December 26, 2024 • 48 mins
Ever wondered how cloud development intersects with network engineering at a tech powerhouse like Cisco? Join us as we chat with Erica Dietrick, a developer advocate at Cisco, who shares her unique perspective on the challenges and rewards of her role. From gearing up for major events such as Cisco Live to engaging with the tech community online, Erica offers an insider's view into the cyclical nature of her work. She also takes us through her personal journey into cloud development, highlighting the continuous learning required in such a fast-paced industry.
As we step further into the world of tech careers, we explore the complex interplay between cloud and DevOps. Whether you're in it for the salary or driven by a genuine passion for technology, the path is fraught with the challenge of understanding a rapidly expanding tech stack. We discuss the increasing accessibility of high-demand roles like cloud engineering and the contrasting motivations of those entering the field. Erica's insights shed light on the overwhelming breadth of knowledge needed today, especially when navigating the unique dynamics of startups, where junior developers can significantly shape the architecture from the ground up.
The conversation takes an innovative turn as we discuss the transformative role of AI in coding education and assessment. With AI tools like GitHub Copilot changing the landscape, we delve into the emerging skill of prompt engineering and the shift towards evaluating problem-solving abilities over traditional coding exercises. Erica shares her thoughts on how AI can enhance learning and streamline complex tasks without undermining foundational problem-solving skills. We reflect on the importance of embracing these tools to boost productivity and deepen our understanding of the technical landscape, making a compelling case for AI as an assistant rather than a replacement.
How to connect with Erika:
https://www.linkedin.com/in/erikadietrick/
https://x.com/Erika_theDev
https://www.youtube.com/@erika_thedev
Purchase Chris and Tim's new book on AWS Cloud Networking: https://www.amazon.com/Certified-Advanced-Networking-Certification-certification/dp/1835080839/
Check out the Fortnightly Cloud Networking News
https://docs.google.com/document/d/1fkBWCGwXDUX9OfZ9_MvSVup8tJJzJeqrauaE6VPT2b0/
Visit our website and subscribe: https://www.cables2clouds.com/
Follow us on BlueSky: https://bsky.app/profile/cables2clouds.com
Follow us on YouTube: https://www.youtube.com/@cables2clouds/
Follow us on TikTok: https://www.tiktok.com/@cables2clouds
Merch Store: https://store.cables2clouds.com/
Join the Discord Study group: https://artofneteng.com/iaatj
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Chris Miles (00:00):
Hey everyone.
It's Chris from Cables toClouds here.
I hope everyone is enjoying anice holiday break, spending
time with loved ones, family,friends, all of the above.
I think this episode's actuallycoming out on Christmas, so if
you celebrate, merry Christmasand upcoming Happy New Year to
everyone as well.
Before we get to the episode, Ijust want to take a second to
let everyone know that Tim and I, along with one of our
(00:22):
critically acclaimed guests, oneof our critically acclaimed
guests, one of our mostdownloaded episodes, I think in
the past with Steve McNutt,we've all come together to
release the first edition of acertification guide for the AWS
Certified Advanced NetworkingSpecialty Exam that is being put
(00:43):
out by PACT.
So the book is actually up forpre-order now PACT.
So the book is actually up forpre-order now.
So if you're interested, ifyou're doing anything with
networking in AWS, I wouldheavily, heavily encourage you
guys to check it out, even ifyou're not preparing for the
exam.
I feel like it's just a goodoverall composition of all the
networking services within AWStoday.
So the book is officially upfor pre-order now and I believe
you get early access throughPACT if you do the pre-order.
(01:04):
So I will put the link in theshow notes and if you're
interested, I encourage you toplease check it out.
And with that out of the waywe'll get to the episode.
Tim McConnaughy (01:11):
Thanks, Welcome to the Cables to Clouds podcast
, your one-stop shop for allthings hybrid and multi-cloud
networking.
Now here are your hosts Tim.
Chris Miles (01:22):
Chris and Alex.
Tim McConnaughy (01:25):
Hello and welcome back to another episode
of the Cables to Clouds podcast.
I'm your host this week, tim atcarpe-dmvpn on Blue Sky, and
with me, as always, is myco-host, chris at the Cloud Main
on Blue Sky.
And, as you may have guessed,we have a guest with us this
(01:47):
week.
We brought Erica Dietrich hereto talk about cloud development.
I'm going to let her introduceherself in a second, but you
know, one thing I've always beenpushing for is that I want to
move across the stack.
I want cloud infrastructure andcloud network people to get
more understanding around what'sriding on the network, because
I think we need to have opinionsand be opinionated about that.
(02:09):
So, yeah, I just wanted tobring Erica on here and talk
about cloud development.
So, erica, go ahead andintroduce yourself.
Erika Dietrick (02:14):
Yeah, so yeah, my name is Erica Dietrich.
I'm a developer advocate atCisco.
I'm part of our developerprogram there and I'm basically
enabling people to learn networkautomation, to use AI tools in
their automation workflow.
And, like you mentioned, tim, alot of our community is also
infrastructure and networkengineers or cloud engineers.
(02:36):
So I get to talk to a lot ofthat community and I spend about
half of my time building things, breaking shit learning, and
half of my time, you know,building things breaking shit
learning and half of my timecommunicating.
Tim McConnaughy (02:48):
So getting presentations being on podcasts
like this, annoying people onsocial media, all that jazz yeah
, no, that's awesome, let's see,so let's, let's just, I guess,
just let's just jump right intoit.
So talk to me a little bitabout, first of all, just about
what your day-to-day looks like,but also about what you've been
(03:09):
doing with cloud development.
Erika Dietrick (03:12):
Sure, day-to-day ranges wildly.
I would say that we work inseasons.
I think that's a developeradvocate thing, not just a me
thing.
Chris Miles (03:21):
It is.
Erika Dietrick (03:24):
We work really hard towards big outputs.
So, whether that's, you know, aawesome YouTube tutorial video
at the end of the month, orwhether it's a Cisco live coming
up, that we're preparing theDevNet zone and a bunch of talks
, it's very periods of head downdelivering and then kind of a
recovery period head downdelivering and then kind of a
recovery period.
(03:44):
So I would say on a typical dayfor me, I'm usually doing my
community interaction firstthing in the morning, posting to
social media whatever it was Ilearned and cobbled together the
day before, try to be somewhatuseful and educational and
interact with the communitythere, and then it varies widely
(04:08):
.
Again, we've got Cisco Livecoming up in February, which
seems like it's really far away,but it's really not.
So usually spend a good chunkof my day trying to ignore
everyone after that and, youknow, become an expert at
whatever it is I'm talking about, because that is a big
misconception is that we're justcomplete experts at everything
we talk about.
We're like everyone else intech.
(04:29):
We have to learn and then weteach, and then the afternoon I
like leaving an ample amount oftime for shit catching on fire.
Chris Miles (04:38):
So honestly, my app is pretty flexible, as it does.
Tim McConnaughy (04:42):
As you do Okay, no, do yeah, okay, no.
That's great.
It's funny because I have manydeveloper advocate friends and
it's pretty simple across all ofthe different silos of
developer advocacy.
That's the main goal, if youwill.
Erika Dietrick (05:01):
Very different philosophies around it, for sure
.
It's just very different fromcompany to company, org to org,
and Cisco in particular, it's alittle more challenging because
we're not a traditionaldeveloper program, that's true.
Tim McConnaughy (05:13):
We're talking about DevNet, of course.
Yeah, yeah, yeah, so that's avery good point.
Actually, most of the developeradvocate types are with a
developer-focused company, likeone of the CSPs, for example
it's a lot of the DAs I know orwith, like you know, like
HashiCorp, or just one of thosedeveloper focused type of
(05:34):
organizations.
Okay, so, all right.
So let's, how did you getstarted with the cloud stuff?
I mean, obviously you probablydidn't, like, like all of us,
you didn't start, you didn'tstart with cloud.
So like, where did how did youget started?
Um to into the cloud, like fromwhere you were?
Erika Dietrick (05:50):
Right, right.
So, uh, and this is.
This is one of the things thatI thought could be insightful or
interesting or maybe it's justgoing to validate what you guys
already know about howdevelopers approach the cloud,
but um, the cloud, but um.
So I was working a lot ofdifferent gigs after my
undergraduate degree, right, um,I was most recently, before
being in tech, working atstarbucks and, uh, managed to
(06:11):
get back into a master's programof software engineering.
Uh.
So, granted, I was looking tohave a challenging and rewarding
career and my dad was asoftware engineer and for some
reason, there was some sort ofdisconnect where I never learned
what he did or why.
I just knew he liked his joband I'm like what the hell?
You know, I'm smart enough,probably.
And so I went into this softwareengineering master's degree at
(06:35):
East Carolina and I would saythat, even though I had a lot of
catching up to do, myexperience when it comes to the
cloud was probably prettytypical with every developer's
learning experience, especiallyif they're going the academic
route, and that is that you knowdevelopment.
We're very much focused in theIDE or in the developer
environment, where it's just youlearn to master a language and
(06:55):
you don't really give a.
I don't know, should I becussing less?
Tim McConnaughy (06:58):
No, that's fine , You're good, you don't?
Erika Dietrick (07:01):
really give a fuck about anything else, right,
you know.
So I think that it's a lot morenaive and siloed than coming
from an IT network engineeringbackground, where you kind of
(07:21):
have a general, well-roundedsense of how technology works
and how things communicate inthe real world.
You're just sitting there inuniversity learning variables
and methods and whatever otherprogramming concepts and being
happy that you are printing alittle line statement to the
console.
Eventually you get moreadvanced, right, depending on
your interest mobile appdevelopment, game development,
whatever.
But what's not taught inuniversity is really anything
beyond just building the code.
Tim McConnaughy (07:42):
The language itself.
Erika Dietrick (07:43):
yeah, so, in terms of how are real-world apps
deployed?
Where are they deployed?
What do you have to considerwhen deploying them?
Everything, I mean I talk aboutthis a lot with DevNet too but
everything from code security tohow you document, how you work
with other people and not juston your homework.
It's a pretty big problem in myopinion, and it really.
I was, I guess, a little bitlucky because in university I
(08:06):
ended up taking a Google CloudPlatform course and it was
because they had some kind ofstudent program that they had
partnered with ECU and they weregiving, you know, a certain
number of free credits.
A professor that I absolutelyadored had been, I think he
applied for this program and waschosen and they gave him some,
you know, PowerPoints andmaterials and he did a great job
(08:27):
with the course.
Um, but so I'm I'm lucky in thesense that I was at least
exposed to it before, you know,the real world hit me in the
face.
But, um, you know this, thisprogram?
It essentially just kind ofthrows you into a series of labs
, and the labs are hands-on,they're well done, they're meant
to get you familiar withdifferent GCP services and using
(08:49):
the CLI and all that.
But again, as a developer, youjust have no idea like why am I
learning?
this.
Why am I?
There's a huge disconnect,right.
You don't have anyunderstanding of networking or
how applications run.
So I took that course and it itwas exposure, but nothing
really stuck.
Um, what's interesting is that?
(09:09):
Um, my experience and stop mewhenever you want my experience,
both as a student and later onin startup world, was all with
Google cloud platform, becausethey really push heavy on those
empowering you to use theplatform free credits, all that
but afterwards I've nevertouched GCP.
(09:30):
It seems to be they're reallytrying to claim the market there
on the little guys and girls.
Chris Miles (09:36):
Well, we've been talking about it on this pod for
a while.
It seems like GCP is, or, as weshould say, google.
Tim McConnaughy (09:43):
Cloud.
It's not GCP anymore, butGoogle Cloud.
Chris Miles (09:47):
It feels like they win the developer battle almost
every time.
They are very attractive todevelopers day in, day out.
I don't know.
Do you have a take on why thatis?
Is there something in thatrealm that was more friendly to
you, that was inviting?
I mean, obviously there'scredits and there's incentives
to use the platform so that it's, you know, cost effective.
But, like the platform itself,what's, what's?
(10:08):
What sets it apart in your mind?
Erika Dietrick (10:10):
It's honestly, it's interesting to hear you say
that, because I okay.
Chris Miles (10:14):
So I think, it's easier to use.
I think it's easier to use anAWS.
Erika Dietrick (10:17):
I don't think there's anything particularly
special about how Google cloudis set up.
I just think there's lessservices to choose from and it's
not as confusing as AWSInteresting okay.
And, you know, in the developerworld too.
I don't know, maybe this is thesame for network engineering,
but there's basically Microsoftshops and non-Microsoft shops.
Tim McConnaughy (10:36):
Oh yeah, right sure.
Erika Dietrick (10:38):
So you know, if you're a Microsoft shop, you're
using Azure, no matter whatRight.
And if you're not, you know,and again, and you've never been
exposed to a cloud platform,you're probably like oh, AWS is
complicated, you know, jump overto Google.
Tim McConnaughy (10:51):
Yeah, what's interesting is something we
observed and I still think it'strue is that I feel like Google
Cloud's networking was builtwith developers in mind, like
the way they do their networkingwas built like you don't have
to care about it, we'll justkind of do it all for you.
Like built towards developersthat won't know or care how the
network works, we'll just kindof figure it out for you, which
(11:14):
confuses a lot of us networkpeople because it doesn't work
like you would expect networkingto work.
Erika Dietrick (11:21):
Well, and I think it's part of that
disconnect to the farther, weabstract what's happening in our
cloud services versus what'sactually happening in the data
center.
Chris Miles (11:30):
You know what?
Erika Dietrick (11:30):
I mean as a developer.
Okay, that's great that you'recatering to us and, like I work
with them, enabling them on thecode side, and you know they
want to take shortcuts or justuse low code, no code or
whatever it's like.
I know this was built forpeople who don't want to learn
(11:53):
and it's easier, but it's notreally great.
Chris Miles (11:55):
Yeah, I totally get that I'm assuming this has to
relate to the fact that there'sjust been like a boom in the
industry and there's more peoplegravitating towards this career
field, right, like I think,like you mentioned before, that
early on you know that you'dlike you didn't have to learn
some of these things.
Like you know, you just likefocus on the code, but then
(12:15):
maybe you start peeling back theonion and learning how
networking works.
I feel like that comes with thegenuine curiosity that was kind
of unique about either workingin infrastructure or working in
software development.
You probably had that genuinecuriosity that was kind of
unique about it, either workingin infrastructure or working in
software development.
You probably had that genuinecuriosity regardless and it kind
of led you down those otheravenues to explore because you
just like couldn't fathom notunderstanding how the entire
(12:36):
system works, right.
But I feel like now that thebar of entry for this stuff is
much lower and it's becoming amuch more standard career field,
like I think people just lackthat curiosity and like I've
been a big proponent about thisfor a long time is like people
need jobs and you have to letpeople be mediocre at their job
sometimes, because to somepeople this is just a fucking
(12:57):
job, right?
This is just how you put inyour 40 hours and you go home
and, like you, take care of yourfamily, and that's that's
what's important to you, right?
You don't give a shit about whatyou're doing for work really.
So, like I don't know, do youthink there's a correlation
there?
It's like just coming from theother side of the stack.
Erika Dietrick (13:11):
A correlation between sorry, can you summarize
the?
Chris Miles (13:15):
question.
Yeah, sorry, I said a lot there.
Just kind of the boost in thecareer field, right, and there's
probably more people justlooking at it from the lens of
like, hey, this is a job morethan this is like something I'm
truly deeply interested in andwant to find out more.
Erika Dietrick (13:29):
I definitely think there's two camps of I'll
just generalize as tech folks.
But you know, there's peoplewho just have a love of learning
and you know, maybe they startout in one discipline but
they're at least willing toexplore in other areas or where
it makes sense and they want tounderstand how things work.
And then I feel like the otheris just looking for that high
salary tech career, right.
(13:50):
They gravitate towards the titleand I'm going to get the
credential for that title, I'mgoing to build for that title
and then I don't care aboutanything else.
So I definitely think cloudengineering is one of those.
I'm not really sure why, butit's one of those titles people
have been really gravitatingtowards lately.
I mean, maybe it feels a littlemore accessible than things
like cybersecurity.
You know, you feel like youneed I don't know more
(14:11):
background to get into stufflike that.
Tim McConnaughy (14:13):
I think if you're starting from nothing,
like, say, you're coming throughcollege right now, for example,
the CSPs have laid a largeadvertising campaign around the
idea that like, hey, we makeeverything extremely easy for
you.
You know we have a high we cangive you.
You know we're involved in high.
You know everybody's moving tothe cloud, so you're going to
find a job as soon as you getthese skills.
(14:34):
But, more importantly, youdon't have to know all that
on-prem stuff.
You don't have to know all thatother garbage.
We're serverless.
Erika Dietrick (15:00):
Like we.
You know we're serverless.
You don't have to care about90% of this, we'll take care of
all of're.
Just ticking boxes, you know,but really I can't think of any
engineering discipline where youreally shouldn't know anything
at all about the otherdisciplines.
Chris Miles (15:14):
And you know I'm a little bit.
Erika Dietrick (15:15):
I will say I am a anomaly, I think, on the
software engineering side,because while I was getting my
master's I did a internship atCisco TAC and you know, I
actually learned networkengineering I mean at least at a
CCNA level at the same time assoftware engineering.
So I had more of anappreciation for oh, this is how
it all comes together.
(15:35):
But most software engineers, Iwould say probably zero exposure
to network engineering and thenjust cloud as needed, or you
know, if you're wearing a lot ofhats at a startup or whatever,
yeah, I definitely don't want tocome across like I'm poo-pooing
and, you know, like the old manshaking his fist.
Chris Miles (15:51):
Like back in my days you had to learn every bit
of the stack or anything likethat.
I think it's just more now.
There's also like a wealth ofthings that you need to learn.
Right.
If you tell someone to learneverything, then they're never
going to start right.
Like AI, like I'm sure AI islike the like learning for that
right now is probably sooversaturated and like people
(16:11):
don't know what to do.
Tim McConnaughy (16:13):
I don't know what you're saying.
Yeah, definitely yeah, but likeyeah.
Chris Miles (16:20):
So that one yeah, definitely like yeah.
Tim McConnaughy (16:21):
So it's kind of like um, freedom of choice is
kind of the natural enemy, Ithink, in the in this in this
situation, I think, yeah, well,and also, uh, one thing that the
cloud has done, that is, thatdevelopment that makes
developers probably need tolearn more about the whole stack
is this concept of loosecoupling, where we're pulling
apart applications into otherpieces, right?
So now it's not just front endand back end.
You've got like databases,you've got web tier, you've got
(16:44):
app tier and like all thesedifferent tiers and all, and
they're all in other places.
Uh, so what's?
What's latency?
Chris Miles (16:49):
dude, who cares?
Erika Dietrick (16:51):
yeah, I mean it's.
It's so difficult too because,um, again, I don't know if this
translates to networkengineering as well, but as a
developer I mean even withindevelopment you can have such
vastly different experience andknowledge.
And even what you're talkingabout with, with the decoupling,
I mean, I know people who spendtheir entire careers building,
(17:12):
you know, three-tierapplications that they're
deploying all in the same placeand you know they do that same
thing over and over and over.
You know they have a hammer andthey see everything as a nail.
And obviously it varies, right,and other people are working
with microservices and stuff.
But, like with my cloudexperience, my first time
actually actually having to useit was when I was hired for my
(17:36):
first software engineering jobat a startup and to be honest,
it was well, I'm thankful forthis job.
In case this person hears I amvery thankful I had this job,
but you know it was a little bittraumatic to have to learn.
You know cloud engineeringwhile really practicing software
engineering professionally forthe first time.
And then also there's that youknow, creating a CID, cicd,
(18:00):
pipeline.
You know who does that fall on?
Well, when there's four peoplein the startup, somebody's got
to volunteer, right and that Ithink that very much.
Even in that scenario, right,you'll find a lot of times.
So there was one other engineer,at least initially, in this
startup.
It was just me doing thebackend, which is just things
you can't see, and then thefront end guy who's doing, you
(18:22):
know, the website, and you knowit's not a matter of, oh, you
know well, we're going to have a, you know, sql server database
and so we should really deploythat in Azure because it's more
cost effective, or this, this orthat.
(18:46):
You know, it's just a matter ofwell, I like this one, you know
well, I knew this one.
This one seemed nice, you know,and just arguing, you know,
back and forth.
That that's how most developerconversations go that I've had,
and cloud was no different.
It did not feel like we choseanything based off of any real
um, expertise or reason besides,eh, I feel like doing this, um,
so anyway, that's how I endedup.
Getting into AWS was, uh, youknow I was.
I was doing a, a Microsoft shopback end, but I was told no,
(19:08):
this needs to be an AWS.
Tim McConnaughy (19:11):
Was.
Was it told to be AWS becausethe front end was being
developed on AWS or because thebusiness had decided like AWS,
is we're going to keep it alltogether, or so it was a React
native front end.
Erika Dietrick (19:25):
But again I really think it just comes down
to that Microsoft anti Microsoftsentiment.
I will say I'm prettyopinionated on how much, how
many problems this causes, thatthis is like our only
decision-making process and somany technical conversations,
but I think it was literallyjust Just anti-Microsoft.
Most people seem to use AWS.
Let's use AWS and screw.
Tim McConnaughy (19:47):
Microsoft.
No, that's interesting for ayou know to make such a big
commitment Because, if you thinkabout it, the business is
making a huge bet.
When you're doing back-enddevelopment, front-end
development, you're doing thewhole thing.
You're developing the revenuegenerating app on this cloud and
it sounds like just because youknow, didn't want to do
Microsoft.
Sometimes there's a lot oftimes, of course we have those
(20:10):
discussions and it comes down tohey, what do you know?
What have you used before?
What's your preference?
But to think of an entirebusiness betting, of course it
was a startup, so everything'salready on fire and it's the
Wild West anyway.
But to bet the whole farm onAWS just because anti-Microsoft,
that's interesting.
Erika Dietrick (20:28):
Well, and you know, when nobody has any true
understanding of these cloudplatforms too, I mean it makes
it almost impossible to make anylogical decisions.
It's a lot of waste of time.
I mean it makes it almostimpossible to make any logical
decisions.
It's a lot of waste of time, Imean.
The easiest simple example Ican think of is you know, we had
a database deployed in AWS RDSI don't even remember what that
(20:49):
stands for anymore and then wehad an EC2 instance for
basically just another portionof the app.
It's not that important.
Yeah not that important, and so,anyway, whenever we were
working with and testing thisapp, right, we had to have the
instances on right.
It takes time for you to turnthe instances on and then for
them to tear back down, and soyou know that costs money.
(21:11):
And you know people ask, youknow leaders were asking well,
why?
You know, why are they takingso long?
Blah, blah, blah, why are webeing charged this?
And I'm like look, whenever Itest it, I have to turn it on.
It takes time to turn on, ittakes time to turn off.
And they're like well, we onlywant it to turn on the instant
an API call is being made, andthen it should turn off after
that.
Tim McConnaughy (21:30):
That's not a thing.
Erika Dietrick (21:39):
We wasted an entire at least an entire sprint
or so you know like a amount oftime in developer land Right,
trying to implement middlewarethat would, on a timer or
trigger, turn on and off theseinstances and it didn't work.
Okay, and again, I was likethis is a really bad idea and
this is not how they're designed.
I mean, this is how they'redesigned, this is just how it
works.
Tim McConnaughy (21:58):
Did they not understand what serverless was
at that point?
Like serverless, you know, likeLambda or API Gateway or any of
those services that arecompletely elastic?
Chris Miles (22:06):
I'm thinking you frontend with API Gateway just
to do a Lambda function to startthe instances, and then you
basically say like the responseto the API call is like try
again in 15 minutes.
So I will say you know as asemi early in career person.
Erika Dietrick (22:23):
I don't get a lot of influence, or anything
that's done, but, um, you know,again, I think it goes back to
with developers.
At least it's either hey, uh,this is what I know and we're
gonna do this period make itwork, or I don't know.
They've learned some cool newtechnology, or like they've got
into Lambda or whatever andthey're like everything's you
know going to be Lambda Hammerand nail.
Tim McConnaughy (22:45):
Yeah, the parallels are uncanny, because
it's like that with I mean, youknow, it's like that with
network engineering too, likenobody ever got fired for buying
Cisco.
I know the Cisco CLI, I'm CCNAcertified.
We're buying Cisco, right?
Or Juniper, or whatever that isright.
Or you know, or hey, you know,I know EIGRP, so that's the
(23:08):
routing protocol we're going touse.
You know whether or not itmakes any sense to do so.
Erika Dietrick (23:12):
It's infuriating to me.
Maybe that's just me, Maybe Ijust like making my life hard
and having to learn new thingsall the time, but I got to try.
I feel like you know.
Tim McConnaughy (23:21):
So I mean, tell us, tell us a little bit more
about the startup gig, causethat sounds pretty interesting.
Specifically like a little bit.
I mean you kind of you kind ofwent into the well, what the
business was looking for, which,of course, was completely
untenable, uh, what they weretrying to get you to build, but,
like, what did you end up doing?
I'm curious, like from adevelopment perspective, like
how I'm interested in the in thearchitecture piece that
developers are I have to come upwith here.
Erika Dietrick (23:42):
So I was hired in.
I feel like I gotta be careful,being too honest, but I was
hired in basically because, um,they had initially outsourced
the work, um to an agency inIndia, and so they had one
on-staff engineer who was doingthe front end, which was the
website and the mobile appportions that you could see, and
(24:04):
then this Indian agency wasdoing the back end.
So they were doing a C-sharpREST server, which just means
that's where the API calls werebeing created, and then the SQL
server database, which is whereall the data was stored.
So the story I was told wasthat the communication didn't go
well, they didn't get exactlywhat they were promised and you
(24:27):
know, for different reasons, youhear that story fairly
frequently.
But as a startup, it's prettyhard to pay market rate to
developers in the US, and so youknow I was cheap and bright
eyed and looking for that firstopportunity, and so I was hired
and I was supposed to just be,and I have a background right In
C sharp and SQL server, atleast from university and so I
(24:52):
had come in and I was supposedto basically fix a bunch of bugs
, is all it was told was.
Fix a bunch of bugs is all itwas told it was fix a bunch of
bugs, implement a few newfeatures Again, the architecture
of this originally was that itwas a React Native front end C
(25:13):
Sharp REST server, SQL serverdatabase and we're making API
calls between each of thosethree tiers.
That's how we communicatebetween the tiers.
So, anywho, we get started andthis is very typical of all
startups.
Every startup I've ever workedat or interacted with.
I get started and it's acompletely different situation
when I start.
We're going to start fromscratch, we're going to do this
way better.
(25:33):
You know it's going to behighly performant, all this
stuff.
So I'm like okay, that's cool.
So we call that greenfieldapplication, doing a greenfield
application when you're doing itfrom scratch.
So I got to actually help, youknow, give a voice on
architectural decisions, which Ithink is kind of unique for a
junior software engineer.
And you know that basicallystarted with the database design
(25:57):
, because you know, with arelational database it's a lot
less forgiving than.
Tim McConnaughy (26:04):
MongoDB or other types of databases Like
NoSQL or something.
Erika Dietrick (26:08):
Exactly so.
You really want to think of itfrom a database perspective and
then kind of build the API callsout from there and then connect
the front end via those APIcalls.
So we started with databasedesign.
I don't know how much you wantto go into that, but it's
literally sitting at a tablewith a piece of paper and we're
(26:28):
going back and forth andscratching stuff out, and it
makes more sense to couple itthis way.
Tim McConnaughy (26:33):
I can actually commiserate with you on this,
because when I did my WGU cloudcomputing degree, one of the
courses one of the not just atheory course but an actual like
do something course where youhave to create a project was
database design and I actuallyhad to build a database schema.
Oh my God, it was like I stillhave nightmares.
(26:54):
I hated it.
Erika Dietrick (26:55):
Well, again, what's funny is that, you know,
in university I love SQL Server.
Something about relationaldatabases just made my day, and
then I realized it was becausein school, you know, I was doing
stupid little.
you know, oh, user IDapplication with three tables
you know and you know twoforeign keys, you know.
But yeah, so an enterprise, orwell, this was an enterprise, I
(27:19):
don't know, using mixing up myterminology, a real world
production database.
It was much, much morecomplicated and in a startup you
don't really have anyone to askquestions to right.
Tim McConnaughy (27:29):
So it's very true here You're the, like it or
not, you're the, the person.
Erika Dietrick (27:34):
Exactly yes, so did the database design.
As the only backend engineerand the only person who
understood C Sharp, I was Iwould say over 50% trusted to
just build it how I thought itneeded to be built, so I used
what's called a model viewcontroller architecture.
It's just a way to separate,yeah.
(27:55):
Yeah, I'm familiar yeah, yeah,I'm familiar, uh, and you know,
beyond that again, and startups,it's usually kind of a cluster.
So you know it was a matter ofdo however you want and then I'd
do it.
No, not like that or no, itneeds to have this too, you know
I wouldn't know anything aboutthat, I just I just like to.
I like to be funny and talkcrap, but I honestly learned a
(28:15):
lot of this experience.
Chris Miles (28:16):
So just you have to laugh about it, or else I mean
well, I'm just keeping my bridgeunbridged.
Tim McConnaughy (28:21):
Well, honestly, where else could you have gone
and get this kind of?
Erika Dietrick (28:25):
Yes.
Tim McConnaughy (28:26):
Experience, even if it is stressful at the
time, Like there's very fewplaces that you could you know,
most places would crib you tothe point where you barely learn
anything for a year.
Erika Dietrick (28:35):
You know, I will say that too.
You know it was a cluster.
But you know, had I been at abig I don't know, a fortune 100,
you know I would have been.
You know I do, you know, workthis one little gear in this
gear only.
So I did get to touch a lot.
Tim McConnaughy (28:47):
It's good after the fact.
It's good for building theportfolio.
It's good for building theportfolio.
Chris Miles (28:52):
Glad you did it, but you wouldn't sign up for it
again.
Erika Dietrick (28:54):
Well, you know, and things happen too right.
I mean in startup world, likeyou have venture capital, you
know, like at Cisco.
I mean obviously it matterswhat we do, deadlines matter,
but at the end of the day, youknow, in startup world you
really can't afford to do that.
That's the death of yourcompany and so on top of that,
(29:24):
you're usually riding on venturecapital.
You usually aren't actuallyprofitable for I don't know
years.
You know many startups outthere are not actually
profitable.
Tim McConnaughy (29:32):
No, that's well .
Well, yeah, agreed.
Erika Dietrick (29:35):
So you know you've got this one little eager
you Agreed.
So you know you've got this onelittle eager.
You know junior engineer whoyou know.
The entire architecture, theCICD pipeline, the way it's
deployed, is riding on this.
Tim McConnaughy (29:46):
Millions of dollars.
Erika Dietrick (29:46):
Six-month-old runway.
And then you know, oh, why isthe service down?
You know emergency.
You know, hello, I don't know.
I'm not going to pretend I canfigure it out.
Tim McConnaughy (29:56):
Yeah, no, that lesson's learned Everybody.
Honestly, the best lessons arelearned in the fire, but as long
as it doesn't actually likemake you want to leave tech
forever, you usually come outthe other side with the kind of
experience that would take youyears otherwise to get.
So there's good and bad to it.
It is a very high-stressenvironment.
Somebody was asking in Discordyesterday you know, hey, I want
(30:18):
to get into tech, but what's thework-life balance and the
stress level like?
And I'm like, do you want toknow the answer?
Do you want to get into tech?
Because I'll tell you Like, soanyway, but that's.
Erika Dietrick (30:30):
You know, but it does get better though, because
it does.
So I'm I have, I will bestarting year six in tech this
year.
I mean, you don't feel like youget any wins Like you feel like
you know nothing.
It's not just oh, you know, I'mgood at some things and I get
that self-esteem boost and I'malso learning and getting
(30:52):
knocked back down again.
It's just getting hit in theface over and over and not
knowing.
So I mean, I feel like thatpart does change and I think
your attitude also changes aboutit.
Right?
Because again, university, youknow, as a developer we're not
even I mean, at least I wasn'tin many developers I know you're
not even taught how to debugyour code, which is highly
(31:12):
unrealistic, right.
Tim McConnaughy (31:13):
Oh really, no unit testing or anything.
That's not part of thecurriculum.
Erika Dietrick (31:17):
So I learned how to do unit tests and how to use
certain mocking frameworks.
But I mean, I think that someprofessors, you know, they kind
of gloss over the debugger.
It's complicated, I won't get.
Tim McConnaughy (31:30):
Like the step debugger where you have to go
like Right.
Erika Dietrick (31:33):
So there's this joke in programming land about
how we debug using printstatements.
I don't know if you guys arefamiliar, okay.
Chris Miles (31:40):
Yep, I write Python , so yes, that's the only way
I've ever debugged.
So yeah, so yes, that's theonly way I've ever debugged.
Erika Dietrick (31:53):
So you know it's fine in a pinch or sometimes,
but it's really not that helpfuland inefficient for real world
development.
But shockingly, it's usedconstantly.
People just do not use thedebugger and that's the real
world.
The real world is that yourcode just doesn't work.
It doesn't magically work.
You mess up all the time.
That's changing a little bitwith AI coding assistants.
Some of the really irritatingthings that you troubleshoot for
hours you know don't reallyhappen anymore, but that's good
(32:14):
news, yeah, yeah, so should we?
Chris Miles (32:16):
should we open that up a little bit?
I know, obviously, the the kindof AI coding assistants or
co-pilots, as we put them um,not to use the specific branded
item, but it's one of the mostpopular ones, so it's not like
we can avoid it.
But I mean, how do you feelthat is changing the kind of
development lifecycle?
For better or worse yeah.
Erika Dietrick (32:37):
I mean, there's a lot of AI copilots out there.
Typically, you know, the onepeople are most familiar with
are code generators like GitHubCopilot.
There's a lot of others outthere and they range from.
You know, I only generate codeand have chat functionality to.
I'm a full on developerenvironment, you know, meant to,
you know, save workflows orsuggest things to you in that
(33:00):
way.
So I will say, at least from aGitHub Copilot perspective,
which is so I will say at leastfrom a GitHub co-pilot
perspective, which is, you know,or at least was, the most
widely adopted AI codingassistant code generator at the
time.
It's what I've done, a bunch ofpresentations on.
It's absolutely revolutionizeddevelopment life cycles in both
(33:23):
good and bad ways, and I dothink that, even though bread
and butter developers wereprobably the first to flock to
these tools, that really we'retalking about something that's
going to change things foranyone who codes, and so I know
we've got some network and cloudpeople.
So I'll speak more to that AtCisco Lives and when I give
talks, the main thing I'm askedagain is you know, do I actually
(33:45):
have to learn how to code now?
Is it just a matter of promptengineering?
Um, and then I, and thenthere's people on the other side
of the fence right, who youknow you're like.
You shouldn't be able to touchan AI coding assistant until I
don't know you're a seniordeveloper or whatever.
Tim McConnaughy (33:59):
Oh yeah, the gatekeeping.
Erika Dietrick (34:21):
And what's interesting at least what I
found through my own researchand opinions is, you know, I
think that it's really going tobe not the code generation per
se, especially for networkengineers, but it's going to be
are taught and hired.
And coincidentally, this bleedsa little bit into teaching
network engineers how to code,and assessing their skill is
essentially, you know, we giveyou some assignment, we see how
you execute said assignment orhow you code it, and then are
you able to explain it right,and then we do some form of that
(34:43):
in the hiring process whenwe're evaluating people too,
right, it might be on awhiteboard, it might be solving
some stupid riddle, butessentially we just give an
assignment and ask to code andthen whatever.
So you know that is completelyblown on its head now that there
are AI coding assistants right,because you have an assignment
(35:04):
and you can generate the answerto that assignment, and you can
hit slash explain.
Literally, it's that easy slashexplain, explain to me what was
just written, explain it to meline by line, and you know
that's your student assignment,you know that's your take home
for your, for your developerinterview.
Um, so it's really not aneffective way to oh, to gauge
ability anyway, right andbecause it's not an effective
(35:24):
way to gauge ability, it poses alot of challenges for teaching.
You know, how do you know whensomebody actually knows how to
code, and how reliant should webe on these tools?
So I feel like I'm rambling alittle bit, I'm sorry.
Tim McConnaughy (35:37):
No no, no this is good.
Erika Dietrick (35:38):
This is actually really good.
I think this is important.
Chris Miles (35:42):
But it's funny because, like I've read on
social media, I've never had togo through one myself, but I've
read a lot about, you know,these software engineering
interviews where people feellike the process and the
interviewing criteria is sostrict and so over the top that
they hate the interview processBecause it's all like, oh, do
this live in front of all of us,you know, write this code, do
this thing on a whiteboard, blah, blah, blah.
And they think that it shouldbe simplified and they should
(36:11):
like, almost almost like the,the, the hiring manager should
be more trustworthy about thingsin a sense.
But now it almost sounds likefrom what you're telling me,
like it's kind of spinning backin the other direction.
Like, how do you validatepeople?
Know what the fuck they'retalking about and they haven't
just done this on, you know,through a chat bot, um, other
than doing that.
Like, where do you think itfits into that?
Erika Dietrick (36:28):
So I won't spill all the tea, but I'm going to
be presenting on this inFebruary oh, all right, yeah a
roadmap how I think that weshould now be evaluating and
teaching and all that.
But yeah, I mean, I think it'san opportunity.
You know, people see it as,like you know, ruining engineers
.
No, they're never, going tolearn blah, blah blah.
But I think it's a greatopportunity to revamp how we
(36:48):
teach people to code and how weassess it.
I think that the kind of likeyou mentioned, I think that the
model was completely broken, Imean even when I was just
learning to code, right, I meanit was I overused the word
traumatic, but it was traumatic,you know.
There were many tears shed, itwasn't fun.
I didn't learn the mostefficient or effective way.
Yeah okay, and so I think thatyou know we can focus more on
(37:10):
things like people's pastexperiences and what they
learned from it and how theywould apply that to new problems
, and talking through people'sthought processes and the way
they think about problems, right, because that's what engineers
do.
It's not about do you know Csharp or Python.
I mean a little bit.
(37:31):
It depends on the, on the levelof the role, um but it's about
can you actually think like, doyou know how to solve a problem?
Um, have you encountered enoughproblems that you can deduce
how to solve a similar problemat our company?
So you know, that is generallythe way, the direction that I
hope, interview assessments aretaken.
And again, I think that from achat perspective, you know,
(37:52):
really it's just Google onsteroids.
So developers learn throughGoogle, right, we do a lot of
teaching ourselves, but Copilotchat has been absolutely
revolutionary.
I taught myself TypeScript in acouple of weeks using Copilot
chat Absolutely phenomenal, inmy opinion, at least, based on
my work speed.
So yeah, a very long storyshort.
(38:14):
I could talk about that foreveris that it's really making us
rethink how we teach people tocode and how we assess it for
network engineers as well.
Tim McConnaughy (38:23):
What do you think about not just AI as a
code assistant, which is kind ofwhat we're really talking about
?
You know, hey, generate thiscode for me you mentioned.
Like you know, it changes theway we teach people, do you
think?
Or what do you think AI has a?
Does AI have a place in theteaching process as well, or is
(38:44):
it really better kept to theassistant level?
And if it does have a place,like where do you see that
fitting in?
Erika Dietrick (38:51):
Well, I can answer yes to both your
questions, because I it shouldalways be kept at the assistant
level.
In my opinion, it is not thebrain, and that is where most
people's disappointment comesfrom.
Again, especially if you don'tknow how to code is just like do
this for me you know, so it ismost effective as an assistant,
and that makes sense, right?
Because if it's the brain, thenwhy are we even here?
And then, yeah, 100%.
(39:12):
I think that you know it'sessentially it's going to be
akin to using a calculator formath right.
So you know whether, again, itdoesn't matter what your
background is you're going towant to learn certain
foundations.
You're going to want to learnhow to think programmatically
because that is a skill rightThinking from like a discrete,
mathematic, logical perspective.
And then again, I won't spillall the tea, but there's certain
(39:35):
concepts.
I think you should learn somebuilding blocks and then I think
from there.
I think you use your calculatorto continue learning, and to I
mean not generate all your code,but at least to continue
learning.
Chris Miles (39:46):
Yeah, no, I think that's valuable.
I mean, I think it's um, I hope.
Tim McConnaughy (39:53):
I hope you call your talk software engineering
re-imagined or something likethat or something.
But just kidding, it's alreadysubmitted.
Chris Miles (39:56):
I should have talked to you, no, but I think
it's.
I think that is a greatapproach because I feel like
that kind of takes a page out ofthe network engineering
interviews that I've done overthe past, Like whether or not
I'm, you know, the intervieweeor the interviewer.
You know, I interview a lot ofpeople and a lot of times when
I'm asking questions, I'm notasking the question to know
(40:20):
whether or not you know theanswer.
I don't care if you know theanswer.
I want to know how you reactwhen you don't know it.
Like how you react when youdon't know it.
like how like how you explain,how like and what you said.
You want to know if people haveseen enough problems to learn
like theory around yeah, whattheir thought process is, how
they troubleshoot you know, ifthey can draw correlations.
I want to see how they buildthat mind map like live, right,
(40:40):
and then it depends on whatquestions they ask me in
response.
Tim McConnaughy (40:44):
Right, that's yeah, that's how you really tell
what that's like.
What questions do they ask mein response?
Right, yeah, that's how youreally tell.
What questions do they ask youback to try to get more
information?
Because you can teach anybodynetwork engineering.
You can teach anybodyTypeScript or code or you know
whatever.
I can teach structured data,json to somebody, like all that
stuff.
It's just stuff in a book thatyou can teach someone.
You can't teach someone how tothink.
(41:05):
You can't teach someone how tohow to solve problems.
Like you know not, not really.
I mean you can, but like theactual thought process of
solving problems is somethingthat people have to kind of come
up with, figure out how itworks for them, and they have to
do it themselves.
Erika Dietrick (41:16):
Yep, 100%.
And I mean these co-pilots,right, I mean we can try to slow
them down and say, rah, no, no,keep them out of coding, but I
mean really they're going toinfiltrate every area of life
like we've.
I went to um my first black hatum this year, which is a cyber
security conference, and theywere talking about how um
microsoft has a co-pilot studionow where anybody of any
(41:38):
non-technical background cancreate their own ai co-pilot and
connect it to, like, forinstance, their hr database and
have it assist them with thingsor whatever it is.
Their profession is right.
Tim McConnaughy (41:47):
That's a great idea.
Erika Dietrick (41:50):
So there's the self-serve no code, low code
options, but there's also justall these products coming out
right.
I mean, we've co -pilot, forAzure was recently announced, so
there's no aspect of tech thatis not going to be touched by
some form of AI assistant thatcan't just parrot out facts to
you, right or lines of code.
(42:12):
So, again, I think we can seeit as a bad thing, but at the
end of the day, you knowcompanies want productivity, so
it's in your best interest tolearn it, and it's an
opportunity to hire people whocan solve problems.
Tim McConnaughy (42:26):
Yeah, I mean, and it can keep people from
having to know everything.
Like you know, for example, I'mworking on a project with
Andrew Brown for his Gen AIbootcamp, and you know we were
working on something yesterdaywhere I needed to get PyDict
data out, which I didn't evenknow.
Pydict was a thing Like it's,you know, python, interpretation
of structured data.
(42:47):
So I went and I like explainedmy problem I think I was using
Claude I explained my problemand like what I was looking to
get back from structured dataand and it it did it.
Now I can't say it's a hundredpercent correct because I don't
know it well enough, but I willsay that it's got me a lot
further along than I would havebeen had I had tried to go read
documentation and tried to learnit all myself, like so, yeah, I
(43:10):
mean, it's like the personalassistant thing is definitely
keeping you from having to learneverything.
Erika Dietrick (43:17):
Well, even just navigating bad documentation.
So I kind of I'll just talkabout it because I kind of axed
the project.
But I was interested in doing aVS code extension at one point,
um this year and, uh, you knowI was not familiar with that
development.
Right, they have their own APIthat you have to learn.
Uh, so I'm going through theAPI documentation and I'm
reading through it and at firstit doesn't make sense because
(43:42):
I'm brand new to me.
I have zero clue what they'retalking about and I was like you
know what I'm just going to askmy GitHub co-pilot.
And GitHub co-pilot just brokeit down Boom, Simple, Summarized
how the API is structured, whatAPI call I need to make to do
certain actions.
It was that easy.
So I didn't even need to reallylearn the API.
I didn't need to struggle overthe documentation or guess or
(44:06):
make the API call and realizeit's actually needs this
additional field that's requiredor whatever.
Tim McConnaughy (44:12):
Yeah, I think people are afraid, though, that
people, people are leapingstraight to oh well, then you
know I don't need then then if Iknow that, then you're going to
replace me with a robot type ofthing, and I think that's a big
.
I think the more I, the more Ilearn about what they are and
are not capable of.
It's a very large leap.
You know, if the only value youbring your organization is
knowing how a certain API callworks, then you're probably not
(44:34):
going to be around very longanyway.
Erika Dietrick (44:36):
Well, you know, I think I lost my train of
thought.
Let me see if I get it back.
I think that it's the sameconversation with automation,
right?
I mean everybody, you know.
They set their sight on somepoint in time and technology and
they're like I'm going to getgood at this, and as soon as
they get good at that, the pointhas moved somewhere else and
they have to know something else, right?
And so they think thatknowledge is useless.
(44:56):
But I mean, it's really notright.
I mean even just from theperspective of problem solving
and learning or how the brainworks, right?
I don't remember all theterminology, but you know, when
you learn something, you knowthe brain can more easily learn
related topics because it hassomething to connect it to.
Tim McConnaughy (45:13):
So that's exactly right.
Erika Dietrick (45:15):
I mean, there's many reasons.
You can go on and on about that, but I think people just get
too laser focused on you know,got to know this.
Tim McConnaughy (45:22):
Right, yeah, well, right, yeah well, the
moving goalpost thing, thetreadmill, the tech treadmill,
is well published and wellunderstood, right, we're always
on a treadmill in tech and it'sit's like the uh, you know the
old cartoons with the, thefishing pole, and it's got the,
the banana or a piece of cake orwhatever it is, the tank, we're
all, we're all just likerunning towards it forever.
Um, but that's, that's a wholeother episode.
(45:43):
Um, okay, so no, this has beengreat.
We need to go ahead and wrap up, but it's been really great
having you on talking to youabout this stuff.
Yeah, I think we covered a lotof good stuff tonight.
Erika Dietrick (45:55):
Yeah, no.
Thanks so much for having me onand for dealing with my Casper
lighting and a greatconversation you guys are doing
a great job on the podcast.
Oh, thank you.
Chris Miles (46:03):
I'll do what I can in post, but I promise all the
listeners or the watchers, Ishould say that Erica does not
look as pale-faced as her camerais made.
Tim McConnaughy (46:10):
She's not quite , as I don't look enough though
it's fairly close.
I mean, we live in NorthCarolina.
There's no way she's walkingaround with skin that pale and
still being alive.
Erika Dietrick (46:19):
I've got a lot of Irish descent, so oh, me too.
Tim McConnaughy (46:21):
Yes, yes, yeah.
So how do people, how do peoplefind you on socials?
How do people connect with you?
Erika Dietrick (46:36):
Yes, yes, so I'm on LinkedIn as my name Erica
Dietrich and I'm on X, Instagramand.
Tiktoks for as long as thatlives, Erica underscore the dev
More platforms to come out, butthat's where you can find me for
now.
Tim McConnaughy (46:44):
Yeah, we might lose two of those things before
too long.
We'll see.
Erika Dietrick (46:46):
Oh man Okay.
Tim McConnaughy (46:49):
It's crashing and burning with the tech
community right now, but anyway,okay, great, so everybody go.
Chris Miles (46:55):
Do you want to plug the talk you have coming up?
Tim McConnaughy (46:58):
What was that?
Anything, anything else youwant to plug?
Erika Dietrick (47:00):
Sure.
So I have four talks coming upat Cisco Live in February, but
the one most relevant to ourdiscussion is called Teaching
Coding Skills in the Age of AI,and I'm actually co-presenting.
It's my first co-presentationand I'm doing it with my dad
he's a university professor he'sbeen a software engineer for 25
years, so I thought it'd be anice blend of perspectives the
(47:23):
somewhat newer developer andseasoned I love it Cool.
Yeah, we're going to take youthrough, you know, the
traditional, where we werebefore and where we're headed
now, provide you with a roadmap.
And again, at the last CiscoLives I've went to.
Every single time I'mapproached by you know, managers
of network engineers saying youknow, I need a blueprint for
(47:45):
teaching my engineers how tocode and we're going to provide
that.
Tim McConnaughy (47:48):
Oh, that's awesome.
I can't wait to see it.
This is an Amsterdam, stillright it's?
Erika Dietrick (47:52):
an Amsterdam.
Hopefully it'll rotate.
If it's popular, it will.
Tim McConnaughy (47:56):
Oh, I love it.
That's a great idea.
I can't wait to catch it on thereplay.
All right, everybody.
Well, this has been the Cablesto Clouds podcast.
Make sure you follow us on allof the social medias that we
still are on, and we'll see younext week.
Bye, hi everyone, it's Tim andthis has been the Cables to
(48:16):
Clouds podcast.
Thanks for tuning in today.
If you enjoyed our show, pleasesubscribe to us in your
favorite podcast catcher, aswell as subscribe and turn on
notifications for our YouTubechannel to be notified of all
our new episodes.
Follow us on socials atCables2Clouds.
You can also visit our websitefor all the show notes at
Cables2Cloudscom.
(48:38):
Thanks again for listening andsee you next time.
Hey everyone.
It's Chris from Cables toClouds here.
I hope everyone is enjoying anice holiday break, spending
time with loved ones, family,friends, all of the above.
I think this episode's actuallycoming out on Christmas, so if
you celebrate, merry Christmasand upcoming Happy New Year to
everyone as well.
Before we get to the episode, Ijust want to take a second to
let everyone know that Tim and I, along with one of our
(00:22):
critically acclaimed guests, oneof our critically acclaimed
guests, one of our mostdownloaded episodes, I think in
the past with Steve McNutt,we've all come together to
release the first edition of acertification guide for the AWS
Certified Advanced NetworkingSpecialty Exam that is being put
(00:43):
out by PACT.
So the book is actually up forpre-order now PACT.
So the book is actually up forpre-order now.
So if you're interested, ifyou're doing anything with
networking in AWS, I wouldheavily, heavily encourage you
guys to check it out, even ifyou're not preparing for the
exam.
I feel like it's just a goodoverall composition of all the
networking services within AWStoday.
So the book is officially upfor pre-order now and I believe
you get early access throughPACT if you do the pre-order.
(01:04):
So I will put the link in theshow notes and if you're
interested, I encourage you toplease check it out.
And with that out of the waywe'll get to the episode.
Tim McConnaughy (01:11):
Thanks, Welcome to the Cables to Clouds podcast
, your one-stop shop for allthings hybrid and multi-cloud
networking.
Now here are your hosts Tim.
Chris Miles (01:22):
Chris and Alex.
Tim McConnaughy (01:25):
Hello and welcome back to another episode
of the Cables to Clouds podcast.
I'm your host this week, tim atcarpe-dmvpn on Blue Sky, and
with me, as always, is myco-host, chris at the Cloud Main
on Blue Sky.
And, as you may have guessed,we have a guest with us this
(01:47):
week.
We brought Erica Dietrich hereto talk about cloud development.
I'm going to let her introduceherself in a second, but you
know, one thing I've always beenpushing for is that I want to
move across the stack.
I want cloud infrastructure andcloud network people to get
more understanding around what'sriding on the network, because
I think we need to have opinionsand be opinionated about that.
(02:09):
So, yeah, I just wanted tobring Erica on here and talk
about cloud development.
So, erica, go ahead andintroduce yourself.
Erika Dietrick (02:14):
Yeah, so yeah, my name is Erica Dietrich.
I'm a developer advocate atCisco.
I'm part of our developerprogram there and I'm basically
enabling people to learn networkautomation, to use AI tools in
their automation workflow.
And, like you mentioned, tim, alot of our community is also
infrastructure and networkengineers or cloud engineers.
(02:36):
So I get to talk to a lot ofthat community and I spend about
half of my time building things, breaking shit learning, and
half of my time, you know,building things breaking shit
learning and half of my timecommunicating.
Tim McConnaughy (02:48):
So getting presentations being on podcasts
like this, annoying people onsocial media, all that jazz yeah
, no, that's awesome, let's see,so let's, let's just, I guess,
just let's just jump right intoit.
So talk to me a little bitabout, first of all, just about
what your day-to-day looks like,but also about what you've been
(03:09):
doing with cloud development.
Erika Dietrick (03:12):
Sure, day-to-day ranges wildly.
I would say that we work inseasons.
I think that's a developeradvocate thing, not just a me
thing.
Chris Miles (03:21):
It is.
Erika Dietrick (03:24):
We work really hard towards big outputs.
So, whether that's, you know, aawesome YouTube tutorial video
at the end of the month, orwhether it's a Cisco live coming
up, that we're preparing theDevNet zone and a bunch of talks
, it's very periods of head downdelivering and then kind of a
recovery period head downdelivering and then kind of a
recovery period.
(03:44):
So I would say on a typical dayfor me, I'm usually doing my
community interaction firstthing in the morning, posting to
social media whatever it was Ilearned and cobbled together the
day before, try to be somewhatuseful and educational and
interact with the communitythere, and then it varies widely
(04:08):
.
Again, we've got Cisco Livecoming up in February, which
seems like it's really far away,but it's really not.
So usually spend a good chunkof my day trying to ignore
everyone after that and, youknow, become an expert at
whatever it is I'm talking about, because that is a big
misconception is that we're justcomplete experts at everything
we talk about.
We're like everyone else intech.
(04:29):
We have to learn and then weteach, and then the afternoon I
like leaving an ample amount oftime for shit catching on fire.
Chris Miles (04:38):
So honestly, my app is pretty flexible, as it does.
Tim McConnaughy (04:42):
As you do Okay, no, do yeah, okay, no.
That's great.
It's funny because I have manydeveloper advocate friends and
it's pretty simple across all ofthe different silos of
developer advocacy.
That's the main goal, if youwill.
Erika Dietrick (05:01):
Very different philosophies around it, for sure
.
It's just very different fromcompany to company, org to org,
and Cisco in particular, it's alittle more challenging because
we're not a traditionaldeveloper program, that's true.
Tim McConnaughy (05:13):
We're talking about DevNet, of course.
Yeah, yeah, yeah, so that's avery good point.
Actually, most of the developeradvocate types are with a
developer-focused company, likeone of the CSPs, for example
it's a lot of the DAs I know orwith, like you know, like
HashiCorp, or just one of thosedeveloper focused type of
(05:34):
organizations.
Okay, so, all right.
So let's, how did you getstarted with the cloud stuff?
I mean, obviously you probablydidn't, like, like all of us,
you didn't start, you didn'tstart with cloud.
So like, where did how did youget started?
Um to into the cloud, like fromwhere you were?
Erika Dietrick (05:50):
Right, right.
So, uh, and this is.
This is one of the things thatI thought could be insightful or
interesting or maybe it's justgoing to validate what you guys
already know about howdevelopers approach the cloud,
but um, the cloud, but um.
So I was working a lot ofdifferent gigs after my
undergraduate degree, right, um,I was most recently, before
being in tech, working atstarbucks and, uh, managed to
(06:11):
get back into a master's programof software engineering.
Uh.
So, granted, I was looking tohave a challenging and rewarding
career and my dad was asoftware engineer and for some
reason, there was some sort ofdisconnect where I never learned
what he did or why.
I just knew he liked his joband I'm like what the hell?
You know, I'm smart enough,probably.
And so I went into this softwareengineering master's degree at
(06:35):
East Carolina and I would saythat, even though I had a lot of
catching up to do, myexperience when it comes to the
cloud was probably prettytypical with every developer's
learning experience, especiallyif they're going the academic
route, and that is that you knowdevelopment.
We're very much focused in theIDE or in the developer
environment, where it's just youlearn to master a language and
(06:55):
you don't really give a.
I don't know, should I becussing less?
Tim McConnaughy (06:58):
No, that's fine , You're good, you don't?
Erika Dietrick (07:01):
really give a fuck about anything else, right,
you know.
So I think that it's a lot morenaive and siloed than coming
from an IT network engineeringbackground, where you kind of
(07:21):
have a general, well-roundedsense of how technology works
and how things communicate inthe real world.
You're just sitting there inuniversity learning variables
and methods and whatever otherprogramming concepts and being
happy that you are printing alittle line statement to the
console.
Eventually you get moreadvanced, right, depending on
your interest mobile appdevelopment, game development,
whatever.
But what's not taught inuniversity is really anything
beyond just building the code.
Tim McConnaughy (07:42):
The language itself.
Erika Dietrick (07:43):
yeah, so, in terms of how are real-world apps
deployed?
Where are they deployed?
What do you have to considerwhen deploying them?
Everything, I mean I talk aboutthis a lot with DevNet too but
everything from code security tohow you document, how you work
with other people and not juston your homework.
It's a pretty big problem in myopinion, and it really.
I was, I guess, a little bitlucky because in university I
(08:06):
ended up taking a Google CloudPlatform course and it was
because they had some kind ofstudent program that they had
partnered with ECU and they weregiving, you know, a certain
number of free credits.
A professor that I absolutelyadored had been, I think he
applied for this program and waschosen and they gave him some,
you know, PowerPoints andmaterials and he did a great job
(08:27):
with the course.
Um, but so I'm I'm lucky in thesense that I was at least
exposed to it before, you know,the real world hit me in the
face.
But, um, you know this, thisprogram?
It essentially just kind ofthrows you into a series of labs
, and the labs are hands-on,they're well done, they're meant
to get you familiar withdifferent GCP services and using
(08:49):
the CLI and all that.
But again, as a developer, youjust have no idea like why am I
learning?
this.
Why am I?
There's a huge disconnect,right.
You don't have anyunderstanding of networking or
how applications run.
So I took that course and it itwas exposure, but nothing
really stuck.
Um, what's interesting is that?
(09:09):
Um, my experience and stop mewhenever you want my experience,
both as a student and later onin startup world, was all with
Google cloud platform, becausethey really push heavy on those
empowering you to use theplatform free credits, all that
but afterwards I've nevertouched GCP.
(09:30):
It seems to be they're reallytrying to claim the market there
on the little guys and girls.
Chris Miles (09:36):
Well, we've been talking about it on this pod for
a while.
It seems like GCP is, or, as weshould say, google.
Tim McConnaughy (09:43):
Cloud.
It's not GCP anymore, butGoogle Cloud.
Chris Miles (09:47):
It feels like they win the developer battle almost
every time.
They are very attractive todevelopers day in, day out.
I don't know.
Do you have a take on why thatis?
Is there something in thatrealm that was more friendly to
you, that was inviting?
I mean, obviously there'scredits and there's incentives
to use the platform so that it's, you know, cost effective.
But, like the platform itself,what's, what's?
(10:08):
What sets it apart in your mind?
Erika Dietrick (10:10):
It's honestly, it's interesting to hear you say
that, because I okay.
Chris Miles (10:14):
So I think, it's easier to use.
I think it's easier to use anAWS.
Erika Dietrick (10:17):
I don't think there's anything particularly
special about how Google cloudis set up.
I just think there's lessservices to choose from and it's
not as confusing as AWSInteresting okay.
And, you know, in the developerworld too.
I don't know, maybe this is thesame for network engineering,
but there's basically Microsoftshops and non-Microsoft shops.
Tim McConnaughy (10:36):
Oh yeah, right sure.
Erika Dietrick (10:38):
So you know, if you're a Microsoft shop, you're
using Azure, no matter whatRight.
And if you're not, you know,and again, and you've never been
exposed to a cloud platform,you're probably like oh, AWS is
complicated, you know, jump overto Google.
Tim McConnaughy (10:51):
Yeah, what's interesting is something we
observed and I still think it'strue is that I feel like Google
Cloud's networking was builtwith developers in mind, like
the way they do their networkingwas built like you don't have
to care about it, we'll justkind of do it all for you.
Like built towards developersthat won't know or care how the
network works, we'll just kindof figure it out for you, which
(11:14):
confuses a lot of us networkpeople because it doesn't work
like you would expect networkingto work.
Erika Dietrick (11:21):
Well, and I think it's part of that
disconnect to the farther, weabstract what's happening in our
cloud services versus what'sactually happening in the data
center.
Chris Miles (11:30):
You know what?
Erika Dietrick (11:30):
I mean as a developer.
Okay, that's great that you'recatering to us and, like I work
with them, enabling them on thecode side, and you know they
want to take shortcuts or justuse low code, no code or
whatever it's like.
I know this was built forpeople who don't want to learn
(11:53):
and it's easier, but it's notreally great.
Chris Miles (11:55):
Yeah, I totally get that I'm assuming this has to
relate to the fact that there'sjust been like a boom in the
industry and there's more peoplegravitating towards this career
field, right, like I think,like you mentioned before, that
early on you know that you'dlike you didn't have to learn
some of these things.
Like you know, you just likefocus on the code, but then
(12:15):
maybe you start peeling back theonion and learning how
networking works.
I feel like that comes with thegenuine curiosity that was kind
of unique about either workingin infrastructure or working in
software development.
You probably had that genuinecuriosity that was kind of
unique about it, either workingin infrastructure or working in
software development.
You probably had that genuinecuriosity regardless and it kind
of led you down those otheravenues to explore because you
just like couldn't fathom notunderstanding how the entire
(12:36):
system works, right.
But I feel like now that thebar of entry for this stuff is
much lower and it's becoming amuch more standard career field,
like I think people just lackthat curiosity and like I've
been a big proponent about thisfor a long time is like people
need jobs and you have to letpeople be mediocre at their job
sometimes, because to somepeople this is just a fucking
(12:57):
job, right?
This is just how you put inyour 40 hours and you go home
and, like you, take care of yourfamily, and that's that's
what's important to you, right?
You don't give a shit about whatyou're doing for work really.
So, like I don't know, do youthink there's a correlation
there?
It's like just coming from theother side of the stack.
Erika Dietrick (13:11):
A correlation between sorry, can you summarize
the?
Chris Miles (13:15):
question.
Yeah, sorry, I said a lot there.
Just kind of the boost in thecareer field, right, and there's
probably more people justlooking at it from the lens of
like, hey, this is a job morethan this is like something I'm
truly deeply interested in andwant to find out more.
Erika Dietrick (13:29):
I definitely think there's two camps of I'll
just generalize as tech folks.
But you know, there's peoplewho just have a love of learning
and you know, maybe they startout in one discipline but
they're at least willing toexplore in other areas or where
it makes sense and they want tounderstand how things work.
And then I feel like the otheris just looking for that high
salary tech career, right.
(13:50):
They gravitate towards the titleand I'm going to get the
credential for that title, I'mgoing to build for that title
and then I don't care aboutanything else.
So I definitely think cloudengineering is one of those.
I'm not really sure why, butit's one of those titles people
have been really gravitatingtowards lately.
I mean, maybe it feels a littlemore accessible than things
like cybersecurity.
You know, you feel like youneed I don't know more
(14:11):
background to get into stufflike that.
Tim McConnaughy (14:13):
I think if you're starting from nothing,
like, say, you're coming throughcollege right now, for example,
the CSPs have laid a largeadvertising campaign around the
idea that like, hey, we makeeverything extremely easy for
you.
You know we have a high we cangive you.
You know we're involved in high.
You know everybody's moving tothe cloud, so you're going to
find a job as soon as you getthese skills.
(14:34):
But, more importantly, youdon't have to know all that
on-prem stuff.
You don't have to know all thatother garbage.
We're serverless.
Erika Dietrick (15:00):
Like we.
You know we're serverless.
You don't have to care about90% of this, we'll take care of
all of're.
Just ticking boxes, you know,but really I can't think of any
engineering discipline where youreally shouldn't know anything
at all about the otherdisciplines.
Chris Miles (15:14):
And you know I'm a little bit.
Erika Dietrick (15:15):
I will say I am a anomaly, I think, on the
software engineering side,because while I was getting my
master's I did a internship atCisco TAC and you know, I
actually learned networkengineering I mean at least at a
CCNA level at the same time assoftware engineering.
So I had more of anappreciation for oh, this is how
it all comes together.
(15:35):
But most software engineers, Iwould say probably zero exposure
to network engineering and thenjust cloud as needed, or you
know, if you're wearing a lot ofhats at a startup or whatever,
yeah, I definitely don't want tocome across like I'm poo-pooing
and, you know, like the old manshaking his fist.
Chris Miles (15:51):
Like back in my days you had to learn every bit
of the stack or anything likethat.
I think it's just more now.
There's also like a wealth ofthings that you need to learn.
Right.
If you tell someone to learneverything, then they're never
going to start right.
Like AI, like I'm sure AI islike the like learning for that
right now is probably sooversaturated and like people
(16:11):
don't know what to do.
Tim McConnaughy (16:13):
I don't know what you're saying.
Yeah, definitely yeah, but likeyeah.
Chris Miles (16:20):
So that one yeah, definitely like yeah.
Tim McConnaughy (16:21):
So it's kind of like um, freedom of choice is
kind of the natural enemy, Ithink, in the in this in this
situation, I think, yeah, well,and also, uh, one thing that the
cloud has done, that is, thatdevelopment that makes
developers probably need tolearn more about the whole stack
is this concept of loosecoupling, where we're pulling
apart applications into otherpieces, right?
So now it's not just front endand back end.
You've got like databases,you've got web tier, you've got
(16:44):
app tier and like all thesedifferent tiers and all, and
they're all in other places.
Uh, so what's?
What's latency?
Chris Miles (16:49):
dude, who cares?
Erika Dietrick (16:51):
yeah, I mean it's.
It's so difficult too because,um, again, I don't know if this
translates to networkengineering as well, but as a
developer I mean even withindevelopment you can have such
vastly different experience andknowledge.
And even what you're talkingabout with, with the decoupling,
I mean, I know people who spendtheir entire careers building,
(17:12):
you know, three-tierapplications that they're
deploying all in the same placeand you know they do that same
thing over and over and over.
You know they have a hammer andthey see everything as a nail.
And obviously it varies, right,and other people are working
with microservices and stuff.
But, like with my cloudexperience, my first time
actually actually having to useit was when I was hired for my
(17:36):
first software engineering jobat a startup and to be honest,
it was well, I'm thankful forthis job.
In case this person hears I amvery thankful I had this job,
but you know it was a little bittraumatic to have to learn.
You know cloud engineeringwhile really practicing software
engineering professionally forthe first time.
And then also there's that youknow, creating a CID, cicd,
(18:00):
pipeline.
You know who does that fall on?
Well, when there's four peoplein the startup, somebody's got
to volunteer, right and that Ithink that very much.
Even in that scenario, right,you'll find a lot of times.
So there was one other engineer,at least initially, in this
startup.
It was just me doing thebackend, which is just things
you can't see, and then thefront end guy who's doing, you
(18:22):
know, the website, and you knowit's not a matter of, oh, you
know well, we're going to have a, you know, sql server database
and so we should really deploythat in Azure because it's more
cost effective, or this, this orthat.
(18:46):
You know, it's just a matter ofwell, I like this one, you know
well, I knew this one.
This one seemed nice, you know,and just arguing, you know,
back and forth.
That that's how most developerconversations go that I've had,
and cloud was no different.
It did not feel like we choseanything based off of any real
um, expertise or reason besides,eh, I feel like doing this, um,
so anyway, that's how I endedup.
Getting into AWS was, uh, youknow I was.
I was doing a, a Microsoft shopback end, but I was told no,
(19:08):
this needs to be an AWS.
Tim McConnaughy (19:11):
Was.
Was it told to be AWS becausethe front end was being
developed on AWS or because thebusiness had decided like AWS,
is we're going to keep it alltogether, or so it was a React
native front end.
Erika Dietrick (19:25):
But again I really think it just comes down
to that Microsoft anti Microsoftsentiment.
I will say I'm prettyopinionated on how much, how
many problems this causes, thatthis is like our only
decision-making process and somany technical conversations,
but I think it was literallyjust Just anti-Microsoft.
Most people seem to use AWS.
Let's use AWS and screw.
Tim McConnaughy (19:47):
Microsoft.
No, that's interesting for ayou know to make such a big
commitment Because, if you thinkabout it, the business is
making a huge bet.
When you're doing back-enddevelopment, front-end
development, you're doing thewhole thing.
You're developing the revenuegenerating app on this cloud and
it sounds like just because youknow, didn't want to do
Microsoft.
Sometimes there's a lot oftimes, of course we have those
(20:10):
discussions and it comes down tohey, what do you know?
What have you used before?
What's your preference?
But to think of an entirebusiness betting, of course it
was a startup, so everything'salready on fire and it's the
Wild West anyway.
But to bet the whole farm onAWS just because anti-Microsoft,
that's interesting.
Erika Dietrick (20:28):
Well, and you know, when nobody has any true
understanding of these cloudplatforms too, I mean it makes
it almost impossible to make anylogical decisions.
It's a lot of waste of time.
I mean it makes it almostimpossible to make any logical
decisions.
It's a lot of waste of time, Imean.
The easiest simple example Ican think of is you know, we had
a database deployed in AWS RDSI don't even remember what that
(20:49):
stands for anymore and then wehad an EC2 instance for
basically just another portionof the app.
It's not that important.
Yeah not that important, and so,anyway, whenever we were
working with and testing thisapp, right, we had to have the
instances on right.
It takes time for you to turnthe instances on and then for
them to tear back down, and soyou know that costs money.
(21:11):
And you know people ask, youknow leaders were asking well,
why?
You know, why are they takingso long?
Blah, blah, blah, why are webeing charged this?
And I'm like look, whenever Itest it, I have to turn it on.
It takes time to turn on, ittakes time to turn off.
And they're like well, we onlywant it to turn on the instant
an API call is being made, andthen it should turn off after
that.
Tim McConnaughy (21:30):
That's not a thing.
Erika Dietrick (21:39):
We wasted an entire at least an entire sprint
or so you know like a amount oftime in developer land Right,
trying to implement middlewarethat would, on a timer or
trigger, turn on and off theseinstances and it didn't work.
Okay, and again, I was likethis is a really bad idea and
this is not how they're designed.
I mean, this is how they'redesigned, this is just how it
works.
Tim McConnaughy (21:58):
Did they not understand what serverless was
at that point?
Like serverless, you know, likeLambda or API Gateway or any of
those services that arecompletely elastic?
Chris Miles (22:06):
I'm thinking you frontend with API Gateway just
to do a Lambda function to startthe instances, and then you
basically say like the responseto the API call is like try
again in 15 minutes.
So I will say you know as asemi early in career person.
Erika Dietrick (22:23):
I don't get a lot of influence, or anything
that's done, but, um, you know,again, I think it goes back to
with developers.
At least it's either hey, uh,this is what I know and we're
gonna do this period make itwork, or I don't know.
They've learned some cool newtechnology, or like they've got
into Lambda or whatever andthey're like everything's you
know going to be Lambda Hammerand nail.
Tim McConnaughy (22:45):
Yeah, the parallels are uncanny, because
it's like that with I mean, youknow, it's like that with
network engineering too, likenobody ever got fired for buying
Cisco.
I know the Cisco CLI, I'm CCNAcertified.
We're buying Cisco, right?
Or Juniper, or whatever that isright.
Or you know, or hey, you know,I know EIGRP, so that's the
(23:08):
routing protocol we're going touse.
You know whether or not itmakes any sense to do so.
Erika Dietrick (23:12):
It's infuriating to me.
Maybe that's just me, Maybe Ijust like making my life hard
and having to learn new thingsall the time, but I got to try.
I feel like you know.
Tim McConnaughy (23:21):
So I mean, tell us, tell us a little bit more
about the startup gig, causethat sounds pretty interesting.
Specifically like a little bit.
I mean you kind of you kind ofwent into the well, what the
business was looking for, which,of course, was completely
untenable, uh, what they weretrying to get you to build, but,
like, what did you end up doing?
I'm curious, like from adevelopment perspective, like
how I'm interested in the in thearchitecture piece that
developers are I have to come upwith here.
Erika Dietrick (23:42):
So I was hired in.
I feel like I gotta be careful,being too honest, but I was
hired in basically because, um,they had initially outsourced
the work, um to an agency inIndia, and so they had one
on-staff engineer who was doingthe front end, which was the
website and the mobile appportions that you could see, and
(24:04):
then this Indian agency wasdoing the back end.
So they were doing a C-sharpREST server, which just means
that's where the API calls werebeing created, and then the SQL
server database, which is whereall the data was stored.
So the story I was told wasthat the communication didn't go
well, they didn't get exactlywhat they were promised and you
(24:27):
know, for different reasons, youhear that story fairly
frequently.
But as a startup, it's prettyhard to pay market rate to
developers in the US, and so youknow I was cheap and bright
eyed and looking for that firstopportunity, and so I was hired
and I was supposed to just be,and I have a background right In
C sharp and SQL server, atleast from university and so I
(24:52):
had come in and I was supposedto basically fix a bunch of bugs
, is all it was told was.
Fix a bunch of bugs is all itwas told it was fix a bunch of
bugs, implement a few newfeatures Again, the architecture
of this originally was that itwas a React Native front end C
(25:13):
Sharp REST server, SQL serverdatabase and we're making API
calls between each of thosethree tiers.
That's how we communicatebetween the tiers.
So, anywho, we get started andthis is very typical of all
startups.
Every startup I've ever workedat or interacted with.
I get started and it's acompletely different situation
when I start.
We're going to start fromscratch, we're going to do this
way better.
(25:33):
You know it's going to behighly performant, all this
stuff.
So I'm like okay, that's cool.
So we call that greenfieldapplication, doing a greenfield
application when you're doing itfrom scratch.
So I got to actually help, youknow, give a voice on
architectural decisions, which Ithink is kind of unique for a
junior software engineer.
And you know that basicallystarted with the database design
(25:57):
, because you know, with arelational database it's a lot
less forgiving than.
Tim McConnaughy (26:04):
MongoDB or other types of databases Like
NoSQL or something.
Erika Dietrick (26:08):
Exactly so.
You really want to think of itfrom a database perspective and
then kind of build the API callsout from there and then connect
the front end via those APIcalls.
So we started with databasedesign.
I don't know how much you wantto go into that, but it's
literally sitting at a tablewith a piece of paper and we're
(26:28):
going back and forth andscratching stuff out, and it
makes more sense to couple itthis way.
Tim McConnaughy (26:33):
I can actually commiserate with you on this,
because when I did my WGU cloudcomputing degree, one of the
courses one of the not just atheory course but an actual like
do something course where youhave to create a project was
database design and I actuallyhad to build a database schema.
Oh my God, it was like I stillhave nightmares.
(26:54):
I hated it.
Erika Dietrick (26:55):
Well, again, what's funny is that, you know,
in university I love SQL Server.
Something about relationaldatabases just made my day, and
then I realized it was becausein school, you know, I was doing
stupid little.
you know, oh, user IDapplication with three tables
you know and you know twoforeign keys, you know.
But yeah, so an enterprise, orwell, this was an enterprise, I
(27:19):
don't know, using mixing up myterminology, a real world
production database.
It was much, much morecomplicated and in a startup you
don't really have anyone to askquestions to right.
Tim McConnaughy (27:29):
So it's very true here You're the, like it or
not, you're the, the person.
Erika Dietrick (27:34):
Exactly yes, so did the database design.
As the only backend engineerand the only person who
understood C Sharp, I was Iwould say over 50% trusted to
just build it how I thought itneeded to be built, so I used
what's called a model viewcontroller architecture.
It's just a way to separate,yeah.
(27:55):
Yeah, I'm familiar yeah, yeah,I'm familiar, uh, and you know,
beyond that again, and startups,it's usually kind of a cluster.
So you know it was a matter ofdo however you want and then I'd
do it.
No, not like that or no, itneeds to have this too, you know
I wouldn't know anything aboutthat, I just I just like to.
I like to be funny and talkcrap, but I honestly learned a
(28:15):
lot of this experience.
Chris Miles (28:16):
So just you have to laugh about it, or else I mean
well, I'm just keeping my bridgeunbridged.
Tim McConnaughy (28:21):
Well, honestly, where else could you have gone
and get this kind of?
Erika Dietrick (28:25):
Yes.
Tim McConnaughy (28:26):
Experience, even if it is stressful at the
time, Like there's very fewplaces that you could you know,
most places would crib you tothe point where you barely learn
anything for a year.
Erika Dietrick (28:35):
You know, I will say that too.
You know it was a cluster.
But you know, had I been at abig I don't know, a fortune 100,
you know I would have been.
You know I do, you know, workthis one little gear in this
gear only.
So I did get to touch a lot.
Tim McConnaughy (28:47):
It's good after the fact.
It's good for building theportfolio.
It's good for building theportfolio.
Chris Miles (28:52):
Glad you did it, but you wouldn't sign up for it
again.
Erika Dietrick (28:54):
Well, you know, and things happen too right.
I mean in startup world, likeyou have venture capital, you
know, like at Cisco.
I mean obviously it matterswhat we do, deadlines matter,
but at the end of the day, youknow, in startup world you
really can't afford to do that.
That's the death of yourcompany and so on top of that,
(29:24):
you're usually riding on venturecapital.
You usually aren't actuallyprofitable for I don't know
years.
You know many startups outthere are not actually
profitable.
Tim McConnaughy (29:32):
No, that's well .
Well, yeah, agreed.
Erika Dietrick (29:35):
So you know you've got this one little eager
you Agreed.
So you know you've got this onelittle eager.
You know junior engineer whoyou know.
The entire architecture, theCICD pipeline, the way it's
deployed, is riding on this.
Tim McConnaughy (29:46):
Millions of dollars.
Erika Dietrick (29:46):
Six-month-old runway.
And then you know, oh, why isthe service down?
You know emergency.
You know, hello, I don't know.
I'm not going to pretend I canfigure it out.
Tim McConnaughy (29:56):
Yeah, no, that lesson's learned Everybody.
Honestly, the best lessons arelearned in the fire, but as long
as it doesn't actually likemake you want to leave tech
forever, you usually come outthe other side with the kind of
experience that would take youyears otherwise to get.
So there's good and bad to it.
It is a very high-stressenvironment.
Somebody was asking in Discordyesterday you know, hey, I want
(30:18):
to get into tech, but what's thework-life balance and the
stress level like?
And I'm like, do you want toknow the answer?
Do you want to get into tech?
Because I'll tell you Like, soanyway, but that's.
Erika Dietrick (30:30):
You know, but it does get better though, because
it does.
So I'm I have, I will bestarting year six in tech this
year.
I mean, you don't feel like youget any wins Like you feel like
you know nothing.
It's not just oh, you know, I'mgood at some things and I get
that self-esteem boost and I'malso learning and getting
(30:52):
knocked back down again.
It's just getting hit in theface over and over and not
knowing.
So I mean, I feel like thatpart does change and I think
your attitude also changes aboutit.
Right?
Because again, university, youknow, as a developer we're not
even I mean, at least I wasn'tin many developers I know you're
not even taught how to debugyour code, which is highly
(31:12):
unrealistic, right.
Tim McConnaughy (31:13):
Oh really, no unit testing or anything.
That's not part of thecurriculum.
Erika Dietrick (31:17):
So I learned how to do unit tests and how to use
certain mocking frameworks.
But I mean, I think that someprofessors, you know, they kind
of gloss over the debugger.
It's complicated, I won't get.
Tim McConnaughy (31:30):
Like the step debugger where you have to go
like Right.
Erika Dietrick (31:33):
So there's this joke in programming land about
how we debug using printstatements.
I don't know if you guys arefamiliar, okay.
Chris Miles (31:40):
Yep, I write Python , so yes, that's the only way
I've ever debugged.
So yeah, so yes, that's theonly way I've ever debugged.
Erika Dietrick (31:53):
So you know it's fine in a pinch or sometimes,
but it's really not that helpfuland inefficient for real world
development.
But shockingly, it's usedconstantly.
People just do not use thedebugger and that's the real
world.
The real world is that yourcode just doesn't work.
It doesn't magically work.
You mess up all the time.
That's changing a little bitwith AI coding assistants.
Some of the really irritatingthings that you troubleshoot for
hours you know don't reallyhappen anymore, but that's good
(32:14):
news, yeah, yeah, so should we?
Chris Miles (32:16):
should we open that up a little bit?
I know, obviously, the the kindof AI coding assistants or
co-pilots, as we put them um,not to use the specific branded
item, but it's one of the mostpopular ones, so it's not like
we can avoid it.
But I mean, how do you feelthat is changing the kind of
development lifecycle?
For better or worse yeah.
Erika Dietrick (32:37):
I mean, there's a lot of AI copilots out there.
Typically, you know, the onepeople are most familiar with
are code generators like GitHubCopilot.
There's a lot of others outthere and they range from.
You know, I only generate codeand have chat functionality to.
I'm a full on developerenvironment, you know, meant to,
you know, save workflows orsuggest things to you in that
(33:00):
way.
So I will say, at least from aGitHub Copilot perspective,
which is so I will say at leastfrom a GitHub co-pilot
perspective, which is, you know,or at least was, the most
widely adopted AI codingassistant code generator at the
time.
It's what I've done, a bunch ofpresentations on.
It's absolutely revolutionizeddevelopment life cycles in both
(33:23):
good and bad ways, and I dothink that, even though bread
and butter developers wereprobably the first to flock to
these tools, that really we'retalking about something that's
going to change things foranyone who codes, and so I know
we've got some network and cloudpeople.
So I'll speak more to that AtCisco Lives and when I give
talks, the main thing I'm askedagain is you know, do I actually
(33:45):
have to learn how to code now?
Is it just a matter of promptengineering?
Um, and then I, and thenthere's people on the other side
of the fence right, who youknow you're like.
You shouldn't be able to touchan AI coding assistant until I
don't know you're a seniordeveloper or whatever.
Tim McConnaughy (33:59):
Oh yeah, the gatekeeping.
Erika Dietrick (34:21):
And what's interesting at least what I
found through my own researchand opinions is, you know, I
think that it's really going tobe not the code generation per
se, especially for networkengineers, but it's going to be
are taught and hired.
And coincidentally, this bleedsa little bit into teaching
network engineers how to code,and assessing their skill is
essentially, you know, we giveyou some assignment, we see how
you execute said assignment orhow you code it, and then are
you able to explain it right,and then we do some form of that
(34:43):
in the hiring process whenwe're evaluating people too,
right, it might be on awhiteboard, it might be solving
some stupid riddle, butessentially we just give an
assignment and ask to code andthen whatever.
So you know that is completelyblown on its head now that there
are AI coding assistants right,because you have an assignment
(35:04):
and you can generate the answerto that assignment, and you can
hit slash explain.
Literally, it's that easy slashexplain, explain to me what was
just written, explain it to meline by line, and you know
that's your student assignment,you know that's your take home
for your, for your developerinterview.
Um, so it's really not aneffective way to oh, to gauge
ability anyway, right andbecause it's not an effective
(35:24):
way to gauge ability, it poses alot of challenges for teaching.
You know, how do you know whensomebody actually knows how to
code, and how reliant should webe on these tools?
So I feel like I'm rambling alittle bit, I'm sorry.
Tim McConnaughy (35:37):
No no, no this is good.
Erika Dietrick (35:38):
This is actually really good.
I think this is important.
Chris Miles (35:42):
But it's funny because, like I've read on
social media, I've never had togo through one myself, but I've
read a lot about, you know,these software engineering
interviews where people feellike the process and the
interviewing criteria is sostrict and so over the top that
they hate the interview processBecause it's all like, oh, do
this live in front of all of us,you know, write this code, do
this thing on a whiteboard, blah, blah, blah.
And they think that it shouldbe simplified and they should
(36:11):
like, almost almost like the,the, the hiring manager should
be more trustworthy about thingsin a sense.
But now it almost sounds likefrom what you're telling me,
like it's kind of spinning backin the other direction.
Like, how do you validatepeople?
Know what the fuck they'retalking about and they haven't
just done this on, you know,through a chat bot, um, other
than doing that.
Like, where do you think itfits into that?
Erika Dietrick (36:28):
So I won't spill all the tea, but I'm going to
be presenting on this inFebruary oh, all right, yeah a
roadmap how I think that weshould now be evaluating and
teaching and all that.
But yeah, I mean, I think it'san opportunity.
You know, people see it as,like you know, ruining engineers
.
No, they're never, going tolearn blah, blah blah.
But I think it's a greatopportunity to revamp how we
(36:48):
teach people to code and how weassess it.
I think that the kind of likeyou mentioned, I think that the
model was completely broken, Imean even when I was just
learning to code, right, I meanit was I overused the word
traumatic, but it was traumatic,you know.
There were many tears shed, itwasn't fun.
I didn't learn the mostefficient or effective way.
Yeah okay, and so I think thatyou know we can focus more on
(37:10):
things like people's pastexperiences and what they
learned from it and how theywould apply that to new problems
, and talking through people'sthought processes and the way
they think about problems, right, because that's what engineers
do.
It's not about do you know Csharp or Python.
I mean a little bit.
(37:31):
It depends on the, on the levelof the role, um but it's about
can you actually think like, doyou know how to solve a problem?
Um, have you encountered enoughproblems that you can deduce
how to solve a similar problemat our company?
So you know, that is generallythe way, the direction that I
hope, interview assessments aretaken.
And again, I think that from achat perspective, you know,
(37:52):
really it's just Google onsteroids.
So developers learn throughGoogle, right, we do a lot of
teaching ourselves, but Copilotchat has been absolutely
revolutionary.
I taught myself TypeScript in acouple of weeks using Copilot
chat Absolutely phenomenal, inmy opinion, at least, based on
my work speed.
So yeah, a very long storyshort.
(38:14):
I could talk about that foreveris that it's really making us
rethink how we teach people tocode and how we assess it for
network engineers as well.
Tim McConnaughy (38:23):
What do you think about not just AI as a
code assistant, which is kind ofwhat we're really talking about
?
You know, hey, generate thiscode for me you mentioned.
Like you know, it changes theway we teach people, do you
think?
Or what do you think AI has a?
Does AI have a place in theteaching process as well, or is
(38:44):
it really better kept to theassistant level?
And if it does have a place,like where do you see that
fitting in?
Erika Dietrick (38:51):
Well, I can answer yes to both your
questions, because I it shouldalways be kept at the assistant
level.
In my opinion, it is not thebrain, and that is where most
people's disappointment comesfrom.
Again, especially if you don'tknow how to code is just like do
this for me you know, so it ismost effective as an assistant,
and that makes sense, right?
Because if it's the brain, thenwhy are we even here?
And then, yeah, 100%.
(39:12):
I think that you know it'sessentially it's going to be
akin to using a calculator formath right.
So you know whether, again, itdoesn't matter what your
background is you're going towant to learn certain
foundations.
You're going to want to learnhow to think programmatically
because that is a skill rightThinking from like a discrete,
mathematic, logical perspective.
And then again, I won't spillall the tea, but there's certain
(39:35):
concepts.
I think you should learn somebuilding blocks and then I think
from there.
I think you use your calculatorto continue learning, and to I
mean not generate all your code,but at least to continue
learning.
Chris Miles (39:46):
Yeah, no, I think that's valuable.
I mean, I think it's um, I hope.
Tim McConnaughy (39:53):
I hope you call your talk software engineering
re-imagined or something likethat or something.
But just kidding, it's alreadysubmitted.
Chris Miles (39:56):
I should have talked to you, no, but I think
it's.
I think that is a greatapproach because I feel like
that kind of takes a page out ofthe network engineering
interviews that I've done overthe past, Like whether or not
I'm, you know, the intervieweeor the interviewer.
You know, I interview a lot ofpeople and a lot of times when
I'm asking questions, I'm notasking the question to know
(40:20):
whether or not you know theanswer.
I don't care if you know theanswer.
I want to know how you reactwhen you don't know it.
Like how you react when youdon't know it.
like how like how you explain,how like and what you said.
You want to know if people haveseen enough problems to learn
like theory around yeah, whattheir thought process is, how
they troubleshoot you know, ifthey can draw correlations.
I want to see how they buildthat mind map like live, right,
(40:40):
and then it depends on whatquestions they ask me in
response.
Tim McConnaughy (40:44):
Right, that's yeah, that's how you really tell
what that's like.
What questions do they ask mein response?
Right, yeah, that's how youreally tell.
What questions do they ask youback to try to get more
information?
Because you can teach anybodynetwork engineering.
You can teach anybodyTypeScript or code or you know
whatever.
I can teach structured data,json to somebody, like all that
stuff.
It's just stuff in a book thatyou can teach someone.
You can't teach someone how tothink.
(41:05):
You can't teach someone how tohow to solve problems.
Like you know not, not really.
I mean you can, but like theactual thought process of
solving problems is somethingthat people have to kind of come
up with, figure out how itworks for them, and they have to
do it themselves.
Erika Dietrick (41:16):
Yep, 100%.
And I mean these co-pilots,right, I mean we can try to slow
them down and say, rah, no, no,keep them out of coding, but I
mean really they're going toinfiltrate every area of life
like we've.
I went to um my first black hatum this year, which is a cyber
security conference, and theywere talking about how um
microsoft has a co-pilot studionow where anybody of any
(41:38):
non-technical background cancreate their own ai co-pilot and
connect it to, like, forinstance, their hr database and
have it assist them with thingsor whatever it is.
Their profession is right.
Tim McConnaughy (41:47):
That's a great idea.
Erika Dietrick (41:50):
So there's the self-serve no code, low code
options, but there's also justall these products coming out
right.
I mean, we've co -pilot, forAzure was recently announced, so
there's no aspect of tech thatis not going to be touched by
some form of AI assistant thatcan't just parrot out facts to
you, right or lines of code.
(42:12):
So, again, I think we can seeit as a bad thing, but at the
end of the day, you knowcompanies want productivity, so
it's in your best interest tolearn it, and it's an
opportunity to hire people whocan solve problems.
Tim McConnaughy (42:26):
Yeah, I mean, and it can keep people from
having to know everything.
Like you know, for example, I'mworking on a project with
Andrew Brown for his Gen AIbootcamp, and you know we were
working on something yesterdaywhere I needed to get PyDict
data out, which I didn't evenknow.
Pydict was a thing Like it's,you know, python, interpretation
of structured data.
(42:47):
So I went and I like explainedmy problem I think I was using
Claude I explained my problemand like what I was looking to
get back from structured dataand and it it did it.
Now I can't say it's a hundredpercent correct because I don't
know it well enough, but I willsay that it's got me a lot
further along than I would havebeen had I had tried to go read
documentation and tried to learnit all myself, like so, yeah, I
(43:10):
mean, it's like the personalassistant thing is definitely
keeping you from having to learneverything.
Erika Dietrick (43:17):
Well, even just navigating bad documentation.
So I kind of I'll just talkabout it because I kind of axed
the project.
But I was interested in doing aVS code extension at one point,
um this year and, uh, you knowI was not familiar with that
development.
Right, they have their own APIthat you have to learn.
Uh, so I'm going through theAPI documentation and I'm
reading through it and at firstit doesn't make sense because
(43:42):
I'm brand new to me.
I have zero clue what they'retalking about and I was like you
know what I'm just going to askmy GitHub co-pilot.
And GitHub co-pilot just brokeit down Boom, Simple, Summarized
how the API is structured, whatAPI call I need to make to do
certain actions.
It was that easy.
So I didn't even need to reallylearn the API.
I didn't need to struggle overthe documentation or guess or
(44:06):
make the API call and realizeit's actually needs this
additional field that's requiredor whatever.
Tim McConnaughy (44:12):
Yeah, I think people are afraid, though, that
people, people are leapingstraight to oh well, then you
know I don't need then then if Iknow that, then you're going to
replace me with a robot type ofthing, and I think that's a big
.
I think the more I, the more Ilearn about what they are and
are not capable of.
It's a very large leap.
You know, if the only value youbring your organization is
knowing how a certain API callworks, then you're probably not
(44:34):
going to be around very longanyway.
Erika Dietrick (44:36):
Well, you know, I think I lost my train of
thought.
Let me see if I get it back.
I think that it's the sameconversation with automation,
right?
I mean everybody, you know.
They set their sight on somepoint in time and technology and
they're like I'm going to getgood at this, and as soon as
they get good at that, the pointhas moved somewhere else and
they have to know something else, right?
And so they think thatknowledge is useless.
(44:56):
But I mean, it's really notright.
I mean even just from theperspective of problem solving
and learning or how the brainworks, right?
I don't remember all theterminology, but you know, when
you learn something, you knowthe brain can more easily learn
related topics because it hassomething to connect it to.
Tim McConnaughy (45:13):
So that's exactly right.
Erika Dietrick (45:15):
I mean, there's many reasons.
You can go on and on about that, but I think people just get
too laser focused on you know,got to know this.
Tim McConnaughy (45:22):
Right, yeah, well, right, yeah well, the
moving goalpost thing, thetreadmill, the tech treadmill,
is well published and wellunderstood, right, we're always
on a treadmill in tech and it'sit's like the uh, you know the
old cartoons with the, thefishing pole, and it's got the,
the banana or a piece of cake orwhatever it is, the tank, we're
all, we're all just likerunning towards it forever.
Um, but that's, that's a wholeother episode.
(45:43):
Um, okay, so no, this has beengreat.
We need to go ahead and wrap up, but it's been really great
having you on talking to youabout this stuff.
Yeah, I think we covered a lotof good stuff tonight.
Erika Dietrick (45:55):
Yeah, no.
Thanks so much for having me onand for dealing with my Casper
lighting and a greatconversation you guys are doing
a great job on the podcast.
Oh, thank you.
Chris Miles (46:03):
I'll do what I can in post, but I promise all the
listeners or the watchers, Ishould say that Erica does not
look as pale-faced as her camerais made.
Tim McConnaughy (46:10):
She's not quite , as I don't look enough though
it's fairly close.
I mean, we live in NorthCarolina.
There's no way she's walkingaround with skin that pale and
still being alive.
Erika Dietrick (46:19):
I've got a lot of Irish descent, so oh, me too.
Tim McConnaughy (46:21):
Yes, yes, yeah.
So how do people, how do peoplefind you on socials?
How do people connect with you?
Erika Dietrick (46:36):
Yes, yes, so I'm on LinkedIn as my name Erica
Dietrich and I'm on X, Instagramand.
Tiktoks for as long as thatlives, Erica underscore the dev
More platforms to come out, butthat's where you can find me for
now.
Tim McConnaughy (46:44):
Yeah, we might lose two of those things before
too long.
We'll see.
Erika Dietrick (46:46):
Oh man Okay.
Tim McConnaughy (46:49):
It's crashing and burning with the tech
community right now, but anyway,okay, great, so everybody go.
Chris Miles (46:55):
Do you want to plug the talk you have coming up?
Tim McConnaughy (46:58):
What was that?
Anything, anything else youwant to plug?
Erika Dietrick (47:00):
Sure.
So I have four talks coming upat Cisco Live in February, but
the one most relevant to ourdiscussion is called Teaching
Coding Skills in the Age of AI,and I'm actually co-presenting.
It's my first co-presentationand I'm doing it with my dad
he's a university professor he'sbeen a software engineer for 25
years, so I thought it'd be anice blend of perspectives the
(47:23):
somewhat newer developer andseasoned I love it Cool.
Yeah, we're going to take youthrough, you know, the
traditional, where we werebefore and where we're headed
now, provide you with a roadmap.
And again, at the last CiscoLives I've went to.
Every single time I'mapproached by you know, managers
of network engineers saying youknow, I need a blueprint for
(47:45):
teaching my engineers how tocode and we're going to provide
that.
Tim McConnaughy (47:48):
Oh, that's awesome.
I can't wait to see it.
This is an Amsterdam, stillright it's?
Erika Dietrick (47:52):
an Amsterdam.
Hopefully it'll rotate.
If it's popular, it will.
Tim McConnaughy (47:56):
Oh, I love it.
That's a great idea.
I can't wait to catch it on thereplay.
All right, everybody.
Well, this has been the Cablesto Clouds podcast.
Make sure you follow us on allof the social medias that we
still are on, and we'll see younext week.
Bye, hi everyone, it's Tim andthis has been the Cables to
(48:16):
Clouds podcast.
Thanks for tuning in today.
If you enjoyed our show, pleasesubscribe to us in your
favorite podcast catcher, aswell as subscribe and turn on
notifications for our YouTubechannel to be notified of all
our new episodes.
Follow us on socials atCables2Clouds.
You can also visit our websitefor all the show notes at
Cables2Cloudscom.
(48:38):
Thanks again for listening andsee you next time.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!