TRASCRIZIONE 🔗
For the last decade, I've taught people how to implement really great quality user experiences on
the web primarily. Mostly it's been JavaScript, TypeScript, and React.
Lots of testing, full stack with TypeScript. but implementation doesn't really matter as much
anymore. I mean, like having those skills, definitely where the agents are currently, you do need
to be a skilled software engineer to build a really great software experience. But who knows how
long that's going to continue to be super valuable. I don't, I certainly don't. It could be for
many years into the future, or it could just like... A year ago, I could never have predicted where
we are now. I can't predict where we're going to be in a year. And so in thinking about what I'm
going to teach next in answer to your question, I have decided that,
like, let's just play the future out. And let's say, like, if agents do end up taking over our
entire industry and make us irrelevant, then, like, I don't know how to plan for that. You can't.
You can't plan for that. And so what if I take one step back from that? what is the last skill that
makes us as humans still be able to provide value? And I decided that it is the ability to know
what to build. Like, what do you tell the agent to build? And that effectively means understanding
problems really, really well. And maybe not necessarily being able to...
articulate the solution, the specific solution super well, I think, like I said, right now that is
important, but having really, really good understanding of the problems and the use cases and your
users, that matters a lot. And in fact, even if I'm wrong and agents are never able to replace us,
this is the skill that we needed 50 years ago and is what I think is the last skill you need to
learn. It's actually a pretty well understood discipline. It's product engineering is the term that
most people use for this. And so I created product or epic product engineer.
That's my new thing. In fact, I'm wearing the logo right here. And yeah,
I'm going to just teach people how to develop product sense,
how to understand and empathize with their users, how to tighten their feedback loop.
Yeah, and develop clarity on the problem, how to talk to stakeholders, all of that. That sounds
like a very, very important mission. I think even more critical for our audience to understand
right now here is that... We believe this is the next future for software engineers, right?
As we come into the world of agentic software development and autonomous development specifically.
So without getting too deep into the details, can you give us a gist of what exactly makes a good
product engineer and where should people start acquiring those skills and how should they go about
it? I think a common mistake... that people make when they're thinking about product engineers and
user empathy and all of that. So they're like, okay, a really good product engineer is one who
listens to their users and does whatever the users say,
Load full transcript
you know, and just like solve the user's
problem all the time. But if you were to do that, like basically turn your user feedback thing into
a prompt for the agent, like, first of all, that's not very valuable. I could automate that process
if that's what this was. The problem with that is users are often going to describe solutions
rather than taking a step back and looking at the product surface as a whole.
So they're going to say, well, I don't really like that I have to click on this button over here
and do these three steps right here. Can you add an additional field so I don't have to do this or
whatever? And if you just do that, you might miss the fact that, well, they shouldn't be doing that
anyway. Instead, we have this other workflow that would be a lot more optimal for what they're
trying to accomplish. And so it's not necessarily about just understanding and doing whatever your
users want. It's truly understanding the job to be done.
Like, what is the user trying to accomplish? And how can you simplify your solution?
to accomplish that better and like really kind of disrupt their experience in solving the problems
that they're trying to solve every day. So it's complicated and it takes some serious domain
knowledge. You have to really love your problem space.
And I think one mistake that a lot of engineers make too is they love their solution. and they just
get so into the weeds of the solution that they miss that that solution shouldn't exist.
And instead, we should kind of go back up what I call the problem tree and go a different direction
that eliminates all the problems that we're having over here. Like no user is going to complain, or
if you continue to iterate on this solution over here and the users are always complaining, you're
iterating and making it better. But then... We find out that like, oh,
if I take in this other direction, we never needed to iterate on that because this other direction
is better anyway. So that's what a product engineer does is they see the product offering as a
whole rather than individual features and iterating on those to make those individual features
better. I used to be a product manager in one of my previous roles and I used to be an engineer as
well. Everything that you just described for me sounds... a lot like the product management role.
Can you draw the distinguishing line a little bit clearer for us? Yeah, I think that's actually
really a good thing to call out. So the way that I see is that with the technology that we have
now, product managers are actually able to do a lot more technical things than they used to. And
engineers don't need to do those things as much and they need to move closer to what product
managers are. And so at some companies, there's no distinction.
Like at a really small company, Sometimes it's just you and your co-founder. And like,
so you're all wearing all the hats, right? But then as the company grows, you start to realize,
okay, so there's some tasks that a
less technical person can do. I can focus on the technical side.
And so I'll be the engineer side of things. I think that agents make it so that you can continue
and develop the solution. longer before you feel that friction that requires another person to do
the management side or even the engineering side. I do think that there is a distinction,
though. And it's possible that in the future, it won't be as that distinction will kind of slowly
get erased. But for right now, at least, engineers and the technical expertise is necessary.
to make the right kinds of decisions around like technical decisions around how uh you architect
the final solution uh it could be that the agent gets good enough that it just asks you a couple
questions and then it's able to say okay well we're gonna obviously we're gonna do this that and
the other We'll deploy it on this infra. We'll use this database. Or like we're in the middle of
another, an existing project. Okay, we're going to do this migration. We're going to, you know,
widen in the narrow, like all the technical decisions that an experienced software engineer would
do. That could be in the future. But for right now, I think there is still a distinction between a
product engineer and a product manager.
where the product manager pulls in feedback from the outside world into like what the product
solution could be. And the product engineer pulls the trade-offs and the capabilities of the
technology into that same space. And then you can communicate in that way. But it requires,
I think on both sides, it requires the product manager to understand the technology and the product
engineer to understand the product, like the problem. So yeah,
that's kind of a long way to answer your question. I do think that currently there's still a
distinction and I think that that will continue, but it's possible that these things could merge
and you could have like a product manager or a product engineer. or like a new product person i
guess is the new thing uh one who is i'm more of a product or a technical kind of product person or
i'm more of a manager product person but like we're both kind of doing similar things i think that
could be something in the future but uh yeah right now there is a distinction personally as as an
engineer by training uh and i spent most of my career in engineering i found myself often in that
trajectory that you just mentioned, which is the trajectory of being so in love with my solution.
And when I got to become a product manager at one of my previous startups, I really needed to
unlearn that type of thinking. Can you give us some advice? How can you best, I would say,
catch yourself almost in that thought? So, yeah, that definitely requires being kind of intentional
and mindful of yourself. But one thing that you can do that's relatively easy,
is watch your users use your software because you might think that like oh this is so obvious or
this is so easy or like what a what a beautiful solution
or especially for the really technical
among us um we get so excited about how we solved this problem under the hood that never actually
surfaces you know in the user experience and so if you're just watching the users use your software
and maybe this is like just after you shipped that optimization that you're just so excited about
and then you see that the user doesn't notice or care or doesn't it doesn't actually impact their
experience yeah it will hurt yeah it'll hurt a bit and and then they start complaining about other
things or you see them maybe they won't complain but you'll see them struggle with other things and
you're like that's the easiest thing like That doesn't even require a lot of work or anything,
but I can make a massive impact on their experience by just doing this one tiny little thing.
Sometimes it'll be like a really significant thing. Like I said, you go back up the problem tree
and you take a different direction and it is a completely different workflow, possibly. You got to
be careful on that too. But watching your users is probably the quickest way.
Another really like a goldmine opportunity, depending on what software you're building.
is support. Go like have lunch with support and ask them about what are people constantly asking
you about? And what are tools that I could build for you to better serve our customers? But also,
what are things that I can build in the products that will make it so customers don't need to reach
out to you ever? What are the like the just most common things? Because if somebody is reaching out
to support, that means two things. It means they're really your customer. Because if they care
enough to reach out to support, then they like they care and they're they're really interested.
Sometimes you get people are just complainers, but like most most of those people are going to be
like, no, I really I want this product to work for me. And then secondly, those are just really
like those are the pain points that they felt like were strong enough to reach out to support.
And if you can eliminate those problems, you free of time for support. users don't like reaching
out to support it's annoying it's very irritating and um and so you can uh alleviate some of their
uh annoyances with your product as well uh so yeah watch your users and have lunch with support
that sounds yeah sounds like a really really good piece of advice i'm just iterating in my head
back all the times i i run a little um educational startup uh as a non-profit with a good friend
of mine and he's really the user empathy person that really drills down on the experience and makes
sure okay this is well suited solving the problem for that person and ruthless on cutting down
engineering our code that i've written and in order to say like this is great but we don't need it
um and i'm the person who always thinks about new features that we could build out or things that
could be amazing and more technical and then uh sometimes i i have to realize that it's not needed
so yeah yeah and it can be hard too because you're like i worked really hard on that And it's
exciting for you, but it really comes down to can you solve a real problem for a real person?
And I think that this transition right now is hard for a lot of people who are in love with that
solution camp. If you've just spent your entire career so far turning a JIRA ticket into an
implementation, then that's super replaceable. And it's kind of sad, especially if you really
enjoyed that, like that was your craft and you took really good care of it. And again,
right now, the agents do need to be guided, like pretty, like depending on a variety of factors.
CodeRabbit helps a lot with that process too, but it is becoming apparent to me that the actual
implementation matters a lot less. It's the... actually solving the problem,
that's the thing that matters most. I agree. Let's talk a little bit about that transition that you
just mentioned. What should people do they're prepared of? I'm thinking, for example, my previous
career was in semiconductors. I had people working for 40 years in that field,
had their PhD in knowing very much about embedded systems, understanding radar technology on a chip
level. If all of that goes away, what do you tell a person that maybe is now in their mid-40s?
still has 10, 15 years of career ahead of themselves. How do they prepare for impact? Yeah.
So first of all, let's get out of the way that nobody can predict the future. And then secondly,
let's say that somehow agents do replace technical people completely.
We can't prepare for that. Like there's just no.
If it gets that good, then robots will be built that are able to do plumbing and stuff too.
So if it really gets that good, you just cannot prepare for that. And you just hope that the people
in power distribute that wealth to humanity.
That's just the hope. It's Star Trek and not the apocalypse, right? With all of that context,
let's back up from that and say that's not going to happen. So how do you prepare for the
disruption that will happen? I think that the thing that's the most human is understanding problems
and directing the AI to solve those problems.
And so if you're in your 40s, you're like really smart with semiconductors or something.
I don't know that world at all, so I can't give you a specific example of what you could do. You
probably like that's this is what humans get paid to do is to solve problems. So focus on
developing your ability to get clarity on the problem, to really understand the problem and how the
skills that you have, which agents are now adopting, I guess, how those skills are applicable to
that problem space. what things are missing in that problem space and how do you fill in those
gaps. When an implementation or a solution for a problem is very cheap and very fast,
the bottleneck becomes identifying the right problem to solve. Well said. I think that is a
beautiful note to finish our conversation here. Thank you so much, Ken, for taking the time. Thank
you, Henrik. Appreciate it.