You can feel the pressure in trucking right now—tight margins, constant market shifts, and more work piling onto the same people. So instead of chasing hype, we sat down with Ken McLoud from Laconic Tech for a real conversation about what AI actually looks like in logistics today.
Ken is an engineer turned founder building AI tools for “dirt under your fingernails” businesses—freight brokerages, carriers, manufacturing, and heavy equipment—and he brings a practical, no-BS approach to automation.
We start with a real-world use case that just makes sense: an AI agent that monitors load boards, factors in truck availability and fuel costs, applies your definition of a “good load,” and places bids fast enough to actually win. From there, we expand into a simple but powerful mindset shift—stop asking “Where can we use AI?” and start asking “What’s the bottleneck?”
Once you identify the constraint, you map the workflow, build it into an SOP, and automate the repetitive pieces—keeping your people focused on what matters most: relationships, problem-solving, and service.
We also get real about the downsides. AI between you and your customer can hurt trust if used wrong. And it’s not free—API costs stack up quickly when systems run nonstop. Ken breaks down how to stay flexible, avoid vendor lock-in, and build smarter by mixing models like Lego blocks—using cheaper tools where possible and stronger ones where needed.
If you’re serious about applying AI in trucking, freight brokerage, and operations without getting lost in the buzzwords, this episode is for you.
Key Takeaways👇
✅ Start with the bottleneck, not the technology—solve real constraints first
✅ AI agents can monitor load boards, evaluate data, and place competitive bids in real time
✅ Turn workflows into SOPs before automating—clarity comes before efficiency
✅ Keep humans focused on relationships, exceptions, and service recovery
✅ Be careful putting AI between you and customers—it can hurt trust if overused
✅ AI isn’t cheap—token-based pricing can add up fast if not managed
✅ Avoid vendor lock-in by treating AI models like Lego blocks—mix and match tools
✅ Use lower-cost models as filters before escalating to more powerful (and expensive) ones
Interested in being a guest on the podcast? Click here to join the conversation!
Need help finding the perfect truck seat? Let us find one for you! Call 269-692-5770, email: sales@seat-specialists.com ,or fill out form and use code TR104U.
Get your Trucking Ray merch to show that you roll with Trucking Ray and That’s Delivered Podcast! click: here
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_01 (00:00):
What's going on, everybody?
It's your boy, Truckin' Ray, andwe have you guys tap into
another episode of the DastDeliver Podcast where we talk
about people moving the industryforward.
Today we're going to guess withus is bringing tech, AI, and
innovation into a space thatdoesn't always move fast but
needs to.
Ken McLeod from Laconic Tech isin the building.
(00:20):
Ken, welcome to the show.
Ken, thank you so much forjoining us today and being on
the show.
I'm really glad to have youhere.
How are you doing?
SPEAKER_00 (00:32):
Awesome.
SPEAKER_01 (00:33):
You love?
Yeah, doing really well.
I'm looking forward to spring.
Kind of kind of done withwinter.
Nice.
Yeah, man.
So uh I got a lot of listenerswho want to know what you guys
are doing over there.
Laconic, uh uh tell us a littlebit about how you got started.
How do you jump into this spaceand you know where did it all
take you?
SPEAKER_00 (00:53):
Yeah, so uh I'm an engineer by training.
I spent most of my uh careerworking in factories and normal
nine to five kind of jobs.
Uh so obviously we're dealing alot with logistics and moving
parts uh back to suppliers andvendors and customers there.
Uh I started side hustling doingcustom software just before the
(01:15):
pandemic, so like 2019, uh doingkind of back office automation
back in the pre-AI days.
And then when the AI stuffstarted exploding in like 2023,
like that just very clearlybecame a superpower that we
could add to these applicationsin order to make them way more
flexible than the old schoolif-then stuff.
(01:39):
So the the side hustling wasdoing pretty well.
I started thinking about maybe Ishould make that my full-time
gig.
And meanwhile, at the day job,uh I had survived three rounds
of layoffs, and then the fourthround of layoffs got me and kind
of made the decision for me thatit was time to take the side
hustle full-time.
And it's been the best thingthat ever happened in my career.
(02:00):
It's been uh nothing but sunnyskies since then.
SPEAKER_01 (02:04):
That's amazing.
I mean, a lot of times you youget a little fearful, but hey,
you got a little nudge to say,hey, this is where the where
you're yeah, awesome.
Uh did you always know that youend up in tech or um was there
anything that pulled you there?
SPEAKER_00 (02:20):
I'm a 90s kid and I've been a computer geek since
back when uh you know we had thebig wooden cabinet in the corner
of the living room with thegiant PC in it.
SPEAKER_01 (02:29):
Nice, nice, yeah.
Yeah, I love technology as well.
Um seemed like it was alwayssomething cutting edge coming
out, and but uh yeah, AI is justtaking everybody by storm.
Um it's a hot topic to talkabout.
Um, and so you guys are in theright space right now.
Uh, what was the first versionthat you built, or how was it
(02:50):
kind of different from what itis today?
SPEAKER_00 (02:52):
Yeah, so the the very first tool in the logistics
space that I built out uh is fora company called Coolmore
Logistics out of Memphis,Tennessee, their brokerage.
And the tool there is uhbasically watching load boards,
looking for opportunities thatlike fit their envelope to bid
on, right?
So they have a couple ofcustomers that post like
(03:15):
semi-public load boards, like aninvite-only load board sort of
situation.
And there's a lot of brokeragesand carriers who have access to
these loadboards.
So, in particular, when like ajuicy, profitable load comes
out, it gets bid on fast andthen it'll be off the board uh
because you know the customersgotta have that load moved.
(03:36):
So the first good-looking bid isprobably gonna get accepted.
So they had this problem wherein order to be grabbing those,
you'd have to have somebody likesit in there hammer and refresh
on the load board.
Uh, and then that person isn'treally doing anything else
because they're just hammeringrefresh on the load board.
And even then, it's like reallyhard to get a person to do a
(03:57):
decent amount of due diligenceon every single load when you're
looking at hundreds of loads aday.
So, what we set up is an AIsystem to watch these load
boards, and every time anopportunity comes across, it
will uh make a few API calls toget stuff like truck
availability, uh current dieselprices is obviously a big deal
(04:18):
for everyone right now.
Uh and then it goes through aseries of rules that we've
written up with them on whatsort of loads kind of fit their
envelope and are good loads forthem.
And then if that's a good load,it uses the information it got
from calling out to the APIs tocalculate a good price, and it
goes ahead and places the bid ina matter of call it a minute by
(04:40):
the time it's done with allthat.
So now without having to havesomebody sitting there hammering
the refresh button on the loadboard, uh, they just
automatically, without a humanhaving to do a thing, get bids
in on all the profitable loadsthat fit their envelope well.
SPEAKER_01 (04:57):
Nice.
So, what kind of companiesbenefit from the solutions?
SPEAKER_00 (05:01):
Yeah, so it's really everybody who has got uh
repetitive tasks like that,where you can write out uh well,
I guess what you're doing with acomputer, right?
We're not driving a truck withthis stuff.
At least we're not, anyways.
I know there's people working onthat sort of thing.
Uh anything where you couldwrite out an SOP document on how
(05:22):
we're doing this, right?
So that's a perfect example oflike that watching the load
board.
You know, you're gonna refreshthe load board all day.
When a new load comes up, you'regonna go look for available
trucks, you're gonna check forhistorical uh data along that
lane, like our what are ourrecent shipments that we've
booked along that lane cost us.
Uh, you're gonna go checkcurrent diesel prices in that
(05:45):
area.
Uh and then you're going to, solong as you're gonna check these
three criteria, and ifeverything looks good, then
you're gonna bid it according tothis formula, right?
You could imagine writing thatout as like an SOP document and
then handing that to an internand sticking them in front of a
computer and going like, okay,buddy, here's what you're doing
all day, is following.
(06:06):
So in any job where you can dothat sort of thing, we can
probably now build an AI agentin order to do that.
And then we take the people andwe move them on to more high
value stuff, like the managingby exception, where like, oh no,
we had a truck breakdown, we hada customer wasn't ready for the
(06:26):
pickup, that screws up thiswhole schedule, and now we got
to move around and shuffle allthe dominoes.
That's the sort of thing weprobably don't want in AI
handling in real time, but wefree up all of this monotonous
work so that the humans can gofocus on the like actual
emergency stuff and the buildingrelationships, right?
Like the sales, uh, both carrierand customer sales side of it,
(06:50):
where relationships matter.
SPEAKER_01 (06:52):
I like that.
Yeah, that's awesome.
You're solving problems thatpeople didn't even realize they
probably had.
So, what's one of the biggestmistakes companies make before
they come to you?
SPEAKER_00 (07:02):
Oh, this one's definitely uh you've probably
seen somebody do it, like walkinto a room in one of these
companies and say somethinglike, uh, we need to be doing
more AI, or we've got to beusing AI.
Where can we use AI?
And like this is classicsolution in search of a problem
kind of thinking, right?
It's like a hammer looking for anail.
(07:26):
Uh, you don't want to bethinking about it as where can
we use AI, because you're almostcertainly going to wind up
picking something uh that isn'tgonna show a strong ROI that's
not actually solving a realproblem.
You're just gonna find somethingthat like fits with your
conception of AI.
What you want to do instead isthink about what are the real
(07:46):
constraints of the businessright now.
So, like what's preventing usfrom hitting our goals?
Whether that's a revenue goal ora profitability goal, or uh,
we're gonna move this many loadsthis month, kind of goal,
whatever it is.
Think about okay, what's thebottleneck in this pipe that's
stopping us from reaching thatgoal?
And then once you identify whatthat is, now we can take this
(08:09):
step back and look at it and go,we've got a bunch of tools
available to us now that wedidn't have six months ago
because of how fast all thistechnology's moving.
Can we attack that bottleneckwith one of these tools in a new
way that we weren't thinkingabout before?
SPEAKER_01 (08:24):
There you go.
Nice.
Yeah, that's great.
Because uh yeah, you can waste alot of money um trying to do
things uh cheaper and then itends up being more expensive.
So um let's talk about AI.
Um, it's the buzzword,everybody's talking about it.
Uh, how does AI actually uhbeing used in trucking or
logistics today uh that you feelis uh not just hype?
SPEAKER_00 (08:46):
Yeah, so it's on workflow stuff like that first
example we were talking about,where uh it's difficult to write
that out in old school kind ofif-then programming language.
Uh because you have things likeis you know, is this let's say
you've got a particular city andyou've got to know, is it near
(09:09):
this metro area?
You know, you've got some townin Texas, you want to know like
would people consider this nearDallas?
Let's say.
Uh just knowing a straight likewhat we do in the old school
programming days, where you'dtake a distance from that uh
that lat longitude for the citythat you're looking at to the
(09:29):
center of Dallas, and you go,okay, we're gonna call near 50
miles.
So it's 53, so we'll say no,it's not near.
However, like I picked Dallasspecifically because I know it's
got huge sprawl, right?
Like 50 miles from the center ofDallas.
Anyone who lives around theredefinitely still considers
Dallas metro area because itjust sprawls out forever.
(09:51):
That's a great example of thesort of thing that's like hard
to capture if we're doing oldschool programming, right?
You're not gonna have this giantdatabase of the radius from each
city center as what counts asnear that city.
But that's exactly the sort ofstuff that LLMs are great at.
You give it, you know, you giveit a location and say, is this
near, would somebody considerthis near Dallas or Dallas
(10:13):
Metro?
And now you can, and they cantell you yes or no, and then you
can do flows based off that.
SPEAKER_01 (10:21):
Huh.
All right, that's great.
Um, so I mean you're making uhreal world use um for a lot of
people out there, so you can seethat it actually works.
Uh um doing AI right now, um,there's something I really
wanted to get into.
A lot of people are uh talkingabout AI to, you know, what what
is the effectiveness of AI?
(10:41):
You know, is that actuallyhelping companies and some red
flags that companies are wastingmoney or uh business owners are
asking to for AI to do thehiring, but yet is it actually
helping out in that area?
Um, what are some simplestarting points that small
trucking companies uh can use AIfor?
Maybe um some data points thatthey could look at to say, hey,
(11:06):
this is where AI is really beingeffective.
SPEAKER_00 (11:09):
Yeah, absolutely.
So one of the things uh to bethinking about here, like one of
the mistakes people make a lotis putting AI between themselves
and their customers.
Right.
So here, and if it's if it's atransactional sort of thing,
this might be okay.
But if we're talking about morerelationship building or sales,
(11:30):
uh like it's not gonna be asgood as a human is at building
that relationship to make a saleright now.
So that's a perfect example ofwhere you might save money on
getting rid of a sales guy andtell yourself that you're gonna
have AI do it.
But I'll guarantee if the salesguy was any good, he was gonna
bring in a lot more businessthan the AI is for now, right?
(11:53):
We're recording this in in uhearly 2026, maybe a year from
now, I'll be in these words.
Uh but it's in the situationwhere there, like in a sales
role, let's say, if the AI iseven a little bit less effective
than the sales guy, uh like thisthing is not gonna rely.
It's gonna do the exactopposite, and you're gonna wind
(12:14):
up money in the hole.
Uh so instead you want to try tofind places where, you know,
maybe like uh in that load boardexample, maybe we're dropping
the ball right now, and there'sa bunch of loads that we're not
bidding that we could bebidding, and then we implement a
system like this, and now all ofa sudden we can be bidding them.
Like even if the whole thingstopped working, we're still
(12:36):
just falling back to where weare now, right?
So it's thinking through thecost benefit uh kind of analysis
of stuff like that.
And then the other big piece iswe don't want to be thinking
about replacing peoplewholesale, like in that salesman
example that I gave, because thetech isn't there yet.
And like, boy, if you want yourteam to not support you, tell
(12:57):
them you're replacing peoplewith AI.
Uh instead, what we want to dois think about individual
workflows that people are doing.
Like, you know, like uh likewatching a load board or like
finding a carrier for a loadthat we just got tendered, or uh
finding a replacement truck forone that just went down, this
(13:18):
kind of stuff.
Identify those out as workflows,and then we can probably
automate the heck out of thoseworkflows, but that workflow
wasn't all that person wasdoing, right?
They had 20 otherresponsibilities too.
So now we just took that offtheir plate and freed them up to
do the more high value stuff.
So we're given our existing teamsuperpowers, we're not replacing
(13:40):
humans with robots.
SPEAKER_01 (13:42):
Nice, nice.
I can see the passion as well.
Um, you know, you think about umthese detectives out there
working cases and things likethat.
Do we want a AI to be solving acase?
Um, I think I listened to onetrue crime thing, and they
talked about how the humaninstinct is faster than any
computer.
So if we can keep training ourinstincts to be faster, to pick
(14:04):
up on those areas where thecompany needs to grow, where
they need to be successfulversus getting bogged down with
a lot of uh processing.
So I think you're on the righttrack.
Sounds great to me.
Um, using AI to make us betteris going to be the key into um
bringing a better product, abetter um service for our for
people to have and to provide.
SPEAKER_00 (14:24):
So yeah, another great example there I was
thinking about when you broughtup the detective example.
I was thinking of like weedingthrough uh like reams and reams
of documents, boxes of boxes ofpapers, like in an old school
detective show.
Uh another place where we canget a lot of value out of AI is
in data analysis, right?
So, like talking and logisticscompanies, just by the nature of
(14:46):
what you do, build these hugevolumes of data, right?
We know about all the loads thatwe ran, the dates they were run
on, the customers they were runfor, uh, what the costs wound up
being, what the profit was.
We just, by virtue of operating,you wind up with a huge data
set.
And what most operations aren'tlarge enough to have uh like a
(15:08):
geek on the staff who can thendive into that data and ask
things about like, hey, whichkinds of runs are actually most
profitable for us?
You know, what kind ofcustomers, what kind of loads
are we moving, what kind ofgeographies are we going to to
from?
Uh, or or maybe it's other dataabout drivers or particular
trucks, or you can think of allsorts of things you might want
(15:28):
to ask.
Uh and a small operation can'thave a geek on hand to go jump
into the data and answer that.
But almost everybody now canwire up AI to those data
sources, and now the AI can beyour on-staff data geek to go
answer those questions.
So if you want to know, uh, youknow, are runs like XYZ more
(15:52):
profitable than runs like ABC,they can go dive into the data,
uh, run some simple scripts,query the database a few times,
and come back with the answerfor you.
SPEAKER_01 (16:02):
Perfect.
I like that.
Um, man, you got to think aboutall the work you're putting into
uh for your company.
Um, what I'm sure there'sstruggles and challenges, some
lessons that we could talkabout.
Maybe people can learn from yourexperience.
Uh, what's been one of thehardest lessons you've learned
uh building this and um scalingit out for you for yourself?
SPEAKER_00 (16:23):
Yeah, so I think that the one of like the most
interesting technical challengesand one of the stuff that I and
my team have to chew on the mostis uh like everybody's used to
thinking of AI as free, right?
Because we all practically allfirst get exposed to it in one
of these free subscriptions, orI should even say subscription,
(16:44):
free membership, right?
Uh where they're limiting yourusage usually.
And when you're doing businessuse cases over the API, there's
like there's no free tier atall.
You're paying right from thefirst uh token that you're
passing through the system.
And it it's like you're paying ahandful of dollars per million
tokens, right?
And you can roughly think of atoken like a word, like a word
(17:06):
of input or a word of output.
So at small scale where you'reasking it to look at one
particular job, uh oneparticular document, it's still
pennies, it's not significant.
But then you kind of between thefree tier stuff that we're used
to and the fact that any one jobwe ask it to do is gonna be a
few pennies, you kind of getused to thinking of it as free,
(17:31):
but then it can creep up on youreally fast if you've got
something that you're doing overand over and over again, right?
If it's something that's runningall the time, all day long, and
you're pushing huge dreams ofdata through it, all of a sudden
the pennies start becoming dimesand then dollars and they start
mattering.
Now, none of these get to likehuge, you know, scary numbers
(17:52):
for a big company, but bigenough that it matters.
And then we wind up having to goin and do a little more careful
engineering about like whatwhich of this data do we
actually need to pass throughthe AI?
You know, are we wasting a bunchof tokens on stuff that doesn't
matter and doing things like canwe go find a dumber, cheaper AI
(18:13):
to act as a filter for our bigbrain, expensive AI, so that
we're only sending stuff throughthe big brain that we actually
need to, and we're kicking outall the kind of useless nonsense
stuff.
SPEAKER_01 (18:25):
Man, so break that down for somebody.
So you have uh a more of amainframe AI.
So for somebody that maybe islooking into expand or uh so you
have a source that you go to.
Um uh can you tell us a littlebit more about that?
SPEAKER_00 (18:39):
Yeah, no, so we're this is another important bit.
So we use the AIs from all themajor providers uh that you've
heard of, right?
Like Google, enthropic, open AI,uh Grok.
Uh, and we'll try really hard tonot get locked in on any one of
them.
There you go.
Because then you're locked in,like, which is never a good
(19:00):
thing.
And also, uh like thesecompanies are all spending
billions of dollars to competewith each other.
So just because one of them isthe best right now, that doesn't
mean they're gonna be the bestthree months from now or three
weeks from now when the next bigrelease comes out, right?
So all these systems are builtto very much treat the AIs like
Lego bricks so that as the newones come out, we can plug the
(19:25):
hot new AI into the system.
And the other thing that lets usdo is for particular jobs,
certain AIs are better thanothers.
SPEAKER_01 (19:35):
Yeah, isn't that weird?
I I try to explain that somepeople they're like, Well, I
don't I like you know, I likeGronk, I like uh Chat GPT, um
Gemini.
And I'm like, Well, did you didyou run it across all of them?
Because sometimes you may youcould get a different answer.
Um is it true or is it we goingdown a conspiracy lane?
SPEAKER_00 (19:54):
No, no, no.
Like they're definitely they'rebuilt different, they're like
different, you think of it likedifferent personalities or
different people, and they eachhave certain things that they're
better at.
Like one of the clearestexamples of this right now, with
the way things sit in March 26,is anything that's got to do
with visual, right?
Like where we want to feed it.
(20:14):
Uh maybe we want to feed it apicture of the cargo in the back
of a van trailer and ask it somequestion, you know, is this
loaded or empty?
Or is this secured or insecured,right?
You want to ask some kind ofquestion like that, or you're
actually feeding it the imageand asking it a question about
the image.
Uh Gemini is hands down betterthan all the other ones at this
(20:37):
kind of deal.
I've got engineering customerswhere we're feeding Gemini
images of engineering drawings,and it's doing a pretty darn
good job of understanding theengineering drawings from
looking at an image of it.
And it's like it's just head andshoulders above the other ones
right now.
But there's other stuff likefollowing long complex lists of
(20:58):
instructions, like those SOPs wewere talking about earlier,
where like it's clearly a littlebehind, let's say uh quad, which
is phonanthropic.
So it it and then like I'd sayprobably the hottest image
generation model right now tolike make new AI pictures like
everybody sees on uh socialmedia, that's probably with open
(21:19):
AI.
So right there, we've got likethree very common things that
we're doing all the time in thisworld, and like the best the
model that's best at it isdifferent for each of them.
SPEAKER_01 (21:28):
Yeah, exactly.
Um, if you get to use Gronk atall, yeah, yep.
SPEAKER_00 (21:34):
I've used Gronk quite a bit, actually.
And in that uh the project thatwe started this talking about,
that's been powered by Gronk.
SPEAKER_01 (21:43):
Wow.
Yeah, because you're you'retesting the limits of it.
And that's kind of nice to hearthat feedback.
You're doing more volume than anormal person would do.
And I just kind of um playaround with it here and there.
And um, yeah, sometimes it seemslike Like I get in an argument
with it sometimes.
It wants to do what it wants todo.
And I'm like, let me go to theother one here and we try that
(22:05):
one.
SPEAKER_00 (22:05):
Yeah, I tell you, it's not so applicable to like
the load board stuff we weretalking about earlier.
But I use grok a lot personallywhen I want to know some bit of
information, right?
Like something happened in thenews or uh or there's some rumor
about a football player orsomething, and you want to know
is this trade gonna go down.
The thing that grok is supergood at because it's so
(22:26):
integrated with Twitter or X nowis it's both very capable of and
very eager to go dig aroundsocial media and even YouTube
and uh blogs and Reddit and gofind like what the latest
information floating around outthere on the internet is.
Those are capabilities thattheoretically they all have,
(22:49):
they can do that, but Grock isboth really eager to do it and
darn good at it.
SPEAKER_01 (22:55):
Okay, that's what I was thinking.
Yeah, because uh, you know, uh Xdoesn't really hold back on free
speech, it just lets it go.
It's got the fire holes openwide open there.
So um man, you know, that'samazing.
Um uh sharing that with us.
Appreciate that.
I know a lot of people canprobably use that in their back
office or just for themselvespersonally to advance, um, maybe
(23:19):
in their career, what have you.
So I appreciate you sharingthat.
Now, also, too, um, when we talkabout the future trucking AI,
where do you see AI helping ortaking logistics in the next
three to five years?
SPEAKER_00 (23:32):
Yeah, I think it's it's this kind of stuff we've
been talking about aboutidentifying workflows that like
that we can write a good SOP forhow to do this.
So it's that kind of likenon-emergency situation in the
normal flow of things, uh, andthen automating those so that
then the humans are freed up todo what humans do best, like
(23:56):
maintain relationships withother humans, handle like
unusual emergency situations.
So when you put all thattogether, it's gonna mean that
uh companies can grow a lotbigger than they used to be able
to on a given team size.
So, like where uh you knowdoubling in size used to take
doubling your team, now we'regoing to be able to grow uh
(24:18):
without having to add so muchhead gap.
SPEAKER_01 (24:21):
Nice.
Okay, yeah, and then they canswings and these unstable
markets that we have going on.
I mean, that is that is hugeright now.
I mean, it seems like you can'tcatch a break.
Uh one month or one week,everything's changing.
Um, if you you know you thoughtyou knew what was going on
today, give it just a couplehours, right?
SPEAKER_00 (24:40):
Not this past month or two, that's what I'm trying.
SPEAKER_01 (24:43):
Wow.
Uh so what should owners andfleet uh carriers out there
companies are preparing for, youknow, they're trying to have a
projection, they're trying tolook out to the future, see
what's going on.
Is there anything coming up thatpeople are not ready for?
SPEAKER_00 (25:00):
Yeah, I mean, I think that a lot of this stuff
that we've been talking abouthere, I think most people in the
sort of normal, like dirt underyour fingernails business world
are probably not ready for thiskind of stuff that we've already
been talking about, right?
And not thinking thinking aboutall the stuff that's got to
happen in the company asworkflows.
(25:21):
And then which of theseworkflows could we write out a
standard operating procedure forand automate?
Like, granted, like nothing Ijust said there is super
cutting-edge Star Trek stuff.
Uh, but like that's where allthe leverage is in the next few
years.
And sure, there's lots ofSilicon Valley people talking
about it, but I don't thinkthere's a lot of main street
(25:41):
people talking that way yet.
SPEAKER_01 (25:43):
Yeah, I gotta kind of take some time to just hunker
down and really go at thosethose processes to make sure
they're smoother and moresimplified and more automated,
and then you'll be more agile.
SPEAKER_00 (25:57):
Uh and you're again like you're thinking about it
more like you're a businessconsultant than like you're an
AI guy, right?
You're not looking for a placeto do AI.
Instead, you're you're layingout all of the processes and
workflows that we're currentlydoing, and then you're finding
the one that's the bottleneckand saying, okay, there, how can
we attack that?
SPEAKER_01 (26:18):
All right.
So tell us a little bit moreabout the company, maybe people
that aren't aware or they'rehearing us from the first time.
Um, what is it that uh peoplecan look to you for if they were
to just uh hey, go to yourwebsite and uh look for a
service right now, today?
SPEAKER_00 (26:33):
Yeah, absolutely.
So we're a team of all Americanengineers.
Uh everybody's getting a littlebit of gray in their beard right
now.
And uh basically what we do iswe partner with companies, uh
most of the time, those kind ofmain street businesses with dirt
under their fingernails, stufflike logistics, manufacturing,
(26:54):
uh heavy equipment, these kindof companies uh in order to
identify where the opportunitiesare to build stuff like this out
and then actually build them outand implement and hold your hand
through the follow-throughprocess of actually getting the
team using these tools and thenadjusting as uh kind of the plan
collides with the real world andwe realize that we've got to uh
(27:18):
to tweak the way we were doingthings in order to get the
output that we're wanting.
SPEAKER_01 (27:22):
I like that, you know, because if you don't like
what your productivity is, youknow, let's say proofs in the
pudding.
So if you don't like the recipe,you don't like the way it
tastes, you gotta try to changeit up.
And I think engineers, man, thethings that they come up with is
beyond some things we canimagine.
So thank you guys for doingthat.
Um, what's one of the mostchallenging things that your
(27:43):
team has been able to accomplishthat you you would say or want
to share?
SPEAKER_00 (27:52):
Yeah, so I think that the a lot of times the most
challenging I mentioned the thecost thing earlier, like when
we've got something that's justtaking a fire hose of data and
winds up costing a lot more thanwe were expecting.
But we already talked about thatone.
So the the next one comes inthese situations where we just
wind up with a huge list ofrules that we're gonna follow,
(28:14):
right?
So uh you could imagine one ofthese things like bidding on a
load board.
This didn't happen in that case,but this sort of thing could
easily happen uh where you justwind up with, you know, right
now Sally's sitting in the frontoffice and she just knows in her
head all the stuff to bid on andnot bid on when it comes up on
the load board.
So if we actually manage to getall that stuff out of Sally's
(28:36):
head, you might wind up in asituation where there's 200
rules, and you know, and it'sstuff like, oh, well, you never
want to pick up from that dockon a Friday because then
everybody's leaving work andthen the truck can't get out and
you can't make the delivery.
So if we're ever picking upthere, you got to tell them that
you're picking up Mondaymorning, never Friday afternoon,
and it's like, but that's justthat one location, and there's
(28:56):
200 of these.
Uh so just like you can imagineit's super hard for Sally to
teach all that stuff to a newperson who's gonna come take
that job when we actually do getall those rules written down and
we try to pass it to an AI, it'salso pretty darn hard for the AI
to hold all that stuff in itshead and like be diligent about
(29:18):
evaluating every single one ofthose.
So there's a bunch of differentways we can uh we can skin that
cat and attack it, like breakingthem up into chunks, using
different more advanced AImodels, uh trying to reword the
list of rules in order to try tomake them more general.
But it's tackling stuff likethat is probably some of the
(29:39):
biggest challenges that we runinto.
SPEAKER_01 (29:42):
Yeah, it makes sense.
I mean, that's you hit the nailon the head there.
That's uh sounds like trucking,all those different things.
You know, driver goes throughthe wrong door.
You're wasting time, you'rewasting money.
Oh, it's huge.
Um, so man, this has been apowerful conversation.
I think a lot of people can havesome takeaways from there.
Um, before we wrap it up, Ialways like to ask um, where are
you headed next?
SPEAKER_00 (30:05):
Oh, I think it's growing this thing, right?
So I've got I've already got acouple engineers underneath me,
and we're kind of on a missionto take these real mainstream
businesses and uh and drag thempicking and screaming into the
future.
SPEAKER_01 (30:18):
Nice, nice.
I like the um, you know, wethat's we got no choice.
Um seems like that's where we'reheaded, that's where a lot of
the money is going.
Um we think about a lot of theadvancements that these uh AI
companies are gonna invest in,uh, this country.
Uh if they keep that, if theykeep those words and those
contracts, I mean, that's gonnabe a busy, busy time.
(30:38):
I mean, just think of the influxof uh revenue that's gonna be
coming through for a lot ofthese companies, that's gonna be
insanity.
SPEAKER_00 (30:47):
It'll be an exciting few years, no matter what
happens, that's for sure.
SPEAKER_01 (30:51):
Yeah, I know the world is very unstable with a
lot of things, but um, if you goback into the past, I mean that
that's instability was alwaysthere as well.
So um if people keep focused on,you know, looking ahead, like
said, keep them keep themfocused on that.
And you're gonna you're doing agreat service for for a lot of
people, a lot of companies outthere to be successful.
So thank you.
(31:12):
Yeah, so yeah.
And so where can people connectwith you guys and learn more?
Um, where would you direct themto go so that they can um not
get lost out there and and go tothe wrong place?
SPEAKER_00 (31:23):
Yeah, so the website is laconictech.com.
That's l-a c o-n-i-c t-e-c-h dotcom.
And my email is Ken atlaconictech.com.
Uh, you can also find me onLinkedIn under Ken McLeod.
SPEAKER_01 (31:38):
Yeah, I was I was looking for it.
I said, hey, there you are.
I found chance got a lot of goodcontent on there as well.
So yeah, well, thank you forsharing that.
And uh, so we appreciate youcoming on the show, Ken.
Uh, for anybody that's listeningout there that has uh something
they need to tackle, somethingthey need to make sure that gets
done, uh look for Laconic, man.
They're they're doing big thingsout there.
(31:59):
So we appreciate you coming onthe show and talking about that.
And if you you guys also enjoythat's delivered podcast, uh
please sure to subscribe, shareit with someone you know that
can use this information.
And uh, we hope that it alsohelps a lot of professionals out
there to be successful, not justin trucking, but also in life.
And so be sure to subscribe,share with somebody.
And uh, I'm your go to guy,trucking Ray.
(32:20):
That's what I like to do is toshare these stories and
successes of a lot ofprofessionals out there.
Uh so here's another episode ofLast Delivery.
What's going on, everybody?
It's your boy, Truckin' Ray, andwe have you guys tap into
another episode of the DastDeliver Podcast where we talk
about people moving the industryforward.
Today we're going to guess withus is bringing tech, AI, and
innovation into a space thatdoesn't always move fast but
needs to.
Ken McLeod from Laconic Tech isin the building.
(00:20):
Ken, welcome to the show.
Ken, thank you so much forjoining us today and being on
the show.
I'm really glad to have youhere.
How are you doing?
SPEAKER_00 (00:32):
Awesome.
SPEAKER_01 (00:33):
You love?
Yeah, doing really well.
I'm looking forward to spring.
Kind of kind of done withwinter.
Nice.
Yeah, man.
So uh I got a lot of listenerswho want to know what you guys
are doing over there.
Laconic, uh uh tell us a littlebit about how you got started.
How do you jump into this spaceand you know where did it all
take you?
SPEAKER_00 (00:53):
Yeah, so uh I'm an engineer by training.
I spent most of my uh careerworking in factories and normal
nine to five kind of jobs.
Uh so obviously we're dealing alot with logistics and moving
parts uh back to suppliers andvendors and customers there.
Uh I started side hustling doingcustom software just before the
(01:15):
pandemic, so like 2019, uh doingkind of back office automation
back in the pre-AI days.
And then when the AI stuffstarted exploding in like 2023,
like that just very clearlybecame a superpower that we
could add to these applicationsin order to make them way more
flexible than the old schoolif-then stuff.
(01:39):
So the the side hustling wasdoing pretty well.
I started thinking about maybe Ishould make that my full-time
gig.
And meanwhile, at the day job,uh I had survived three rounds
of layoffs, and then the fourthround of layoffs got me and kind
of made the decision for me thatit was time to take the side
hustle full-time.
And it's been the best thingthat ever happened in my career.
(02:00):
It's been uh nothing but sunnyskies since then.
SPEAKER_01 (02:04):
That's amazing.
I mean, a lot of times you youget a little fearful, but hey,
you got a little nudge to say,hey, this is where the where
you're yeah, awesome.
Uh did you always know that youend up in tech or um was there
anything that pulled you there?
SPEAKER_00 (02:20):
I'm a 90s kid and I've been a computer geek since
back when uh you know we had thebig wooden cabinet in the corner
of the living room with thegiant PC in it.
SPEAKER_01 (02:29):
Nice, nice, yeah.
Yeah, I love technology as well.
Um seemed like it was alwayssomething cutting edge coming
out, and but uh yeah, AI is justtaking everybody by storm.
Um it's a hot topic to talkabout.
Um, and so you guys are in theright space right now.
Uh, what was the first versionthat you built, or how was it
(02:50):
kind of different from what itis today?
SPEAKER_00 (02:52):
Yeah, so the the very first tool in the logistics
space that I built out uh is fora company called Coolmore
Logistics out of Memphis,Tennessee, their brokerage.
And the tool there is uhbasically watching load boards,
looking for opportunities thatlike fit their envelope to bid
on, right?
So they have a couple ofcustomers that post like
(03:15):
semi-public load boards, like aninvite-only load board sort of
situation.
And there's a lot of brokeragesand carriers who have access to
these loadboards.
So, in particular, when like ajuicy, profitable load comes
out, it gets bid on fast andthen it'll be off the board uh
because you know the customersgotta have that load moved.
(03:36):
So the first good-looking bid isprobably gonna get accepted.
So they had this problem wherein order to be grabbing those,
you'd have to have somebody likesit in there hammer and refresh
on the load board.
Uh, and then that person isn'treally doing anything else
because they're just hammeringrefresh on the load board.
And even then, it's like reallyhard to get a person to do a
(03:57):
decent amount of due diligenceon every single load when you're
looking at hundreds of loads aday.
So, what we set up is an AIsystem to watch these load
boards, and every time anopportunity comes across, it
will uh make a few API calls toget stuff like truck
availability, uh current dieselprices is obviously a big deal
(04:18):
for everyone right now.
Uh and then it goes through aseries of rules that we've
written up with them on whatsort of loads kind of fit their
envelope and are good loads forthem.
And then if that's a good load,it uses the information it got
from calling out to the APIs tocalculate a good price, and it
goes ahead and places the bid ina matter of call it a minute by
(04:40):
the time it's done with allthat.
So now without having to havesomebody sitting there hammering
the refresh button on the loadboard, uh, they just
automatically, without a humanhaving to do a thing, get bids
in on all the profitable loadsthat fit their envelope well.
SPEAKER_01 (04:57):
Nice.
So, what kind of companiesbenefit from the solutions?
SPEAKER_00 (05:01):
Yeah, so it's really everybody who has got uh
repetitive tasks like that,where you can write out uh well,
I guess what you're doing with acomputer, right?
We're not driving a truck withthis stuff.
At least we're not, anyways.
I know there's people working onthat sort of thing.
Uh anything where you couldwrite out an SOP document on how
(05:22):
we're doing this, right?
So that's a perfect example oflike that watching the load
board.
You know, you're gonna refreshthe load board all day.
When a new load comes up, you'regonna go look for available
trucks, you're gonna check forhistorical uh data along that
lane, like our what are ourrecent shipments that we've
booked along that lane cost us.
Uh, you're gonna go checkcurrent diesel prices in that
(05:45):
area.
Uh and then you're going to, solong as you're gonna check these
three criteria, and ifeverything looks good, then
you're gonna bid it according tothis formula, right?
You could imagine writing thatout as like an SOP document and
then handing that to an internand sticking them in front of a
computer and going like, okay,buddy, here's what you're doing
all day, is following.
(06:06):
So in any job where you can dothat sort of thing, we can
probably now build an AI agentin order to do that.
And then we take the people andwe move them on to more high
value stuff, like the managingby exception, where like, oh no,
we had a truck breakdown, we hada customer wasn't ready for the
(06:26):
pickup, that screws up thiswhole schedule, and now we got
to move around and shuffle allthe dominoes.
That's the sort of thing weprobably don't want in AI
handling in real time, but wefree up all of this monotonous
work so that the humans can gofocus on the like actual
emergency stuff and the buildingrelationships, right?
Like the sales, uh, both carrierand customer sales side of it,
(06:50):
where relationships matter.
SPEAKER_01 (06:52):
I like that.
Yeah, that's awesome.
You're solving problems thatpeople didn't even realize they
probably had.
So, what's one of the biggestmistakes companies make before
they come to you?
SPEAKER_00 (07:02):
Oh, this one's definitely uh you've probably
seen somebody do it, like walkinto a room in one of these
companies and say somethinglike, uh, we need to be doing
more AI, or we've got to beusing AI.
Where can we use AI?
And like this is classicsolution in search of a problem
kind of thinking, right?
It's like a hammer looking for anail.
(07:26):
Uh, you don't want to bethinking about it as where can
we use AI, because you're almostcertainly going to wind up
picking something uh that isn'tgonna show a strong ROI that's
not actually solving a realproblem.
You're just gonna find somethingthat like fits with your
conception of AI.
What you want to do instead isthink about what are the real
(07:46):
constraints of the businessright now.
So, like what's preventing usfrom hitting our goals?
Whether that's a revenue goal ora profitability goal, or uh,
we're gonna move this many loadsthis month, kind of goal,
whatever it is.
Think about okay, what's thebottleneck in this pipe that's
stopping us from reaching thatgoal?
And then once you identify whatthat is, now we can take this
(08:09):
step back and look at it and go,we've got a bunch of tools
available to us now that wedidn't have six months ago
because of how fast all thistechnology's moving.
Can we attack that bottleneckwith one of these tools in a new
way that we weren't thinkingabout before?
SPEAKER_01 (08:24):
There you go.
Nice.
Yeah, that's great.
Because uh yeah, you can waste alot of money um trying to do
things uh cheaper and then itends up being more expensive.
So um let's talk about AI.
Um, it's the buzzword,everybody's talking about it.
Uh, how does AI actually uhbeing used in trucking or
logistics today uh that you feelis uh not just hype?
SPEAKER_00 (08:46):
Yeah, so it's on workflow stuff like that first
example we were talking about,where uh it's difficult to write
that out in old school kind ofif-then programming language.
Uh because you have things likeis you know, is this let's say
you've got a particular city andyou've got to know, is it near
(09:09):
this metro area?
You know, you've got some townin Texas, you want to know like
would people consider this nearDallas?
Let's say.
Uh just knowing a straight likewhat we do in the old school
programming days, where you'dtake a distance from that uh
that lat longitude for the citythat you're looking at to the
(09:29):
center of Dallas, and you go,okay, we're gonna call near 50
miles.
So it's 53, so we'll say no,it's not near.
However, like I picked Dallasspecifically because I know it's
got huge sprawl, right?
Like 50 miles from the center ofDallas.
Anyone who lives around theredefinitely still considers
Dallas metro area because itjust sprawls out forever.
(09:51):
That's a great example of thesort of thing that's like hard
to capture if we're doing oldschool programming, right?
You're not gonna have this giantdatabase of the radius from each
city center as what counts asnear that city.
But that's exactly the sort ofstuff that LLMs are great at.
You give it, you know, you giveit a location and say, is this
near, would somebody considerthis near Dallas or Dallas
(10:13):
Metro?
And now you can, and they cantell you yes or no, and then you
can do flows based off that.
SPEAKER_01 (10:21):
Huh.
All right, that's great.
Um, so I mean you're making uhreal world use um for a lot of
people out there, so you can seethat it actually works.
Uh um doing AI right now, um,there's something I really
wanted to get into.
A lot of people are uh talkingabout AI to, you know, what what
is the effectiveness of AI?
(10:41):
You know, is that actuallyhelping companies and some red
flags that companies are wastingmoney or uh business owners are
asking to for AI to do thehiring, but yet is it actually
helping out in that area?
Um, what are some simplestarting points that small
trucking companies uh can use AIfor?
Maybe um some data points thatthey could look at to say, hey,
(11:06):
this is where AI is really beingeffective.
SPEAKER_00 (11:09):
Yeah, absolutely.
So one of the things uh to bethinking about here, like one of
the mistakes people make a lotis putting AI between themselves
and their customers.
Right.
So here, and if it's if it's atransactional sort of thing,
this might be okay.
But if we're talking about morerelationship building or sales,
(11:30):
uh like it's not gonna be asgood as a human is at building
that relationship to make a saleright now.
So that's a perfect example ofwhere you might save money on
getting rid of a sales guy andtell yourself that you're gonna
have AI do it.
But I'll guarantee if the salesguy was any good, he was gonna
bring in a lot more businessthan the AI is for now, right?
(11:53):
We're recording this in in uhearly 2026, maybe a year from
now, I'll be in these words.
Uh but it's in the situationwhere there, like in a sales
role, let's say, if the AI iseven a little bit less effective
than the sales guy, uh like thisthing is not gonna rely.
It's gonna do the exactopposite, and you're gonna wind
(12:14):
up money in the hole.
Uh so instead you want to try tofind places where, you know,
maybe like uh in that load boardexample, maybe we're dropping
the ball right now, and there'sa bunch of loads that we're not
bidding that we could bebidding, and then we implement a
system like this, and now all ofa sudden we can be bidding them.
Like even if the whole thingstopped working, we're still
(12:36):
just falling back to where weare now, right?
So it's thinking through thecost benefit uh kind of analysis
of stuff like that.
And then the other big piece iswe don't want to be thinking
about replacing peoplewholesale, like in that salesman
example that I gave, because thetech isn't there yet.
And like, boy, if you want yourteam to not support you, tell
(12:57):
them you're replacing peoplewith AI.
Uh instead, what we want to dois think about individual
workflows that people are doing.
Like, you know, like uh likewatching a load board or like
finding a carrier for a loadthat we just got tendered, or uh
finding a replacement truck forone that just went down, this
(13:18):
kind of stuff.
Identify those out as workflows,and then we can probably
automate the heck out of thoseworkflows, but that workflow
wasn't all that person wasdoing, right?
They had 20 otherresponsibilities too.
So now we just took that offtheir plate and freed them up to
do the more high value stuff.
So we're given our existing teamsuperpowers, we're not replacing
(13:40):
humans with robots.
SPEAKER_01 (13:42):
Nice, nice.
I can see the passion as well.
Um, you know, you think about umthese detectives out there
working cases and things likethat.
Do we want a AI to be solving acase?
Um, I think I listened to onetrue crime thing, and they
talked about how the humaninstinct is faster than any
computer.
So if we can keep training ourinstincts to be faster, to pick
(14:04):
up on those areas where thecompany needs to grow, where
they need to be successfulversus getting bogged down with
a lot of uh processing.
So I think you're on the righttrack.
Sounds great to me.
Um, using AI to make us betteris going to be the key into um
bringing a better product, abetter um service for our for
people to have and to provide.
SPEAKER_00 (14:24):
So yeah, another great example there I was
thinking about when you broughtup the detective example.
I was thinking of like weedingthrough uh like reams and reams
of documents, boxes of boxes ofpapers, like in an old school
detective show.
Uh another place where we canget a lot of value out of AI is
in data analysis, right?
So, like talking and logisticscompanies, just by the nature of
(14:46):
what you do, build these hugevolumes of data, right?
We know about all the loads thatwe ran, the dates they were run
on, the customers they were runfor, uh, what the costs wound up
being, what the profit was.
We just, by virtue of operating,you wind up with a huge data
set.
And what most operations aren'tlarge enough to have uh like a
(15:08):
geek on the staff who can thendive into that data and ask
things about like, hey, whichkinds of runs are actually most
profitable for us?
You know, what kind ofcustomers, what kind of loads
are we moving, what kind ofgeographies are we going to to
from?
Uh, or or maybe it's other dataabout drivers or particular
trucks, or you can think of allsorts of things you might want
(15:28):
to ask.
Uh and a small operation can'thave a geek on hand to go jump
into the data and answer that.
But almost everybody now canwire up AI to those data
sources, and now the AI can beyour on-staff data geek to go
answer those questions.
So if you want to know, uh, youknow, are runs like XYZ more
(15:52):
profitable than runs like ABC,they can go dive into the data,
uh, run some simple scripts,query the database a few times,
and come back with the answerfor you.
SPEAKER_01 (16:02):
Perfect.
I like that.
Um, man, you got to think aboutall the work you're putting into
uh for your company.
Um, what I'm sure there'sstruggles and challenges, some
lessons that we could talkabout.
Maybe people can learn from yourexperience.
Uh, what's been one of thehardest lessons you've learned
uh building this and um scalingit out for you for yourself?
SPEAKER_00 (16:23):
Yeah, so I think that the one of like the most
interesting technical challengesand one of the stuff that I and
my team have to chew on the mostis uh like everybody's used to
thinking of AI as free, right?
Because we all practically allfirst get exposed to it in one
of these free subscriptions, orI should even say subscription,
(16:44):
free membership, right?
Uh where they're limiting yourusage usually.
And when you're doing businessuse cases over the API, there's
like there's no free tier atall.
You're paying right from thefirst uh token that you're
passing through the system.
And it it's like you're paying ahandful of dollars per million
tokens, right?
And you can roughly think of atoken like a word, like a word
(17:06):
of input or a word of output.
So at small scale where you'reasking it to look at one
particular job, uh oneparticular document, it's still
pennies, it's not significant.
But then you kind of between thefree tier stuff that we're used
to and the fact that any one jobwe ask it to do is gonna be a
few pennies, you kind of getused to thinking of it as free,
(17:31):
but then it can creep up on youreally fast if you've got
something that you're doing overand over and over again, right?
If it's something that's runningall the time, all day long, and
you're pushing huge dreams ofdata through it, all of a sudden
the pennies start becoming dimesand then dollars and they start
mattering.
Now, none of these get to likehuge, you know, scary numbers
(17:52):
for a big company, but bigenough that it matters.
And then we wind up having to goin and do a little more careful
engineering about like whatwhich of this data do we
actually need to pass throughthe AI?
You know, are we wasting a bunchof tokens on stuff that doesn't
matter and doing things like canwe go find a dumber, cheaper AI
(18:13):
to act as a filter for our bigbrain, expensive AI, so that
we're only sending stuff throughthe big brain that we actually
need to, and we're kicking outall the kind of useless nonsense
stuff.
SPEAKER_01 (18:25):
Man, so break that down for somebody.
So you have uh a more of amainframe AI.
So for somebody that maybe islooking into expand or uh so you
have a source that you go to.
Um uh can you tell us a littlebit more about that?
SPEAKER_00 (18:39):
Yeah, no, so we're this is another important bit.
So we use the AIs from all themajor providers uh that you've
heard of, right?
Like Google, enthropic, open AI,uh Grok.
Uh, and we'll try really hard tonot get locked in on any one of
them.
There you go.
Because then you're locked in,like, which is never a good
(19:00):
thing.
And also, uh like thesecompanies are all spending
billions of dollars to competewith each other.
So just because one of them isthe best right now, that doesn't
mean they're gonna be the bestthree months from now or three
weeks from now when the next bigrelease comes out, right?
So all these systems are builtto very much treat the AIs like
Lego bricks so that as the newones come out, we can plug the
(19:25):
hot new AI into the system.
And the other thing that lets usdo is for particular jobs,
certain AIs are better thanothers.
SPEAKER_01 (19:35):
Yeah, isn't that weird?
I I try to explain that somepeople they're like, Well, I
don't I like you know, I likeGronk, I like uh Chat GPT, um
Gemini.
And I'm like, Well, did you didyou run it across all of them?
Because sometimes you may youcould get a different answer.
Um is it true or is it we goingdown a conspiracy lane?
SPEAKER_00 (19:54):
No, no, no.
Like they're definitely they'rebuilt different, they're like
different, you think of it likedifferent personalities or
different people, and they eachhave certain things that they're
better at.
Like one of the clearestexamples of this right now, with
the way things sit in March 26,is anything that's got to do
with visual, right?
Like where we want to feed it.
(20:14):
Uh maybe we want to feed it apicture of the cargo in the back
of a van trailer and ask it somequestion, you know, is this
loaded or empty?
Or is this secured or insecured,right?
You want to ask some kind ofquestion like that, or you're
actually feeding it the imageand asking it a question about
the image.
Uh Gemini is hands down betterthan all the other ones at this
(20:37):
kind of deal.
I've got engineering customerswhere we're feeding Gemini
images of engineering drawings,and it's doing a pretty darn
good job of understanding theengineering drawings from
looking at an image of it.
And it's like it's just head andshoulders above the other ones
right now.
But there's other stuff likefollowing long complex lists of
(20:58):
instructions, like those SOPs wewere talking about earlier,
where like it's clearly a littlebehind, let's say uh quad, which
is phonanthropic.
So it it and then like I'd sayprobably the hottest image
generation model right now tolike make new AI pictures like
everybody sees on uh socialmedia, that's probably with open
(21:19):
AI.
So right there, we've got likethree very common things that
we're doing all the time in thisworld, and like the best the
model that's best at it isdifferent for each of them.
SPEAKER_01 (21:28):
Yeah, exactly.
Um, if you get to use Gronk atall, yeah, yep.
SPEAKER_00 (21:34):
I've used Gronk quite a bit, actually.
And in that uh the project thatwe started this talking about,
that's been powered by Gronk.
SPEAKER_01 (21:43):
Wow.
Yeah, because you're you'retesting the limits of it.
And that's kind of nice to hearthat feedback.
You're doing more volume than anormal person would do.
And I just kind of um playaround with it here and there.
And um, yeah, sometimes it seemslike Like I get in an argument
with it sometimes.
It wants to do what it wants todo.
And I'm like, let me go to theother one here and we try that
(22:05):
one.
SPEAKER_00 (22:05):
Yeah, I tell you, it's not so applicable to like
the load board stuff we weretalking about earlier.
But I use grok a lot personallywhen I want to know some bit of
information, right?
Like something happened in thenews or uh or there's some rumor
about a football player orsomething, and you want to know
is this trade gonna go down.
The thing that grok is supergood at because it's so
(22:26):
integrated with Twitter or X nowis it's both very capable of and
very eager to go dig aroundsocial media and even YouTube
and uh blogs and Reddit and gofind like what the latest
information floating around outthere on the internet is.
Those are capabilities thattheoretically they all have,
(22:49):
they can do that, but Grock isboth really eager to do it and
darn good at it.
SPEAKER_01 (22:55):
Okay, that's what I was thinking.
Yeah, because uh, you know, uh Xdoesn't really hold back on free
speech, it just lets it go.
It's got the fire holes openwide open there.
So um man, you know, that'samazing.
Um uh sharing that with us.
Appreciate that.
I know a lot of people canprobably use that in their back
office or just for themselvespersonally to advance, um, maybe
(23:19):
in their career, what have you.
So I appreciate you sharingthat.
Now, also, too, um, when we talkabout the future trucking AI,
where do you see AI helping ortaking logistics in the next
three to five years?
SPEAKER_00 (23:32):
Yeah, I think it's it's this kind of stuff we've
been talking about aboutidentifying workflows that like
that we can write a good SOP forhow to do this.
So it's that kind of likenon-emergency situation in the
normal flow of things, uh, andthen automating those so that
then the humans are freed up todo what humans do best, like
(23:56):
maintain relationships withother humans, handle like
unusual emergency situations.
So when you put all thattogether, it's gonna mean that
uh companies can grow a lotbigger than they used to be able
to on a given team size.
So, like where uh you knowdoubling in size used to take
doubling your team, now we'regoing to be able to grow uh
(24:18):
without having to add so muchhead gap.
SPEAKER_01 (24:21):
Nice.
Okay, yeah, and then they canswings and these unstable
markets that we have going on.
I mean, that is that is hugeright now.
I mean, it seems like you can'tcatch a break.
Uh one month or one week,everything's changing.
Um, if you you know you thoughtyou knew what was going on
today, give it just a couplehours, right?
SPEAKER_00 (24:40):
Not this past month or two, that's what I'm trying.
SPEAKER_01 (24:43):
Wow.
Uh so what should owners andfleet uh carriers out there
companies are preparing for, youknow, they're trying to have a
projection, they're trying tolook out to the future, see
what's going on.
Is there anything coming up thatpeople are not ready for?
SPEAKER_00 (25:00):
Yeah, I mean, I think that a lot of this stuff
that we've been talking abouthere, I think most people in the
sort of normal, like dirt underyour fingernails business world
are probably not ready for thiskind of stuff that we've already
been talking about, right?
And not thinking thinking aboutall the stuff that's got to
happen in the company asworkflows.
(25:21):
And then which of theseworkflows could we write out a
standard operating procedure forand automate?
Like, granted, like nothing Ijust said there is super
cutting-edge Star Trek stuff.
Uh, but like that's where allthe leverage is in the next few
years.
And sure, there's lots ofSilicon Valley people talking
about it, but I don't thinkthere's a lot of main street
(25:41):
people talking that way yet.
SPEAKER_01 (25:43):
Yeah, I gotta kind of take some time to just hunker
down and really go at thosethose processes to make sure
they're smoother and moresimplified and more automated,
and then you'll be more agile.
SPEAKER_00 (25:57):
Uh and you're again like you're thinking about it
more like you're a businessconsultant than like you're an
AI guy, right?
You're not looking for a placeto do AI.
Instead, you're you're layingout all of the processes and
workflows that we're currentlydoing, and then you're finding
the one that's the bottleneckand saying, okay, there, how can
we attack that?
SPEAKER_01 (26:18):
All right.
So tell us a little bit moreabout the company, maybe people
that aren't aware or they'rehearing us from the first time.
Um, what is it that uh peoplecan look to you for if they were
to just uh hey, go to yourwebsite and uh look for a
service right now, today?
SPEAKER_00 (26:33):
Yeah, absolutely.
So we're a team of all Americanengineers.
Uh everybody's getting a littlebit of gray in their beard right
now.
And uh basically what we do iswe partner with companies, uh
most of the time, those kind ofmain street businesses with dirt
under their fingernails, stufflike logistics, manufacturing,
(26:54):
uh heavy equipment, these kindof companies uh in order to
identify where the opportunitiesare to build stuff like this out
and then actually build them outand implement and hold your hand
through the follow-throughprocess of actually getting the
team using these tools and thenadjusting as uh kind of the plan
collides with the real world andwe realize that we've got to uh
(27:18):
to tweak the way we were doingthings in order to get the
output that we're wanting.
SPEAKER_01 (27:22):
I like that, you know, because if you don't like
what your productivity is, youknow, let's say proofs in the
pudding.
So if you don't like the recipe,you don't like the way it
tastes, you gotta try to changeit up.
And I think engineers, man, thethings that they come up with is
beyond some things we canimagine.
So thank you guys for doingthat.
Um, what's one of the mostchallenging things that your
(27:43):
team has been able to accomplishthat you you would say or want
to share?
SPEAKER_00 (27:52):
Yeah, so I think that the a lot of times the most
challenging I mentioned the thecost thing earlier, like when
we've got something that's justtaking a fire hose of data and
winds up costing a lot more thanwe were expecting.
But we already talked about thatone.
So the the next one comes inthese situations where we just
wind up with a huge list ofrules that we're gonna follow,
(28:14):
right?
So uh you could imagine one ofthese things like bidding on a
load board.
This didn't happen in that case,but this sort of thing could
easily happen uh where you justwind up with, you know, right
now Sally's sitting in the frontoffice and she just knows in her
head all the stuff to bid on andnot bid on when it comes up on
the load board.
So if we actually manage to getall that stuff out of Sally's
(28:36):
head, you might wind up in asituation where there's 200
rules, and you know, and it'sstuff like, oh, well, you never
want to pick up from that dockon a Friday because then
everybody's leaving work andthen the truck can't get out and
you can't make the delivery.
So if we're ever picking upthere, you got to tell them that
you're picking up Mondaymorning, never Friday afternoon,
and it's like, but that's justthat one location, and there's
(28:56):
200 of these.
Uh so just like you can imagineit's super hard for Sally to
teach all that stuff to a newperson who's gonna come take
that job when we actually do getall those rules written down and
we try to pass it to an AI, it'salso pretty darn hard for the AI
to hold all that stuff in itshead and like be diligent about
(29:18):
evaluating every single one ofthose.
So there's a bunch of differentways we can uh we can skin that
cat and attack it, like breakingthem up into chunks, using
different more advanced AImodels, uh trying to reword the
list of rules in order to try tomake them more general.
But it's tackling stuff likethat is probably some of the
(29:39):
biggest challenges that we runinto.
SPEAKER_01 (29:42):
Yeah, it makes sense.
I mean, that's you hit the nailon the head there.
That's uh sounds like trucking,all those different things.
You know, driver goes throughthe wrong door.
You're wasting time, you'rewasting money.
Oh, it's huge.
Um, so man, this has been apowerful conversation.
I think a lot of people can havesome takeaways from there.
Um, before we wrap it up, Ialways like to ask um, where are
you headed next?
SPEAKER_00 (30:05):
Oh, I think it's growing this thing, right?
So I've got I've already got acouple engineers underneath me,
and we're kind of on a missionto take these real mainstream
businesses and uh and drag thempicking and screaming into the
future.
SPEAKER_01 (30:18):
Nice, nice.
I like the um, you know, wethat's we got no choice.
Um seems like that's where we'reheaded, that's where a lot of
the money is going.
Um we think about a lot of theadvancements that these uh AI
companies are gonna invest in,uh, this country.
Uh if they keep that, if theykeep those words and those
contracts, I mean, that's gonnabe a busy, busy time.
(30:38):
I mean, just think of the influxof uh revenue that's gonna be
coming through for a lot ofthese companies, that's gonna be
insanity.
SPEAKER_00 (30:47):
It'll be an exciting few years, no matter what
happens, that's for sure.
SPEAKER_01 (30:51):
Yeah, I know the world is very unstable with a
lot of things, but um, if you goback into the past, I mean that
that's instability was alwaysthere as well.
So um if people keep focused on,you know, looking ahead, like
said, keep them keep themfocused on that.
And you're gonna you're doing agreat service for for a lot of
people, a lot of companies outthere to be successful.
So thank you.
(31:12):
Yeah, so yeah.
And so where can people connectwith you guys and learn more?
Um, where would you direct themto go so that they can um not
get lost out there and and go tothe wrong place?
SPEAKER_00 (31:23):
Yeah, so the website is laconictech.com.
That's l-a c o-n-i-c t-e-c-h dotcom.
And my email is Ken atlaconictech.com.
Uh, you can also find me onLinkedIn under Ken McLeod.
SPEAKER_01 (31:38):
Yeah, I was I was looking for it.
I said, hey, there you are.
I found chance got a lot of goodcontent on there as well.
So yeah, well, thank you forsharing that.
And uh, so we appreciate youcoming on the show, Ken.
Uh, for anybody that's listeningout there that has uh something
they need to tackle, somethingthey need to make sure that gets
done, uh look for Laconic, man.
They're they're doing big thingsout there.
(31:59):
So we appreciate you coming onthe show and talking about that.
And if you you guys also enjoythat's delivered podcast, uh
please sure to subscribe, shareit with someone you know that
can use this information.
And uh, we hope that it alsohelps a lot of professionals out
there to be successful, not justin trucking, but also in life.
And so be sure to subscribe,share with somebody.
And uh, I'm your go to guy,trucking Ray.
(32:20):
That's what I like to do is toshare these stories and
successes of a lot ofprofessionals out there.
Uh so here's another episode ofLast Delivery.
Popular Podcasts
Hey Jonas!
Hey Jonas! The official Jonas Brothers podcast. Hosted by Kevin, Joe, and Nick Jonas. It’s the Jonas Brothers you know... musicians, actors, and well, yes, brothers. Now, they’re sharing another side of themselves in the playful, intimate, and irreverent way only they can. Spend time with the Jonas Brothers here and stay a little bit longer for deep conversations like never before.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by Audiochuck Media Company.
Betrayal Weekly
Betrayal Weekly is back for a new season. Every Thursday, Betrayal Weekly shares first-hand accounts of broken trust, shocking deceptions, and the trail of destruction they leave behind. Hosted by Andrea Gunning, this weekly ongoing series digs into real-life stories of betrayal and the aftermath. From stories of double lives to dark discoveries, these are cautionary tales and accounts of resilience against all odds. From the producers of the critically acclaimed Betrayal series, Betrayal Weekly drops new episodes every Thursday. If you would like to share your story, you can reach out to the Betrayal Team by emailing them at betrayalpod@gmail.com and follow us on Instagram at @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.