April 29, 2025 • 50 mins
Machine learning, generative AI and other technology innovations have the ability to drive safer and more efficient supply chains. Freight-transportation markets are in the early stages of leveraging AI’s capabilities which will be transformational in the years to come. In this Talking Transports podcast, Dr. Yossi Sheffi, director of the MIT Center for Transportation & Logistics, joins Lee Klaskow, Bloomberg Intelligence senior transportation and logistics analyst, to share his insights about the opportunities these new technologies can bring and what it takes for wider adoption. Dr. Sheffi also talks tariffs, autonomous vehicles, warehouse robots, value-added tax (VAT) and how his real-world experience as an entrepreneur helps him to be better in the classroom.
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:07):
Hi everyone, this is Lee Clasgow and we're Talking Transports.
Welcome to Bloomberg Intelligence Talking Transports podcast. I'm your host,
Lee Claskow, Senior free transportation logistics analyst at Bloomberg Intelligence,
Bloomberg's end house research arm of almost five hundred analysts
and strategists around the globe. Before diving in a little
public service announcement, your support is instrumental to keep bringing
(00:28):
great guests onto the podcast like the one we have today,
So we need your support, so please if you enjoy
the podcast, share it, like it, and leave a comment. Also,
if you have any ideas for future episodes or just
want to talk transports, please hit me up on the
Bloomberg terminal, on LinkedIn or on Twitter at Logistics.
Speaker 2 (00:47):
Lee.
Speaker 1 (00:47):
Now onto our episode We're delighted to have with us today.
Speaker 2 (00:51):
Yoz Sheffey.
Speaker 1 (00:52):
He's a director of the Center for Transportation Logistics at MIT.
He's an expert in system optimization, risk analysis, and supply
chain management. Doctor Sheffee has authored a number of textbooks
on these topics. His ninth and latest book is The
Magic Conveyor Belt, Supply Chains, AI and the Future of Work.
(01:14):
He consults with leading enterprises and founded or co founded
five successful companies. Yosi or doctor Chffee, Welcome to the podcast.
Can I call you Yosi?
Speaker 2 (01:25):
Is that all right? That's the only name I answer to.
Speaker 1 (01:28):
So it's fine, okay, fantastic. So can you just a
little background about the MIT Logistics program. Can you just
talk a little bit about what you guys do over there.
Speaker 2 (01:40):
The center itself is now fifty two years old, so
we have started the as I said, fifty two years
mostly about public transportation. But they remember a little over
thirty years when I took it over thirty three years.
(02:02):
Actually I moved it to logistics and supply chain management
and freight transportation basically, rather than transit and urban planning.
Even though my first book was on urban transportation networks,
how you analyze urban transportation networks, but then my next
eight books were on supply chain logistics and related issues.
(02:25):
The center itself is an interdepartment center. It's a house
in the School of engineering because we are engineers, but
it's it's employee faculty from across MIT. That's the nature
of Interdepartmental Center c MIT. The departments are you think
(02:47):
about it, it's silos, it's vertical. There are cent into department.
Center go horizontal across the department in the schools, so
we have people from the School of Management, from the
School of Engineering, from the School of Urban Planning, School
of Science, and so forth. We have several programs. In
one sense, we are the most complex unit or the
(03:08):
most interesting unit that they might tea in that we
have everything that one can have. We have our education program.
We offer a master program in supply chain management, a
pH d program only graduate program. We have a very
extensive research program, many research programs. We have a you know,
(03:31):
extensive industrial partnership program, industry partnership program when we work
with industry. In fact, we can talk more about it later.
And we have five international centers that we set up
in Columbia, Spain, Luxembourg, China, Malaysia and now the UK.
(03:55):
We just launched the UK. So these are stuff as
a copy of our center, but we run it all
as a big network. We also have a very large
online programs and we can talk more about depending what
you're interested in.
Speaker 1 (04:11):
Sure, Just so for for context, about how many kids
graduate in the undergraduating undergraduate class each year.
Speaker 2 (04:18):
Around around eighty master students, eighty students from our program
and now they are one hundred and fifty from the
programs around the globe. This is a master student. Plus
we have about four or five PhD students every year
studying here. So we have about I think twelve or
(04:39):
fourteen PhD students right now in our program and it
takes them about four years, so about three four graduating
every year.
Speaker 1 (04:50):
And how many undergraduate We don't have undergraduate programs. You
don't have undergraduate Okay, well that would make sense. So
it's just such a graduate and.
Speaker 2 (04:58):
My undergraduate program only managed by department, not by center.
The other centers that are might be not as big
as as ours, but centers are managing only graduate program.
Not because it involved people from all kinds. In fact,
we believe in this. In this we want people to study,
to have a profession before they coming to study engineering systems,
(05:22):
industrial engineering, economics, some subjects that gives them some management,
some you know, ability to do some analytics, and some breath.
So that's basically our main input to the program. And
(05:44):
we have eight We have a believable number of applications,
very hard, but we have eighty. They are divided forty
and forty forty are here for the master program for
the air here in the residential program, they come, they
study in Gusts and they end up in May or
sometimes beginning of June. So it's about ten months program.
(06:05):
And we have what we call the blended program which
people take half the program online. It takes them about
eighteen months to three years. And then we take the
cream of the crop of this. And when I say
cream of the crop, this has one point two million
(06:27):
learners in our learness. In our program we have a
lot to truth from. Turns out the best students that
we have and they come to enmity of only one
for five months, only for one semester. This is the
eighty student. And then we have it's more complex because
we have so many students online that we have dozens
(06:49):
and dozens university who accept this online our online program
and give them one semester or two semesters sometimes to
finish and get a full master degree, because if you
just take the online program, you get a certificate from
m I T, but not the full masters degree.
Speaker 1 (07:05):
Right okay, And so what are what are your students
or what are you focusing on right now? Obviously there's
a lot to talk about, you know, whether it's technology,
the economy. So you know, what are you guys focusing
on right now?
Speaker 2 (07:20):
Okay? In the educational program, it does not change that
quickly because it focused on fundamentals, focused on the ability
to design a network, to optimize inventory all you know,
work on advanced procurement distribution is a transportation. Uh, these
(07:41):
issues don't change so because it's based on principles and
they don't. However, students also have to do a busis
or you know, final project, and this topic change from
year to year because they have to work industry. We
insist on students working with real data, we real companies
(08:04):
on real problems. So it's not in that sense there's
not theory here. They have to work on real problem
and real data and then and understand, the data is
never cleaned and never you know the way you want it,
and you have to clean it. You have to work
at it and you have to use it. So every
one of our students when they finished and so coming
(08:27):
back these the topic are usually more up to date topics.
So my guess is next year there will be we'll
see a lot of things about tariffs. The last the
last two years, every other thesis was about AI. Before
that it was blockchain or Alfred or you know, whatever
(08:50):
is the whatever the companies are interested in because they
marry with the students to do to do a project.
So the companies in August we have all our partners
present areas that they want to investigate. They want to
employ a student for ten months. Now I should say
(09:13):
these are limited engagement because it's especially when new companies
are join our group. They think that they fail on
gold within ten months. You know, a student will solve
problems that the mckinseye was not able to do in
five years and then five million dollars. It's not quite.
(09:33):
So we have we go to a process of understanding
the scope and what can be done in terms what
cannot be done or so the student work is an
advisor work with them. Somebody at the company is responsible
for it. So in that sense, these issues are more
up to date. We also have lots of outside presentations
(09:54):
and lectures and about ongoing research, ongoing issues with industry.
These are more up to date. But the basic program
program is based on principle. It does not change quickly
from year to year.
Speaker 1 (10:09):
You mentioned AI, so is in your view you know
you mentioned a couple other things like blockchain, which really
kind of fizzled out. You know, it seems like AI
is truly one of the transformational things that could impact
not only transportation, but obviously the broader economy and other industries.
Can you talk about, you know, where you're seeing AI
(10:30):
uh and supply chains and where do you think, uh
it could go?
Speaker 2 (10:35):
Okay, where do I think it can go? Is the
you know, hard to predict. The technology is still evolving
so quickly and the capabilities are changing so fast that
it's hard to imagine all the changes. However, we see
already cases where it is very helpful. For example, in procurement,
(11:02):
we see companies implementing machine learning and generative AI in
making sure that the whole process from generating you know,
requests for information and to requests for proposal, for requests
(11:23):
for quote, to contracting to signing contract to then following
what's going on is done on semi automatically at this point,
but going to become more and more automated using these
advanced AI methods. We see, of course, you know that
the things that you don't even think about, but when
(11:45):
you talk to your favorite customer service representative, you know,
used to say press one to take hear, press to
to get ear prestificate here. Now you just talk and
you know, the the chatbot talks back to you. So
it's done in the natural language. Lots of the communication
between people and computer is now done with language. There's
(12:10):
a lot of a lot of work going on on
a network design that try to do it more, you know,
with more AI. This is not going this is not
yet widespread. There is a lot of work on infusing
more and more robots with more and more capabilities. So
(12:33):
if you think about a warehouse when robots are running
around and you have to worry about safety because there
are people there, and you're also worry about the robots themselves.
So think about it that robots have a if each
robots run around the warehouse a space of let's say
(12:55):
a miter a miter and a half around there that
if anything enters this space, the robot stops. Now we
have prograt projects basically a lot of AI in this
lots of sensors and AI and combination that try to
limit it to let's say a lot less maybe a foot.
(13:17):
This means that the robots can be a lot more efficient,
more of them can run around and not run into
each other or running too people. Some of these technology
are the same as technology using autonomous vehicles. Also, it's
a combination. It's not just AI. It's a combination of
(13:39):
sensors and leaders and radars and some smart to process
all this and demonstrate and get into a conclusion. Upcoming.
What we see coming down the pulic is agent AI,
(14:00):
when there'll be a piece of so software that will
go out and perform tasks on an individual level. You
can ask even today, you can ask some of the
leading AI provider to design a trip to Venice and
(14:22):
they'll come up with just like a travel agent. They'll
come up with the flights and hotel and suggestions where
to go. I didn't use it myself, so I don't
know how good they are, but it's an example of
what's coming. Clearly. People are if you say a law firm,
(14:44):
people can send and it's being done, sending an agent
to collect all the president cases for some lawsuits and
summarize them and present them. Now, the problem is sometimes
to hallucinate and come up with nonsense. But this problem
(15:04):
is being you know, address better and better. It's not
perfect yet by any stretch of the imagination. We see
even you know, when we use Google Map, it's becoming
better and better. Translation is becoming better and better because
of the use of AI. So we see it happening
in all parts of life and certainly in supply chain management.
Speaker 1 (15:26):
And do you guys at the center use AI.
Speaker 2 (15:29):
Yes, we do. In fact, we encourage our students to
use and AI. We went to go to go through
the understanding that you're give an assignment or a case
study and people use chet, gipt in or to give
you an answer. This changed how we teach and how
we work. We separated and put a lot more emphasis
(15:50):
on the assessment park. So, for example, assume there's a
case study. You send students at Hon't to do a
case study, and you don't know if they come up
with che jipt or some other states or they actually
do the work. What you do is in the class
you don't talk about it. You just tell them, okay,
I'm going to point at you and you have to
tell me what's going on with the case. Just calling
(16:16):
they induce fear make sure that they do the work.
So we had to adapt ourselves to for example, in
the teaching, we don't want them not to use AI
because they'll have everything work, they'll have it everywhere. So
we want them, we want them to learn how to
use it and want them to get better at it.
But of course we tell them that when you submit,
(16:39):
when the jgbts submit nonsense, it's you get an f
not open eye. So anyway, it's honestly it's still work
in progress. We're still learning to work with it. There now.
For example, pieces of software that when the students submit
somebody something you can analyze it and see what percentage
(17:02):
of it was written by AI. Turns out you can
do it Amazon for example, if you submit you know,
I did several of my books with Amazon. If you
submit a text for a book to Amazon, they'll analyze
it and if more than fifty percent is written by AI,
they're not going to publish it. Kidney is not going
(17:23):
to publish it. So it starts to be an arms race,
you know stuff and anti stuff and we'll see.
Speaker 1 (17:34):
So do you view like in your career and supply
chains and transportation, do you view AI as the most
transformative change or technology that you've seen in your career?
Speaker 2 (17:47):
Well, that's being old means that it's hard to make
this statement. I was still alive when containers came on.
You know, containers change everything, change the whole. So the
Internet change. It's hard to say, you know, the Internet
(18:10):
come on, change how we work, how we do stuff,
how we communicate. I think that AI is a change
on that scale. It will change how people work, it
will change relationship. There's an interesting article when it is
the worst with journal or the economists, that they talk
about the CEOs that this is the last crop of
(18:34):
CEOs who still manage only humans. They'll have to manage
humans working with you know, digital agents. And actually there
there are age R programs that are starting to think
about how do we manage agents. I'm not sure that
(18:54):
HR professionals are because they are taught to show em
and then and work on the regulation for human resources.
But it just so to show you how people are
start to think about it. This is still we're talking
about the future, really when you're not there yet. In fact, interestingly,
(19:19):
I'm now working on a book just it's not even
close to being done because it's a moving target. I'm
working about what is actually being used in supply chain management,
because when everything is said and done, you know a
lot more is said than done. So people are trying
(19:40):
to do some testing, some experiment here and there. There
are some people who are starting to put it to
work at scale and commercial use. So I'm trying to
find out where these nuggets are, talking to a lot
of companies, trying to find what is being done. Mostly
(20:03):
when I find out, of course Chegibt and just talking
the tech is everywhere interesting, the stuff that works. So
I'll give you some of the things that I've found
the stuff that works. Let me take a step back.
You know why chess and games and where software AI
(20:32):
was very successful in playing chess, in playing goal, which
is much more complicated. Given all this, because these our
system with rules, the objectives is clear, it's you can
repeat it, and it's the same thing. The world is
not like this. You were very far from being able
to ask Chegibt or any other of the program to
(20:57):
design us a global networks that we will respond to
risks and be efficient and you know, make sure that
we not burn the planet, worry about sustainability and risk
and you know, and be a low cost and efficient
and fast can't be done. Were not there because these systems,
(21:21):
there are so many objectives sometimes the contradict with each other.
The world is changing all the time. So this is
what you might call open system versus chess. That's closed system,
but certainly chess. Let's say you go to your favorite
you know, windows in a drive through to buy coffee.
(21:44):
In many of these places, you're not talking to a
person when you order. You're talking to chipetitos or to
a chat bot. And it works. Why does it work?
Why does it work? Because you are not going to
go to the window and ask questions like is their god?
You would ask questions can I have almond milk with
(22:04):
my coffee? I mean? Which is easy to train on
the questions are the same. You do machine learning on
all the data that you have, and it works. It
was pretty good. So so when the problem, when the
problem is contained in several on several dimensions, it's very
it's we're there, we can do it. The the hope
(22:28):
is the work on problem that I'm not contained are
really big and ugly and the changing all the time,
But we're not there yet.
Speaker 1 (22:37):
Gotcha. And you know, and you mentioned another technology earlier
when you were talking about robots you talked you mentioned
it's the same technology or similar technology for safety with
autonomous vehicles. What what's your take on like autonomous vehicles
on the roadway. Do you think that it's the future
is closed or do you think this is something that
(22:59):
is a lot of further are out then maybe those
that are pretty bullish on the autonomous vehicles out there,
whether it's trucks are card i.
Speaker 2 (23:08):
Would saw to some extent the future is here. You
go to San Francisco to Phoenix, you can have an
autonomous taxi. And in the South you can have autonomous trucks,
mostly still running in training, but some of them are
running limited commercial use, very limited and only in the
(23:29):
South when the roads are open, when you don't have snow,
when you don't have the cowpath that define Boston roads,
it's you know, it works there. There are, however, issues
with this. The issues are with every system that involves
human and robots, human and autonomy. There's a question of
(23:55):
public acceptance. Let me give you an extreme example. Today
modern aircraft seven eighty seven eight three fifty is actually
a drone. It can fly by itself. Not too many
people would fly at the aluminum tube, but thirty five
thousand feet across the Atlantic with nobody in the front.
(24:19):
But it's possible. There's the technologies there. You know, these
are really drones. I mean you see what drones do
and autonomous they do. They bomb, they go back, they
collect intelligence. It can be done, but public acceptance is
not there. So even though by the way that people
(24:41):
working on it, on getting we used to have five
people in the cockpit of an airliner and now that too,
and there's a lot of work and getting into one
and the other one is an AI helper and it's
and at the end of the day it will become autonomous.
So the question is how do people feel about a
(25:03):
large truck, weird looking track running at the one hundred
miles per hour next to them on the highway with
nobody in the front? Would they be you know, and
we will have accidents and who knows how it will
play people. No, this is too dangerous. We cannot have it.
(25:24):
It may come because because there'll be impacts on jobs.
You know, truck driver jobs are in thirty thirty some
states are the number one job, I mean the number
one job category. So it's lots of trucking jobs. So
(25:45):
going away, we'll see, we see what happens in the
port very hard to automate because of the union. So
and if you go to ports, as you must know,
you go to buy or Singapore or Rotterdam, and then
you go to an American port, then you get depressed.
(26:08):
It is what it is. I mean, I I understand,
but there's a lot of fame pushback and on automation.
It also it also takes a longer time than people think,
only a longer time than people think. Give you give
you one example. You know we used to have telephone
(26:30):
exchange operators. In eighteen ninety the AT and T came
up with an automatic exchange just based on numbers. You
didn't have to call susan in the exchange and ask
it to find where mister Glasgow is having lunch today
(26:53):
so we can talk to him. It was all automated. Now,
honestly think we lost some level of service, but we
got we were an efficiency. Turns out it took by
nineteen fifty we still had three hundred and fifty thousand
exchange operators. Only by nineteen eighty, which is nine decades later,
(27:19):
these people disappeared. So it takes time until technologies catch
not to be sure, I was working a lot faster
because it starts And the reason is because most of
today's technology starts with the consumers and they pick it
up because it's useful, it's interesting, it's cool, so they
(27:40):
pick it up and then it goes to companies.
Speaker 1 (27:43):
You mentioned like acceptance of autonomous vehicles. It was in
a way about cab in southern California a couple of
months ago, and it was definitely cool. My anxiety was
definitely higher than a normal taxi. But yeah, I can
I can see people, you know, not being very comfortable
(28:05):
going into a plane without at least two people in
the cockpit. And so you know, another thing I'm assuming
you guys are looking at and we're in extremely early
early innings and we don't really know how this is
going to play out. But all these tariffs that are
coming out of the US, I mean, are supply chain
is going to be redrawn or it's just just going
(28:25):
to be inflationary and we're all just going to deal
with higher prices and things are still going to be
you know, stickers are still going to be manufactured in
Vietnam and you know those those jobs aren't obviously coming
to the US. So so so how do you see
this playing out from a supply chain standpoint.
Speaker 2 (28:43):
Let me start my more general Okay, during the decades
of globalization, most of the population around the world benefit
with much higher start of living. Right, those are small
minority that extraly suffer. The small minority people who lost
manufacturing job in the industrial heartland the United States, for example,
(29:07):
ghost towns that the lost employment, which in some sense
led to the Fentanel crisis later on. So we had
most people, you know, enjoyed it, few people suffered. With
the current tarists, we are turning it upside down. Most
(29:29):
of people are going to suffer, a few people will benefit.
The people who benefit, for example, are domestic manufacturers who
will be able to have less competition, be able to
raise the price, increase their margins, and they will keep,
by the way, lobbying to continue this. So every economic
(29:51):
dislocation creates winner and loser, and the winners will keep
making it hard to dial it down. Now, let's talk
about what happened to supply chain. In some sense, I'm
less worried after COVID and Ukraine and the Middle East
and all this. In many ways, it's just another day
(30:14):
of the office. It is I was just with a
group of fifty supply chain officers and this was the
basically the the main response this is I was with
them last week when the when it was announced, nobody
was panicking. Some people immediately said, okay, even fifty tariffs
(30:40):
in now China is sixty percent or something. When you
enter everything up, it is still we cannot move to
the United States. The unfortunate thing that the administration does
seem to pay attention to our people, not that they
did right media to pay attention to is the fact
(31:02):
that it takes three years to be a manufacturing plant.
You cannot just come to deal. And then, by the way,
we don't have workers. You know, when DSMC build a
plant in a in Phoenix, in the Phoenix area, they
said that they're making the chips in the US will
cost fifty percent more. Why because we don't have enough
(31:23):
engineering talent and they have to train them. Every time
they're bringing engineers on, they have to train them. So
it's they said, it's not we Our biggest failure is
the education system. The US education system failed miserably. We
don't have an you know, MIT is not an example.
(31:45):
We get out of three hundred million people we get,
you know, where the MIT or Stanford or Harved we
get the top students even in STEM in science, technology, engineering, mathematics.
But the average level of the US is going down significantly.
And this is to me, this is the biggest fear.
(32:08):
And by the way, one of the main reasons that
we have this, you know, deficits in the balance of payment,
because otherwise we would have attracted more companies. But companies
know that there's no the tenant here is lucky and
the administration rather than invest mightily in education, I mean,
(32:32):
they try to root out DEI and the other things
from the education. Okay, fine, but make sure that they
focus on education or real education. I'm not sure they're
doing this, I mean the sleep service. But they're not
doing this because we are graduating kids who can hardly
(32:55):
read and write, forget being able to do real engineerings.
It's just or the number that we the number of
people who can do advanced engineering is way too small
because and it's not because of the university, it's because
of the input that comes into the universities. That's not
(33:18):
enough anyway. So that's a long answer to a short question.
Speaker 1 (33:24):
Yeah, I'm sure we can probably talk a long time
about it. Do you see any things that are any
industries or manufacturing processes that are easy to bring.
Speaker 2 (33:38):
Back to the United States?
Speaker 1 (33:39):
You know, we mentioned chips, we mentioned sneakers. You know,
I can't see you know people you know want of
those the manufacturing sneakers. And I don't think Americans are
going to pay five hundred dollars for a pair of Nikes.
Speaker 2 (33:54):
The problem with American liberates and American you know, oh,
environmental rules and all kinds of regulations, the prices will
be very high. It's so the question for a sneaker
company is do we pay sixty percent more and get
the sneakers vesty for three hundred dollars, or we bring
(34:15):
it to the United States, wait three years, build a plant,
don't find the workers for it, and then when we
finally automate it and get it to work with minimum
number of workers, it's five hundred dollars per sneaker and
we can't sell them anyway. So it's not clear. Look,
(34:36):
the main is it a problem with that you know,
automotive industry. They're being hammered because many of the parts
cross Canada and Mexico to United Several times, and each
time they'll be will be tariff on this. So it's
a it's getting ridiculous. The only way to do it is,
(34:58):
you know, but we should if we want to do something,
we should adopt value edit tacks. Just like most of
the world has value edit tax. It's a tax that
it is the least disruptive tax, but the US never
(35:18):
never went to it. It's much more efficient than sales
text that we charge here because it charged and it
also means very hard. It has so many other pluses.
It's it's very hard to use, you know, to do
stuff under the table because every stage of production must
report it and then get the money back. It's the
(35:40):
whole way that it's working along the supply chain.
Speaker 1 (35:45):
So can you, I guess can you explain that a
little bit? So why is the value added tacks the
bad tax.
Speaker 2 (35:52):
More fashion? So let's say you have a twenty percent tax.
It's most Europe is twenty percent tax. And let's say
if very simple thing, you get one supplier and one manufacturer.
The supplier pays twenty five percent tax. It comes to
the manufacturer. The manufacturers now has value added. He paid
(36:18):
twenty five percent only on the value added part, and
then on the part that comes from a supplier he
can get the money back. He applies to get the
money back, so everything has to be upfront in order
to get the money back. Now, finally, the one who
pays the full freight so to speak, is the consumer.
(36:40):
The consumer cannot get anything back. So for the consumer
it's the consumer paid twenty percent. But along the way
companies paid just for the value added actually that they
add add to it, and then for anything that the
supplier paid, the buyer can charge back at value. So
(37:05):
it's it kind of moves along the supply chain and
people are charging and charging and getting the money bag,
charging and getting part of the money bag. So it
requires significant you know management, but hey, we have a
big IRS, so's it can be done. It's a question
(37:26):
and now it's required on software and all this. But
it's a much more efficient way of taxation than sales
tax for example Celtics, it's not doesn't have any of
these qualities. But by the way, the administration makes a
mistake by looking at value added tax as if it's
(37:47):
distorting the market. First of all, the Europeans cannot get
rid of it because it finds most of the government.
It's a huge part of funding in a local government
in every state. They are in every country. I think
the only three or four countries in the world, and
the US is one of them. They do not have
value added tax. It is every economist will tell you
(38:10):
it's the most efficient way to tax, but we don't.
So we don't anyway. But it's so now they're the
h The administration took value added tax into account, thinking
that it's it's like a tariff. It's not, absolutely not.
(38:30):
But anyway, the administration is very simple calculation. But let
me add something just for the record, so to speak.
It is clear that some of what Trump is saying
that we were taking advantage of is correct. It's absolutely correct.
(38:50):
And by the way, every administration, the last I don't
know five six the administration was complaining about it. Right.
Trump is trying to do something about it now the
way he works in everything. If you look at how
its university, first of all, take a few hundred million
dollars and then let's talk, so you certainly get people's attention,
(39:13):
and just like what it does now is getting He
said that already fifty countries came back to renegotiate. Okay,
we'll see. We'll see how that goes. I hope they'll negotiate.
What can also happen that in a few months, when
the full branch of these teriffs will hit consumers, the
(39:36):
outcry will reach Congress, will reach the Senate, and we
may have, you know, congressional action. They'll simply outlaw some terrasts,
which is again a bad way of doing it without analysis,
without it. It should be the executive branch doing this,
but it should be done with a lot of analysis.
(39:58):
What is it should be? You know, which items should
be thanks at, what which how ter should should replace?
Should be a lot more sophisticated than just blend xpresent
on everything and just the simplical collation that the administration did.
(40:20):
But and again I am you know, Trump said he's
gonna do it. In some sense, he was elected based
on what he said. He never hid the fact that
he's in labor terist. I mean, he never hed it.
He was talking about it in the campaign trail. He
(40:41):
votes people voted him in. So it's a it's hard
to complain about it because we got what we voted for.
When I said we, I mean the majority of the
population got what they voted for. So here we are.
Speaker 1 (40:59):
So let's shift gears a little bit. So you're not
only UH an academia, You're you're also been involved with
a number of companies. Looking at your bio, you founded
or co founded five companies that were acquired by other companies.
So that's that's a pretty darn good track record. Can
you talk about the kind of like the lessons learned
(41:21):
from founding companies UH in supply chain and sure, you know,
maybe maybe talk to those experiences a little bit.
Speaker 2 (41:31):
Sure, So some companies were software companies and some were
maybe one of the more interesting companies was called Logicorp.
It's a I was consulting at the time to Rockwell
International and befriended the vice president logistic the trock Oil
(41:52):
Roco was selling truck parts UH trucking company everything under
the truck brank, breaks, excels, what have you. At that time,
it was deregulation of the transportation industry came into being,
and we just look at each other and say, okay,
we can now serve every route without getting the the
(42:16):
government to prove it. So let's do what's called a
market test auction. We did we got the right to
do under Rockwell still got the right to do auction,
and the results were astounding prices when half became half
the price or you know, forty percent better and with
(42:38):
the same carriers as before. And then so we look
at it and said, okay, there's a business here. So
we convinced Rockwall to start the separate unit and then
basically we bought it from Rockwall and this started growing
like weeds in We ran it three years under rock Oil,
(43:02):
which it didn't grow much because the carriers were afraid
that we have biasing. We did these auctions, but the
carriers were afraid of buiasing. The results for rockcol for
a Rockcay customers, and we said we must get it out.
We found the an angel. We took it out and
it grew from it was about forty million when it
(43:22):
took it out of rock Oil. It grew within three
years to six hundred million, and it grew like weeds.
We were an example of outsourcing. We had just logistical engineer.
Everything else was every other corporate function was outsourced. We
tried to tell people they can outsource the traffic manage
transportation management to us and to give you an idea.
(43:47):
There was two quarters when we said we cannot take
any new business, and we still doubled just existing business.
Speaker 1 (43:55):
It's good problem to have.
Speaker 2 (43:56):
Well, the wheels are coming off the wagon in terms
of very hard thisason before the internet was it not
easy to manage a business like this? I mean the
papers and stuff and showed with mis solid to Writer.
It's now a three billion you know dollars annually part
of Writer. You know. The chairman of Writer called me,
(44:19):
he retired. He said, you know what we did one
hundred and fifty six acquisition. This was the best one
that we ever did. So so it was and I
got on the Writer board and stuff like this. But
let's a so this was one company. We had a
(44:39):
software company that did that started by doing software for
tracking for routing, scheduling, assignment of drivers and TL truck
load load load matching, a load planning for LTL and
then change it or grew it to do procurement basically
for for shippers. And this company was sold to Manhattan Associates.
(45:07):
H There were lots of stories, you know, along ever,
but I tell you what I always stayed also my
t I usually well what I did is started a company,
spent a year at the company and then just became
chairman just you know, found management team. And then because
I I could not stay them T and have b
(45:29):
a line officer because I understand it, I see it
in front of the class. You don't want to get
a phone call certainly and run to the company. Okay,
that's an m T rule that made that's reasonable. So
I work rule and either took of absence of sabbatical
whatever it was, and start a company and then acquire
(45:53):
They all acquired just putunistically, they're all all acquired by
by the company. But what did you ask for? What
I learned? I learned that that became an amazingly good teacher.
I talked about my own experience and there's nothing you know.
I worked directly with companies, with a tracking company, ships
(46:15):
with carrier. We developed one of the first TMS's transportation
management systems. So it's a It was one thing to
talk to students about stuff that I had experienced with
that they're not talking about what they reading the book,
what they do a case study. I developed case studies
(46:37):
that according that I know everything about it because I
lived it. So I'm also, of course consulting. Now the
last company was so years ago. I'm consulting to lots
of budding entrepreneurs because after that I joined some venture
capital boards, but not anymore. But a lot of people
(47:01):
they might want to start companies, they want to study this.
I'm the informal consult that to those people stry to
try to start companies.
Speaker 1 (47:09):
Well, well that must be extremely rewarding. You know, we're
coming up towards the end of our time and I
just always like to ask this, and maybe I'll phrase
the question all differently. Since you know, giving that your
mi T, do you have like a required reading for
all your students. It might not be, it might not
be a textbook, but it's something that you know, would
(47:31):
prepare them for a career in supply chain.
Speaker 2 (47:37):
That's very you know, there are some textbooks in supply chain.
My own books are not textbooks. There are business books
and look for example, my book on name it's called
Sustainability in Supply Chain. It's called the Bus and Green.
(48:02):
When to apply Sustainability in supply gen and when not
to is used by several schools as textbook, including the Hours.
Now we don't have a class on separate issues so
we usually don't assign books. We are command books, so
books on specific specific subject. On AI, we don't have
(48:27):
a book because it is we just talk about the
latest and every semester the the lectures change and the
the work the homework changes because we students. So for example,
our students now have to not Python, which is the
(48:47):
language that you know, the computer language used mostly for APIs,
and but every year they have to do more or
different because the abilities are a different, you know, the
new software coming, the latest version of GPT of you know, Copilot, Jemini,
(49:12):
whatever it is, it's getting better. I can do more
and so you work with it.
Speaker 1 (49:19):
Well you sa I really want to thank you for
your time. This was a really great conversation and thanks
for your insights.
Speaker 2 (49:26):
All right, thank you very much for having me appreciate it.
Speaker 1 (49:29):
It's my pleasure and hope to have you back again
some time. And I want to thank you for tuning in.
If you liked the episode, please subscribe and leave a review.
We've lined up a number of great guests for the podcast,
so please check back to your conversations with c suite executives, shippers, regulators,
and decision makers within the freight markets. Also, if you
have any ideas for a future episodes, just hit me
(49:51):
up on the Bloomberg terminal or on Twitter at logistics Late.
Thanks so much and take care everyone,
Hi everyone, this is Lee Clasgow and we're Talking Transports.
Welcome to Bloomberg Intelligence Talking Transports podcast. I'm your host,
Lee Claskow, Senior free transportation logistics analyst at Bloomberg Intelligence,
Bloomberg's end house research arm of almost five hundred analysts
and strategists around the globe. Before diving in a little
public service announcement, your support is instrumental to keep bringing
(00:28):
great guests onto the podcast like the one we have today,
So we need your support, so please if you enjoy
the podcast, share it, like it, and leave a comment. Also,
if you have any ideas for future episodes or just
want to talk transports, please hit me up on the
Bloomberg terminal, on LinkedIn or on Twitter at Logistics.
Speaker 2 (00:47):
Lee.
Speaker 1 (00:47):
Now onto our episode We're delighted to have with us today.
Speaker 2 (00:51):
Yoz Sheffey.
Speaker 1 (00:52):
He's a director of the Center for Transportation Logistics at MIT.
He's an expert in system optimization, risk analysis, and supply
chain management. Doctor Sheffee has authored a number of textbooks
on these topics. His ninth and latest book is The
Magic Conveyor Belt, Supply Chains, AI and the Future of Work.
(01:14):
He consults with leading enterprises and founded or co founded
five successful companies. Yosi or doctor Chffee, Welcome to the podcast.
Can I call you Yosi?
Speaker 2 (01:25):
Is that all right? That's the only name I answer to.
Speaker 1 (01:28):
So it's fine, okay, fantastic. So can you just a
little background about the MIT Logistics program. Can you just
talk a little bit about what you guys do over there.
Speaker 2 (01:40):
The center itself is now fifty two years old, so
we have started the as I said, fifty two years
mostly about public transportation. But they remember a little over
thirty years when I took it over thirty three years.
(02:02):
Actually I moved it to logistics and supply chain management
and freight transportation basically, rather than transit and urban planning.
Even though my first book was on urban transportation networks,
how you analyze urban transportation networks, but then my next
eight books were on supply chain logistics and related issues.
(02:25):
The center itself is an interdepartment center. It's a house
in the School of engineering because we are engineers, but
it's it's employee faculty from across MIT. That's the nature
of Interdepartmental Center c MIT. The departments are you think
(02:47):
about it, it's silos, it's vertical. There are cent into department.
Center go horizontal across the department in the schools, so
we have people from the School of Management, from the
School of Engineering, from the School of Urban Planning, School
of Science, and so forth. We have several programs. In
one sense, we are the most complex unit or the
(03:08):
most interesting unit that they might tea in that we
have everything that one can have. We have our education program.
We offer a master program in supply chain management, a
pH d program only graduate program. We have a very
extensive research program, many research programs. We have a you know,
(03:31):
extensive industrial partnership program, industry partnership program when we work
with industry. In fact, we can talk more about it later.
And we have five international centers that we set up
in Columbia, Spain, Luxembourg, China, Malaysia and now the UK.
(03:55):
We just launched the UK. So these are stuff as
a copy of our center, but we run it all
as a big network. We also have a very large
online programs and we can talk more about depending what
you're interested in.
Speaker 1 (04:11):
Sure, Just so for for context, about how many kids
graduate in the undergraduating undergraduate class each year.
Speaker 2 (04:18):
Around around eighty master students, eighty students from our program
and now they are one hundred and fifty from the
programs around the globe. This is a master student. Plus
we have about four or five PhD students every year
studying here. So we have about I think twelve or
(04:39):
fourteen PhD students right now in our program and it
takes them about four years, so about three four graduating
every year.
Speaker 1 (04:50):
And how many undergraduate We don't have undergraduate programs. You
don't have undergraduate Okay, well that would make sense. So
it's just such a graduate and.
Speaker 2 (04:58):
My undergraduate program only managed by department, not by center.
The other centers that are might be not as big
as as ours, but centers are managing only graduate program.
Not because it involved people from all kinds. In fact,
we believe in this. In this we want people to study,
to have a profession before they coming to study engineering systems,
(05:22):
industrial engineering, economics, some subjects that gives them some management,
some you know, ability to do some analytics, and some breath.
So that's basically our main input to the program. And
(05:44):
we have eight We have a believable number of applications,
very hard, but we have eighty. They are divided forty
and forty forty are here for the master program for
the air here in the residential program, they come, they
study in Gusts and they end up in May or
sometimes beginning of June. So it's about ten months program.
(06:05):
And we have what we call the blended program which
people take half the program online. It takes them about
eighteen months to three years. And then we take the
cream of the crop of this. And when I say
cream of the crop, this has one point two million
(06:27):
learners in our learness. In our program we have a
lot to truth from. Turns out the best students that
we have and they come to enmity of only one
for five months, only for one semester. This is the
eighty student. And then we have it's more complex because
we have so many students online that we have dozens
(06:49):
and dozens university who accept this online our online program
and give them one semester or two semesters sometimes to
finish and get a full master degree, because if you
just take the online program, you get a certificate from
m I T, but not the full masters degree.
Speaker 1 (07:05):
Right okay, And so what are what are your students
or what are you focusing on right now? Obviously there's
a lot to talk about, you know, whether it's technology,
the economy. So you know, what are you guys focusing
on right now?
Speaker 2 (07:20):
Okay? In the educational program, it does not change that
quickly because it focused on fundamentals, focused on the ability
to design a network, to optimize inventory all you know,
work on advanced procurement distribution is a transportation. Uh, these
(07:41):
issues don't change so because it's based on principles and
they don't. However, students also have to do a busis
or you know, final project, and this topic change from
year to year because they have to work industry. We
insist on students working with real data, we real companies
(08:04):
on real problems. So it's not in that sense there's
not theory here. They have to work on real problem
and real data and then and understand, the data is
never cleaned and never you know the way you want it,
and you have to clean it. You have to work
at it and you have to use it. So every
one of our students when they finished and so coming
(08:27):
back these the topic are usually more up to date topics.
So my guess is next year there will be we'll
see a lot of things about tariffs. The last the
last two years, every other thesis was about AI. Before
that it was blockchain or Alfred or you know, whatever
(08:50):
is the whatever the companies are interested in because they
marry with the students to do to do a project.
So the companies in August we have all our partners
present areas that they want to investigate. They want to
employ a student for ten months. Now I should say
(09:13):
these are limited engagement because it's especially when new companies
are join our group. They think that they fail on
gold within ten months. You know, a student will solve
problems that the mckinseye was not able to do in
five years and then five million dollars. It's not quite.
(09:33):
So we have we go to a process of understanding
the scope and what can be done in terms what
cannot be done or so the student work is an
advisor work with them. Somebody at the company is responsible
for it. So in that sense, these issues are more
up to date. We also have lots of outside presentations
(09:54):
and lectures and about ongoing research, ongoing issues with industry.
These are more up to date. But the basic program
program is based on principle. It does not change quickly
from year to year.
Speaker 1 (10:09):
You mentioned AI, so is in your view you know
you mentioned a couple other things like blockchain, which really
kind of fizzled out. You know, it seems like AI
is truly one of the transformational things that could impact
not only transportation, but obviously the broader economy and other industries.
Can you talk about, you know, where you're seeing AI
(10:30):
uh and supply chains and where do you think, uh
it could go?
Speaker 2 (10:35):
Okay, where do I think it can go? Is the
you know, hard to predict. The technology is still evolving
so quickly and the capabilities are changing so fast that
it's hard to imagine all the changes. However, we see
already cases where it is very helpful. For example, in procurement,
(11:02):
we see companies implementing machine learning and generative AI in
making sure that the whole process from generating you know,
requests for information and to requests for proposal, for requests
(11:23):
for quote, to contracting to signing contract to then following
what's going on is done on semi automatically at this point,
but going to become more and more automated using these
advanced AI methods. We see, of course, you know that
the things that you don't even think about, but when
(11:45):
you talk to your favorite customer service representative, you know,
used to say press one to take hear, press to
to get ear prestificate here. Now you just talk and
you know, the the chatbot talks back to you. So
it's done in the natural language. Lots of the communication
between people and computer is now done with language. There's
(12:10):
a lot of a lot of work going on on
a network design that try to do it more, you know,
with more AI. This is not going this is not
yet widespread. There is a lot of work on infusing
more and more robots with more and more capabilities. So
(12:33):
if you think about a warehouse when robots are running
around and you have to worry about safety because there
are people there, and you're also worry about the robots themselves.
So think about it that robots have a if each
robots run around the warehouse a space of let's say
(12:55):
a miter a miter and a half around there that
if anything enters this space, the robot stops. Now we
have prograt projects basically a lot of AI in this
lots of sensors and AI and combination that try to
limit it to let's say a lot less maybe a foot.
(13:17):
This means that the robots can be a lot more efficient,
more of them can run around and not run into
each other or running too people. Some of these technology
are the same as technology using autonomous vehicles. Also, it's
a combination. It's not just AI. It's a combination of
(13:39):
sensors and leaders and radars and some smart to process
all this and demonstrate and get into a conclusion. Upcoming.
What we see coming down the pulic is agent AI,
(14:00):
when there'll be a piece of so software that will
go out and perform tasks on an individual level. You
can ask even today, you can ask some of the
leading AI provider to design a trip to Venice and
(14:22):
they'll come up with just like a travel agent. They'll
come up with the flights and hotel and suggestions where
to go. I didn't use it myself, so I don't
know how good they are, but it's an example of
what's coming. Clearly. People are if you say a law firm,
(14:44):
people can send and it's being done, sending an agent
to collect all the president cases for some lawsuits and
summarize them and present them. Now, the problem is sometimes
to hallucinate and come up with nonsense. But this problem
(15:04):
is being you know, address better and better. It's not
perfect yet by any stretch of the imagination. We see
even you know, when we use Google Map, it's becoming
better and better. Translation is becoming better and better because
of the use of AI. So we see it happening
in all parts of life and certainly in supply chain management.
Speaker 1 (15:26):
And do you guys at the center use AI.
Speaker 2 (15:29):
Yes, we do. In fact, we encourage our students to
use and AI. We went to go to go through
the understanding that you're give an assignment or a case
study and people use chet, gipt in or to give
you an answer. This changed how we teach and how
we work. We separated and put a lot more emphasis
(15:50):
on the assessment park. So, for example, assume there's a
case study. You send students at Hon't to do a
case study, and you don't know if they come up
with che jipt or some other states or they actually
do the work. What you do is in the class
you don't talk about it. You just tell them, okay,
I'm going to point at you and you have to
tell me what's going on with the case. Just calling
(16:16):
they induce fear make sure that they do the work.
So we had to adapt ourselves to for example, in
the teaching, we don't want them not to use AI
because they'll have everything work, they'll have it everywhere. So
we want them, we want them to learn how to
use it and want them to get better at it.
But of course we tell them that when you submit,
(16:39):
when the jgbts submit nonsense, it's you get an f
not open eye. So anyway, it's honestly it's still work
in progress. We're still learning to work with it. There now.
For example, pieces of software that when the students submit
somebody something you can analyze it and see what percentage
(17:02):
of it was written by AI. Turns out you can
do it Amazon for example, if you submit you know,
I did several of my books with Amazon. If you
submit a text for a book to Amazon, they'll analyze
it and if more than fifty percent is written by AI,
they're not going to publish it. Kidney is not going
(17:23):
to publish it. So it starts to be an arms race,
you know stuff and anti stuff and we'll see.
Speaker 1 (17:34):
So do you view like in your career and supply
chains and transportation, do you view AI as the most
transformative change or technology that you've seen in your career?
Speaker 2 (17:47):
Well, that's being old means that it's hard to make
this statement. I was still alive when containers came on.
You know, containers change everything, change the whole. So the
Internet change. It's hard to say, you know, the Internet
(18:10):
come on, change how we work, how we do stuff,
how we communicate. I think that AI is a change
on that scale. It will change how people work, it
will change relationship. There's an interesting article when it is
the worst with journal or the economists, that they talk
about the CEOs that this is the last crop of
(18:34):
CEOs who still manage only humans. They'll have to manage
humans working with you know, digital agents. And actually there
there are age R programs that are starting to think
about how do we manage agents. I'm not sure that
(18:54):
HR professionals are because they are taught to show em
and then and work on the regulation for human resources.
But it just so to show you how people are
start to think about it. This is still we're talking
about the future, really when you're not there yet. In fact, interestingly,
(19:19):
I'm now working on a book just it's not even
close to being done because it's a moving target. I'm
working about what is actually being used in supply chain management,
because when everything is said and done, you know a
lot more is said than done. So people are trying
(19:40):
to do some testing, some experiment here and there. There
are some people who are starting to put it to
work at scale and commercial use. So I'm trying to
find out where these nuggets are, talking to a lot
of companies, trying to find what is being done. Mostly
(20:03):
when I find out, of course Chegibt and just talking
the tech is everywhere interesting, the stuff that works. So
I'll give you some of the things that I've found
the stuff that works. Let me take a step back.
You know why chess and games and where software AI
(20:32):
was very successful in playing chess, in playing goal, which
is much more complicated. Given all this, because these our
system with rules, the objectives is clear, it's you can
repeat it, and it's the same thing. The world is
not like this. You were very far from being able
to ask Chegibt or any other of the program to
(20:57):
design us a global networks that we will respond to
risks and be efficient and you know, make sure that
we not burn the planet, worry about sustainability and risk
and you know, and be a low cost and efficient
and fast can't be done. Were not there because these systems,
(21:21):
there are so many objectives sometimes the contradict with each other.
The world is changing all the time. So this is
what you might call open system versus chess. That's closed system,
but certainly chess. Let's say you go to your favorite
you know, windows in a drive through to buy coffee.
(21:44):
In many of these places, you're not talking to a
person when you order. You're talking to chipetitos or to
a chat bot. And it works. Why does it work?
Why does it work? Because you are not going to
go to the window and ask questions like is their god?
You would ask questions can I have almond milk with
(22:04):
my coffee? I mean? Which is easy to train on
the questions are the same. You do machine learning on
all the data that you have, and it works. It
was pretty good. So so when the problem, when the
problem is contained in several on several dimensions, it's very
it's we're there, we can do it. The the hope
(22:28):
is the work on problem that I'm not contained are
really big and ugly and the changing all the time,
But we're not there yet.
Speaker 1 (22:37):
Gotcha. And you know, and you mentioned another technology earlier
when you were talking about robots you talked you mentioned
it's the same technology or similar technology for safety with
autonomous vehicles. What what's your take on like autonomous vehicles
on the roadway. Do you think that it's the future
is closed or do you think this is something that
(22:59):
is a lot of further are out then maybe those
that are pretty bullish on the autonomous vehicles out there,
whether it's trucks are card i.
Speaker 2 (23:08):
Would saw to some extent the future is here. You
go to San Francisco to Phoenix, you can have an
autonomous taxi. And in the South you can have autonomous trucks,
mostly still running in training, but some of them are
running limited commercial use, very limited and only in the
(23:29):
South when the roads are open, when you don't have snow,
when you don't have the cowpath that define Boston roads,
it's you know, it works there. There are, however, issues
with this. The issues are with every system that involves
human and robots, human and autonomy. There's a question of
(23:55):
public acceptance. Let me give you an extreme example. Today
modern aircraft seven eighty seven eight three fifty is actually
a drone. It can fly by itself. Not too many
people would fly at the aluminum tube, but thirty five
thousand feet across the Atlantic with nobody in the front.
(24:19):
But it's possible. There's the technologies there. You know, these
are really drones. I mean you see what drones do
and autonomous they do. They bomb, they go back, they
collect intelligence. It can be done, but public acceptance is
not there. So even though by the way that people
(24:41):
working on it, on getting we used to have five
people in the cockpit of an airliner and now that too,
and there's a lot of work and getting into one
and the other one is an AI helper and it's
and at the end of the day it will become autonomous.
So the question is how do people feel about a
(25:03):
large truck, weird looking track running at the one hundred
miles per hour next to them on the highway with
nobody in the front? Would they be you know, and
we will have accidents and who knows how it will
play people. No, this is too dangerous. We cannot have it.
(25:24):
It may come because because there'll be impacts on jobs.
You know, truck driver jobs are in thirty thirty some
states are the number one job, I mean the number
one job category. So it's lots of trucking jobs. So
(25:45):
going away, we'll see, we see what happens in the
port very hard to automate because of the union. So
and if you go to ports, as you must know,
you go to buy or Singapore or Rotterdam, and then
you go to an American port, then you get depressed.
(26:08):
It is what it is. I mean, I I understand,
but there's a lot of fame pushback and on automation.
It also it also takes a longer time than people think,
only a longer time than people think. Give you give
you one example. You know we used to have telephone
(26:30):
exchange operators. In eighteen ninety the AT and T came
up with an automatic exchange just based on numbers. You
didn't have to call susan in the exchange and ask
it to find where mister Glasgow is having lunch today
(26:53):
so we can talk to him. It was all automated. Now,
honestly think we lost some level of service, but we
got we were an efficiency. Turns out it took by
nineteen fifty we still had three hundred and fifty thousand
exchange operators. Only by nineteen eighty, which is nine decades later,
(27:19):
these people disappeared. So it takes time until technologies catch
not to be sure, I was working a lot faster
because it starts And the reason is because most of
today's technology starts with the consumers and they pick it
up because it's useful, it's interesting, it's cool, so they
(27:40):
pick it up and then it goes to companies.
Speaker 1 (27:43):
You mentioned like acceptance of autonomous vehicles. It was in
a way about cab in southern California a couple of
months ago, and it was definitely cool. My anxiety was
definitely higher than a normal taxi. But yeah, I can
I can see people, you know, not being very comfortable
(28:05):
going into a plane without at least two people in
the cockpit. And so you know, another thing I'm assuming
you guys are looking at and we're in extremely early
early innings and we don't really know how this is
going to play out. But all these tariffs that are
coming out of the US, I mean, are supply chain
is going to be redrawn or it's just just going
(28:25):
to be inflationary and we're all just going to deal
with higher prices and things are still going to be
you know, stickers are still going to be manufactured in
Vietnam and you know those those jobs aren't obviously coming
to the US. So so so how do you see
this playing out from a supply chain standpoint.
Speaker 2 (28:43):
Let me start my more general Okay, during the decades
of globalization, most of the population around the world benefit
with much higher start of living. Right, those are small
minority that extraly suffer. The small minority people who lost
manufacturing job in the industrial heartland the United States, for example,
(29:07):
ghost towns that the lost employment, which in some sense
led to the Fentanel crisis later on. So we had
most people, you know, enjoyed it, few people suffered. With
the current tarists, we are turning it upside down. Most
(29:29):
of people are going to suffer, a few people will benefit.
The people who benefit, for example, are domestic manufacturers who
will be able to have less competition, be able to
raise the price, increase their margins, and they will keep,
by the way, lobbying to continue this. So every economic
(29:51):
dislocation creates winner and loser, and the winners will keep
making it hard to dial it down. Now, let's talk
about what happened to supply chain. In some sense, I'm
less worried after COVID and Ukraine and the Middle East
and all this. In many ways, it's just another day
(30:14):
of the office. It is I was just with a
group of fifty supply chain officers and this was the
basically the the main response this is I was with
them last week when the when it was announced, nobody
was panicking. Some people immediately said, okay, even fifty tariffs
(30:40):
in now China is sixty percent or something. When you
enter everything up, it is still we cannot move to
the United States. The unfortunate thing that the administration does
seem to pay attention to our people, not that they
did right media to pay attention to is the fact
(31:02):
that it takes three years to be a manufacturing plant.
You cannot just come to deal. And then, by the way,
we don't have workers. You know, when DSMC build a
plant in a in Phoenix, in the Phoenix area, they
said that they're making the chips in the US will
cost fifty percent more. Why because we don't have enough
(31:23):
engineering talent and they have to train them. Every time
they're bringing engineers on, they have to train them. So
it's they said, it's not we Our biggest failure is
the education system. The US education system failed miserably. We
don't have an you know, MIT is not an example.
(31:45):
We get out of three hundred million people we get,
you know, where the MIT or Stanford or Harved we
get the top students even in STEM in science, technology, engineering, mathematics.
But the average level of the US is going down significantly.
And this is to me, this is the biggest fear.
(32:08):
And by the way, one of the main reasons that
we have this, you know, deficits in the balance of payment,
because otherwise we would have attracted more companies. But companies
know that there's no the tenant here is lucky and
the administration rather than invest mightily in education, I mean,
(32:32):
they try to root out DEI and the other things
from the education. Okay, fine, but make sure that they
focus on education or real education. I'm not sure they're
doing this, I mean the sleep service. But they're not
doing this because we are graduating kids who can hardly
(32:55):
read and write, forget being able to do real engineerings.
It's just or the number that we the number of
people who can do advanced engineering is way too small
because and it's not because of the university, it's because
of the input that comes into the universities. That's not
(33:18):
enough anyway. So that's a long answer to a short question.
Speaker 1 (33:24):
Yeah, I'm sure we can probably talk a long time
about it. Do you see any things that are any
industries or manufacturing processes that are easy to bring.
Speaker 2 (33:38):
Back to the United States?
Speaker 1 (33:39):
You know, we mentioned chips, we mentioned sneakers. You know,
I can't see you know people you know want of
those the manufacturing sneakers. And I don't think Americans are
going to pay five hundred dollars for a pair of Nikes.
Speaker 2 (33:54):
The problem with American liberates and American you know, oh,
environmental rules and all kinds of regulations, the prices will
be very high. It's so the question for a sneaker
company is do we pay sixty percent more and get
the sneakers vesty for three hundred dollars, or we bring
(34:15):
it to the United States, wait three years, build a plant,
don't find the workers for it, and then when we
finally automate it and get it to work with minimum
number of workers, it's five hundred dollars per sneaker and
we can't sell them anyway. So it's not clear. Look,
(34:36):
the main is it a problem with that you know,
automotive industry. They're being hammered because many of the parts
cross Canada and Mexico to United Several times, and each
time they'll be will be tariff on this. So it's
a it's getting ridiculous. The only way to do it is,
(34:58):
you know, but we should if we want to do something,
we should adopt value edit tacks. Just like most of
the world has value edit tax. It's a tax that
it is the least disruptive tax, but the US never
(35:18):
never went to it. It's much more efficient than sales
text that we charge here because it charged and it
also means very hard. It has so many other pluses.
It's it's very hard to use, you know, to do
stuff under the table because every stage of production must
report it and then get the money back. It's the
(35:40):
whole way that it's working along the supply chain.
Speaker 1 (35:45):
So can you, I guess can you explain that a
little bit? So why is the value added tacks the
bad tax.
Speaker 2 (35:52):
More fashion? So let's say you have a twenty percent tax.
It's most Europe is twenty percent tax. And let's say
if very simple thing, you get one supplier and one manufacturer.
The supplier pays twenty five percent tax. It comes to
the manufacturer. The manufacturers now has value added. He paid
(36:18):
twenty five percent only on the value added part, and
then on the part that comes from a supplier he
can get the money back. He applies to get the
money back, so everything has to be upfront in order
to get the money back. Now, finally, the one who
pays the full freight so to speak, is the consumer.
(36:40):
The consumer cannot get anything back. So for the consumer
it's the consumer paid twenty percent. But along the way
companies paid just for the value added actually that they
add add to it, and then for anything that the
supplier paid, the buyer can charge back at value. So
(37:05):
it's it kind of moves along the supply chain and
people are charging and charging and getting the money bag,
charging and getting part of the money bag. So it
requires significant you know management, but hey, we have a
big IRS, so's it can be done. It's a question
(37:26):
and now it's required on software and all this. But
it's a much more efficient way of taxation than sales
tax for example Celtics, it's not doesn't have any of
these qualities. But by the way, the administration makes a
mistake by looking at value added tax as if it's
(37:47):
distorting the market. First of all, the Europeans cannot get
rid of it because it finds most of the government.
It's a huge part of funding in a local government
in every state. They are in every country. I think
the only three or four countries in the world, and
the US is one of them. They do not have
value added tax. It is every economist will tell you
(38:10):
it's the most efficient way to tax, but we don't.
So we don't anyway. But it's so now they're the
h The administration took value added tax into account, thinking
that it's it's like a tariff. It's not, absolutely not.
(38:30):
But anyway, the administration is very simple calculation. But let
me add something just for the record, so to speak.
It is clear that some of what Trump is saying
that we were taking advantage of is correct. It's absolutely correct.
(38:50):
And by the way, every administration, the last I don't
know five six the administration was complaining about it. Right.
Trump is trying to do something about it now the
way he works in everything. If you look at how
its university, first of all, take a few hundred million
dollars and then let's talk, so you certainly get people's attention,
(39:13):
and just like what it does now is getting He
said that already fifty countries came back to renegotiate. Okay,
we'll see. We'll see how that goes. I hope they'll negotiate.
What can also happen that in a few months, when
the full branch of these teriffs will hit consumers, the
(39:36):
outcry will reach Congress, will reach the Senate, and we
may have, you know, congressional action. They'll simply outlaw some terrasts,
which is again a bad way of doing it without analysis,
without it. It should be the executive branch doing this,
but it should be done with a lot of analysis.
(39:58):
What is it should be? You know, which items should
be thanks at, what which how ter should should replace?
Should be a lot more sophisticated than just blend xpresent
on everything and just the simplical collation that the administration did.
(40:20):
But and again I am you know, Trump said he's
gonna do it. In some sense, he was elected based
on what he said. He never hid the fact that
he's in labor terist. I mean, he never hed it.
He was talking about it in the campaign trail. He
(40:41):
votes people voted him in. So it's a it's hard
to complain about it because we got what we voted for.
When I said we, I mean the majority of the
population got what they voted for. So here we are.
Speaker 1 (40:59):
So let's shift gears a little bit. So you're not
only UH an academia, You're you're also been involved with
a number of companies. Looking at your bio, you founded
or co founded five companies that were acquired by other companies.
So that's that's a pretty darn good track record. Can
you talk about the kind of like the lessons learned
(41:21):
from founding companies UH in supply chain and sure, you know,
maybe maybe talk to those experiences a little bit.
Speaker 2 (41:31):
Sure, So some companies were software companies and some were
maybe one of the more interesting companies was called Logicorp.
It's a I was consulting at the time to Rockwell
International and befriended the vice president logistic the trock Oil
(41:52):
Roco was selling truck parts UH trucking company everything under
the truck brank, breaks, excels, what have you. At that time,
it was deregulation of the transportation industry came into being,
and we just look at each other and say, okay,
we can now serve every route without getting the the
(42:16):
government to prove it. So let's do what's called a
market test auction. We did we got the right to
do under Rockwell still got the right to do auction,
and the results were astounding prices when half became half
the price or you know, forty percent better and with
(42:38):
the same carriers as before. And then so we look
at it and said, okay, there's a business here. So
we convinced Rockwall to start the separate unit and then
basically we bought it from Rockwall and this started growing
like weeds in We ran it three years under rock Oil,
(43:02):
which it didn't grow much because the carriers were afraid
that we have biasing. We did these auctions, but the
carriers were afraid of buiasing. The results for rockcol for
a Rockcay customers, and we said we must get it out.
We found the an angel. We took it out and
it grew from it was about forty million when it
(43:22):
took it out of rock Oil. It grew within three
years to six hundred million, and it grew like weeds.
We were an example of outsourcing. We had just logistical engineer.
Everything else was every other corporate function was outsourced. We
tried to tell people they can outsource the traffic manage
transportation management to us and to give you an idea.
(43:47):
There was two quarters when we said we cannot take
any new business, and we still doubled just existing business.
Speaker 1 (43:55):
It's good problem to have.
Speaker 2 (43:56):
Well, the wheels are coming off the wagon in terms
of very hard thisason before the internet was it not
easy to manage a business like this? I mean the
papers and stuff and showed with mis solid to Writer.
It's now a three billion you know dollars annually part
of Writer. You know. The chairman of Writer called me,
(44:19):
he retired. He said, you know what we did one
hundred and fifty six acquisition. This was the best one
that we ever did. So so it was and I
got on the Writer board and stuff like this. But
let's a so this was one company. We had a
(44:39):
software company that did that started by doing software for
tracking for routing, scheduling, assignment of drivers and TL truck
load load load matching, a load planning for LTL and
then change it or grew it to do procurement basically
for for shippers. And this company was sold to Manhattan Associates.
(45:07):
H There were lots of stories, you know, along ever,
but I tell you what I always stayed also my
t I usually well what I did is started a company,
spent a year at the company and then just became
chairman just you know, found management team. And then because
I I could not stay them T and have b
(45:29):
a line officer because I understand it, I see it
in front of the class. You don't want to get
a phone call certainly and run to the company. Okay,
that's an m T rule that made that's reasonable. So
I work rule and either took of absence of sabbatical
whatever it was, and start a company and then acquire
(45:53):
They all acquired just putunistically, they're all all acquired by
by the company. But what did you ask for? What
I learned? I learned that that became an amazingly good teacher.
I talked about my own experience and there's nothing you know.
I worked directly with companies, with a tracking company, ships
(46:15):
with carrier. We developed one of the first TMS's transportation
management systems. So it's a It was one thing to
talk to students about stuff that I had experienced with
that they're not talking about what they reading the book,
what they do a case study. I developed case studies
(46:37):
that according that I know everything about it because I
lived it. So I'm also, of course consulting. Now the
last company was so years ago. I'm consulting to lots
of budding entrepreneurs because after that I joined some venture
capital boards, but not anymore. But a lot of people
(47:01):
they might want to start companies, they want to study this.
I'm the informal consult that to those people stry to
try to start companies.
Speaker 1 (47:09):
Well, well that must be extremely rewarding. You know, we're
coming up towards the end of our time and I
just always like to ask this, and maybe I'll phrase
the question all differently. Since you know, giving that your
mi T, do you have like a required reading for
all your students. It might not be, it might not
be a textbook, but it's something that you know, would
(47:31):
prepare them for a career in supply chain.
Speaker 2 (47:37):
That's very you know, there are some textbooks in supply chain.
My own books are not textbooks. There are business books
and look for example, my book on name it's called
Sustainability in Supply Chain. It's called the Bus and Green.
(48:02):
When to apply Sustainability in supply gen and when not
to is used by several schools as textbook, including the Hours.
Now we don't have a class on separate issues so
we usually don't assign books. We are command books, so
books on specific specific subject. On AI, we don't have
(48:27):
a book because it is we just talk about the
latest and every semester the the lectures change and the
the work the homework changes because we students. So for example,
our students now have to not Python, which is the
(48:47):
language that you know, the computer language used mostly for APIs,
and but every year they have to do more or
different because the abilities are a different, you know, the
new software coming, the latest version of GPT of you know, Copilot, Jemini,
(49:12):
whatever it is, it's getting better. I can do more
and so you work with it.
Speaker 1 (49:19):
Well you sa I really want to thank you for
your time. This was a really great conversation and thanks
for your insights.
Speaker 2 (49:26):
All right, thank you very much for having me appreciate it.
Speaker 1 (49:29):
It's my pleasure and hope to have you back again
some time. And I want to thank you for tuning in.
If you liked the episode, please subscribe and leave a review.
We've lined up a number of great guests for the podcast,
so please check back to your conversations with c suite executives, shippers, regulators,
and decision makers within the freight markets. Also, if you
have any ideas for a future episodes, just hit me
(49:51):
up on the Bloomberg terminal or on Twitter at logistics Late.
Thanks so much and take care everyone,
Host
Lee Klaskow
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com