Episode 36 | F*ck Around and Find Out: The Recession & Engineering Leadership in 2023

Jason: So there's only two ways to look at the future when it comes to software developers. Either we'll have more or we'll have less than we have today. And when I say more or less, I mean dramatically more or dramatically less. And my view is that we will actually have dramatically more software developers because everyone will become a software developer at that point.

And it is not in the same definition by which we have software development today. And that's, I think, an important distinction. But I think what will happen is that people will be creating more and more things. 

VO: Welcome to Developing Leadership, the podcast for engineering leaders where Eiso Kant and Jason Warner share their lessons on the ins and outs of managing software teams.

Happy 2023, or is it? On today's episode, Jason and Eiso talk about the economic recession and share their predictions for the year ahead. They chatted about what happened in 2022 that set the stage for 2023, what the hiring halt means for software engineers, what will happen to those that didn't prepare, and more? 

As always, this episode comes with accompanying show notes with a deep dive into the main topics, mental models, and key moments from this episode. Find them at developingleadership.co and linked in the episode description 

Eiso: Hi everyone. Welcome back again to another episode of Developing Leadership.

Today you have just Jason and I together going on one of our good old rants. And Jason, first of all, how are you?

Jason: I'm good, I'm good. 

Eiso: So when we talked about this episode, it was the episode that, you know, we wanted to go out at the beginning of the year and think it'll be titled somewhere along the lines of like the recession and 2023 predictions.

So we're gonna probably have to try really hard to keep that somewhat on topic related to engineering leader. But it seems that people sometimes enjoy our rants. So why don't you kick it off. Feel free to start anywhere. Chat G P T. Recession Predictions. You go. 

Jason: Well, I think 2023 is gonna be an interesting year.

An interesting in all of the various ways in which you wanna take that word. Like, you know, we, we can look at the stock market, we can look at startup funding, we can look at potential layoffs, we can look at new advancements on the horizon. We can talk about the fact that everyone has to mute basically everything on Twitter at this point to stay sane.

 2023 is gonna be a kind of a interesting year. And as they say, you know, it's only gonna get weirder from here. And that's, that's what 2023 feels like, doesn't it? 

Eiso: It does. So let's, let's make some predictions, right? Let's make some that people are gonna probably write us about on Twitter in six months and say, you guys were really dumb, you were wrong.

Or they'll probably do that anyways. But I would say if we start with the recession, if I look at the last, kind of, since September till now and all the conversations I've been having with engineering orgs, there's a couple of interesting things that have happened in 2022 that I think set the stage for 2020.

In the very beginning of 2022, we had that like nice little freak out moment. Stock markets crashing, recession is coming, and there were a small number of companies who in my opinion had the balls and rightfully so to say, "we're going to act now and we're gonna act big". Like cut burn, cut layoffs, get back to like normal. It's gonna be tough times to coming years. And then we had everybody else, which feels like 95% of like tech companies, I see at least, they just ignored that moment. 

Jason: Or wait and see what happens. 

Eiso: Exactly. Wait and see, which by now people like wait and see after a hundred plus years of recorded economic history and recessions and bubbles, like it doesn't make sense, right?

We've all know 2008 where we all have at least some sense of the 2000 crash, like, wait and see is never the right approach. 

Jason: Wait and see is more appropriately the term "fuck around and find out." And that's what we're in at the moment. 

By the way, anytime you're actually doing this, let's keep this on track for leadership. Wait and see. That is definitely what your view should think of it as. Like, okay, we're about to fuck around. How are we gonna find out about this? You know? 

Eiso: Exactly. Exactly. And so what was interesting for me in Q3 is it was the first time where I started seeing engineering org saying, we're not hiring anymore.

Yeah. and Q4 is the first quarter where I'm seeing engineering org saying, we're firing and we're reorging. And that was a very stark difference. And I guess I have this weird visibility because I see dozens of, of engineering orgs, different ones every week. Q4 is the first time where we're starting to see like we're now really like letting go of people, and it feels like it's the beginning.

But if you ask my prediction, I think by the end of Q1 coming into maybe mid Q2, we've all gone through that. We'll, we will have actually made those decisions. We've moved through it, we've made the tough calls. What do you think? 

Jason: So it's tough, cause I think I've already been, I've already been wrong on this one, which is I had thought that all of the, let's, let's back up for a second. I thought that we would see all of the pipe and ARR heartburn shake out of the system by the end of Q4. 

You know, Red Point and what we talked about internally, we, we've been preparing for this since last year when valuations were crazy. We had taken a stance internally. It's not sustainable. We understand what's about to happen, all that sort of stuff.

So we have already been in this mode where we've been talking to our companies and et cetera, et cetera, and a lot of them have already made corrections a long time ago, as you mentioned. And internally, we were saying, "okay, so here's what we think might happen" end to Q3, going into Q4, all of the pipe, and ARR stuff is gonna shake out of the system, you know, we'll, we'll know better where we stand in Q1.

That doesn't seem to be the case. It seems like we're looking into 2023 where all of this ARR and pipe is shaking out in the middle of, you know, late into the Q1, into Q2 of, of 2023. 

So I think I've already been wrong cause I was had hoping to be working its way outta the system right now. It looks like it's gonna be in Q1 and Q2. 

What do I think is actually gonna happen is I think that there's is, we're basically gonna start entering the long tail of people who have not been prepared for this, and they're gonna have to react real suddenly. And that will take place in Q1 into Q2. I hope by the, the summer of next year everyone knows what the bottom of this looks like for their companies and for their, their particular instances. That said, it never shakes out that quickly. You have to hit it and then you have to like kind of get through it and you have to stabilize and all that sort of stuff.

So that's why everyone's kind of been saying, "Hey, with the coming recession, 2023, what's this gonna look like?" And if again, you go back to a hundred years in historical data, what does that mean for us? Well, it means that we're gonna be talking about 2024. Before you know it as when things, you know, might start turning or maybe tail into 2024. And for us, you know, we're in 2023 recording this right now. You're probably listening to this or 2022, recording, in 2023 listening to this 2024 feels like it's really far away. Yeah. We gotta get, we gotta get through that.

If you, you have burn that doesn't take you to 2024. Start figuring it out. 

Eiso: So when we came in the beginning of this year, I called the following, I said, it's 18 months from what I think it was February, like the freak out.

And that's again, I think, uh, is it Warren Buffet who said something along the lines of like, "a minute spent with an economist is already a minute wasted" . And so I think any, any of these predictions are shaky at best. 

Jason: Yeah. And the point of a prediction is not to be right in my opinion. The point is to understand directionally what's about to happen. 

Eiso: Exactly, to me, I took the 18 month figure because if we look at most even the 2008 crisis, it was about 18 months till we hit rock bottom, which feels really long. And I also know why, like, it's funny you said like the end of last year, because I remember we were raising our round the end of last year and we were telling investors "we don't want this high valuation because it doesn't make sense when the markets crash."

And I remember the amount of times hearing, ah, "but it's a new paradigm now. It's not gonna happen like that." Which of course is just like anybody, by the way, ever tells you something is a new paradigm, runaway. Like runaway. 

Jason: It's, it's rather, I don't wanna say amusing cause it's not amusing, but it's actually kind of stark to me to also how many people want to shake off and kind of dismiss this type of thing. And so, you know, I see it, I see it sometimes from founders particularly, obviously younger founders, maybe first time founders, which is, they're not aware about. What's actually happening in the market overall. They might understand that like, "oh, it's harder to raise around right now," or something like that. But they don't understand the mechanics of it fully. 

But even so, other people are still like, ah, like hand wave it away. It's, it's all gonna be fine. It's all gonna dismiss it. Yes, in the fullness of time everything is always fine, but the point is like you you have to get through these periods to get to the other side and yes, revenue multiples will increase on the stock market. Companies will trade better at some point in the near future. Yes. We're gonna have another bubble. Yes, we're gonna. Yes. You gotta get through this. Companies out there have to survive this, people have to stay employed or find new jobs. There is a moment in time that's probably gonna be for the next 18 months, right now, maybe going into two years where it's gonna get really, really, really rocky.

Eiso: I fully agree with you and to me what's been very interesting is that we're seeing for the first time something that we haven't expected for a very long time, and I think we haven't, I don't think we felt this yet. And I think, we'll, and if you ask me for a bold prediction, we're gonna start feeling this towards the end of 2023.

We've spent the last decade talking about a shortage of engineers, but I have a really hard time today seeing significant number of software engineering organizations that are hiring. And so common sense to me plays out. Which is, if we look at that, we're gonna have a weird moment in time where for the first time in a decade, we're gonna have great engineers that are not finding themselves the next place to be at the salary that they were before.

And I know this is a pretty bold prediction, and it's one that I hope doesn't play out that way just because of my own personal love for engineers and engineering leaders, but, I see the flip side of who's hiring and it's a very, very small number of companies that I expect are gonna be doing so throughout next year.

And again, like you said, it all plays out over time. Perfectly fine. But it's that weird moment in between where all of a sudden you're gonna have a hackathon event where like there's gonna be engineers who are great, that should be employed, that aren't. 

Jason: What we'll shake out this way I think is like, again, it's historical. Think about other places we've seen these things and all over time. 

So, we had a 10 year period where let's use the word power here, and I don't like, I understand, blah, blah, blah, et cetera, et cetera. But let's say the power was more in the employees hands, like the, the engineer's hands, they could dictate terms, they could dictate salaries, they could companies bent over backwards to try to attract and retain and all that sort of stuff.

It's flipped or maybe flipping, but I think it's actually flipped in the moment here to employers now have a lot more say here. They're gonna say, "No, we don't, we want the person to maybe be physically located in this place. No, that salary is too high for us," et cetera, et cetera. 

And here's where it gets nuanced is at the extreme end of a spectrum, maybe for the best of the best, they will always, the best of the best always get to dictate their own terms. It's, that's the nature of the world, but it's the average and slightly above average, or the deviation from the means, they know that that person or a group of people will no longer get to dictate them, and they will be beholden to what has become the norm or expected for the employer.

You know, five years ago, everyone could be remote or everyone could dictate 50% higher than normal salary for being an average employee who is essentially a seat filler or someone who could like increment a product, which 20 years ago was only allowed for the top three, four, maybe 5% of all people on the planet in terms of what their skillset was.

Eiso: I think that's a great way of looking at it. So what does that mean for engineering leaders, right? If we, if we try to pull ourselves a little bit back, is that all of a sudden all that time that was spent on hiring, which was if, let's be honest, a huge part of a lot of engineering leaders calendar is now being told to shift to we need to be more efficient. Right? 

And I think we had Rands who said, you know, no one, no one gets excited about being more efficient. The other thing that I'm seeing, which I think is in line with that, and I've, I've seen it probably about a dozen plus companies in the last three, four weeks, don't all say the same thing. We're focusing on our core products. 

Yeah. 

We're cutting everything. We're back to like, we're focused now. It's like, but what were you doing before? Like, no, you were unfocused before you were doing 50 things. Like now we're focused. This is, this is shaking out for engineering leaders to be very different responsibilities that they're gonna hold.

Jason: I mean, this, this looks as, there's a little thing, there's like wartime CEOs and peacetime CEOs and stuff, and like, obviously anything as cliched as that doesn't actually mean anything in reality. And you have to understand exactly what they're actually saying here. But in, in the scenario here, we're entering, there's, there's going to be a bunch of things that I think shake out.

One is your primary focus was hiring, but it was expansion. It was hiring more people to do more things, et cetera. Now, I still think hiring will happen as you should be recruiting, but what you're actually trying to do is raise your average. And by that is when someone leaves or you move somebody out, you're cycling in somebody who has higher skillset or larger talent or blah, blah, blah, et cetera, et cetera, et cetera.

So like if you're, if you've got a five out of 10 on the team and you're going to be replacing them, you're replacing them with an eight outta 10. And that's what your goal is. Your goal is to continue to do that. 

And in the best of times, people say they were doing that, but they would just add the 8, instead of replacing the five, because there's that old saying, which is like, "well, it's better to have a five in seat than have zero person", blah, blah, blah. No, that doesn't work now, CFOs are gonna call Engineering Managers and Engineering Leaders and say, no, you don't get that anymore. 

The other is cost savings will come in vogue again. So, you know, "Hey, what's your AWS bill? What's your Azure bill? What's your GCP bill? What's your, what's your SAS spend?"

All that sort of stuff. Like internal tools like, oh, is uh, do you actually need five different products along this path, or could we just pick the, it's okay, but kind of sucks solution that's bundled together with something else. So you're gonna start seeing some more tool consolidation on that front. And I, you know, I think as you mentioned, all these kind of special projects internally that a lot of companies have, they're gonna go, you know, they're gonna have to, they're gonna focus on the core product and basically it's gonna be like, does this make us money.

Eiso: Which to be honest, is also I think, uh, well an extreme to your side, a healthy regression to the fact of like, we are here all to build businesses and businesses fundamentally like your local restaurant and your neighborhood knows, need to make money, need to be able to grow if they're in tech, or at least need to be able to sustain themselves. And I think there's a lot of stuff that we've seen that was completely unsustainable, that's gladly shaking out. 

Jason: So this is a weird one for me because I do think that for the most part, tech has been insulated from the real world for a long time, and they haven't actually, we haven't actually had to understand that. So I'm, I'm old enough to be of a generation where you actually had still had to fight to survive. And I don't mean it just individually, which is true, but the company had to fight to survive. But there's a period of of time probably going on like maybe 13, 14 years, where effectively it was like there's, there's no mechanism by, by which you were going to ultimately fail It was like let the good times roll type of deal. And what that led to was this sort of inevitability attitude. And I'm not saying it's bad, I'm not saying it's good, I'm just saying what it, the knock on effect of it is, it breeds a certain type of, behavior set. 

And if that's the case, you don't really know how to deal with adversity in the moment. We know how to deal with a set of adversity, whether it's like, "hey, maybe the product's not resonating, or maybe we gotta expand faster, or maybe we gotta sell better." All that sort of stuff. But it's different when it's existential as in like, you know, "Hey, a year from now we're gonna run outta runway and we're not gonna be able to raise money."

If you look at the totality of companies funded between 2017 and 2021, the vast majority of them will not be able to raise another round. And I mean that like if you think about what your, your revenue is, if you think about what your expansion is, you think about how your growth rate like in the market that exists today, right now, you're not gonna be able to raise another round, or the company will have to take dramatic steps to raise another round. We're talking dramatic down rounds. We're talking liquidation preferences. We're talking things that startup employees have not had to think about for a decade at this point. 

Eiso: And they're the kind of things that end up in many times killing startups that, yeah, and I think this is the, and I think if I was anywhere today where the valuation of the last round was a crazy excessive multiple on the, on the revenue, I would be very doubtful of the, if this is where I want to be for the next five years, 10 years.

Jason: It's dramatically different for a founder versus an employee of a company. Dramatically different for a founder versus an employee of a company. And I say this talking to a founder of a company, and I understand that, but if there is a company out there that has, I'm just gonna pick out numbers out of the hat, but let's say you have a multiple billion dollar valuation, let's just pick, a 2 billion valuation and your revenue run rate right now is maybe 15, going 40. You probably don't get to raise another round. 

Like that, Hey, that company's growing, that company's doing incredibly well, blah, blah, blah, all sorts of stuff. There's not a sane growth investor on the market who's going to fund that company at some point. That company has years to grow into that 2 billion valuation. Go look at public market comps right now, understand what's happening here. 

So employees are effectively, they have to work 3, 4, 5 years, year over year, over year, a hundred percent growth just to earn into that valuation at the moment, founders have a different story that's, they might have taken some money off the table. It's a different valuation structure. Maybe even if they sell at seven 50, they're gonna do well when you know, all that sort of stuff. But it looks very different. So it's, it's going to be a really tough market. So if I'm an employee, you know, one, you know, I, I have to evaluate that if I'm making a decision, you know, maybe they don't have that luxury, maybe they just need a job. That's going to be a situation too. 

Two is look at how capitalized those companies are. So the good news is, and those companies that raise those valuations that typically might have a massive balance sheet, they might have amazing amount of runway if they didn't over hire. 

You know, some of those companies dramatically overhired and that's going to lead to another set of things. And if they had not corrected for that overhire, they need to understand that, "okay, I thought I was gonna be able to raise another round. I dramatically overhired, I hired, you know, two extra, three x amount of people I have, I'm gonna have to go back to, oh shit, this is my last dollar in and I need to get to profitability. Or, you know, some sort of, uh, sane efficiency number that's going to make sure that I can get my next round." 

Eiso: I'm smiling because I'm going to use this to segment into something that you probably hope I would've never segmented into Twitter and all the layoffs. And Independent. Let, let's, let's completely take away all-

Jason: I gotta loosen up for this one-

Eiso: I was about to say, so let's, like, you know, this is being listened to in 2023. So let's, let's, you know, let's segment away from the how and everything that, like the shit show of the, the how. 

Yeah. 

How does Twitter play out in 2023? And do you think that the, the headcount the company is right now is significantly healthier than before?

Because I do, I, I don't believe in how they got there. 

Yeah. 

But I do think the company has a real chance in this market now compared to if it would've gone on the track that it, it stayed on as a public company. 

Jason: I am on the record that most of the large tech companies are overstaffed. I don't know the percentage that they're overstaffed, but they're overstaffed. There's various reasons for it. There's good reasons for it, there's bad reasons for it. It doesn't really matter. But they're overstaffed. And in particular, when you get to the megacorp size ones they're dramatically overstaffed. They're dramatically overstaffed. Again, like, you know, my feelings on this, and in fact, I have recently had a couple of different posts on Twitter about all of these various things, so we don't need to rehash any of those sorts of things.

I definitely think that it's a healthier number for them. Seven plus thousand people at Twitter didn't make any sense. Now when you hear some people saying that, "Hey, it only takes 20 people to run Twitter," like, yeah, you don't have, you don't have any idea what you're doing if you make, you literally have no context to make that.

Now does it take a couple hundred people to run Twitter? Yeah, it takes a couple hundred people to run Twitter. It takes probably up to, up to a thousand people to run the Twitter the service and some of the internal operations of the company. It probably takes another couple of hundred to do things well in terms of like enforcement and moderation and um, legal capacities and, you know, all these various things that kind of go on to do this, but you know, it, it doesn't take several thousand people. 

Now that said, Twitter is also benefiting in the moment from years of hardening their systems with so many people. It's different to build a company versus to operate a company. And right now also this is December 9th, 2022 when we're recording this, Twitter hasn't released literally anything since the acquisition, like it tried to and it rolled back. What it obviously is focusing on is stability, it's trying to keep it up with less people, it's trying to harden the systems, it's very obvious what's happening here. 

So what do I think is gonna happen? It'll, it'll find a healthier homeostasis. It'll work with this and it will, you know, it'll go down once or twice or we, we've already seen a bunch of the issues with-

Eiso: And it's non-critical, so.

Jason: It's not critical. It's a distraction mechanism. We're all fine with it, et cetera, et cetera. So it'll just play out and it'll go about its business and that sort of thing. You know, I don't expect any major disruptions in Twitter. I don't expect anything really major or revealing, you know, any new features to be released though either like they're in this kind of this weird Twitter is what it is in the moment.

Once it figures that out, maybe it'll break off and kind of do some more interesting things in the future. Maybe towards like the first half of next year, they'll like release a couple of features or two that are kind of interesting, you know, back half of the year and stuff. But it's, it's in this like weird stabilization phase at the moment and it'll continue to be in a weird stabilization phase for a while.

Eiso: Yeah, I can't argue with you on any of that , but I think it's, I think it's, you know, worth as a 2023 prediction. Because of what you said right there. There are so many crazy extreme opinions that, that are floating around on what happened there and it's, it's worth for a lot of, I think, engineering leaders listening to this to understand like what is our equivalent in the companies we're at to being overstaffed, because it's not just Twitter that is overstaffed, and it's also not just the mega corps right, there's a lot of places where if we take a, if we take an honest look, the default was let's add engineers, let's add engineers, and I think a lot of organizations can get to that homeostasis after those cuts. 

Jason: Here's where I, I think Twitter, unfortunately is gonna break people's brains. A lot of people are gonna draw the wrong lesson from Twitter and, and it is particularly a lot of people who want to draw the wrong lessons are going to draw the wrong lessons from Twitter. And so there's gonna be a set, I've already seen this happening. There's people saying like, "the only way you can be successful is in extreme micromanagement" like, well, no, that's not true. In fact, like it actually is counter the difference being like, you know, you have to understand how to actually be efficient at your jobs and all that sort of stuff. 

And you know, there's a bunch of people saying like, "Hey, we need to, you know, one quarter of the size of these companies. That's actually what is healthy." Like, well again, no, you have to understand what the context of these situations are. That said, like you asked the question, what do I think engineering managers should take away from this? Well, I think they should take a hard look at what is actually being done in their teams and how much is getting done and understand, hey, this, it's a different job, you know? 

 Let's take a step back and say a different way too. I think for like the last 10 years, we've also said, "Hey, all employees want to be enabled and empowered in such a way that you just need to give them this grand vision and they'll just get it done and they'll organize and all that sort of stuff." that's like been a common trope throughout Silicon Valley. 

And in reality that's not true. In reality, you have to be directive, you have to be directed. You have to say, "we need to get stuff done and here's what we need to get done." And on this podcast we talk a ton about focus and prioritization. And too often people will say, "somebody will do that on their own." Sometimes you have to tell people to go do something and you have to tell them when it needs to get done. And I think it's more obvious in an environment we're gonna be in, in 2023 and possibly into 2024. But that is a way in which I think engineering managers need to take a kind of a sober cold look at their operations and in particular VPs and Heads of which is how are they organized, how are they operating, how are they executing? What dates are they holding organizations to, what quality standards are they holding organizations to? All that sort of stuff. And I think this, this is probably where a lot of folks who are in first time management roles are gonna really struggle coming up here because it's, it's not the way in which they've been taught to operate.

Eiso: I think you say something that I, I know it sounds super obvious to us from all of our discussions, but, uh, the fact of this is what I expect to be done by this date, and this is what we expect from the organization to have as our standard of quality and time to resolve bugs. All of these kind of things is in most modern engineering organizations, not the standard today.

Jason: In fact, like I think you, like if you went back to 2020 and said that, people would be like, that's not how we operate. 

Eiso: Yeah. And and I think this is something where like, it feels so funny because we keep coming back to the things that we've already learned from, I almost hate myself for saying this, but what was probably on Harvard Business Review in the nineties on a blog post. Right. 

Yeah. 

And I think this is the part from like, from engineering that like, and you guys listen to this, know that I love engineering leaders. I love engineers, I'm the first person to wish we could live in a, in a fairy land where we could all just, you know, do everything that's interesting and don't have like, heart dates to work towards and not get stressed out. But like, the truth is, like, in this environment, we're going back to what should have always been far more normal, which is like, "we're here to deliver things and we need to actually commit to them and we need to have clear expectations." 

Jason: I think that there is this, again, like I'm more most familiar with Silicon Valley, obviously, and like the tech companies that operate in the Silicon Valley ecosystem and stuff. And I think that the zeitgeist in the Silicon Valley ecosystem for so long had been dominated by what some of the darling startups did. Like, you know, hey, what is Dropbox doing? What is GitHub doing? What is Strike doing? Et cetera, et cetera. And then Google. And Google, you know what Google did became this norm. 

And I equate it to, you know, what happened in the university system in the United States, which if you're not familiar, you're listening to it outside the United States, like these random kind of universities would attract liberal arts majors and charge 'em exorbitant amounts of money. But they would, they would have like lazy, they would have water parks, like thematic water parks on campus with lazy rivers and like all these, these student unions that were spectacular and, and like they spent all this money on these facilities to attract people. Same, same thing has happened in athletics for decades, but it was happening for the average student to attract them. 

And it created these environments, which everyone basically just, just like laissez, fared the entire thing. We are not here to learn. We're not here to study or et cetera. We're here to have a good time. Yes, you're, yeah, you're supposed to enjoy college, but the fact that like a, a university is gonna attract somebody by having a lazy river like this is this weird environment and happened at Google. Google is, I think, the first one that did this. And I think that unfortunately we're, have to work our way out of this too and break some people's brains to understand that this is weird. This is actually weird. 

And my last statement on this topic would also be, I think, I think people expect too much outta work. I think over the course of the last decade, people wanted work to be everything to them. They wanted it to be work. So fulfillment and meaning and purpose for your career, but also like their friends, your activities, their distractions, food, , their food, their health and wellness. Particularly in the United States, like work became like everything like who your employer you picked became kind of like everything to you. And you know, let's just say 50 years ago, that would never have been the case 20 years ago. That would've been exceedingly odd too. 10 years ago started to become kind of the norm and Silicon Valley and I think people expected it. And I find that to be really weird. And again, I'm, I'm a little bit older, but I find it to be actually really bad for an individual to do this, they think it's great for them. I can go to work and get my breakfast and stay, get my lunch, and get my dinner, and my friends are there and I'm getting my massage and I'm seeing my therapist and blah, blah, blah. No. 

And it's so bad too because I think it causes a younger generation of people not to actually grow up and I'm being very pointed about this, it causes a generation of people to not have to grow up and understand how to do those things. I talk a lot about not hiring from big tech for like high-end executive roles because they don't know how to, to fend for the themselves in the wild and grow in a startup. Well, in reality, you're supposed to have to figure out how to go find your own therapist. You're supposed to figure out how to find your own physical therapist if you got something broken or to get in shape or what meals to eat and all that sort of stuff. And to have that, all that stuff taken away from you, your agency, taken away from you to go figure that stuff out. It's stunting your own growth. 

Eiso: I couldn't agree more with you and, and I'm glad we got here and, and I think there's something worth noting, right? The reason Google could do this, hell, the reason GitHub could do this is the fact that there was some cash cow, right? There's something that worked so well that despite of how the company would operate. 

Yeah. 

Was going to make it successful, right? It, the ads business for Facebook or Google, the massive virality of, of the growth and becoming, you know, the daily usage tool. GitHub, even if there was a period where the organization couldn't ship, like the product itself was just so, like the, the original core of it was just so strong that it can pretty much subsidize everything else.

Jason: There, there's a term for this, a Peter Levine who was on the board of GitHub, but he used for a lot of the Silicon Valley terms, and he called it cockroach companies. He's like, no matter what, you can't destroy that cash cow. Therefore, it allows you to do all sorts of different things. 

Eiso: And I think the causation correlation mix up that happened in tech is, "oh, doing all those things leads you to build a Google." No. It's when you get lucky to have found a golden gem like a Google. You can do all those things and then I'm, and then even then I understand it like if you have that cash cow and you want to create the best environment and attract the best people, and what I personally like, I find it a dystopian novel as well, but it's not the other way around.

Jason: So this is incre again, like we talked about nuance on this too. One of those those not lead to the other, right? So this is where we have to get really, really pointed and it's luck, it's smart, it's timing, it's, there's a lot of things involved. If you happen to capture lightning in a bottle and get that, there's, what you actually need to think about is maximize.

And those, those things that we're talking about might help you maximize, but they do not cause they do not cause that to be the case. So this is where everyone in Silicon Valley saying like, well, shit, Google did that, therefore it's gonna No, no, Google, Google. The first 10 engineers at Google and what they did in the silicone that they built and all that sort of stuff, that was not because of the benefits in the freaking, in the, in the freaking commissary. That was, that was, it was different. 

So, you have to understand that. So if you're in a mode. Trying to figure this out. Like, yeah, we need to have a bar, we need to have this like a hundred thousand dollars bar because GitHub had a hundred thousand dollars bar and that's what made their no, no, that in fact that, you know, was counter, that that worked against their success.

That actually limited them because of the, the effect that was happening on, on their productivity and stuff. It was just, again, not understanding these things and kind of cargo culting is something I'm so against, because if you don't understand what's, why something was successful in the first place, you're gonna learn the wrong lessons. Going back to Twitter. 

Eiso: And that for me is, well, recessions are horrible and have, you know, have lots of negative effects, is the good thing of recession because anybody who still mixes up that causation and correlation is gonna be gone. Right. This is why the companies that we look at, you know, that we admire the UI paths of the world and, and even the Dropboxes and the Googles in the early days, they weren't built in those times with that mix up of causation and correlation. They were built in the times of, we had to really get tough. Right? And it's the quote and you know, forget about the gender. But like that you were saying earlier that really stuck with me which is: "hard times create strong men, strong men create good times, good times create weak men." right? 

Jason: This is one of those situations where like I'm, you know, we go back to sports and every single one of these, but let's just do it. Let's just get out of the way. Right? 

Let's just get it out the way. 

So imagine you're this profoundly talented athlete, 13 years old, just unbelievably head and shoulders above everyone else. And just on the development curve, you're, so, you're six foot whatever and you know, you're just dominating basketball, football, whatever it is. What are the chances that person's going to make it in the NBA or the NFL or whatever? Well, if they continue to grow physically, I bet they have a good chance. But they're not gonna, they're not learning lessons. They're not learning the lessons in, in all reality, because it comes so easily to them in that moment.

And this is where you see so many people fail in athletics. And every time you, you see these, these supremely talented physical specimens wash outta the NBA or NFL. But the NBA in particular, it's probably because they had easy times getting there. There was no challenges whatsoever. They didn't have to learn. They could just go on just pure athleticism. 

And we've seen it all the time. All these people that are quote unquote, the next Jordans of the world, and they just really never amount that much. And in reality, if, if they didn't have all of these people sitting in their corner at camp telling 'em they were God's gift to athletics and they actually had adversity getting to that point, it would be much, it likely would be a much very much better outcome. Much different story for them. You know, if you have to work at something, yeah, it's always gonna be better than if it comes to you naturally, easily. 

Eiso: I couldn't agree more, cause I know we can go down this path for a long time and let's shift 2023 predictions and I'll, I'll start with one. For those of you who know a little bit about my background, having listened to a lot of episodes. I started six years ago obsessed with the applications for machine learning for source code. And dedicated a lot of years of my life to it and uh, and Jason and I almost ended up working together at some point because of it as well. That's a bit our origin story of how we met that we'll, we'll have to fully share one day, but we are now in a place where there's been a couple of fundamental shifts.

One is connected experience of a developer like my editor, connects to the internet. I know this sounds like the very normal thing, these days, but like four or five years ago that was completely unimaginable to get past people. And now we have these very large language models that are starting to kind of show us the fact that we can be assisted. And assisted, I think is really the key word here in terms of, well, not having to Google as much stuff. Not having to spend as much time in IRC as we did in the old days or in Discord these days, like asking questions like we're all of a sudden not as much time that we need to spend writing documentation.

Like we're accelerating the curve at which we can do things as developers that we did not know how to do before. And that to me is kind of, if I look at 2023, we're on a path that we are going to just allow more people to become faster at being able to actually produce something that works. And we're going to allow people as engineers to become faster at being able to actually do their job. And I'm kind of curious, like since we're talking about 2023, where would you put the boundary at the end of the year, right? We're here December 12 months from now, what does the boundary look like to you, Jason? 

Jason: I think it's pretty obvious. You know, the last thing that I incubated at, at in Okta, the office of the CTO at GitHub with the amazing team that was there was copilot.

So we've been on this for a little while and I think it's quite obvious with things that you're gonna have assistance that help you create, debug and understand your code. And right now they're kind of specific, right? Like, "hey, like, help me create in these specific ways," they'll get more general and get smarter.

But you know, think about what Rails did with its rails generate type of approach. Like you can say like, "Hey, I just wanna scaffold this whole crud base thing" and everyone looked at that and he goes,"ah, wow, that's amazing." That famous David Heiner Hansen blog post with the, the actual video with the blog that they generated.

Well, imagine you're able to do that on the fly saying, I want that same thing, but I wanna react front end. I want Python backend, I wanna talk into a Mongo plus, uh, cash plus a blah blah, blah. And it kind of structures this for you. It's going to get smarter. It's gonna get bigger and better. It's gonna create a lot more opportunity.

2023, I think, like we're gonna see the- we're gonna either see that whole thing right there fully fleshed out, or it's gonna, someone's gonna actually create the very first X, Y, Z that does that for you. I also think that the understanding of code, which is incredibly important for this, will become a lot more mainstream. 

Right now I think I've seen it in two or three different places. Obviously you can, you can ask chat GPT questions about code snippets, I believe, replet has a code understanding block. And so you, we started to see that pop out too. And I think that's fascinating. I think that's, that's gonna be, that's actually probably gonna unlock more than the generation, cause I think generation is a simple step.

Um, but debugging, debugging is gonna blow my mind. Debugging is where I think a lot of this kind of fun stuff happens. Like, "oh shit, what the. Like, what is happening in this? I don't really-" and that kind of goes hand in hand with understanding, but it's like it's gonna get better at that for you. I think that will happen.

Oh, one thing I do wish, if anyone's listening to some wants to build a Startup, I think that one thing that's gonna get lost in this, and this gets lost in every single abstraction that happens and that's what AI is. It's another abstraction on top of co-generation is that our understanding of the previous generation goes away over time and it's lost, and there's a, like a set of old bearded voodoo wizards who understand that that space, and you know, if you think about the semblance of competing as being a layer of abstractionist, like I don't even fully understand stuff that happened 25 years ago because I started 20 years ago and et cetera, et cetera.

In this wave, what I think will happen is people's understanding of architecture's gonna go away because you can just generate a whole bunch of this stuff and like interactions with various systems. And I kind of hope that we can find a way to make it so that people still understand that and help understand like, "okay, so here's what's actually happening. This is interacting with this, but it's being lost because of these things over here." Like stuff that we all know. Even if web development has gotten really kind of laborious at the moment, we still kind of understand what's happening with shadow dams over here and how it interacts with databases over here that's gonna get lost. Stuff like that is gonna get lost on people's understandings. It's gonna look a lot more like magic. 

Eiso: I think your last phrase, I, I agree with almost everything you're saying. I think the last phrase is, is the important one, right? We're moving development more and more towards magic and talking, you mentioned front end development, talk about, you know, what's happening with React or any front end like library compared to 10, 15 years ago. There's a lot of developers out there right now to whom what's happening at that shadow DOM and below already feels like magic.

Jason: Even to me, right? Like I, I'm an old school web developer, but I'm an infrastructure developer, right?

So I look at what's happening on that side of the fence and I think like, is AI wave, is it gonna be magic for me because I'm gonna be able to, say, generate a front end for me, given in my, my preference over this, and I want to add this. Tell me where to go point to go do this. And because I have the fundamentals, I'm gonna be able to do that pretty quickly, but it's gonna speed up my entire access.

I, I can't tell you how many side projects have abandoned at some point because I'm like, "Ugh, I don't wanna learn react" because I don't like the way that react does something, you know? And I don't care to learn it. 

Eiso: Okay. We have that discussion on that one day. Cause I fully agree with you. It's, it's exactly that, right?

We're getting into a world where, and we've always done this, right, we build abstraction on top of abstraction. What I find very interesting about the abstractions that are happening with, with large language models and everything is that we're, we're bridging across the stack. Right. We've gone for like the last decade. 

Yeah. 

And have extracted so much on front end and so much on back and so many on different places that like you had to become an expert on one thing. Yeah. Like it, it's hard today, like, and you can do it in your personal project, but it's hard today to really fully go into any part of a code base of a company and, and be able to contribute there quickly if your expertise went very far in one dimension.

But we're kind of going back to now like the, you know, the 12 year olds and 14 year olds that are all on re by the way, these days to them this is being brought back again to like, Hey, here's magic. But it's magic across the whole stack. And, and that I find is actually very cool. I think it's about scary and it's also pretty exciting cuz we haven't had that in, uh, in the last 10 years.

Jason: Well, I think so another prediction. This is a generalized AI prediction. I'll try to find the version of this for 2023, but I have that in my will AI replace developers Twitter thread that you know, again, like you're re you're hearing this in 2023, this is earlier December, 2022. There's a lot of ceremony that's involved with development, as I say, in that thread. And I think that there's a lot of people who actually are paid to deal with the ceremony. And part of what's gonna happen with AI coming in and this abstraction layer is a lot of that ceremony goes away. So if you're looking, yeah, we're back to billing, so of like, that stuff will go away. If your job is to basically deal with ceremony, you gotta figure out what that means for you.

But I personally think this is an amazing, amazingly good development because frankly, this generation of software development has way too much of that. 

It's not fun anymore.

It's, it's not, and I think it's, it's led us to these weird leaky abstractions. We talk about that all the time. It's weird led to systems that are kind of brittle in some ways.

Also this, I do think this contributes to some of the bloat that we've seen inside of companies, cuz you have people who are dealing with very, very, very, very specific portions of the system.

Eiso: One of the things that, and, and again, it's probably this goes a little bit beyond 2023, but what all of this does and, and cements for me, is just even more centralization.

And I don't like it , but I like it. What I, I mean in the sense of, I like it as an end user. I don't like it when I look at the, the implications of what software development used to be, but if we really play this out right, like the, the system speaking to each other, making it even easier, the, it all means at the end of the day, The abstractions we're gonna see are not just at how we build, we write code and how to see each other. Is that the systems in itself that are gonna get most advantage from this are the systems that are running on cloud companies. 

Jason: There's, there's no doubt if anyone's doubting whether cloud's a real thing. Obviously you're not even doing a serious job these days. If, if you have the last enter your mind.

So where is all the value going to accrue in the next 20 years? Obviously it's going to be an interconnected. Whether that's, hey, in somebody's public cloud hyperscale or somebody's VPC private cloud, or these things that talk to each other, I, it's, it's the only way the world's gonna work. 

Eiso: Let me throw something at you like, and I'm very curious if you ask me, and, and I don't think this is even 10 years, any real understanding or needing to understand a database, I think we'll fall away for 99% developers and I, and I mean, I mean even to the point of like everything that we care about today from understanding an index to even how you structure your data, you know, the, the best principles of, of asset relational databases, like all of that I think goes away.

It's just, I put something here and I have a way to ask for it probably will still be sequel that probably will never go away , but, and like I have access to it. Everything else no one ever is gonna likely have to think about anymore. I mean in 99% of developers. 

Now, everything that's already for most of the applications that we're thinking about, right? Like everything that's authentication, building, like, I mean, really feature sets. Yeah. Already getting abstracted away. 

Jason: Yeah, I mean the fact that like, you shouldn't have to make a login page. Yeah. Like login pages just work these days, like 

Eiso: Right, exactly. Right. Like odd zero proof that like, we're getting to this point where for the, let's call it the average or above average developer, not the, the crazy small niche that's working on all the, the cool stuff at the cloud companies and in the tooling, we're getting to the point that, That all that is left is thinking about the actual ui, the logic, the what are we trying to do here in 95% of the cases?

Right? And, and this is the part to me that, like, I still don't think we've really realized the implications of, because nobody is willing to speak up and say that 95% of what we do is crude applications. Right. Or web or mobile. So if we play this out, Jason, and this is not a, this is definitely not a one year prediction, but, but a 10 year prediction or maybe 15 year prediction, how many software engineers in the world, will still need the background that a software engineer has today. 

Jason: So there's only two ways to look at the future when it comes to software developers. Either we'll have more or we'll have less than we have today. And when I say more or less, I mean dramatically more or dramatically less. And my view is that we will actually have dramatically more software developers because everyone will become a software developer at that point. And, and it's not in the same definition by which we have software development today. And that's, I think, an important distinction. But I think what will happen is that people will be creating more and more things. 

So like let's just even take today, like would you, like, I think part of this is a little bit of a, our own wanting to keep the mystique of software development, but would you call someone who's developing on Webflow or Wix or Squarespace or something like that, releasing something, talking to Shopify, talking to whatever, and running a business. Would you call them a software developer? I don't think you would today because you'd say like, "Hey, they're an entrepreneur building a business and they're using low-code, no-code tools," and there's this dismissiveness that happens with that with, "well, I'm a, I'm a Kubernetes."

We're the old guard. 

Yeah, yeah. We're the old guard. And I think what will happen is that dramatically more people will basically not give a shit about some definition of developer, and they're like, whatever. I'm creating a thing. That's all I care about. And that will become a thing. 

Now, I also think that they're going to always, always, always, always be a subset of people who need to maintain the systems and all that sort of stuff, blah, blah, blah. And there's always gonna be that. And I think unfortunately, there's going to be some sort of like elitist attitudes that emerge over that too. I'm like, well, I mean, I'm maintaining the AI systems. I'm a real developer. I'm a real developer and all that sort of stuff.

But I think the average person just like me as the average developer from the late nineties, early two thousands into the, you know, I'm a distributed systems engineer. There was a time when the older guard from my generation were like, yeah, but you're not writing Assembler, dude. You're not writing the, you know, that sort of, you're not like the kernel engineer and all that sort of stuff.

And now like if I were to allow myself to become a certain category of person too, I'd be, I'd have the same sort of attitude, about people who are doing all the React stuff and I'm not, because what I look at that is like, "Hey, that's kind of magic" what they do, because I don't understand it. You know, so you've gotta take that attitude of like, Hey, a, a single person created this company that's worth 500 million because they just co, they cobbled together all these things to like, and then AI helped 'em do that. Like, maybe I wanna rethink a little bit of my elitist attitude about what it means to be a software developer. 

Eiso: I couldn't agree more. I think that's a, that's a fantastic place to leave today at. I think we could probably go for hours and we've done more prediction episodes and we'll do more, but I think to sum it up the end of 2023, we're expecting the crazy reorgs to firings to to have passed at least one or two quarters before that.

There's a lot of companies that will have had to dramatically change or maybe even, you know, be on the cusp of no longer existing because they didn't get the fundamentals right. We will see a lot of developers in the world now having access to tools that help them in a system, in debugging, in reviewing, in understanding code, and even suggesting code.

And Jason and I, uh, probably when we're doing this episode at the end of 2023, we'll again, make predictions that we're probably half wrong about. 

Jason: Well if I could give anybody advice walking into this next 2023 year, if you're a founder listening to this is, it's for real, like take a look at your burn, take a look at your runway, take a look at those things, get efficient, that sort of thing.

If you're an engineering leader looking at this and you're trying to figure out what's coming for you, take our advice to heart about, "Hey, this is what we need to do. This is when we need to do it," be ruthless about that sort of focus and maniacal attention to, to value. And if you're an engineer listening to this, understand the companies that you're working for, understand the value, uh, the valuations that they have and the, the burn multiples and the revenue that they have gotta work into. And if you wanna harden your skills, just understand also what is attractive. You have to be. You have to be really good at your job, and you have to be efficient, and you have to be good at multiple jobs. 

VO: Thank you for listening to developing leadership. Make sure you subscribe to follow us on our journey to meaningful engineering leadership. If you have any challenges or topics you would like us to explore in an upcoming episode, tweet or the MS at dev leadership underscore on Twitter to learn more about data enabled engineering and how metrics can help your teams and improve your process. Go to athenian.com. See you in two weeks for another episode of Developing Leadership.