TRANSKRIP 🔗
WEBVTT
00:00:00.000 --> 00:00:03.439
A healthy organization needs to have a substantial
00:00:03.439 --> 00:00:06.580
digital workforce. You could foster bad habits
00:00:06.580 --> 00:00:08.500
that you've been fostering for the last five
00:00:08.500 --> 00:00:10.400
years in your organization, right? You could
00:00:10.400 --> 00:00:13.080
exponentially be making things worse. 18 months
00:00:13.080 --> 00:00:15.800
at the current rate of change is equivalent to
00:00:15.800 --> 00:00:18.140
five years a few years ago. What is the most
00:00:18.140 --> 00:00:21.039
exciting thing you're seeing in emerging technology?
00:00:21.079 --> 00:00:22.640
I'm going to try to take a swing at this one
00:00:22.640 --> 00:00:25.039
and not get too excited or aggressive, Tarek.
00:00:40.390 --> 00:00:44.259
Yo, yo, yo. Welcome to Ready, Set, Runtime. a
00:00:44.259 --> 00:00:47.979
podcast from ahead. I'm Michael Kulbrenner, Build
00:00:47.979 --> 00:00:51.950
CTO. People call me MK. I've spent more than
00:00:51.950 --> 00:00:54.689
20 years delivering enterprise technology, and
00:00:54.689 --> 00:00:57.789
AI is the most significant innovation we've seen
00:00:57.789 --> 00:01:00.719
since the introduction of the internet. AI is
00:01:00.719 --> 00:01:03.420
reshaping not just what software does, but how
00:01:03.420 --> 00:01:07.400
we plan it, build it, secure it, and pay for
00:01:07.400 --> 00:01:10.750
it. Over the next eight episodes, we'll follow
00:01:10.750 --> 00:01:13.870
that change across the full stack. talking with
00:01:13.870 --> 00:01:17.629
engineers, architects, operators, and the leaders
00:01:17.629 --> 00:01:21.430
doing the work. Joining me today are my colleagues,
00:01:21.790 --> 00:01:25.109
Josh Perkins and Craig Martin. Guys, please introduce
00:01:25.109 --> 00:01:28.230
yourselves. I'm Josh Perkins. I look after our
00:01:28.230 --> 00:01:31.010
emerging technologies ecosystem at AHAD, which
00:01:31.010 --> 00:01:32.590
is really just the way that we try to work with
00:01:32.590 --> 00:01:34.409
our customers and partners to figure out what's
00:01:34.409 --> 00:01:36.829
next and what do we need to do as an organization.
00:01:37.209 --> 00:01:39.890
And I'm Craig Martin, VP of Advanced Engineering
00:01:39.890 --> 00:01:43.659
here. I focus on accounts that are Deep in trying
00:01:43.659 --> 00:01:45.500
to modernize how they think and work and get
00:01:45.500 --> 00:01:48.079
more value from from tech I work really closely
00:01:48.079 --> 00:01:50.980
with Josh on a on a daily basis of just trying
00:01:50.980 --> 00:01:53.319
to drive our customers forward Yeah, this is
00:01:53.319 --> 00:01:55.879
an amazing time for the first time I think in
00:01:55.879 --> 00:01:59.620
a long time and I would say 20 years We're rethinking
00:01:59.620 --> 00:02:02.519
about how we deliver technology to the enterprise
00:02:02.519 --> 00:02:05.159
and AI is a big component of that But there's
00:02:05.159 --> 00:02:08.280
a process part. There's a human part I'd love
00:02:08.280 --> 00:02:11.080
for you guys to talk about what you think that
00:02:11.080 --> 00:02:14.270
expression means Because people talk about accelerated
00:02:14.270 --> 00:02:17.289
development, but I'm not sure it's well understood.
00:02:18.030 --> 00:02:21.099
Josh, do you want to start with that? Yeah. I
00:02:21.099 --> 00:02:23.180
mean, this is maybe where Craig and I will be
00:02:23.180 --> 00:02:25.860
most interesting kind of overlap on what do we
00:02:25.860 --> 00:02:28.659
think the scope of this is. I think that broadly
00:02:28.659 --> 00:02:32.639
speaking, when we say AI accelerated development
00:02:32.639 --> 00:02:35.460
or AIS DLC, what we're really talking about is
00:02:35.460 --> 00:02:40.360
the reformation of the practices and ultimately
00:02:40.360 --> 00:02:43.650
the tool chains involved with the creation. and
00:02:43.650 --> 00:02:47.030
operationalization of digital products that we
00:02:47.030 --> 00:02:49.069
built by nature of the work that we're doing
00:02:49.069 --> 00:02:51.669
in an organization. And I think that the reason
00:02:51.669 --> 00:02:55.800
I'm a little bit more wide open on that is because
00:02:55.800 --> 00:02:58.139
I think the scope of what we mean by this is
00:02:58.139 --> 00:03:01.639
grossly changing versus what we meant in traditional
00:03:01.639 --> 00:03:05.439
SDLC parlance back in the day. And I think people
00:03:05.439 --> 00:03:08.400
participating in that process is grossly changing
00:03:08.400 --> 00:03:12.599
as a consequence of this as well. And I think
00:03:12.599 --> 00:03:15.340
the AI component of it is really not just about
00:03:15.340 --> 00:03:19.520
how do we improve, you know, front end code generation
00:03:19.520 --> 00:03:23.639
or dynamic design, like, or even something like
00:03:23.639 --> 00:03:25.819
AI assisted testing. It's literally how do we
00:03:25.819 --> 00:03:28.180
look at every particular phase gate in the inner
00:03:28.180 --> 00:03:30.580
and outer loop and say, what, what, what makes
00:03:30.580 --> 00:03:32.759
sense in this new world to maintain velocity
00:03:32.759 --> 00:03:36.500
across this in such a way that I can start measuring
00:03:36.500 --> 00:03:39.560
outcomes instead of inputs. Right. Like, and
00:03:39.560 --> 00:03:41.879
I think that's ultimately where the organizations
00:03:41.879 --> 00:03:44.939
want to go. And I think that the other big thing
00:03:44.939 --> 00:03:47.699
for me in this situation is I think the idea
00:03:47.699 --> 00:03:50.900
of like software as the basis of this has a very
00:03:50.900 --> 00:03:54.000
like anchoring mental effect in people's minds
00:03:54.000 --> 00:03:55.900
around like we're going to go build an application.
00:03:56.520 --> 00:03:58.180
When I really think what we're talking about
00:03:58.180 --> 00:04:02.020
here is how do we create a pipeline for the creation
00:04:02.020 --> 00:04:04.039
and assimilation of
Load full transcript
digital labor in the long
00:04:04.039 --> 00:04:06.960
term for our businesses. And that is a massively
00:04:06.960 --> 00:04:10.139
different dynamic than I want to go build a new
00:04:10.139 --> 00:04:12.840
enterprise application. Not surprising. Josh
00:04:12.840 --> 00:04:16.980
and I are very aligned on this. I probably won't
00:04:16.980 --> 00:04:19.740
be enough on AI accelerated dev, but the way
00:04:19.740 --> 00:04:24.199
I see it is we are headed towards. A healthy
00:04:24.199 --> 00:04:27.560
organization needs to have a substantial digital
00:04:27.560 --> 00:04:30.339
workforce. One that's out there working on their
00:04:30.339 --> 00:04:32.920
behalf, helping to make decisions, maybe making
00:04:32.920 --> 00:04:37.819
decisions and executing. And many customers that
00:04:37.819 --> 00:04:39.459
I'm working with are saying, how do I do this?
00:04:39.620 --> 00:04:42.110
How do I have... Hundreds, if not thousands of
00:04:42.110 --> 00:04:44.730
agents and out and working on my behalf. And
00:04:44.730 --> 00:04:46.970
what do I, how do I secure it? How do I bring,
00:04:47.110 --> 00:04:48.850
when do I decide to bring in humans? How do they
00:04:48.850 --> 00:04:51.689
orchestrate that? How do I measure that? And,
00:04:51.689 --> 00:04:56.629
and, and, and, and do I, do I apply organizational
00:04:56.629 --> 00:04:59.069
structures to this? Or do I go back to what we
00:04:59.069 --> 00:05:01.889
learned in microservices 10, 15 years ago about
00:05:01.889 --> 00:05:03.769
domain -driven design and start to build your
00:05:03.769 --> 00:05:05.970
organizations around these domains. And maybe,
00:05:06.029 --> 00:05:09.069
maybe my agents fit in, in, in, in those domains.
00:05:09.069 --> 00:05:13.009
And, and, and. And that to me is a really exciting
00:05:13.009 --> 00:05:16.060
time, not purely to like... you know, reduce
00:05:16.060 --> 00:05:19.639
the cost of labor or, or reduce workforces, but
00:05:19.639 --> 00:05:22.360
to have humans get more value. And then, then
00:05:22.360 --> 00:05:25.180
these digital personas are continuing to do things
00:05:25.180 --> 00:05:27.779
that, that, that maybe isn't as fulfilling for
00:05:27.779 --> 00:05:30.399
humans to do. And I, and I think, I think that's
00:05:30.399 --> 00:05:33.740
that we're seeing that a lot where an entry point
00:05:33.740 --> 00:05:36.240
for this. And I think what Josh alluded to it
00:05:36.240 --> 00:05:39.120
is software development is a very easy place
00:05:39.120 --> 00:05:41.600
to start to bring this style of thinking, but
00:05:41.600 --> 00:05:46.420
it's kind of the tip of the spear to. fraud detection,
00:05:47.620 --> 00:05:52.519
credit scoring, to employee onboarding, to all
00:05:52.519 --> 00:05:55.120
sorts of plethora of use cases. And if we could
00:05:55.120 --> 00:05:59.800
do this in our SDLC, then we certainly can do
00:05:59.800 --> 00:06:02.319
it in all other aspects of the business. So I
00:06:02.319 --> 00:06:05.050
look at it as... You know i'll be a beach head
00:06:05.050 --> 00:06:07.050
i don't know the right word is for this but but
00:06:07.050 --> 00:06:11.350
it's it's way bigger than just software development
00:06:11.350 --> 00:06:15.290
and sdlc and bringing ai thinking to that it's
00:06:15.290 --> 00:06:18.310
it's just tends to be where we often start it's
00:06:18.310 --> 00:06:21.850
an obvious place to start to because. The people
00:06:21.850 --> 00:06:26.029
involved tend to be extremely open to technology
00:06:26.029 --> 00:06:27.949
adoption and change quickly right and that's
00:06:27.949 --> 00:06:30.949
not a. negative to other portions of the business
00:06:30.949 --> 00:06:33.790
greater proximity right to that and that change
00:06:33.790 --> 00:06:36.250
rate so like they tend to be a little bit more
00:06:36.250 --> 00:06:39.110
open to self disruption in in that segment and
00:06:39.110 --> 00:06:41.750
familiarity with the tools and technical fluency
00:06:41.750 --> 00:06:44.430
tends to be thrive right in that particular dimension
00:06:44.430 --> 00:06:46.990
the businesses but you're right i think like
00:06:46.990 --> 00:06:49.209
the two things that can't be understated here
00:06:49.209 --> 00:06:53.730
like. One the practices and procedures will translate
00:06:53.730 --> 00:06:56.350
really well to other pork like parts of the business
00:06:56.350 --> 00:06:58.370
like there's corollaries everywhere we don't.
00:06:58.319 --> 00:07:00.860
choose to look at it that way, but we will write
00:07:00.860 --> 00:07:03.279
like as we start decompiling other business process.
00:07:03.879 --> 00:07:06.459
And then I think the other part that's like super
00:07:06.459 --> 00:07:09.459
critical on this is we have all this multi stage,
00:07:09.660 --> 00:07:13.420
multi step flow in the development process, which
00:07:13.420 --> 00:07:15.480
like by and large means we're going to have to
00:07:15.480 --> 00:07:18.019
get really good at like how we verify things,
00:07:18.100 --> 00:07:20.639
how, how we build trust and confidence and delegation,
00:07:20.639 --> 00:07:22.360
like all these things that we're going to have
00:07:22.360 --> 00:07:25.180
to do in every other meaningful business process.
00:07:25.550 --> 00:07:28.949
we can start here and it's highly, highly verifiable
00:07:28.949 --> 00:07:31.870
versus like, I have to always think about what's
00:07:31.870 --> 00:07:33.629
the definition of done in software development.
00:07:33.930 --> 00:07:35.990
I know what the definition of done is. Right.
00:07:36.110 --> 00:07:39.730
So like it's, there's just, it's a great like
00:07:39.730 --> 00:07:42.350
catalyst for like the getting a more robust set
00:07:42.350 --> 00:07:43.750
of capabilities together for the rest of the
00:07:43.750 --> 00:07:47.160
organization. And I want to beat this point a
00:07:47.160 --> 00:07:50.319
little bit further. I'm very literal. And so
00:07:50.319 --> 00:07:52.819
I think SDLC software development life cycle,
00:07:52.939 --> 00:07:55.560
but I agree with what you're both saying, it's
00:07:55.560 --> 00:07:59.480
really the X DLC. It's really delivery across
00:07:59.480 --> 00:08:01.839
the organization. And it's true that we're able
00:08:01.839 --> 00:08:04.480
to demonstrate it, I think, in accelerated software
00:08:04.480 --> 00:08:07.860
first, but we're starting to see it adopted across
00:08:07.860 --> 00:08:11.199
the organization. And as importantly, I feel
00:08:11.199 --> 00:08:13.939
like we are starting to model the behavior of
00:08:13.939 --> 00:08:17.899
how to use AI so that our partners can see these
00:08:17.899 --> 00:08:21.459
examples and say, oh, I have a pattern similar
00:08:21.459 --> 00:08:25.160
to that. I can apply the same principles for
00:08:25.160 --> 00:08:28.680
a different outcome. And I think that's a key
00:08:28.680 --> 00:08:31.000
part of what we're all doing right now is modeling
00:08:31.000 --> 00:08:34.700
this behavior. That paradigm shift of I have
00:08:34.700 --> 00:08:39.470
a problem, let's bring some sort of... AI or
00:08:39.470 --> 00:08:42.590
agentic thinking to that problem, I think is
00:08:42.590 --> 00:08:44.870
a very healthy paradigm shift. And I'm watching
00:08:44.870 --> 00:08:47.129
so many different organizations go through that
00:08:47.129 --> 00:08:49.070
where at first they've got a problem. Okay, well,
00:08:49.250 --> 00:08:52.070
how can we automate this? And let's just think
00:08:52.070 --> 00:08:55.029
about this through, you know, Terraform or some
00:08:55.029 --> 00:08:58.009
sort of workflow engine or, or, or, and then,
00:08:58.149 --> 00:09:01.259
and then it's, let's step back. What if we had
00:09:01.259 --> 00:09:03.200
a digital worker who could, who could do this
00:09:03.200 --> 00:09:05.580
and listen to events and, and when we trust them
00:09:05.580 --> 00:09:08.240
and verify and, and, and then, and that, that,
00:09:08.240 --> 00:09:10.980
that paradigm shift is a very healthy paradigm
00:09:10.980 --> 00:09:13.279
shift for organizations to go through. When you
00:09:13.279 --> 00:09:16.120
think about the changes that are occurring, what
00:09:16.120 --> 00:09:18.840
has occurred? What has been the main triggering
00:09:18.840 --> 00:09:21.679
factor that has changed the way we think about
00:09:21.679 --> 00:09:25.240
AI or delivery in the enterprise? Josh, let's
00:09:25.240 --> 00:09:28.259
start with you. I think, you know, everything
00:09:28.259 --> 00:09:31.399
that. the enterprise is currently focused on
00:09:31.399 --> 00:09:34.620
is predicated on speed, right? Like the reality
00:09:34.620 --> 00:09:38.299
of trying to move quicker and ultimately spend
00:09:38.299 --> 00:09:42.100
more of their time on what's next and not necessarily
00:09:42.100 --> 00:09:44.899
the huge amount of burden that's existed on what
00:09:44.899 --> 00:09:47.120
they've previously done and accumulated over
00:09:47.120 --> 00:09:50.799
time. And so I think this is a, you know, the
00:09:50.799 --> 00:09:55.220
focus on the SDLC is really like, it's a multi
00:09:55.220 --> 00:09:57.740
-variable. thing for most enterprises. I think
00:09:57.740 --> 00:10:00.259
at first it was like, how do we do more new stuff
00:10:00.259 --> 00:10:02.779
faster, right? Like the, how do we accelerate
00:10:02.779 --> 00:10:05.399
our ability to write code? But I think increasingly
00:10:05.399 --> 00:10:07.639
we're watching people realize that it's also
00:10:07.639 --> 00:10:11.179
a huge opportunity to rethink and retire technical
00:10:11.179 --> 00:10:13.179
bet. And so they're kind of approaching it in
00:10:13.179 --> 00:10:16.960
multiple different ways. Um, and that at some
00:10:16.960 --> 00:10:19.179
point, like if we can do the ladder really well,
00:10:19.179 --> 00:10:21.220
it actually frees up more time and effort for
00:10:21.220 --> 00:10:23.700
the former. And so there's a bit of this like,
00:10:23.889 --> 00:10:27.110
If we can do some of these things like go slow
00:10:27.110 --> 00:10:28.870
a little bit at the beginning be very deliberate
00:10:28.870 --> 00:10:31.470
about what we do it could be a massive acceleration
00:10:31.470 --> 00:10:34.649
right over time for the organization. I think
00:10:34.649 --> 00:10:37.009
the big challenge though at the moment is like.
00:10:37.769 --> 00:10:40.309
We can probably go much faster with the technology
00:10:40.309 --> 00:10:43.049
than we can with the adoption and the people
00:10:43.049 --> 00:10:45.250
components of the orcs that we're deploying these
00:10:45.250 --> 00:10:48.429
things into and so there's like this very natural.
00:10:49.320 --> 00:10:52.240
kind of gravity, right, that's offsetting some
00:10:52.240 --> 00:10:55.159
of the potential speed benefit, which is just
00:10:55.159 --> 00:10:58.360
like, could we do it? Should we do it? And we
00:10:58.360 --> 00:11:00.840
do it right at the same speed that we adopt normally
00:11:00.840 --> 00:11:03.120
in the enterprise. And they didn't get big by
00:11:03.120 --> 00:11:06.200
being stupid. So the reality is like these things
00:11:06.200 --> 00:11:09.139
can naturally and intentionally go slowly. And
00:11:09.139 --> 00:11:11.480
that creates a velocity mismatch to a large degree.
00:11:12.480 --> 00:11:14.659
Yeah, I agree. I mean, I'll add a little bit
00:11:14.659 --> 00:11:18.149
to that. I think AI is moving faster than a lot
00:11:18.149 --> 00:11:21.330
of other epochs of computing and changes than
00:11:21.330 --> 00:11:27.559
we've seen in the past. Uh, but the story of
00:11:27.559 --> 00:11:30.320
the tech is amazing. We can do amazing things
00:11:30.320 --> 00:11:33.580
that we couldn't do before. Oh gosh, how do I
00:11:33.580 --> 00:11:36.700
think about business processes and ways of working
00:11:36.700 --> 00:11:39.460
and, and that's lagging behind. Like that's not
00:11:39.460 --> 00:11:42.559
really a
new story. I think the speed at which,
00:11:42.559 --> 00:11:46.000
uh, AI is changing is really accentuating that
00:11:46.000 --> 00:11:47.639
story. It's really easy to just look at the tech,
00:11:47.919 --> 00:11:49.440
but we start to look at, okay, how do I start
00:11:49.440 --> 00:11:53.039
to organize and not just automate or bring AI
00:11:53.039 --> 00:11:56.840
to existing processes. or rethink them, fundamentally
00:11:56.840 --> 00:11:59.440
overhaul how they work and operate. Related to
00:11:59.440 --> 00:12:06.399
that with AI and AIS DLC, it is, I believe, a
00:12:06.399 --> 00:12:09.159
huge differentiator for organizations in the
00:12:09.159 --> 00:12:12.159
next five years, or maybe even the next year.
00:12:12.639 --> 00:12:17.120
If you can figure out how to do agentic thinking,
00:12:17.519 --> 00:12:21.879
AI engineering, and get value from your tools,
00:12:22.179 --> 00:12:24.860
better and more effectively than your competition,
00:12:25.100 --> 00:12:26.879
you're probably going to win and you're going
00:12:26.879 --> 00:12:28.460
to win in the next couple of years. And, and
00:12:28.460 --> 00:12:30.980
I, and I think I don't want to make, you know,
00:12:31.059 --> 00:12:33.440
be that Darwinian about it, but, but I think
00:12:33.440 --> 00:12:36.360
that mentality is, is a healthy one. And many
00:12:36.360 --> 00:12:39.460
organizations are realizing they need to fight
00:12:39.460 --> 00:12:43.299
for that. If, if it is so is how I think of it.
00:12:43.450 --> 00:12:47.159
I think like even to that point, I think like
00:12:47.159 --> 00:12:50.399
the lack of the lack of newness is also the,
00:12:50.399 --> 00:12:53.820
like the reality that this is an extension of
00:12:53.820 --> 00:12:56.500
the last 10 years of work and the people that
00:12:56.500 --> 00:12:59.039
we see. Like the people that I work with every
00:12:59.039 --> 00:13:01.419
day that are like making the fastest strides
00:13:01.419 --> 00:13:03.799
are the ones that were working on the precursory
00:13:03.799 --> 00:13:06.659
elements to this well before accelerated STLC
00:13:06.659 --> 00:13:10.080
was even a thing right like and and we talk about
00:13:10.080 --> 00:13:12.820
all the time this idea that you know what are
00:13:12.820 --> 00:13:15.360
the what are the things that are like positive
00:13:15.360 --> 00:13:16.860
indicators of whether or not you're going to
00:13:16.860 --> 00:13:19.080
be successful in this like a gentically driven
00:13:19.080 --> 00:13:22.320
or AI driven era and it's like well. How good
00:13:22.320 --> 00:13:24.740
were you on the data readiness side of the house,
00:13:24.879 --> 00:13:26.899
right? How good were you at the DevOps insect
00:13:26.899 --> 00:13:28.679
DevOps transformation that you were supposed
00:13:28.679 --> 00:13:31.480
to previously make? Like how good are you? Were
00:13:31.480 --> 00:13:34.860
how good were you at completing the digital transformation
00:13:34.860 --> 00:13:37.899
tasks that pre were predated this this transformation
00:13:37.899 --> 00:13:41.059
phase? Because if like the ingredients in this
00:13:41.059 --> 00:13:43.779
phase are good, like you will go faster. And
00:13:43.779 --> 00:13:46.059
the other consequence is you're not trying to
00:13:46.059 --> 00:13:48.559
change your culture and adopt to the technology
00:13:48.559 --> 00:13:50.000
at the same time, because you already worked
00:13:50.000 --> 00:13:52.700
on the culture over the last decade. And so I
00:13:52.700 --> 00:13:55.440
do see this as like very combinatorial, right?
00:13:55.440 --> 00:13:58.419
Where you're like, you know, it's a it's a mix
00:13:58.419 --> 00:14:01.600
of like preparedness. And then NKU and I have
00:14:01.600 --> 00:14:03.340
had this discussion numerous. And it's also a
00:14:03.340 --> 00:14:06.500
mix of like the deep rooted culture and nature
00:14:06.500 --> 00:14:08.340
of the business that we're in, which is why we
00:14:08.340 --> 00:14:12.480
see like massive strides in the AI natives first,
00:14:12.480 --> 00:14:14.990
because it's like da. Right. But then immediately
00:14:14.990 --> 00:14:16.889
upon that, you're like, well, who's making the
00:14:16.889 --> 00:14:19.070
biggest progress? And it's like, well, dude,
00:14:19.350 --> 00:14:21.750
the front end of capital markets and financial
00:14:21.750 --> 00:14:23.730
services are moving very, very quickly in this
00:14:23.730 --> 00:14:26.649
space. The biopharma organizations are moving
00:14:26.649 --> 00:14:29.129
very quickly in these spaces. And that's because
00:14:29.129 --> 00:14:31.230
like they're used to throwing 80 percent of the
00:14:31.230 --> 00:14:33.509
work on the cutting room floor and moving on
00:14:33.509 --> 00:14:36.169
because the juice is worth the squeeze. And most
00:14:36.169 --> 00:14:38.929
of our enterprise clients, that's just like the
00:14:38.929 --> 00:14:41.629
antithesis of the way they operate. And so you
00:14:41.629 --> 00:14:43.850
do have this like big, again, like the people
00:14:43.850 --> 00:14:47.250
process, cultural change that comes from doing
00:14:47.250 --> 00:14:50.250
this, I think is not one to be like taken too
00:14:50.250 --> 00:14:52.509
lightly. And we need to cascade the lessons learned
00:14:52.509 --> 00:14:55.789
as quickly as possible. You know, if you are
00:14:55.789 --> 00:14:59.769
too fast in your adoption of accelerated development,
00:14:59.929 --> 00:15:02.990
then as a result, you could foster bad habits
00:15:02.990 --> 00:15:04.870
that you've been fostering for the last five
00:15:04.870 --> 00:15:06.769
years in your organization, right? You could
00:15:06.769 --> 00:15:09.309
exponentially be making things worse to your
00:15:09.309 --> 00:15:11.669
point about. How are you getting things done
00:15:11.669 --> 00:15:14.450
previous to your deployment of accelerated tools?
00:15:15.009 --> 00:15:17.389
You know, were you standardizing on a common
00:15:17.389 --> 00:15:22.049
stack? Were you using like good DevOps hygiene?
00:15:22.230 --> 00:15:24.830
Those things all start to become bigger problems
00:15:24.830 --> 00:15:27.230
if you use accelerated dev. You know, a lot of
00:15:27.230 --> 00:15:30.330
times we talk about best practice. I'm curious,
00:15:30.350 --> 00:15:33.889
I'd like to turn that around. What are the worst
00:15:33.889 --> 00:15:39.100
practices you've seen companies or people do
00:15:39.100 --> 00:15:42.450
that? sort of were obvious errors in adopting
00:15:42.450 --> 00:15:45.129
accelerated development. And then we'll talk
00:15:45.129 --> 00:15:47.990
about sort of the good side of that coin. What
00:15:47.990 --> 00:15:50.690
are the best things we've seen? I'll let Josh
00:15:50.690 --> 00:15:55.629
go first on that one. I think like to our earlier
00:15:55.629 --> 00:15:58.169
point on velocity matching, I think the biggest
00:15:58.169 --> 00:16:02.639
early thing that we saw was people invoked tools
00:16:02.639 --> 00:16:05.840
or gave people tools that massively changed part
00:16:05.840 --> 00:16:09.480
of the overall process. but just created more
00:16:09.480 --> 00:16:12.679
downstream strife as we went, right? So like,
00:16:12.879 --> 00:16:14.860
if I just give everyone code generation, which
00:16:14.860 --> 00:16:16.700
is very naturally what happened at the beginning
00:16:16.700 --> 00:16:20.779
of a lot of this, maybe 18 months ago, you know,
00:16:20.840 --> 00:16:23.000
all I did was create a massive top of funnel
00:16:23.000 --> 00:16:26.620
challenge with kind of poor quality code. And
00:16:26.620 --> 00:16:29.220
it became a pretty untenable scenario for teams
00:16:29.220 --> 00:16:30.759
that were supposed to be responsible for trying
00:16:30.759 --> 00:16:35.070
to translate that into like business value. I
00:16:35.070 --> 00:16:37.389
think the other piece here is like we saw people
00:16:37.389 --> 00:16:39.590
kind of conflate the idea that code generation
00:16:39.590 --> 00:16:42.269
equals software engineering expertise and excellence
00:16:42.269 --> 00:16:46.129
and they are not the same thing. And so the early
00:16:46.129 --> 00:16:49.269
narratives created. Green negative reinforcement
00:16:49.269 --> 00:16:51.690
loops in my mind that I've seen in certain organizations
00:16:51.690 --> 00:16:54.110
around the deterioration of the value of the
00:16:54.110 --> 00:16:58.820
role. um, the ideas that, you know, citizen development
00:16:58.820 --> 00:17:01.120
is going to become this de facto standard across
00:17:01.120 --> 00:17:03.240
the organization for like the fifth time, by
00:17:03.240 --> 00:17:05.359
the way, that we've gone through this, this motif.
00:17:06.079 --> 00:17:08.660
Um, and I think that we'll get there, but it's
00:17:08.660 --> 00:17:11.619
like, you can't skip the steps, right? Like everything
00:17:11.619 --> 00:17:13.980
that organizations that are functional on this
00:17:13.980 --> 00:17:16.740
are doing is they're now looking at every particular
00:17:16.740 --> 00:17:19.539
piece of a process. They're like six sigma to
00:17:19.539 --> 00:17:22.940
a degree, the entire developmental flow and saying
00:17:22.940 --> 00:17:27.170
like, how do I make sure that No new 10 X in
00:17:27.170 --> 00:17:29.690
bottlenecks can occur through this process and
00:17:29.690 --> 00:17:31.450
if anything like how do I continue to look at
00:17:31.450 --> 00:17:35.509
this for like infinite scale principles and this
00:17:35.509 --> 00:17:38.190
is really important because of it it it take
00:17:38.190 --> 00:17:39.849
us out of the idea that we're just gonna have
00:17:39.849 --> 00:17:44.190
a you know. strategically isolated pipeline that's
00:17:44.190 --> 00:17:45.910
focused on enterprise application development
00:17:45.910 --> 00:17:48.190
and turns it into something as a democratized
00:17:48.190 --> 00:17:51.970
centralization and pipeline element for everyone
00:17:51.970 --> 00:17:55.250
coming into the top of that scenario. It's not
00:17:55.250 --> 00:17:56.869
only a flow matching for the engineering teams,
00:17:56.990 --> 00:17:58.750
it becomes a flow matching for the organizational
00:17:58.750 --> 00:18:01.730
desire to build and their ability to interpret
00:18:01.730 --> 00:18:04.829
that into production grade trustable outcomes.
00:18:05.750 --> 00:18:08.049
But early on, man, I'll tell you what, it was
00:18:08.740 --> 00:18:11.299
People can make very convincing front -end things
00:18:11.299 --> 00:18:15.000
with very bad backends, very bad security parameters.
00:18:15.599 --> 00:18:20.099
They weren't software engineers. Much like our
00:18:20.099 --> 00:18:23.700
chatbot phase in AI, what we saw was a bunch
00:18:23.700 --> 00:18:28.220
of really convincing answers that deteriorated
00:18:28.220 --> 00:18:30.960
and hallucinated in the background very, very
00:18:30.960 --> 00:18:33.220
quickly. The same was true for code. It's just
00:18:33.220 --> 00:18:36.680
that code looked syntax verifiable as a consequence.
00:18:37.549 --> 00:18:39.750
Yeah, and I'll add, I mean, maybe even echo a
00:18:39.750 --> 00:18:42.289
little bit. The biggest couple of mistakes I
00:18:42.289 --> 00:18:44.970
see. One, and I'm seeing this a little less now,
00:18:45.210 --> 00:18:49.710
but many organizations were afraid to go forward
00:18:49.710 --> 00:18:52.369
because they were afraid to make mistakes. The
00:18:52.369 --> 00:18:56.109
landscape was changing so rapidly. And so they
00:18:56.109 --> 00:18:58.450
sat there and waited and watched. And I think
00:18:58.450 --> 00:19:02.630
that's a big mistake. I think every organization's
00:19:02.630 --> 00:19:05.660
going to make mistakes and learnings along the
00:19:05.660 --> 00:19:12.380
way. We're going to bring in AI tools and agentic
00:19:12.380 --> 00:19:14.940
thinking to existing processes and it's going
00:19:14.940 --> 00:19:16.920
to uncover new bottlenecks and
we'll have to
00:19:16.920 --> 00:19:22.880
course correct and fix that. And that fear scares
00:19:22.880 --> 00:19:25.759
many organizations from taking action. And I
00:19:25.759 --> 00:19:28.220
think what's really important is take action,
00:19:28.599 --> 00:19:31.670
make mistakes. fix, correct, take action, make
00:19:31.670 --> 00:19:35.750
mistakes, fix and correct. That's been that way
00:19:35.750 --> 00:19:41.009
for a while and before AI, but it is very scary
00:19:41.009 --> 00:19:44.069
with the rate of change that was going on. I
00:19:44.069 --> 00:19:46.069
think that was a mistake. I'm seeing that a little
00:19:46.069 --> 00:19:50.269
bit less. I think many people overindex to the
00:19:50.269 --> 00:19:53.500
tool. And here's the tool we use. Well, that
00:19:53.500 --> 00:19:57.859
tool, it needs to be used properly. We need checks
00:19:57.859 --> 00:20:00.380
and balances in place. We need potentially agents
00:20:00.380 --> 00:20:05.200
to validate our security. We need strongly defined
00:20:05.200 --> 00:20:09.740
architectural patterns that these skills or agents
00:20:09.740 --> 00:20:14.119
will leverage and ensure that we adhere to. And
00:20:14.119 --> 00:20:16.019
it's not just, hey, look how quickly I can generate
00:20:16.019 --> 00:20:19.710
code. That's great, but Now look how quickly
00:20:19.710 --> 00:20:22.509
I can generate well -written code that adheres
00:20:22.509 --> 00:20:25.049
to my architectural patterns and anti -patterns
00:20:25.049 --> 00:20:27.130
and security standards. Well, now we're actually
00:20:27.130 --> 00:20:31.349
cooking with gas. And I think that comes in...
00:20:31.309 --> 00:20:33.549
thinking about it not from a tool perspective,
00:20:34.069 --> 00:20:36.509
but from a workflow perspective. And how do I
00:20:36.509 --> 00:20:39.750
share these agents and their capabilities? And
00:20:39.750 --> 00:20:42.269
how do I measure them? And how do I know that
00:20:42.269 --> 00:20:46.089
the response that came out is actually adhering
00:20:46.089 --> 00:20:50.789
to my standards? And so what we often find and
00:20:50.789 --> 00:20:55.509
what I preach towards is not necessarily anything's
00:20:55.509 --> 00:20:58.329
fundamentally changed, but with AI tooling and
00:20:58.329 --> 00:21:02.170
agentic thinking, we can get more draconian about
00:21:02.430 --> 00:21:05.829
what our architectures are and define them and
00:21:05.829 --> 00:21:09.150
measure those. And then on the right -hand side
00:21:09.150 --> 00:21:12.130
of development, we have to be more opinionated
00:21:12.130 --> 00:21:14.950
on how we test and maybe do more fitness functions
00:21:14.950 --> 00:21:16.950
than we've ever done before. If I'm gonna let
00:21:16.950 --> 00:21:20.009
the people in the middle here, the digital workers,
00:21:20.369 --> 00:21:23.390
and do more and own more, I have to be more opinionated
00:21:23.390 --> 00:21:25.490
on the left -hand side and I have to be even
00:21:25.490
--> 00:21:27.710
more opinionated and validated on the right -hand
00:21:27.710 --> 00:21:31.529
side. And so I think those are mistakes that
00:21:31.529 --> 00:21:36.980
I'm... But that desire for perfection will kill
00:21:36.980 --> 00:21:40.099
organizations. Go make mistakes, embrace them,
00:21:40.140 --> 00:21:44.079
and adjust. And it is, I think, a healthy mindset
00:21:44.079 --> 00:21:46.759
to have with the landscape that's changing the
00:21:46.759 --> 00:21:49.140
way it is. Yeah. I think battling fast, right?
00:21:49.240 --> 00:21:52.500
How do we move forward? Similarly, disagree and
00:21:52.500 --> 00:21:54.539
commit, right? We might not all agree, but how
00:21:54.539 --> 00:21:57.059
do we continue to always make progress? So this
00:21:57.059 --> 00:21:59.539
conversation naturally segued into best practice.
00:21:59.799 --> 00:22:02.160
best practices, which is what our business is.
00:22:02.539 --> 00:22:04.359
When you think about going to meet with a client,
00:22:04.440 --> 00:22:07.480
if you had an opportunity with a client to greenfield,
00:22:08.200 --> 00:22:12.680
describe, define, help them understand what their
00:22:12.680 --> 00:22:15.859
adoption cycle and pattern is, what would you
00:22:15.859 --> 00:22:18.119
say? If somebody comes and says, you know what,
00:22:18.700 --> 00:22:22.920
we're flexible and open about how we do this.
00:22:23.279 --> 00:22:26.200
We'd like some guidance on tools, on process
00:22:26.200 --> 00:22:29.170
and people. How would you craft a message for
00:22:29.170 --> 00:22:31.869
that discussion? So I think there's twofold a
00:22:31.869 --> 00:22:36.170
little bit. First is, I think there is a natural
00:22:36.170 --> 00:22:40.579
progression of AI -enabled tooling, where I can
00:22:40.579 --> 00:22:43.960
use tools, I get more value from it, it's helped
00:22:43.960 --> 00:22:48.980
my development cycle, and I bring that to all
00:22:48.980 --> 00:22:53.559
the key roles of my STLC. I think that's an entry
00:22:53.559 --> 00:22:57.279
point that we see all the time is, I know I can
00:22:57.279 --> 00:23:03.000
use Cursor or Copilot or any of these tools.
00:23:03.160 --> 00:23:05.339
I want to get value from it, help me get value
00:23:05.339 --> 00:23:07.450
from that. That's an entry point that... that
00:23:07.450 --> 00:23:09.930
many organizations start with. But that progresses
00:23:09.930 --> 00:23:14.730
to more agentic thinking, where actually my agents
00:23:14.730 --> 00:23:17.430
are writing user stories, responding to user
00:23:17.430 --> 00:23:22.529
stories, responding to events that are happening
00:23:22.529 --> 00:23:25.150
behind the scenes, and now humans are starting
00:23:25.150 --> 00:23:29.710
to orchestrate those agents. But it's a progression
00:23:29.710 --> 00:23:34.910
from AI enabled to autonomous agentic thinking.
00:23:35.190 --> 00:23:38.980
But I would say i don't always lead that way
00:23:38.980 --> 00:23:42.079
the reason i don't i
don't like the conversation
00:23:42.079 --> 00:23:44.940
always there is i still feel like that's too
00:23:44.940 --> 00:23:48.680
much of a tech focused tool focused approach
00:23:48.680 --> 00:23:54.150
where. What I tend to prefer is, what's the business
00:23:54.150 --> 00:23:57.269
problem? We have, Josh alluded to earlier, there's
00:23:57.269 --> 00:23:59.730
a slew of tech debt that we have and we have
00:23:59.730 --> 00:24:03.549
to resolve that. That's a real security risk,
00:24:03.869 --> 00:24:07.589
cost risk. I'm taking human capital to manage
00:24:07.589 --> 00:24:10.450
different ways of technologies. I need to fix
00:24:10.450 --> 00:24:14.369
that. There's a real outcome that I'm interested
00:24:14.369 --> 00:24:16.990
in. Well, let's AI enable that. That's maybe
00:24:16.990 --> 00:24:21.599
bring that more to... autonomous agentic thinking,
00:24:22.019 --> 00:24:25.420
but there's a real case. Maybe it's a new application
00:24:25.420 --> 00:24:29.579
that we're trying to build and for a new market.
00:24:30.500 --> 00:24:32.720
And so then bring this style of thinking to a
00:24:32.720 --> 00:24:38.259
real business case. So those, I tend to really
00:24:38.259 --> 00:24:41.759
try to hook into those because I think IT projects
00:24:41.759 --> 00:24:45.529
for the sake of IT projects. tend to really struggle
00:24:45.529 --> 00:24:47.970
and we have to have business outcomes for them.
00:24:48.069 --> 00:24:51.349
And so whether it's tech debt, new market, maybe
00:24:51.349 --> 00:24:56.230
a full modernization of an organization to be
00:24:56.619 --> 00:25:02.140
more focused on, you know, agentic operating
00:25:02.140 --> 00:25:06.859
and whether we see that in healthcare or finance
00:25:06.859 --> 00:25:10.559
or energy industry, like there's so many use
00:25:10.559 --> 00:25:13.680
cases around those. hook into those because those
00:25:13.680 --> 00:25:15.460
are outcomes that the business is gonna care
00:25:15.460 --> 00:25:19.740
about. The larger the organization, the more
00:25:19.740 --> 00:25:22.079
important that we get away from the tool being
00:25:22.079 --> 00:25:26.000
like a critical path at all, right? As sparkly
00:25:26.000 --> 00:25:28.660
what's going on. I think one of the big things
00:25:28.660 --> 00:25:31.259
there is we probably need to use all the tools
00:25:31.259 --> 00:25:33.259
for a period of time. Like at some point in time,
00:25:33.380 --> 00:25:35.079
a different team within these organizations is
00:25:35.079 --> 00:25:38.519
gonna touch every tool. The tools hop each other
00:25:38.519 --> 00:25:40.579
so quickly right now it's more about like the
00:25:40.579 --> 00:25:42.759
what's it doing to change our process right and
00:25:42.759 --> 00:25:45.500
what's it like and almost like i wanna swap tools
00:25:45.500 --> 00:25:47.519
on people because i want them to decouple from
00:25:47.519 --> 00:25:50.000
the technology and think more about how we're
00:25:50.000 -->
00:25:53.079
changing the directive moving forward for the
00:25:53.079 --> 00:25:55.960
org. The other the other thing i would say or
00:25:55.960 --> 00:25:59.579
two is i as much as we complain about the delta
00:25:59.579 --> 00:26:02.339
between the speed of the technology and adoption
00:26:02.339 --> 00:26:05.990
curve. We have to be mindful that it takes people.
00:26:06.250 --> 00:26:10.950
prime and exposure to embody change. And so like,
00:26:11.049 --> 00:26:13.289
there's something okay about the idea that we
00:26:13.289 --> 00:26:16.259
start by like, Looking at your current processes
00:26:16.259 --> 00:26:18.480
and amplifying your current processes and helping
00:26:18.480 --> 00:26:20.380
you like learn how that thing works and then
00:26:20.380 --> 00:26:22.700
what it breaks you go attack a different piece
00:26:22.700 --> 00:26:25.200
and then you get to the like logical conclusions
00:26:25.200 --> 00:26:27.460
that you want to orchestrate digital personas
00:26:27.460 --> 00:26:29.279
that represent different roles and that allows
00:26:29.279 --> 00:26:31.940
you to do new work right like. So the humans
00:26:31.940 --> 00:26:34.420
not just totally removing themselves from this
00:26:34.420 --> 00:26:37.440
process on my one fall swoop. And eventually
00:26:37.440 --> 00:26:39.619
the trust gets built in such a way that we can
00:26:39.619 --> 00:26:41.460
allow it to do more and more of the things we
00:26:41.460 --> 00:26:44.140
feel good about delegating work to it as an organization.
00:26:44.660 --> 00:26:46.460
That is where we ultimately want to go, but it
00:26:46.460 --> 00:26:49.440
is very hard. If that's what you're not stepping
00:26:49.440 --> 00:26:51.960
into, but like you've had 20 years of working
00:26:51.960 --> 00:26:54.720
in a different way, like you got to give people
00:26:54.720 --> 00:26:56.839
a little bit of time to like work into that and
00:26:56.839 --> 00:26:59.640
the organization time to consume it. So I would
00:26:59.640 --> 00:27:01.500
just say like a lot of the talk track right now
00:27:01.500 --> 00:27:04.390
is. Do we really aligned on the vision like do
00:27:04.390 --> 00:27:06.509
we agree that the end state is the end state
00:27:06.509 --> 00:27:08.509
like where you're trying to go as an organization?
00:27:08.910 --> 00:27:11.930
And then how do we define the right program exposure
00:27:11.930 --> 00:27:15.230
enablement right tool varieties for what you're
00:27:15.230 --> 00:27:18.190
trying to accomplish and I also think like Craig
00:27:18.190 --> 00:27:22.029
you're spot -on like I think everything in AI
00:27:22.029 --> 00:27:24.549
should be tied to like what is something valuable
00:27:24.549 --> 00:27:28.069
enough to begin the financing process and the
00:27:28.069 --> 00:27:31.089
human effort process to accomplish it and then
00:27:31.089 --> 00:27:34.430
all of these motions whether it's data readiness
00:27:34.430 --> 00:27:37.930
developer pipeline structure like. Even security
00:27:37.930 --> 00:27:40.289
programs have to be predicated on how do they
00:27:40.289 --> 00:27:42.549
contribute right to the to the growth extension
00:27:42.549 --> 00:27:46.329
and continued value for that that effort and
00:27:46.329 --> 00:27:50.109
I think is like unsexy as it may be. Like technical
00:27:50.109 --> 00:27:53.190
debt represents a much larger amount of operating
00:27:53.190 --> 00:27:55.470
costs in a business than software development
00:27:55.470 --> 00:27:58.910
traditionally has. And as a consequence of that
00:27:58.910 --> 00:28:02.779
like it may not be fun but. It's hugely valuable
00:28:02.779 --> 00:28:05.559
to organizations and i think if we're being candid
00:28:05.559 --> 00:28:09.250
and like. Authentic about where head plays really
00:28:09.250 --> 00:28:12.250
we spend eighty percent of our time and regulated
00:28:12.250 --> 00:28:15.670
complex enterprises like the technical that is
00:28:15.670 --> 00:28:20.170
profuse in that space it's everywhere and i think
00:28:20.170 --> 00:28:23.160
you know. Our first points around verifiability
00:28:23.160 --> 00:28:26.019
and logical kind of growth coupled with the like
00:28:26.019 --> 00:28:28.059
this huge nut that's sitting out there that's
00:28:28.059 --> 00:28:29.740
saying like we've got all this work to do and
00:28:29.740 --> 00:28:32.380
it's never been economically viable to do it.
00:28:32.839 --> 00:28:35.920
And now this thing that we can pump out at a
00:28:35.920 --> 00:28:38.579
high velocity creates economic viability to the
00:28:38.579 --> 00:28:41.900
retirement of those things. It is like it's where
00:28:41.900 --> 00:28:44.200
I point a lot of our customers in the direction
00:28:44.200 --> 00:28:47.240
of. Because I'd rather work on something that
00:28:47.240 --> 00:28:49.559
immediately pays back value and shows people
00:28:49.559 --> 00:28:51.779
that it bought back time and budget and effort
00:28:51.779 --> 00:28:54.119
for the next thing than go in the other direction,
00:28:54.200 --> 00:28:55.980
which is let's go create new technical debt.
00:28:56.099 --> 00:28:57.920
Just do it fast than you've previously done it
00:28:57.920 --> 00:29:00.680
before. Yeah, I think the equations changed there
00:29:00.680 --> 00:29:03.779
in many ways. Like a couple of years ago, the
00:29:03.779 --> 00:29:06.740
idea of getting to a common platform and a common
00:29:06.740 --> 00:29:10.460
set of architectures and technologies and patterns
00:29:10.460 --> 00:29:14.880
and was desirable. But the idea of, okay, I'm
00:29:14.880 --> 00:29:17.480
going to spend X amount of time, months, maybe
00:29:17.480 --> 00:29:20.759
even years, going rebuilding yesterday's applications.
00:29:20.940 --> 00:29:23.220
And the answer was, hey, we'll bring humans to
00:29:23.220 --> 00:29:25.279
come do this. While you're fighting for your
00:29:25.279 --> 00:29:27.140
life out in the market space, what we need to
00:29:27.140 --> 00:29:30.480
do is spend more time and energy over here. is
00:29:30.480 --> 00:29:34.400
not a good business decision. But now with agentic
00:29:34.400 --> 00:29:40.640
workforces and automation we can drive, the equation's
00:29:40.640 --> 00:29:43.220
fundamentally changed. Like we can really do
00:29:43.220 --> 00:29:46.480
that with a lot less human capital than ever
00:29:46.480 --> 00:29:52.119
before. And so that to me is an exciting use
00:29:52.119 --> 00:29:56.039
case that I'm seeing many customers recognizing
00:29:56.039 --> 00:30:00.309
that I can get to that. nirvana of a standardized
00:30:00.309 --> 00:30:03.549
platform with heavily opinionated patterns that
00:30:03.549 --> 00:30:07.390
my entire organization follow. And that's an
00:30:07.390 --> 00:30:11.549
exciting time for me. I feel I put on my architect
00:30:11.549 --> 00:30:16.269
hat for a second. I also think the best practices,
00:30:16.470 --> 00:30:20.769
it's really important for organizations to Once
00:30:20.769 --> 00:30:22.890
again, not talk around the tech, but talk around
00:30:22.890 --> 00:30:25.450
the capability that the platform provides. Talk
00:30:25.450 --> 00:30:28.710
around the capability that the business is providing,
00:30:29.430 --> 00:30:32.049
and then start to organize around that. Whether
00:30:32.049 --> 00:30:34.630
it's capability of automated deployments, the
00:30:34.630 --> 00:30:39.349
capability of observability, the capability of
00:30:39.349 --> 00:30:44.869
user story writing in the SDLC, and how we build
00:30:44.869 --> 00:30:48.000
those and have... Owners that own those capabilities
00:30:48.000 --> 00:30:51.079
and then they all get to evolve as the tech evolves
00:30:51.079 --> 00:30:54.480
But we're the way we communicate it to the business
00:30:54.480 --> 00:30:57.299
and the consumers is around the capability that
00:30:57.299 --> 00:30:59.720
it's providing and and I think architecturally
00:30:59.720 --> 00:31:02.559
that's a very healthy mindset to have because
00:31:02.559 --> 00:31:05.619
the underpinning tech is going to change and
00:31:05.619 --> 00:31:08.579
But but how we've organized and grouped and think
00:31:08.579 --> 00:31:11.599
about it that that To me, hasn't fundamentally
00:31:11.599 --> 00:31:15.460
changed in the 15 years. The rate of the pace
00:31:15.460 --> 00:31:18.599
of the tech underneath it is certainly evolving,
00:31:18.880 --> 00:31:24.319
but I'm not seeing the SDLC fundamentally be
00:31:24.319 --> 00:31:27.119
different in terms of gates and how we do it.
00:31:27.359 --> 00:31:30.599
Just what we're automating and what's owned by
00:31:30.599 --> 00:31:33.759
humans and then the underlying platform, the
00:31:33.759 --> 00:31:36.079
same thing. Like we still have. We still need
00:31:36.079 --> 00:31:38.440
observability, still need SRE. We still need
00:31:38.440 --> 00:31:41.799
a control plane. We need a delivery plane. Just
00:31:41.799 --> 00:31:45.720
the
tooling underneath it is fundamentally changing
00:31:45.720 --> 00:31:49.190
at a rapid pace. And there's irony, I think.
00:31:49.309 --> 00:31:52.829
You both talked about the need to always align
00:31:52.829 --> 00:31:57.009
the technical work with either increased revenue,
00:31:57.269 --> 00:32:01.329
improved efficiency, regulatory, and risk management.
00:32:03.769 --> 00:32:06.210
new technology comes out, sometimes there's pressure
00:32:06.210 --> 00:32:09.549
on us. We want you to use AI. And what you just
00:32:09.549 --> 00:32:12.029
described is the irony, which is we're reminding
00:32:12.029 --> 00:32:13.950
our partners that this still has to be about
00:32:13.950 --> 00:32:17.589
outcomes. I'm not building AI. I'm building a
00:32:17.589 --> 00:32:19.589
solution that satisfies a problem that's meaningful
00:32:19.589 --> 00:32:21.829
to the business. And if we're fortunate enough
00:32:21.829 --> 00:32:24.930
that AI makes that better, let's do that. But
00:32:24.930 --> 00:32:28.730
it's not the reverse. Let's use AI as the magnifying
00:32:28.730 --> 00:32:31.480
glass to find the problem. We know what the problems
00:32:31.480 --> 00:32:33.779
are. We live in the enterprise. So how do we
00:32:33.779 --> 00:32:37.019
reverse that? And I expect in the next 18 months,
00:32:37.099 --> 00:32:39.940
we'll start to hear the C -suite who in the last
00:32:39.940 --> 00:32:42.400
two or three years have been like, are you using
00:32:42.400 --> 00:32:45.799
AI? Are you adopting AI? How do you use AI to
00:32:45.799 --> 00:32:49.279
saying it the way you've both just said it? How
00:32:49.279 --> 00:32:52.339
do we ensure that we're using AI to affect the
00:32:52.339 --> 00:32:53.900
outcomes that are important to our business?
00:32:54.460 --> 00:32:56.720
What do you think is changing in the next six,
00:32:56.740 --> 00:33:00.829
12 and 18 months? We've seen. good, strong adoption
00:33:00.829 --> 00:33:03.230
of Accelerated DevTools. You guys talked about
00:33:03.230 --> 00:33:06.390
it as it relates to individuals. You talked about
00:33:06.390 --> 00:33:10.680
it as it relates to teams of agents. What do
00:33:10.680 --> 00:33:12.940
you think happens in the next six, 12, 18 months
00:33:12.940 --> 00:33:15.720
where people start to recognize the value of
00:33:15.720 --> 00:33:17.619
the thing we're talking about? I'll give you
00:33:17.619 --> 00:33:21.799
my thoughts. And I think Josh and I will maybe
00:33:21.799 --> 00:33:23.859
have different opinions on this, but equally
00:33:23.859 --> 00:33:27.299
valuable. I'll let you guys be the judge. But
00:33:27.299 --> 00:33:35.180
I think The shift for me is really around context
00:33:35.180 --> 00:33:38.339
engineering. And if I am truly going to have
00:33:38.339 --> 00:33:43.420
a digital workforce and I've got hundreds if
00:33:43.420 --> 00:33:47.079
not thousands of agents out there, how do I give
00:33:47.079 --> 00:33:50.430
each of them the right context and knowledge
00:33:50.430 --> 00:33:53.009
at the right time, given the task that it's trying
00:33:53.009 --> 00:33:56.650
to do. And I think we have ample amounts of data.
00:33:56.750 --> 00:34:00.109
And I see this as a signal versus noise problem,
00:34:00.450 --> 00:34:03.170
where noise is data. We've got data lake houses,
00:34:03.329 --> 00:34:06.130
we've got Italian architectures, we've got events
00:34:06.130 --> 00:34:08.449
flowing through event buses all over the place,
00:34:08.550 --> 00:34:11.289
we've got customers coming in and doing stuff.
00:34:11.889 --> 00:34:16.019
That's a lot of useful... useful data, but there's
00:34:16.019 --> 00:34:18.900
a signal, what is truly a signal I wanna go off
00:34:18.900 --> 00:34:22.380
of and how do I take all that data, turn it into
00:34:22.380 --> 00:34:25.159
knowledge and start to organize that. And we
00:34:25.159 --> 00:34:28.539
end up going down knowledge graphs and storing
00:34:28.539 --> 00:34:31.480
that information, but making sure that if I have
00:34:31.480 --> 00:34:36.260
someone who's gonna, a digital agent who's gonna...
00:34:36.119 --> 00:34:38.920
build something for this business unit, then
00:34:38.920 --> 00:34:42.500
what does that mean? What does good look like?
00:34:42.679 --> 00:34:45.460
What regulations do they adhere to or not adhere
00:34:45.460 --> 00:34:48.619
to? And what were the last things that were deployed
00:34:48.619 --> 00:34:51.300
and went well and didn't go well? And how do
00:34:51.300 --> 00:34:54.239
I, as I build this, do I build it in more resilient
00:34:54.239 --> 00:34:58.019
ways? And so taking all of this knowledge or
00:34:58.019 --> 00:35:00.139
all this data, excuse me, and turning it into
00:35:00.139 --> 00:35:04.440
meaningful knowledge, for my agents to leverage
00:35:04.440 --> 00:35:08.059
and consume will drive down token costs, will
00:35:08.059 --> 00:35:12.519
drive up the speed at which I'm getting value
00:35:12.519 --> 00:35:14.679
from it. Cause I can write agents that continue
00:35:14.679 --> 00:35:16.139
to iterate through and through and through and
00:35:16.139 --> 00:35:18.460
takes forever. It's expensive. It'll eventually
00:35:18.460 --> 00:35:21.300
get there. But if I needed seven humans in the
00:35:21.300 --> 00:35:24.659
loop and it had to iterate through 57 different
00:35:24.659 --> 00:35:28.440
times, that costs money and time. And so I think
00:35:28.440 --> 00:35:32.730
having, I think The next year is going to be
00:35:32.730 --> 00:35:37.150
a lot about organizing my knowledge in meaningful
00:35:37.150 --> 00:35:42.550
ways so that I can get more value from all of
00:35:42.550 --> 00:35:45.449
our agents. And we talk about it as context engineering
00:35:45.449 --> 00:35:47.570
and giving the right context at the right time
00:35:47.570 --> 00:35:50.190
and not too much information, not too little.
00:35:51.210 --> 00:35:55.630
And so I think that's a big part of what the
00:35:55.630 -->
00:35:57.889
future is going to be. Coupled to that, I don't
00:35:57.889 --> 00:36:01.119
want to just throw. buzzwords, but you can't
00:36:01.119 --> 00:36:05.739
do proper context engineering without strong
00:36:05.739 --> 00:36:08.559
tokenomics, knowing where my value is going and
00:36:08.559 --> 00:36:12.300
where my spend is. And then the data and observability
00:36:12.300 --> 00:36:15.820
of how did it run the last time? And what can
00:36:15.820 --> 00:36:18.699
I do to improve itself? So that short -term memory
00:36:18.699 --> 00:36:22.400
versus that long -term memory. And so to me,
00:36:22.800 --> 00:36:25.360
that space, I'm seeing many organizations well
00:36:25.360 --> 00:36:28.659
beyond the, okay, this tool can do this. I've
00:36:28.659 --> 00:36:31.420
orchestrated a whole bunch of workflows to, okay,
00:36:31.519 --> 00:36:34.679
now I've got all this data, turn it into meaningful
00:36:34.679 --> 00:36:39.480
knowledge and make it useful. That to me is,
00:36:39.480 --> 00:36:42.400
I think, a very exciting space and one I'm pushing
00:36:42.400 --> 00:36:46.210
very hard into. I think I would just say like,
00:36:46.309 --> 00:36:49.940
one, 18 months at the current rate of change
00:36:49.940 --> 00:36:52.639
is equivalent to like five years, a few years
00:36:52.639 --> 00:36:55.920
ago. So like, it's super hard to project what's
00:36:55.920 --> 00:36:59.789
like. possible when we have kind of Exponential
00:36:59.789 --> 00:37:02.570
forces colliding at the same time within a space
00:37:02.570 --> 00:37:06.550
I think what Craig is describing is very to me
00:37:06.550 --> 00:37:09.909
is almost like the natural flow has been for
00:37:09.909 --> 00:37:13.920
a long time intact, which is eventually it outpaces
00:37:13.920 --> 00:37:16.039
our ability to be efficient with it and we have
00:37:16.039 --> 00:37:18.099
to like go back and look at like all the stuff
00:37:18.099 --> 00:37:19.719
that was hard that we didn't do because the tech
00:37:19.719 --> 00:37:21.400
made everything look easy like that's what's
00:37:21.400 --> 00:37:24.230
happening again and context engineering is years
00:37:24.230 --> 00:37:27.130
and years of lacking of data stewardship and
00:37:27.130 --> 00:37:30.409
governance and appropriate curation of information.
00:37:31.289 --> 00:37:34.389
The irony is how many decades have we talked
00:37:34.389 --> 00:37:36.550
about the concept of tribal knowledge across
00:37:36.550 --> 00:37:38.849
an organization? And what we're talking about
00:37:38.849 --> 00:37:43.349
context engineering is the overlaying of process
00:37:43.349 --> 00:37:47.230
understanding and tribal knowledge and what drives
00:37:47.230 --> 00:37:50.349
high efficacy through work. Instantiated on top
00:37:50.349 --> 00:37:53.210
of massive isolated volumes of data that now
00:37:53.210 --> 00:37:56.090
have to be. Not isolated in the way that they
00:37:56.090 --> 00:37:59.949
work right and and in textualize that in such
00:37:59.949 --> 00:38:02.349
a way that a new employee right coming on board
00:38:02.349 --> 00:38:04.489
can use it very very quickly because that's what
00:38:04.489 --> 00:38:07.769
an agent ultimately is in this space. And I think
00:38:07.769 --> 00:38:10.570
you're by the way he's 100 % right in that you
00:38:10.570 --> 00:38:12.949
know all of our kind of more mature clients and
00:38:12.949 --> 00:38:15.880
many of the. Code generation firms in general
00:38:15.880 --> 00:38:17.980
are pointing to context compression being the
00:38:17.980 --> 00:38:20.760
number one indicator of of tokenomic kind of
00:38:20.760 --> 00:38:22.840
going out of control within the space, right?
00:38:23.280 --> 00:38:25.059
We don't know what to give something so we feed
00:38:25.059 --> 00:38:26.719
it a lot more than it needs to accomplish the
00:38:26.719 --> 00:38:29.400
same task. And it's increasingly a problem because
00:38:29.400 --> 00:38:32.039
as the models are improving and they will materially
00:38:32.039 --> 00:38:34.539
like stepwise improve over the next 6 to 12 months.
00:38:35.219 --> 00:38:39.179
How good we need to be at getting them the generalized
00:38:39.179 --> 00:38:41.820
data is decreasing. So we actually need to provide
00:38:41.820 --> 00:38:44.079
them less generalized data and more specific
00:38:44.079 --> 00:38:46.360
data because they already know how to leverage
00:38:46.360 --> 00:38:48.400
the general knowledge that we were filling in
00:38:48.400 --> 00:38:52.139
the gaps and the context on. So I think context
00:38:52.139 --> 00:38:55.840
compression and tokenomics, very big threads
00:38:55.840 --> 00:38:58.800
for the enterprise, probably for the foreseeable
00:38:58.800 --> 00:39:02.880
future, by the way. Unless they become non -items
00:39:02.880 --> 00:39:05.280
in general, they're going to be a challenge.
00:39:06.559 --> 00:39:10.400
I think the other piece here is trust. I think
00:39:10.400 --> 00:39:12.420
there's gonna be a huge focus on how do we generate
00:39:12.420 --> 00:39:15.000
trust. And I don't mean that as like security,
00:39:15.179 --> 00:39:17.900
by the way, that's not a euphemism for like,
00:39:18.219 --> 00:39:20.500
we ran checks and we did security analysis or
00:39:20.500 --> 00:39:23.179
something. I mean, like we actually build a level
00:39:23.179 --> 00:39:27.360
of like, comfort and confidence in our ability
00:39:27.360 --> 00:39:31.480
to delegate complex long horizon work to these
00:39:31.480 --> 00:39:36.860
abstract digital concepts in such a way that.
00:39:37.130 --> 00:39:39.530
What today feels like we're breaking things into
00:39:39.530 --> 00:39:43.570
very very micro segmented box like very bound
00:39:43.570 --> 00:39:46.670
boxed ideas with almost no agency by the way
00:39:46.670 --> 00:39:48.469
like that's the irony of the whole space right
00:39:48.469 --> 00:39:51.250
now is. It's worth treating it like it's the
00:39:51.250 --> 00:39:54.170
antithesis of of what we call the actual tool
00:39:54.170 --> 00:39:57.469
there's no self determination but as we get more
00:39:57.469 --> 00:39:59.590
trust and more confidence in the models continue
00:39:59.590 --> 00:40:02.940
to improve like word. we're going to have a level
00:40:02.940 --> 00:40:04.599
of self -determination through some of these
00:40:04.599 --> 00:40:06.159
tools. And I think that's going to be light bulb
00:40:06.159 --> 00:40:08.280
moments in a lot of organizations. I think you
00:40:08.280 --> 00:40:11.260
already see this improving. Like what we trusted
00:40:11.260 --> 00:40:15.059
to do is already increasing. And the reality
00:40:15.059 --> 00:40:16.860
is like 12 months from now, I think it's going
00:40:16.860 --> 00:40:19.699
to look pretty unfamiliar, like unrecognizable
00:40:19.699 --> 00:40:23.400
to 12 months ago where we were. And what we tell
00:40:23.400 --> 00:40:25.360
most organizations is like, don't focus on the
00:40:25.360 --> 00:40:26.780
model layer. Like you really aren't going to
00:40:26.780 --> 00:40:28.760
participate that broadly at the model layer at
00:40:28.760 --> 00:40:32.710
this point. But I think the structure you build
00:40:32.710 --> 00:40:34.809
around it, like this concept of how do we think
00:40:34.809 --> 00:40:37.269
about a harness and what's unique to us? Like,
00:40:37.510 --> 00:40:38.889
I don't think everyone's going to go build a
00:40:38.889 --> 00:40:40.429
harness, but I do think people are going to go
00:40:40.429 --> 00:40:43.690
customize harnesses that they can consume because
00:40:43.690 --> 00:40:45.530
that's where we're going to instantiate a lot
00:40:45.530 --> 00:40:47.630
of the technical context for the organization
00:40:47.630 --> 00:40:50.329
beyond just the data level context components.
00:40:51.789 --> 00:40:54.489
And I think we're going to get more and more
00:40:54.489 --> 00:40:57.369
kind of bent around this idea that the intelligence
00:40:57.369 --> 00:41:00.030
needs to be a composable component and that the
00:41:00.030 --> 00:41:02.690
real Organizational intelligence exists in the
00:41:02.690 --> 00:41:04.690
context layer, exists in the harness layer, et
00:41:04.690 --> 00:41:08.730
cetera. And you are just renting the technical
00:41:08.730 --> 00:41:11.530
intelligence from the model layer. And we create
00:41:11.530 --> 00:41:14.289
composability behind it. I think that's reinforced,
00:41:14.309 --> 00:41:18.269
by the way, through Stripe acquiring OpenRouter,
00:41:18.449 --> 00:41:21.590
through Ramp, another fintech running a new router
00:41:21.590 --> 00:41:26.130
service. It's rapidly becoming the topic du jour
00:41:26.130 --> 00:41:29.690
in this space. And as the proliferation of agents
00:41:29.690 --> 00:41:32.349
go, so will the need to understand that align
00:41:32.349 --> 00:41:35.070
the right types and size of models to the type
00:41:35.070 --> 00:41:37.389
of work that we're doing. And that will also
00:41:37.389 --> 00:41:39.829
play a role in the tokenomics discussion moving
00:41:39.829 --> 00:41:43.119
forward. And this is a really great discussion.
00:41:43.320 --> 00:41:48.099
I appreciate both of you. You are daily learning,
00:41:48.619 --> 00:41:52.239
delivering, fighting these opportunities. Um,
00:41:52.239 --> 00:41:54.619
I want to close with a kind of a softball, which
00:41:54.619 --> 00:42:00.199
is for each of you. Regardless of SDLC and what
00:42:00.199 --> 00:42:02.579
we've been talking about, what is the most exciting
00:42:02.579 --> 00:42:05.900
thing right now you're seeing in emerging technology?
00:42:06.019 --> 00:42:08.760
What is the thing that you're following and tracking
00:42:08.760 --> 00:42:11.909
that you really get excited about? I'm going
00:42:11.909 --> 00:42:14.349
to try to take a swing at this one and not get
00:42:14.349 --> 00:42:19.829
too excited or esoteric. I would say it's two
00:42:19.829 --> 00:42:22.630
things going at the same time, but that massively
00:42:22.630 --> 00:42:26.230
impact one another. One is I think we're starting
00:42:26.230 --> 00:42:30.079
to See signs that we're making the transition
00:42:30.079 --> 00:42:34.219
from purely training to actual learning and and
00:42:34.219 --> 00:42:37.599
obviously that's coming by way of CPT and things
00:42:37.599 --> 00:42:39.579
like Reinforcement learning through verifiable
00:42:39.579 --> 00:42:42.960
rewards, but I'm very optimistic about the level
00:42:42.960 --> 00:42:44.920
and pace of self -improvement And I'm not going
00:42:44.920 --> 00:42:47.139
to go as far as to say recursive self -improvements
00:42:47.139 --> 00:42:49.539
like tomorrow But I do think like we're starting
00:42:49.539 --> 00:42:52.679
to see a time compression as a consequence of
00:42:52.679 --> 00:42:56.000
very rapid domain skill set expansion understanding
00:42:56.469 --> 00:42:58.849
And I think why that's really important or what
00:42:58.849 --> 00:43:01.409
I'm following the most, it's now how AI is starting
00:43:01.409 --> 00:43:05.550
to contribute to other technology areas. So like
00:43:05.550 --> 00:43:08.449
not AI for the sake of AI, but what's AI's impact
00:43:08.449 --> 00:43:11.050
on the pace of improvement in quantum computing
00:43:11.050 --> 00:43:14.070
research? What's AI's impact in material sciences?
00:43:14.590 --> 00:43:16.429
By the way, what's going on in mathematics right
00:43:16.429 --> 00:43:18.210
now, if you're not watching along at home, it's
00:43:18.210 --> 00:43:22.030
absolutely absurd, right? So like, and that may
00:43:22.030 --> 00:43:24.210
seem very elitist in that we don't know what
00:43:24.210 --> 00:43:27.170
those proofs are really for. But they have widespread
00:43:27.170 --> 00:43:29.829
implications on physics and our understanding
00:43:29.829 --> 00:43:31.929
of the world around us and our ability to do
00:43:31.929 --> 00:43:35.090
chemistry, our ability to build things, our ability
00:43:35.090 --> 00:43:39.269
to test things. So I just think we're rapidly
00:43:39.269 --> 00:43:42.289
hitting that point where it's not about you,
00:43:42.489 --> 00:43:44.690
your daily life being easier or an enterprise
00:43:44.690 --> 00:43:47.349
building a digital product faster. It's like,
00:43:47.650 --> 00:43:50.030
how are we about to do science and research at
00:43:50.030 --> 00:43:52.929
a pace that we've never done before in as a society,
00:43:53.110 --> 00:43:57.340
as a race? And if that's not exciting, I don't
00:43:57.340 --> 00:44:00.619
know what is, to be very honest. It feels like
00:44:00.619 --> 00:44:02.159
we're going to hit an inflection point, and it's
00:44:02.159 --> 00:44:05.539
not pure Silicon Valley hype. The proof is starting
00:44:05.539 --> 00:44:08.659
to actually emerge, not just fun articles and
00:44:08.659 --> 00:44:10.840
blogs of people trying to project into the future.
00:44:11.909 --> 00:44:14.250
Yeah, I'll add a little to that. So twofold.
00:44:14.510 --> 00:44:18.650
One is I do find it unbelievably exciting that
00:44:18.650 --> 00:44:21.119
the The, my customers that I'm working with who
00:44:21.119 --> 00:44:24.840
are really are embracing this and it's having
00:44:24.840 --> 00:44:27.679
material results, like very material results.
00:44:27.760 --> 00:44:30.340
And I, and I, I love that. I love seeing it.
00:44:30.480 --> 00:44:33.659
I love fighting their fights with them and bringing
00:44:33.659 --> 00:44:36.260
teams of people to help fight those fights and
00:44:36.260 --> 00:44:38.360
not just for the sake of fighting, but actually
00:44:38.360 --> 00:44:42.960
winning. And that, that to me is a really exciting
00:44:42.960 --> 00:44:47.869
to, to, to feel. a part of that and be a part
00:44:47.869 --> 00:44:51.590
of that and be a huge catalyst to how they're
00:44:51.590 --> 00:44:54.489
performing in the marketplace. I think, and I'll
00:44:54.489 --> 00:44:59.789
put my architect hat back on too, I love that,
00:44:59.949 --> 00:45:02.550
I don't want to say before AI, five, six years
00:45:02.550 --> 00:45:05.710
ago, it was getting a little stale, but I'll
00:45:05.710 --> 00:45:08.349
say that where like, Okay, we were talking about
00:45:08.349 --> 00:45:11.469
the same cloud native architectural patterns
00:45:11.469 --> 00:45:17.690
and let's build some microservices and I love
00:45:17.690 --> 00:45:20.940
those and love talking about orchestration and
00:45:20.940 --> 00:45:23.219
choreography and saga patterns and things like
00:45:23.219 --> 00:45:27.000
that. But I'm so excited personally to see all
00:45:27.000 --> 00:45:29.380
those architectural patterns get rechallenged.
00:45:29.880 --> 00:45:32.619
And do they even make sense anymore? Do we build
00:45:32.619 --> 00:45:35.320
them for humans so that humans could maintain
00:45:35.320 --> 00:45:37.659
architectures and code bases? And do we even
00:45:37.659 --> 00:45:39.900
need those anymore? And how do we leverage those?
00:45:40.699 --> 00:45:45.579
And so it feels like, for me, I'll maybe share
00:45:45.579 --> 00:45:47.840
my age a little bit here, but You know, when
00:45:47.840 --> 00:45:50.349
I started coming to life in tech. It's three
00:45:50.349 --> 00:45:52.869
tier architectures and, you know, and, and, you
00:45:52.869 --> 00:45:56.309
know, monoliths and, uh, and, uh, you know, large
00:45:56.309 --> 00:45:59.449
class structures and, and, and, uh, soap and
00:45:59.449 --> 00:46:02.449
some basic rest web services. And then microservices
00:46:02.449 --> 00:46:05.550
clink came out and with cloud. And we said, everything
00:46:05.550 --> 00:46:08.210
you learned there, forget that you can denormalize
00:46:08.210 --> 00:46:10.610
your data, break into microservices. It's totally
00:46:10.610 --> 00:46:13.469
different. And, uh, and, and so we spent years
00:46:13.469 --> 00:46:15.550
like rethinking the architecture patterns around
00:46:15.550 --> 00:46:18.329
that and how to, how to do circuit breaker patterns
00:46:18.329 --> 00:46:21.719
and. and orchestration at scale and self -healing.
00:46:22.239 --> 00:46:25.000
And now in the last three to four years, everything
00:46:25.000 --> 00:46:27.300
you learned there, we might wanna forget all
00:46:27.300 --> 00:46:30.360
of that. And maybe we need to bring them back
00:46:30.360 --> 00:46:32.300
in and maybe we need a knowledge graph layer.
00:46:32.760 --> 00:46:35.519
And when do we use, you know, PEV flow versus
00:46:35.519 --> 00:46:39.760
react versus reflection. And all those architectures
00:46:39.760 --> 00:46:43.659
are being re -challenged. And as the total tech
00:46:43.659 --> 00:46:46.719
software nerd inside of me, I love that, right?
00:46:46.780 --> 00:46:49.900
Like I really love that. Everything I got, I
00:46:49.900 --> 00:46:52.340
spent so much time learning. I get to rethink
00:46:52.340 --> 00:46:55.920
it and relearn it. And that's maybe why I love
00:46:55.920 --> 00:46:57.780
this industry and I'm kind of romantic about
00:46:57.780 --> 00:47:02.519
it, but like it just never ends. And so that
00:47:02.519 --> 00:47:05.880
to me is personally very exciting. Well, thank
00:47:05.880 --> 00:47:08.920
you again. The two of you really have amazing
00:47:08.920 --> 00:47:11.659
experience. I appreciate you participating in
00:47:11.659 --> 00:47:15.760
this debut episode of the Ahead podcast, Ready,
00:47:15.840 --> 00:47:19.800
Set, Runtime. Our next episode, we will continue
00:47:19.800 --> 00:47:24.179
the discussion, talking to partners about practitioner
00:47:24.179 --> 00:47:28.659
examples on accelerated dev. Josh, Craig, thank
00:47:28.659 --> 00:47:30.579
you so much for your time today. I really appreciate
00:47:30.579 --> 00:47:34.840
it. Thank you for listening to Ready, Set, Runtime.
00:47:35.059 --> 00:47:37.840
If something in today's conversation sparked
00:47:37.840 --> 00:47:40.619
an idea or a question about your own environment,
00:47:41.199 --> 00:47:43.619
a HEADS team would love to hear from you. Reach
00:47:43.619 --> 00:47:46.820
out anytime at ahead .com and if you haven't
00:47:46.820 --> 00:47:49.519
already, follow the show wherever you're tuning
00:47:49.519 --> 00:47:52.500
in. The views shared on this podcast are those
00:47:52.500 --> 00:47:54.780
of the speakers and don't necessarily reflect
00:47:54.780 --> 00:47:57.099
those of a head. Until next time.