:45:02]
Speaker 4
Do a thing.
[01:40:45:02 - 01:41:04:04]
Speaker 3
Find the fun. Do the thing. Find the fun. Because if you can tap into it, you'll stop feeling like, oh, the cool kids are over there and doing this thing. And it's going to come get me. And I'm just resentful about it because I spent years learning that thing. And it's terrible. And what is going to happen? And is singularity going to come? And is transhumanism going to be like,
[01:41:05:15 - 01:41:18:12]
Speaker 3
can I just say once I do not believe that AG AGI is coming. I if I improve it wrong, I will happily say it. But like, look, all the computing power in the world can't tell us what the weather is going to be like next month.
[01:41:19:15 - 01:41:42:20]
Speaker 3
Can't predict the outcome of the social interaction with three people. That is it. That is a different thing. Like the reason that software is a killer app for generative AI is because it is something that we built that is like for human beings to understand what their opinions on human minds, you know. And so I
[01:41:43:23 - 01:41:51:09]
Speaker 3
I read a really great book last weekend by Matt Becker, and it's called More Everything Forever.
[01:41:52:10 - 01:41:56:15]
Speaker 3
He is a science journalist with a PhD in astrophysics.
[01:41:57:21 - 01:42:19:13]
Speaker 3
And he just very methodically picks up and addressed everything from like, altruism and long termism and all these silicon dollar pretty much like this is all these things are the one thing that we know about exponential changes is they don't last forever. Or it's like, it's the universe.
[01:42:19:13 - 01:42:29:04]
Speaker 1
It's a legit. It's a logistic curve. There's some reinforcing loop. And then there's some inhibiting loop that keeps it from growing forever.
[01:42:29:04 - 01:42:49:21]
Speaker 3
So I think I think for people who I think there is an especially broad spectrum of possible features right now. Yes. But I also think it's not as broad as the people who are getting paid a lot of money to talk about all these beautiful possibilities.
[01:42:49:21 - 01:42:52:09]
Speaker 1
Or the doom and gloom.
[01:42:52:09 - 01:42:54:05]
Speaker 3
Or the doom. Yeah.
[01:42:55:13 - 01:43:05:00]
Speaker 3
If anyone builds it, everyone, come on, take a logic class. Like, this is the biggest bullshit I have ever seen.
[01:43:06:05 - 01:43:06:21]
Speaker 3
Ridiculous.
[01:43:06:21 - 01:43:16:15]
Speaker 1
It's interesting how people can be very smart in some ways and very not smart in other ways. I mean, I include myself in that 100%.
[01:43:16:15 - 01:43:23:23]
Speaker 3
Same. The oldest fantasy of our species is what if I could never die? And it's just.
[01:43:25:10 - 01:43:32:12]
Speaker 1
Oh, so many business models have been based on that. So many. And people are still dying. So still dying.
[01:43:32:12 - 01:43:38:00]
Speaker 3
You know, maybe AGI will have apparently the smart money is on 2027.
[01:43:40:01 - 01:43:53:08]
Speaker 3
I, you can quote me on this and I love being wrong. So for Mama, I do predict that in 2027, we start to see a lot of claims that, oh, that was API. That just happened. We just didn't define it. You just didn't understand the definition.
[01:43:53:08 - 01:43:54:04]
Speaker 1
The goalposts.
[01:43:54:04 - 01:44:05:21]
Speaker 3
You know, it can, yeah, no, I don't think it's gonna, I don't think it's gonna happen. The history of technological innovations tell us that low hanging fruit gets snapped up first.
[01:44:07:15 - 01:44:28:09]
Speaker 3
This is what's happening. I think that we are. And diminishing returns. And diminishing returns after that. I think that software is the killer app for generative AI. I think that what we've seen, you know, coding being basically solved. This is, this is, this is going to be something to deal with the agentic stuff is just kicking off. But like,
[01:44:29:13 - 01:44:34:10]
Speaker 3
I think that's the pig in the Python. I think that's the tip in the Python. And I think it's going to take,
[01:44:35:13 - 01:44:37:21]
Speaker 3
I think it's going to take years for us to digest it.
[01:44:37:21 - 01:44:41:17]
Speaker 1
Okay. Well, you heard it here first. You heard it here first.
[01:44:41:17 - 01:44:49:19]
Speaker 1
I, I have one more question for you, which is, what frightens you about the current situation? What wakes you up at night?
[01:44:53:15 - 01:44:56:15]
Speaker 3
Frightens me about the current situation is
[01:45:09:18 - 01:45:19:04]
Speaker 3
I am a capitalist. I think any kid who grew up third floor has some respect for capital, money in general.
[01:45:20:19 - 01:45:29:13]
Speaker 3
But I, I just worry about, worry about our governance. I worry about,
[01:45:35:14 - 01:45:40:17]
Speaker 3
worry about this sort of Ponzi, the unification of just that, is that,
[01:45:44:10 - 01:45:46:13]
Speaker 3
this is very original, but
[01:45:50:17 - 01:45:55:10]
Speaker 3
the whole model of first pass the post voting, that's a sub for extremism.
[01:45:56:18 - 01:45:58:10]
Speaker 3
I don't see how that,
[01:46:01:06 - 01:46:24:12]
Speaker 3
I don't see how that diverging of language and values. I don't see how it comes to an end. And that makes it very hard to build the kind of predictable, stable society that can make long-term decisions about our shared history.
[01:46:26:12 - 01:46:30:15]
Speaker 1
What are you doing about it?
[01:46:35:18 - 01:46:35:22]
Speaker 3
So,
[01:46:38:10 - 01:46:46:23]
Speaker 3
I, last election day, I was in Scandinavia over a election day. I thought Kamala had it in the bank.
[01:46:48:10 - 01:46:53:23]
Speaker 3
I'm just like, damn, she is good. She's incredible. That first debate, you know, I'm just like,
[01:46:55:06 - 01:46:59:00]
Speaker 3
surely everyone's here in Scandinavia. Not too worried. And I was there,
[01:47:01:07 - 01:47:11:06]
Speaker 3
obviously, it's their day. It's nice to like, while I'm both being counted overnight, I'm there and I'm just like, the red wave, you know, and I felt very alone.
[01:47:13:02 - 01:47:22:13]
Speaker 3
And I got up the next morning and I went, I was supposed to keynote at another conference and I was crying and I couldn't stop crying. And I just,
[01:47:25:00 - 01:47:30:14]
Speaker 3
I got up on stage and I was still crying. I was just like, I don't know what to say. I'm sorry.
[01:47:33:01 - 01:47:33:17]
Speaker 3
I'm sorry.
[01:47:34:22 - 01:47:39:22]
Speaker 3
And some woman pipes up. We get through this together.
[01:47:39:22 - 01:47:41:00]
Speaker 4
Yeah.
[01:47:41:00 - 01:47:46:02]
Speaker 3
It was like shit. And I found out afterward, she was from Ukraine. So like, I'm like,
[01:47:47:06 - 01:47:55:20]
Speaker 3
what standing do I have? You know? So I came home, but that was when I was just like,
[01:47:58:19 - 01:48:18:23]
Speaker 3
I've always been kind of a compulsive read. I read journalism. I read the news. I'm always picking up my phone. I'm like, why is it like, I can't actually, I'm not going to be able to something this year. I do that. So the goal that I set for myself in 2025 was to stop picking up my phone, start picking up my Kindle,
[01:48:20:05 - 01:48:36:03]
Speaker 3
because I don't know what I can do about politics, but I know that if I can't function, it's not going to hurt the people who I don't think we should about going to hurt the people who are dependent on me.
[01:48:37:22 - 01:48:42:08]
Speaker 3
So I try to manage my attention and focus.
[01:48:42:08 - 01:48:51:03]
Speaker 1
I think just that moment of vulnerability on stage is a thing you're doing about it. Do not pretend like this is normal.
[01:48:51:03 - 01:49:04:13]
Speaker 3
I think we have a responsibility when we need to function, you know, but I also think that you can take that too far
[01:49:06:20 - 01:49:15:05]
Speaker 3
and not expose yourself to the things that are happening in the world, endose it, you know? And so that's the balance that I think we need to
[01:49:16:06 - 01:49:18:19]
Speaker 3
feel things when, you know, the Minneapolis,
[01:49:19:20 - 01:49:24:13]
Speaker 3
you need to let yourself fall apart sometimes and just to understand what's happening.
[01:49:24:13 - 01:49:26:12]
Speaker 1
Be angry or whatever your anger is.
[01:49:26:12 - 01:49:32:08]
Speaker 3
And then what am I feeling? What can I do with that? But budget limit it, you know?
[01:49:33:20 - 01:49:35:03]
Speaker 3
Yeah. It just flattens you.
[01:49:35:03 - 01:49:35:17]
Speaker 1
Right.
[01:49:38:14 - 01:49:41:01]
Speaker 1
But we can always start another side project.
[01:49:44:06 - 01:49:50:02]
Speaker 1
Charity, what an absolute pleasure to talk with you. Thank you so much for coming up here. I really appreciate it.
[01:49:50:02 - 01:49:53:04]
Speaker 3
Thanks for having me. Very nice to see you. Bye.
[01:49:53:04 - 01:49:55:19]