This video discusses essential skills for developers in the age of AI code. The speakers, Trisha G and Kent Beg, explain that while AI can handle code creation, developers still need strong decision-making, design, and planning abilities. They emphasize the importance of understanding the business, fostering curiosity, and using advanced testing techniques, as well as crucial 'soft skills' like communication and prioritization, which become even more critical with accelerated development cycles.
Key Takeaways
1
The traditional focus on memorizing syntax and API details for developers is becoming less important, as AI can generate code.
2
Developers now need to prioritize skills like design taste, knowing when to refactor, and understanding when to ask clarifying questions about business needs.
3
The ability to identify and address what slows down the development process, especially with AI's accelerated pace, is a critical skill.
4
Cultivating confidence in one's ability to learn new skills as technology changes is a fundamental skill for developers in the AI era.
5
Understanding and applying advanced testing techniques, like mutation testing and fuzz testing, becomes more valuable as AI can assist in generating and evaluating tests.
6
Formulating clear, high-level goals and prompts for AI tools helps maintain focus and can even lead to unexpected, valuable suggestions from the AI.
7
Curiosity and the willingness to explore new concepts and tools are highly leveraged skills, as AI can act as an infinitely patient tutor.
8
So-called 'soft skills,' such as diverse communication abilities, business analysis, saying 'no,' prioritizing, and setting expectations, are more important than ever for developers acting as a bridge between technology and users.
9
Developers need to embrace the fast feedback loop enabled by AI, exercising skills like decision-making, listening to feedback, and communicating more frequently.
10
Understanding one's own learning style and tailoring AI's teaching approach to individual needs is an overlooked but important aspect of working with AI.
No matching segments found
0:00
Welcome to the modern software
0:01
engineering channel. Today's one big
0:03
question is what skills do devs actually
0:06
need in the age of AI code. I
0:10
>> Oh, yeah. Yeah. Who are you?
0:11
>> Yeah, that's where I was going to go.
0:12
>> Oh, I'm sorry. I already interrupted.
0:15
>> Do you want to keep going or should we
0:17
start?
0:17
>> No, no, no. Go, go right ahead. This is
0:19
Trisha G, fabulous thinker about
0:23
software development, amazing presenter
0:25
and uh glad to be here with you today.
0:28
>> Thank you. This is Kent Beg. Amazing
0:30
person. Someone who I learned about
0:33
test-driven development from who I was
0:35
very privileged to meet in person last
0:37
year. Uh lots of opinions on software
0:40
development developers and and genies.
0:43
So I really wanted to ask you this
0:45
question about what skills do devs need
0:46
in the age of AI code because I see that
0:48
you're actually using um agents and AI
0:51
coding to to do code. As someone who's
0:54
who's been how do I put this politely?
0:56
uh you've got lots of experience in
0:59
writing code for a long time.
1:01
>> You can say all this fine
1:03
>> you're not old you're very experienced.
1:05
Um I was really interested in your
1:06
viewpoint on because you have adopted uh
1:09
AI development and I think people might
1:11
have expected you to be somewhat grumpy
1:13
about these new tools but in actual fact
1:15
I think you've kind of embraced them. So
1:16
I'm interested in in your viewpoint on
1:18
using AI.
1:19
>> Well one of the beauties of being old is
1:21
that you can be grumpy and enthusiastic
1:23
at the same time. It's a it's a form of
1:25
multi emotional multitasking and uh I've
1:28
gotten gotten good at it with lots of
1:30
practice. So people look at augmented
1:34
development developing with the genie
1:36
and they think well this is completely
1:38
different because my activities are very
1:40
different. I'm not typing all the time.
1:42
I'm reading way more. I'm doing more
1:45
kind of review more project management
1:47
kind of stuff more planning looking at
1:51
more feedback. and they think, well,
1:53
okay, like all of my skills must just be
1:56
gone. And uh my experience of it is that
2:00
that's not at all true. That that 90% of
2:02
what I used to pride myself, you know,
2:05
knowing all of the keyboard shortcuts
2:07
for the refactorings in all of my idees,
2:10
you know, that that was really important
2:12
when I'm doing hundreds of refactorings
2:15
in a day. Just doesn't matter if I don't
2:18
do refactorings by hand anymore. But
2:21
knowing to do a refactoring that becomes
2:24
really important. That sense of timing,
2:27
the sense of taste, the sense of
2:30
short-term versus long-term, the sense
2:32
of now it's time to introduce this
2:35
architectural principle. All those
2:37
become much more important because
2:39
you're making those kinds of decisions
2:41
so much more often than you used to. So
2:44
for that kind kind of code I do, those
2:46
are the skills that really have a lot of
2:48
leverage. This kind of brings me to
2:50
something I've been thinking for a long
2:51
time. Like I always felt like we're not
2:53
looking for the we've never been looking
2:55
for the right skills when we're trying
2:56
to hire developers anyway. we often used
2:58
to test them on syntax and APIs and you
3:01
know weird compiler tricks and things
3:04
like that when in actual fact the skills
3:06
that we've always used are the sorts of
3:08
things you're talking about like design
3:10
taste knowing when to slow down knowing
3:12
when to to refactor knowing for me it's
3:15
things like knowing when to ask
3:17
clarifying questions of the business and
3:20
knowing when to push back and say hey is
3:22
that what you really want and and I
3:24
think these are the sorts of skills that
3:25
really good developers have got that are
3:27
never really going to go away if you're
3:28
using a different way of creating code.
3:30
>> Yeah. So, the creation of the code, the
3:33
cost of that is dropped to to zero
3:37
except that the way the genie naturally
3:40
creates code creates a lot of friction
3:43
in the process. So, if you just kind of
3:46
let the genie do its thing, it's going
3:48
to go slower and slower and slower over
3:50
time. Now, we spent decades of our
3:53
lives. I did. Maybe it's just years for
3:56
you, but uh working with teams of humans
3:59
and they also got slower and slower and
4:01
slower over time. The process just
4:04
happens so much faster with a genie
4:06
involved. And so being aware of what's
4:09
slowing you down, what you should do
4:11
about it now, what you can do about it
4:14
in the future, that becomes a much more
4:17
important skill.
4:18
>> Yeah, I I think you're absolutely right.
4:20
Like I I can't think of a single team
4:22
I've ever joined where whether we were
4:25
near the beginning of a project or it
4:26
was a legacy project, there was still
4:28
bits of code which were gnarly and
4:31
difficult to change and areas where we
4:33
had tech debt, whatever that means, you
4:35
know, where it was slow to to add stuff
4:38
and the business never gave you time to
4:39
fix it and that's that's just always
4:42
been the way. I was interested in what
4:43
you were saying when I when I went to
4:44
your um tidy code workshop in go to
4:47
Copenhagen when you were talking about
4:48
this and how code is just naturally
4:51
going to get more difficult to change if
4:53
you don't pay attention to to the the
4:56
cleanliness of the code and how easy it
4:58
is to make changes to.
4:59
>> Yeah. I I like to talk in terms of
5:02
optionality instead of debt. I I think
5:04
there's a deep analogy with debt that
5:09
people understand if they are like
5:12
financial traders but usually people
5:15
think debt that's a bad why why did you
5:17
make that mistake in the first place? So
5:19
I I think the the positive spin on it is
5:23
we have options and we can do things
5:26
that increase our features or we can do
5:29
things that increase our options and
5:31
both of those create value. The features
5:33
are really visible, very legible, easy
5:36
to count, easy to point to. The options
5:38
are harder because it's kind of like,
5:40
well, the next time we add a feature
5:43
like this, it's going to be much easier
5:44
than it was the last time.
5:47
Maybe. And how much of that do you do?
5:50
And you imagine some big change you can
5:53
make that will give us all the options
5:55
in the world, but does it really and can
5:58
we afford to spend that money right now?
6:00
it. Those are all still difficult
6:02
questions to answer. And because the
6:05
whole timeline is compressed, they
6:09
become more valuable because we're
6:11
making those decisions again much more
6:13
frequently than we used to be. You know,
6:15
well, I I want to highlight another one
6:18
which is something I thought of in an
6:20
earlier conversation today. geeks. The
6:23
way you become a geek is you dig deep
6:26
into technical topics and you get this
6:29
confidence like yeah if it's things
6:32
change I have the confidence that I can
6:34
learn the next thing and a just a a
6:37
sense that I have of geeks around even
6:40
the even the people who are embracing
6:42
the genie are starting to lose
6:45
confidence that they'll be able to learn
6:48
the next set of skills and confidence is
6:51
a skill and you can cultivate it and you
6:54
can practice it and I think we all need
6:56
to get better at it. In fact, we're
6:59
building it's people building tools for
7:01
people. However,
7:04
we're one more level of abstraction
7:06
removed from what we're doing. And so
7:08
that confidence like yeah things are
7:11
going to change and things are going to
7:12
get different. That's okay. We're we'll
7:15
be able to figure it out. Just having
7:17
that confidence is a basic skill that
7:20
everybody needs to cultivate. Yeah, I
7:22
agree with that. I I have to say I think
7:23
my confidence has been somewhat hit
7:25
because I spent 10 years of my career
7:26
talking about IDE shortcuts, right? And
7:28
now I'm like are we even using ideides
7:30
anymore? Like what's the point of me?
7:33
But that's why I wanted to have this
7:34
conversation because I think the sorts
7:35
of skills that we can take confidence
7:37
from, especially if we've got experience
7:39
in the industry, are still applicable
7:42
things like thinking through what could
7:44
go wrong with this with this code or
7:46
what could go wrong with this use case
7:48
or which comes to testing which I want
7:49
to talk about in a minute or like you
7:51
were saying optionality being able to
7:52
recognize optionality or being able to
7:54
recognize that you're closing off
7:55
options and understand what implications
7:57
that might have. These are things that I
7:59
think we've we've been doing as
8:00
developers for a long time. I don't
8:02
recognize it as skills because you don't
8:03
put it on your CV or resume as I'm able
8:06
to recognize optionality and when not to
8:08
do certain things, right? You say I can
8:10
use Spring Boot and Java, which is, you
8:12
know, not not the same thing. So, I I
8:14
did want to talk about testing because
8:16
I'm a big fan of testing. You're a big
8:18
fan of testing. And I think that the
8:20
skills that we learn writing automated
8:22
tests, particularly if we TDD, are still
8:24
really valid in the age of AI, even if
8:27
we don't necessarily handw write all of
8:29
our unit tests the way we might have
8:31
done in the past. So like what do you
8:33
think about that?
8:34
>> Yeah. So, I I had a project I have a
8:37
series of live videos called the Genie
8:40
sessions where I'm just streaming live
8:42
coding with a genie and was looking at
8:45
one of the projects I did in that sess
8:47
in one of those sessions and I realized,
8:49
oh, there's a bunch of uncovered code
8:51
here. There were a bunch of corner cases
8:54
that aren't covered. I I felt pretty
8:56
good at the time it was going, but I
8:59
needed the awareness to go and review
9:02
the tests that were there. I can also
9:04
the nice thing about having the genie
9:06
around is you can always ask the genie.
9:08
So knowing about something like coverage
9:11
testing where I can say to the genie run
9:15
a coverage tester and propose to me
9:18
tests that will only pass if the
9:22
uncovered code works the way it's
9:24
supposed to.
9:26
You know, whoa, that's so vague to say
9:29
all of that. If you're used to, you
9:31
know, writing spec every character of
9:34
every test yourself, that seems like
9:36
it's culplessly vague. But in fact, that
9:38
does a pretty good job of filling in the
9:41
holes in a test suite. All those te all
9:44
those tricks that you know mutation
9:46
testing, fuzz testing, you know, sending
9:50
random inputs in when you get a crash,
9:53
you reduce the set of random inputs to
9:55
the minimum set that still reproduces
9:57
the crash. Oh, if you have those tools
10:00
in your box, they're more valuable than
10:03
they used to. They used to be
10:05
>> because you can say run a coverage
10:07
tester and tell, you know, propose the
10:11
tests. Now you can review the result to
10:13
that. You can say, well, what part of
10:15
the code does that cover? You know, why
10:18
do we miss this in the first place? Does
10:20
this really matter? Does this make
10:22
sense? Is it really redundant with the
10:25
other test we have or is this something
10:27
genuinely new? You can do all that with
10:29
the genie's help in minutes in a way
10:32
that would have taken hours before. And
10:35
so every trick you have in your box of
10:38
tricks, every tool you have in your
10:40
toolbox, there we go. That's better.
10:42
Every tool that you have in your
10:43
toolbox, you can pull it out and use it
10:45
multiple times a day now in a way that
10:48
maybe you would do the same thing once a
10:49
year before.
10:51
>> Yeah.
10:52
>> And I like um I like what you were
10:53
saying about uh one of the things I
10:55
quite like about working with the Genie
10:57
or with an LLM or chat system is that it
11:00
does force you to formulate your
11:02
thoughts. Right. Like you were saying in
11:04
the past, we might have looked at a
11:05
piece of uncovered code and gone, "Oh,
11:07
that's a piece of uncovered code. I
11:08
should write a test." But by forcing
11:10
yourself to think what are the
11:12
activities that need to happen in order
11:13
to make sure that gets covered and what
11:15
am I looking for and why am I looking
11:16
for it? You know, that's that helps you
11:19
I think that helps you write potentially
11:21
a more resilient system because instead
11:23
of just going I'm going to write the
11:25
dumbest unit test possible to make sure
11:26
the code has coverage.
11:28
>> Yeah. No assertions. Uh
11:30
>> yeah, exactly. And you're testing your
11:31
getters and setters or whatever, you're
11:33
thinking, what is this code? What is it
11:34
for? What do I want to prove or
11:36
disprove? and and you might even come to
11:38
the conclusion the code's not even
11:39
necessary. For example, the the
11:41
production code or whatever. So, I think
11:43
that for me, what I found really helpful
11:44
is just writing the prompt more often
11:47
than not helps me understand what it is
11:48
I'm trying to do. Whereas in the past, I
11:51
it helped when I was doing pair
11:52
programming because when I was
11:54
navigating, I'd write down the notes of
11:55
the things that I need to do next or the
11:57
steps that we need to take later. But
11:59
before I did pair programming, I would
12:01
quite often do what a lot of us do and
12:03
just sit down and write code. And then
12:05
at some point you get to a point where
12:06
you're like what was I doing and why am
12:08
I doing this and what was the point? One
12:10
of the reasons I like TDD is because it
12:11
forced me to do it in smaller steps. But
12:14
the working with the genie you're not
12:16
working in small steps is still useful.
12:18
It's one of the things I wrote down like
12:20
breaking stuff into small things is a
12:21
good thing. But having your eye on the
12:23
prize of what are we trying to do here?
12:25
I think in like typing that out to the
12:28
genie and say I'm trying to do this
12:30
thing helps us keep clear where we're
12:32
going and what we're trying to achieve.
12:34
>> Yeah. I I've found surprising
12:37
influence by by stating highlevel goals
12:42
even if they aren't really relevant to
12:45
the details like we would like people to
12:48
have confidence in the performance of
12:50
this data structure. Oh, you know, okay.
12:54
So, that's a highlevel goal. That's not
12:56
like a tool invocation. That's not even
12:59
a programming technique that I apply to
13:02
it, but helps me remember what I'm
13:06
doing. And it also influences the genie.
13:09
So, I think there's a that's maybe a new
13:12
skill in the augmented coding toolbox,
13:15
which is remembering to state highle
13:18
goals of what you're doing. every once
13:20
in a while it'll come up with uh okay so
13:22
since that's what we're trying to do let
13:24
me try and achieve this sub goal and
13:26
you're like oh I wouldn't have thought
13:28
of that and that's a that's a fantastic
13:30
moment I mean sometimes it suggests
13:32
things to do that are just infuriating
13:35
and then I'm ever more addicted to this
13:38
slot machine nature of this stupid tool
13:40
I'm using but uh
13:41
>> yeah and but I think that's where
13:43
sometimes um experience comes in right
13:45
because you can look at the result and
13:46
just go this like this is clearly
13:48
nonsense go away and do it again. But I
13:50
also like the the fact that it could
13:52
suggest things to you that you might not
13:53
have thought about. So for example, you
13:54
mentioned mutation testing earlier. Um
13:56
when I was giving a talk about testing a
13:58
couple of years ago, I went off and I
14:00
spent two days researching mutation
14:01
testing because I kind of vaguely
14:03
theoretically knew what it was, but I
14:04
didn't know how it worked, what the pros
14:06
and cons were, how to do it, what the
14:08
frameworks were, any of that stuff. But
14:10
with a if a if a genie kind of said back
14:12
to you some or generated some mutation
14:14
tests or talked to you about mutation
14:16
testing, you can ask it, hey, like what
14:18
is that? what is it for? What are the
14:19
pros and cons? Why do I need it? Or
14:21
maybe I don't need it. So, you can have
14:24
those kind of you can educate yourself
14:26
while you're kind of going along. And I
14:28
I quite like that because it's one of
14:29
the reasons I like the IDE because your
14:31
education is usually like contextually
14:33
based on the things that you care about
14:35
right now. And and the genie is like now
14:37
is a good time to learn about mutation
14:39
testing. Like, oh, good. Tell me what I
14:40
need to know about mutation testing and
14:42
write one for this thing and tell me
14:43
what's good about it. So, I I'll add to
14:46
the list curiosity, which is my
14:48
restatement of what what you you I think
14:51
you just said, which is cultivating that
14:54
sense of curiosity where you're like,
14:57
"What is mutation testing?" Well, you
14:59
have an infinitely patient tutor.
15:02
>> Yeah.
15:02
>> A little bit maybe
15:04
little uh tipsy, but yet there's a tutor
15:08
right there. You can ask them all kind
15:09
of questions. You can ask about the
15:12
programming language you're using. You
15:13
can ask about the tools internals. You
15:17
can ask about the code that you're
15:19
working on. You can ask about
15:20
alternatives. You can ask about the
15:23
techniques that it pulled out. And
15:26
cultivating that sense of curiosity,
15:28
learning to listen to your own ideas and
15:30
and give them appropriate weight is
15:33
that's a highly leveraged skill that
15:36
we're all just learning. So I did this
15:38
this project in a live coding session
15:41
and at the end of it so it was a it was
15:43
doing a a key value store on a on GPUs
15:47
and I'd never written any GPU code
15:49
before so I thought this will be fun and
15:51
it were I got something was working and
15:54
okay well that's interesting but how
15:56
does this GPU stuff work and I don't
15:59
know where the idea came from just
16:01
whatever random photons mashing in my
16:04
brain I said uh write me a fairy tale
16:08
that explains the operation of explains
16:11
the fetch operation and it came up with
16:14
this uh there's a queen and she lives in
16:16
a castle and the messenger comes and she
16:18
goes to the library and tells the pages
16:21
and there's the whole thing written up
16:24
as a fairy tale and I thought oh okay
16:27
now I'm not going to say write your
16:29
documentation in the form of fairy tales
16:31
or haikus or raps or Shakespearean
16:34
sonnetss or whatever But being aware
16:38
that you could gives you all kinds of
16:41
options for understanding your world.
16:43
And we're all in this this efficiency
16:47
mania where we just want to do the next
16:48
thing and as soon as you get a thing
16:50
done then you have to be starting the
16:51
next thing right away. and learning to
16:55
see the space between the features and
16:58
learning to use our new tools to nourish
17:01
ourselves in that process to learn more
17:04
things to communicate more clearly with
17:06
our our colleagues. That's a hugely
17:09
transformative skill because, you know,
17:12
I'm endlessly curious, but I also have
17:14
to get stuff done. So, I'm looking at
17:16
your bookshelf behind you right now, for
17:18
example, and I'm like, "Oh, I'd like to
17:20
read that one. That's that looks
17:22
interesting. And with the genie, I still
17:25
don't have infinite time to do that. But
17:28
if I choose to use it, I have time to
17:31
the word indulge my curiosity. I don't
17:33
want I don't invest in my curiosity
17:36
>> or sate it to some extent.
17:38
>> Yeah. I would never sate it.
17:40
>> Okay. Go towards sating it a tiny bit.
17:44
>> Yeah. Another line of skills that I'd
17:47
love your take on is
17:50
soft skills. Soft skills. Do soft ski
17:54
interpersonal skill like intrapersonal
17:56
and intrapersonal skills. I'll give you
17:58
the softball setup. Are those more or
18:00
less important? Trisha?
18:03
>> Yes. I my co-author Helen Scott wrote a
18:06
blog post about this. She wrote about
18:07
how soft skills makes it sound bad as if
18:10
they're just not as important as the
18:11
hard skills. And I I think I can't
18:14
remember if I've done a video about it,
18:15
but I've talked before about how I hate
18:18
the term good communication skills. So
18:21
I've done a lot of CV and resume
18:22
screening and all developers put at the
18:24
top of it good communication skills and
18:26
then they spend like three paragraphs
18:28
explaining all their technical skills.
18:30
So they've put all of their soft skills
18:32
in good communication skills. And yet
18:34
that even if you're talking about just
18:35
the communication bit, let alone all the
18:37
other soft skills, you've got all sorts
18:38
of different types of communication. You
18:40
have asynchronous communication via
18:42
email like written communication. You've
18:44
got like spoken communication. You've
18:46
got team communication. You've got
18:47
talking up to your boss. You've got
18:48
talking down to your minions. You know,
18:50
you got customer speak. Some people are
18:52
good at one particular set of
18:55
communication skills and not everything.
18:57
And that's fine. So, but we super
18:59
underrate all of these skills. We just
19:01
group it as like I have soft skills.
19:03
Like no, you don't. You're a developer.
19:05
Like I'm going to be rude.
19:07
Whoa, whoa, whoa.
19:09
Shots fired. Oh, we're almost out of
19:12
time.
19:13
>> Oh,
19:14
>> so actually I do want to explore the
19:16
fact that we do have soft skills as
19:18
developers and we we chronically
19:20
underrate them. We chronically don't um
19:24
like you're saying invest in them. We
19:26
don't put them on our resumes. We don't
19:29
uh look for them in in interviews. And
19:31
these soft skills are things like things
19:33
I mentioned before like asking the
19:35
business did you really want a big red
19:36
button which prints this thing or what
19:38
was it you were really trying to do the
19:40
sort of business an analysis skills that
19:41
most developers need things like the
19:44
ability to say no that's a skill that as
19:46
you get more senior you really need to
19:48
be able to prioritize like direct your
19:51
time say no and you know and communicate
19:54
with stakeholders set expectations all
19:56
these sorts of things even if you're
19:57
still an IC writing code you still need
19:59
to be able to do all those things. And
20:01
now that we're talking to the genie,
20:02
these skills are even more important.
20:04
Like, I'm trying to do a thing. This is
20:06
important. This is less important.
20:08
Please prioritize these things. I um you
20:10
know that the answer is always it
20:11
depends for a bunch of stuff. So, being
20:13
able to list out it depends on these
20:15
things and these are my answers to these
20:17
kinds of like knobs and twiddles. And
20:19
these are things that we're we're as
20:21
developers are always doing in our
20:22
heads, but we're not we don't put it
20:24
down on paper that this is the the skill
20:27
that we use. And we don't say you know
20:29
we don't have names for half of these
20:31
skills but we are talking about
20:32
communication skills soft skills people
20:35
skills the main thing I think a
20:37
developer does especially these days
20:39
when developers don't just write code
20:41
preai they don't just write write code
20:43
they do talk to business um stakeholders
20:45
they do do devops they do potentially
20:47
talk to customers like these these
20:50
people developers are a bridge between
20:53
the the silicon and the end user or the
20:56
customer who asks for Correct. That's a
20:58
really difficult thing to do. No matter
21:01
how you get the operations onto the
21:03
silicon and where the silicon is hosted,
21:05
it's very difficult to take that
21:07
customer requirement of I want a big red
21:10
button that does a report and come up
21:12
with something which really meets their
21:13
needs. And that requires a particular
21:15
set of skills. And it doesn't matter if
21:17
you write that, handwrite it in an IDE
21:19
with Java or translate it into correct
21:22
prompts in the in the Genie that the the
21:26
skill is being able to take that fluffy
21:28
thing and turn it into something that a
21:30
user can really use. And that's what
21:34
developers do. That's what developers
21:36
have always done. And that's why things
21:38
like human readable business language
21:41
has never really taken off and why
21:43
things like the Genie is not going to
21:44
replace developers.
21:45
>> Yeah. Agreed.
21:47
>> We the the age of augmented development
21:50
means that we need to exercise those
21:54
skills more frequently. So there's more
21:57
leverage in knowing how to say we can do
22:00
anything but we can't do everything.
22:02
Having a sense of slicing. Okay. So you
22:07
want all all of these things in this
22:09
bag. Uh which one would you like first
22:11
sir for ma'am?
22:13
>> Yeah. Um, and we're exercising those
22:16
skills once a day instead of once a a
22:19
week, once a month, the way that we had
22:21
had been before, because we can get to a
22:23
point where we're ready for feedback.
22:25
We've made decisions, we have something
22:27
concrete, we're ready for feedback. You
22:29
know, if that happens once a day instead
22:32
of once a month, then anything around
22:36
getting that feedback, listening to it,
22:38
absorbing it, communicating it, all of
22:41
those things become more highly
22:42
leveraged than before. I want to be
22:44
aware of time. There's the clock behind
22:46
you, so I figure you can't see it, but I
22:49
can. Can you uh can you close us out?
22:51
>> Like, so we were talking about what
22:53
skills do devs actually need in the age
22:55
of AI code. You mentioned confidence
22:57
which is actually something which is
22:59
eroding somewhat as we worry about our
23:01
skills. You mentioned
23:03
uh curiosity which I really love and and
23:06
I wanted to I forgot I wanted to follow
23:08
up on something there which was one of
23:10
the things as someone who tries to
23:11
educate developers one of the things we
23:13
know is that developers learn in
23:14
different ways depending on who you are.
23:16
And I I love the idea of asking the
23:19
genie to translate its teaching style
23:22
into something that works for you. But
23:24
that means that you have to understand
23:25
yourself a little bit better and you
23:27
have to learn a bit more about yourself
23:29
which I think we have not always been
23:31
good at recognizing that we're not
23:33
autonomous. We are humans and we have
23:35
special individual needs which need to
23:37
be catered for. Uh so you mentioned
23:39
curiosity. I mean all the skills we
23:42
developed with testing in terms of like
23:44
knowing what good tests look like,
23:45
what's what why unhappy patterns matter.
23:48
uh like how we might want to make sure
23:51
that certain bits of the code are tested
23:52
and and why we might want to do that,
23:54
that kind of thing. And stuff we were
23:56
talking about at the end, which I can't
23:56
remember what that was already cuz it's
23:58
like my brain went black.
23:58
>> Uh soft skills.
24:00
>> Soft skills, of course. Communication
24:02
and being able to break problems down
24:03
into smaller pieces and being able to to
24:06
push back and say no and prioritize and
24:08
and decide on the the different values
24:11
of certain things and embracing this
24:13
fast feedback loop as well.
24:14
>> Oh, cool.
24:15
>> Well, well, what are we going to do next
24:17
week? What are we going to do next week?
24:18
Well, you and I will be in Utah next
24:20
week talking about the future of
24:21
software development.
24:22
>> Absolutely. I'm looking forward to that.
24:25
Trisha, thank you so much for your time
24:26
today. Thank you very much for being
24:28
here and sharing your thoughts with
Start a conversation
Ask anything about What Skills Do Developers NEED In An AI Future?
100/100 free messages remaining
AI responses may be inaccurate. Verify important information.
What Skills Do Developers NEED In An AI Future?
Continuous Delivery