Collaborative Innovation, AI, & Cultural Dynamics w/ Dylan DuFresne

 How can the collaboration between AI and cultural understanding create optimal performance in projects?

Join us with Solutions Architect, Dylan DuFresne, as we navigate the complex landscape of digital transformation, shedding light on the often-overlooked human elements that can make or break technology implementations. 

We reflect on the evolving intersection of electrical engineering and software, and how the collaboration of AI and cultural understanding are keys to unlocking the full potential of technology in today’s rapidly changing world. 

Support the show

__________________________________________________________________

Co-Hosts are Alicia Gilpin Director of Engineering at Process and Controls Engineering LLC, Nikki Gonzales Director of Business Development at Weintek USA, and Courtney Fernandez Robot Master at FAST One Solutions.

Follow us on Linkedin and YouTube for live videos, demos, and other content!

Subscribe to our weekly newsletter for episode updates, job announcements, and more!

Get in touch with us at automationladies.io!

P.S. - Help our podcast grow with a 5-star podcast review if you love us!

Episode Transcript

Speaker 1 (00:00):
Cool.
Okay, it is the day beforeThanksgiving, wednesday,
november 27th.
This is a pre-recorded episodeand I actually, dylan, I'm not
sure if it's going to go out inthis season or if we're going to
have to wait to publish it nextseason, but I'm really glad we
caught you before a barrage ofend of year on-site projects and

(00:22):
stuff.
I know that your schedule waspretty full for the rest of the
year, yeah, so I guess I shouldback up and say welcome to
another episode of Automation.
Ladies Folks, thank you forlistening.
And we have, myself and Allie,today with our guest, dylan
Dufresne, which we have tried acouple of times to get on and we
ended up having to reschedulebecause Allie couldn't make it

(00:45):
last time and Dylan and I endedup talking for the whole hour
anyway without recordinganything.
So to officially get thisconversation out in front of our
audience and to get them to letyou know, get to know you a
little bit better, dylan, Iguess I'll go ahead and ask our
first question, and that is canyou tell us a little bit about
yourself and how you got to bedoing what you're doing right

(01:07):
now in the automation space?

Speaker 2 (01:09):
Yeah, I guess I could talk for hours on that part, so
we'll keep it short on that.
But a lot of it happened justkind of by accident and where I
ended up Started off in a techschool, relatively local, kind
of an easy option and good place.

Speaker 3 (01:28):
I was I was stalking you.
How did you find a mechatronicsprogram?
And then I know now that you'rejust like really like software
heavy.
How did you like were you maybea computer guy before that?
Like, how did you get so deepinto the software when you
started in a mechatronicsprogram?

(01:48):
And how did you find themechatronics program, Start with
that and then go to the otherone?

Speaker 2 (01:53):
Yeah, so the funny thing with the robotics program
is the school that I went to isa pretty well-known regional one
, not like a big name in thecountry, but well-known with a
lot of the local companies, andI fell into it because my
grandparents lived in town and Ididn't have to pay rent for the
first year they had a robotprogram and robots sounded cool.

(02:19):
I was on the robotics team inhigh school, did a little bit of
CAD and stuff like that, butnothing like real industry type
stuff.
And one thing led to another.
We went through the program andwhen is this?
This is in.

Speaker 3 (02:30):
Minnesota.

Speaker 2 (02:31):
Yeah, this is in Minnesota.
Yeah, yeah, One thing led toanother got the first job there
working as a service tech,started out with an internship
and got a job there for a OEMmaking packaging machines.
While I was there we did a lotof everything.
They kind of bounced me arounddepartment to department so I
worked with the mechanicalassembly guys, the electrical

(02:54):
assembly panel shop.
I worked with the controlsengineers, the electrical
designers, startups andtroubleshooting calls, things
like that for about a year,Ended up getting laid off from
that job.
That's always fun, oh yeah,especially right out of school,
and I did the dumb thing andjust bought my first real car

(03:15):
and it was a fun time.
It didn't last long, Turns outit was an in-demand skill set.
So about a month after I gotlaid off I think I had like six
or seven job offers.
In that time it wasn't like uhthat's a sell.
That's a sell for us, yeah andended up getting a job as an

(03:35):
electrical drafter with aelectrical contractor so still
not really software.
Um, ended up going and startedto do, just because that came up
, started doing some controlsprojects with them, got into one
really, really big one way inover my head, sweet um.
It was probably 50 or 60different pieces of equipment we

(03:56):
were controlling I think it was15 and a whole grain elevator
into feed mill situation pl PLC,scada project and that was a
good start for your project.

Speaker 3 (04:06):
A bunch of cabinets.

Speaker 2 (04:08):
All MCC, yep, nice.
So we had one panel and a bunchof MCC sections, a bunch of
VFDs and feeder screws.
We had a 400-horse VFD controland hammer mill so that we could
ramp it up and take care of theinrush current a little bit and
then rotate the direction everytime you started it up.
It was a lot for the firstproject.

(04:30):
We'll put it that way To thispoint I had only troubleshot
ladder logic, I had never doneanything real in a PLC.
Yet, baller, that's how youlearn.

Speaker 3 (04:36):
By fire.

Speaker 2 (04:38):
Oh yeah, that project also, ironically, was the first
time I had heard of Ignition.
It was a Rockwell PLC atIgnition and that was my
introduction to the industry,early days of Ignition.
And then from there thecontrols team that did exist all

(04:58):
ended up quitting or gettingfired.
So then for a while there itwas just me and I was taking
care of all the service callsand everything across the pretty
much across the Midwest.
So one day I'd be in prettymuch South Dakota and the next
day I'd be on the other side ofWisconsin and the next day after
that I'd be somewhere in Iowadoing a service call, or

(05:19):
Nebraska and just bouncingaround driving all over the
Midwest.

Speaker 3 (05:23):
Would you notice that when we would like?
When you go out to dinner,right, you see, like other
people that are completely bythemselves and you're just like
I don't know, you like nod atthem because you're like, you're
here in business and you don'tknow anyone, do you?
It's really easy to spot thosepeople because you are them oh,
yeah, definitely and yeah.

Speaker 2 (05:43):
So through that troubleshooting I definitely
learned a lot, but that wasagain still mostly PLCs, HMIs,
some SCADA stuff and a lot ofinstrumentation and working very
closely with electricians.
So a lot of focus on theinstall side as well, but
learning everything I mean.
There was one project we did.
It was a pretty small project,adding a bin to a grain site,
and the owner said, hey, we gotthis job, Go do it, Figure it

(06:07):
out.
And my first question was allright, what's Profibus?
And his response was you'llfigure it out, Gave me the keys
to the truck and sent me on myway.
Fantastic.

Speaker 3 (06:19):
And you did, didn't you?
Oh yeah, I did.

Speaker 2 (06:26):
I don't know how I did it at the time, but I always
figured it out one way oranother.

Speaker 3 (06:28):
That's what I love about this job and then, yeah,
that company.

Speaker 2 (06:31):
It grew from that where I think I could be wrong
on my numbers, but when Istarted there there was less
than 30 employees for sure, andthen when I left that company
about eight years later, therewas almost 300.
So then I got through theentire phase of that integrator.
About eight years later therewas almost 300.
So then I got through theentire phase of that integrator.
That was.
Their primary business waselectrical services, power
services, things like that, somesolar work and all that.

(06:51):
And then we had the integratorwithin that and, yeah, so I
learned a lot watching thatcompany grow from the inside,
seeing how things worked.
I mean, I went from the drafterwho got thrown on a PLC project
to, by the time I was done, Iwas the manager of the entire
engineering team and saw thatalong the way, made my own
mistakes along the way andlearned a lot on that side.

(07:14):
And over the time there, thedifferent projects we were on,
we got more and more into SCADA.
I accidentally fell into MESjust because that was customer
request.
I didn't know the word for ituntil we were already done with
the project, but they just asked, hey, can we track this.
Can we log that?
Let's put a database in forthis and we need to be able to
track this stuff and schedulework orders and all that.

(07:35):
And then, a few months afterthat project, I learned what MES
was and I'm like, oh, I'vealready done this, that's
hilarious.
All at MES in their industry Yep, um, that's hilarious.
All at mes in their industryyep, um.
And yeah, so through all that,I mean I pretty much got
everything from hands in panelsinstalling sensors and
troubleshooting the installationall the way through all the

(07:56):
starting to get into thesoftware and the mes side of
things and pretty much everylayer of the stack through there
, integrating with the erpsystems and as the company grew.
Those are the jobs we got.
So I just kind of went wherethe work did and learned what I
needed to to keep up and thenmanaging the team and all that
was a whole different story, awhole other skill set to learn.

(08:16):
And yeah, but along the way,the big one for me was where I
really enjoyed it was thesoftware side.
That's where I was interactingmore with people and not just
sitting by myself in a controlroom, and for me personally that
was kind of what I enjoyed morethan anything.

Speaker 3 (08:38):
The police.

Speaker 2 (08:41):
They'll pass, you'll be able to hear me again, yeah.
So through all those differenttypes of projects I kind of
learned what I enjoyed, more Igot I'm still probably I'm still
really good with and I justcan't seem to get away from the
plc side of stuff.
But I really enjoy the softwareand the people aspect of it
really more so than any of that,and being on the software side

(09:04):
and almost even to a point ofmore like consulting work and
things like that.
That's where I really kind offound my enjoyment.
So the last couple of years hasbeen spent really trying to
find that niche and figure outwhat I want to do going forward.

Speaker 3 (09:20):
Very cool very cool.

Speaker 2 (09:25):
So, yeah, I mean I could go round and round at a
lot of this, but I don't know ifyou've got more questions or
ali, I'm gonna give you theopportunity to ask follow-up
questions, otherwise I'll divein.

Speaker 1 (09:36):
But I I like tried to take a screenshot just now and
I did something funny with mycomputer, where everything is
dark relatively dark I can stillsee it, except for your face.
Dylan is nice and bright, but,like when I change my screens,
that that little rectangle stays.
So I'm in this stage of my liferight now where I'm seem to be
in between being good atanything, including this, um,

(10:01):
but aside from from thosedistractions, uh, I'm I'm
curious.
Well, ali, what do you?
Do you have any follow-ups tothat, or do you want me?

Speaker 3 (10:08):
to um, yeah, like how how did covid go?

Speaker 2 (10:13):
for us it was probably our busiest season.
Interesting, um, it might havebeen the phase of growth in the
company.
It might have been more workcoming in because things were
kind of shut down otherwise.
But I mean again, our primarybusiness was electrical
contractors were you travelingduring that time?
Oh yeah, I spent pretty muchthe entire winter of I believe
it was 2020 in arizona, back andforth, and I think I flew more

(10:36):
during covid than I have beforeor since interesting like
reverse oh yeah, are you tiredof traveling?
uh, sometimes it comes in waves.
If I don't travel too long, Iwant to travel again.
If I travel too long, I want tobe home.

Speaker 1 (10:53):
It's hard to find that balance, but do you think
it has something to do with thenature of the work, the fact
that a lot of us, I think thatare drawn to it like need that
variety somehow?

Speaker 2 (11:04):
I think so the people , people, um.

Speaker 3 (11:10):
So so when did you first like get into anything
with erp?
So it looks like mes.

Speaker 2 (11:17):
It's like you went from skater into mes on accident
and then yeah so erp was kindof part of that accidental mes
piece where we were interfacingwith, like recipe systems and
reporting actual run reports andthings like that back into the
ERP system.
So we would have a couple ofdifferent side ones, some big,

(11:38):
real ERP systems, and at thetime most of it was a file
transfer, some legacy systemwhere we didn't really have good
access to stuff, but some of itwas like custom homebrewed
Microsoft access systems thatthey called their ERP functions
and various types of them.
But it was always kind offeature requests where we want

(11:58):
to do this and we'd figure outhow to do it.

Speaker 3 (12:03):
Yeah, how to do it.
Um, yeah, it's really cool thatyou, that you like started your
career kind of with theelectricians.
Um, because I was going to askyou, like on the people side,
like, have you seen, you know,implementations of you?
Know, I guess industry 4.0 ordigital transformation go wrong
because nobody cares about you,the people that they're pushing

(12:26):
it onto, that have to make itsuccessful to go like to move on
or to like, yeah, after theimplementation.

Speaker 2 (12:35):
Yeah, definitely.
I mean it's all, at the end ofthe day, it's there to make
things easier for people.
I mean easier for everybody,and if they're not thinking
about everybody it doesn'treally work very well Most of
the time.
What I've worked on has begunpretty well, but that's also
because I've been embedded onthe plant floor from day one.

(12:56):
I mean there's projects whereI'll be quoting it with the
owner of a local company andthen I'll be doing the project
on the floor with themaintenance team owner of a
local company and then I'll bedoing the project on the floor
with the maintenance team.
So it's kind of it's.
It's it's kind of skewed.
My perspective there Causethat's kind of what I've seen
more than not, is the smallercompanies and the more
integrated approach where thepeople who are running the

(13:16):
business still day-to-dayinteract with the people, and
there's those smaller businessesthat really have that luxury
afforded to them.

Speaker 3 (13:24):
Yeah, because the ones that are really kicking
butt are like the Amazons of theworld and Teslas.
But you say you've worked withsmaller companies Like can you
talk to?
Like how do you, can youencourage, I guess, small
companies to like get into this,get into this connecting all of
your systems, and then what are?

(13:45):
You know, what are steps youcan take so you don't have to
like take it all at once,because I think that's what
these companies are afraid ofand they're just happy to, you
know, be at the size thatthey're at, but you know it will
come down to competitioneventually.

Speaker 2 (14:01):
Right, and I think the big one there is.
The small companies areactually at a luxury there
because it's a lot easier tostart small when you're already
small.
Oh, and it's didn't think aboutthat.
Everything's dependent onexisting systems too.
But it's also a lot easier ifyou have one brownfield site
versus 100 or 200 or a thousand,and so there's always the

(14:24):
difference between brownfieldand greenfield.
Most of the stuff's going to bebrownfield and you don't want
to rip and replace things thatare already working more often
than not.
But if we're talking moretheoretical, we're just going to
get started and some of themore greenfield type stuff it's
easy enough just to throw in asingle ignition gateway or
something and just get startedwith that.
Um, one of the some of theprojects we do I mean it's very

(14:48):
much you hear a little bit aboutthat land and expand model and
one of the projects we're doingrecently is we came in to quote
a process system mostly justpumps and valves transferring
between tanks and fill lines andthings like that, and we came
in as just the PLC and the HMIof this project.

(15:12):
We were able to work with thecustomer and to get the HMI
ended up being ignition in thisscenario, which opened the door
for a lot of other stuff, andthen throwing in a little bit of
data logging features and eventlogs and things like that.
So they had a very data-centricHMI that was logging events and
data that otherwise reallywouldn't have been possible with

(15:32):
some of the more traditionalHMI platforms.
Then, through that, I mean atthe beginning of the project,
they said you can use Ignitionbut you can't call it a SCADA
system.
We're not ready for that.
And then, by by now closer tothe end of the project, they're
looking for ms systems, they'relooking for some data collection
and enterprise wide kpis andthings like that.

(15:53):
So a lot of it is just showingthem what's possible early on,
and it doesn't have to be bigvery cool.

Speaker 1 (16:05):
I like that a lot because, like you said, you know
existing companies, likeespecially the small ones, right
, they always like don't havebudget right and maybe not
always, like some companies incertain niches are, you know,
very profitable.
But I'd say they the averagemanufacturing operation, right,
they don't have, you know,slushy budgets and they have a
lot of things to considerongoing maintenance, just

(16:27):
keeping their operation goingright.
So it can be hard to A fit inthe time, b find the right
partner and then C, likeactually overhaul anything.
So I think, just from you knowmy perspective and the things
that I've built in the past, itmakes a lot of sense to try to

(16:49):
or to be able to like startsmall, like that, right.
And so you mentioned that, yeah, you came in and you called it
the HMI, but you guys threw insome some data logging and stuff
.
Was that just on yourprerogative, wanting to be like,
hey, we want to show thiscustomer that there's more here,
and so you kind of add that tothe project, or do you try to
get them to agree up front that,hey, we should add this, we
should like and I know the waythat I would do.
It would just be like slightlyput it in there and be like look

(17:10):
, how cool this is, here'ssomething that you can do.
But like, how do you approachthat conversation?

Speaker 2 (17:14):
Well, part of it is just our base standards, and a
lot of that comes down to yes,we want to show them what's
possible.
Yes, we want to show themwhat's possible, and that is
part of it.
That's probably half of it.
But you also a project likethis, you also can't lose.
I mean, as an integrator,you've still got to make your
own money too, You've got tokeep your doors open and your
employees paid, so you can'tgive away everything for free,

(17:35):
especially as a small integrator.
Some of the bigger ones theymight be able to afford to just
do the whole $100,000 project tostart with and prove it and
then get paid later.
The smaller ones, like us, wereally can't afford that.
So we got to find ways ofgetting the little bits we can
in and then working with thepeople and the culture to really
kind of prove what's possibleand shift that and the other
piece the other half of thatpuzzle really is too, though it

(17:58):
makes our job that much easier.

Speaker 1 (18:00):
Yeah.

Speaker 2 (18:01):
I mean, one thing I learned very early on in my
career is if I start loggingwho's pressing buttons and if I
have trend logs and things likethat, I get the call that says,
hey, something happened at 2 am,we don't know what, and then we
can figure out what.
We're not sitting there waitingto do it, we're not, and so a
lot of it just comes intovarious practices.

(18:22):
We have that keep maketroubleshooting easier, make
commissioning easier, validationeasier, and so, yeah, 50%.
We definitely want to show themwhat's possible.
But the project actually getsexponentially more expensive
when we don't have these tools,because they also help our
efficiencies as an integrator,which definitely helps that
balance of how can we afford togive these things away, to show

(18:45):
the art of the possible, and allthat without having that budget
.
Because we're a smallintegrator, we use those same
tools to increase our ownefficiencies and that makes the
projects on our side that muchmore effective as well, which
kind of hits both sides andeverybody's happy.
And if they don't want it, wecan delete the database when
we're done commissioning it, butwe're still going to use it as
our own tool, just like we wanta meter or anything else.

Speaker 1 (19:11):
Right, right, yeah, okay, that's really smart.

Speaker 3 (19:16):
Let's talk about the magic of Ignition, because I
thought Ignition for the longesttime.
I just thought it was like areally cool um skater platform.
And it is not a skater platform, it is a iiot like mega giant
um and it can do so much morethan I ever thought.
But, um, kind of, maybe, sinceyou like saw it like back when

(19:40):
it was, you know, growing, canyou talk about, like, how much
more it is than a skater systemand what, and like uh, how easy
it is, and like what, what theother, because I've heard of
other platforms, likeperspective, and like I'm not
even sure you know thedifference, uh, but I know
there's like a lot more computerprogramming on that side.
But yeah, like, yeah, can youtalk on that point of, like the

(20:01):
magic disney world of ignition?

Speaker 2 (20:04):
Yeah.
So at the risk of sounding likea salesperson here, but it is
from my experience I do believeit.
So the big thing for ignitionfor me is if you strip away all
the other features, if you stripaway every module and function,
what you get in that first coremodule for it's like 1100 bucks

(20:24):
or something I have to look attheir website again for sure but
even headless just that coremodule and nothing else, gives
you the ability to connect to aplc of almost any type and brand
.
Amazing, uh, use the udts andscripting to do a lot with that.
I mean even headless.
At that base price you canconnect a SQL database, a

(20:47):
Rockwell PLC, and start loggingdata in ways that are kind of
hard.
Otherwise.
You can accomplish a lot ofthat at that small level with
stuff like Node-RED and that too, but then that doesn't scale
very well.
So then you're ripping out thework you already did rather than
building upon it.
So when we can start with evenjust that base headless ignition
, it gives us that platform tostart collecting that data and

(21:10):
then you can add on, like theweb dev module.
Now we've got an API, we canuse something else to visualize
it.
Still for relatively cheap.
You can start adding MQTT.
We can publish that to a brokerand now we've got these data
models published to a brokerwith the API and we don't have

(21:32):
any visualization in Ignitionyet.
But we're doing all the backend, all the data modeling and
connecting to the edge devices.
And the big one is there too wecan contextualize edge devices.
Maybe we've got a bunch ofsensors that are all Modbus, tcp
that aren't really part ofcontrol but they're part of how
we're monitoring the environmentand we've got a Rockwell PLC in
one corner and maybe a SiemensPLC in the next and we can
connect all three of thoseseamlessly, build up UDTs and

(21:53):
structure that data and thenfrom there we can share that
data with the rest of the world.
And we haven't even gotten toanything that's remotely HMI or
SCADA yet.
I mean, once you start adding,in the past you in the early
days it was just the visionmodule, which was a more
traditional SCADA.
It was all a Java based dragand drop screens.

(22:15):
You bind properties and it'svery traditional SCADA.
It's just a lot easier to workwith than most of them and I
have worked with most of themhere to work with than most of
them and I have worked with mostof them, um, and so from there,
it was just the easiest gatorto work with.
It was both the cheapest, theeasiest and allowed us to
deliver the best results.
And this was back in 2015, 2016probably as mature as it is um

(22:40):
and then over the years theyadded more functions, more, more
features, things like that.
Yeah, there were some bugs towork out.
As they grew, their supportkind of came and went, sometimes
as their company too, and I'msure they were growing pains on
that side.
And that's all resolved now,which is awesome.
And over the years, now we'vegot the Perspective module.
The MQTT stuff's gotten evenbigger.
The industry is really lookingfor more of this and Perspective

(23:06):
is much more of your webdevelopment type side of things.
So it's kind of a you could useit for SCADA and I think they
would tell you to.
But my personal preference tothis day is, if I'm doing an HMI
and maybe a SCADA system I'mEnvision If it's a SCADA system
that's more on the MES or IOTside, then I might use
Perspective and a lot of yourbig applications use both for

(23:29):
different purposes, maybe evenon different gateways, as we
keep scaling up and up, andPerspective is more of a
web-based interface.
It's kind of a drag and dropway to make web pages with those
same tags that we built in thatfirst version when we first got
started.
That's not to mention the somore for mes and erp for my

(23:50):
personal experience, I would sayyes, but that's also changing
as that platform is getting moremature as well.
So watching this a year or twofrom now, this could all be
different yeah, how do you seeum ai helping the systems
integrators?

Speaker 3 (24:07):
I think because it's being used for everything, but
like yeah, let's start with like, where different pieces it can
be used yeah, there's a lot ofthings we can do with ai.

Speaker 2 (24:16):
I think there's a lot of steps before we should
really get to that conversation.
As most integrators, um, if youdon't have the support and the
culture and the attitude and thetechnology and the data already
in place, it's not going to dothat much for you.
I mean, even before AI was onthis train, I went down the AI

(24:36):
rabbit hole a little bit andpulled back, but I'm much more a
proponent of let's get somebetter practices and standards,
let's make more reusable code,let's write more code generation
without AI, let's get thingsrepeatable and reusable, and
then we can start talking abouthow AI is going to help and
stuff too.
I mean, I really do like it,for I mean, it's really nice to

(24:59):
help comment code things likethat.
I use it almost every day, butI also as much as I think it's a
really, really great benefit tothe industry, I also think it's
pretty overhyped as well.

Speaker 1 (25:15):
Yeah.
I will say, my recentexperience with Copilot has left
me very underwhelmed, has leftme very, uh, underwhelmed back
where I was about a year agothinking like, yeah, this sounds

(25:35):
kind of cool, but honestly,these like the real value add
for me, for me is very low.
Um, and unfortunately, I wasthinking that a ai tool built
into a tool suite would be moreeffective than something like a
generic, like a chat gpt maybeit will be one day, because it's
inside the application andshould be then tailored to be

(25:57):
useful to the use cases of thatapplication.
Um, but I've kind of found it tobe the exact opposite, like
it's less capable than a generictool and the fact that it's
tied to the application for asuite that has multiple
different tools and they don'ttalk to each other and there's
no like recognition between them.
Um, so I can only imagine, likenot having touched any of the

(26:20):
co-pilots inside any kind ofengineering tools, that they
would be equally limited at thispoint to like very, very small
use cases, like you say, maybethe note annotating, or for us
it's like meeting notes.
But even then I can't even tellCopilot in what format I
normally like my meeting notes,like in ChatGPT.

(26:40):
It can remember, I can say, hey, I like my meeting notes
formatted this way, and it willremember that Copilot was like
oh, I have no memory there is a.
We want to do it all.
Do you want me to summarizethis email for you?
I'm like no, the email isalready two, two sentences long.

Speaker 2 (26:57):
Like I don't need a summary of this email there is
an important distinction toobetween general ai, just as a
conversation and topic, and llms, which is what we're talking
about with like copilot and chatgpt.
Yeah yeah, there's a whole lotof different specific
implementations.
Well then, just as aconversation and a topic, and
LLMs, which is what we'retalking about with like Copilot
and ChatGP.

Speaker 1 (27:11):
Yeah, there's a whole lot of different specific
implementations.

Speaker 3 (27:14):
Well then, what's machine learning All over this
AI umbrella.

Speaker 2 (27:17):
Yeah, and there's I'm not going to repeat one here
because I'll probably be wrongbecause I've seen 20 different
versions of definitions of thedifference between AI and ML.
But there is the AI side andthe machine learning side.
I mean it doesn't have to bethese big chat, gpt, llm type
models, it doesn't have to bethese big diffusions and things

(27:38):
like that.
It's just.
I mean a simple forecastalgorithm can do a lot of good
for somebody.
A simple search algorithm on adata set can do a lot.
I mean you could do before westart talking about all the LLMs
and stuff.
I see there a lot of benefitwith what the software world
would consider traditional andold school.
But what can we do with graphsand search algorithms and stuff

(28:02):
before we even really talk aboutLLMs and all that?

Speaker 1 (28:06):
And there's a lot of them what's been traditionally
called like operations research,that they use a lot of machine
learning, um, in, yeah,statistical modeling and things
like that, right, uh, anddistributions like I think this
is my uneducated or or out ofdate opinion of when I used to

(28:29):
like play more in that space.
But most people that want AI,they really want the
functionality that can beaccomplished by standard machine
learning, statistical modelingusing the right data, because
that's also going to be a lotmore likely to like, because
it's working with a specificdata set and it's trained on
that and it's a computationaltype model.

(28:52):
I'm probably saying this allwrong, but like it's actually uh
, trying to just give you theright answer, versus, or to be
trained to produce a betteranswer, um, or better prediction
, right, whereas, like the lomsare kind of they're still a bit
of a black box and they all theydo is predict, and they predict
based on large, based onlanguage, like how it sounds,

(29:13):
right.
So I think that's a big sourceof confusion for a lot of people
is that they're never, they'venever been designed to give you
anything close to accuracy oryou know anything of use other
than sounding good, which is whythose like use cases.
Okay, you can summarize anemail right like that's a pretty
decent use case and you knowthat it can be, you know, fairly

(29:34):
predictable and do fairly wellwith that.
But you really have toarchitect a whole host of models
if you really want it to doanything that gives you any kind
of like output that you canrely on.
Right, and then an lln may bethe input mechanism for reaching
that model if you want to beable to talk to it, yeah, and

(29:55):
right now.

Speaker 2 (29:55):
Well, again, I'll say it's overhyped, and even the
llm side of things, I meantechnology is moving at an
incredibly fast pace.
So if you're not thinking aboutit right now, you are probably
going to fall behind at somepoint.
But I do caution against justthrowing it at everything and
considering it's at the bestoption without looking at others
.

Speaker 1 (30:13):
I think right now, everybody should be playing
around with it a little bit,even if just to be like me, to
be like you know what, yeah,you're, you're absolutely like.
My instinct was correct thishype and this marketing does not
live up to what I actuallyexpect, but it takes me actually
going and trying it out, to belike, yes, I feel comfortable
that I know that currentcapabilities of this technology

(30:34):
don't do what I want or it's notthe right fit, right, I think
if you're not touching it at all, then you're just susceptible
to listening to the hype andthinking it's going to solve
your problems and then it reallywon't.
I think in most cases it won'tat all.
Or, yeah, yeah, like me, likeI've been kind of following from
the sidelines, knowing a lotabout the previous models and

(30:57):
implementations of AI.
Before LLMs came on the scene,I was skeptical but optimistic.
Right, I'm excited for thepotential of this but at the
same time, like I know howoverhyped the previous, you know
version of things was, evenback then, um, working on like
proof of concepts where peoplewanted ai to do all these things

(31:19):
.
And then eventually it's like,yeah, we actually, you know,
presented and solved the problemwith a, you know, random forest
, uh model, architected with acouple of other things that
really would be considered morelike basic machine learning,
right, right, um, but whatmatters is the output, like
people want the output that theywant.
How we get there doesn't reallymatter, but I think if you're
not on that side of the equation, it's way too easy right now to

(31:42):
think that current, like hypedtechnology will just all of a
sudden fix all these things thatwe couldn't do before, or on
the other side, you can just belike completely pessimistic and
not look into it at all, andthen I think you are like things
are moving so quickly thatyou're kind of going to miss the
boat because you'll have noclue what's going on when things
actually are getting to a pointwhere they are really useful

(32:03):
yeah, I mean I remember back.

Speaker 2 (32:04):
Was it two or three years ago?
I'd have to look at the exacttimeline again.
It might be close to three bynow.
But when Dolly Mini reallyfirst came out and I mean it was
just we'd around the office wewould print out pictures as
memes and throw them on each, onpeople's cubes and on the walls
and by the printer, and it wasjust fun.

Speaker 1 (32:21):
Yeah.

Speaker 2 (32:22):
But it was.
We weren't getting anything outof it, it was just office jokes
.
And then all of a sudden, chat,gpt came out of nowhere and it
was kind of cool, but it wasn't.
We weren't doing too much crazystuff with it.
And then you started learningoh, we can help you with the
code and things like that.
And in my experience it's a lotbetter on the software side of
things like if you want toprogram something in python or
something like that than it ison the plc side, just because

(32:43):
there's a larger knowledge baseout there for it to learn from
yeah um, but then it went fromthat to all what we know now,
all overnight.
I mean watching the differentimage generators come and go and
watching the technology changeand I mean, if I remember, back
to just those pictures where itcouldn't even do a face without
making it look like a swirl,that we were laughing and joking

(33:04):
about in the office, to whatwe're doing now, where it is
ubiquitous and a lot of people'slives.

Speaker 1 (33:09):
it's crazy how fast it's moving yeah, so I think we
will have.
Potentially I guess we'll stilldetermining this and, ali, you
can speak to this better than Ican we're thinking about having
an ai section.
Uh, at ot skater con right, andthat that's prompted by a
couple of people in our networkthat attended last year that are
actually working on somethingin the systems integration realm

(33:33):
using ai, and that is likebuilding specific applications
to to solve a customer problemthat they have, or a um, a
process that they see as being arepeatable part of their
business that they can offer totheir customers.
That has some value for.

Speaker 3 (33:49):
AI.
Like Wireshark with AI.
Huh, like Wireshark with AI.

Speaker 1 (33:55):
Yeah, yeah, okay, I don't know about that one, but
that would be something thatwe'd be interested in.
You know, kind of likeexploring in the community, but
with, like, what are we actuallydoing?
Don't give me the sales pitch.
Like, having been to some ofthe press conferences for the

(34:16):
you know, release of theseco-pilot things, like I just
don't see a whole lot of valuein like hearing that it's like,
okay, brass tacks, so like, canthis actually do my thing?
Or, if it can't do my thing,what can it do?
What is it actually doing?

Speaker 3 (34:30):
I think it can just do anything that we can do, but
way faster than us.
I think the idea is that it'snot going to build anything
magical, but we are the magicand if we fed it to these
systems, we can do our magicfaster.
So really it's just aboutbringing our magic to market

(34:51):
faster.
I mean, I think that's reallyall we're trying to do with the
ai right now.
Um is speed, um, because I Imean, I don't think, yeah, ai
isn't magical enough to justmake shit up, but, like we have
so many things that we can, justthat we do so slow and yeah,
it's great job security.
But, like, if we want toactually compete with, like

(35:14):
large firms bigger than us,cause we're little right, like
if we want to compete and allthese other small firms, like,
how do you leverage?
And especially because we aresmall, we can and we're like
startups and we're new, we canbring all this stuff in and
force all our people to notforce, but you know what I mean?
Like bring it, make it part ofthe culture, whereas these
giants can't do that, and so itreally can be a means of, uh,

(35:39):
competition between the littlefirms and the giants.
Um, so we can take a realmarket share if we can use this
magic.
But we have to direct the magicbecause it absolutely is
garbage in, garbage out, um, andso we we have to find the, the
magical people to to feed.
I think I mean nikki was talkingabout doing this, where it was

(36:00):
like people needed to make theirown gpts.
So it was like courtney couldmake a courtney gpt and that gpt
could like help you size arobot and like I could make you,
you could make an alley GPT andthat will help you, like I
don't know, size a pump orwhatever, do some piping, um,
just like, take all ourindividual skills and like make
these, you know GPTs.

(36:20):
And then I guess, if we could,I I really want to use it to
like quote systems, because Ihate making custom, you know,
custom control panel quotes.
It's like there's only so manydrawings, there's only so many
control narratives and like oemmanuals, there's only so many
like deliverables, right, likeon the integration side, and

(36:42):
then the installation side hasall of its own and you know, we
partner with like electriciansand then we just build something
that just like makes this somuch faster because, like we do,
we spend a lot of overhead, uh,creating these freaking
proposals, and I think some Imean other people do it better
than I do it or we do it, butlike, yeah, that's where I'm

(37:02):
like I feel like we could reallyuse it.
Um is, yeah, just sizing, sizinglike all the crap that we
already know how to size, andyeah, kind of just if we taught
it from the you know best peoplein our organizations.
But then how do you implementthat?
Because that just seems likereally hard, like, unless you

(37:25):
can take, unless the person'smind is of the type that can
like, like regurgitateeverything they know into a list
or a flow chart, which I feel Ican, but, like, I feel like
most people can't, um, and theyjust don't know how to like dump
whatever it is they want tohappen into a sequence, or into
a flow chart or into a whateverlot sequence of logic gates,

(37:48):
yeah, um, and so like, maybe, um, are we thinking?
Or into a whatever sequence oflogic gates, and so, like, maybe
are we thinking of, like, maybewe need to start a new business
?

Speaker 1 (37:57):
No, I'm just kidding.
The inverse of that also islike that is all possible
without AI.
It's just it takes building andI think maybe what the barrier
that may be being removed hereis somebody that's not technical
enough to build this into astandard, like in a standard
software type environment, or aplugin like build it drop tool

(38:18):
Could do this now potentiallyright by just dumping all that
information into a GPT and liketrying to train it.
But, dylan, is that somethingthat you guys like that you're
already familiar with, kind ofbuilding these things?

Speaker 2 (38:27):
I see you nodding like a little bit technology I'm
more nodding because thispivots back to an earlier piece.
Here there's two points off ofthat to bring up.
Is one just like anything else,whether or not we use ai, what,
what tool it is, whether it'sml or just a human process, it

(38:49):
goes back to the conversation ofwhat are we trying to
accomplish?
Then let's find a tool to do it.
Ai may or may not be that tool,and it could be a lot of times.
But if we start with aconversation with, well, I need
AI, you're not going to get theright tool.

Speaker 3 (39:02):
Okay, touche.

Speaker 2 (39:04):
So the big one for me is starting with the problem,
starting with the need, and thenfinding the tools and the
technologies that can supportthat.
And then, if we follow thatlogic back to kind of what Ali
was saying, I was really noddingalong with the point of well,
that takes the right type ofperson, who can document that,
who can think that way, who can.
And even if you take A out ofthe equation, you need that,

(39:27):
that somebody in theorganization with that skill to
get that knowledge transfer tohappen.
Because a lot of the engineersare really good at documenting
in very technical ways, or someof them are really good at
figuring stuff out, but theycan't document anything for
their life.
Some people are really good atdocumenting but they don't have
any of the technical experienceand so they're really good at

(39:49):
pulling that out of people andthen documenting it for them,
and those are all good ways ofdoing it for different people.
And that comes back to evenbefore you start talking about
the needs and the problems.
That's much larger culturalthings to think about and that's
part of where the further I gotin with the technology,
starting even with theelectricians and wires and
terminations in the field, allthe way through almost the

(40:11):
entire automation stack.
The conclusion I came to isnone.
And AI is included.
None of these technologies aregoing to solve your problems
until you've identified theproblems and identify the need
to do something about it.
That's where a lot of mycurrent time and thought is
spent on the people side and theculture, much more than it is

(40:32):
on any specific technology orimplementation, because they're
always different, but the peopleare the same, the attitudes are
the same, the culture is thesame, and it always goes back to
that question of whether I wantto put a wiki together and put
it in a Teams channel as aknowledge base.
Well, I got to get my employeesto do that.
They're not all good writers.
Some of them are really goodand some of them are good at

(40:53):
other things, and we were inthat same problem that we just
came to when we were talkingabout documenting GPTs as well.
So you have a lot of thisoverlap, and it all happens way
before we start talking abouttechnology.

Speaker 3 (41:08):
People are everything .

Speaker 2 (41:10):
Mm-hmm technology.
People are everything.
So, just like, if I mean thequestion comes down to do I need
an MQTT broker in my IOT stack?
Maybe let's talk about it.
What are you trying toaccomplish?
Do I need AI in my corporateinitiative?
Maybe let's figure out why.
But if you start saying I needan MES system or I need a SCADA

(41:32):
system or I need AI, it's goingto be a problem.
And then your digitaltransformation, your little
project, whatever the scope, isgoing to fail.
The one might be successful,the second one or the third one
you're eventually going to findyourself in a bottleneck.
You're going to have sometechnical debt.
Things are going to fall apartsomewhere along the line.
So if you really want long-termsustainable solutions rather

(41:55):
than just a one-off project,that's where these questions
come in, where it's not about Ineed an ERP system.
It's not a part.
I need MES.
It's I want a way to schedulemy work orders better, because
the people on the floor arelosing the papers.
And now we start talking abouthow can we do that?
And oh, by the way, now we havean MES system, or maybe we're
going to add some more featuresthan eventually we do, or I need

(42:17):
to know the state of thismachine that's running so that I
can calculate this KPI overhere.
All right, let's throw in someIO link sensors and maybe that
master talks MQTT, so we throw abroker in there and then that
broker goes to whatever'scalculating that kpi and all
right, yeah, cool, now we've gotsome iot going.
But that was never theintention.

Speaker 1 (42:36):
The intention should always be to solve a problem
yeah, I think we get a lot of onthe vendor side of our industry
.
We get lost in our ownbuzzwords and marketing and we
think that everybody needs oursolution.

Speaker 3 (42:50):
Right, they should well, somebody brought up to me
recently the difference betweenengineering, like selling
engineering, and sellingsolutions, and I was like, oh,
that was like an epiphany for me.
I was like, oh crap, like,because engineering doesn't, you
know, is repeatable.
Everybody can, a lot of peoplecan do engineering, but like

(43:12):
providing a solution, I meanit's like when did when did
engineering not become solutionsbased?
Like, I don't know whathappened, but it definitely
happened Because, you know, Ialways thought engineering was,
you know, we're engineeringbecause we're trying to solve
something, but that's notnecessarily what's going on.
And so you should.
I mean, at least in terms oflike, when people are looking
for solutions, they're notlooking for engineering.

Speaker 2 (43:34):
Right, and that's where I'm in.
Going back to the project I wastalking about, where we start
with ignition.
As an HMI, I mean projects likethat and I get myself in
trouble sometimes because it'snot always great for the
companies I work for usuallyends up good in the end, but
some, depending on how you'rethinking about it, right yeah Is
I'll be sitting down with thecustomer and clients and I will

(43:55):
be sitting in this meeting.
We'll be talking about thescope and what we need to do and
I'll start putting a quotetogether and I'll see, well, we
hard specced a panel view andI'll go back and said, well, you
said you want to do all this,but you told me you want to use
that.
That's not going to work.
And so then we start to kind ofthe conversation and sometimes

(44:16):
I have to just quote the panelview and it is what it is and
we'll do the job.
And most of the time it doesn'twork.
If that was what they weretalking about, if it's a
standalone little panel on theside, a little skid system or
something, there's no problem.
But if you're trying to thenconnect to that later, then the
PLC becomes a bottleneck becausenow you've got the HMI and some
other device and your IoTsystem all pulling tags out of

(44:37):
the thing, and now you just runout of memory.
The CPU can't keep up, so thenokay, well, if we know we're
going to do this data initiativeno-transcript, the plc like let

(45:18):
it go, like, oh yeah and I meanat the same time they're,
they're, they're, they're goingto argue.

Speaker 3 (45:23):
Well, our people all know how to use, like they know
how to do mer files, or theyknow how to use panel view, and
that's why we don't want to doyour solution.
We want to do what our peopleknow.
Um, I mean, how do you reallycombat that?
Or do you just put that gatewayin and you're just like fine,
we'll use your stuff, but we'regoing to make it all talk
through this and then we all win, even though you're still

(45:44):
you're spending more money butyou are making your maintenance
people happy right and there's acompromise.

Speaker 2 (45:48):
I mean, sometimes there is a very good reason for
this is what the people know andwe can't spare parts agreements
.
I mean, that's used as a verybad argument a lot of times, but
there are times where it'svalid and it's identifying those
and being able to put togetherthe team.
And again that's where I keepfalling back to the culture
piece much more in thetechnology, because you can't

(46:09):
actually affect any change forwhat I'm trying to look for and
where I'm passionate in withjust putting in the panel view.
So then it's well, andsometimes the panel view is
great.
I'm not saying that's alwaysthe bad solution, but in the

(46:33):
case where you want this bigdata architecture, you want this
other stuff and you're going toput in the panel view, but you
won't even at least put in anOPC server or something on the
side, then you start havingproblems with scalability and
maintainability.
Things really don't scale theway that, through the
conversation with them, havealready identified that they're
looking for.
And so then you guys starttalking about the bigger picture
stuff and you start talkingabout strategies and roadmaps
and how are we going to getthere?
And well, this is part of abigger thing and you lose a lot
of people there.
So then you got to startfiguring out how to feed the

(46:55):
information at the right timesand make sure things are
successful, and it just becomesa whole dance of how to.
I mean, it's all about theculture, more so than anything,
and that's a whole differentgame than the technology it's.

Speaker 1 (47:13):
That's so true, because you can take the same
technology stack and implementit in two different companies
that have two different cultures, and you'll have very, very,
very different outcomes.

Speaker 2 (47:22):
Yep, and I'll recommend a fully different
technology stack to those twodifferent cultures yeah I mean
you might have a brownfield sitethat already has plant packs
installed on everything andthere's no reason to rip it out
for data stuff, but you're notgetting much of your data out of
it.
So then maybe we don't evenlook too much at ignition right
away.
Maybe we just start with likeflow software or high byte or
something to pull data out offact talk historian.

(47:44):
Or if we're doing a site wherethey don't already have this
entrenched stuff, or maybe it isa rip and replace type thing
where the old system's notworking well or it's a legacy
thing that needed to go anyway,then we can start with something
big like Ignition and build thewhole thing from there.
But there's always differentpeople involved and as long as
there's different people, you'regoing to have different
solutions and different needsand different wants and

(48:05):
different perspectives.
Everybody's different, everyteam is unique and what they
need to be successful is unique.

Speaker 3 (48:13):
Well, I think the one non-unique thing is that
everybody needs to like focus ondata, and if you're blind to
data, you're kind of like wellwhat, I don't know what's gonna
happen to you.
I worked with a machine shopthat I will not name whose
control engineer used panelviews often and never, never,

(48:34):
used the like alarm historianand like alarm alarming, like
you know the table yep um, andso there would just be, you know
, like if there was an alarmstate, you could see it, but
there was no history of it,right.
so it was like we didn't havetime stamped alarms and I was
like what in the hell is goingon, um, and so it was just like
welcome to you know the 21stcentury.

(48:55):
No, but, um, uh, that'sincredibly important and people
don't know they need that.
They don't even know whatthey're missing if they've never
seen it, and some of that stuffis just like built into even
things like a panel view andyeah.
So there's like those standards, you know, if they don't know

(49:17):
about it, uh, data, datastandards, um, and and like how
you get like what is the generalinformation that you need to
show to like the operator of amachine, like there's minimums
and you can go really deep, um,and you can create all these you
know, like compounded kind ofsituations, um, but at the

(49:39):
minimum, you know every sensoryou need to know like is, is it
in a faulted state, like um, andor you could just not look into
that at all and not show thosenumbers, and and then you don't
know what you don't know, Iguess, yeah um, and that goes
back not just to the customerside too, but also as we're

(49:59):
doing commissioning and thingslike that, or as we get service
calls.

Speaker 2 (50:02):
So going back to do we put these things in, at whose
cost and how much are we givingaway for free?
This is part of theefficiencies that we're gaining
on our side, too is now we getthe service call for this
machine that we've warrantiedfor a year or whatever, and with
Ignition as the HMI.
One of the things that I dowith very minimal effort, now
that I've done it enough timesis we've got the data log to

(50:26):
tell you which operator waslogged into which HMI when they
pressed the button to put themachine into hand mode and
manually jog this thing, atwhich alarm showed up at that
time nice, the general equipmentstate and all the other stuff.
That and the other machinesthat might also have that right.
And once you're on that levelof stuff, I mean that's when the

(50:50):
doors really start to open.
People know what they, peopledon't know what they don't know.
They don't know.
This is even possible, letalone that it's easy.
And so the big one there is.
I mean my favorite story onthat is I was at a customer site
where they had three shifts andthe manager came in.
Manager and I both came in themorning we were still
commissioning it, but they werealready running it.

(51:10):
It was pretty much that lastweek where you're kind of
babysitting and training right.
And we came in in the morningand there had been a big problem
, a big pile up of spilt rawmaterials on the floor on third
shift and nobody knew why.
Manager just got upset andstarted yelling at the operator
this is what's going on, and allthat.
And I pulled him back and said,hey, before you start yelling,

(51:32):
let's actually figure out whathappened.
And so I pulled him in towardwe were looking at that.
This is a larger SCADA system,not so much just an HMI.
So we pulled them in front ofthe computer and pulled up the
keyboard and threw together someof the ad hoc trending tools we
had and let him drag and droptags and history into a chart
and some trends and all that.
And we were able to identifythat the operator actually did

(51:52):
exactly what they were trainedto do in that scenario and he
realized they probably shouldn'tbe trained to do that.
So I actually don't do thatanymore.
That's on me.
And now nobody was screaming atanybody, nobody was angry, and
everybody got better for it, andthat's not an, that's more
that's.

Speaker 3 (52:11):
That's a life jab at me because I always spiral and
really I just need to not dothat.

Speaker 2 (52:17):
It's like calm down let's just let's just figure it
out, yeah but even I mean Iguess I don't know that about
you or not, but specificallylike if you had the option to
come in, you saw something waswrong and you knew the data was
there.
Now that gives you the chanceto go yeah absolutely, even if
you were right now you've gotthat cool down period where

(52:37):
you're no longer hotheaded, justjumping into the situation.
You've had 20 minutes removedtrying to solve it and identify
it, but also now you know whathappened.

Speaker 3 (52:45):
whether it was what you thought.
Absolutely, I've never notregretted it.
It's just something that's hardto control.
But yeah, no, the data.
The data speaks way louder thananything else and that's why
you need to have the data.
And yeah, I've, you know,through my career, have, you
know, let others pressure me.

(53:06):
You know, like project managersand like other people in the
project that have a, you knowthey want to save money on a
project.
So it's like they make decisionsand or they push decisions onto
the designers that aren't thebest decisions, and I've watched
, you know, the repercussions ofthat, and so it's just, you
know, nice to see that like, ohcrap, Like I've seen we're not

(53:30):
going to put limit switches inon our valves.
So so when we go to, like youknow, and we had like
temperature control valves andwe're like, well, it's sending
out, like full open for thecooling valve, and like you go
there and you just hear yep, andI'm like, no, it's because air
is like not being delivered,like, yes, the solenoids open,

(53:53):
uh, but there is, yeah, the airwas not connected.
Somebody disconnected air.
And so, if you don't know that,by having all of the proper
feedback, feedback costs moremoney because it, because it's
usually another set of wires oranother wire, um, and so that's
where people are like, well,let's use you know as much as we
can.
You know communication orremote ios, so we can do you

(54:15):
know less, less wiring orwhatever.
But yeah, I don't know where Iwas going with that.

Speaker 2 (54:22):
I'll make an assumption where you're going
with that and correct me if I'mwrong.
But I mean that data just isn'tfor the big MES data type, erp,
digital transformation things.
We need more data for thecontrol systems too.
I mean I've been there whereyou've got the projects, like
you just said, where you don'thave positions on the valves and
the operator's saying, well,the valve's open, why isn't it

(54:43):
open?
Screen says it is.
You're clearly wrong and I'mlike, no, did you turn on the
air?
And yeah, whole thing'savoidable by just putting the
switches on.

Speaker 3 (54:50):
Yeah, the screen is it got an orange or a green
valve because the signal is on,but that doesn't mean that it's
actually turned on.
It just means you're trying toturn it on.
The command is on.

Speaker 2 (55:02):
Right, Green doesn't.
If it's green or white orwhatever your color scheme, that
doesn't mean that it's on.
It means that we're assumingit's on.

Speaker 3 (55:09):
Yeah, it means that we're trying to turn it on.

Speaker 2 (55:12):
And then I mean it goes to the same thing.
Do you have auxiliary contactscoming into inputs off your
e-stops?
Do you have disconnects?
Do you know which disconnect istripped, or is the motor just
going to not run?
So now we're telling it to runand it's not running.
Well, do we assume it's a badmotor because or, as we're
troubleshooting, we find thedisconnect?

Speaker 3 (55:31):
was off, and how much time was wasted there.

Speaker 2 (55:37):
And even outputs on cards can go out, or inputs,
which is crazy.
And so it's.
The data isn't just for the bigpicture IoT world and the
digital transformation.
We need more data to makebetter control systems too,
before we even talk aboutlogging in and doing stuff with
it outside of controls.
And again that goes back to theculture piece, because if the
bottom line says, well, we cansave money by pulling, not
pulling these two wires, andthen a year later they realized
how much more time they'respending in maintenance and

(55:58):
troubleshooting because theydidn't have those wires, and I
don't think anybody in thecontrols world has ever not had
this thought of, well, weweren't consulted early enough
in the project.
But it's true, yeah, becausenobody thought about the
ramifications of taking outthose wires because they didn't
know it's not well.

Speaker 1 (56:13):
That's why I think it's so unique and so valuable
to have someone like you, withyour background, or Allie in the
sense of like, where the timethat she spent, like with the
electricians and in thesedifferent types of like,
different areas of the business,right, because you start to
realize that you should think ofthings that other people that
just haven't had that experiencewith it, right, and I know we

(56:35):
don't have much time left sopeople that just haven't had
that experience with it right,and, uh, I know we don't have
much time left, so, um, I'llmention a tangent path that we
won't go down but otherwisewould be a whole.
Another episode um, I noticedsome chatter in our ot skater
con uh group about, uh, I guessvlad from manufacturing hub
shared an article or a podcast,I think, think, and this was um,

(56:55):
and I know Alicia Lomas uhchimed in but about kind of
software engineering and howthat's really kind of taken over
in a lot of the greenfield andlike the startup space in
particular.
And I'm quoting something Ihaven't listened to, but the
discussion seemed to be that,you know, they're kind of
finding that, yes, people wantto do this with more traditional
software approaches and getaway from, like the old school,

(57:18):
plc stuff, right, so you canhave more software engineers
instead of, like, lateral logiccontrols people.
But they seem to be missing theboat on this success.
I think in large part and thisis again just my you know,
throwing my opinion at the wallwithout knowing much that you
have software engineers that canprogram processes all day long,

(57:39):
but they don't know whetherthere should be a limit switch
on something or whether there'smissing data input from the way
that the processes actually workor what cavitation is, or
anything.

Speaker 3 (57:50):
Yeah, and this is, I think.

Speaker 1 (57:51):
Ali and correct me if I'm wrong, but one of the
reasons why you put together OTSkatecon was exactly for this
reason to try to get all ofthose people together in the
same room and to like not somuch that everyone has to know
all the things, but you shouldknow what you don't know to a
degree and like who you need tobe thinking of and maybe who to
call if you do need thatperspective.

(58:14):
Knowing what you don't know isincredibly powerful today yeah
and except that many people likedylan around that, like you
know, you started an electricalinstallation and now you do.
You know software programmingfor these big systems like that.
I don't think that, like Idon't know any old electricians

(58:34):
and then now works in mes.

Speaker 3 (58:35):
Like that's crazy, which is really cool I mean,
naomi does too.
But like, yeah, like it's very,very, very rare to be able to
to speak both of those languages.
Like because electricians dospeak their own language.
And if you do not know how todesign motor control circuits,
good luck.
Uh, actually like arguing withthem in a commissioning job, I

(58:56):
find that right now, like withmy own engineers, I'm like they
are going to, they're going tosay a lot of things and the only
way to prove it isn't us is toprove that it is them.
So you have to be able to like,see, like look at their
circuits and be like, well, whatabout this?
And once they know that youspeak their language, like they
do, they stop with the crap.

Speaker 2 (59:17):
And and.
Once they know that you speaktheir language like they do,
they stop with the crap.
And to touch on a few of thepoints we just hit there real
quick before we run out of timeis the first one going backwards
in order here as I rememberthem.
But the funny thing for me toois I didn't just start with and
know a lot of electricians.
I was taught electricalengineering by electricians.

Speaker 3 (59:33):
That's baller.

Speaker 2 (59:34):
And that comes from a whole different perspective
than a traditional educationwould.
Yeah, then the other one is toowith the software piece, and we
could have a whole, anotherhour or two long conversation on
the nuances there.
But just to say it a little bitis again, it comes down to the
people and the needs.
I mean, I know plenty ofcontrols engineers who don't

(59:54):
have the process knowledgeeither, and so it's not a unique
thing with the softwareengineers, it's just a thing
with the industry.
You need to have both theprocess and some of the
programming knowledge, or atleast a team that has all of
that.

Speaker 1 (01:00:05):
Yeah.

Speaker 2 (01:00:06):
And going back with different needs, different tools
, different technologies fordifferent people and cultures
and things like that is you getinto, it's always going to be
you need both.
Nowadays, there's never anargument, I'll say, where you
don't need software or you don'tneed control as a traditional
engineering.
But how much of each is goingto depend on the team.
And there are applicationswhere a fully software-defined

(01:00:28):
PLC will do the job better thana regular one, and there's times
where you need the reliabilityand the safety of a more
traditional one.
And as more technology comesout, I'm sure that line is going
to get more and more gray as itgoes.
But it always comes down to,just with any other technology,
it's we can use this and that,not this, or yeah.

Speaker 1 (01:00:48):
So, yeah, my, my life motto seems to apply here, just
like it does to everything else, which is why it's still my
motto, which is the answer toeverything, is it depends.
Yep, you're like that's acop-out, but it's not in the
sense well, yeah, no, I mean, itjust feels like one, yeah, well

(01:01:12):
, yeah, and.
And the lack of a cop-out isclear if you say, well, it
depends, and now let's dig intowhat those things are, versus
just using the word it dependsas a way to avoid getting any
kind of answer I've been accusedof that myself.
I say it depends a little bittoo much, especially with
technical stuff and especiallyon forums and stuff where you
can't really have the wholestory up front yeah it's always

(01:01:35):
it's why it's so hard to havethese discussions, and I don't
want to venture into othertopics like politics or whatever
, but it's like when peopledon't appreciate nuance, then
what the heck?
Why are you even talking?
Now?
You're just like you're.
You're saying one thing and itclearly can't be like correct
and you just have childhoodtrauma at that point like, and

(01:01:58):
that's not our problem, but Ithink that's one of the things
that we struggle with, too, ascontent creators, if you want to
call it that.
Right, like we were havingthese discussions, and in public
, and the reason we did was wewere, we thought there was value
in the discussions we werehaving in small groups online,
and we were like you know what?
There's probably people thatwould love to be part of this
small group if they knew that weexisted, and vice versa.

(01:02:18):
So let's put this out there andlet those people find these
discussions or whatever we'retalking about, right, but we're
so not like focused on theheadlines or the clickbaits, or
even when people, when they wantto have an episode with us and
I love the way you put it waslike here, an episode with us,

(01:02:41):
um, and I love the way you putit was like here's a bunch of
things I can talk about and oranything.
Um, because that's also how weare because when people want to
like, oh, we want to talk aboutthis specific thing and the
headline is going to be this, Ijust feel like it loses meaning
for me.
Like, if you're trying topredetermine in a can what
you're saying, then really whatyou're doing is messaging you're
not discussing anything if allyou want to do is, like you know
, talk like throw out yourbuzzwords and not get into the
meat of actually how and why,and those things are always

(01:03:03):
messy and complicated and theydon't got into some buzzwords
when they were playing meeting.
Um, it just yeah, it's, it's.
It's tricky to then be like youhave confidence in what you know
, um, but standing next tosomeone that has like unlimited
confidence in the thing theydon't know, they can come off as

(01:03:27):
more than you because youactually care to like really
give the the better answer orthe more correct answer or the
more feasible answer, versusjust the thing that I've been
taught and like I've been guilty.
My first job out of college wasfor a, you know, manufacturer of
sensors like kians, and they'revery they sell only their own

(01:03:49):
stuff and they they're you know,japanese kind of version of
doing things is when they, whenthey come out with a new version
, they want to make sure thatthey're the best at something
right.
So every catalog that comes outof key is that I know from at
least back in my day was likeit's the fastest, the world's
smallest, the whatever.
Like a very definitive, veryflashy statement which has some

(01:04:12):
truth to it, because at the timeit was released it probably was
the smallest in its class,based on the design
considerations and like theapplication that they're
tackling.
But you know, when I come outwith that brochure two years
later and I go to my customerand I'm like it's the smallest
in the world and it's like, well, with a lot of asterisks, right
, like you start to realize, themore you learn how stupid it is

(01:04:35):
to say something like that.

Speaker 2 (01:04:37):
Most of the smallest in the world.
But if you've got a bunch ofempty space now because you
didn't need the smallest, itcomes back down to yeah, and
also it probably won't be thesmallest for very long, right?

Speaker 1 (01:04:47):
So your talking point is immediately outdated as soon
as somebody else makes asmaller one.
And, yeah, I think you knowfrom a systems integrator point
of view too, like, how do youcommunicate the value of the
fact that you know what?
We're not going to try to sellyou a solution and a can
solution that won't actually bethe best for your problem.
We have, you know, being ableto say, like that, something

(01:05:17):
that's like super and I guessthis is why I'm not in
copywriting or marketing, reallybut like you have to be able to
translate that into somethingthat sounds as flashy as as the
people that are super confidentin their, in their like major
feature that's just going tolike be the silver bullet for
everything.

Speaker 2 (01:05:34):
Yeah, and that's as you said.
Nuance and intent matters, andone of the scary parts of public
conversations like this isintent matters, and somebody
listening to this later or witha different background may or
may not pick up on that intent.
And then you said somethingthat was right for this scenario
, but that doesn't mean it'sright for every scenario, and
you can only say so much in aone-hour conversation, so you

(01:05:55):
can't cover your bases of well,in this scenario, it's this and
I like that over there, andmaybe this one instead, and
there's a lot of nuance thatit's hard to convey.

Speaker 1 (01:06:04):
Absolutely Well, I think, and for those of you that
are listening unless you're,like a first time listener to
Automation Ladies, hopefully youknow our intent and the fact
that we appreciate this nuance,and we hope that you do too.
Uh, but we would, I guess.
Yeah, let's, let's wrap this up, ali.
Do you have any last questions,since I kind of hijacked the
last few minutes here before Igo into my standard closing

(01:06:26):
question?
No, I'm good, this was awesome,thank you.
Thank you, yeah, so I guesswe'll just end with dylan.
Can you tell our listeners, um,where, if, if you know you're
open to connecting, where peopleshould find you, what they
should do, if they like what youhave to say today, either from
a, you know they want to workwith you or they just want to,

(01:06:48):
you know, be in yourprofessional circles.
Where do they find you and isthere anything that we should be
looking forward to seeing fromyou in the near future?

Speaker 2 (01:06:56):
So the first one is easiest place to find me is
going to be LinkedIn.
Other links should be availablethrough there just as easy as
anywhere else.
Yep, the if you need to get intouch and want to work with us,
things like that.
Fishbone Technical is a systemsintegrator, so we're doing a
lot of the stuff I talked abouthere.

(01:07:17):
That's what we're doing and wedo everything from machine and
process control to data-centricIoT stuff and process
implementations and consultingand things like that.
And as far as what's comingnext, I don't have too many
announcements, but 2025 will bea fun year.

Speaker 1 (01:07:38):
Okay.
So I think everybody, everybody, let's follow dylan.
Um, we are going to be seeingyou at ot skate icon and I think
, ali, we are working on, Ithink we got you to agree, we
just now.
Now we need to go implement,but to to help us with some sort
of discord, uh chat for thosethat are coming.
Um, and, yeah, I think weshould make one of the rules for
that is that you mustappreciate nuance and discussion

(01:07:59):
.
If you want to be here, yeah,it might as well be said.
You know, to be clear that, andI tried to put this caveat like
you know, we're not just goingto cancel you if you say
something wrong, but you kind ofhave to be open to the feedback
as well, like if you're goingto come in and like in any
public form or semi you knowpublic thing, because even we

(01:08:26):
all know, like private groupsare still not private.
Somebody can always share whatyou say or whatever, but it's
really important to be able toappreciate people's intent and
the nuance in the discussion.
And I think in most cases wherepeople end up fighting and you
know mudslinging and stuff, likeyou really don't need to do
that.
You probably were trying to saythe same thing in the beginning

(01:08:47):
but you just, you know, didn'ttake the time to try to get
there, to realize where, whereyou're at compared to another
person, and so many of theselike interpersonal and career
and like things that we thinkabout there are so reflected on
the same side of the technology,because we're always
integrating, right, we'reintegrating different processes

(01:09:08):
with different companies and youknow you've got the end user
and the systems integrator andall the associated contractors
and like it kind of goes hand inhand.
Like you, really you just gotto understand who you're working
with, like try to understand,like where to anchor the
conversation, what people know,what they don't know, and then
like just proceed with kindnessand the intent to get something

(01:09:30):
done of value together.

Speaker 2 (01:09:33):
Integration is a much , much larger category than just
automation, and it's easy toforget that in our field.

Speaker 1 (01:09:39):
That's true, um, and I guess it's been hard for us to
capture the in, in the name ofthe event, at exactly what it is
, because some people just lookat skater and they're like, well
, I don't do skater, this isn'tfor me, um, or you know that
sort of stuff.
So we we could change that, wecould put a motto and it's just
like the yeah, like a littletagline or something that

(01:10:01):
solutions community, that we'rea lot more inclusive than just
the, the ot and the skater uhterminology.
But anyway, we're very happy tohave you as part of our
community, dylan.
Thank you so much for the time.
I think our listeners are goingto either already know you or
be very interested in getting toknow you and following you.

(01:10:23):
And yeah, I think that's all.
And happy Thanksgiving.
I know whoever's listening tothis it will not be Thanksgiving
, but we are recording this,everybody that listens and
supports our community, and fromus you should probably expect
some more video content as wetry to shift a little bit just
from the audio side of things tovideo going into 2025.

(01:10:47):
But this one will come out onthe audio podcast either way.
So with that I'm going to stoprambling and say goodbye
everybody.
Thank you, thank you, thank you.